Tuesday, November 2, 2004
Part III
Department of Housing and Urban Development 24 CFR Part 81 HUD’s Housing Goals for the Federal National Mortgage Association (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac) for the Years 2005–2008 and Amendments to HUD’s Regulation of Fannie Mae and Freddie Mac; Final Rule
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00001 Fmt 4717 Sfmt 4717 C:\02NOR2.SGM 02NOR2 63580 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
DEPARTMENT OF HOUSING AND Government Sponsored Enterprises/ Fannie Mae and Freddie Mac engage URBAN DEVELOPMENT RESPA, Office of the General Counsel, in two principal businesses: (1) Room 9262, telephone 202–708–3137. Purchasing and otherwise investing in 24 CFR Part 81 The address for all of these persons is residential mortgages, and (2) [Docket No. FR–4790–F–03] Department of Housing and Urban guaranteeing securities backed by Development, 451 Seventh Street, SW., residential mortgages. As a result of RIN 2501–AC92 Washington, DC, 20410. Persons with their status as GSEs, Fannie Mae and hearing and speech impairments may Freddie Mac receive significant explicit HUD’s Housing Goals for the Federal access the phone numbers via TTY by National Mortgage Association (Fannie benefits that are not enjoyed by fully calling the Federal Information Relay private shareholder-owned corporations Mae) and the Federal Home Loan Service at (800) 877–8399. Mortgage Corporation (Freddie Mac) in the mortgage market. These benefits SUPPLEMENTARY INFORMATION: for the Years 2005–2008 and include: Amendments to HUD’s Regulation of I. General • Conditional access to a $2.25 billion Fannie Mae and Freddie Mac A. Authority line of credit from the U.S. Treasury (see section 306(c)(2) of the Freddie Mac Act AGENCY: Office of the Assistant HUD’s authority to regulate the GSEs and section 304(c) of the Fannie Mae Secretary for Housing—Federal Housing is established under: Charter Act); Commissioner, HUD. (1) The Federal National Mortgage • ACTION: Final rule. Association Charter Act (‘‘Fannie Mae Exemption from the securities Charter Act’’), which is Title III of the registration requirements of the U.S. SUMMARY: Through this final rule, the National Housing Act, section 301 et Securities and Exchange Commission Department of Housing and Urban seq. (12 U.S.C. 1716 et seq.); and the State securities regulatory Development establishes new housing (2) The Federal Home Loan Mortgage agencies (see section 306(g) of the goal levels for the Federal National Corporation Act (‘‘Freddie Mac Act’’), Freddie Mac Act and section 304(d) of Mortgage Association (Fannie Mae) and which is Title III of the Emergency the Fannie Mae Charter Act); 1 and the Federal Home Loan Mortgage Home Finance Act of 1970, section 301 • Exemption from all State and local Corporation (Freddie Mac) (collectively, et seq. (12 U.S.C. 1451 et seq.); taxes except property taxes (see section the government sponsored enterprises, (3) FHEFSSA, enacted as Title XIII of 303(e) of the Freddie Mac Act and or GSEs) for calendar years 2005 the Housing and Community section 309(c)(2) of the Fannie Mae through 2008. The new housing goal Development Act of 1992 (Pub. L. 102– Charter Act). levels are established in accordance 550, approved October 28, 1992) (12 with the Federal Housing Enterprises U.S.C. 4501–4641); and While the securities that the GSEs Financial Safety and Soundness Act of (4) Section 7(d) of the Department of guarantee, and the debt instruments 1992 (FHEFSSA) and govern the Housing and Urban Development Act they issue, are explicitly not backed by purchase by Fannie Mae and Freddie (42 U.S.C. 3535(d)). the full faith and credit of the United Mac of mortgages financing low- and States, and nothing in this rule should B. Background: Fannie Mae and Freddie be construed otherwise, such securities moderate-income housing, special Mac affordable housing, and housing in and instruments trade at yields only a central cities, rural areas and other Fannie Mae and Freddie Mac were few basis points over those of U.S. underserved areas. This rule also chartered by the Congress as GSEs. Treasury securities with comparable establishes new subgoals for the GSEs’ Pursuant to section 301 of the Fannie terms. These securities also offer yields acquisitions of home purchase loans Mae Charter Act (12 U.S.C. 1716) and lower than those for securities issued by that qualify for each of the housing section 301(b) of the Freddie Mac Act fully private firms that are more highly goals. The final rule also establishes a (12 U.S.C. 1451), the GSEs were capitalized but otherwise comparable. new regulatory section relating to GSE chartered expressly to: In addition, the market does not require data integrity, amends and adds certain (1) Provide stability in the secondary that individual GSE securities be rated definitions, provides a method for market for residential mortgages; by a national rating agency. (2) Respond appropriately to the imputing the distribution of GSE- Consequently, the GSEs are able to fund private capital market; their operations at lower cost than other purchased mortgages that lack income (3) Provide ongoing assistance to the private firms with similar financial data, prohibits goals credit for purchases secondary market for residential characteristics. In a recent report, the of loans in transactions with an option mortgages (including activities relating Congressional Budget Office (CBO) to dissolve the purchase in less than one to mortgages on housing for low- and year, and makes a technical change to moderate-income families involving a estimated that this funding advantage the counting rules to clarify HUD’s rules reasonable economic return that may be for the year 2003 resulted in a $19.6 on double counting of loans. less than the return earned on other billion annual combined subsidy for EFFECTIVE DATE: January 1, 2005. activities) by increasing the liquidity of both GSEs. Of this amount, CBO FOR FURTHER INFORMATION CONTACT: mortgage investments and improving estimated that the GSEs retained about Sandra Fostek, Director, Office of the distribution of investment capital $6.2 billion, or approximately one-third Government Sponsored Enterprises, available for residential mortgage of the subsidy, for their officers and Office of Housing, Room 3150, financing; and telephone 202–708–2224. For questions (4) Promote access to mortgage credit 1 Fannie Mae and Freddie Mac have both announced their intention voluntarily to register on data or methodology, contact John L. throughout the nation (including central their common stock with the Securities and Gardner, Director, Financial Institutions cities, rural areas, and other Exchange Commission (SEC) under section 12(g) of Regulation Division, Office of Policy underserved areas) by increasing the the Securities Exchange Act of 1934. Fannie Mae’s Development and Research, Room 8212, liquidity of mortgage investments and registration became effective March 31, 2003. Freddie Mac has stated that it will complete the telephone (202) 708–1464. For legal improving the distribution of process of voluntarily registering its common stock questions, contact Paul S. Ceja, Deputy investment capital available for once it resumes timely reporting of its financial Assistant General Counsel for residential mortgage financing. results.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00002 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63581
shareholders, while the remainder loans that qualify for each of the 2008, to achieve the ultimate objective accrued to borrowers.2 housing goals. for the GSEs to lead the market under In return for the public benefits they In addition to soliciting public a range of foreseeable economic receive, Congress has mandated in the comments on the proposed goal levels circumstances by 2008. GSEs’ Charter Acts that the GSEs carry and new subgoals, the rule solicited The staged increases established by out public purposes not required of public comments on several other issues this rule, are consistent with the other private sector entities in the related to the housing goals, including: statutory requirement that HUD housing finance industry. These (1) Provisions relating to GSE data consider the past performance of the statutory mandates obligate the GSEs to integrity, such as verification, GSEs in setting the Housing Goals. work to ensure that everyone in the certification, treatment of errors, Staged annual increases in the Goals nation has a reasonable opportunity to omissions or discrepancies, and other will provide the GSEs with the enjoy access to the mortgage financing enforcement authority; (2) amended opportunity to adjust their business benefits resulting from the activities of definitions of ‘‘underserved area,’’ models, so as to meet the required 2008 these enterprises. ‘‘metropolitan area’’ and ‘‘minority,’’ levels without compromising other and a new definition of the term ‘‘home business objectives and requirements. With respect to these public purposes, purchase mortgage’’; (3) a method for The Department believes that the Congress does not simply expect the imputing the distribution of GSE- Home Purchase Subgoals established by GSEs to strive toward achievement of purchased mortgages that lack income this final rule are necessary and these purposes but rather to ‘‘lead the data; and (4) other changes related to the warranted. Increasing homeownership mortgage finance industry’’ and to GSEs’ bulk purchases of seasoned loans. is a national priority. The past average ‘‘ensure that citizens throughout the More detailed information about HUD’s performance of the GSEs in the home country enjoy access to the public proposals can be found in the preamble purchase market has been below market benefits provided by these federally to HUD’s May 3, 2004, proposed rule. levels. As further discussed below, the related entities.’’ (See S. Rep. No. 102– GSEs must apply greater efforts to 282, at 34 (1992).) E. This Final Rule—Overview increasing homeownership for low- and C. Statutory and Regulatory Background Under this 2004 rulemaking, the moderate-income families, families Department is setting new, higher levels living in underserved areas, and very- The statutory and regulatory for the Housing Goals, accompanied by low income families and low-income background applicable to the chartering subgoals under each of the Housing families living in low-income areas. The of Fannie Mae and Freddie Mac and Goals for purchases of home purchase addition of Home Purchase Subgoals to HUD’s regulatory authority over these mortgages (i.e., excluding refinance the regulatory structure will serve to two GSEs were set out in detail in the mortgages) on owner-occupied better focus the GSEs’ efforts in a clear preamble to HUD’s proposed rule properties in metropolitan areas. (The and transparent manner. The Home published on May 3, 2004 (69 FR subgoals are referred to in this rule as Purchase Subgoals will better allow the 24228). Therefore, this background the ‘‘Home Purchase Subgoals.’’) government and public alike to monitor information is not repeated here in the The Department’s purpose in setting the GSEs’ efforts in meeting the nation’s preamble to this final rule. Interested higher Housing Goals and in homeownership needs. The increases in members of the public should refer to establishing new Home Purchase the levels of the Housing Goals, and the Section I.A. of the preamble to the Subgoals in this final rule is to addition of the new Home Purchase proposed rule at pages 69 FR 24228 encourage the GSEs to facilitate greater Subgoals, are predicated upon the through 69 FR 24230 for this financing and homeownership Department’s recognition that the GSEs information. opportunities for families and not only have the ability to achieve D. The Proposed Rule neighborhoods targeted by the Housing these Housing Goals and Subgoals but, Goals. The final rule establishes levels also, that they are fully consistent with On May 3, 2004, HUD published a of the Housing Goals that will bring the the statutory factors established under proposed rule setting forth new housing GSEs to a position of market leadership FHEFSSA. In addition, this rule is goal levels for Fannie Mae and Freddie in a range of foreseeable economic supported by the Department’s Mac. (See 69 FR 24228.) HUD’s rule circumstances related to the future comprehensive analyses of the size of proposed to increase the level of the course of interest rates and consequent the mortgage market, the opportunities housing goals (‘‘Housing Goals’’) for the fluctuations in origination rates on available to the GSEs, America’s unmet purchase by Fannie Mae and Freddie home purchase and refinance housing needs, and identified credit Mac of mortgages financing low- and mortgages—both multifamily and gaps. moderate-income housing, special single-family. In addition to the establishment of affordable housing, and housing in For each goal, HUD has projected higher Housing Goals for the years 2005 central cities, rural areas, and other goal-qualifying percentages of mortgage through 2008, and the establishment of underserved areas. The rule also originations in terms of ranges that Home Purchase Subgoals, specific proposed to establish new subgoals for cover a variety of economic scenarios. changes included in the final rule from the GSEs’ acquisitions of home purchase The objective of HUD’s Housing Goals is the provisions included in the May 3, to bring the GSEs’ performance to the 2004, proposed rule are as follows: 2 ‘‘Updated Estimates of the Subsidies to the upper end of HUD’s market range (1) The final rule expands the existing Housing GSEs,’’ attachment to a letter from Douglas estimate for each goal, consistent with provisions to permit the GSEs to impute Holtz-Eakin, Director, Congressional Budget Office, the requirement in FHEFSSA that HUD incomes or rents when data are missing to the Honorable Richard C. Shelby, Chairman, Committee on Banking, Houseing, and Urban should consider the GSEs’ ability to lead for some purchases, addressing the Affairs, United States Senate, April 8, 2004. A the market for each goal. market’s expanding use of low related recent study is Wayne Passmore, ‘‘The GSE To enable the GSEs to achieve this documentation mortgages; Implicit subsidy and Value of Government leadership, the Department has (2) The final rule provides that goals Ambiguity,’’ Board of Governors of the Federal Reserve System. Finance and Economics Discussion established staged increases in Housing credit is available for purchases of loans Series, FEDS Working Paper 2003–64, December Goal levels for 2005, which will in transactions involving seller 2003. increase further, year-by-year through dissolution options, such as repurchase
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00003 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63582 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
agreements, only when the option Home Purchase Subgoals to the and Freddie Mac. Neither GSE was provides for a minimum one-year regulatory structure provides an supportive of the higher goal levels lockout period; additional means of encouraging the proposed for 2005–2008, nor did either (3) The final rule clarifies the GSEs’ affordable housing activities to support the creation of HUD’s proposed proposed provisions regarding HUD’s address identified, persistent credit Home Purchase Subgoals. The GSEs procedures for correcting errors, needs while leaving to the GSEs the stated, among other comments that they omissions and discrepancies in current specific approaches used to meet these made on the rule, that the effect of many year-end data and in remedying material needs. goals and subgoals would be overstatements of housing goals (3) Discrimination in lending micromanagement of the GSEs. With performance for prior years; continues to limit access to credit for their comments, the GSEs provided (4) The final rule changes the scope of purchasing homes by racial and ethnic several appendices that provided the proposed certification statement that minorities. Troublesome gaps in alternative analyses of data and the GSEs must provide to make it closer homeownership remain for minorities questioned the Department’s to the certification used by the Office of even after record growth in affordable methodology in determining market Federal Housing Enterprise Oversight lending and homeownership during the share for the three affordable housing (OFHEO), the GSEs’ financial safety and nineties. Studies indicate that, over the goals, a key component for establishing soundness regulator; and next few years, minorities will account the appropriate level of the housing (5) The final rule makes a technical for a growing share of the families goals and the subgoals. correction to the special counting rules seeking to buy their first home. HUD’s The GSEs did not object to HUD’s prohibiting double counting of GSE analyses indicate, however, that Fannie special affordable multifamily subgoal purchases of seasoned mortgages toward Mae and Freddie Mac account for a levels for 2005–2008, but other the housing goals. disproportionately small share of the commenters (mostly public advocacy In developing these regulations, the minority first-time homebuyer market. groups) recommended that HUD Department was guided by, and re- The GSEs have a responsibility to increase the levels of these subgoals. affirms, the following principles promote access to capital for minorities In addition to the GSEs, the established in the Housing Goals 1995 and others who are seeking their first commenters included national and final rule (published on December 1, homes, and to demonstrate the benefits regional housing industry organizations, 1995 at 60 FR 1846): of such lending to industry and nonprofit organizations, alliances, (1) The GSEs should fulfill councils, and advocacy organizations FHEFSSA’s intent that they lead the borrowers alike. The GSEs also have an integral role in eliminating predatory involved in housing or housing issues, industry in ensuring that access to lenders, academic researchers, Members mortgage credit is made available for mortgage lending practices. (4) In addition to the GSEs’ purchases of Congress, state and local government very low-, low- and moderate-income officials, and two individuals. families and residents of underserved of single-family home mortgages, the GSEs also must continue to assist in the In large measure, except for several areas. HUD recognizes that, to lead the nonprofit organizations and public creation of an active secondary market mortgage industry over time, the GSEs advocacy groups that favored higher for mortgages on multifamily rental will have to stretch to reach certain goals, the majority of commenters were housing. Affordable rental housing is Housing Goals and to close gaps not supportive of HUD’s proposed goals, essential for those families who cannot between the secondary mortgage market especially in the outer years when the afford to become, or who choose not to and the primary mortgage market for goal levels would reach their highest become, homeowners. For this reason, various categories of loans. This levels. A particular concern cited by a the GSEs must assist in making capital recognition is consistent with the number of commenters was the available to assure the continued Congressional directive that ‘‘the potential for adverse impact on middle- development of single-family and enterprises will need to stretch their income borrowers, particularly higher multifamily rental housing. efforts to achieve’’ the goals. (See S. interest rates and fees. Another concern Rep. No. 102–282, at 35 (1992).) II. Discussion of Public Comments raised by the commenters was the (2) The Department’s role as a possibility of unintended consequences A. Overview of Public Comments regulator is to set broad performance for the industry. Many commenters, standards for the GSEs through the At the close of the public comment including the GSEs, urged HUD to Housing Goals, but not to dictate the period on July 16, 2004, which was exclude all single-family refinances specific products or delivery extended an additional two weeks from the calculation of the goals. mechanisms the GSEs will use to beyond the original public comment The Department received fewer achieve a Housing Goal. Regulating two deadline of July 2, 2004, HUD had comments that addressed other exceedingly large financial enterprises received 302 comments, which are in proposals in the rule, such as those in a dynamic market requires that HUD HUD’s docket file for this rule. In regarding data integrity, large-scale provide the GSEs with sufficient addition to the public comments transactions involving seasoned loans, latitude to use their innovative received on the rule, during the public the treatment of missing income data, capacities to determine how best to comment period, HUD met with and modifying the definition of rural develop products to carry out their representatives of several organizations, underserved areas. For those respective missions. HUD’s regulations including Fannie Mae and Freddie Mac, commenters who submitted comments are intended to allow the GSEs the to accommodate oral presentation of on these proposals, the reactions were flexibility to respond quickly to market concerns about the rule. HUD’s docket generally mixed. opportunities. At the same time, the file for this rule contains information on With respect to HUD’s proposals for Department must ensure that the GSEs’ the dates of these meetings, the new data integrity provisions, the strategies address national credit needs, attendees, and the subject discussed. majority of those who commented on especially as they relate to housing for Of the public comments received on the new data integrity proposals were low- and moderate-income families and the proposed rule, the most detailed generally supportive of the concept and housing located in underserved comments were those submitted by the acknowledged the need for some sort of geographical areas. The addition of two directly affected GSEs, Fannie Mae data verification process. However, two
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00004 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63583
industry-related commenters expressed public comments received and incorporates a determination that concern about the potential for HUD’s acknowledges the value of all of the mortgage purchases will count toward proposals to result in increased comments submitted in response to the the Underserved Areas Housing Goal reporting burdens for lenders. The proposed rule. where such purchases finance GSEs’ comments also reflected several properties that are located in B. Subpart A—General concerns about the data integrity underserved census tracts. These are provisions, mainly with respect to In the May 3, 2004, rule, HUD defined as census tracts where either: definitions, procedures, and proposed to add a definition of ‘‘home (1) the median income in the tract does enforcement. purchase mortgage’’ in connection with not exceed 95 percent of the greater of The GSEs favored generous proxy its proposal to specify Home Purchase the median income for the non- provisions for the treatment of missing Subgoals under each of the three metropolitan portions of the state or the income data and submitted several Housing Goals, to revise the definitions median income of the non-metropolitan suggestions. The majority of of ‘‘metropolitan area’’ and ‘‘minority’’ portions of the nation as a whole; or (2) commenters on this issue, consisting to conform HUD’s regulations to minorities comprise at least 30 percent chiefly of nonprofit and advocacy changes in data collection practices of the residents and the median income organizations, opposed using proxies, made by the Office of Management and in the tract does not exceed 120 percent and several favored an outright ban on Budget (OMB), and to modify the of the greater of the median incomes for purchasing ‘‘no income’’ subprime current definition of ‘‘underserved area’’ the non-metropolitan portions of the mortgages. with respect to the delineation of state or of the nation as a whole. With regard to large-scale transactions underserved portions of non- HUD originally adopted its current involving seasoned loans, the GSEs metropolitan areas. county-based definition for targeting commented that they should receive 1. Home Purchase Mortgage GSE purchases to underserved non- housing goals credit and that no change metropolitan areas primarily based on in HUD’s current definition of HUD proposed to insert a definition of information that rural lenders did not ‘‘mortgage purchase’’ was warranted. ‘‘home purchase mortgage’’ for purposes perceive their market areas in terms of However, a group of industry-related of specifying the Home Purchase census tracts, but rather, in terms of organizations opposed providing goals Mortgage Subgoals. Since no comments counties. A further concern was an credit for seasoned loans, as did several bearing directly on this definition were apparent lack of reliability of geocoding advocacy organizations. Commenters received and the Department has software applied to non-metropolitan offered no alternative definitions for retained the subgoal concept in this areas. ‘‘mortgage purchase’’ in HUD’s final rule, the definition is adopted. regulations. Thirteen commenters endorsed HUD’s All but one commenter who 2. Metropolitan Area proposed change in definition, addressed the issue of HUD’s rural HUD proposed to alter the definition observing that the change would underserved area definition favored of ‘‘metropolitan area’’ to reflect a produce more precise targeting and changing this definition to one that is change in the definition of improved service toward underserved census tract-based, rather than county- ‘‘metropolitan area’’ recently segments of the market within counties. based. Those commenters favoring promulgated by OMB, in which the One banking trade association conversion to a tract-based definition concept of ‘‘Primary Metropolitan advocated continuation of a county- believed that county-level data do not Statistical Area’’ was removed. No based definition, stating that because show disparities in service that the comments were received on this the business perspective of community GSEs should address. The dissenting proposed change; accordingly, it is banks in rural areas is geared toward commenter felt that lenders serving adopted. entire counties, there would be costs rural areas would face operational associated with monitoring the tract difficulties and expenses in shifting to 3. Minority location of loans, and therefore, a tract-based orientation. HUD proposed to alter the definition marketing toward borrowers at the tract In addition to comments on its of ‘‘minority’’ to reflect changes in level would be difficult. proposals related to housing goals, HUD standards for the classification of federal Recent research summarized in received other comments on subjects data on race and ethnicity previously Appendix B to this rule indicates that a pertaining to HUD’s regulatory authority promulgated by OMB and implemented tract-based system will improve the over the GSEs but which were not in the 2000 census and in data extent to which the underserved area related to the rule’s proposals on collection under the Home Mortgage definition distinguishes areas by key housing goals (for example, comments Disclosure Act in 2004. No comments socioeconomic and demographic on new program authority, monitoring were received on this proposed change; characteristics such as median family and reporting procedures, and public accordingly, it is adopted. income, poverty, unemployment, school access to GSE mortgage data). Because dropout rates, and minority these comments raised issues outside 4. Underserved Area populations. Under a tract-based the scope of the May 3, 2004, proposed HUD proposed to alter the definition definition underserved areas stand out rule, they are not addressed in this final of ‘‘underserved area’’ to provide for the more as areas of lower income and low rule. specification of underserved areas economic activity and as having A discussion of the general and outside of metropolitan areas at the somewhat larger minority population specific comments on the rule, as well census tract level rather than at the proportions. A tract-based definition as HUD’s responses to these comments, county level. will also improve the targeting of the follows in subsequent sections in this For properties in non-metropolitan goal to areas with relatively greater preamble, as well as in the Appendices (rural) areas, mortgage purchases have housing needs. Based on these findings, to this Final Rule. While comments are counted toward the Underserved Areas which are detailed in Appendix B to summarized, not all the comments are Housing Goal where such purchases this rule, HUD is adopting a re- addressed explicitly in this preamble. finance properties that are located in specification of underserved areas HUD is appreciative of the full range of underserved counties. This final rule within non-metropolitan (rural) areas to
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00005 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63584 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
be based on census tracts rather than help offset declines in the demand for completions, and affordability—reached counties. housing caused by the aging of the levels that were among the highest in population. the past two decades. C. Subpart B—Housing Goals The continued influx of immigrants The Administration’s forecast for real 1. Overview will increase the demand for rental GDP growth is 3.7 percent for 2005 and A substantial majority of the housing, while those who immigrated 3.1–3.4 percent in 2006–2009, while comments received criticized HUD’s during the 1980s and 1990s will be in CBO projects that real GDP will grow at proposed levels of the housing goals on the market for homeownership. an average rate of 4.1 percent in 2005 the basis that they would be difficult for Immigrants and minorities—who and annual rates of 2.9–3.2 percent in 3 the GSEs to achieve, particularly in accounted for nearly 40 percent of the 2006 through 2009. The growth in the nation’s homeownership periods of high refinance activity when Administration projects the 10-year rate over the past five years—will be higher-income borrowers comprise a Treasury rate to average 5.1 percent in responsible for almost two-thirds of the relatively high proportion of mortgage 2005 and 5.4–5.8 percent between 2006 growth in the number of new borrowers. Several types of adverse and 2009 compared to its average of 4.6 households over the next 10 years. consequences of such high goals were percent in 2002 and 4.0 percent in 2003. Non-traditional households have Standard & Poor’s expects housing forecast, including diminution of become more important, as overall availability of mortgage credit to some starts to average 1.8 million units in household formation rates have slowed. 2004–2005. Fannie Mae projects sectors of the mortgage market, With later marriages, divorce, and non- unfavorable effects on neighborhood existing home sales for 2004 at 6.1 traditional living arrangements, the million units, and for 2005 at 5.8 housing quality, and other adverse fastest growing household groups are effects discussed below. This section of million, compared to their record level single-parent and single-person in 2003 of 6 million units. the final rule reviews the statutory households. By 2025, non-family factors the Department must consider in households will make up one-third of (iii) Mortgage Market Conditions setting the level of the housing goals all households. The role of traditional Low interest rates and record levels of and the Department’s determinations 25-to-34 year-old married, first-time refinancing caused mortgage with regard to the levels of each of the homebuyers in the housing market will originations to soar from $2.0 trillion in housing goals as well as the proposed be smaller in the current decade due to 2001 to $2.6 trillion in 2002 and around Home Purchase Subgoals. the aging of the population. Between $3.8 trillion in 2003. The Mortgage 2. Statutorily Required Factors in 2000 and 2025, the Census Bureau Bankers Association projects that Setting the Levels of the Housing Goals projects that the largest growth in mortgage originations will drop to $2.7 and Subgoals households will occur among trillion in 2004 and $1.8 trillion in 2005, householders who are age 65 and older. The Housing Goals and Home as refinancing returns to more normal As these demographic factors play 4 Purchase Subgoals being implemented levels. out, the overall effect on housing The volume of home purchase by this final rule were established demand will likely be continued growth mortgages was $910 billion to $1.1 following consideration of the six and an increasingly diverse household trillion between 1999 and 2001 before factors required by statute to be population from which to draw new jumping to $1.2 trillion in 2002 and $1.3 considered in establishing goal levels renters and homeowners. A greater trillion in 2003. As with housing starts, and establishing subgoals. A summary diversity in the housing market will, in the home purchase origination market is of HUD’s findings relative to each of the turn, require greater adaptation by the expected to exhibit sustained growth. six statutory factors follows. More primary and secondary mortgage detailed discussion of these points is markets. b. National Housing Needs included in Appendices A, B, and C to (i) Affordability Problems this rule. (ii) Economic and Housing Conditions Data from the 2000 Census and the While most other sectors of the a. Demographic, Economic, and Housing American Housing Survey demonstrate economy were weak or declining during Conditions that there are substantial housing needs 2001 and 2002, the housing sector among low- and moderate-income (i) Demographic Trends showed remarkable strength. The families. Many of these households are Changing population demographics housing market continued at a record burdened by high homeownership costs will result in a need for the primary and pace during 2003. or rent payments and, consequently, are secondary mortgage markets to meet In 2002, the U.S. economy moved into facing serious housing affordability nontraditional credit needs, respond to recovery, with real Gross Domestic problems. There is evidence of diverse housing preferences and Product (GDP) growing 2.2 percent, persistent housing problems for overcome information and other barriers although measures of unemployment Americans with the lowest incomes. that many immigrants and minorities continued to rise before declining again Since 1977 the percentage of U.S. currently face. in 2003. In October 2002, the average The U.S. Census Bureau has projected 30-year home mortgage interest rate households with worst case needs has that the U.S. population will grow by an slipped below 6 percent for the first hovered around five percent, with the average of 2.5 million persons per year time since the mid-1960s. Favorable worst year being 1983 (6.03 percent) and between 2000 and 2025, resulting in financing conditions and solid increases the best being 1999 (4.72 percent). The about 1.2 million new households per in house prices were the key supports 3 Fiscal Year 2005 Budget of the U.S. year. The aging of the baby-boom to record housing markets during both Government: Mid-Session review (July 30, 2004). generation and the entry of the baby- 2002 and 2003. By the end of 2003, the Office of Management and Budget, also posted at bust generation into prime home-buying industry had set new records in single- http://www.whitehouse.gov/omb. The Budget and age will have a dampening effect on family home permits, new home sales, Economic Outlook: An Update, Washington, Congressional Budget Office, September 2004, also housing demand. Growing housing existing home sales, low interest rates, posted on http://www.cbo.gov. demand from minorities, immigrants and rates of homeownership. Other 4 Mortgage Bankers Association of America, MBA and non-traditional homebuyers will indicators—total permits, starts, Mortgage Finance Forecast, September 17, 2004.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00006 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63585
proportion in 2001 was 4.77 percent, conclude that African Americans and Between 1993 and 2002, conventional which is not significantly different from Hispanics risk unequal treatment when loans to low-income and minority the 1999 figure. HUD’s analysis of they visit mainstream mortgage lenders. families increased at much faster rates American Housing Survey data reveals Studies reveal higher mortgage denial than loans to upper-income and non- that, in 2001, 5.1 million unassisted rates for African Americans and minority families. Conventional home very-low-income renter households had Hispanics, even after controlling for purchase originations to African ‘‘worst case’’ housing needs, defined as applicant income and a host of Americans more than doubled between housing costs greater than 50 percent of underwriting characteristics, such as the 1993 and 2002, and those to Hispanic household income or severely credit record of the applicant.6 borrowers more than tripled during this inadequate housing. Among these However, substantial progress has been period. Home loans to low-income households, 90 percent had a severe made since 1989. borrowers and to low-income and high- rent burden, 6 percent lived in severely The existence of substantial minority census tracts also more than inadequate housing, and 4 percent neighborhood disparities in doubled during this period. suffered from both problems. Among the homeownership and mortgage credit is Thus, the 1990s and the early part of 34 million renters in all income also well documented for metropolitan the current decade have seen the categories, 6.3 million (19 percent) had areas. HUD’s analysis of Home Mortgage development of a strong affordable a severe rent burden and over one Disclosure Act (HMDA) data shows that lending market. Homeownership million renters (3 percent) lived in mortgage credit flows in metropolitan statistics show similar trends. After housing that was severely inadequate. areas are substantially lower in high- declining during the 1980s, the minority and low-income homeownership rate has increased (ii) Disparities in Housing and Mortgage every year since 1994, reaching a record Markets neighborhoods and that mortgage denial rates are much higher for residents of mark of 69.2 percent in the second Despite the strong growth in these neighborhoods. Studies have also quarter of 2004. affordable lending over the past ten documented that mainstream lenders The number of households owning years, there are families who are not often do not operate in inner-city their own home increased by 13.3 being adequately served by the nation’s minority neighborhoods, leaving their million between 1994 and 2003. Gains housing and mortgage markets. Serious residents with only high-cost lenders as in homeownership rates during the racial and income disparities remain. options. Too often, residents of these period of 1994 to 2003 have been The homeownership rate for minorities same neighborhoods have been widespread, with the homeownership is 25 percentage points below that for subjected to the abusive practices of rate for African-American households whites. A major HUD-funded study of predatory lenders. increasing from 42.5 percent to 48.8 discrimination in the sales and rental These troublesome disparities mostly percent, for Hispanic households from markets found that discrimination still affect those families (minorities and 41.2 percent to 46.7 percent, for non- persists in both rental and sales markets immigrants) who are projected to Hispanic white households from 70.0 of large metropolitan areas nationwide, account for almost two-thirds of the percent to 75.4 percent, and for central although its incidence has generally growth in the number of new city residents from 48.5 percent to 52.3 declined since 1989. The most prevalent households over the next 10 years. percent. form of discrimination observed in the Despite the record gains in study against Hispanic and African- (iii) Single-Family Market: Trends in homeownership since 1994, a gap of American home seekers was Hispanics Affordable Lending and approximately 25 percent in the and African Americans being told that Homeownership homeownership rate prevails for housing units were unavailable when Many younger, minority and lower- African-American and Hispanic non-Hispanic whites found them to be income families did not become households as compared to white non- available. Levels of consistent adverse homeowners during the 1980s due to Hispanic households. Studies show that treatment experienced by the nation’s slow growth of some earnings, high real these lower homeownership rates are largest minority groups when they interest rates, lower inflation, and only partly accounted for by differences inquire about a unit advertised for sale continued increases in housing prices. in income, age, and other in metropolitan areas nationwide in Over the past 10 years, economic socioeconomic factors. 2000–2001 were: African Americans expansion, accompanied by low interest In addition to low income, barriers to 16.8 percent, Hispanics 18.3 percent, rates and increased outreach on the part homeownership that disproportionately and Asians and Pacific Islanders 20.4 of the mortgage industry, has improved affect minorities and immigrants percent. affordability conditions for these include: lack of capital for The study also found other worrisome families. downpayment and closing costs; poor trends of discrimination in metropolitan As this preamble and the appendices credit history; lack of access to housing markets that persisted in 2000. note, there has been a ‘‘revolution in mainstream lenders; little Examples include geographical steering affordable lending’’ that has extended understanding of the home buying experienced by African-American homeownership opportunities to process; a limited supply of modestly homebuyers, and real estate agents who historically underserved households. priced homes in locations where these provided less assistance in obtaining The mortgage industry, including the populations reside; and continued financing for Hispanic homebuyers than GSEs, has offered more customized discrimination in housing markets and 5 for non-Hispanic whites. Racial mortgage products, more flexible mortgage lending. These barriers are disparities in mortgage lending are also underwriting, and expanded outreach to discussed in Appendix A to this rule. well documented. HUD-sponsored low-income and minority borrowers. studies of the pre-qualification process (iv) Single-Family Market: Potential HMDA data suggest that the industry Homeowners and GSE initiatives are increasing the 5 Margery Austin Turner, All Other Things Being flow of credit to underserved borrowers. As already noted, the potential Equal: A Paired Testing Study of Mortgage Lending homeowner population over the next Institutions, The Urban Institute Press, April 2002. Appendix A includes further discussion of this 6 These studies are discussed in section B.1 of decade will be highly diverse, as study. Appendix B. growing housing demand from
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00007 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63586 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
immigrants (both those who are already half of Hispanic renters cite different segments of the single-family in this country and those who are homeownership as being ‘‘one of their mortgage market.) projected to arrive), minorities, and non- top priorities.’’ (v) Multifamily Mortgage Market traditional homebuyers will help to Despite these trends, potential offset declines in the demand for minority and immigrant homebuyers see The market for financing of housing caused by the aging of the more obstacles to buying a home than multifamily apartments has reached population. does the general public. Typically, the record volume. The favorable long-term Studies cited in Appendix A to this primary barriers to homeownership are prospects for apartments, combined rule reveal that increased immigration credit issues and a lack of funds for a with record low interest rates, have kept during the 1990s directly accounted for downpayment and closing costs. investor demand for apartments strong 35 percent of the nation’s rise in However, other barriers also exist, such and have supported property prices population during that decade, as a as a lack of affordable housing, little despite recently high vacancy rates. result of which the foreign-born understanding of the home buying Fannie Mae and Freddie Mac have population of the United States was 31.1 process, and language barriers. Thus, been among those boosting their million in 2000. These trends do not the new group of potential homeowners volumes of multifamily financing and depend on the future inflow of new will have unique needs. both have introduced new programs to serve the multifamily market. Fannie immigrants, as immigrants do not, on The GSEs can play an important role Mae and, especially (considering its average, enter the home purchase in tapping this potential homeowner earlier withdrawal from the market), market until they have been in this population. Along with others in the Freddie Mac have rapidly expanded country for eleven years. Fannie Mae industry, they can address these needs staff have noted that there are enough their presence in the multifamily on several fronts, such as expanding mortgage market under the Housing immigrants already in this country to education and outreach efforts, keep housing demand strong for several Goals. introducing new products, and Freddie Mac has successfully rebuilt years. adjusting current underwriting Thus, the need for the GSEs and other its multifamily acquisition program, as standards to better reflect the special industry participants to meet reflected by the increase in its purchases circumstances of these new households. nontraditional credit needs, respond to of multifamily mortgages: from $27 These efforts are necessary for achieving diverse housing preferences, and to million in 1992 to $3.9 billion in 1998 the Administration’s goal of expanding overcome the information barriers that and then rising to $9.5 billion in 2001, minority homeownership by 5.5 million many immigrants face will take on $10.7 billion in 2002, and $21.5 billion families by the end of the decade. added importance. A new or recent in 2003. Multifamily units accounted for immigrant may have no credit history The single-family mortgage market 9.0 percent of all dwelling units (both or, at least, may not have a credit history has been very dynamic over the past few owner and rental) financed by Freddie that can be documented by traditional years, experiencing volatile swings in Mac between 1999 and 2003. Concerns methods. In order to address these originations (with the 1998 and 2001– regarding multifamily capabilities no needs, the GSEs and the mortgage 2003 refinancing waves), witnessing the longer constrain Freddie Mac’s industry have been developing rapid growth in new types of lending performance with regard to the Housing innovative products and seeking to (such as subprime lending), Goals. extend their outreach efforts to attract incorporating new technologies (such as Although Fannie Mae never withdrew these homebuyers, as discussed in automated underwriting systems), and from the multifamily market, it has Appendix A to this rule. facing serious challenges (such as stepped up its activities in this area In addition, the current low predatory lending). Fannie Mae and substantially, with multifamily homeownership rates in inner cities (as Freddie Mac have played a major role in purchases rising from $3.0 billion in compared with the suburbs) also suggest the ongoing changes in the single-family 1992 to $10.0 billion in 1999, and $19.1 that urban areas may be a potential market and in helping the industry billion in 2001, then declining slightly growth market for lenders. As explained address the problems and challenges to $16.6 billion in 2002, and then rising in Appendix A to this rule, lenders are that have arisen. markedly to $30.9 billion in 2003. beginning to recognize that urban The appendices to this final rule Multifamily units accounted for 8.8 borrowers and properties have different discuss the various roles that Fannie percent of all dwelling units (both needs than suburban borrowers and Mae and Freddie Mac have played in owner and rental) financed by Fannie properties. the single-family market. A wide range Mae between 1999 and 2003. Surveys indicate that these of topics is examined, including the The increased role of Fannie Mae and demographic trends will be reinforced GSEs’ automated underwriting Freddie Mac in the multifamily market by the fact that most Americans desire, technology used throughout the has major implications for the Low- and and plan, to become homeowners. industry, their many affordable lending Moderate-Income Housing and Special According to Fannie Mae’s 2002 partnerships and underwriting Affordable Housing Goals, since high National Housing Survey, Americans initiatives aimed at extending credit to percentages of multifamily units have rate homeownership as the best underserved borrowers, their affordable-level rents and can count investment they can make, far ahead of development of new targeted low- toward one or both of these Housing 401(k) plans, other retirement accounts, downpayment products, their entry into Goals. However, the potential of the and stocks. Forty-two percent of new markets such as the subprime GSEs to lead the multifamily mortgage African-American families reported that market, and their attempts to reduce industry has not been fully developed. they were ‘‘very or fairly likely’’ to buy predatory lending. As that discussion The GSEs’ purchases between 1999 and a home in the next three years, up from emphasizes, the GSEs have the ability to 2002 accounted for less than 40 percent 38 percent in 1998 and 25 percent in bring increased efficiencies to a market of the multifamily units that received 1997. Among Hispanics and Hispanic and to attract mainstream lenders into financing during this period. Certainly immigrants, the numbers reached 37 markets. (Readers are referred to there are ample opportunities and room percent and 34 percent, respectively. Appendices A, B, and C to this rule for for expansion of the GSEs’ share of the The survey also reported that more than further discussion of the GSEs’ role in multifamily mortgage market.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00008 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63587
The GSEs’ size and market position historically experienced difficulty (ii) The GSEs’ Efforts in the Home between loan originators and mortgage gaining access to mortgage financing, Purchase Mortgage Market investors make them the logical and the flow of capital into multifamily The Appendices to this final rule institutions to identify and promote housing for seniors has been historically include a comprehensive analysis of needed innovations and to establish characterized by volatility. The GSEs each GSE’s performance in funding standards that will improve market can play a role in promoting liquidity home purchase mortgages for borrowers efficiency. As their role in the for multifamily mortgages and and neighborhoods targeted by the three multifamily market continues to grow, increasing the availability of long-term, Housing Goals—special affordable and the GSEs will have the knowledge and fixed-rate financing for these properties. low- and moderate-income borrowers market presence to push simultaneously and underserved areas. The GSEs’ role for standardization and for c. GSEs’ Past Performance and Effort in the first-time homebuyer market is programmatic flexibility to meet special Toward Achieving the Housing Goals also analyzed. Because homeownership needs and circumstances, with the opportunities are integrally tied to the ultimate goal of increasing the Since the enactment of FHEFSSA and HUD’s establishment in 1993 of the ready availability of affordable home availability and reducing the cost of purchase loans, the main findings from Housing Goals, both Fannie Mae and financing for affordable and other that analysis are provided below. multifamily rental properties. Freddie Mac have improved their • Both Fannie Mae and Freddie Mac The long-term outlook for the affordable housing loan performance. have increased their purchases of multifamily rental market is sustained, However, the GSEs’ mortgage purchases affordable home purchase mortgages moderate growth, based on favorable have generally lagged, and not led, the since the Housing Goals were put into demographics. The minority population, overall primary market in providing effect, as indicated by the increasing especially Hispanics, provides a financing for affordable housing to low- share of their business going to the three growing source of demand for affordable and moderate-income families and goals-qualifying categories. Between rental housing. ‘‘Lifestyle renters’’ underserved borrowers and their 1992 and 2003, the special affordable (older, middle-income households) are neighborhoods, indicating that there is share of Fannie Mae’s business almost also a fast-growing segment of the rental more that the GSEs can do to improve tripled, rising from 6.3 percent to 17.1 population. At the same time, the provision of their performance. percent, while the underserved areas share increased more modestly, from affordable housing units will continue (i) Performance on the Housing Goals to challenge suppliers of multifamily 18.3 percent to 26.8 percent. The figures rental housing as well as policy makers The year 2001 was the first year under for Freddie Mac are similar. The special at all levels of government. Low the higher levels of the Housing Goals affordable share of Freddie Mac’s incomes, combined with high housing established in the Housing Goals 2000 business rose from 6.5 percent to 15.6 expenses, define the difficult situation final rule. Fannie Mae met all three percent, while the underserved areas share also increased but more modestly, of millions of renter households. Housing Goals in 2001, 2002, and 2003. Housing cost reductions are constrained from 18.6 percent to 24.0 percent. Freddie Mac met all three Housing • While both GSEs improved their by high land prices and construction Goals in 2001 and 2003. However, in costs in many markets. Regulatory performance, they have historically 2002 HUD discovered that Freddie Mac lagged the primary market in providing barriers at the state and local level have had counted 22,371 housing units an enormous impact on the affordable home purchase mortgage towards the Low- and Moderate Income loans to low-income borrowers and development of affordable rental Goal even though it had previously housing. Government action—through underserved neighborhoods. Freddie counted these same housing units land use regulation, building codes, and Mac’s average performance, in towards the same goal in 2001. Freddie occupancy standards—is a major particular, fell far short of market contributor to high housing costs. Mac also counted 22,424 housing units performance during the 1990s. Fannie Since the early 1990s, the multifamily towards the Underserved Area Goal Mae’s average performance was better mortgage market has become more even though these units had also been than Freddie Mac’s during the 1993– closely interconnected with global credited towards the same goal in 2001. 2003 period as well as during the 1996– capital markets, although not to the HUD’s regulations prohibit double 2003 period, which covers the period same degree as the single-family counting. To correct for these double- under HUD’s currently-defined Housing mortgage market. Loans on multifamily counting errors, the Department has Goals. Between 1993 and 2003, 12.2 properties are still viewed as riskier by adjusted its official performance results percent of Freddie Mac’s mortgage some than are mortgages on single- for Freddie Mac in 2002 by deducting purchases were for special affordable family properties, despite delinquency the double-counted housing units, borrowers, compared with 13.3 percent rates that in recent quarters have been including all bonus point credit that had of Fannie Mae’s purchases, 15.4 percent lower than those on single-family been awarded for most of these units, of loans originated by depositories, and mortgages. from the official performance results it 15.5 percent of loans originated in the There is a need for an ongoing GSE had previously reported publicly. As a conventional conforming market (without estimated B&C subprime presence in the multifamily secondary result of these adjustments, Freddie Mac loans).7 market, both to increase liquidity and to continued to meet the Low- and advance affordable housing efforts. The Moderate-Income Goal in 2002. potential for an increased GSE presence 7 Unless otherwise noted, the conventional However, Freddie Mac fell short of the is enhanced by the fact that an conforming market data reported in this section 31 percent target for the Underserved exclude an estimate of B&C loans; the less-risky A- increasing proportion of multifamily Areas goal by 90 units or 0.002 percent. minus portion of the subprime market is included mortgages are now originated in in the market definition. See section d below and accordance with secondary market Freddie Mac’s 2002 goal performance Appendix D for a discussion of primary market standards. Small multifamily properties, results are described more fully in definitions and the uncertainty surrounding Tables 4, 6 and 8 in this preamble, estimates of the number of B&C loans in HMDA and multifamily properties with data. As noted there, B&C loans are much more significant rehabilitation needs, have including the accompanying footnotes. Continued
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00009 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63588 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
• Between 2001 and 2003, both underserved areas market in 2003 (a low loan-to-value ratios and Fannie Mae and Freddie Mac fell 26.8 percent share for Fannie Mae consequently high downpayments, significantly below the market in compared with a 27.6 percent share for which may explain the GSEs’ limited funding affordable home purchase the market). These data are based on the role in the first-time homebuyer market. mortgage loans. During this period, ‘‘purchase year’’ approach, that is, Appendix A to this rule provides special affordable loans accounted for Fannie Mae’s performance is based on evidence that there is a significant 15.1 percent of Fannie Mae’s purchases, comparing its purchases of all home population of potential homebuyers 14.7 percent of Freddie Mac’s purchase loans (both seasoned loans who are likely to respond well to purchases, and 16.2 percent of loans and newly-originated mortgages) during increased homeownership opportunities originated in the market; thus, the a particular year with loans originated produced by increased GSE purchases ‘‘Fannie-Mae-to-market’’ ratio was 0.93 in the market in that year. When Fannie in this area. Immigrants and minorities, and the ‘‘Freddie-Mac-to-market’’ ratio Mae’s performance is measured on an in particular, are expected to be a major was also 0.91. During the same period, ‘‘origination year’’ basis (that is, source of future homebuyers. underserved area loans accounted for allocating Fannie Mae’s purchases in a 24.7 percent of Fannie Mae’s purchases, particular year to the year that the d. Size of the Mortgage Market That 23.1 percent of Freddie Mac’s purchased loan was originated), Fannie Qualifies for the Housing Goals purchases, and 26.2 percent of loans Mae also led the 2003 market in funding The Department has estimated the originated in the market; the ‘‘Fannie- special affordable and low- and size of the conventional, conforming Mae-to-market’’ ratio was 0.94 and the moderate-income loans, and lagged the market for loans that would qualify ‘‘Freddie-Mac-to-market’’ ratio was only market in funding underserved area under each Housing Goal category based 0.88. loans. on 2000 Census data and geography. • While Freddie Mac has improved • Appendix A compares the GSEs’ These estimates, which are changed its affordable lending performance in funding of first-time homebuyers with slightly from estimates reported in the the past two years, it has continued to that of primary lenders in the proposed rule, are as follows: lag the conventional conforming market conventional conforming market. Both • 51–56 percent for the Low- and in funding affordable home purchase Fannie Mae and Freddie Mac lag the Moderate-Income Housing Goal loans for special affordable and low- market in funding first-time • 23–27 percent for the Special and moderate-income borrowers and homebuyers, and by a rather wide Affordable Housing Goal underserved neighborhoods targeted by margin. Between 1999 and 2002, first- • 35–39 percent for the Underserved the Housing Goals. In 2003, Freddie time homebuyers accounted for 27 Areas Housing Goal Mac’s performance on the underserved percent of each GSE’s purchases of areas goal was particularly low relative home purchase mortgages, compared These market estimates exclude the to both the performances of Fannie Mae with 38 percent for home purchase B&C (i.e., subprime loans that are not A- and the market; in that year, mortgages originated in the minus grade) portion of the subprime underserved area loans accounted for conventional conforming market. For market. The estimates, expressed as only 24.0 percent of Freddie Mac’s minority first-time homebuyers, the GSE ranges, allow for economic and market purchases compared with 26.8 percent ratio was 6.2 percent, compared to a affordability conditions that are more of Fannie Mae’s purchases and 27.6 market originations ratio of 10.6 adverse than recent conditions. The percent of market originations. percent. For African-American and market estimates are based on several • As noted above, Fannie Mae’s Hispanic first-time homebuyers, the mortgage market databases such as average performance during past GSE ratio was 3.8 percent, compared to HMDA and American Housing Survey periods (e.g., 1993–2003, 1996–2003, a market originations ratio of 6.9 data. The Department’s estimates of the 1999–2003) has been below market percent. For first-time homebuyers, size of the conventional mortgage levels. However, it is encouraging that particularly first-time minority market for each Housing Goal are Fannie Mae markedly improved its homebuyers, both GSEs substantially lag discussed in detail in Appendix D to affordable lending performance relative the private conventional conforming this rule. to the market during 2001, 2002, and market. The GSEs have room for growth in 2003, the first three years under the • The GSEs account for a small share serving the affordable housing mortgage higher housing goal targets that HUD of the market for important groups such market. The Department estimates that established in the GSE Final Rule dated as minority first-time homebuyers. the two GSEs’ mortgage purchases October 2000. Over this three-year Considering all mortgage originations accounted for 55 percent of the total period, Fannie Mae led the primary (both government and conventional) (single-family and multifamily) market in funding special affordable and between 1999 and 2001, it is estimated conventional, conforming mortgage low- and moderate-income home that the GSEs purchased only 14 percent market between 1999 and 2002. In purchase mortgage loans but lagged the of all loans originated for African- contrast, GSE purchases comprised 48 market in funding underserved areas American and Hispanic first-time percent of the low- and moderate- home purchase loans. In 2003, Fannie homebuyers, or one-third of their share income market, 48 percent of the Mae’s increased performance placed it (42 percent) of all home purchase loans underserved areas market, and a still significantly above the special originated during that period. smaller 41 percent of the special affordable market (a 17.1 percent share Considering conventional conforming affordable market. Thus, the remaining for Fannie Mae compared with a 15.9 originations during the same time 52–59 percent of the Goals-qualifying percent share for the market) and the period, it is estimated that the GSEs markets have not yet been touched by low-mod market (a 47.0 percent share purchased only 31 percent of loans for the GSEs. for Fannie Mae compared with a 44.6 African-American and Hispanic first- The GSEs’ presence in mortgage percent share for the market). However, time homebuyers, or about one-half of markets for rental properties, where Fannie Mae continued to lag the their share (57 percent) of all home much of the nation’s affordable housing purchase loans in that market. A large is concentrated, is below that in the likely to be refinance loans rather than home percentage of the lower-income loans single-family-owner market. The GSEs’ purchase loans. purchased by the GSEs had relatively share of the total rental market
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00010 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63589
(including both single-family and increase their presence in the single- relative to market projections for 2005– multifamily) was also less than 40 family rental and multifamily rental 2008 and the Housing Goal levels for percent between 1999 and 2002. markets. 2005–2008. Obviously, there is room for the GSEs to Table 1 summarizes the Department’s BILLING CODE 4210–27–P findings regarding GSE performance
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00011 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63590 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00012 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.000
The analysis for 2005 and later Income Housing Goal is slightly above range for 2005, and to the upper end of reflected in Table 1 is based on 2000 the low-end (51 percent) of HUD’s the market range projection by 2008. Census data on area median incomes market estimate range. An analysis of the GSEs’ mortgage and minority concentrations, using the Second, the 2005 Underserved Areas purchases by property type shows that metropolitan area boundaries specified Housing Goal level (37 percent) is they have had much less presence in the by OMB in June 2003. This affects the consistent with the market range (35–39 ‘‘goals-rich’’ rental segments of the market percentages for all three Housing percent) now projected by HUD for the market, as compared with the ‘‘less- Goals, as well as the figures on area Housing Goals using 2000 Census data. goals-rich’’ owner segment of the median incomes and minority market. As shown in Figure 1, GSE percentage figures that will be used to Third, the GSEs’ performance on all of mortgage purchases represented 37 measure GSE performance on the the Housing Goals was significantly percent of single-family and multifamily Housing Goals beginning in 2005. The below the market averages for 1999– rental units financed between 1999 and greatest effect of the updated data is on 2002. Appendix D to this rule provides 2002. This figure is much lower than the Underserved Areas Housing Goal. market estimates for the years 1999– their 61 percent market share for single- Expressing this goal in terms of 2000 2002 under different assumptions about family owner-occupied properties. Census data adds approximately 5 the multifamily mix (i.e., newly- (Figure 2 provides unit-level detail percentage points to the Housing Goal mortgaged multifamily units as a share comparing the GSEs’ purchases with and market levels, compared with of all financed dwelling units). The originations in the conventional analysis using 1990 Census data with estimates differ between the two home conforming market.) Metropolitan Statistical Areas (MSAs) as purchase years (1999 and 2000) and the Typically, about 90 percent of rental defined prior to 2000. heavy refinance years (2001 and 2002). units in single-family rental and The GSEs’ baseline performance For the low-mod goal, the estimates multifamily properties qualify for the figures in Table 1 exclude the effects of average approximately 56 percent for Low- and Moderate-Income Housing the bonus points for small multifamily the two home purchase years and 52 Goal, compared with about 44 percent and single-family two-to-four unit percent for the two heavy refinance of owner units. Corresponding figures owner-occupied properties and the years, with an overall 1999–2002 low- for the Special Affordable Housing Goal Temporary Adjustment Factor (TAF) for mod average of 54 percent (five are almost 60 percent of rental units and Freddie Mac that were applied in percentage points above Fannie Mae’s 16.4 percent of owner units. Thus, one official scoring toward the Housing performance and seven percentage reason that the GSEs’ performance Goals in 2001–2003. The Department points above Freddie Mac’s under the Low- and Moderate-Income did not extend these adjustments performance). The market estimates for Housing and Special Affordable beyond 2003. the underserved areas goal average Housing Goals has fallen short of HUD’s Table 1 reveals several features of slightly over 37 percent (38 percent market estimates is that the GSEs have HUD’s Housing Goals. First, it is evident during the two home purchase years had a relatively small presence in the from this table that the 2005 level (22 and 36 percent during the two heavy two rental market segments, percent) for the Special Affordable refinance years), or approximately 2–4 notwithstanding that these market Housing Goal is below the low end (23 percentage points above the GSEs’ segments are important sources of percent) of HUD’s projected market performance (see Table 1). The higher affordable housing and important range for 2005–2008. The 2005 level (52 Housing Goals are intended to move the components in HUD’s market estimates. percent) of the Low- and Moderate- GSEs closer to or within the market BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00013 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63592 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00014 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.001 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63593
In the overall conventional establishment of partnerships and As noted above, the GSEs have been conforming mortgage market, rental alliances with local communities and much less active in providing financing units in single-family properties and in community groups, leadership must for the rental housing market. Between multifamily properties represented always involve increasing the 1999 and 2002, the GSEs financed 4.5 approximately 25 percent of the overall availability of financing for million rental dwelling units, which mortgage market between 1999 and homeownership and affordable rental represented 37 percent of the 12 million 2002, 42 percent of the units that housing. Thus, the GSEs’ obligation to single-family and multifamily rental collateralize mortgages qualifying for ‘‘lead the industry’’ entails leadership in dwelling units that were financed in the the Low- and Moderate-Income Housing facilitating access to affordable credit in conventional market during this period. Goal, and 56 percent of the units that the primary market for borrowers at Thus, the GSEs’ share of the rental collateralize mortgages qualifying for different income levels, and with mortgage market was just three-fifths of the Special Affordable Housing Goal. different housing needs, as well as in their share of the market for mortgages Yet between 1999 and 2002, units in underserved urban and rural areas. on single-family owner-occupied such properties accounted for only 17 Because the GSEs’ market presence properties. percent of the GSEs’ overall purchases, varies significantly by property type, the Clearly there is room for the GSEs to 32 percent of the GSEs’ purchases Department examined whether the GSEs increase their presence in the single- meeting the Low- and Moderate-Income have led the industry in three different family rental and multifamily rental Housing Goal, and 44 percent of the market sectors served by the GSEs: markets. As explained above, these GSEs’ purchases meeting the Special single-family-owner, single-family markets are an important source of low- Affordable Housing Goal.8 Continuing rental (those with at least one rental unit and moderate-income housing since weakness in GSE purchases of and no more than four units in total), these units qualify for the Housing mortgages on single-family rental and and multifamily rental. Goals in a greater proportion than do multifamily properties has been a The GSEs’ purchases between 1999 single-family owner-occupied significant factor underlying the and 2002 financed almost 61 percent of properties. Thus, Fannie Mae and shortfall between GSE performance and the approximately 36 million owner- Freddie Mac can improve their that of the primary mortgage market. occupied units financed in the performance on each of the three conventional conforming market during Housing Goals if they increase their e. Ability of the GSEs To Lead the that period. The GSEs’ state-of-the-art purchases of mortgages on rental Industry technology, staff resources, share of the properties. An important factor in determining total conventional conforming market, As discussed below in Section II.C.4 the overall Housing Goal level is the and financial strength strongly suggest of this preamble with respect to the ability of the GSEs to lead the industry that they have the ability to lead the Home Purchase Subgoals, both GSEs in making mortgage credit available for industry in making home purchase should be able to lead the market for Housing Goals—qualifying populations credit available for low-income families single-family owner-occupied properties and areas. and underserved neighborhoods. From in all three housing goal categories— The legislative history of FHEFSSA the analysis in Appendices A-D to this special affordable, low- and moderate- reflects Congress’s strong concern that rule, it is clear that the GSEs are able to income, and underserved areas. The the GSEs need to do more to benefit improve their performance and lead the GSEs are already dominant players in low- and moderate-income families and primary market in financing Housing this market, which, unlike the rental residents of underserved areas that lack Goals—qualifying home purchase markets, is their main business activity. access to credit. (See, e.g., S. Rep. No. mortgages. In fact, Fannie Mae’s However, as already discussed, research 102–282, at 34.) The Senate Report on improved performance in 2003 is studies conducted by HUD and FHEFSSA emphasized that the GSEs evidence of this potential, as it led the academic researchers conclude that should ‘‘lead the mortgage finance market in funding home purchase loans except for Fannie Mae’s recent industry in making mortgage credit for special affordable and low- and performance on the special affordable available for low- and moderate-income moderate-income families. and low- and moderate-income families.’’ (Id.) As discussed in Appendix A to this categories, the GSEs have not led the Thus, FHEFSSA specifically requires rule, there are a wide variety of primary market in financing owner- that HUD consider the ability of the quantitative and qualitative indicators GSEs to lead the industry in establishing that demonstrate that the GSEs have reported in Section G of Appendix A and Sections the level of the Housing Goals. ample, indeed robust, financial strength F–H of Appendix D, HUD also conducted sensitivity analyses that reduced its 1999–2002 FHEFSSA also clarified the GSEs’ to improve their affordable lending performance. For example, the multifamily shares for the market by approximately responsibility to complement the two percentage points. As a result, 1999–2002 requirements of the CRA (see section combined net income of the GSEs has multifamily units decreased from 7,018,044 units to 1335(a)(3)(B) of FHEFSSA, 12 U.S.C. risen steadily over the last 15 years, 5,991,036 units (reducing the multifamily share from 14.8 percent to 12.6 percent). With these 4565(a)(3)(B)), and fair lending laws (see from $888 million in 1988 to $12.7 billion in 2003. This financial strength reduced multifamily market numbers, the GSEs’ section 1325 of FHEFSSA, 12 U.S.C. share of the multifamily market increased from 35 4545) in order to expand access to provides the GSEs with the resources to percent to 41 percent. The GSEs also accounted for capital to those historically underserved lead the industry in making mortgage higher shares of the goals-qualifying multifamily market: 42 percent for low-mod units, 34 percent by the housing finance market. financing available for families and neighborhoods targeted by the Housing for underserved area units, and 37 percent for While leadership may be exhibited 9 special affordable units. In this case, the GSEs’ through the GSEs’ introduction of Goals. shares of the overall goals-qualifying markets innovative products, technology, and (including single-family-owner, single-family- 9 As discussed in Appendix D, the GSEs rental, and multifamily mortgages) increased as processes, and through their questioned HUD’s historical estimates of the follows: low-mod—from 48 percent (see right multifamily market as too high. Section C of column of Table A.30 in Appendix A) to 50 percent 8 These percentage shares are computed from Appendix D discusses these comments and (see right column of Table A.31b in Appendix A); Table A.30 in Appendix A. Note that B&C loans are responds. As indicated in Table A.30, multifamily underserved areas—from 48 percent to 50 percent; excluded from these data. See also Table A.31b in loans accounted for 14.8 percent of all financed and special affordable—from 41 percent to 43 Appendix A. units in the market, excluding B&C loans. As percent.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00015 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63594 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
occupied housing for low-income condition of the GSEs. Further sizing analysis and highlights the need families, for first-time homebuyers, or discussion of this issue is found in the for sensitivity analyses to show the for properties located in underserved Economic Analysis. effects of different multifamily mixes on areas. the size of the goals-qualifying markets. 3. Determinations Regarding the Levels As discussed above, the Housing As explained below, HUD’s market of the Housing Goals Goals established by this rule are analysis focused on multifamily mixes quantitative measures of how well the There are several reasons why the between 13.5 percent and 16.0 percent, GSEs are serving low- and moderate- Department, having considered all the with a baseline of 15 percent. This range income homebuyers. HUD received statutory factors as well as the and baseline is consistent with HUD’s comments on this factor from Freddie comments on the May 3, 2004, proposed historical estimates of the multifamily Mac and one other commenter. The rule, is increasing the levels of the mix reported in Table D.5b of Appendix commenter stated that, in addition to Housing Goals. The following sections D. For example, between 1995 and 2002, measuring leadership through the describe these reasons and discuss and HUD estimated that the multifamily mix purchase of goal-qualifying mortgages, respond to comments received by HUD was in the 14–16 percent range. Fannie Mae and Freddie Mac’s regarding the levels of the housing In its comments, Fannie Mae leadership should be measured in more goals. estimated a multifamily mix of 12.3 qualitative ways such as their a. HUD’s Market Analysis percent, stating that HUD’s range is too ‘‘development of products and high for current conditions in the technologies that the private sector may Summary of Comments and HUD’s multifamily market. Fannie Mae cited not be willing or able to do as well.’’ Determination. As part of the process of the current high vacancy rates for This commenter asserted that through establishing goals, HUD estimates the multifamily properties and the fact that the qualitative leadership of the GSEs, size of the conventional conforming the population aged 20 to 34 will not homeownership opportunities are mortgage market. In this process, HUD begin to increase until after 2007; this expanded and costs lowered for all separately analyzes the markets for age group tends to be predominantly potential purchasers, including those in several different categories of mortgage renters. Fannie Mae also projected a low more affordable markets. loans: single-family owner-occupied multifamily refinance volume, because With respect to the issue of housing units, rental units in two-to- of a recent peak in multifamily leadership, Freddie Mac contended in four unit properties where the owner originations; these recent originations its comments on the proposed rule that occupies one unit, rental units in one- will not be able to refinance easily HUD misinterpreted the ‘‘leading the to-four-unit investor-owned properties, under their current contracts until 2008 industry’’ statutory factor and asserted and rental units in multifamily (five or or later. that ‘‘[t]here is no intimation in the Act more units) properties. This At Freddie Mac’s request, ICF or its legislative history that Congress categorization is necessary because the Consulting also calculated the intended industry leadership to be data sources differ for the various multifamily mix. In its best estimate, determined based on the enterprises categories, and it is also desirable ICF projected an average of 14.2 percent purchases of goal-qualifying mortgages.’’ because goals-qualifying shares of units over the 2005–2008 period, ranging Moreover, Freddie Mac commented that vary markedly by category. HUD between 13.7 percent and 14.7 percent the GSEs are statutorily mandated to described its methodology for analyzing in individual years, while recognizing ‘‘facilitate the financing of affordable each category in Appendix D to the that the actual outcomes may be higher housing for low- and moderate-income proposed rule, and the GSEs or lower. ICF projected multifamily families in a manner consistent with commented on that analysis. Other refinancings based on the number of their overall public purposes.’’ Freddie commenters expressed concern about units financed eight, nine, and ten years Mac stated that the overall public the magnitude of the goals, but did not ago, because 10-year balloon mortgages purpose of the GSEs is to facilitate the discuss the analysis on which the goals are the most common multifamily operation of, and provide ongoing calculations were based. mortgages, and prepayment possibilities assistance to, the secondary market for are limited by yield maintenance (i) Multifamily Share of the Mortgage residential mortgages. To the extent that agreements in their current mortgage Market the proposed goals inhibit or endanger contracts. Freddie Mac’s ability to accomplish its An important component of HUD’s In Appendix D to this rule, HUD general purpose of bringing liquidity calculation process is estimating the reviews the evidence provided by the and stability to the residential mortgage number of multifamily units financed GSEs in their comments. HUD notes that market, Freddie Mac contended that its each year as a percentage share of the the 2001 Residential Finance Survey ability to ‘‘lead the market’’ is in total number of dwelling units financed (RFS) has recently been published by jeopardy. While the Department (often referred to as the ‘‘multifamily the Census Bureau, and that the RFS recognizes the degree of qualitative mix’’); this is important because of the provides higher estimates of the leadership provided by the GSEs, the high proportions of multifamily units, multifamily mix for 1999–2001 (the Department also believes that their which qualify for credit under all three most recent years available) than either expertise and substantial financial goals. Section C of Appendix D to this Fannie Mae or ICF. The RFS data and resources allow them to lead Final Rule provides a detailed other data analyzed in Appendix D to quantitatively as well. discussion of estimates of the size of the this rule suggest that 15.0 percent is a multifamily mortgage market, including reasonable baseline, particularly in a f. Need To Maintain the Sound estimates by HUD, the GSEs, and other home purchase mortgage market Financial Condition of the GSEs researchers. As explained in Appendix environment, with a relatively small Based on HUD’s economic analysis D, comprehensive data on the annual volume of refinanced mortgage prepared for this final rule (Economic volume of multifamily mortgage originations. HUD also notes that the Analysis) and review by OFHEO, the originations are much less available ICF average of 14.2 percent is fairly Department has concluded that the than similar data on single-family close to HUD’s estimate of 15.0 percent. Housing Goals in this final rule will not mortgage originations. This introduces a HUD therefore continues to use 15.0 adversely affect the sound financial degree of uncertainty into the market percent as the best estimate of the
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00016 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63595
projected share of multifamily loans; it was 9.6 percent over 1993–2003 analyses for higher and lower investor mortgages over the 2005–2008 period. and 11.2 percent over 1999–2003. The shares of 8.0 and 9.5 percent. Using this HUD reports the goals-qualifying shares RFS for 2001 reported a larger share of range of single-family investor share of mortgage originations on the basis of investor loans than HMDA, 13.4 percent estimates, the estimate of the goals- this estimate in Appendix D to this rule. compared to 7.8 percent. The RFS also qualifying share of mortgage HUD also publishes sensitivity analyses reported larger investor shares for 1999 originations varies by about 1.5 using other estimates of the multifamily and 2000. percentage points for the low-mod goal, mix, including 12.3 percent (Fannie In the proposed rule, HUD estimated and by 0.5 percent or less for the other Mae estimate), 13.5 percent (low end of the investor share of the single-family two goals. The estimate varies HUD’s range), 14.2 percent (ICF’s best market at 10 percent, based on HMDA depending on other market factors. estimate), and 16.0 percent (high end of data and the 2001 RFS, which was then In the proposed rule, HUD estimated HUD’s range). Using this range of the most recent available. HUD also that the share of the single-family multifamily mix estimates, the estimate considered alternatives of 8 percent and market consisting of two-to-four units of the goals-qualifying share of mortgage 12 percent. Both GSEs and ICF properties with one unit owner- originations varies by about 1.5 to 2.5 commented that HUD should use occupied was 2.0 percent of all single- percentage points for the low-mod goal, HMDA data rather than RFS data, and family mortgages. This category is by about 1.0 percentage point for the should use a lower investor share in reported only in the RFS. The 2001 RFS underserved areas goal, and by about 1.2 setting the goals. While they agreed with reports that this category comprised 1.5 to 1.7 percentage points for the special HUD that the RFS provides the most percent of all single-family mortgages. affordable housing goal. The estimate accurate estimate of the true investor Because the RFS calculates a higher varies depending on other market share of the market, they stated that share of investor mortgages in the factors. lender reporting of investor loans to the single-family market (13.4 percent) than As also discussed in Appendix D to GSEs is conceptually closer to HMDA HUD employs in this rule (8.5 to 9.0 this final rule, the multifamily mix is data, which are based on lender reports. percent), it is necessary to adjust the even lower during heavy refinance They commented that the actual 2001 RFS figure upward. environments, as single-family owner opportunities available to the GSEs in The RFS reports that 85.1 percent of refinance loans dominate both the the single-family investor loan market all single-family mortgages were for market and the GSEs’ purchases. This are best measured by data that lenders owner-occupied homes. The estimated makes it more difficult for the GSEs to report, based on actual loan share of two-to-four units properties meet specific Housing Goal targets. As applications. with one unit owner-occupied in the discussed in section b below of this Fannie Mae stated that HUD’s two single-family market is calculated at preamble, HUD is soliciting public highest alternatives exceed the highest 1.73 percent (i.e., 1.5 percent/[1.5 comments on how to structure and investor share ever reported in HMDA. percent + 85.1 percent]). This figure lies implement a regulatory provision to Fannie Mae cited research indicating a between Fannie Mae’s share of about 2.0 take account of the effects of high reporting bias in HMDA, due to ‘‘hidden percent over 1999–2003 and Freddie volumes of refinance loans in some investors.’’ At the time of loan Mac’s share of about 1.5 percent. In this years on the GSEs’ ability to achieve the origination, a property may be owner- final rule, HUD uses a share of 1.6 Housing Goals. occupied or intended for owner- percent. Sensitivity analyses for 2.0 occupancy, but may become rental percent are reported in Appendix D to (ii) Single-Family Rental Share of the shortly after origination. Fannie Mae this rule. Mortgage Market stated that the same bias exists in its Similarly, the single-family owner- HUD also estimated the distribution own reporting. The hidden investors occupied share is adjusted upward to of mortgage originations for single- cannot be identified at the time of take account of the lower share of family properties, defined as structures origination. investor loans, from 85.1 percent to 89.9 with one-to-four units. In Appendix D to Freddie Mac stated that investors percent. this rule, HUD disaggregates single- have an incentive to claim falsely that The estimated market share range for family mortgage originations into three they are owner-occupants because each of the three goals categories is as categories: those on owner-occupied investor properties are subject to higher follows: 51–56 percent for the Low- and single-family homes, those on structures underwriting standards and loans tend Moderate-Income Goal, 35–39 percent with two to four units having one unit to carry higher interest rates. Freddie for the Underserved Areas Goal, and 23– owner-occupied, and those on Mac concluded that HUD should 27 percent for the Special Affordable structures with one to four rental units measure the opportunities that are Goal. These estimates are one owned by investors. HUD bases this actually available in the market to the percentage point below the market categorization on the fact that the rental GSEs, which are best measured by ranges reported in the Proposed Rule, units in the latter two categories qualify lender-reported HMDA data. for the reasons discussed above and at much higher rates for the housing In this rule, HUD has adopted HMDA detailed in Sections F–H of Appendix D. goals. data as the basis for its calculation of the The top ends of the market ranges were HUD uses two data sources in investor share of single-family mortgage reduced as follows: from 57 percent to Appendix D to estimate the size of the originations. The GSEs make a valid 56 percent for the low- and moderate- investor category, the Residential argument that lender-reported data at income market; from 40 percent to 39 Finance Survey (RFS) and the Home the time of origination measures the percent for the underserved areas Mortgage Disclosure Act database investor loans that are available for market; and from 28 percent 27 percent (HMDA). HMDA provides data only on them to purchase; HMDA provides that for the special affordable market. the investor category. The investor share data. As discussed in Appendix D to Accordingly, the 2008 goals were also of HMDA single-family loans averaged this rule, HUD projects the investor reduced by one percentage point from 7.8 percent over 1993–2003, and 8.3 share to be 8.5 to 9.0 percent (based on those included in the Proposed Rule. In percent over the recent period of 1999– HMDA) during the 2005–2008 home the Final Rule, the Low- and Moderate- 2003. The share of investor loans has purchase environments, rather than 10 Income Goal increases from 52 percent also been rising for home purchase percent. HUD also reports sensitivity in 2005 to 56 percent in 2008, as
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00017 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63596 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
compared with an increase of 52 percent interest rate stability. One suggestion recognizes and takes into consideration to 57 percent in the Proposed Rule. In that was offered for dealing with market the impact of high volumes of refinance the Final Rule, the Underserved Areas mix fluctuations (i.e., between home transactions on the GSEs’ ability to Goal increases from 37 percent in 2005 purchase and refinance loans) was to achieve the housing goals in certain to 39 percent in 2008, as compared with remove from both the numerator and years, and solicits proposals on how an increase of 38 percent to 40 percent denominator ‘‘any mortgage activity in such a provision should be structured in the Proposed Rule. In the Final Rule, excess of the percentage of home and implemented. HUD believes that it the Special Affordable Goal increases refinance loans used by HUD for would benefit from further from 22 percent in 2005 to 27 percent estimating the size of this market (i.e., consideration and additional public in 2008, as compared with an increase above 35%).’’ input on this issue. HUD also notes that of 22 percent to 28 percent in the Another commenter stated that ‘‘HUD FHEFSSA provides a mechanism by Proposed Rule. should simply set goals that require the which HUD can take into consideration GSEs to lead the market, whatever the market and economic conditions that b. Attainability of the Goals in a High market turns out to be.’’ This may make the achievement of housing Refinance Environment commenter explained that ‘‘if 50% of goals infeasible in a given year. (See 12 Summary of Comments. A common home purchase loans are to low- U.S.C. 4566(b).) theme of many of the public comments moderate income borrowers in 2005, c. Bonus Points was concern about the volatility of the then HUD should expect that a slightly mortgage market and how such higher percentage than this, say 51%, of The Housing Goals 2000 final rule volatility makes setting Housing Goals a Fannie’s and Freddie’s home purchase provided for the award of bonus points delicate and risky proposition. loans should fit in the purchase category (double credit) toward the Housing These commenters indicated that the of loans to low-moderate income Goals for both GSEs’ mortgage goals proposed by HUD would be borrowers.’’ purchases that financed single-family, unattainable, particularly in a high HUD’s Determination. This final rule owner-occupied two-to-four unit refinance environment when a large retains the approach of the May 3, 2004, properties and 5–50 unit multifamily portion of the mortgage market is proposed rule, in which the level of properties. The rule also established a comprised of refinance loans rather than each Housing Goal will increase year- temporary adjustment factor (TAF) that home purchase mortgages. by-year so that by 2008 each goal will awarded Freddie Mac 1.2 units credit Fannie Mae and others suggested that match the top of the market range for each multifamily unit in properties including single-family refinance established in section 2.d, above. over 50 units for calendar years 2001 mortgages in goals calculations creates The last three years have shown through 2003. (Congress increased the tension between liquidity goals and unprecedented volumes of refinance level of the TAF to 1.35 per unit under affordable housing goals by taking the activity. For both GSEs, refinance loans section 1002 of Public Law 106–554.) emphasis away from increasing accounted for 64 percent of all loans on The Housing Goals 2000 final rule purchase money mortgages (and single-family owner-occupied properties made clear that both of these measures therefore homeownership) and placing in 2001.10 The refinance shares were temporary, intended to encourage the focus instead on meeting high goals. increased to 70 percent for Fannie Mae the GSEs to increase their efforts to meet Freddie Mac, several trade and 73 percent for Freddie Mac in 2002, financing needs that had not been well associations, a financial organization and rose even further last year, to 79 served. During the three years for which and consumer advocacy groups also percent for Fannie Mae and 82 percent the temporary bonus points and TAF expressed concern that inclusion of for Freddie Mac. These unexpected were established, HUD expected the single-family refinances jeopardizes the record refinance rates made it more GSEs to develop new, sustainable GSEs’ abilities to increase challenging for the GSEs to attain the business relationships and purchasing homeownership through acquisitions of housing goals in the past few years, as strategies for the targeted needs. Data purchase money mortgages because the discussed elsewhere in this Preamble. indicate that, because both GSEs did focus would be on attaining goals rather The goals in HUD’s proposed rule for increase their financing of units targeted than providing affordable home the latter part of the 2005–2008 period by the bonus points and the TAF, the purchases for the target population. would be even more challenging if original objectives were met. The One trade association, however, Department determined at the end of the (contrary to current expectations) very asserted that removing refinance three years (2001–2003) not to extend high refinance rates are experienced in mortgages from the goals calculations the bonus points or the TAF. those years. Summary of Comments. A number of would only serve to encourage the GSEs HUD received a number of public non-GSE commenters, including to buy refinance loans instead of home comments seeking a regulatory solution organizations representing affordable purchase loans. By buying refinance to the issue of the ability of the GSEs to housing and consumer groups, trade loans, the GSEs could effectively ignore meet the housing goals during a period associations, organizations representing housing goals and both ‘‘jeopardize the when refinances of home mortgages racial and ethnic minorities, other safety and soundness of the GSEs due to constitute an unusually large share of organizations, and both Fannie Mae and the higher default rate of refinance loans the mortgage market. HUD is not Freddie Mac, recommended that the and increase the minority housing gap addressing the refinance issue as a Department reinstate the award of bonus due to the lower rate of minority regulatory change in this final rule. points for the GSEs which were borrowers for refinance loans.’’ Elsewhere in today’s Federal Register, Other commenters suggested that the established for 2001–2003 but which the HUD is publishing an Advance Notice final rule should include mechanisms Department did not continue after the of Proposed Rulemaking that advises the for making adjustments to the goals if end of 2003. Various non-GSE public of HUD’s intention to consider by there are changes in market conditions commenters, in addition to separate rulemaking a provision that including a surge or drop in refinance recommending reinstatement, also volume. These commenters asserted that 10 By way of comparison, the refinance rate was suggested that HUD develop new bonus the GSEs’ ability to successfully meet 29 percent for both Fannie Mae and Freddie Mac point incentives for other unmet the goals should not be contingent upon in 2000. housing needs, such as manufactured
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00018 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63597
housing, rural housing, or tax credit will take away from the broad middle Using historical trends in prices, the properties or for particular groups, e.g., class, especially in high-housing cost GSEs asserted that their presence in the Native Americans, other minority regions. For example, Fannie Mae mortgage market explains why populations, or persons with asserted that if the special affordable mortgage-backed securities have a more disabilities. housing goal had been set at 28 percent stable price trend than commodity Fannie Mae recommended that HUD in 2003, then it would have needed to markets. They warned that because of provide bonuses for targeted business greatly curtail support to the overall denominator management resulting such as extremely low-income market to meet that goal. Fannie Mae from unrealistic goals, they could not households, i.e., those with incomes concluded that such manipulation does buy mortgage-backed securities and less than 30% of area median income not promote stability and limits encourage stability in a financial crisis. (AMI); first-time homebuyers; liquidity, and that it can shut out The GSEs further contended that if manufactured housing; rural areas; and working middle class borrowers, they reduce their willingness to buy small multifamily properties. Freddie contribute to higher interest rates and non-goals eligible mortgages, it will be Mac suggested that instead of purchase lower conforming loan limits. harder for borrowers whose incomes money subgoals, the Department could Many commenters, including Freddie marginally exceed goals eligibility provide bonus point incentives for these Mac, also claimed that setting the goals requirements to obtain financing since mortgages. Freddie Mac stated that the at a high percentage may lead to the two income-based Housing Goals bonus points for two-to-four unit and 5– denominator management. They state compare the incomes of the borrower or 50 unit properties provided an that denominator management would resident to area median income. For extremely effective incentive. Freddie occur if a GSE purposely abstained from example, the combined incomes in a Mac indicated that other markets that buying mortgages in the markets that are working family may just disqualify that could be assisted by bonus points are not goals-eligible, rather than increasing family’s loan for eligibility under the rural and manufactured housing. its purchases in markets that are goals- low- and moderate-income goal even Freddie Mac noted that the eligible. Freddie Mac contended that though each individual’s income would Department’s concern that bonus points this may be necessary if goals are set not be considered to be affluent. The obfuscate the GSEs’ actual goals- above the market percentage of available GSEs and other commenters provided qualifying performance is easily goal-qualifying loans. One financial examples of working families in the remedied by having the GSEs report two institution observed that denominator middle class, such as ‘‘teacher/fireman’’ numbers, one with and one without the management ‘‘will be exacerbated by the households, that could encounter bonuses. fact that the GSEs do not operate in the difficulties in financing a home. HUD’s Determination. The primary market and do not have any Moreover, the commenters asserted Department has fully considered the direct control over the origination that non-goal qualifying households comments suggesting the re- strategies of their customers.’’ may have higher costs associated with introduction of bonus points, as well as In addition to the allocation problems available financing since these other types of targeted incentives for the discussed above, the GSEs stated that mortgages would be less likely to be GSEs’ mortgage purchases, and has the liquidity requirements in their purchased by a GSE. Freddie Mac determined not to reinstate the bonus charters imply that they must stand asserted that HUD did not take this into points for the years covered by this rule. ready to buy any and all conventional, account in its cost/benefit analysis. The position of the Department conforming residential mortgages. They Furthermore, commenters claimed discussed in the preamble of the contend that denominator management that denominator management may proposed rule (see 69 FR 24228, 24232) is in direct conflict with these contribute to higher interest rates and, remains unchanged; that is, the provisions, and goals set higher than as a result, harm the precise borrowers continued use of the bonus points market originations could force the that HUD is trying to help. These ‘‘would only result in misleading GSEs to refuse to purchase mortgages commenters stated that if denominator information about the extent to which that are not eligible. This, in turn, could management reduces liquidity then the the GSEs are, in fact, meeting the reduce liquidity in the market. Knowing supply of mortgage funds will decline Housing Goals.’’ In addition, the that the GSEs would no longer stand and interest rates will rise. The GSEs Department reiterates that the ‘‘decision willing and able to purchase all contended that if they are less willing to to increase the levels of the Housing conventional, conforming mortgages, buy mortgages under all conditions, Goals substantially in a staged manner other market participants might be less then investors will be less willing to * * * and, at the same time, not renew willing to hold these mortgages in their provide funds to the market. As a result, the bonus points or TAF, will ensure portfolios, and general liquidity would the GSEs claimed that as investors seek that the GSEs continue to address the decline. The GSEs further asserted that out safer instruments, home mortgage areas formerly targeted by these changing market forces could cause interest rates will rise, and this rise in measures’’ (see 69 FR 24228, 24232). swings in prices and trading volumes, home mortgage rates will harm even and these temporary disturbances could those borrowers that are still goals- d. Appropriate Levels of the Goals create unstable markets, increase risk, eligible. In the May 3, 2004, proposed rule, and reduce the willingness of investors Several commenters expressed HUD set the Goals to increase to levels to invest in the sector. Thus, the GSEs concern about the effect of the goals on at or near the high end of the estimated maintained that denominator high cost markets. One commenter market range for each goal category by management decreases market stability. explained that while the goals are set 2008. A large number of commenters The GSEs pointed to specific with a national standard, a market level expressed concern that these goal levels historical examples that describe their analysis ‘‘reveals a pronounced shortage were set too high, and could have positive influence on stability. They of affordable mortgages in high cost deleterious consequences for the maintained that during the 1990–1991 housing markets.’’ Commenters stated mortgage market as a whole, or for recession, the GSEs advised that they that the GSEs’ current loan purchasing specific sectors of the market. stood ready to purchase mortgages patterns demonstrate that market Fannie Mae commented that a high while many industry participants affordability already has an impact on allocation of affordable mortgage credit curtailed their purchase programs. goals-related purchases. The
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00019 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63598 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
commenters expressed concern that With regard to the effects of the goals subsidies to pay claims, effectively high cost markets could see even tighter on high-cost markets, HUD notes that making FHA the lender of last resort,’’ credit if the proposed goals are enacted. the overall presence of the GSEs in these said one trade association. The GSEs note that under HUD’s May markets depends on the conforming One financial institution stated that 3, 2004, proposed rule, the goal levels loan limit, which has been established the so-called competition for goals- rise to levels at the top of HUD’s market by Congress for all states, including qualifying loans would not be between range in 2008 and stabilize there. They states deemed to be ‘‘high-cost areas.’’ traditional conventional lenders vying state that the projected market range With regard to HUD’s housing goals for loans with a separate group of concept is one in which HUD projects more specifically, the low- and traditional FHA lenders, but rather an market levels of loans generated within moderate-income and special affordable acceleration of product competition each goal category will fluctuate within goals are based on borrower income within a single group of existing lenders the range, depending on relative relative to area median income, thus a who originate for both markets. This volumes of single-family refinance loans mortgage for a lower-income family in a commenter stated that 12 of the top 15 relative to other loans, interest rates and high-income metropolitan area will (by volume) FHA/VA lenders are also other macroeconomic and housing count towards the goals in the same among the top 15 conventional lenders market conditions. The GSEs express manner as a mortgage for a lower- and indicated that the increased product the concern that, in any particular year, income family in a low-income area. competition would not result in a net they could be confronted with goal Underserved areas are defined in terms increase in goals-qualifying loans, but in levels that are several percentage points of median family income in a census a shift from FHA to the GSEs of FHA’s higher than the market percentages of tract relative for median income in the relatively lower risks. goal-qualifying loans, or goal levels that area; thus a mortgage for a family living HUD’s Determination. The are at the market percentages. The GSEs in a lower-income tract in a high- Department agrees with many of these state that if HUD’s proposed Housing income metropolitan area will count commenters that improvements in Goals are retained, they foresee years towards the goals in the same manner as technology, such as the widespread use when the goal levels will be attainable a mortgage for a family living in a lower- of commercial credit scores, mortgage only by means of ‘‘denominator income tract in a low-income area. Thus scores, and automated underwriting management’’ in which they limit their HUD concludes that its housing goals systems, have fundamentally changed purchases of loans that do not qualify will have no adverse impact on the way lenders process loan for the goals. borrowers or neighborhoods in areas applications in recent years. Where once HUD’s Determination. Many of the with high housing costs. rules-based underwriting distinctions comments expressed concern about the between prime conventional and FHA goal levels established for the last year e. Consequences of the Goals for FHA loans were fairly clear, in recent years, or two of the period covered by this Fannie Mae, Freddie Mac, several with the new technology, these rule. In these years, the goals are set at trade associations, two advocacy groups distinctions have become blurred. For the market levels estimated by HUD. and two financial institutions expressed example, loan applications with Also, since they are the later years, concern over the impact HUD’s payment-to-income ratios above market projections are necessarily more proposed goals would have on the conventional market guidelines were imprecise. In particular, the possibility future solvency of the FHA program. once clearly candidates for FHA of a decline in mortgage interest rates in One trade association asserted that financing because FHA would accept those years raises the possibility of ‘‘excessive goals will push GSEs to applicants with higher payment-to- another boom in refinancing, and thus expand into the least-risky part of the income ratios. However, today, the same greater difficulty for the GSEs to meet FHA market and put into question application would be processed using the housing goals without denominator FHA’s long-term viability.’’ an automated underwriting system management. The comments relating to The aforementioned commenters (AUS) that scores the application based middle-income borrowers are reiterated this point by stating that on the totality of the application’s risk predicated on the difficulty of unrealistically high goals would compel factors. What once may have been an foreseeing refinance volatility. Recent the GSEs to increase competition with unacceptable payment-to-income ratio years have seen large unexpected home FHA for higher credit quality borrowers for a prime conventional loan may now refinance rates. Since higher income and would therefore further undermine be acceptable if the application contains homeowners disproportionately engage the FHA program in the long-run. One offsetting low risks in other key areas in refinancing, inclusion of refinance advocacy group asserted that not only such as borrower cash reserves, loan-to- loans in the denominator increases the will these goals encourage the GSEs to value ratio, or commercial credit scores. difficulty of GSE goals performance. A compete with FHA more in the single In addition to these technological middle-income borrower just above the family sector but in the multifamily changes, FHA made several changes to low/mod bracket would be less sector as well. its underwriting guidelines in FY 1995 attractive to the GSEs in high refinance Freddie Mac and Fannie Mae agreed in order to promote increased years. As noted in section II.C.3.b., HUD that they would be compelled to more homeownership opportunities among is considering in a separate rulemaking aggressively compete with FHA in low-income and minority homebuyers. a provision that recognizes and takes procuring top-quality borrowers. By doing so, FHA modestly increased into consideration the impact of high Freddie Mac stated that the GSEs would the risk characteristics of its post-1995 volumes of refinance transactions on the take as many as ‘‘1⁄3 of all FHA books of business, but it succeeded in GSEs’ ability to achieve the housing borrowers.’’ Freddie Mac and two trade raising FHA’s proportion of first-time goals in certain years. HUD also notes associations further contended that such homebuyers from 60.9 percent in fiscal that FHEFSSA provides a mechanism by a loss to the FHA program would be year 1994 to 73.0 percent in fiscal year which HUD can take into consideration seen in the increasing expenses to the 2003. During the same period (fiscal market and economic conditions that remaining FHA borrowers. As the FHA years 1994 to 2003), FHA’s proportion may make the achievement of housing program loses better quality loans to the of minority borrowers increased from goals infeasible in a given year. (See 12 GSEs, the result would be ‘‘higher fees 24.8 percent to 33.0 percent, and has U.S.C. 4566(b).) to FHA borrowers or government since remained at this level, or higher.
VerDate jul<14>2003 19:33 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00020 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63599
The new technology may allow the respectively. The increases in both substantial rehabilitation, and Section conventional market to identify lower measures for Fiscal Year 2003 were 223(f). FHA endorsed over $2.1 billion risk loan applications that historically driven by the large positive economic in Section 223(a)(7) loans during Fiscal have come to FHA. However, the ability value the actuary placed on a record Year 2003, and is expected to endorse to identify risks does not, in and of dollar volume of new loans FHA about $1.4 billion during Fiscal Year itself, equate to shifts in market share insured in FY 2003 along with the rapid 2004. As with the Section 223(f) from FHA to conventional lenders. prepayment of older loans, keeping the program, FHA’s Section 223(a)(7) Better pricing for borrowers by the end-of-year insurance-in-force program is also profitable to FHA— conventional market is required to lure (denominator of the capital ratio) down. operating with an estimated negative lower risk borrowers from FHA. If With regard to the GSEs taking credit subsidy of 2.2%. conventional lenders use the new multifamily business away from FHA, If FHA does lose some multifamily technology to not only evaluate risks but the Department notes that there are market share from its purchase or also to price according to risk, then many differences between the types of refinance programs for existing housing there may be some shift from FHA to the multifamily mortgages FHA insures and as a result of the goals, it would not conventional market. Such a shift can those the GSEs purchase. For newly likely have any significant impact on produce tangible benefits for borrowers constructed multifamily properties, FHA overall. in the form of lower cost mortgage FHA insures the loan from the start of construction while GSE multifamily f. Consequences of the Goals for the financing. Multifamily Market The Department does not believe it is loan products generally do not. The FHA’s mission to compete with the GSEs do have forward commitment Summary of Comments. Several private sector. Rather, FHA’s mission is programs that can be used for new organizations commented on potential to complement the conventional market, construction, but the purchase of the adverse consequences if the housing using FHA’s cost of capital advantage permanent loan by the GSEs generally goals are set too high. Fannie Mae and where it can have the most benefit in requires the property to achieve Freddie Mac, among others, cited the creating homeownership opportunities minimum sustained occupancy levels, recent high vacancy rates for for those households who might not whereas FHA does not have this multifamily rental housing as an otherwise be served by the prime requirement. However, it is possible that example that increased lending by the conventional market. the new goals will provide incentives GSEs at this time would encourage HUD gauges the soundness of FHA’s for the GSEs to expand and refine their overbuilding. insurance funds in several ways. The forward commitment products to be Others stated that the multifamily statutorily mandated annual more attractive in the market for new market is already flush with capital and independent actuarial review of FHA’s multifamily housing. This could be a that inappropriate goals could promote principal single-family insurance fund, benefit to the market. overly aggressive bidding for loans and the Mutual Mortgage Insurance Fund The greatest potential impact of the reckless lending. (MMIF), provides the Department, and higher housing goals on FHA’s One trade association stated that the the public, with an outside expert’s multifamily business may come from a increased presence of the two GSEs estimate of the capital ratio of the reduction in two of FHA’s programs that would promote a duopsony (a market overall fund, and the economic value of address the purchase or refinance of with only two buyers) that would new business coming into the fund. The existing properties. The first is the hinder competition in the multifamily capital ratio indicates whether the Section 223(f) program, which insures mortgage market. Other commenters suggested that existing books of business (current mortgages for the purchase or refinance increased loan purchases by the GSEs portfolio) are financially sound, while of existing (over three year old) properties that are not currently would skim the highest credit-quality the economic value estimates of new financed with an FHA mortgage. This loan from other mortgage lenders, and business show whether if the marginal program accounted for about $0.8 reduce the credit quality of multifamily impact of new loans insured is adding billion in endorsements for FHA during loans remaining in the portfolios of or detracting from the financial health of Fiscal Year 2003, and is expected to pension funds or insured through FHA. the fund.11 Specifically, the Fiscal Year produce about $0.5 billion in Another commenter stated that 2003 actuarial review estimated the endorsements during Fiscal Year 2004. increased goals pressure on the GSEs economic value of the MMIF at the end FHA’s 223(f) business is estimated to be would cause them to concentrate on of Fiscal Year 2003 to be $22.7 billion profitable to FHA—it is estimated to large properties, where a single loan and the fund’s capital ratio to be 5.21 have a credit subsidy (net present value would contribute more toward goal percent—the eighth full year this ratio of all cash flows from the insurance attainment. has exceeded the Congressionally contract at the time of endorsement) of HUD’s Determination. One of HUD’s mandated minimum of 2.0 percent. The negative 3.0%.12 The second is the objectives in promulgating this final economic value of new loans endorsed Section 223(a)(7) program, which rule is to promote the availability of for insurance during 2003 was estimated insures mortgages for FHA-to-FHA mortgage credit to affordable properties by FHA’s independent actuary to be refinances—that is, the refinance of an at the lowest possible cost. It is not the $2.8 billion, indicating new business existing FHA-insured mortgage. Section intent of this rule to promote the coming into FHA is further contributing 223(a)(7) is used, for example, to maximum flow of credit to this market, to FHA’s reserves. refinance loans previously insured regardless of housing and mortgage In comparison, the Fiscal Year 2002 under FHA’s most used programs—i.e., market conditions. actuarial review estimated the economic Section 221(d)(4) new construction/ Increased competition for business, as value and capital ratio of the MMIF at intended by the rule, should bring $22.6 billion and 4.52 percent, 12 A negative credit subsidy of 3.0 percent means benefits to borrowers, and therefore that the net present value of FHA’s revenues renters, through lower interest rates and 11 ‘‘Economic value’’ is the net present value of (premiums, fees, recoveries from claims paid, etc.) the fund’s reserves plus expected future cash flows, will exceed the net present value of FHA’s program more attractive non-price terms. Such and the ‘‘capital ratio’’ is economic value divided costs (claims and related expenses) by 3.0 percent increased competition does not imply by insurance-in-force. of the total insured mortgage amount. impaired credit quality or lax
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00021 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63600 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
underwriting. As the GSEs compete Regarding the market structure family rentals are another source of more aggressively for multifamily implications of increased GSE affordable housing. Also, the capital business and gain market share, the multifamily activity, HUD estimates that provided by investors can help maintain market will not necessarily grow one- the GSEs purchased slightly less than demand for single-family homes in for-one with every additional loan one-third of the dollar volume of underserved neighborhoods. While purchased by the GSEs. It is likely that conventional multifamily loan some commenters complained that this the market impacts will be more on the originations during 2001–2003 (see raises the cost to first-time homebuyers, pricing of multifamily credit and less on Table D.2). There is room for increasing investors also help to maintain the the volume of credit supplied. Lower this market share without producing the liquidity and value of owner-occupied pricing of credit in and of itself does not duopsony alluded to in the previously homes. Further, there are some investors promote overbuilding; its one cited comments. Furthermore, if the who make it their business to renovate unambiguous effect is to reduce the cost GSEs do increase their market the housing stock and resell the of supplying housing to consumers. penetration, it is because they are properties. On balance, HUD found no Demand for multifamily mortgages offering multifamily borrowers more compelling evidence that single-family will be responsive to cyclical attractive products or pricing than are rentals should be excluded from goals macroeconomic factors. Beyond these their competitors, including the pension eligibility. funds and FHA programs alluded to by influences, demand for multifamily h. Consequences of the Goals for the some commenting organizations. The housing will be supported by favorable Subprime Market demographics. In its comments on the borrower and, ultimately, the rent- Summary of Comments. Both GSEs proposed rule, Fannie Mae highlighted paying affordable housing resident indicated that they would need to the prospective growth in the number of benefit from these more attractive increase their purchase of subprime people ages 20 through 34 in arguing products and pricing. In summary, the Department’s loans to meet the higher goals. Freddie that the demographics do not become Mac stated that the increased affordable clearly favorable to rental demand until determination is that the Housing Goal levels established by this rule are housing goals created tension in its late in this decade. But fewer than half prudent and will improve the business practices between meeting the of all renter households are headed by availability and pricing of credit for goals and conducting responsible someone of this age, and more affordable multifamily properties. For lending practices. comprehensive estimates and the reasons stated above, it is the In the past, Fannie Mae and Freddie projections suggest a steadier path of Department’s view that the rule will not Mac have voluntarily decided not to moderate growth in the demographic have the adverse consequences purchase subprime loans with features component of demand for multifamily mentioned in some comments on the such as single-premium life insurance housing. proposed rule. and prepayment penalty terms that Interest rates clearly will be important exceed three years, or to purchase loans for the future path of mortgage lending, g. Consequences of the Goals for the subject to the Home Ownership and as noted by Fannie Mae, Freddie Mac, Single-Family Rental Market Equity Protection Act (HOEPA). Freddie and other commenters. The historically Summary of Comments. Several Mac indicated that the increased goals low interest rates of recent years have community organizations raised would limit its ability to influence spurred lending in both the multifamily concerns about encouraging the single- subprime lending practices. More and single-family markets. If interest family rental market. They asserted that specifically, Freddie Mac claimed that, rates should rise in the future, the the goals should target families who to meet the higher housing goals, it volume of mortgage lending presumably want to live in the financed houses, as might not have the option in the future would be lower than if rates were to opposed to the investors who purchase of turning away subprime loans that remain at current levels. But the effect these homes. In these commenters’ have less desirable loan terms than the of higher rates on the GSEs’ ability to view, investors take affordable housing subprime business it currently achieve the housing goals is less clear. stock off the market, which raises the purchases. Because the goals are established in price for low and moderate-income first- Several commenters suggested that if terms of shares of the GSEs’ business, time homebuyers. They claimed that the GSEs are pushed to serve more of rather than levels, a key question is how homeownership should be stressed the subprime market, they will skim a higher interest rates would affect the because home equity is a large significant portion of the lower-risk relative demand for single-family and component of the disparity that exists in borrowers from that market. The multifamily mortgage credit. Because of household net wealth between ethnic resulting smaller subprime market differences in prepayment provisions groups. would include the neediest borrowers. and other characteristics between Some commenters cited studies that The commenters stated that these higher single-family and multifamily mortgage suggest homeownership has beneficial risk borrowers would pay more because lending, multifamily credit demand neighborhood effects relative to lower risk borrowers would not be might drop off proportionally less than investor-owned properties. According to present to subsidize them, and the would single-family credit demand in one cited study, absentee landlords are market’s high fixed costs would be response to higher rates.13 This in turn much more likely to let housing stock distributed across fewer borrowers. would make it easier to attain the goal decline but homeowners are much more One industry group also suggested levels if interest rates were to increase likely to invest in the upkeep of their that a significantly smaller subprime from current levels. homes. In the view of one of these market for private lenders would drive organizations, the incentives that the some lenders out of business and 13 This is suggested by recent experience of GSEs receive for rental housing should translate into less competition. below-average multifamily mix in years where the be to promote multifamily While some industry commenters volume of single-family refinancings has been high. developments, not single-family homes. welcomed the entrance of the GSEs into Further support is provided by evidence of a relationship between interest rates and the HUD’s Determination. HUD the subprime market because their multifamily share of the net change in residential considered many factors related to the presence would bring stability and mortgage debt between 1975 and 2002. single-family rental market. Single- standardize business practices, the
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00022 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63601
commenters also expressed concern that risk profiles differ markedly from neighborhood can lead to serious blight unrealistically high goals could force borrowers who qualify for prime and disinvestment in the community. the GSEs to jump into the market in a mortgage products. Millions of One commenter recommended that manner that negatively distorts Americans with less than perfect credit HUD establish safeguards against underwriting and pricing. These or who cannot meet some of the tougher aggressive affordable products. The commenters contended that the GSEs underwriting requirements of the prime commenter suggested that HUD deny could bring capital and standards up, market for reasons such as inadequate Housing Goal credit for GSE mortgage but that they must gradually and income documentation, limited purchases that experience early-term carefully enter the subprime market to downpayment or cash reserves, or the serious defaults (e.g., delinquent 90 desire to take more cash out in a have a positive effect. They strongly days or longer within 12 months of the urged HUD to lower the goals to refinancing than conventional loans date of origination). encourage the GSEs to expand their allow, rely on subprime lenders for subprime activities at a measured pace. access to mortgage financing. If the The GSEs and community groups Some commenters suggested that GSEs reach deeper into the subprime cautioned that the struggle to meet high bonus points, or other incentives for the market, more borrowers will benefit goals for low-income groups could GSEs’ purchases of certain nonprime from the advantages that greater stability cause the GSEs to relax underwriting loans, could foster more deliberate and and standardization create. standards and/or extend loans to people prudent purchases by the GSEs of i. Consequences of the Goals for who are unprepared. For example, the subprime loans. One lender also commenters pointed out that FHA suggested that incentives could be Mortgage Defaults; Neighborhood Impacts default rates are higher than the granted to the GSEs for other conventional conforming market. High Summary of Comments. HUD underserved market segments, such as goals would encourage the GSEs to enter received several comments concerning manufactured homes, minority first time markets served by FHA. This incentive buyers, and nonprime first-time buyers. the impact of mortgage default rates on to extend credit to unprepared low- HUD’s Determination. To date, the neighborhoods. Comments from GSEs’ involvement in the subprime mortgage insurance companies income people would rise if unexpected market has benefited two types of highlighted that the higher goals will refinances decreased the proportion of borrowers: ‘‘A’’ risk and ‘‘near A’’ risk. likely lead to more expanded affordable goals-eligible units produced in the The first group consists of borrowers housing products as well as higher market. with risk profiles similar to ‘‘A’’ foreclosures. Affordable products HUD’s Determination. HUD carefully borrowers, but receive mortgages from a present challenges to borrowers and reviewed the comments regarding subprime lender. The GSEs’’ outreach lenders. For borrowers, qualifying for an mortgage default rates. The Department and education efforts increase the affordable mortgage does not insure they believes that the GSEs’ presence in likelihood that ‘‘A’’ borrowers will use have a clear understanding of the risks underserved markets will be beneficial cheaper prime lenders for refinance of homeownership. Where aggressive for neighborhoods. The GSEs have mortgages, and reduce their reliance on affordable products are aimed at improved their underwriting methods to subprime firms. The second group, qualifying borrowers for home loans better identify risks in these markets, borrowers who are near A credit risks, rather than qualifying families for and also have instituted homebuyer has growing access to mortgage products homeownership, lenders need to be education programs. An increased role offered by the GSEs as these borrowers cautious of products that test the limits are increasingly served by GSE seller/ of borrowers’ credit capabilities. for the GSEs’ seller-services in inner- servicers. Affordable products that have been city neighborhoods will improve The GSEs have been prudent in their introduced into the market under competition, reduce high-cost lending, pursuit of subprime lending, focusing favorable economic conditions can and reduce predatory lending. As on the top part of the market, the ‘‘A- experience increasing defaults and described in Appendix A, families minus’’ and ‘‘Alt A’’ segments. A-minus foreclosures during periods of higher living in inner-city, high-minority mortgages are typically those where interest rates, higher unemployment neighborhoods often have to rely on borrowers have less than perfect credit. and/or lower house price appreciation subprime lenders as their main source Alt A mortgages are originated to rates. One commenter indicated that 15 of mortgage credit. Studies indicate that borrowers who cannot document all of percent or more of borrowers in some many of these borrowers obtaining high- the underwriting information in the affordable housing products could cost loans could qualify for lower-cost, application but generally have FICO experience default in an economic prime mortgage credit. An active GSE scores similar to those in the prime downturn. effort in these neighborhoods will market. The GSEs’ subprime products As defaults on affordable products encourage traditional, mainstream are integrated into their automated rise, inner city neighborhoods can be lenders to increase their lending underwriting systems and are approved especially hard hit. A large number of activities in these historically based on mortgage scoring models. foreclosures in an area may lead to underserved areas. This will offer These models have proven over the abandoned properties. While additional funding options for those years to be an effective tool in limiting foreclosures devastate borrowers who lower-income and minority borrowers risk layering. The GSEs charge lenders lose their homes and damage borrowers’ who today may have to take out a high- higher fees for guaranteeing these loans. credit history, foreclosures also weaken As a result these higher risk loans are the neighborhoods where the properties cost loan in order to purchase or priced above those offered to prime are located. renovate a home or to refinance an borrowers but below what subprime The potential for affordable lending existing mortgage. Reductions in lenders would otherwise charge for products to result in higher foreclosure predatory lending reduce the costs of these loans. during a less prosperous economic mortgages and the chances of default. The GSEs’ presence in the subprime environment was echoed in Freddie As a result, the Department believes that mortgage market benefits many low- Mac’s comments. Its comment discussed GSE participation is a net benefit to income and minority borrowers whose how too many defaults in one lower income neighborhoods.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00023 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63602 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
j. Consequences of the Goals for purchase market has been below market another market. Fannie Mae stated that Residents of Puerto Rico levels, and the Administration’s conflicts between the goals arise Summary of Comments. Several emphasis on increasing homeownership because the goals are set as a percentage associations stated that HUD’s proposed opportunities, including those for low- of business, and fulfilling the numerator affordable housing goals could be and moderate-income and minority of one goal adds to the denominator of disadvantageous to residents of Puerto borrowers, HUD proposed to set Home the other goals. Fannie Mae asserted Rico, alleging that less than 10 percent Purchase Subgoals for GSE mortgage that the GSEs could be forced to abstain of loans that are originated in the Puerto purchase activities to increase financing from buying non-goal eligible mortgages Rican market would qualify for the opportunities for low- and moderate- that would count in the denominator, goals. These commenters were income, underserved, and special but that would not benefit its concerned that the GSEs might be affordable borrowers who are calculation of goals performance in the unable to buy loans from Puerto Rico, purchasing single-family homes. numerator. In Fannie Mae’s view, its Specifically, the Department proposed and urged HUD to take special measures own abstention from buying implies an Home Purchase Subgoals for home to ensure that owner and rental housing illiquid market. purchase loans that qualify for the Other commenters affirmed Fannie production are not deleteriously Housing Goals. The purpose of the Mae’s comments and expressed concern affected by the demographic and Home Purchase Subgoals is to ensure that, given the market leadership of the economic differences that exist between that the GSEs focus on financing home GSEs, the manner in which home the mainland markets and the Puerto purchases for the homeowners targeted purchases are counted toward the Rico market. HUD’s Determination. Loans by the Housing Goals. The Department Subgoals could distort the lending believes that the establishment of Home market. purchased by the GSEs for properties in Purchase Subgoals will place the GSEs In addition, both Fannie Mae and Puerto Rico are counted in the same in an important leadership position in Freddie Mac asserted that FHEFSSA manner as loans purchased on the Housing Goals categories, while also requires that HUD consider each of the properties in any other location. Since facilitating homeownership. The GSEs six statutory factors set forth in sections underserved areas are defined as low- have years of experience in providing 1332(b) and 1334(b) of the statute in income and/or high-minority census secondary market financing for single- setting the levels of any Subgoals within tracts in metropolitan areas or counties family properties and are fully capable those Housing Goals. Freddie Mac in non-metropolitan areas, the of exerting such leadership. objected to the home purchase Subgoals overwhelming majority of loans The focus of these Subgoals on home because it claimed these Subgoals purchased by the GSEs on properties in purchase loans meeting the Housing would constitute micromanagement of Puerto Rico count toward that goal. In Goals will also help address the racial the GSEs’ business decisions. Freddie fact, in 2003, Fannie Mae reported that and income disparities in Mac also noted that, in the past, HUD 95 percent of the units it financed in homeownership that exist today. As has declined to implement subgoals for Puerto Rico qualified for the noted earlier, although minority that very reason. underserved areas goal; the homeownership has grown, the Several commenters expressed the corresponding figure for Freddie Mac homeownership rate for African- view that HUD had overestimated was 98 percent. American and Hispanic families is still available purchase money mortgages Relatively few of the loans in Puerto approximately 25 percentage points and noted that if Subgoals on these Rico that are purchased by the GSEs below that for non-Hispanic white types of mortgages are set too high, qualify for the two income-based goals. families. The focus of the Subgoals on adverse market distortions will occur. Despite this, HUD does not believe that home purchase will also increase the Other commenters contended that, the final housing goals will adversely GSEs’ support of first-time homebuyers, regardless of the level of the Subgoals, affect Puerto Rico. In 2003, Puerto Rico a market segment where they have a subgoal that targets home purchase accounted for only 0.2 percent of all lagged primary lenders. mortgages unfairly allocates credit units financed by Fannie Mae and only Summary of Comments. Fannie Mae toward home buying rather than 0.1 percent of all units financed by claimed that the proposed Subgoals are mortgage refinances. These commenters Freddie Mac. Thus overall performance not necessary and are, in fact, asserted that this credit allocation is on these broad national goals is not duplicative of the broader goals unfair in that it penalizes borrowers materially affected by the characteristics structure. Fannie Mae asserted that it is who want to lower mortgage costs or of loans purchased by the GSEs in already a leader in financing home improve their homes. They also Puerto Rico. purchases, even in a period of contended that credit allocation that Apparently many lower-income aggressive refinancings. In addition, promotes purchase mortgages could families in Puerto Rico rely on Fannie Mae stated that subgoals add push refinance borrowers into high-cost consumer finance companies for complexity to the mortgage market and loans rather than conforming, GSE- financing their homes. Since such contribute to a loss of liquidity, and eligible mortgages. To combat such financing is typically more expensive to suggested that the proposed Subgoals do effects, one organization suggested borrowers than traditional mortgages, not reflect recent market experience separate subgoals for both purchase this suggests that the GSEs could play because affordability may decline and money mortgages and refinances, with an important role, working with HUD may mistreat missing data when the overall low- and moderate-income mortgage originators, to better develop formulating subgoals. Fannie Mae also goal as the weighted average of the the mortgage market in Puerto Rico. stated that HUD improperly exercised different subgoals. 4. Determinations Regarding the its authority in proposing the Subgoals. Commenters also objected to mortgage Specification and Levels of the Home Specifically, Fannie Mae contended purchase subgoals targeting only those Purchase Subgoals that a complex subgoal structure harms loans originated in metropolitan areas liquidity and that when Fannie Mae because this geographic limitation a. Overview needs to stretch in one market to meet allocates credit at the expense of Given that the past average a goal, it may have to reduce its residents of rural communities. The performance of the GSEs in the home willingness to purchase mortgages in commenters stated that Congress
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00024 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63603
charged the GSEs in their charters to Several commenters asked HUD to while Freddie Mac lagged the market in ‘‘promote access to mortgage credit extend the counting for the home all three goals-qualifying categories. The throughout the Nation (including purchase subgoal to rural areas even Department’s analysis reveals that there central cities, rural areas, and though data for rural areas is sparse. is ample room for both Fannie Mae and underserved areas).’’ One commenter HUD disagrees. Although HMDA data Freddie Mac to improve their stated that the lack of detailed HMDA for rural areas has improved, it is still performance in purchasing home loans data in rural areas makes market size too incomplete to support extending the that qualify for the Housing Goals, estimates difficult, but suggested that counting system. Alternative sources particularly in important market other data from private vendors could from private lenders are similarly segments such as the minority, first-time provide acceptable measures (without flawed. While HMDA’s reporting of homebuyer market. offering any specific sources). non-metropolitan areas has improved HUD’s Determination. Home purchase over the years, it continues to be Both GSEs’ funding of mortgages for is a high national priority. The unreliable. In 2001, 3,757 (3,280 of first-time homebuyers lags the market’s comments received and research which were small banks) of the 4,394 provision of funding for these families, reviewed document many studies non-metropolitan-area banks did not and the lag is particularly large for first- revealing the desire of Americans to report under HMDA. In that same year, time minority homebuyers. Table 2 own their own home. HUD finds that 324 (246 of which were small thrifts) of compares the GSEs’ funding of the proposed home purchase subgoal the 458 non-metropolitan-area thrift mortgages for first-time homebuyers furthers the statutory objectives of institutions did not report under with market loan originations for first- FHEFSSA. HUD set the level of the HMDA. time homebuyers. This table shows that home purchase subgoal prudently. Except for Fannie Mae’s recent first-time homebuyers represented 37.6 Details of HUD’s methodology are found performance in the Special Affordable percent of market loan originations, in Appendices A and D of this final rule and Low- and Moderate-Income compared with 26.5 percent of the and in chapter 3 of the Economic categories, the GSEs have lagged the GSEs’ purchases; thus, the GSEs fell Analysis that accompanies the rule. market in purchasing single-family, substantially short of the market Rather than distorting the market, the owner-occupied loans that qualify for originations ratio for first-time home purchase subgoal facilitates the the Housing Goals. In 2003, Fannie Mae homebuyers, over the period 1999– desire of many Americans to use the continued to lag the market in financing 2001. market to acquire their own home. properties located in underserved areas BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00025 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63604 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00026 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.002 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63605
For minority first-time homebuyers, each respective goal; no mention is continues to lag behind the primary the GSE ratio was 6.2 percent, compared made of the subgoals. However, despite market on all housing goal categories. to a market originations ratio of 10.6 the absence of any statutory requirement The subgoals will ensure that Freddie percent. For African-American and to consider the listed factors in setting Mac erases its gaps with the market and Hispanic first-time homebuyers, the the levels of the subgoals, HUD has takes a leadership position as well. The GSE ratio was 3.8 percent, compared to nevertheless carefully considered each type of improvement needed for Freddie a market originations ratio of 6.9 of these factors in setting the subgoal Mac to meet these new subgoals was percent. For first-time homebuyers, levels in this final rule. demonstrated by Fannie Mae during particularly first-time minority The following sections provide an 2001–2003. For example, Fannie Mae homebuyers, both GSEs substantially lag overview of HUD’s reasons for increased its low-mod purchases from the private conventional conforming establishing the Subgoals, which are 40.8 percent of its single-family-owner detailed in the Appendices to this rule. market. business in 2000 to 45.3 percent in 2002 As detailed in Appendix A to this (i) The GSEs Have the Ability To Lead to 47.0 percent in 2003. rule, evidence suggests that there is a the Market significant population of potential (iii) Disparities in Homeownership and homebuyers who are likely to respond The GSEs have the ability to lead the Credit Access Remain well to increased homeownership primary market for mortgages on single- opportunities produced by increased family owner-occupied properties, HUD notes that there remain GSE purchases in this area. Immigrants which are the GSEs’ principal line of troublesome disparities in our housing and minorities, in particular, are business. Both GSEs have long and mortgage markets, even after the expected to be a major source of future experience in the home purchase ‘‘revolution in affordable lending’’ and homebuyers. Furthermore, studies mortgage market, and therefore there is the growth in homeownership that has indicate the existence of a large no issue of the degree to which they taken place since the mid-1990s. As untapped pool of potential homeowners have penetrated this market. In noted previously in the discussion of among the rental population. Indeed, addition, because the Subgoals focus on the goals, the homeownership rate for the GSEs’ recent experience with new homeownership opportunities and, African-American and Hispanic outreach and affordable housing thus, do not include refinance loans, households remains 25 percentage initiatives confirms the existence of this there is no issue regarding potentially points below that of white households. potential. large year-to-year changes in refinance In 2002, the mortgage denial rate for The Department therefore is mortgage volumes, which affect the African-American borrowers was over establishing through this rule Subgoals magnitude of the denominator in twice that for white borrowers, even for home purchase loans that qualify for calculating performance percentages after controlling for the income of the the three Housing Goals to encourage under the Housing Goals, as borrower. the GSEs to take a leadership position experienced in the heavy refinance in creating homeownership financing years of 1998 and 2001–2003. HUD also notes that there is growing opportunities within the categories that Both GSEs have not only been evidence that inner city neighborhoods Congress expressly targeted with the operating in the single-family owner are not always being adequately served Housing Goals. mortgage market for years, they have by mainstream lenders. Some have been the dominant players in that concluded that a dual mortgage market b. HUD’s Determinations Regarding the market, funding 57 percent of mortgages has developed in our nation, with Home Purchase Subgoals on single-family owner-occupied conventional mainstream lenders Under FHEFSSA, HUD is authorized residences financed between 1999 and serving mainly white families living in to establish nonenforceable Subgoals 2002. As discussed in Section G of the suburbs and FHA and subprime within the Low- and Moderate-Income Appendix A to this rule, their lenders serving minority families Housing Goal and the Underserved underwriting guidelines are industry concentrated in inner city Areas Housing Goal. HUD also is standards and their AUS are widely neighborhoods. In addition to the authorized under FHEFSSA to establish used in the mortgage industry. unavailability of mainstream lenders, enforceable Subgoals within the Special families living in high-minority Affordable Housing Goal. The (ii) The GSEs’ Performance Relative to the Market neighborhoods generally face many Administration has proposed, as part of additional hurdles, such as lack of cash GSE regulatory reform, that Congress Even though the GSEs have had the for a downpayment, credit problems, authorize HUD to establish a separate ability to lead the home purchase and discrimination. Home Purchase Goal that would include market, their past average performance enforceable components. Pending the (1993–2003, 1996–2003, and 1999– Immigrants and minorities are enactment of any such legislation, HUD 2003) has been below market levels. projected to account for almost two- is establishing the Home Purchase During 2002 and 2003, Fannie Mae thirds of the growth in the number of Subgoals described in this final rule improved its performance enough to new households over the next ten years. under its current statutory authority. lead the special affordable and low-mod As emphasized throughout this HUD stated in the preamble to the markets for home purchase loans, but preamble and the Appendices to this proposed rule that in setting a subgoal, Fannie Mae continued to lag the rule, changing population demographics ‘‘[c]urrent law does not require that primary market in funding homes in will result in a need for the primary and HUD consider the statutory factors set underserved areas. The subgoals will secondary mortgage markets to meet forth in FHEFSSA prior to establishing ensure that Fannie Mae maintains and nontraditional credit needs, respond to or setting the level of Subgoals.’’ (69 FR further improves its above-market diverse housing preferences and 24244.) HUD’s interpretation of this performance in the special affordable overcome information and other barriers portion of FHEFSSA is unchanged. Each and low-mod markets, and also becomes that many immigrants and minorities of the subsections identifying the factors a market leader in funding underserved face. HUD finds that the GSEs must for consideration indicates that the areas. Freddie Mac, although it has also increase their efforts towards providing factors are to be considered in setting improved its recent performance, financing for these families.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00027 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63606 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
(iv) There Are Ample Opportunities for cities suggests that there will be c. Structure and Levels of the Home the GSEs To Improve Their Performance considerable growth in the origination Purchase Subgoals in the Home Purchase Market of CRA loans in urban areas. For banks and thrifts, selling their CRA Under this rule, performance on the Home purchase loans that qualify for Home Purchase Subgoals will be the Housing Goals are available for the originations will free up capital to make new CRA loans. As a result, the CRA calculated as Housing Goal-qualifying GSEs to purchase, which means they percentages of the GSEs’ total purchases can improve their performance and lead market segment provides an opportunity for the GSEs to expand their affordable of mortgages that finance purchases of the primary market in purchasing loans lending programs. As explained in single-family, owner-occupied for lower-income borrowers and Appendix A to this rule, Fannie Mae properties located in metropolitan areas, properties in underserved areas. Three and Freddie Mac have already started based on the owner’s income and the indicators of this have already been developing programs to purchase CRA- location of the property. Specifically, for discussed. type loans on a flow basis as well as each GSE the following Subgoals would First, the affordable lending market after they have seasoned. apply. (A ‘‘home purchase mortgage’’ is has shown an underlying strength over While the GSEs can choose any defined as a residential mortgage for the the past few years that is unlikely to strategy for leading the market, this purchase of an owner-occupied single- vanish (without a significant increase in leadership role can likely be family property.) interest rates or a decline in the accomplished by building on the many • economy). Since 1999, the shares of the 45 percent of home purchase initiatives and programs that the mortgages purchased by the GSE in home purchase market accounted for by enterprises have already started, the three Housing Goal categories are as metropolitan areas must qualify under including: (1) Their outreach to the Low- and Moderate-Income Housing follows: 16.3 percent for special underserved markets and their affordable, 31.4 percent for underserved Goal in 2005, with this share rising to partnership efforts that encourage 46 percent in 2006 and 47 percent in areas, and 44.1 percent for low- and mainstream lenders to move into these moderate-income. both 2007 and 2008; markets; (2) their incorporation of • Second, market share data reported in greater flexibility into their purchase 32 percent of home purchase section G of Appendix A to this rule and underwriting guidelines, (3) their mortgages purchased by the GSE in show that almost half of newly- development of new products for metropolitan areas must qualify under originated loans that qualify for the borrowers with little cash for a the Underserved Areas Housing Goal in Housing Goals are not purchased by the downpayment and for borrowers with 2005, with this share rising to 33 GSEs. As noted above, the situation is credit blemishes or non-traditional percent in both 2006 and 2007 and 34 even more extreme for special sub- credit histories; (4) their targeting of percent in 2008; and markets, such as the minority first-time important markets where they have had • 17 percent of home purchase homebuyer market where the GSEs have only a limited presence in the past, such mortgages purchased by the GSE in only a minimal presence. In terms of the as the markets for minority first-time metropolitan areas must qualify under overall mortgage market (both homebuyers; (5) their purchases of both the Special Affordable Housing Goal in conventional and government), the newly-originated and seasoned CRA both 2005 and 2006, with this share GSEs funded only 24 percent of all first- loans; and (6) their use of automated rising to 18 percent in both 2007 and time homebuyers and 17 percent of underwriting technology to qualify 2008. minority first-time homebuyers between creditworthy borrowers that would have Calculation of performance under the 1999 and 2001. Similarly, during the been deemed not creditworthy under Home Purchase Subgoals will be in same period, the GSEs funded only 40 traditional underwriting rules. terms of numbers of mortgages, not percent of first-time homebuyers in the The experience of Fannie Mae and conventional conforming market, and Freddie Mac in the subprime market numbers of units. This is consistent only 33 percent of minority first-time indicates that they have the expertise with the basis of reporting in HMDA homebuyers in that market. and experience to develop technologies data, which were HUD’s point of Finally, the GSEs’ purchases that can and new products that allow them to reference in establishing the Home count toward the Subgoal are not enter new markets in a prudent manner. Purchase Subgoal levels. HMDA data limited to new mortgages that are Given the innovativeness of Fannie Mae are reported in terms of numbers of originated in the current calendar year. and Freddie Mac, other strategies will mortgages in metropolitan areas. The GSEs can purchase loans from the be available as well. In fact, a wide These Home Purchase Subgoals are substantial, existing stock of affordable variety of quantitative and qualitative shown in Table 3, along with loans held in lenders’ portfolios, after indicators suggest that the GSEs have information on what the GSEs’ these loans have seasoned and the GSEs the expertise, resources and financial performance on the Subgoals would have had the opportunity to observe strength to improve their affordable have been if they had been in effect for their payment performance. In fact, lending performance enough to lead the 1999–2003 (under the proposed based on Fannie Mae’s recent home purchase market for special counting rules for 2005–2008). Table 3 experience, the purchase of seasoned affordable, low- and moderate-income, also presents HUD’s estimates of the loans is at present one strategy and underserved areas loans. The recent average shares of mortgages on owner- employed for purchasing Housing improvement in the affordable lending occupied single-family properties in Goals-qualifying loans and meeting the performance of the GSEs, and metropolitan areas that were originated goals. particularly Fannie Mae, further in 1999–2003 that would have qualified The current low homeownership rate demonstrates the GSEs’ capacity to lead for these Home Purchase Subgoals. of minorities and others living in inner the home purchase market. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00028 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63607
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00029 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.003
d. Counting Mortgages Toward the statutory factors in arriving at, and the reasonably be anticipated. HUD Home Purchase Subgoals comments received on, the new housing estimates that the low-and-moderate- The Department is amending 24 CFR goal level for the Low- and Moderate- income share of the market averaged 57 81.15 to add a new paragraph (i) that Income Housing Goal, which targets percent between 1999 and 2002. mortgages on housing for families with would clarify that the procedures in b. Past Performance of the GSEs Under § 81.15 generally govern the counting of incomes at or below the area median income. After consideration of these the Low- and Moderate-Income Housing home purchase mortgages toward the Goal Home Purchase Subgoals in §§ 81.12, factors, this final rule establishes the 81.13 and 81.14. The new paragraph goal for the percentage of dwelling units A number of changes in goal-counting provides, however, that the numerator to be financed by each GSE’s mortgage procedures were adopted as part of and denominator for purposes of purchases at 52 percent for 2005, 53 HUD’s Housing Goals final rule counting performance under the percent for 2006, 55 percent for 2007, published on October 31, 2000 (65 FR Subgoals are comprised of numbers of and 56 percent for 2008. 65044) (Housing Goals 2000 final rule). home purchase mortgages in Additional information analyzing Thus, it is necessary to provide metropolitan areas, rather than numbers each of the statutory factors is provided information using several different in Appendix A, ‘‘Departmental of dwelling units. Paragraph (i) also measures in order to track performance Considerations to Establish the Low- provides that, for purposes of on the Low- and Moderate-Income and Moderate-Income Housing Goal,’’ addressing missing data or information Housing Goal over the 1996–2003 and Appendix D, ‘‘Estimating the Size for each Subgoal, the procedures in period. Table 4 shows performance of the Conventional Conforming Market § 81.15(d) shall be implemented using under these measures.14 numbers of home purchase mortgages in for each Housing Goal.’’ BILLING CODE 4310–27–P metropolitan areas and not single- a. Market Estimate for the Low- and family, owner-occupied dwelling units. Moderate-Income Housing Goal 14 The Freddie Mac 2002 figures in Table 4 differ Finally, the new paragraph provides from the corresponding figures in Table 3 in HUD’s that where a single home purchase The Department estimates that Proposed Rule. Subsequent to publication of the mortgage finances the purchase of two dwelling units serving low- and Proposed Rule, HUD discovered that HUD had or more owner-occupied units, the moderate-income families will account credited some units toward Freddie Mac’s Low- and for 51–56 percent of total units financed Moderate-Income Housing Goal in 2002 that had mortgage shall count once toward each been previously counted toward the goal in 2001. Subgoal that applies to the GSE’s in the overall conventional conforming The units were associated with a large year-end mortgage purchase. mortgage market during the period 2005 Freddie Mac mortgage purchase transaction in through 2008. HUD has developed this 2002. Because HUD’s regulations prohibit double 5. Low- and Moderate-Income Housing range, rather than a specific point counting, HUD has recalculated Freddie Mac’s 2002 Goal, § 81.12 Low- and Moderate-Income Housing Goal estimate, to account for the projected performance. The recalculation also reflects This section discusses the effects of different economic and correction of some coding errors discovered in Department’s consideration of the affordability conditions that can HUD’s recent review.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00030 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63609
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00031 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.004 63610 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Specifically, the following changes terminated at the end of 2003) are not HUD is also increasing the amount of were made in counting procedures for applied. These figures are termed the the maximum allowed for affordability measuring performance on the Low- and ‘‘2001–2003 baseline assumptions.’’ For estimation for multifamily units. Moderate-Income Housing Goal for 1996–2000 these figures differ from the Beneath the 2001–2003 baseline 2001–2003. HUD: official performance figures because figures in Table 4 is another row of (1) Established ‘‘bonus points’’ they incorporate the revised counting figures designated ‘‘With 2005 (awarding double credit) for purchases procedures described under point (c), of low- and moderate-income mortgages above, which were not reflected in the Assumptions.’’ These figures show the on small (5–50 unit) multifamily official performance figures at that time. effects of applying 2000 Census data properties and, above a threshold level, For 2001–2003 both sets of figures and the new specification of MSAs mortgages on two-to-four unit owner- incorporate the revised counting released by OMB in 2003 to the occupied properties; procedures, but the baseline does not measurement of Low- and Moderate- (2) Established a ‘‘temporary incorporate the bonus points and the Income purchase percentages with the adjustment factor’’ (1.35 units credit, as Freddie Mac TAF. same counting rules that were used for revised by Congress for 2001–2003 from In terms of the 2001–2003 baseline the 2001–2003 baseline in Table 4. The HUD’s 1.2 unit credits in the Housing measure, both Fannie Mae’s and Freddie effect is to reduce the Goal-qualifying Goals 2000 final rule) that applied to Mac’s low- and moderate-income percentage by an average of 0.6 Freddie Mac’s purchases (but not performance reached its maximum in percentage points for Fannie Mae and Fannie Mae’s purchases) of low- and 2000 (Fannie Mae at 51.3 percent and 0.7 percentage points for Freddie Mac, moderate-income mortgages on large Freddie Mac at 50.6 percent). Baseline over the 1999–2002 period. (more than 50-unit) multifamily performance fell somewhat for both properties; and GSEs in 2001, 2002, and 2003. Fannie However, for 2003, the effects are just (3) Revised procedures that HUD had Mae’s baseline performance last year the opposite—these assumptions instituted regarding the treatment of exceeded the level attained in 1999, but increased Fannie Mae’s performance by missing data on unit affordability, the Freddie Mac’s performance fell to the 0.8 percentage point (from 48.7 percent use of imputed or proxy rents for lowest level since 1998. to 49.5 percent) and Freddie Mac’s determining goal credit for multifamily Overall, both GSEs’ performance performance by 0.3 percentage point mortgages, and the eligibility for goals exceeded HUD’s Low- and Moderate- (from 45.0 percent to 45.3 percent). The credit for certain qualifying government- Income Housing Goals by significant difference in the direction of this impact backed loans. margins in 1996–1999, and by wide between 1999–2002 and 2003 may be Based on the counting rules in effect margins in 2000. New, higher goals were due to the need to apply estimation at that time for 1996–2000, as shown established for 2001–2003, and despite techniques in 1999–2002 but not in under ‘‘official performance’’ for 1996– somewhat lower performance than the 2003. For 1999–2002 HUD had to 2000 in Table 4, Low- and Moderate- level attained in 2000, both GSEs’ estimate the effect based on data Income Housing Goal performance for official performance exceeded the new geocoded according to 1990 census tract Fannie Mae was consistently in the 44– goal levels in each year 2001–2003, with definitions, while for 2003 the data were 46 percent range over the 1996–1999 the inclusion of the bonus points and geocoded to 2000 census tracts. Further period, before jumping to a peak of 49.5 the TAF. insight will be provided by analysis of percent in 2000. Freddie Mac’s The decline in baseline performance data for 2004 and further years. performance started at a lower level, but in 2001–2003 can be attributed in large then increased in several steps, from measure to the mortgage refinance wave c. Low- and Moderate-Income Home 41–43 percent in 1996–1998 to 46.1 that occurred in those years. Fannie Purchase Subgoal percent in 1999, and a record level of Mae’s overall volume of mortgage 49.9 percent in 2000. That was the only purchases (in terms of numbers of The Department has determined to year prior to 2001 in which Freddie housing units) rose from 2.2 million in establish a Subgoal of 45 percent for Mac’s performance exceeded Fannie 2000 to 4.7 million in 2001, 6.4 million each GSE’s purchases of home purchase Mae’s performance on this goal. in 2002, and then to 10.1 million in mortgages on single-family owner- Based on the then current counting 2003. Similarly, Freddie Mac’s volume occupied properties in metropolitan rules, including the bonus points and rose from 1.6 million in 2000 to 3.3 areas which are for low- and moderate- TAF, as shown under ‘‘official million in 2001, 4.3 million in 2002, income families in 2005, with this performance’’ in Table 4, Low- and and then to 5.8 million in 2003. For Subgoal rising to 46 percent in 2006 and Moderate-Income Housing Goal each GSE the increase in volume each 47 percent in both 2007 and 2008. performance was 51.5 percent for year can be largely attributed to The purpose of this Subgoal is to Fannie Mae in 2001, 51.8 percent in increases in purchase volumes for encourage the GSEs to increase their 2002, and 52.3 percent in 2003. For refinance mortgages relative to home acquisitions of home purchase loans for Freddie Mac, performance was 53.2 purchase mortgages. For each GSE, the low- and moderate-income families, percent in 2001, 50.5 percent in 2002, fraction of mortgages that qualified as many of whom are expected to enter the and 51.2 percent in 2003. Low- and Moderate-Income was less for Immediately beneath the official Low- refinance mortgages than for home homeownership market over the next and Moderate-Income Housing Goal purchase mortgages. few years. Table 5 provides basic performance percentages in Table 4 are For 2005–2008, HUD is expanding the information on both the GSEs’ low-mod figures showing the GSEs’ low- and affordability estimation of units with performance and the primary market’s moderate-income purchase percentages missing affordability information. In low-mod performance for the years 1999 on a consistent basis for the entire addition to multifamily units, the GSEs to 2003. Since the same format will be 1996–2003 period. The assumptions will also be able to use estimates of followed for the other housing subgoals, used were the counting rules affordability for single-family rental several points are made about the established in HUD’s Housing Goals units with missing rents and owner- information in the Table 5, prior to 2000 final rule except that bonus points occupied units with missing borrower discussing the low-mod subgoal. and the Freddie Mac TAF (which were incomes for determining goal credit. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00032 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63611
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00033 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.005 63612 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Average Performance Data. In 0.2 percent (market without both B&C rule, discuss the reasons why the addition to individual year data, various and small loans) below peak market Department is establishing the Subgoal averages of annual performance are performance. The 46-percent subgoal for for low- and moderate-income loans, as provided at the bottom of Table 5 2006 would add one percentage point to follows: (1) The GSEs have the resources (1999–2003, 2001–2003, and 2002– these comparisons, while the 47-percent and the ability to lead the market in 2003); these averages provide a useful subgoal for 2007 and 2008 would add providing mortgage funding for low- context for examining the feasibility of two percentage points. For example, the and moderate-income families; (2) the subgoals and the degree to which 47-percent subgoal is approximately except for Fannie Mae’s recent they call for performance that is above three percentage points above 2002– performance, the GSEs have historically past market levels. This table provides 2003 average market performance, and (over periods such as 1993–2003, 1996– a picture of how much the low-mod 1.4 percent (market without B&C loans) 2003, and 1999–2003) not led the subgoal targets move the GSEs above to 1.8 percent (market without both B&C market, even though they have had the past market levels and how much of a and small loans) above peak market ability to do so; (3) troublesome stretch each subgoal will be for each performance. disparities in our housing and mortgage GSE (as compared with that GSE’s past markets indicate a continuing need for Low-Mod Subgoals Compared With Past performance). As will become clear Freddie Mac Performance. To reach the 45- increased GSE activity; and (4) there are below, Fannie Mae and Freddie Mac percent 2005 subgoal, Freddie Mac would ample opportunities for the GSEs to have shown different past performances, have to improve its performance by 3.0 improve their low- and moderate- which means that the subgoal targets percentage points over its 2001–2003 average income performance in the home will appear to have different impacts on low-mod performance of 42.0 percent, by 1.8 purchase market. these two institutions. percentage points over its 2002–2003 average Although single-family owner- Definitions of Primary Market. HUD’s low-mod performance of 43.2 percent, and by occupied mortgages comprise their basic market definition is the 0.8 percent over its previous peak principal line of business, Freddie Mac conventional conforming market performance of 44.2 percent in 2003. To has always lagged behind the primary reach the 47-percent subgoal, Freddie Mac without B&C loans; in other words, the would have to improve its performance by market in financing mortgages for low- A-minus loans in the subprime market 3.8 percentage points over its 2002–2003 and moderate-income families. Over the are included in the market definition average low-mod performance, and by 2.8 past three years Fannie Mae has closed but the more risky B&C portion is not percent over its previous peak performance. its historical gap with the market and included (see Appendix D of the final Low-Mod Subgoals Compared With Past now leads the primary market in rule for further discussion of this). In its Fannie Mae Performance. To reach the 45- funding mortgages for low- and report for Freddie Mac, ICF indicated percent 2005 subgoal, Fannie Mae would moderate-income families. Because that small loans (those less than have to improve its performance by 0.7 home purchase loans account for a $15,000) should be excluded from any percentage points over its 2001–2003 average major share of the GSEs’ purchases, the analysis that dealt with loans that might low-mod performance of 44.3 percent; Fannie establishment of this Subgoal will aid Mae would meet the 45-percent subgoal be available for purchase by the GSEs. based on its 2002–2003 average low-mod their performance under the overall Therefore, data are provided in Table 5 performance of 45.6 percent and its previous Low- and Moderate-Income Housing for (a) the market without B&C loans peak low-mod performance of 47.5 percent in Goal. and (b) the market without both B&C 2003. To reach the 47-percent subgoal, For the foregoing reasons, the and small loans less than $15,000. As Fannie Mae would have to improve its Department believes that the GSEs, and shown in Table 5, dropping small loans performance by 2.7 percent over its 2001– particularly Freddie Mac, can do more reduces the low-mod share of the 2003 average performance and by 1.4 to raise the share of their home loan conventional conforming market by percentage points over its 2002–2003 average purchases serving low- and moderate- about one-half percentage point. performance; Fannie Mae would meet the 47- income families. This can be Projected 2000-Based Data. Table 5 is percent subgoal based on its previous peak accomplished by building on efforts that based on projected data that performance of 47.5 percent in 2003. the enterprises have already started, incorporates both 2000 Census The low-mod subgoal targets will be including their new affordable lending geography and the new OMB more challenging for Freddie Mac than products, their many partnership efforts, definitions. Thus, the goals-qualifying Fannie Mae. The type of improvement their outreach to inner city percentages in this table differ from needed to meet the new low-mod neighborhoods, their incorporation of those reported earlier in this Preamble, subgoal targets was demonstrated by greater flexibility into their the latter being historical, 1990-Census- Fannie Mae during 2001–2003, as underwriting guidelines, and their based percentages. HUD had to Fannie Mae increased its low-mod purchases of seasoned CRA loans. A reapportion the data for the years prior purchases from 40.1 percent of its wide variety of quantitative and to 2003. For 2003, both HMDA and GSE single-family-owner business in 2000 to qualitative indicators indicate that the data were defined in terms of 2000 43.6 percent in 2002 to 47.5 percent in GSEs have the resources and financial Census geography, so no 2003, as shown in Table 5. The strength to improve their affordable reapportionment was necessary; for this approach taken is for the GSEs to obtain lending performance enough to lead the reason, the 2003 data are probably the their leadership position by staged market serving low- and moderate- most accurate. With these basics, the increases in the subgoals; this will income families. results for the low-mod subgoal can now enable the GSEs to take new initiatives be briefly summarized as follows: in a correspondingly staged manner to d. Summary of Comments Low-Mod Subgoals Compared With achieve the new subgoals each year. The majority of comments that Market. The 45-percent subgoal for the Thus, the increases in the housing addressed the housing goals focused on first year (2005) is approximately two subgoals are sequenced so that the GSEs the highest goal in year 2008 for the percentage points above 1999–2003 and can gain experience as they improve Low- and Moderate-Income Housing 2001–2003 average market performance, and move toward the new higher Goal. While some commenters, such as one percentage point above 2002–2003 subgoal targets. affordable housing policy advocacy average market performance, and 0.6 Section 4.b. above of this preamble, groups and housing and consumer percent (market without B&C loans) to and Section I.3 of Appendix A to this coalitions, expressed support for more
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00034 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63613
aggressive goals, stating that the goals stated that HUD should withdraw the performance in that market. A wide should be set to challenge the GSEs to home purchase subgoals or HUD should variety of quantitative and qualitative do more, most commenters expressed re-estimate the market using reasonable indicators demonstrate that the GSEs concerns about possible adverse affects assumptions and set both the goal and have the expertise, resources and on middle-income borrowers, including subgoal levels no higher than the financial strength to improve their low- the potential for higher costs and for midpoint of the resulting ranges. and moderate-income lending performance, including lending for low- unrealistic goals to lead to credit e. HUD’s Determination allocation to the lower end of the and moderate-income home purchases, housing market, thereby hindering the The Low- and Moderate-Income and achieve the levels of the goals being GSEs’ ability to serve all homebuyers. Housing Goal established in this final established. rule is reasonable and appropriate Other concerns included issues related 6. Central Cities, Rural Areas, and Other to HUD’s market share methodology having considered the factors set forth in FHEFSSA. For 2001–2003, HUD set Underserved Areas Housing Goal, analysis and the effects of single-family § 81.13 refinance loans in high refinance years the level of the housing goal on the GSEs’ ability to meet the higher conservatively, relative to the This section discusses the goals. Many commenters recommended Department’s market share estimates, in Department’s consideration of the that HUD exempt refinances from the order to accommodate a variety of statutory factors in arriving at, and the goals performance calculation. As economic scenarios. Moreover, current comments received on, the new housing described earlier in this rule, HUD is examination of the gaps in the mortgage goal levels for the Central Cities, Rural seeking public comments on how to markets, along with the estimated size Areas, and Other Underserved Areas address the effects of refinance loans of the market available to the GSEs, Goal, which focuses on areas currently when this annual volume is high. In demonstrate that the number of underserved by the mortgage finance addition, some expressed the belief that mortgages secured by housing for low- system. After consideration of the overly aggressive goals could weaken and moderate-income families is more factors and the comments received, this the FHA insurance program and could than sufficient for the GSEs to achieve final rule establishes the goal for the encourage over-investment in rental the new goal. percentage of dwelling units to be Therefore, having considered all the housing at a time when multifamily financed by each GSE’s mortgage statutory factors including housing vacancy rates are high. HUD has purchases at 37 percent in 2005, 38 needs, projected economic and percent in 2006 and 2007, and 39 addressed these concerns in earlier demographic conditions for 2005 to percent in 2008. sections of this final rule preamble. 2008, the GSEs’ past performance, the The 1995 final rule provided that Others felt that higher goal levels will size of the market serving low- and mortgage purchases count toward the encourage more investor-owned rental moderate-income families, and the Underserved Areas Housing Goal if such units that harm communities. Both GSEs’ ability to lead the market while purchases finance properties that are Fannie Mae and Freddie Mac objected maintaining a sound financial located in underserved census tracts. At to the higher goal level for the Low- and condition, HUD has determined that the 24 CFR 81.2 of HUD’s current Moderate-Income Goal. Each disputed annual goal for mortgage purchases regulations, HUD defines ‘‘underserved HUD’s market share analysis, citing the qualifying under the Low- and areas’’ for metropolitan areas (in central uncertainty of data, for example the size Moderate-Income Housing Goal will be cities and other underserved areas) as of the multifamily market, and the 52 percent for 2005, 53 percent in 2006, census tracts where either: (1) The tract uncertainty about future economic 55 percent in 2007, and 56 percent in median income is at or below 90 percent conditions. Freddie Mac stated that 2008. This reflects a reduction in the of the area median income (AMI); or (2) HUD overestimated the low/mod market upper end of the market share range the minority population is at least 30 share by 4 percent. Both GSEs also from 57 percent to 56 percent since percent and the tract median income is stated that it was inappropriate to base HUD’s publication of its proposed rule, at or below 120 percent of AMI. The the goals at the high end of market share resulting from changes in estimating AMI ratio is calculated by dividing the ranges. Freddie Mac stated that this market share as described at the end of tract median income by the MSA approach ignores the year-to-year section 3 (a), above, and in section F of median income. The minority variability of the market. Appendix D to Appendix D to this rule. percentage of a tract’s population is this rule responds to these market issues Further, the Department is calculated by dividing the tract’s raised by the GSEs. establishing a Subgoal for each GSE’s minority population by its total With regard to the Low- and purchases of home purchase mortgages population. For properties in non- Moderate-Income Home Purchase on single-family owner-occupied metropolitan (rural) areas, mortgage Subgoal, most commenters did not properties in metropolitan areas which purchases have counted toward the address the subgoal levels proposed by are for low- and moderate-income Underserved Areas Housing Goal where HUD, and none specifically addressed families of 45 percent in 2005, with this such purchases finance properties that the proposal levels for the Low- and Subgoal rising to 46 percent in 2006, are located in underserved counties. As Moderate-Income Subgoal. For those and 47 percent in both 2007 and 2008. discussed above under the heading that did mention the subgoals, the The reasons for increasing the Low- and ‘‘Definitions’’ in this final rule, HUD is comments were mixed with about half Moderate-Income Housing Goal are changing this specification from the supportive of the subgoal proposals in discussed in sections a and b, above, county level to the census tract level. general and half believing the subgoal and the reasons for establishing a Home Mortgages will count toward the levels were too high. Both GSEs Purchase Subgoal at the stated levels are Underserved Areas Housing Goal where commented on HUD’s proposed set forth in section c. such purchases finance properties that subgoals. Fannie Mae stated that the While the GSEs have lagged the are located in census tracts were either levels were higher than any values primary market in financing owner and (1) the median income in the tract does observed in HMDA from 1999–2002, rental housing for low- and moderate- not exceed 95 percent of the greater of and that the concept was duplicative of income families, they appear to have the median incomes for the non- the overall goal structure. Freddie Mac ample room to improve their metropolitan portions of the state or the
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00035 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63614 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
non-metropolitan portions of the nation properties located in underserved areas the projected effects of different as a whole, or (2) minorities comprise at for 2005, 38 percent for 2006 and 2007, economic and affordability conditions least 30 percent of the residents of the and 39 percent for 2008; (b) establishing that can reasonably be anticipated. HUD tract and the median income in the tract census tracts as the spatial basis for estimates that the underserved areas does not exceed 120 percent of the establishing whether properties in non- market averaged 39 percent between greater of the median incomes for the metropolitan (rural) areas count toward 1999 and 2002. non-metropolitan portions of the state or the Underserved Areas Housing Goal, in the non-metropolitan portions of the place of counties as in the definition b. Past Performance of the GSEs Under nation as a whole. stated above, for the reasons described the Underserved Areas Housing Goal The level for the Underserved Areas below; and (c) also establishing a As discussed above, a number of Subgoal of 32 percent of the total Housing Goal is based on 2000 Census changes in goal-counting procedures number of dwelling units financed by data on area median incomes and were adopted as part of HUD’s Housing each GSE’s purchases of home purchase minority percentages for census tracts, Goals 2000 final rule. Thus it is mortgages in metropolitan areas for MSAs, and the non-metropolitan necessary to provide information using properties located in underserved areas portions of states and of the entire several different measures in order to nation. HUD’s analysis, which is set of metropolitan areas for 2005, rising to 33 percent for 2006 and 2007, and 34 track changes in the GSEs’ performance forth below and described in greater on the Underserved Areas Housing Goal detail in Appendix B to this rule, is percent for 2008. A short discussion of the statutory over the 1996–2003 period. These are based on 2000 census data. The effect of 15 factors reviewed follows. Additional shown in Table 6. The same changes using 2000 census data rather than 1990 information analyzing each of the in counting rules described for the Low- data to determine whether areas are statutory factors is provided in and Moderate-Income Housing Goal are underserved increases the percentage of Appendix B to this rule, ‘‘Departmental applicable to the Underserved Areas the GSEs’ mortgage purchases in Considerations to Establish the Housing Goal. underserved areas by an estimated Underserved Areas Housing Goal,’’ and BILLING CODE 4210–27–P average of 5 percentage points for Appendix D to this rule, ‘‘Estimating the Fannie Mae and 4 percentage points for 15 Size of the Conventional Conforming The Freddie Mac 2002 figures in Table 6 differ Freddie Mac, based on the geographic Market for each Housing Goal.’’ from the corresponding figures in Table 4 in HUD’s locations of properties financed by the Proposed Rule. Subsequent to publication of the GSEs’ mortgage purchases in 1999 a. Market Estimate for the Underserved Proposed Rule, HUD discovered that HUD had through 2003. This change reflects Areas Housing Goal credited some units toward Freddie Mac’s geographical shifts in population Underserved Areas Housing Goal in 2002 that had The Department estimates that been previously counted toward the goal in 2001. concentrations by income and minority dwelling units in underserved areas will The units were associated with a large year-end status from 1990 to 2000. account for 35–39 percent of total units Freddie Mac mortgage purchase transaction in After analyzing the statutory factors, financed in the overall conventional 2002. Because HUD’s regulations prohibit double HUD is: (a) establishing a Goal of 37 conforming mortgage market during the counting, HUD has recalculated Freddie Mac’s 2002 percent for the percentage of the total period 2005 through 2008. HUD has Underserved Areas Housing Goal performance. The number of dwelling units financed by developed this range, rather than a recalculation also reflects correction of some coding errors discovered in HUD’s recent review. With the each GSE’s mortgage purchases for specific point estimate, to accommodate recalculation, Freddie Mac fell slightly short of its 2002 Underserved Areas Housing Goal.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00036 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63615
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00037 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.006 63616 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Based on the counting rules in effect performance than the level attained in been based on area median incomes and at that time, as shown under ‘‘official 2000 (largely due to the 2001–2003 area minority percentages from the 1990 performance’’ for 1996–2000 in Table 6, refinance wave), both GSEs’ Census. HUD applied the existing Underserved Areas Housing Goal performance exceeded the new Goal numerical thresholds for minority performance for Fannie Mae generally levels in 2001 and 2003; Fannie Mae percentages and median incomes to fluctuated between 27 and 29 percent also exceeded its goal in 2002, while 2000 Census data and ascertained that over the 1996–1999 period, before rising Freddie Mac fell slightly short. the proportion of underserved census to a peak of 31.0 percent in 2000. Appendix B to this rule includes a tracts and the proportion of housing Freddie Mac’s performance started at a comprehensive analysis of the GSEs’ units in underserved census tracts in lower level, but then increased in performance in funding mortgages for metropolitan areas both have increased several steps, from 25–26 percent in single-family-owner properties in significantly from 1990 levels: from 47.6 1996–1998, to 27.5 percent in 1999, and underserved areas. (The data reported percent to 51.3 percent of census tracts a record level of 29.2 percent in 2000. there are based on 2000 Census underserved and from 44.3 percent to Freddie Mac’s performance in 1999 was geography, which produces underserved 48.7 percent of population in the only year prior to 2001 in which it area figures slightly over five percentage underserved census tracts (including the exceeded Fannie Mae’s performance on points higher than 1990-based effects of the 2003 re-specification of this Goal. geography.) Both GSEs have lagged the Metropolitan Statistical Areas). Based on counting rules in effect for market in funding properties located in Comparable shifts at the county level 2001–2003, including the bonus points underserved neighborhoods. Between in non-metropolitan areas were found to and the TAF, as shown under ‘‘official 1999 and 2003, 28.3 percent of Freddie be of much smaller magnitude. Further, performance’’ in Table 6, Underserved Mac’s purchases of home loans financed HUD estimated the spatial distribution Areas Housing Goal performance for properties in underserved of GSE mortgage purchases across Fannie Mae was 32.6 percent in 2001, neighborhoods, as did 30.0 percent of metropolitan census tracts and non- 32.8 percent in 2002, and 32.1 percent Fannie Mae’s purchases—compared metropolitan counties for recent years. in 2003. Performance for Freddie Mac with 31.4 percent of home purchase The findings were that for 2000, 2001, was 31.7 percent in 2001, slightly less loans originated in the conventional 2002, and 2003, Fannie Mae’s than 31.0 percent in 2002, and 32.7 conforming market (excluding B&C performance figures are an estimated 7.2 percent in 2003. loans). Thus, Freddie Mac performed at percentage points, 6.0 percentage Immediately beneath the official 90 percent of the market level, while points, 5.5 percentage points, and 5.1 Underserved Areas Housing Goal Fannie Mae performed at 96 percent of percentage points higher in terms of performance percentages in Table 6 are the market level. In 2003, underserved 2000 Census geography than with 1990 figures showing the GSEs’ purchase areas accounted for 29.0 percent of Census geography. The corresponding percentages under this Goal on a Freddie Mac’s purchases, 32.0 percent figures for Freddie Mac are 5.6 consistent basis for the entire 1996– of Fannie Mae’s purchases, and 32.5 percentage points, 5.1 percentage 2003 period. The assumptions used percent of market originations. points, 5.1 percentage points, and 3.9 were the counting rules established in In evaluating the GSEs’ past percentage points larger, respectively. HUD’s Housing Goals 2000 final rule, performance, it should be noted that With a further shift to tract-based except that bonus points and the while borrowers in underserved definitions, the figures for Fannie Mae Freddie Mac TAF (which terminated at metropolitan areas tend to have much are reduced by 0.7 percentage point in the end of 2003) are not applied. These lower incomes than borrowers in other 2000, 2001, and 2002, and for Freddie figures are termed the ‘‘2001–2003 areas, this does not mean that GSE Mac by 0.7, 0.8, and 0.7 percentage baseline’’ assumptions. For 1996–2000 mortgage purchases in underserved point, respectively. The differences these figures differ from the official areas must necessarily be mortgages on between county-based performance and performance figures because they housing for lower income families. tract-based performance were much incorporate the revised counting Between 1999 and 2001, housing for smaller in 2003, with the latter falling procedures, which were not reflected in above median-income households below the former by only 0.2 percentage the official performance figures at that accounted for nearly 60 percent of the point for Fannie Mae and exceeding the time. For 2001–2003 both sets of figures single-family owner-occupied mortgages former by 0.1 percentage point for incorporate the revised counting that the GSEs purchased in underserved Freddie Mac last year. As previously procedures, but the baseline does not areas. noted in the discussion of the Low- and incorporate the bonus points and Beneath the 2001–2003 baseline Moderate-Income Housing Goals, the Freddie Mac TAF. figures in Table 6 are two additional smaller differences between these two In terms of the 2001–2003 baseline rows of figures designated ‘‘2005 approaches in 2003 than in 2000–2002 measure, both Fannie Mae and Freddie Assumptions.’’ These figures show the may be due to the need to apply Mac’s Underserved Areas Housing Goal effects of applying 2000 census data and estimation techniques in 2000–2002 but performance reached its maximum in the new specification of MSAs released not in 2003. 2000 (Fannie Mae at 31.0 percent and by OMB in 2003 to the identification of Freddie Mac at 29.2 percent) before underserved areas for purposes of c. Underserved Areas Home Purchase declining somewhat over the 2001–2003 measuring historical GSE goal Subgoal period. Both GSEs’ baseline performance. The second of the two The Department believes the GSEs performance in 2001–2003 exceeded the rows also incorporates the effects of the can play a leadership role in level attained in 1999. Department’s proposed change from underserved markets. To facilitate this Overall, both GSEs’ official counties to census tracts as the basis for leadership, the Department is performance exceeded their identifying underserved areas outside of establishing a Subgoal of 32 percent for Underserved Areas Housing Goal by metropolitan areas beginning in 2005. each GSE’s acquisitions of home significant margins in 1996–1999, and HUD’s determination of underserved purchase mortgages on properties by wide margins in 2000. New, higher areas for purposes of computing the located in the underserved census tracts Goals were established for 2001–2003, GSEs’ performance on the Underserved of metropolitan areas for 2005, rising to and despite somewhat lower Areas Housing Goal has, through 2003, 33 percent in 2006 and 2007, and 34
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00038 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63617
percent in 2008. The purpose of this Even though they have the ability to following points can be made about the Subgoal is to encourage the GSEs to lead the market, they have not done so, data presented in Table 7 regarding the improve their purchases of mortgages as both GSEs have lagged behind the underserved areas subgoal: for homeownership in underserved primary market in serving underserved BILLING CODE 4210–27–P areas, thus providing additional credit areas. As shown in Table 7, underserved and capital for neighborhoods that areas (based on 2000 Census geography) and 2002, HUD used various apportionment historically have not been adequately accounted for 29.4 percent of Freddie techniques to re-allocate 1990-based GSE and served. As discussed in Appendix A to HMDA data into census tracts as defined by the Mac’s purchases of home purchase 2000 Census. (Since 2003 HMDA and GSE data this rule, the GSEs have the ability to mortgages in 2003, 32.0 percent of were gathered in terms of 2000 Census geography, lead the primary market for single- Fannie Mae’’ purchases, and 32.5 no apportionment was required for that year.) family-owner loans, which is their percent of market originations.16 The Switching to the 2000-based tracts increases the ‘‘bread-and-butter’’ business. Both GSEs underserved area share of market originations by have been dominant players in the 16 HUD will begin defining underserved areas about five percentage points. Between 1999 and home purchase market for years, based on 2000 Census geography and new OMB 2002, 30.3 percent of mortgage originations funding 61 percent of the single-family- definitions of metropolitan areas in 2005, the first (without B&C loans) were originated in underserved year of the proposed rule. As explained in tracts based on 2000 geography, compared with owner mortgages financed between 1999 Appendix B of the proposed GSE Rule, the 2000- 25.2 percent based on 1990 geography. As shown and 2002. Through their many new based definition of underserved areas includes in Table B.8 of Appendix B of this Final Rule, the product offerings and their various 5,372 more census tracts in metropolitan areas than underserved areas share of each GSE’s purchases the 1990-based definition, which means the GSE- also rises by approximately five percentage points. partnership initiatives, the GSEs have market comparisons had to be updated to shown that they have the capacity to incorporate tract designations from the 2000 Thus, conclusions about the GSEs’ performance operate in underserved neighborhoods. Census. Therefore, for the years 1999, 2000, 2001, relative to the market are similar whether the analysis is conducted in terms of 2000 Census geography or 1990 Census geography.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00039 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63618 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00040 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02no04.007 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63619
Underserved Areas Subgoals As with the other two home purchase their underwriting guidelines, and their Compared With Market. The 32-percent subgoals, the underserved areas subgoal purchases of seasoned CRA loans. subgoal for the first year (2005) is targets will be more challenging for A wide variety of quantitative and approximately one percentage point Freddie Mac than Fannie Mae, qualitative indicators demonstrate that above 1999–2003 and 2001–2003 particularly given Freddie Mac’s low the GSEs have the resources and average market performance (based on performance (29.0 percent) during the financial strength to improve their the market defined without B&C and most recent year (2003). Again, the type affordable lending performance enough small loans) and approximately at the of improvement needed to meet the new to lead the market in underserved areas. 2002–2003 average market performance underserved areas subgoal targets was and the previous peak market demonstrated by Fannie Mae during d. Summary of Comments performance. The 33-percent subgoal for 2001–2003, as Fannie Mae increased its The Department received no 2006 and 2007 would add one underserved areas purchases from 29.0 comments that specifically addressed percentage point to these comparisons, percent of its single-family-owner the level of the Underserved Areas Goal. while the 34-percent subgoal for 2008 business in 2000 to approximately 32 The majority of commenters that offered would add two percentage points. For percent in both 2002 and 2003. As noted opinions on the level of the housing example, the 34-percent subgoal is above for the low-mod subgoals, staged goals focused on the high year (2008) of approximately three percentage points increases in the underserved areas the Low- and Moderate-Income Goal. above both 1999–2003 and 2001–2003 subgoal enable the GSEs to obtain their Where commenters did mention the average market performance, 1.8 percent leadership position by gaining Underserved Area Goal, their remarks (market without B&C loans) to 2.4 experience as they improve and move were in the context of better targeting percent (market without both B&C and toward the new higher subgoal targets. through changes in the definition of small loans) above 2002–2003 average The type of improvement needed to underserved areas. HUD also received meet this new underserved area subgoal market performance, and 1.5 percent no comments specific to the was demonstrated by Fannie Mae (market without B&C loans) to 1.8 Underserved Area Home Purchase during 2001 and 2002. During 2001, percent (market without both B&C and Subgoal. Both Fannie Mae and Freddie underserved area loans declined as a small loans) the market’s previous peak Mac commented on the level of the percentage of primary market performance in 2003. Underserved Area Goal. Fannie Mae originations (from 31.7 to 30.7 percent), stated that its replication of HUD’s Underserved Areas Subgoals Compared but they increased as a percentage of market sizing assumptions did not With Past Freddie Mac Performance. To Fannie Mae’s purchases (from 29.0 to reach the 32-percent 2005 subgoal, Freddie justify an Underserved Area Goal of 38 29.8 percent); and during 2002, they Mac would have to improve its performance or 40 percent. For example, Fannie Mae increased further as a percentage of by 2.7 percentage points over its 2001–2003 noted that in reaching a goal level of 40 average underserved areas performance of Fannie Mae’s purchases (from 29.8 to 32.3 percent), placing Fannie Mae at the percent, HUD relied on the most 29.3 percent, by 1.6 percentage points over unlikely owner-occupied underserved its 2002–2003 average underserved areas market level. performance of 30.4 percent, and by 0.3 Section 4.b. above of this preamble share of 30 percent, a level reached only percent over its previous peak performance and Section I.4 of Appendix B to this once in the past 11 years. With respect of 31.7 percent in 2002. To reach the 34- rule discuss the reasons why the to the Underserved Area Subgoal, percent subgoal, Freddie Mac would have to Department is establishing a Subgoal for Fannie Mae stated generally that improve its performance by 3.6 percentage home purchase mortgages in subgoals risk unintended consequences points over its 2002–2003 average underserved areas, namely: (1) the GSEs and that HUD has proposed subgoals in underserved areas performance, and by 2.3 have the resources and the ability to excess of the opportunity and business percent over its previous peak performance. lead the market in providing funding in mix seen in the market. Freddie Mac As noted in Table 7, Freddie Mac’s underserved neighborhoods; (2) the commented in general that all the goals performance jumped from 27.3 percent in and subgoals were set beyond what the 2001 to 31.7 percent in 2002, only to fall back GSEs lag the underserved areas market, primary market is likely to originate. to 29.0 percent in 2003. Thus, the 32-percent even though they have the ability to subgoal for 2005 is three percentage points lead; (3) troublesome disparities in our With respect to the underserved areas above Freddie Mac’s most recent experience housing and mortgage markets indicate market share, Freddie Mac estimates (29.0 percent). However, as noted above, a continuing need for increased GSE that the core ranges are 3–4 percentage Freddie Mac’s 31.7-percent performance in activity; and (4) there are ample points below the upper limits of the 2002 is only 0.3 percentage points below the opportunities for the GSEs to improve Department’s projected ranges. 32-percent subgoal for 2005. their underserved area performance in e. HUD’s Determination Underserved Areas Subgoals Compared the home purchase market. With Past Fannie Mae Performance. To reach Although single-family owner- The Underserved Areas Housing Goal the 32-percent 2005 subgoal, Fannie Mae occupied mortgages are the GSEs’ established in this final rule is would have to improve its performance by 0.6 percentage points over its 2001–2003 principal line of business, the GSEs reasonable and appropriate having average underserved areas performance of have lagged behind the primary market considered the factors set forth in 31.4 percent; Fannie Mae would meet the 32- in financing properties in underserved FHEFSSA. For 2001–2003, HUD set the percent subgoal based on its 2002–2003 areas. For the foregoing reasons, HUD level of the housing goal conservatively, average underserved areas performance of believes that the GSEs can do more to relative to the Department’s market 32.2 percent and its previous peak raise the share of their home loan share estimates, in order to underserved areas performance of 32.3 purchases in underserved areas. This accommodate a variety of economic percent in 2002. To reach the 34-percent can be accomplished by building on scenarios. Moreover, current subgoal, Fannie Mae would have to improve efforts that the GSEs have already examination of the gaps in the mortgage its performance by 2.6 percent over its 2001– 2003 average performance, by 1.8 percentage started, including their new affordable markets, along with the estimated size points over its 2002–2003 average lending products, their many of the market available to the GSEs, performance, and by 1.7 percent over its partnership efforts, their outreach to demonstrate that the number of previous peak performance of 32.3 percent in inner city neighborhoods, their mortgages secured by housing in 2003. incorporation of greater flexibility into underserved areas is more than
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00041 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63620 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
sufficient for the GSEs to achieve the improve their low- and moderate- mortgage purchases that are for new goal. income lending performance, including multifamily housing in 2005–2008. Therefore, having considered all the lending for home purchases in A short discussion of the statutory statutory factors including housing underserved areas, and achieve the factors for establishing the Special needs, projected economic and levels of the goals being established. Affordable Housing Goal follows. demographic conditions for 2005 to Additional information analyzing each 2008, the GSEs’ past performance, the 7. Special Affordable Housing Goal, § 81.14 of the statutory factors is provided in size of the market serving low- and Appendix C, ‘‘Departmental moderate-income families, and the This section discusses the Considerations to Establish the Special Department’s consideration of the GSEs’ ability to lead the market while Affordable Housing Goal,’’ and statutory factors in arriving at, and the maintaining a sound financial Appendix D, ‘‘Estimating the Size of the comments received on, the new housing condition, HUD has determined that the Conventional Conforming Market for goal level for the Special Affordable annual goal for mortgage purchases each Housing Goal.’’ qualifying under the Underserved Areas Housing Goal, which targets mortgages Housing Goal will be 37 percent for on housing for very low-income families a. Market Estimate for the Special 2005, 38 percent for 2006 and 2007, and and low-income families in low-income Affordable Housing Goal 39 percent for 2008. areas. After consideration of these Further, the Department is statutory factors and the comments The Department estimates that establishing a Subgoal of 32 percent for received, this final rule establishes the dwelling units serving very low-income each GSE’s acquisitions of home goal for the percentage of dwelling units families and low-income families living purchase mortgages on properties to be financed by each GSE’s mortgage in low-income areas will account for located in the underserved census tracts purchases at 22 percent in 2005, 23 23–27 percent of total units financed in of metropolitan areas for 2005, rising to percent in 2006, 25 percent in 2007, and the overall conventional conforming 33 percent in 2006 and 2007, and 34 27 percent in 2008. mortgage market during the period 2005 percent in 2008. This reflects a After analyzing the statutory factors, through 2008. HUD has developed this reduction in the upper end of the HUD has determined to establish: (a) a range, rather than a point estimate, to market share range from 35 percent to Goal of 22 percent for the percentage of account for the projected effects of 34 percent since HUD’s publication of the total number of dwelling units different economic conditions that can its proposed rule, resulting from financed by each GSE’s mortgage reasonably be anticipated. HUD also changes in estimating market share as purchases that are for special affordable estimates that the special affordable described at the end of Section 3.a. housing, affordable to very low-income market averaged 28 percent between above, and in Section G of Appendix D families and families living in low- 1999 and 2002. to this rule. income areas for 2005, rising to 23 b. Past Performance of the GSEs under The reasons for increasing the percent in 2006, 25 percent in 2007, and the Special Affordable Housing Goal Underserved Areas Housing Goal are 27 percent in 2008; (b) a Subgoal of 17 discussed in Sections a. and b. above, percent of the total number of each As discussed above, a number of and for establishing a Home Purchase GSE’s purchases of home purchase changes in goal-counting procedures Subgoal at the stated levels in section c. mortgages in metropolitan areas that are were adopted as part of HUD’s Housing While the GSEs have lagged the primary for housing affordable to very low- Goals 2000 final rule. Thus, it is market in funding loans in underserved income families and low-income necessary to provide information using areas, they appear to have ample room families in low-income areas for 2005 several different measures in order to to improve their performance in that and 2006, rising to 18 percent in 2007 track changes in performance on the market. A wide variety of quantitative and 2008; and (c) a Subgoal of 1 percent Special Affordable Housing Goal over and qualitative indicators demonstrate of each GSE’s combined annual average the 1996–2003 period. These are shown that the GSEs have the expertise, mortgage purchases in 2000, 2001, and in Table 8.17 resources, and financial strength to 2002, for each GSE’s special affordable BILLING CODE 4210–27–P
17 The Freddie Mac 2002 figures in Table 8 differ from the corresponding figures in Table 5 in HUD’s Proposed Rule. Subsequent to publication of the Proposed Rule, HUD discovered that HUD had credited some units toward Freddie Mac’s Special Affordable Housing Goal in 2002 that had been previously counted toward the goal in 2001. The units were associated with a large year-end Freddie Mac mortgage purchase transaction in 2002. Because HUD’s regulations prohibit double counting, HUD has recalculated Freddie Mac’s 2002 Special Affordable Housing Goal performance. The recalculation also reflects correction of some coding errors discovered in HUD’s recent review.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00042 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63621
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00043 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.008 63622 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Based on the counting rules in effect GSEs maintain a consistent focus on special affordable market was only 52 at that time, as shown under ‘‘official serving the low- and very low-income percent. As noted above, Fannie Mae performance’’ for 1996–2000 in Table 8, portion of the housing market where led the primary market in funding Special Affordable Housing Goal housing needs are greatest. Appendices special affordable home loans during performance for Fannie Mae generally A and C to this rule use HMDA data and 2003. On the other hand, Freddie Mac fluctuated in the range between 14 and GSE loan-level data for home purchase continued to lag that market in 2003. 17 percent over the 1996–1999 period, mortgages on single-family owner- The data indicate that there is room for before rising to a peak of 19.2 percent occupied properties in metropolitan Freddie Mac to improve its performance in 2000. Freddie Mac’s performance areas to compare the GSEs’ performance in purchasing affordable home loans at started at a lower level, but then in special affordable lending to the the lower-income end of the market. increased in several steps, from 14–16 performance of depositories and other The rental market (including both 1- percent in 1996–1998 to 17.2 percent in lenders in the conventional conforming to 4-family rental properties and 1999, and to a record level of 20.7 market. There are two main findings percent in 2000. That was the only year with respect to the special affordable multifamily rental properties) is prior to 2001 in which Freddie Mac’s category. especially important in the performance exceeded Fannie Mae’s First, Freddie Mac and Fannie Mae establishment of the Special Affordable performance on the Special Affordable have historically lagged depositories Housing Goal for Fannie Mae and 9Housing Goal. and the overall market in providing Freddie Mac because of the relatively Based on counting rules in effect for mortgage funds for special affordable high percentage of rental units meeting 2001–2003, as shown under ‘‘official borrowers over periods, such as 1993– the Special Affordable Housing Goal. performance’’ in Table 8, Special 2003, 1996–2003, and 1999–2003. For example, between 1999 and 2002, Affordable Housing Goal performance Between 1993 and 2003, 12.2 percent of 51 percent of units financed by Fannie for Fannie Mae was 21.6 percent in Freddie Mac’s mortgage purchases were Mae’s rental mortgage purchases met the 2001, 21.4 percent in 2002, and 21.2 for special affordable borrowers, 13.3 Special Affordable Housing Goal, percent in 2003. Official performance percent of Fannie Mae’s purchases, 15.4 representing 46 percent of units counted for Freddie Mac was 22.6 percent in percent of loans originated by toward the Special Affordable Housing 2001, 20.4 percent in 2002, and 21.4 depositories, and 15.5 percent of loans Goal, during a period when rental units percent in 2003. originated in the conventional represented only 18 percent of its total Immediately beneath the official conforming market (without estimated purchase volume. For Freddie Mac, 50 Special Affordable Housing Goal B&C loans). During the period between percent of units financed by rental performance percentages in Table 8 are 1999 and 2003, the GSEs’ performance mortgage purchases met the Special figures showing the GSEs’ special was approximately 90 percent of the Affordable Housing Goal, representing affordable purchase percentages on a market’special affordable loans 41 percent of units counted toward the consistent basis for the entire 1996– accounted for 15.1 percent of Fannie Special Affordable Housing Goal, during 2003 period. The assumptions used Mae’s purchases, 14.5 percent of a period when rental units represented were the counting rules established in Freddie Mac’s purchases, and 16.2 only 16 percent of its total purchase HUD’s Housing Goals 2000 final rule, percent of loans originated in the volume. except that bonus points and the conforming market. (See Table 9, which Freddie Mac TAF (which were is based on 2000 Census geography.) c. Special Affordable Home Purchase terminated at the end of 2003) are not Second, while both GSEs have Subgoal applied. These are termed the ‘‘2001– improved their performance over the The Department believes the GSEs 2003 baseline’’ assumptions. In terms of past few years, Fannie Mae has made can play a leadership role in the special more progress than Freddie Mac in this measure, both Fannie Mae and affordable market generally, and the erasing its gap with the market. During Freddie Mac’s special affordable home purchase special affordable 2003, the special affordable share of performance reached its maximum in market in particular. Thus, the Fannie Mae’s purchases was 17.7 2000 (Fannie Mae at 21.4, percent and Department is establishing a Subgoal of percent, which was above the market Freddie Mac at 21.0 percent) before 17 percent for each GSE’s purchases of share of 16.8 percent. In 2003, the declining somewhat in 2001, and then home purchase mortgages for special special affordable share of Freddie declining further in 2002 and 2003. affordable housing located in Both GSEs’ baseline performance in Mac’s purchases was 16.2 percent. Section G in Appendix A to this rule metropolitan areas for 2005 and 2006, 2003 exceeded the level attained in rising to 18 percent in 2007 and 2008. 1999. discusses the role of the GSEs both in Overall, both GSEs’ performance the overall special affordable market The purpose of this Subgoal is to exceeded HUD’s Special Affordable and in the different segments (single- encourage the GSEs to improve their Housing Goals by significant margins in family owner, single-family rental, and purchases of home purchase mortgages 1996–1999, and by wide margins in multifamily rental) of the special on special affordable housing, thus 2000. New, higher Goals were affordable market. The GSEs’ special expanding homeownership established for 2001–2003, and despite affordable purchases accounted for 41 opportunities for very-low-income somewhat lower performance than the percent of all special affordable owner borrowers and low-income borrowers in level attained in 2000 (largely due to the and rental units that were financed in low-income areas, including minority 2001–2003 refinance wave, as discussed the conventional conforming market first-time homebuyers who are expected under the Low- and Moderate-Income between 1999 and 2002. The GSEs’ 41- to enter the housing market over the Housing Goal), both GSEs’ performance percent share of the special affordable next few years. Table 9 provides exceeded the new Goal levels in 2001– market was below their 55-percent share information needed to compare the 2003. of the overall market. Even in the owner special affordable subgoal targets with The Special Affordable Housing Goal market, where the GSEs account for 61 past market and GSE performance. is designed, in part, to ensure that the percent of the market, their share of the BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00044 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63623
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00045 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02no04.009 63624 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Special Affordable Subgoals based on its peak performance of 17.7 omission indicates that to the extent Compared With Market. The 17-percent percent in 2003. that subgoals or subcategories are subgoal for the first year (2005) is As with the low-mod and promulgated for the Special Affordable approximately one percentage point underserved areas subgoals, the special Housing Goal, no bar exists to enforcing above the 1999–2003, 2001–2003, and affordable subgoal targets will be more them.’’ (60 FR 61860.) The 1995 2002–2003 average market performance. challenging for Freddie Mac than Housing Goals final rule established an The 17-percent subgoal is at the Fannie Mae. But, as with other goals, enforceable subgoal for multifamily previous peak market performance (the the type of improvement needed to meet mortgages within the Special Affordable 1999, 2000, and 2003 markets without the new special affordable subgoal Housing Goal; this subgoal has been in B&C loans were about 17 percent) or targets was demonstrated by Fannie Mae place each year since then. This final slightly below the previous peak market during 2001–2003, as Fannie Mae rule does not change this longstanding performance (based on 2003 market increased its special affordable agency interpretation. without both B&C and small loans). The purchases from 13.4 percent of its d. Special Affordable Housing Goal: 18-percent subgoal for 2007 and 2008 single-family-owner business in 2000, to Multifamily Subgoals would add one percentage point to these 15.8 percent in 2002, to 17.7 percent in figures. Thus, the 18-percent subgoal is 2003, as shown in Table 9. This subgoal Based on the GSEs’ past performance approximately two percentage points is designed to encourage Fannie Mae on the Special Affordable Multifamily above the 1999–2003, 2001–2003, and and Freddie Mac to lead the special Subgoals, and on the outlook for the 2002–2003 average market performance affordable market. As noted earlier, the multifamily mortgage market, HUD of approximately 16 percent. The 18- approach taken is for the GSEs to obtain proposed that these Subgoals be percent subgoal is one percentage point their leadership position by staged retained for the 2005–2008 period. above the previous peak market increases in the subgoals to enable the Unlike the overall Goals, which are performance (the 1999, 2000, and 2003 GSEs to gain experience as they improve expressed in terms of minimum Goal- markets without B&C loans were about and move toward the new higher qualifying percentages of total units 17 percent) or 1.5 percentage points subgoal targets. financed, these Subgoals for 2001–2003 above the previous peak market The section above on considerations and in prior years have been expressed performance based on the 2003 market in establishing the Low- and Moderate- in terms of minimum dollar volumes of without both B&C and small loans. Income Home Purchase Subgoal and Goal-qualifying multifamily mortgage Special Affordable Subgoals Section D of Appendix C to this rule purchases. Specifically, each GSE’s Compared With Past Freddie Mac further discuss reasons why the special affordable multifamily Subgoal Performance. To reach the 17-percent Department set the Subgoal for special is currently equal to 1.0 percent of its 2005 subgoal, Freddie Mac would have affordable loans. average total (single-family plus to improve its performance by 1.9 Both Fannie Mae and Freddie Mac multifamily) mortgage volume over the percentage points over its 2001–2003 questioned HUD’s authority under 1997–1999 period. Under the proposal, average special affordable performance FHEFSSA to establish any subgoals the GSEs’ purchases of mortgages of 15.1 percent, by 1.3 percentage points within the Special Affordable Housing financing dwelling units in multifamily over its 2002–2003 average special Goal. The GSEs noted that both sections housing for calendar years 2005–2008 affordable performance of 15.7 percent, establishing the Low- and Moderate- will be 1.0 percent of the GSEs’ average and by 0.8 percent over its previous Income and the Underserved Areas annual dollar volume of mortgage peak performance of 16.2 percent in Housing Goals include language that purchases in the calendar years 2000, 2003. To reach the 18-percent subgoal, HUD ‘‘may establish separate specific 2001, and 2002. The proposal would Freddie Mac would have to improve its subgoals within the goal under this increase the subgoal levels by roughly performance by 2.9 percentage points section and such subgoals shall not be 90 percent compared to their current over its 2001–2003 average special enforceable * * * .’’ No such language levels. Specifically, Fannie Mae’s total affordable performance, 2.3 percent over appears in the section establishing the eligible multifamily mortgage purchase its 2002–2003 average performance, and Special Affordable Housing Goal. The volume increased from $4.6 billion in by about 1.8 percent over its previous GSEs asserted that this omission is an 1993 to $12.5 billion in 1998, and then peak performance. indication that Congress intended to jumped sharply to $18.7 billion in 2001, Special Affordable Subgoals prohibit HUD from establishing any $18.3 billion in 2002, and $33.3 billion Compared With Past Fannie Mae subgoals within the Special Affordable in 2003. As shown in Table 8, special Performance. To reach the 17-percent Housing Goal. affordable multifamily mortgage 2005 subgoal, Fannie Mae would have HUD has also considered the GSEs’ purchases followed a similar path, to improve its performance by 0.9 claim that HUD lacks the statutory rising from $1.7 billion in 1993 to $3.5 percentage points over its 2001–2003 authority to impose any subgoals within billion in 1998 and $4.1 billion in 1999, average special affordable performance the Special Affordable Housing Goal. and also jumping sharply to $7.4 billion of 16.1 percent; Fannie Mae would These same arguments were presented in 2001, $7.6 billion in 2002, and $12.2 essentially meet the 17-percent subgoal by the GSEs during HUD’s 1995 billion in 2003. As a result of its strong based on its 2002–2003 average special rulemaking establishing the housing performance, Fannie Mae’s purchases affordable performance of 16.8 percent goals. (See Housing Goals 1995 have been at least twice its minimum and would surpass the 17-percent proposed rule published on February subgoal in every year since 1997—247 subgoal based on its peak special 16, 1995 at 60 FR 9154, and the final percent of the Subgoal in that year, 274 affordable performance of 17.7 percent rule published on December 1, 1995 at percent in 1998, 315 percent in 1999, in 2003. To reach the 18-percent 60 FR 1846.) 294 percent in 2000, and, under the new subgoal, Fannie Mae would have to At that time, HUD stated that the Subgoal level, 258 percent in 2001, 266 improve its performance by 1.9 percent absence of a similar subgoal provision percent in 2002, and 426 percent in over its 2001–2003 average performance under the Special Affordable Housing 2003. and by 1.2 percentage points over its Goal section ‘‘is not an indication that Freddie Mac’s total eligible 2002–2003 average performance; Fannie subgoals or subcategories within the multifamily mortgage purchase volume Mae would meet the 18-percent subgoal overall goal are prohibited; rather, such increased even more sharply, from $0.2
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00046 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63625
billion in 1993 to $6.6 billion in 1998, them to back away from this market, that market. The GSEs’ mortgage and then jumped further to $11.8 billion and that the situation had changed purchases between 1999 and 2002 in 2001, $18.3 billion in 2002, and $21.5 greatly since then. The organization accounted for 55 percent of the total billion in 2003. As shown in Table 8, stated that the overall goals now (single-family and multifamily) special affordable multifamily mortgage provided sufficient incentive for the conforming mortgage market, but they purchases followed a similar path, GSEs to focus on multifamily mortgage accounted for only 41 percent of the rising from $0.1 billion in 1993 to $2.7 purchases. One multifamily lender special affordable market. A wide billion in 1998, and also jumping expressed concern that increasing the variety of quantitative and qualitative sharply to $4.6 billion in 2001, $5.2 Multifamily Special Affordable Subgoal indicators demonstrate that the GSEs billion in 2002, and $8.8 billion in 2003. will push the GSEs to extend credit to have the expertise, resources, and As a result of its strong performance, unqualified borrowers with poor quality financial strength to improve their Freddie Mac’s purchases have also been properties that should not be eligible for special affordable lending performance at least twice its minimum Subgoal in long-term, low-cost financing. However, and to close their gap with the market. every year since 1998—272 percent of other commenters, including multiple Further, the Department is the Subgoal in that year, 228 percent in public advocacy groups and a local establishing a Subgoal of 17 percent for 1999, 242 percent in 2000, and, under government official, recommended that each GSE’s acquisitions of home the new Subgoal level, 220 percent in HUD increase the level of this subgoal. purchase mortgages for special 2001, 247 percent in 2002, and 417 Several commenters specifically affordable housing in 2005 and 2006, percent in 2003. recommended that HUD set this subgoal rising to 18 percent in 2007 and 2008. The Special Affordable Multifamily between 2.5 percent and 3 percent of the The special affordable home purchase Subgoals set forth in this final rule are GSEs’ purchases in preceding years. subgoal will ensure that Freddie Mac reasonable and appropriate based on the They noted that the GSEs have far improves its performance enough not Department’s analysis of this market. exceeded the subgoal levels in recent only to close its current gap with the The Department’s decision to retain years and said that a higher subgoal primary market but also to place itself these Subgoals is based on HUD’s level is needed to promote additional in a leadership position. The subgoal analysis, which indicates that multifamily lending. will also encourage Fannie Mae to multifamily housing still serves the improve further its current market- housing needs of lower-income families f. HUD’s Determination leading performance. A wide variety of and families in low-income areas to a HUD concludes that the Special quantitative and qualitative indicators greater extent than single-family Affordable Housing Goal established in demonstrate that the GSEs have the housing. By retaining the Special this final rule is reasonable and expertise, resources, and financial Affordable Multifamily Subgoal, the appropriate having considered the strength to improve their special Department ensures that the GSEs factors set forth in FHEFSSA. Current affordable lending performance, continue their activity in this market, examination of the gaps in the mortgage including lending for home purchases and that they achieve at least a markets, along with the estimated size for special affordable housing, and to minimum level of special affordable of the market available to the GSEs, achieve the levels of the subgoals being multifamily mortgage purchases that are demonstrates that the number of established. affordable to lower-income families. mortgages secured by special affordable Finally, the Department is housing is more than sufficient for the establishing each GSE’s Special e. Summary of Comments GSEs to achieve the new goal. Affordable Multifamily Subgoal at 1.0 Comments regarding the Special Therefore, having considered all the percent of its average annual dollar Affordable Goal were received from statutory factors including housing volume of total (single-family and numerous public advocacy groups and needs, projected economic and multifamily) mortgage purchases over one trade association; however, only demographic conditions, the GSEs’ past the 2000–2002 period. In dollar terms, one public advocacy group commented performance, the size of the market the level of the subgoal is $5.49 billion on the level of the goal. The serving low- and moderate-income per year in special affordable commenting group recommended that families, and the GSEs’ ability to lead multifamily mortgage purchases for the 2004 Special Affordable Goal be the market while maintaining a sound Fannie Mae and $3.92 billion per year maintained for the years 2005–2008. financial condition, HUD has in special affordable multifamily No comments specific to the Special determined that the Special Affordable mortgage purchases for Freddie Mac. Affordable Home Purchase Subgoal Housing Goal will be 22 percent for These Subgoals would be less than the were received from the public. Fannie 2005, 23 percent for 2006, 25 percent for actual special affordable multifamily Mae provided an analysis as part of its 2007, and 27 percent for 2008. This mortgage purchase volume in 2001– comments that illustrated, for the years reflects a reduction in the upper end of 2003 for both GSEs. Thus, the 1999 through 2002, that the market did the market share range from 28 percent Department believes that they would be not perform up to the level of HUD’s to 27 percent since HUD’s publication of feasible for the 2005–2008 period. proposed Special Affordable Home its proposed rule, resulting from HUD believes that the proposed Purchase Subgoal. changes in estimating market share as increase in the dollar level of the Regarding the Multifamily Special described at the end of section 3.a, Special Affordable Multifamily Subgoal Affordable Subgoal, neither GSE above, and in Section H of Appendix D balances the need to promote GSE objected to HUD’s proposed subgoal to this rule. activity in this segment with the need to levels for 2005–2008. One trade The reasons for increasing the Special provide some protection in the event of organization suggested that the subgoal Affordable Housing Goal are discussed a decline in overall mortgage market has outlived its original purpose and above in this preamble. Since the GSEs activity. Because this goal is set as a should be discontinued. This have historically lagged the primary dollar amount rather than as a share of organization stated that the subgoal was market in purchasing loans on owner business, overall declines in residential established to induce the GSEs to and rental properties that qualify as mortgage lending would make this goal purchase multifamily loans at a time special affordable, they have ample harder to achieve. Setting the subgoal when heavy credit losses had caused room to improve their performance in level based on the GSEs’ record
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00047 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63626 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
multifamily loan purchases during In view of the increasing use of loans Several other organizations endorsed 2000–2002 sets an appropriately high made without obtaining income a standardized procedure for estimating level for the next several years, in the information from the borrower, there is affordability for those units missing rent Department’s view. In recent years a question whether HUD’s existing or income data, including an Fannie Mae and Freddie Mac have each counting rules for missing-data econometrically based methodology. purchased multifamily mortgages in at situations are adequately reliable and Two commenters stated that HUD least twice the subgoal amount. The create no more than a negligible should require only actual data for increase in that subgoal dollar level statistical bias in the GSEs’ Housing determining whether a unit is affordable should serve to provide a more Goals performance figures relative to the or not. In addition, some commenters meaningful floor to the level of values that they would have if complete strongly recommended that HUD multifamily lending during the 2005– income data could be obtained, and disallow goals credit for all no- 2008 period. whether a more precise method for documentation subprime loans because imputing incomes could be employed. such loans are likely to be predatory. 8. Missing Data/No-Doc Loans For this reason, HUD requested HUD’s Determination. Having Overview. Accurate measurement of comments from the public about the considered the comments received, the GSEs’ performance under the three desirability and feasibility of HUD has determined that permitting Housing Goals depends on the implementing a standard some level of estimation for affordability completeness of data on borrower econometrically based method for data is reasonable and consistent with income (or, in the case of non-owner- imputing the income distribution of statutory intent that the GSEs serve the occupied units, the rent) and property mortgages purchased by each GSE that affordable housing needs of families location. With respect to property lack income data, based on known even if actual data are not available. location data, there was a less than one characteristics of the loan and the With regard to some commenters’ percent incidence of missing or census tract. objections that HUD should not permit incomplete geographical data between Summary of Comments. Fannie Mae the use of estimated data for—or even 2000 and 2002 for mortgages purchased supported expanding affordability allow goals credit for—any loans that by the GSEs. The incidence of missing estimation to single-family rental and were underwritten for approval without borrower income data has been owner-occupied goal performance borrower income data due to the greater—on the order of several percent calculations and favored a more potential for these loans to have each year. complex econometrically based predatory features, the Department does One reason for the increase in missing affordability estimation methodology. not find that these loans are inherently income data is the market’s recent For owner-occupied units Fannie Mae predatory in nature. Also, both GSEs increased use of mortgages, commonly suggested a method based on the have publicly announced that they will called low documentation (Low Doc) probability of mortgages/units not finance any loans with predatory and no documentation (No Doc) loans. qualifying for a goal based on census features, and the Department expects These loans do not require the borrower tract location. Fannie Mae stated that that they will continue to vigorously to provide income information. In some the multifamily affordability estimation enforce these policies. Accordingly, this cases, the borrower provides methodology could also be applied to final rule implements several changes to information on assets but not income single-family rental units. Fannie Mae the treatment of missing data. The first because of circumstances that make commented that if HUD were to adopt change amends § 81.15(d) of the General assets easier to document. In other an econometrically based methodology, Requirements to provide an alternative instances, mortgages are originated no limit should be placed on its treatment for single-family owner- entirely on the basis of a credit report, implementation. With the current occupied units where the mortgagor’s property appraisal, and cash for the methodology, Fannie Mae requested income is missing. As provided in downpayment. These mortgages that the limit for rental units be § 81.15(d), the GSEs may continue to typically require relatively large increased to 10 percent of total rental exclude such units from the downpayments and may also require a unit acquisitions. denominator as well as the numerator higher interest rate than fully Freddie Mac commented that HUD when they are located in census tracts documented mortgages. should adopt a simpler approach to with median income less than or equal The Housing Goals 2000 Final Rule missing data. For example, HUD should to area median income according to the provided that the GSEs may exclude allow the GSEs to remove units with most recent census, up to a ceiling of from the denominator owner-occupied missing incomes from the calculation of one percent of total eligible units. units which lack mortgagor income data the housing goals. Freddie Mac Purchases in excess of the ceiling will and which are located in low- or reasoned that the market numbers used be included in the denominator and moderate-income census tracts, i.e., in establishing the Housing Goals omit excluded from the numerator if they are tracts whose median income is no missing data and that omitting missing missing data. greater than the median income of the data from a GSE’s performance would However, in lieu of using this metropolitan area, or for properties be consistent. Also, Freddie Mac stated procedure, HUD is making available to located outside of metropolitan areas, that it historically has had a lower the GSEs in § 81.15(d) an alternative the larger of the median incomes of the missing data rate than the market and method for missing income treatment county or the statewide non- that it has sufficient business related that provides the GSEs with the ability metropolitan area (see 24 CFR incentives to reduce missing data. to apply a HUD-approved affordability 81.15(d)).18 Freddie Mac commented that any limits estimation methodology to all single- on adjustments for missing data should family owner-occupied units with 18 For rental units, the 2000 Housing Goals Final be related to overall missing data rates missing borrower income data up to a Rule also established counting rules that allow the in the market, estimation parameters specified maximum. This alternative GSEs to estimate rents or exclude units from the should be available at the beginning of provision specifies an approach that denominator when rent data are missing. See 24 CFR 81.15(e)(6)(i) on the rules applicable to the performance year, and estimation recognizes the distribution of borrower multifamily units and 24 CFR 81.15(e)(6)(ii) on the procedures should be simple and incomes within census tracts in rules for single-family rental units. straightforward to implement. determining how to treat loans with
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00048 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63627
missing income data. Goal-qualifying multifamily properties with missing 9. Double Counting of Seasoned units, by census tract, are estimated by rent data. This methodology is the same Mortgages multiplying the number of single-family methodology that has been used in past In addition to the preceding changes owner-occupied units with missing years to estimate affordability data for being made at this final rule stage, HUD borrower income information in multifamily properties with missing is making a technical change to properties securing mortgages rent data. § 81.16(c)(6) for purposes of clarity. purchased by the GSE, by the With regard to single-family one-to- Paragraph (c)(6) addresses the treatment percentage of all single-family owner- four unit rental properties financed with of seasoned mortgages. The paragraph, occupied units from originations that as currently codified, is a long one- would count toward achievement of the loans that are missing affordability data, the Department finds that a lack of data sentence paragraph. HUD believes that goal, as determined by HUD based on dividing this paragraph into two the most recent HMDA data available, should not act as a disincentive for the GSEs to serve markets that historically subparagraphs would improve for each census tract where the GSE comprehensibility and clarity. This acquired mortgage units. In establishing are important sources of affordable housing. Under HUD’s 2000 Rule, change is intended to clarify the the maximum number of units where restriction on double counting of § 81.15(e)(6)(ii) permits the GSEs to borrower income may be estimated seasoned mortgages in § 81.16(c)(6), i.e., exclude these units from both the under this alternative provision, HUD the restriction that prohibits the will apply two factors. The first of these numerator and the denominator when counting of a GSE’s purchase of a is the rate of missing borrower income neither income nor rental data are seasoned mortgage toward a goal where data for each census tract. This is available. While this provision does not such mortgage has already been counted calculated using HMDA data for the penalize the GSEs for financing these by the GSE toward the goal. This change most recent years for which comparable properties by requiring that they be makes clear that the restriction applies data are available. The second factor is counted in the denominator towards to all seasoned mortgages, regardless of the number of single-family owner- goal calculation, it also does not allow whether any other counting rules under occupied units purchased by a GSE them to obtain Housing Goals credit for § 81.16(c) also apply. Section 81.16(c)(6) during the performance year, by census financing mortgages that tend in this final rule reflects this technical tract. The maximum is calculated by disproportionately to serve affordable change. multiplying the HMDA percentage of housing. In this final rule, HUD is 10. Bulk Purchases/Counting of missing income data by the number of retaining the exclusion provision at Seasoned Loans units that a GSE purchased in each tract. § 81.15(e)(6)(ii) but is also adding an This number is summed up for all tracts alternative provision that will permit Overview. In its May 3, 2004, to obtain the overall nationwide the use of the same estimation proposed rule, HUD sought comment on maximum for that GSE. HUD will methodology now used for multifamily whether its current definition of a provide each GSE with a dataset ‘‘mortgage purchase’’ should be revised containing applicable tract-based loans with missing rent data. However, HUD is imposing separate maximum to ensure that transactions, especially HMDA missing income rates prior to the large transactions, are appropriately start of each year. The GSEs may choose rates for the new provision as follows: a 5 percent maximum on unseasoned counted under the law and in which provision of § 81.15(d) they will accordance with the purposes of single-family rental units originated in use in any year. However, they may not FHEFSSA and the GSEs’ charter acts. the current year and a 20 percent combine the options available under HUD also sought comment on whether this provision. If the maximum on maximum for seasoned loan units, that it should amend its counting rules at 24 missing single-family owner-occupied is, for loans that were originated more CFR 81.15 and 81.16 to ensure that the unit incomes is exceeded, the estimated than 365 days prior to the date of GSEs’ large-scale transactions further goal-qualifying units will be adjusted by acquisition by the GSE. HUD recognizes the requirements and purposes of the the ratio of the maximum amount the greater difficulty of obtaining rent Housing Goals. divided by the total number of units information on units from mortgages For example, HUD asked if with missing income information. originated a year or more prior to commenters believe the current Under each provision of § 81.15(d), acquisition by the GSE. Therefore, HUD counting rules are specific enough to units in excess of the specified is allowing the higher maximum on determine which seasoned mortgage maximum as well as units where affordability estimation for these units. transactions, including large-scale affordability information is not available As with the estimating provisions transactions, are substantially will remain in the denominator when permitted under § 81.15(d), the GSEs equivalent to mortgage purchases. HUD calculating goal performance. may use only one of the provisions sought these comments primarily in HUD is also in this final rule revising permitted under § 81.15(e)(6)(ii) in any response to certain large-scale § 81.15(e)(6) to change the current year. transactions of seasoned loans maximum on the use of HUD-approved undertaken by both GSEs in late 2003 In addition to the changes described multifamily rent estimation data from 5 for the purpose of meeting the 2003 herein, HUD is adding a provision to percent to 10 percent. In analyzing the Housing Goals. HUD questioned GSEs’ multifamily purchases for the §§ 81.15(d)(2)(i), 81.15(e)(6)(i) and (ii) whether such transactions furthered the past several years, HUD has determined that permits the use of such other data purposes of FHEFSSA, especially since that this change is statistically source or methodology as may be the transactions, including a transaction insignificant and will serve to promote approved by HUD. HUD is also between Freddie Mac and Washington further the financing of rental units that clarifying that owner occupied units Mutual Bank (WaMu), contained an would otherwise be eligible for credit that exceed the maximum established option for dissolution in the following under the Housing Goals. In this final under § 81.15(d)(2) for using any year. HUD sought public comment on rule, HUD is also specifying a estimation methodology will remain in its counting rules and definitions to methodology that may be used to the denominator of the respective goal ascertain the effect of the GSEs’ bulk estimate affordability data for calculation. purchases, including those with special
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00049 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63628 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
terms or conditions, on the market and Fannie Mae characterized the purchase volatile. In real time, it is extremely on affordable housing. of seasoned loans as an important difficult to predict the volume and Summary of Comments. HUD component of the liquidity of current ‘‘mix’’ or proportion of goals-eligible received several suggestions for revising mortgages. Knowing that there is a ready mortgages those markets will produce. its current definitions and counting market allows financial institutions to Market refinance forecasts for 2003 by rules. A trade association commented hold some of their assets in the form of Economy.com and Freddie Mac were off that HUD should specify the definition mortgages, and affords them the by over $2 trillion. Large transactions of of mortgage purchase so as not to count opportunity to sell these mortgages later mortgage purchases are essential transactions that are goals-oriented in to manage liquidity, improve because forecasts are not precise. form but not in substance. Some profitability, strengthen their capital With respect to its transaction in 2003 organizations commented that seasoned position, and manage certain risks. with WaMu, Freddie Mac stated that it loans should be excluded from counting In addition to the market benefits of engaged in this transaction because towards the goals altogether because seasoned mortgages, Fannie Mae also HUD took a number of steps to strongly they do not directly fund new housing discussed the practical relationship of encourage the GSEs to participate in the supply. Likewise, some commenters seasoned loan treatment and goals small 5–50 multifamily mortgage believed that these transactions are performance. The GSEs need bulk market, including bonus points. The contrary to the Charter requirement that purchases of seasoned loans to meet the GSEs can only purchase on terms that the GSEs provide assistance to the goals in years when the mix of business sellers are willing to accept. Freddie secondary market on an on-going basis. in the primary market deviates from the Mac further stated that goals that force One policy group asked that HUD business mix anticipated at the time the the GSEs to stretch their business mix exclude loans with recourse clauses goals were set. Fannie Mae pointed out in uncertain market conditions must because these purchases do not alleviate that HUD cited late-year purchases of eventually cause the GSEs to value some risk from the market. Other commenters seasoned loans in the proposed rule as mortgages more than sellers do. Under took the opportunity to request that the a useful method to meet the goals when these conditions, sellers will negotiate definitions and counting rules more market conditions change unexpectedly. for more favorable terms. Freddie Mac closely match CRA loan definitions. Fannie Mae also discussed the attributes stated that the seller ‘‘put’’ option in the These commenters did not suggest of dissolvable securities, stating that WaMu transaction and a similar specific regulatory language for the lenders sometimes request the option to transaction with Citibank exemplify definitions. dissolve securities swapped with the pro-seller terms and that these HUD also received comments that GSEs. Fannie Mae said that dissolution transactions advance the GSE’s supported counting bulk purchases that options are common terms in the regulatory purposes as well as meet the occur late in the year towards the goals. marketplace because dissolution options letter of the law. One trade association described the grant lenders greater control over their In response to concerns about the efficiencies gained from large-scale balance sheets, capital position, and options included in the swap, Freddie transactions. For example, the market other financial concerns. Fannie Mae Mac stated that ‘‘it is the GSE’s for multifamily units is large and indicated that lenders request these affordable housing goal requirements, fragmented, and seasoned portfolio options because they obtain more among other things, that give the sellers transactions are an efficient means for favorable rates and can make more the negotiating power to obtain such the GSEs to acquire smaller loans in the loans. options.’’ Both Fannie Mae and Freddie under 50-unit segment of the market. Freddie Mac made many of the same Mac concluded that HUD’s definition of Some commenters cautioned that points about bulk purchases of seasoned a mortgage purchase and the counting changing the definition of mortgage purchases as Fannie Mae and also rules should not be changed. purchase or the counting rules to clarify discussed its recent bulk transaction HUD’s Determination. HUD the treatment of large-scale seasoned with WaMu. For example, Freddie Mac considered the comments received, with mortgage transactions could have commented that bulk purchases and particular focus on the GSEs’ comments negative unintended consequences. dissolution options are common regarding transactions that include The GSEs responded to this issue industry practices. Freddie Mac also dissolution options. HUD is concerned with detailed comments. Fannie Mae stated that counting seasoned loans that transactions of this type, which stated that every mortgage purchase, increased the value and liquidity of both GSEs undertook in 2003 to achieve whether executed through flow, large or current loans. Knowledge that the GSEs their affordable housing goals, are not seasoned transactions, contributes to its stand ready to purchase mortgages fully consistent with the purposes of housing mission, and therefore, HUD under all market conditions gives other FHEFSSA, which are to award goals should not change the qualification of investors greater confidence because credit for mortgage purchases that mortgage purchases either for the size of they have a viable exit strategy when increase market liquidity for affordable the transaction or for the amount of providing funds to the real estate housing. When a seller can exercise its seasoning involved. Fannie Mae also market. option to reverse or unwind a stated that large-scale mortgage Freddie Mac indicated that bulk transaction and take back the mortgages purchases lower transactions costs for purchases are an essential means of within a specified time period, the both the buyers and sellers of mortgages. achieving the goals when market transaction appears temporary in nature, Some lenders offer to sell the GSEs conditions take an unexpected turn, and the liquidity that might result from mortgages on a flow basis, but others such as the conditions leading to its the transaction also appears transitory. prefer to bundle mortgages together and transaction with WaMu in 2003. Freddie The drafters of FHEFSSA intended sell to the GSEs from their portfolios. Mac pointed out that, unlike FHA, that the GSEs provide liquidity for Bulk transactions also serve the which can manage its business to the affordable housing where such liquidity business needs of lenders who do not cap on insurance commitments set would otherwise not exist or where it have a direct relationship with Fannie annually by Congress, Freddie Mac would be less reliable. HUD is aware Mae. Fannie Mae said that two-thirds of instead must respond to a dynamic that even short-term liquidity, as may its bulk purchases between 2001 and market in which the nature and occur with dissolution options, can be 2003 were not for seasoned loans. magnitude of loan originations are of value to mortgage sellers, especially
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00050 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63629
for balance sheet management or other defaults. The provision also does not regulations limits a GSE’s ability to purposes, but sellers seeking such apply to repurchase and resale request HUD to examine whether a options are generally not constrained in agreements where the GSE is the particular goal may be infeasible. If locating short-term liquidity solutions, purchaser of the security. Rather, the HUD determines that a GSE has failed especially when these solutions are transactions addressed by HUD’s to meet a housing goal, or that there is backed by seasoned mortgage loans. regulation provide, as a term of the a substantial probability that a GSE will Further, HUD believes that placing no transaction, the mortgage lender/seller— fail to do so, HUD must notify the GSE constraints on goals eligibility for and not the GSE—with the option of and provide an opportunity for the GSE transactions with dissolution options dissolving the transaction and having to respond. HUD must then determine would have the effect of encouraging the mortgages returned to the mortgage whether or not the goal was feasible. If transactions that are so short-term as to lender/seller. HUD determines that the goal was be dissolvable almost immediately after HUD believes the one-year lockout infeasible, then no further HUD action they have been counted towards the period will prevent potential misuse of to enforce the goal is authorized. housing goals. Such an outcome is these transactions but will still allow HUD’s proposed rule did not make clearly at odds with FHEFSSA. sellers of mortgages to manage their Therefore, HUD has determined to any changes to the process for portfolios in the medium and long term. determining whether a goal was or was amend its counting rules to provide that The limit on dissolution options applies for units acquired in transactions with not feasible. However, HUD still to all transactions because it is the received comments from both Fannie seller dissolution options to count potential for misuse, not the size of the toward the housing goals, such options Mae and Freddie Mac regarding those transaction that could conflict with provisions. must provide for a lockout period that FHEFSSA. HUD will continue to prohibits the exercise of the dissolution monitor the GSEs’ use of dissolution Summary of Comments. Fannie Mae option for at least one year from the date options to ensure that the one-year commented that ‘‘uncertainty regarding on which the transaction was entered minimum lockout requirement is HUD’s potential feasibility into and the transaction cannot be accomplishing its intended purpose. If determination would lead Fannie Mae dissolved during the one-year period. there is a question about whether a and Freddie Mac to engage in whatever The Secretary may grant an exception to particular transaction complies with the means necessary to meet the goals, the minimum lockout period, in one-year minimum lockout requirement, potentially resulting in market response to a written request from a HUD expects that the GSE will seek distortions.’’ Fannie Mae recommended GSE, if the Secretary determines that the clarification from HUD regarding the that the goals be set at levels that are transaction furthers the GSE’s statutory appropriate treatment of that transaction more likely to be seen in the purposes and the purposes of FHEFSSA. under the counting rules. marketplace, rather than at the high end Where a mortgage purchase involving a With regard to modifying its of market estimates. seller dissolution option has been definition of a ‘‘mortgage purchase,’’ Freddie Mac commented that an after- counted toward the housing goals under HUD has determined that defining the-fact finding of ‘‘infeasibility’’ or an a transaction subject to this provision, mortgage purchases in terms of market adjustment to the goals would not the transaction may not be dissolved effects would be cumbersome. The alleviate the burden imposed by (either by the exercise of the seller definition would have to be broad unreasonable goals. Freddie Mac noted dissolution option, or by separate enough to encompass all of the statutory that it is very difficult to estimate the agreement entered into by the GSE and purposes, including market liquidity size and composition (or ‘‘goal mix’’) of the seller) during the one-year minimum and market stability, and still narrow the mortgage market in advance. lockout period. If the seller of the enough to exclude transactions that are Freddie Mac also expressed concern mortgages and the GSE dissolve the legitimate in form but not in substance. that an after-the-fact feasibility transaction before that time, the Similarly, while some commenters determination would require HUD to transaction may no longer be counted suggested that HUD exclude seasoned second-guess innumerable business toward the housing goals and the GSE’s mortgages from its definition or that decisions made by the GSEs, with no performance must be adjusted in HUD impose a credit risk threshold for certainty as to how HUD would make accordance with this rule. The Department defines seller awarding goals credit, HUD believes such determinations. Finally, Freddie dissolution option as an option for a that these measures could have Mac stated that its reputation would seller of mortgages to the GSEs to unintended consequences that could suffer great harm during the time HUD dissolve or otherwise cancel a mortgage potentially harm market liquidity for considered its feasibility determination, purchase agreement or loan sale. The affordable housing. For example, HUD and that this harm could not be undone. Department, however, wishes to fully has encouraged the GSEs to buy HUD’s Determination. The final rule distinguish the arrangements seasoned portfolios of CRA loans as an does not make any changes to the established in these seller dissolution important source of liquidity for these process for determining whether a goal options from other types of agreements loans. is infeasible for a particular year. involving repurchases of securities or 11. Responses to Other Issues Raised by Although HUD has never had to make mortgages that involve the GSEs. For Commenters Relating to the Housing a determination that a goal is infeasible, example, the GSE, as seller of a security, Goals HUD believes that the process that is may agree to repurchase, or buy back, a currently in place provides an effective previously sold mortgage-backed a. Feasibility Determinations framework for making a timely security on a negotiated basis from the Overview. Section 1336(b) of determination of infeasibility. If in the holder of the security. HUD’s regulation FHEFSSA, together with HUD’s current future it is necessary to make a does not address that practice. Likewise, regulations, provides a process for determination of whether a goal is or it does not address arrangements determining that one or more goal levels was infeasible, HUD will make every whereby a mortgage lender agrees to are infeasible. This process may be effort to expedite the process in an effort repurchase or replace a mortgage upon initiated either by HUD or by a GSE; to minimize any potential costs and demand of the GSE if the mortgage nothing in FHEFSSA or in HUD’s uncertainty associated with the process.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00051 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63630 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
b. Specification of Underserved Areas 80 percent of area median income, between home improvement loans and Summary of Comments. Several while for CRA purposes, the definition small business loans secured by housing commenters suggested that HUD should of ‘‘low-income’’ is a borrower at or may not match HUD’s definitions of redefine the Underserved Areas Goal. A below 50 percent of area median mortgage purchases. In contrast, HUD consensus of these commenters stated income. Similarly, the affordable does not use a system of examiners to that lowering the tract income criteria housing goal definition of a ‘‘moderate determine the goals eligibility of sellers from 90 (120) percent to 80 (100) income’’ borrower is at or below 100 dealing with the GSEs, and comparison percent of area median income, while areas are established through regulation. percent would make the Underserved for CRA purposes, ‘‘moderate income’’ In light of these legal constraints, Areas Goal consistent with CRA. Several is defined as at or below 80 percent of HUD will not make any changes to the of the commenters also stated that the median area income. housing goals to address CRA concerns current definition is too broad and that Commenters pointed out that these at this time. lowering the tract income criteria to 80 definitional discrepancies create a d. Predatory Lending percent or 100 percent when the mismatch between the loans made by minority population is greater than 50 the primary market institutions and Summary of Comments. Certain percent (as opposed to 30 percent those purchased by the GSEs to meet commenters urged the Department to currently) of the tract would focus the affordable housing goals. The result is adopt predatory lending safeguards in goal on truly underserved areas. One that the GSEs can meet their goals by the final rule that would prohibit commenter suggested including a purchasing loans to borrowers in higher Housing Goals credit for purchases of borrower income criteria, such as less income ranges than those mandated loans that included mandatory than 80 percent of area median income, under CRA, which may result in less arbitration clauses or loans with in the Underserved Areas Goal to liquidity available to primary mortgage prepayment penalties beyond three further focus the goal on the market lenders to make additional low years towards the goals. The GSEs did underserved. and moderate income loans. not specifically mention this issue in HUD’s Determination. As discussed in These commenters recommended that their comments to HUD. HUD’s Appendix B to this rule, HUD has HUD find a way to resolve the apparent proposed rule did not suggest changes determined not to go forward with contradiction between the definitions. to its existing GSE regulations that redefining the Underserved Areas Goal One commenter suggested that HUD has address predatory lending practices. at this time. the authority to align the affordable HUD’s Determination. The Department continues to vigorously c. Reconciling the CRA and the housing goals with the CRA definitions oppose specific lending practices that Affordable Housing Goals without additional legislation. This commenter recommended that HUD are predatory or abusive in nature. As Summary of Comments. Several require the GSEs to report low-income stated in the 2000 rulemaking, the GSEs commenters from trade associations and loans in two categories—‘‘low income’’ should seek to ensure that they do not policy organizations suggested that HUD and ‘‘very low income’’—and conform purchase loans that actually harm could more sharply focus GSE activity the definitions of low-income and borrowers and support unfair lending on low- and moderate-income moderate income to the CRA practices. In that rulemaking, the homebuyers by encouraging greater definitions. Department determined that the GSEs purchases of CRA loans. According to Other commenters however, indicated should not receive the incentive of goals these commenters, this could be that legislative correction would be credit for purchasing high cost accomplished by establishing a new needed to accomplish such alignment. mortgages, including mortgages with CRA goal or by establishing CRA These commenters recommended that unacceptable features. subgoals under each of the current until that time, HUD should consult The Department is authorized under Housing Goals. with federal bank and thrift regulators to 24 CFR 81.16 to determine whether to The CRA requires depository determine the CRA-eligible market share provide full, partial, or no credit toward institutions to help serve the credit and adjust the affordable housing goals achievement of any of the housing goals needs of their communities and for Fannie Mae and Freddie Mac for any transaction. The Department’s authorizes federal regulators to examine accordingly. existing rules contain strong safeguards the level of lending, investment, and Several commenters recommended against abusive lending by excluding service that these institutions provide. that HUD should consider establishing certain types of mortgages from Commenters noted that under section specific ‘‘CRA loan sub-goals’’ under the counting towards the affordable housing 1335 of FHEFSSA, Fannie Mae and existing goals for the GSEs. One goals. These include loans with Freddie Mac are directed to ‘‘take commenter suggested that HUD could excessive fees, and prepayment affirmative steps to assist insured create a new goal that requires the GSEs penalties in certain loans. depository institutions to meet their to purchase stated amounts of CRA- The Department is aware that certain obligations under the CRA which shall eligible home purchase mortgages, with practices that were not enumerated in include developing appropriate and low and moderate income subgoals the regulations adopted in 2000, such as prudent underwriting standards, based on the CRA measures. loans with prepayment penalties after business practices, repurchase HUD’s Determination. After close three years and loans with mandatory requirements, pricing, fees, and review of this issue, HUD has arbitration clauses, often lock borrowers procedures.’’ These commenters noted, determined that full harmonization into disadvantageous loan products. The however, that under FHEFSSA, the between affordable housing goals and Department will rely on existing definitions for key categories of CRA definitions will require legislative regulatory authorities to monitor the borrowers served through affordable action. Income brackets for the goals GSEs’ performance in this area. Should housing goals differ from those under FHEFSSA and under CRA are the Department later determine that established for borrowers served under statutorily defined, and CRA definitions there is a need to specifically enumerate CRA. allow for much greater discretion by additional prohibited predatory Under FHEFSSA, the definition for examiners to determine CRA scoring. practices, it will address such practices ‘‘low income’’ is a borrower at or below For example, under CRA, the distinction at a future time.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00052 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63631
e. Minority Subgoals/Goals its regulations to incorporate certain consistent with the customary practice Summary of Comments. Among the data integrity procedures designed to of regulators of financial institutions, many suggestions HUD received for ensure the accuracy, completeness, and the Department proposed that the GSEs subgoals and bonus points, several timeliness of housing goal information be required to provide a certification advocacy groups recommended that submitted by the GSEs to the with their AHAR reports and such other HUD directly target minority mortgage Department. These procedures report(s), data submission(s), or purchases such as those made to Native included: (1) A requirement that the information for which HUD requests Americans. These groups note that GSEs provide a certification with their certification in writing. HUD proposed a homeownership rates are not equal Annual Housing Activity Reports certification that consisted of the across ethnic groups. Fewer Blacks and (AHAR) and such other reports, data following four parts: (1) The GSE Hispanics own their own homes than submissions, and information that the Certifying Official has reviewed the the general population. Although the Department may request in writing be particular AHAR, other report(s), data GSEs have made some progress in this certified; (2) a procedure to adjust submission(s) or information; (2) to the area, the GSEs are still less likely to current year-end errors, omissions, and best of the GSE Certifying Official’s serve high minority areas than other discrepancies in data submissions to knowledge and belief, the particular lenders. In the view of these HUD; and (3) a procedure for correcting AHAR, other report(s), data commenters, the absence of the GSEs prior year overstatements of submission(s), or information are performance due to reporting errors, current, complete, and do not contain has led to higher borrowing costs and omissions, or discrepancies in a GSE’s any untrue statement of a material fact; harsher borrowing terms for minority AHAR. HUD also restated in the (3) to the best of the GSE Certifying borrowers because they are more likely proposed amendment to § 81.102 the Official’s knowledge and belief, the to deal with nontraditional and enforcement options and remedies AHAR or other report(s), data predatory lenders. HUD’s Determination. Under under FHEFSSA and HUD’s regulations submission(s), and information fairly FHEFSSA, HUD does not have statutory that could result from a determination present in all material respects the authority to establish goals beyond that a GSE’s data submissions, GSE’s performance, as required to be information, or reports were not current, reported; and (4) to the best of the those enumerated in the statute. were incomplete, or otherwise Certifying Official’s knowledge and FHEFSSA directs HUD to establish a contained an untrue statement of belief, the GSE has identified in writing goal for underserved areas, and HUD’s material fact. any areas in which the GSE’s particular goal includes census tracts with high In addition to comments provided by AHAR, other report(s), data concentrations of minority households the GSEs, HUD received comments from submission(s), or information may differ (and with median income below a groups that included mortgage lenders, from HUD’s written articulations of its certain level) as one category of non-profit and policy advocacy counting rules including, but not underserved area. The statute does not organizations, and trade associations. limited to, the regulations under 24 CFR empower HUD to establish a goal based Most commenters supported the data part 81, and any other areas of on the characteristics of borrowers, verification provisions of the proposed ambiguity. other than by income of borrower. rule. However, one mortgage lender Summary of Comments. Fannie Mae Even without an explicit subgoal, stated that the proposed certification and Freddie Mac commented on this HUD believes that the goals structure would impose a severe burden on the proposal. Each expressed many similar will address the concerns of minority GSEs and lenders. Another suggested objections to the certification language borrowers. As discussed in the that the data integrity process should as proposed and offered many similar introduction, minorities and immigrants include some leeway for unintentional recommendations. For example, both are a growing percentage of homebuyers mistakes so that it does not become GSEs stated that the certification and many more aspire to home burdensome. A trade association stated language was overly broad and should ownership. Demographics dictate that that HUD should not enact regulations be modified to the form authorized in these buyers will become increasing that would put additional data integrity FHEFSSA for submissions to OFHEO; shares of the conventional conforming burdens on lenders. Fannie Mae and namely, that the report is true and market. Requiring the GSEs to lead the Freddie Mac provided detailed correct to the best of such officer’s market will encourage them to do even comments on each proposal. These knowledge and belief. Each more to reach out to minorities. comments are discussed more fully in recommended that the words ‘‘fairly f. Technical Change to § 81.16(c)(7) the following sections. present’’ be deleted from the third proposed certification statement stating In addition to the preceding changes 2. Independent Verification Authority— that these words are meaningful only in being made at this final rule stage, HUD § 81.102(a) the context of Generally Accepted is making a technical change to As it proposed, the Department is Accounting Practices (GAAP), which § 81.6(c)(7) to correct a cross-reference. retaining and recodifying the provisions defines standards of determining Paragraph (c)(7) addresses the treatment of the current § 81.102(a) that provide ‘‘fairness’’ in financial reporting, but not of refinanced mortgages. The paragraph that HUD may independently verify the performance reporting. includes a reference to § 81.14(f), which accuracy and completeness of data, In addition, both GSEs questioned is not related to refinanced mortgages. information and reports submitted by a HUD’s authority to impose a Section 81.16(c)(7) in this final rule is GSE in addition to the Department’s certification requirement, but stated that revised to correct this cross-reference. existing authority to conduct on-site to the extent HUD does impose this D. Subpart I—Other Provisions verifications and performance reviews. requirement, it should be the HUD is redesignating this section, as certification used by OFHEO. They also 1. Overview—Verification and HUD proposed, as § 81.102(a). stated that the phrases ‘‘errors, Enforcement To Ensure GSE Data omissions, discrepancies, and Integrity 3. Certification—§ 81.102(b) ambiguities,’’ ‘‘written articulations of HUD proposed to amend § 81.102 To ensure the highest degree of its counting rules,’’ and ‘‘any other areas (Independent Verification Authority) of corporate accountability, and to be of ambiguity’’ are vague and undefined,
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00053 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63632 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
and that this vagueness makes it The Department has also included internal review processes by the GSEs. possible for HUD to arbitrarily language to clarify that it may pursue A certification regarding the entire implement the certification provision by enforcement action against a GSE that report helps to ensure the GSEs’ interpreting it in a way that is not fails to provide the certification required accountability for the information that known by the GSEs. Freddie Mac also under § 81.102(b). In addition, the they are required to report accurately stated that HUD’s informal written Department may pursue enforcement under their charters. articulations are not enforceable and action if a GSE submits the certification Although Fannie Mae recommended that it may not know about all of HUD’s required under § 81.102(b), but the that the certification should apply only informal articulations. Both GSEs also Secretary later determines that the data, to the tables in the AHARs and Freddie stated that it is difficult to certify to the information or report(s) are not true, Mac recommended that the certification accuracy of information that must be correct and complete. For data, should apply only to the data tables in included in the reports that they receive information and report(s) subject to the AHAR and the loan-level data it from third parties. § 81.102(c) or (d), the final rule makes submits with its AHAR, from time to Freddie Mac suggested that the clear that the Department will only time HUD requires one or both GSEs to subject of the certification be limited to pursue enforcement action against a submit other report(s), information, or the year-end annual data tables and GSE in connection with material errors, data submission(s) that rise to a computerized loan-level data that it omissions or discrepancies, as those sufficient level of importance to HUD’s submits with its AHARs, and should not terms are defined therein. oversight work that a certification cover any narrative portions of the The GSEs have asserted that HUD statement is warranted. The final rule, AHARs. Fannie Mae suggested that the may not require certification of any therefore, retains this provision and certification should focus on the process information they submit because the further provides that the Secretary will it follows for generating its submissions Department has no express statutory issue a written notification to the GSE and should cover only the final tables in authority to do so. The Department’s whenever such a certification is the AHAR that it submits each year. authority to require certification of required. HUD expects that any Both GSEs stated that no certification information submitted by the GSEs is additional certification requirements should be required for reports-in- authorized under HUD’s general will be the exception rather than the progress, such as the housing goals regulatory power over the GSEs under rule to ensure that the routine and progress reports each submits to HUD section 1321 of FHEFSSA as well as its necessary flow of information is not on a quarterly basis. authority to monitor and enforce the impeded. A policy advocacy group commented GSEs’ compliance with the housing Both GSEs recommended that HUD that the certification should be limited goals under section 1336. (See the not impose a certification on any to reporting processes of the GSEs, not preamble of HUD’s proposed rule at 69 progress reporting, such as the quarterly the accuracy of the underlying data FR 24247–24248 for a full discussion of housing goals performance reports each obtained from individual lenders. A HUD’s authority to require certification.) submits to HUD. HUD did not propose trade association commented that HUD In requiring this certification, HUD is that such reports be certified and should not put additional data integrity fully aware that the GSEs collect reiterates that certification statements burdens on lenders. millions of data elements from will not be required for the GSEs’ first HUD’s Determination. HUD has hundreds of sources and that the GSEs three quarterly housing goals reports considered the comments received and must depend upon these sources to and any other report(s), data has determined to modify its proposal. provide accurate data. In requiring a submission(s) or information that HUD’s reasons for requiring a certification, HUD intends that the GSEs represent incomplete ‘‘snapshots’’ or certification were not disputed by will use and rely upon their internal information that is being gathered but commenters. However, HUD has revised controls and other due diligence which is not in final form. Certification the proposed rule language to address processes and procedures for collecting, will be required for the fourth quarter commenters’ concerns regarding clarity. compiling, verifying the accuracy of, report, i.e., the Annual Mortgage Report. HUD has also included alternative and reporting the data received from 4. Adjustment To Correct Current Year- language in the final rule that would sellers. HUD will evaluate the end Errors, Omissions or specifically define terms as well as sufficiency of this certification Discrepancies—§ 81.102(c) eliminate the language that the GSEs beginning with the 2005 fourth quarter and others found to be ambiguous. As Annual Mortgage Report and the AHAR HUD routinely conducts a result, the final rule includes a to determine whether it is serving its computerized consistency checks of simplified certification that is much function of providing adequate loan-level data received from the GSEs closer to the certification used by assurance as to the accuracy and as part of their AHAR reporting. This OFHEO. Section 81.102(b) has been completeness of information. data are received on March 15th of each amended to require the senior officer of With respect to the scope of the year for the previous year’s each GSE who is responsible for certification, HUD believes it is performance. These reviews verify that submitting to HUD the fourth quarter appropriate and reasonable that the the GSEs have applied HUD’s counting Annual Mortgage Report and the AHAR certification statement apply to the rules and goals eligibility standards under sections 309(m) and (n) of the entire AHAR submission, including the appropriately in determining their year- Fannie Mae Charter Act or sections narrative text, data tables, and end performance. A key procedure 307(e) and (f) of the Freddie Mac Act, computerized loan-level data. Section involves applying HUD’s counting rules as applicable, or for submitting to the 309(n) of Fannie Mae’s Charter Act and to the GSEs’ loan-level data for the Secretary such other report(s), data section 307(f) of the Freddie Mac Act purposes of replicating the performance submission(s), or information for which specify the types of information each figures computed by the GSEs in their certification is requested in writing by GSE is required to report, including AHARs. Also, in conjunction with other the Secretary to state that: ‘‘To the best narrative descriptions as well as data. reports provided by the GSEs, including of my knowledge and belief, the HUD expects that all of the required a report that reconciles all adjusted information provided herein is true, information, not just the data and data mortgage purchases (the denominator) correct and complete.’’ tables, will be subjected to appropriate with the GSE’s total business volume as
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00054 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63633
reported in the annual report to give the public the ability to comment HUD’s Determination. HUD has shareholders or other information on the GSEs’ AHARs. considered the comments and filings, HUD’s reviews also verify the Both GSEs commented in detail on determined that a provision specifying completeness of the data. If HUD finds HUD’s proposal. Both expressed what procedures HUD will use in discrepancies between its results and concerns about the scope of this developing its official performance those reported by the GSEs, HUD works provision and questioned what numbers for the immediately preceding with appropriate GSE staff to resolve the procedures, especially adjustment year is necessary. HUD notes that many discrepancies after which HUD makes a notification and enforcement of the concerns expressed by final determination of year-end results procedures, would be associated with commenters, especially the GSEs, and publishes these as HUD’s official its implementation. Freddie Mac involve the lack of definition of the performance figures for the year. augmented its comments with a legal terms ‘‘errors, omissions or HUD’s proposed rule provided for a opinion from outside counsel. discrepancies’’ and a lack of clarity timeframe within which the GSEs may With respect to the words ‘‘errors, regarding how the regulation will be comment or otherwise respond to omissions and discrepancies,’’ the GSEs enforced. Accordingly, in the final rule, HUD’s findings of errors, omissions, or contended that these terms were vague HUD has added a paragraph that defines discrepancies with additional and needed further definition. Freddie an ‘‘error’’ as a technical mistake, such information. If a GSE did not respond Mac stated that without such further as a mistake in coding or calculating with information to correct or explain definition, HUD could disallow data. Mistakes of this type may also the error, omission, or discrepancy to counting of units based upon include, but not be limited to, systems HUD’s satisfaction within five working interpretations of its rules of which errors, such as those related to days of HUD’s initial notification, then Freddie Mac was unaware, and thus geocoding or misapplication of HUD’s violate the fair notice doctrine. Freddie HUD would notify the GSE in writing most current data regarding median Mac suggested that if HUD retained the and seek clarification or additional income or underserved areas. An use of these words in its regulation, it information. At this point, the GSE ‘‘omission’’ is defined as a GSE’s failure should explain how their meanings would have 10 working days in which to count units in the denominator. A differ. Fannie Mae stated that potential to respond and could request an ‘‘discrepancy’’ is defined as any adjustments should apply only to extension of an additional 20 working difference between HUD’s analysis of situations where the GSE failed to days from HUD. If the GSE still did not data and the analysis contained in a follow HUD’s rules for data collection respond in a manner that corrected the GSE’s submission of data, including a and reporting, and not where it failed to error, omission, or discrepancy, then discrepancy in goal and/or Special follow its own rules for procedures in HUD would determine the appropriate Affordable subgoal performance. data collection and reporting. Fannie The Department also clarifies in adjustment to the numerator and Mae also contended that adjustments § 81.102(c)(5) of this final rule that an denominator of the applicable goal and/ should be made only where the error, error, omission or discrepancy is or subgoal. Currently, there are no omission or discrepancy was in a data ‘‘material’’ if it results in an required time limits within which the field that affected scoring and where it overstatement of credit for a housing GSEs must respond to HUD’s inquiries also had a material effect on compliance goal or Special Affordable subgoal and, for additional information, and there is with a housing goal. Freddie Mac stated without such overstatement, the GSE no procedure by which HUD can bring that adjustments should be made only would have failed to meet such housing the process of reviewing a GSE’s current for material errors or omissions. Fannie goal or Special Affordable subgoal for year submission to closure absent Mae stated that a GSE should be subject the immediately preceding year. Finally, voluntary assistance from the GSEs. The to additional enforcement action only the rule defines the term ‘‘year-end practical effect of not codifying a when an error, omission or discrepancy data’’ to mean data that HUD receives timetable for completion of this process is due to intentional or bad faith action. from the GSEs related to housing goals is that HUD could be delayed in Both GSEs stated that HUD’s performance in the immediately fulfilling its responsibilities to issue a regulations should provide that HUD preceding year and covering data timely, official report on the GSEs’ will issue a written determination to a reported in the fourth quarter Annual performance for the year most recently GSE when it determines that an Mortgage Report and the GSE’s AHAR. ended and to produce the public use adjustment is necessary, that HUD With respect to procedures for database. should specify which official within notifying a GSE of any suspected error, Summary of Comments. In addition to HUD is authorized to issue orders under omission or discrepancy, HUD is the GSEs, many organizations, including proposed § 81.102(c) and (d), and that responding to the concerns raised by the policy advocates, trade associations, and the rule should provide for more lenient commenters by amending the proposed one non-profit group, commented on the time frames for responding to HUD’s rule to: (1) Provide that, with regard to data verification provisions of HUD’s inquiries. In addition, Freddie Mac each initial notification by HUD to a proposed rule. Nearly all of these commented that the regulations should GSE, HUD may, in its own discretion, or comments supported implementation of state that an order requiring an upon a request by a GSE, extend the some type of data verification adjustment constitutes ‘‘final agency initial five working day response period procedures. One trade group stated that action’’ for purposes of judicial review for up to 20 additional working days; (2) data verification regulations should be under the Administrative Procedure Act establish that any person with delegated enforced to get more accurate and that judicial review is immediately authority from the Secretary, or the information. However, another trade available. Director of HUD’s Financial Institution group expressed concern that the data Fannie Mae also commented on the Regulation Division, or his or her integrity process should include some title of HUD’s provision stating that a designee, is responsible for issuing leeway for unintentional mistakes to provision to correct ‘‘current year end initial notifications regarding errors, avoid becoming burdensome. Two errors’’ is confusing because HUD omissions, or discrepancies, making advocacy organizations supported the cannot correct errors for a current year determinations on the adequacy of proposed provisions regarding data when it does not receive the data about responses received, approving any verification but thought HUD should any current year until the next year. extensions of time permitted under this
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00055 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63634 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
provision, and generally managing the procedures. Examples of errors, and securities issuers and dealers and data verification process; (3) establish omissions, or discrepancies that could may include additional procedures to that the Secretary or his designee will rise to this level under these criteria test random samples of data for inform a GSE in writing of HUD’s include a GSE counting units that are accuracy and completeness. To determination of official performance not eligible under HUD’s rules for goals supplement HUD’s on-site performance figures, including any adjustments, five credit or a GSE underreporting units in review work, the Department has working days prior to HUD’s release of the denominator. With respect to implemented specialized reporting by its official performance figures to the Freddie Mac’s suggestion that HUD’s which each GSE informs HUD on a public; (4) provide that during the five regulations should state that this scheduled basis of key issues and working days prior to such public determination is ‘‘final agency action’’ findings relevant to goals reporting. For release, a GSE may request for purposes of the Administrative example, the GSEs report to HUD reconsideration in writing of HUD’s Procedure Act and is immediately quarterly on the results of their own final determination of its performance subject to judicial review, FHEFSSA internal reviews and self-assessments in which case the Secretary will decide already provides that the GSEs may related to housing goals. These reports whether to grant the request for obtain judicial review in connection cover all actions taken by the GSE to reconsideration, and if the request is with proceedings to enforce the housing remove any findings related to granted, make a final determination on goals, and that those proceedings shall weaknesses in controls or procedures, the request for reconsideration within be governed by the Administrative including those findings identified by 10 working days of the Secretary’s Procedure Act. Therefore, the HUD. granting of the GSE’s request for Department declines to adopt Freddie Because of the complexity of each reconsideration; and (5) provide that, Mac’s suggestion. GSE’s business, as well as the with the exception of the written To more clearly define the scope of complexity of many of the transactions determination of HUD’s official this provision, HUD has renamed this that the GSEs undertake to meet their performance figures, all other provision in the final rule as housing goals, there is a possibility that notifications under this provision may Verification Procedure and Adjustment HUD may discover, during a be by electronic mail. to Correct Errors, Omissions, or performance review, that a serious HUD has also clarified through its Discrepancies in AHAR Data for the overstatement of credit towards one or definitions of errors, omissions and Immediately Preceding Year. more housing goals occurred in the discrepancies, that an ‘‘adjustment’’ will reported prior year under review. 5. Procedures for Prior Year Reporting be made in situations where a GSE Currently, HUD has no procedure for Errors—§ 81.102(d) failed to follow correct procedures in ensuring that any such overstatement is data compilation and reporting and/or The annual data verification review corrected or otherwise adjusted in some where it failed to comply with HUD’s for the immediately preceding year manner unless the overstatement is regulation for determining eligible units. described in § 81.102(c) was designed to discovered in the review of the As has been the case in the past, HUD ensure that reported goals performance immediately past year’s data during the expects that any adjustments that it may was correctly calculated in accordance replication review described in make to the numerator or denominator, with HUD’s regulations. Although these § 81.102(c). To remedy this, HUD that result in a difference between the reviews can test for the reasonableness proposed a procedure that would adjust GSE’s performance as stated in the of some reported data, the reviews a GSE’s current year performance by GSE’s AHAR for the immediately cannot generally determine the accuracy deducting from the numerator of the preceding year and HUD’s official of the underlying loan-level data. To relevant housing goal or subgoal the performance figures, will be well monitor data accuracy, HUD has number of overstated units from a prior understood by the GSE because implemented a second type of year. A prior year was any one of the adjustments of this type occur routinely procedure, called performance reviews. two years preceding the current during HUD’s verification work. Performance reviews are especially reporting year. HUD is also clarifying that it intends necessary because housing goals are Summary of Comments. Many to treat a GSE’s material errors, calculated from information (e.g., organizations commented on HUD’s omissions or discrepancies in, or failure number of dwelling units) that is not data integrity provisions in general and to certify, data submissions under reported in the GSEs’ financial nearly all of these organizations § 81.102(c) as a failure to submit statements and is, therefore, not subject expressed support for data verification. information that the GSE is required to to all GSE procedures designed to The GSEs commented more specifically submit under its charter. Accordingly, ensure the accuracy and completeness on HUD’s proposals for adjustments to the Department may pursue the of reported financial information. HUD’s make up prior years’ overstatements. additional enforcement remedies performance reviews ensure that The GSEs asserted that the Department authorized under § 81.102(e). rigorous audit procedures, either similar does not have authority to either deduct With respect to events that could or identical to those used to monitor the credit from a current year’s purchase trigger enforcement under this integrity of financial data, are also used that is entitled to credit under HUD’s provision, HUD does not intend that in monitoring the accuracy, regulations or add to a current year’s routine technical errors or omissions completeness, and timeliness of the data housing goal to compensate for the would warrant such enforcement. In each GSE submits to HUD. Performance GSE’s failure to meet its goals in a prior order to trigger the enforcement reviews include, but are not limited to, year. They also had other objections, provision, errors, omissions or evaluating the GSEs’ internal controls including the objection that the only discrepancies discovered during review over the collection, management and remedy provided in FHEFSSA for any of the immediately preceding year’s reporting of loan-level mortgage data failure to meet housing goals is the performance must be material, as HUD used in calculating housing goals imposition of a housing plan, which has defined that term. The error, performance. Performance reviews may may address only a probable failure to omission or discrepancy also must be also focus on the GSEs’ quality control meet housing goals in the current year one that indicates to HUD a serious standards and procedures for or actual failure to meet goals in a problem in the GSE’s internal information received from loan sellers current year in the next calendar year.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00056 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63635
The GSEs stated that the Congress agrees that the only remedy Congress set Congress as expressed in the FHEFSSA intended the statute to provide no out in FHEFSSA for failing to meet a and its legislative history. FHEFSSA remedy for their failure to meet a prior housing goal is a housing plan under and its legislative history indicate that year’s housing goal and, therefore, that section 1336, and as the statute is Congress established a comprehensive the Department has no authority to written, the housing plan addresses only regulatory scheme under which HUD fashion such a remedy. Based on this a current year’s failure, either in that would establish and enforce the line of reasoning, they concluded that year or in the next calendar year. Housing Goals through strict and HUD may not take any action against a Therefore, the statute does not pervasive regulation. GSE when it discovers that it failed to specifically address a GSE’s failure to Furthermore, there is absolutely no meet a housing goal in a prior year, even meet a housing goal in a prior year, i.e., statement in FHEFSSA or its legislative though HUD could have taken action if a failure occurring in any one of the two history to suggest that Congress the failure had been discovered within years immediately preceding the latest intended that HUD must ignore or one year after the year in which it year for which data on housing goals forgive a GSE’s failure to meet its occurred. performance was reported to HUD. housing goals in any year for any Both GSEs also objected to the policy However, the Department does not agree reason, including the passage of a basis for HUD’s proposal. For instance, that Congressional silence on this certain amount of time before HUD Fannie Mae wanted the time period precise issue means either that Congress discovers this failure. The fact that within which HUD might impose a intended the GSEs to be allowed to fail FHEFSSA is silent on the issue of how prior-year correction shortened from up to meet their housing goals as long as to address a GSE’s failure to meet a prior to 24 months to three months after the Department does not discover that year’s housing goal means that there is HUD’s receipt of AHAR loan-level data, failure within a specific time or that the a gap in FHEFSSA’s enforcement which HUD receives on March 15th of Department may not fashion a remedy scheme regarding this precise issue. each year. Fannie Mae cited OFHEO’s to address this issue. This conclusion Under Chevron v. NRDC, 467 U.S. 837 ability to render a decision on its final runs counter to Congress’s purposes in (1984), the Department has discretion to capital classification within 90 days of enacting FHEFSSA, which directs HUD fashion an appropriate remedy to fill the reporting quarter as evidence that to establish and monitor the GSEs’ this gap, and it has done so in complex determinations can be made compliance with the Housing Goals. § 81.102(d). Moreover, the Department within short time frames. Freddie Mac Section 1336 of FHEFSSA provides has the discretion to fashion a remedy saw no reason why the necessary that the Secretary shall ‘‘monitor and to correct prior year overstatements evaluations could not be accomplished enforce’’ the GSEs’’ compliance with the without which a GSE would have failed within six months after the close of the housing goals set by the Department. to meet a housing goal or Special immediately preceding year. Freddie According to FHEFSSA’s legislative Affordable subgoal under its general Mac stated that HUD’s policy history, in enacting FHEFSSA Congress regulatory powers under section 1321 of justification does not support the intended ‘‘to establish a comprehensive FHEFSSA. proposal and that HUD did not point to framework of goals, data collection, However, in light of the objections any instance where the increasing reporting requirements and enforcement raised to the proposed regulation in the comments discussed above, HUD has complexity of transactions has led to provisions.’’ S. Rep. No. 102–282, at 34 revised § 81.102(d) to remove provisions overstatements in performance. Freddie (1992)(emphasis added). Mac also commented that the When discussing the GSEs’ duties to that either provide for deduction of Housing Goals credit in a current year Department already has the option of meet housing goals set for low- and from purchases that qualify for credit, or publicizing the discovery of any prior moderate-income housing and housing that add requirements to a current year’s year mistakes—by press release, news in central cities and rural areas, Housing Goals due to errors, omissions conference or its Web site information— Congress stated: and of making Congress aware of these or discrepancies in a prior year’s data mistakes. The GSEs need to provide more leadership submissions. The final rule provides Freddie Mac requested that HUD in all of these areas, and they have indicated instead that the Secretary may require withdraw the proposal entirely. If HUD a desire to do so. But direct and potentially the GSEs to make up any overstatements opted to proceed to implement the forceful federal oversight is the only way to of goal performance without which a ensure that it will happen. Id. at 11. proposal, then Freddie Mac suggested GSE would have failed to meet a prior that HUD amend the provision to: (1) Under the GSEs’ suggested year’s Housing Goal, no later than the Limit application of the rule to large construction of FHEFSSA, HUD’s ability year following the year in which HUD prior year overstatements that affect a to enforce the housing goals is totally first notifies the GSE of this failure. (The material number of units under a goal dependent upon only one factor, namely rule now defines the term ‘‘prior year’’ (e.g., five percent); (2) provide that HUD how quickly HUD discovers that a GSE to mean any one of the two years will apply the rule only when a GSE has failed to meet a goal. In order to immediately preceding the latest year acted in bad faith; (3) provide that HUD determine whether a GSE has failed a for which data on housing goals will not apply the rule cumulatively; goal, HUD must receive, verify and performance was reported to HUD.) that is, that HUD will not accumulate analyze massive amounts of data, as In order to remedy this failure, the several years’ over-counts and then described above. Under the GSEs’ Secretary may require the GSEs to deduct a cumulative total from the suggested construction of FHEFSSA, purchase additional mortgages that current year; and (4) clarify in the final only if HUD discovers that a GSE has finance a number of units that either (a) rule which official within HUD will failed to meet a housing goal or subgoal equal the number of units overstated in make decisions under this provision in the nine month period that runs from the prior year’s goal performance, or for and provide that the basis for decisions March 15th, when the GSEs submit the Special Affordable subgoals the be explained. current year-end data, to the end of that number or dollar amount, as applicable, HUD’s Determination. HUD has year—may HUD enforce the housing of mortgage purchases that the Secretary carefully considered both GSEs’ goals for that year. Such a construction has determined were overstated, or (b) comments, including their legal and is not only unreasonable on its face but that equal the percentage of the policy arguments. The Department it is contrary to the plain intent of overstatement in the prior year’s goal
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00057 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63636 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
performance as applied to the most treating a GSE’s material errors, data accuracy in performance reviews, current year-end performance, omissions or discrepancies in, or failure HUD is concerned with errors of a whichever is less. Units purchased to to certify, a prior year’s data submission substantial nature, such as those that remedy an overstatement must be as a failure to submit information that suggest a larger internal control eligible to qualify under the same goal the GSE is required to submit under its problem, an example of which could be or goals for which the overstatement charter. a pattern of incorrect rental data occurred in the prior year. For example, Both GSEs also questioned the need acquired from or generated by the same a GSE may have overstated a prior year’s for an adjustment period that could source. HUD is also concerned with performance by 5,000 units or .22 extend for up to 24 months from the types of transactions that are either percent under the Low- and Moderate- close of a calendar year’s performance, expressly prohibited from goals Income Goal. To make up this believing instead that any such review eligibility, such as high cost mortgages, overstatement, a GSE may purchase an could be accomplished within six or for which HUD approval may have additional 5,000 units that are eligible months of the close of the previous year, been required but not obtained prior to under the Low- and Moderate-Income which is a time frame similar to that a GSE counting the units, such as the Goal in the year immediately following used by OFHEO to assess the adequacy use of an affordability estimation the year in which HUD notifies the GSE of a GSE’s capital. As HUD has stated methodology. Other similar types of of the overstatement or it may multiply previously, reviews conducted problems may also trigger a HUD the current year’s total eligible immediately upon receipt of a GSE’s determination of error. In the event purchases under the Low- and prior year loan-level data and pursuant HUD supplements its regulations with Moderate-Income Goal by the overstated to § 81.102(c) cannot generally gauge the letters to one or both GSEs regarding percentage from a prior year (e.g., .22 accuracy of the data and cannot always appropriate counting treatment, the GSE percent) to determine the number of determine whether the transaction itself will be responsible for complying with units that must be purchased, provided complies with HUD’s regulations for only the specific directives it has this number is less than 5,000 units. The counting units towards goals received from HUD. In the final rule, same requirement also applies to the performance. Assessments of this type HUD has reiterated that this procedure Special Affordable Home Purchase require the application of procedures, will apply only in those instances where Subgoal. When an overstatement occurs either in whole or in part, that are an overstatement was material in nature; under this Subgoal, the Secretary may characteristic of audit engagements. For that is, the overstated units enabled the require the GSE to make up the number example, it is customary for audits of a GSE to meet a housing goal that it of mortgages that were overstated using previous year’s financial statements to otherwise would not have met. In the the lesser of the two procedures require up to one year or more for event that HUD undertakes a previously described. For completion due to the number of performance review that covers a two- overstatements under the Special procedures involved and the volume of year period and determines that Affordable Multifamily Subgoal, the information to be reviewed, especially material misstatements of housing goals GSE may be required to make up the for exceedingly large and complex or Special Affordable subgoals dollar amount of overstatement by organizations. Similarly, the relatively performance occurred in both years, purchasing qualifying multifamily new field of performance data auditing, then HUD will apply the same mortgages in an amount equal to the including reviews based on some or all procedures as described previously for overstatement. The GSEs will not be of these procedures, also requires a making up the overstatements. Upon a substantial commitment of time and required to make up any errors, written request from a GSE, the resources if meaningful results are to be omissions or discrepancies in prior Secretary may, in his discretion, obtained. For these reasons, years that were not material. As determine to grant an extension of performance reviews are not analogous previously noted, the final rule provides additional time to correct or compensate to OFHEO’s evaluations of capital that an error, omission or discrepancy is for the overstatement. For example, if adequacy. HUD believes that its original material if it results in an overstatement overstatements were discovered for proposal of allowing for up to 24 of credit for a housing goal or Special years 2005 and 2006 and the GSE is Affordable subgoal and, without such months after the close of the year under notified of the overstatements for both overstatement, the GSE would have review is the appropriate time frame for years in 2007, then the GSE could be failed to meet such housing goal or completion of the performance review. required to make up the overstatements Special Affordable subgoal for the prior The GSEs also expressed some for both years in 2008. Similarly, if the year. concerns about the potential for HUD to Also, corrections for overstatement of make determinations of error after the overstatement was discovered for one goals performance under this provision fact and without any prior notice to a year, 2005, and the GSE was notified of will not be counted or reported under GSE that a type of transaction and/or the overstatement in 2006, then the GSE the GSEs’ Annual Housing Activity housing unit would not be eligible for could be required to make up the Report, including calculation of housing goals credit. HUD believes it is useful to overstated units or mortgages in 2007. In goal performance in any year, but rather more fully describe the types of errors both examples, upon receipt of a GSE’s will be managed separately from the likely to trigger a finding that units were written request for an extension of time, housing goals as directed by HUD. overstated. In the context of the Secretary may grant an extension for If the GSE does not purchase a performance reviews, the words ‘‘errors, completing make up of the overstated sufficient number of eligible units or omissions or discrepancies’’ connote units or mortgages. mortgages, as described previously, then serious mistakes, such as those With regard to HUD’s reasons for HUD may issue a notice that the GSE associated with violations of HUD’s implementing a procedure that provides failed a housing goal or subgoal in a counting rules and other goals eligibility a mechanism by which overstated units previous year, or seek additional criteria as set forth in its regulations. of a material nature from a prior year enforcement remedies under § 81.102(e) HUD is aware that in collecting and can be made up in a subsequent year or any other civil or administrative reporting millions of data elements, following the year a GSE is first notified remedies that are available under some level of factual error is probably of the overstated units, for reasons applicable law. The Department is unavoidable. However, with regard to stated above, it is the Department’s view
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00058 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63637
that it has authority to do so, and that civil money penalties, or any other GSE fails to submit the certification the procedure is needed at this time. remedies or penalties that may be required by § 81.102(b) in connection The procedure is the only tool by which available to the Secretary as a result of with data, information or report(s) HUD can meet its statutory the GSE’s failure to provide data required by section 309(m) or (n) of the responsibility to assure the integrity of submissions, information, and/or Fannie Mae Charter Act, section 307(e) all of the housing goal data reported to report(s) in accordance with § 81.102. or (f) of the Freddie Mac Act, or under the public, including the data reported Summary of Comments. Several 24 CFR part 81, subpart E; or (2) when in the GSE public use database and its organizations commented, generally, on a GSE submits the certification required duty to enforce the housing goals. HUD’s proposed requirements in by § 81.102(b) in connection with such § 81.102 for ensuring the accuracy and data, information or report(s), but the 6. Additional Enforcement Option integrity of GSE data and other Secretary later determines that the data, § 81.102(e) submissions, and almost all expressed information or report(s) are not ‘‘true, The Department proposed a new support for HUD’s proposals relating to correct and complete’’ as provided in § 81.102(e) that would provide HUD data verification. The GSEs commented the certification. The final rule provides with additional enforcement options in more specifically on HUD’s proposal in that, under either of the above two the event it determines that a GSE has § 81.102(e) relating to additional circumstances, the Secretary may regard submitted data, information, or report(s) enforcement options. a GSE’s actions as tantamount to a that are not current, are incomplete, or Fannie Mae asserted that HUD’s failure to submit the data, information otherwise contain an untrue statement proposed additional enforcement or report(s) which, in turn, authorizes of material fact. Section 81.102(e) options were overly broad, and the Secretary to take the additional identified the data, information, or exceeded the Department’s authority enforcement remedies described in report(s) that would be subject to HUD’s under FHEFSSA to issue cease and § 81.102(e). additional enforcement authority as desist orders, impose civil money The final rule also clarifies that for those required under section 307(e) and penalties, and to punish GSE data, information or report(s) that are (f) of the Freddie Mac Act, section noncompliance by requiring the subject to § 81.102(c) or (d), the 309(m) or (n) of the Fannie Mae Charter adoption of a housing plan. Fannie Mae Secretary may only pursue additional Act, or under 24 CFR part 81, subpart stated that, if HUD decided to retain enforcement remedies in connection E. § 81.102(e), this provision should be with material errors, omissions or The Department indicated in redrafted more narrowly. discrepancies. Moreover, if the data, proposed § 81.102(e) that it could make Freddie Mac, through a legal opinion information or report(s) are subject to a determination—either under its prepared by outside counsel, asserted § 81.102(d), the rule provides that the independent verification authority in that sections 1341 and 1345 of Secretary may only pursue additional § 81.102(a) or by ‘‘other means’’—that FHEFSSA provide a two-step process enforcement remedies if the GSE has such data, information or report(s) are before a GSE’s failure to submit a failed to purchase a sufficient amount or not current, are incomplete, or housing plan, or its failure to comply type of mortgages as required by the otherwise contain an untrue statement with a feasible housing plan, could Secretary under § 81.102(d)(4). of material fact. This reference to ‘‘other result in the Department’s initiating an It is the Department’s view that means’’ was intended to encompass the action for a cease and desist order or § 81.102(e) is needed so that it can take Secretary’s authority under the three civil money penalties. Freddie Mac appropriate action to ensure the other provisions in § 81.102 that were asserted that HUD’s proposal expanded accuracy and completeness of the GSEs’ also being proposed to ensure the its enforcement authority beyond the submissions to HUD, and also to accuracy, truthfulness and completeness FHEFSSA statutory limits by implement the certification that is now of GSE submissions to HUD: (1) The eliminating this two-step process. established at § 81.102(b) of this final proposed GSE certification in Freddie Mac also contended that HUD’s rule, while providing the Secretary with § 81.102(b); (2) the proposed procedure enforcement powers under sections sufficient flexibility to exercise his or established in § 81.102(c) to verify the 1341 and 1345 of FHEFSSA extend only her discretion to determine whether accuracy and completeness of the GSE’s to instances of intentional non- enforcement action is appropriate in current year-end data; and (3) the compliance by the GSE, and that each instance. proposed procedure established in § 81.102 should be narrowed to reflect The final rule clarifies that the § 81.102(d) to ensure the accuracy and this limitation. proposed rule’s reference in paragraph completeness of the GSE’s prior years’ HUD’s Determination. The (e)(1) to ‘‘other means’’ by which the data. Department has considered the GSEs’ Secretary may determine that a GSE’s The Department further provided in and other comments on § 81.102(e) and data submission(s), information or § 81.102(e) that the Secretary could is making several changes in this final report(s) fail to meet the prescribed regard a GSE’s submission of data, rule in response to these comments. In regulatory standards is meant to refer to information or report(s) that he or she addition, the Department is making a the Secretary’s determinations under determines under § 81.102(a), or by number of conforming changes to paragraphs (b), (c) or (d) of § 81.102 (i.e., ‘‘other means’’ (i.e., pursuant to § 81.102(e) to reflect changes that it has the GSE certification in § 81.102(b), the paragraphs (b), (c) or (d) of § 81.102), are also decided to adopt in connection procedure established in § 81.102(c) to not current, are incomplete, or that with the other provisions in § 81.102 verify the accuracy and completeness of otherwise contain an untrue statement (primarily in paragraphs (b), (c) and (d)), the GSE’s data for the immediately of material fact to be equivalent to the and is also making minor editorial preceding year, and the procedure GSE’s failure to submit such data. As a corrections. established in § 81.102(d) to ensure the result of such failure of submission, Specifically, the Department is accuracy and completeness of the GSE’s proposed § 81.102(e) provided that the providing in this final rule that: prior years’ data). In the final rule, the Department could initiate against the The Department may pursue Department has deleted the reference to GSE, in accordance with the procedures additional enforcement remedies under ‘‘other means’’ and has included a in 24 CFR part 81, subpart G, an order paragraph (e) under either of the specific reference to paragraphs (b), (c) to cease and desist, an action to seek following circumstances: (1) When a or (d) of § 81.102.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00059 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63638 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
The final rule establishes a bifurcated Department of Housing and Urban not have federalism implications and approach with respect to the types of Development, 451 Seventh Street, SW., does not impose substantial direct additional enforcement remedies that Washington, DC. The Economic compliance costs on state and local the Department may pursue under Analysis prepared for this rule is also governments or preempt state law paragraph (e). This bifurcated approach available for public inspection in the within the meaning of the executive recognizes that the Department’s ability Office of the Rules Docket Clerk and on order. to pursue a cease and desist order, or to HUD’s Web site at http://www.hud.gov. Unfunded Mandates Reform Act levy civil money penalties, applies Congressional Review of Regulations specifically to data required by section Title II of the Unfunded Mandates 309(m) or (n) of the Fannie Mae Charter This rule is a ‘‘major rule’’ as defined Reform Act of 1995 (12 U.S.C. 1531– Act or section 307(e) or (f) of the in Chapter 8 of 5 U.S.C. This rule will 1538) (UMRA) establishes requirements Freddie Mac Act. The rule nevertheless be submitted for Congressional review for federal agencies to assess the effects provides that the Department may in accordance with this chapter. of their regulatory actions on state, local, and tribal governments, and the pursue other types of remedies against Paperwork Reduction Act a GSE in connection with data that the private sector. This rule would not GSE is required to submit under 24 CFR HUD’s collection of information on impose any federal mandates on any part 81, subpart E, but that the GSE is the GSEs’ activities has been reviewed state, local, or tribal government, or on not required to submit under section and authorized by OMB under the the private sector, within the meaning of 309(m) or (n) of the Fannie Mae Charter Paperwork Reduction Act of 1995 (44 UMRA. U.S.C. 3501–3520), as implemented by Act or section 307(e) or (f) of the List of Subjects in 24 CFR Part 81 Freddie Mac Charter Act. OMB in regulations at 5 CFR part 1320. The final rule provides that, in The OMB control number is 2502–0514. Accounting, Federal Reserve System, connection with either of the two Environmental Impact Mortgages, Reporting and recordkeeping remedial approaches now described in requirements, Securities. This final rule does not direct, § 81.102(e)(2), the Secretary may pursue I For the reasons discussed in the any civil or administrative remedies or provide for assistance or loan and mortgage insurance for, or otherwise preamble, HUD is amending 24 CFR part penalties against the GSE that may be 81 as follows: available to the Secretary by virtue of govern or regulate real property either of the circumstances described in acquisition, disposition, leasing, PART 81—THE SECRETARY OF HUD’S 81.102(e)(1). If the Department elects to rehabilitation, alteration, demolition, or REGULATION OF THE FEDERAL pursue a cease-and-desist order or civil new construction; or establish, revise, or NATIONAL MORTGAGE ASSOCIATION money penalties against a GSE under provide for standards for construction or (FANNIE MAE) AND THE FEDERAL § 81.102(e)(2)(i)(A) or (B), it will comply construction materials, manufactured HOME LOAN MORTGAGE with the procedures applicable to such housing, or occupancy. Accordingly, CORPORATION (FREDDIE MAC) actions under 24 CFR part 81, subpart under 24 CFR 50.19(c)(1), this rule is I G. Alternatively, if the Department categorically excluded from 1. The authority citation for 24 CFR elects to pursue other civil or environmental review under the part 81 continues to read as follows: administrative remedies against a GSE National Environmental Policy Act of Authority: 12 U.S.C. 1451 et seq., 1716– under either §§ 81.102(e)(2)(i)(C) or 1969 (42 U.S.C. 4321). 1723h, and 4501–4641; 28 U.S.C. 2461 note; 81.102(e)(2)(ii), it will pursue such Regulatory Flexibility Act 42 U.S.C. 3535(d) and 3601–3619. remedies in accordance with applicable The Secretary, in accordance with the I 2. In § 81.2(b), revise the definitions of law. Regulatory Flexibility Act (5 U.S.C. ‘‘Metropolitan area’’ and ‘‘Minority,’’ Finally, the Department is replacing 605(b)), has reviewed this rule before and paragraph (2) of the definition of in paragraph (e) each reference to publication and by approving it certifies ‘‘Underserved area,’’ and add a new ‘‘HUD’’ with a reference to ‘‘the that this rule would not have a definition of the term ‘‘Home Purchase Secretary.’’ This replacement is significant economic impact on a Mortgage,’’ in alphabetical order, to read designed to ensure that any additional substantial number of small entities. as follows: enforcement action that may be pursued This rule is applicable only to the GSEs, under § 81.102(e) will be considered at § 81.2 Definitions. which are not small entities for the highest levels within the purposes of the Regulatory Flexibility * * * * * Department. Act. Therefore, the rule does not have a (b) * * * III. Findings and Certifications significant economic impact on a Home Purchase Mortgage means a substantial number of small entities residential mortgage for the purchase of Executive Order 12866 within the meaning of the Regulatory an owner-occupied single-family The Office of Management and Budget Flexibility Act. property. (OMB) reviewed this final rule under * * * * * Executive Order 12866, Regulatory Executive Order 13132, Federalism Metropolitan area means a Planning and Review, which the Executive Order 13132 (‘‘Federalism’’) metropolitan statistical area (‘‘MSA’’), or President issued on September 30, 1993. prohibits, to the extent practicable and a portion of such an area for which This rule was determined to be permitted by law, an agency from median family income estimates are economically significant under E.O. promulgating a regulation that has published annually by HUD. 12866. Any changes made to this rule federalism implications and either Minority means any individual who is subsequent to its submission to OMB imposes substantial direct compliance included within any one or more of the are identified in the docket file, which costs on state and local governments following racial and ethnic categories: is available for public inspection and is not required by statute, or (1) American Indian or Alaskan between 8 a.m. and 5 p.m. weekdays in preempts state law, unless the relevant Native—a person having origins in any the Office of the Rules Docket Clerk, requirements of section 6 of the of the original peoples of North and Office of General Counsel, Room 10276, executive order are met. This rule does South America (including Central
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00060 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63639
America), and who maintains tribal Subgoal, 45 percent of the total number metropolitan areas financed by that affiliation or community attachment; of home purchase mortgages in GSE’s mortgage purchases shall be home (2) Asian—a person having origins in metropolitan areas financed by that purchase mortgages in metropolitan any of the original peoples of the Far GSE’s mortgage purchases shall be home areas which count toward the Low- and East, Southeast Asia, or the Indian purchase mortgages in metropolitan Moderate-Income Housing Goal in each subcontinent, including, for example, areas which count toward the Low- and of those years unless otherwise adjusted Cambodia, China, India, Japan, Korea, Moderate-Income Housing Goal in the by HUD in accordance with FHEFSSA. Malaysia, Pakistan, the Philippine year 2005 unless otherwise adjusted by I 4. In § 81.13, revise the last sentence of Islands, Thailand, and Vietnam; HUD in accordance with FHEFSSA; paragraph (b) and revise paragraph (c), to (3) Black or African American—a (2) For the year 2006, 53 percent of read as follows: person having origins in any of the the total number of dwelling units black racial groups of Africa; financed by that GSE’s mortgage § 81.13 Central Cities, Rural Areas, and Other Underserved Areas Housing Goal. (4) Hispanic or Latino—a person of purchases unless otherwise adjusted by Cuban, Mexican, Puerto Rican, South or HUD in accordance with FHEFSSA. In * * * * * Central American, or other Spanish addition, as a Low- and Moderate- (b) Factors. * * * A statement culture or origin, regardless of race; and Income Housing Home Purchase documenting HUD’s considerations and (5) Native Hawaiian or Other Pacific Subgoal, 46 percent of the total number findings with respect to these factors, Islander—a person having origins in any of home purchase mortgages in entitled ‘‘Departmental Considerations of the original peoples of Hawaii, Guam, metropolitan areas financed by that to Establish the Central Cities, Rural Samoa, or other Pacific Islands. GSE’s mortgage purchases shall be home Areas, and Other Underserved Areas Housing Goal,’’ was published in the * * * * * purchase mortgages in metropolitan areas which count toward the Low- and Federal Register on November 2, 2004. Underserved area means * * * (c) Goals. The annual goals for each (2) For purposes of the definition of Moderate-Income Housing Goal in the year 2006 unless otherwise adjusted by GSE’s purchases of mortgages on ‘‘Rural area,’’ a whole census tract, a housing located in central cities, rural Federal or State American Indian HUD in accordance with FHEFSSA; (3) For the year 2007, 55 percent of areas, and other underserved areas are: reservation or tribal or individual trust the total number of dwelling units (1) For the year 2005, 37 percent of land, or the balance of a census tract financed by that GSE’s mortgage the total number of dwelling units excluding the area within any Federal or purchases unless otherwise adjusted by financed by that GSE’s mortgage State American Indian reservation or HUD in accordance with FHEFSSA. In purchases unless otherwise adjusted by tribal or individual trust land, having: addition, as a Low- and Moderate- HUD in accordance with FHEFSSA. In (i) A median income at or below 120 Income Housing Home Purchase addition, as a Central Cities, Rural percent of the greater of the State non- Subgoal, 47 percent of the total number Areas, and Other Underserved Areas metropolitan median income or the of home purchase mortgages in Home Purchase Subgoal, 32 percent of nationwide non-metropolitan median metropolitan areas financed by that the total number of home purchase income and a minority population of 30 GSE’s mortgage purchases shall be home mortgages in metropolitan areas percent or greater; or purchase mortgages in metropolitan financed by that GSE’s mortgage (ii) A median income at or below 95 areas which count toward the Low- and purchases shall be home purchase percent of the greater of the State non- Moderate-Income Housing Goal in the mortgages in metropolitan areas which metropolitan median income or year 2007 unless otherwise adjusted by count toward the Central Cities, Rural nationwide non-metropolitan median HUD in accordance with FHEFSSA; Areas, and Other Underserved Areas income. (4) For the year 2008, 56 percent of Housing Goal in the year 2005 unless * * * * * the total number of dwelling units otherwise adjusted by HUD in I 3. In § 81.12, revise the last sentence of financed by that GSE’s mortgage accordance with FHEFSSA; paragraph (b) and revise paragraph (c), to purchases unless otherwise adjusted by (2) For the year 2006, 38 percent of read as follows: HUD in accordance with FHEFSSA. In the total number of dwelling units addition, as a Low- and Moderate- financed by that GSE’s mortgage § 81.12 Low- and Moderate-Income Income Housing Home Purchase purchases unless otherwise adjusted by Housing Goal. Subgoal, 47 percent of the total number HUD in accordance with FHEFSSA. In * * * * * of home purchase mortgages in addition, as a Central Cities, Rural (b) Factors. * * * A statement metropolitan areas financed by that Areas, and Other Underserved Areas documenting HUD’s considerations and GSE’s mortgage purchases shall be home Home Purchase Subgoal, 33 percent of findings with respect to these factors, purchase mortgages in metropolitan the total number of home purchase entitled ‘‘Departmental Considerations areas which count toward the Low- and mortgages in metropolitan areas to Establish the Low- and Moderate- Moderate-Income Housing Goal in the financed by that GSE’s mortgage Income Housing Goal,’’ was published year 2008 unless otherwise adjusted by purchases shall be home purchase in the Federal Register on November 2, HUD in accordance with FHEFSSA; and mortgages in metropolitan areas which 2004. (5) For the year 2009 and thereafter count toward the Central Cities, Rural (c) Goals. The annual goals for each HUD shall establish annual goals. Areas, and Other Underserved Areas GSE’s purchases of mortgages on Pending establishment of goals for the Housing Goal in the year 2006 unless housing for low- and moderate-income year 2009 and thereafter, the annual otherwise adjusted by HUD in families are: goal for each of those years shall be 56 accordance with FHEFSSA; (1) For the year 2005, 52 percent of percent of the total number of dwelling (3) For the year 2007, 38 percent of the total number of dwelling units units financed by that GSE’s mortgage the total number of dwelling units financed by that GSE’s mortgage purchases in each of those years. In financed by that GSE’s mortgage purchases unless otherwise adjusted by addition, as a Low and Moderate purchases unless otherwise adjusted by HUD in accordance with FHEFSSA. In Income Housing Home Purchase HUD in accordance with FHEFSSA. In addition, as a Low- and Moderate- Subgoal, 47 percent of the total number addition, as a Central Cities, Rural Income Housing Home Purchase of home purchase mortgages in Areas, and Other Underserved Areas
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00061 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63640 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Home Purchase Subgoal, 33 percent of and affordable to low-income families in accordance with FHEFSSA. In addition, the total number of home purchase low-income areas and very low-income as a Special Affordable Housing Home mortgages in metropolitan areas families are: Purchase Subgoal, 18 percent of the financed by that GSE’s mortgage (1) For the year 2005, 22 percent of total number of home purchase purchases shall be home purchase the total number of dwelling units mortgages in metropolitan areas mortgages in metropolitan areas which financed by each GSE’s mortgage financed by each GSE’s mortgage count toward the Central Cities, Rural purchases unless otherwise adjusted by purchases shall be home purchase Areas, and Other Underserved Areas HUD in accordance with FHEFSSA. The mortgages in metropolitan areas which Housing Goal in the year 2007 unless goal for the year 2005 shall include count toward the Special Affordable otherwise adjusted by HUD in mortgage purchases financing dwelling Housing Goal in the year 2007 unless accordance with FHEFSSA; units in multifamily housing totaling otherwise adjusted by HUD in (4) For the year 2008, 39 percent of not less than 1.0 percent of the average accordance with FHEFSSA; the total number of dwelling units annual dollar volume of combined financed by that GSE’s mortgage (single-family and multifamily) (4) For the year 2008, 27 percent of purchases unless otherwise adjusted by mortgages purchased by the respective the total number of dwelling units HUD in accordance with FHEFSSA. In GSE in 2000, 2001, and 2002, unless financed by each GSE’s mortgage addition, as a Central Cities, Rural otherwise adjusted by HUD in purchases unless otherwise adjusted by Areas, and Other Underserved Areas accordance with FHEFSSA. In addition, HUD in accordance with FHEFSSA. The Home Purchase Subgoal, 34 percent of as a Special Affordable Housing Home goal for the year 2008 shall include the total number of home purchase Purchase Subgoal, 17 percent of the mortgage purchases financing dwelling mortgages in metropolitan areas total number of home purchase units in multifamily housing totaling financed by that GSE’s mortgage mortgages in metropolitan areas not less than 1.0 percent of the average purchases shall be home purchase financed by each GSE’s mortgage annual dollar volume of combined mortgages in metropolitan areas which purchases shall be home purchase (single-family and multifamily) count toward the Central Cities, Rural mortgages in metropolitan areas which mortgages purchased by the respective Areas, and Other Underserved Areas count toward the Special Affordable GSE in 2000, 2001, and 2002, unless Housing Goal in the year 2008 unless Housing Goal in the year 2005 unless otherwise adjusted by HUD in otherwise adjusted by HUD in otherwise adjusted by HUD in accordance with FHEFSSA. In addition, accordance with FHEFSSA; and accordance with FHEFSSA; as a Special Affordable Housing Home (5) For the year 2009 and thereafter (2) For the year 2006, 23 percent of Purchase Subgoal, 18 percent of the HUD shall establish annual goals. the total number of dwelling units total number of home purchase Pending establishment of goals for the financed by each GSE’s mortgage mortgages in metropolitan areas purchases unless otherwise adjusted by year 2009 and thereafter, the annual financed by each GSE’s mortgage HUD in accordance with FHEFSSA. The goal for each of those years shall be 39 purchases shall be home purchase goal for the year 2006 shall include percent of the total number of dwelling mortgages in metropolitan areas which mortgage purchases financing dwelling units financed by that GSE’s mortgage count toward the Special Affordable units in multifamily housing totaling purchases in each of those years. In Housing Goal in the year 2008 unless not less than 1.0 percent of the average addition, as a Central Cities, Rural annual dollar volume of combined otherwise adjusted by HUD in Areas, and Other Underserved Areas (single-family and multifamily) accordance with FHEFSSA; and Home Purchase Subgoal, 34 percent of mortgages purchased by the respective (5) For the year 2009 and thereafter the total number of home purchase GSE in 2000, 2001, and 2002, unless HUD shall establish annual goals. mortgages in metropolitan areas otherwise adjusted by HUD in Pending establishment of goals for the financed by that GSE’s mortgage accordance with FHEFSSA. In addition, year 2009 and thereafter, the annual purchases shall be home purchase as a Special Affordable Housing Home goal for each of those years shall be 27 mortgages in metropolitan areas which Purchase Subgoal, 17 percent of the percent of the total number of dwelling count toward the Central Cities, Rural total number of home purchase units financed by each GSE’s mortgage Areas, and Other Underserved Areas mortgages in metropolitan areas purchases in each of those years. The Housing Goal in each of those years financed by each GSE’s mortgage goal for each such year shall include unless otherwise adjusted by HUD in purchases shall be home purchase accordance with FHEFSSA. mortgage purchases financing dwelling mortgages in metropolitan areas which units in multifamily housing totaling * * * * * count toward the Special Affordable not less than 1.0 percent of the annual I 5. In § 81.14, revise the last sentence of Housing Goal in the year 2006 unless average dollar volume of combined paragraph (b) and revise paragraph (c), to otherwise adjusted by HUD in (single-family and multifamily) read as follows: accordance with FHEFSSA; mortgages purchased by the respective (3) For the year 2007, 25 percent of GSE in the years 2000, 2001, and 2002. § 81.14 Special Affordable Housing Goal. the total number of dwelling units In addition, as a Special Affordable * * * * * financed by each GSE’s mortgage Housing Home Purchase Subgoal, 18 (b) * * * A statement documenting purchases unless otherwise adjusted by HUD’s considerations and findings with HUD in accordance with FHEFSSA. The percent of the total number of home respect to these factors, entitled goal for the year 2007 shall include purchase mortgages in metropolitan ‘‘Departmental Considerations to mortgage purchases financing dwelling areas financed by each GSE’s mortgage Establish the Special Affordable units in multifamily housing totaling purchases shall be home purchase Housing Goal,’’ was published in the not less than 1.0 percent of the average mortgages in metropolitan areas which Federal Register on November 2, 2004. annual dollar volume of combined count toward the Special Affordable (c) Goals. The annual goals for each (single-family and multifamily) Housing Goal in each of those years GSE’s purchases of mortgages on rental mortgages purchased by the respective unless otherwise adjusted by HUD in and owner-occupied housing meeting GSE in 2000, 2001, and 2002, unless accordance with FHEFSSA. the then-existing, unaddressed needs of otherwise adjusted by HUD in * * * * *
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00062 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63641
I 6. In § 81.15, revise paragraphs (d), (ii) In any calendar year, a GSE may toward achievement of each goal, as (e)(6)(i), and (e)(6)(ii) and add a new use only one of the methods specified determined by HUD based on the most paragraph (i), to read as follows: in paragraph (d)(2)(i) of this section to recent decennial census. For units with estimate affordability information for missing affordability information in § 81.15 General requirements. single-family owner-occupied units. tracts for which such methodology is * * * * * (iii) If a GSE chooses to use an not possible, such units will be (d) Counting owner-occupied units. estimation methodology under excluded from the denominator as well (1) For purposes of counting owner- paragraph (d)(2)(i)(B) or (d)(2)(i)(C) of as the numerator in calculating occupied units toward achievement of this section to determine affordability performance under the respective the Low- and Moderate-Income Housing for owner-occupied units in properties housing goal(s); or Goal or the Special Affordable Housing securing single-family mortgage (2) Such other data source and Goal, mortgage purchases financing purchases eligible to be counted toward methodology as may be approved by such units shall be evaluated based on the respective housing goal, then that HUD. the income of the mortgagors and the methodology may be used up to (B) In any calendar year, a GSE may area median income at the time of nationwide maximums for home use only one of the methods specified origination of the mortgage. To purchase mortgages and for refinance in paragraph (e)(6)(i)(A) of this section determine whether mortgages may be mortgages that shall be calculated by to estimate affordability information for counted under a particular family multiplying, for each census tract, the multifamily rental units. income level, i.e., especially low, very percentage of all single-family owner- (C) If a GSE chooses to use an low, low or moderate income, the occupied mortgage originations with income of the mortgagors is compared to estimation methodology under missing borrower incomes (as paragraph (e)(6)(i)(A) of this section to the median income for the area at the determined by HUD based on the most time of the mortgage application, using determine affordability for rental units recent HMDA data available for home in properties securing multifamily the appropriate percentage factor purchase and refinance mortgages, provided under § 81.17. mortgage purchases eligible to be respectively) by the number of single- counted toward the respective housing (2)(i) When the income of the family owner-occupied units in mortgagor(s) is not available to goal, then that methodology may be properties securing mortgages used up to a nationwide maximum of determine whether an owner-occupied purchased by the GSE for each census ten percent of the total number of rental unit in a property securing a single- tract, summed up over all census tracts. units in properties securing multifamily family mortgage originated after 1992 If this nationwide maximum is mortgages purchased by the GSE in the and purchased by a GSE counts toward exceeded, then the estimated number of current year. If this maximum is achievement of the Low- and Moderate- goal-qualifying units will be adjusted by exceeded, the estimated number of goal- Income Housing Goal or the Special the ratio of the applicable nationwide qualifying units will be adjusted by the Affordable Housing Goal, a GSE’s maximum number of units for which ratio of the nationwide maximum performance with respect to such unit income information may be estimated to number of units for which affordability may be evaluated using estimated the total number of single-family owner- information may be estimated to the affordability information in accordance occupied units with missing income with one of the following methods: information in properties securing total number of multifamily rental units (A) Excluding from the denominator mortgages purchased by the GSE. with missing affordability information and the numerator single-family owner- Owner-occupied units in excess of the in properties securing mortgages occupied units located in census tracts nationwide maximum, and any units for purchased by the GSE. Multifamily with median incomes less than, or equal which estimation information is not rental units in excess of the maximum to, area median income based on the available, shall remain in the set forth in this paragraph (e)(6)(i)(C), most recent decennial census, up to a denominator of the respective goal and any units for which estimation maximum of one percent of the total calculation. information is not available, shall be number of single-family owner- (e) * * * removed from the denominator of the occupied dwelling units eligible to be (6) * * * respective goal calculation. counted toward the respective housing (i) Multifamily. (A) When a GSE lacks (ii) Rental units in 1–4 unit single- goal in the current year. Mortgage sufficient information to determine family properties. (A) When a GSE lacks purchases with missing data in excess of whether a rental unit in a property sufficient information to determine the maximum will be included in the securing a multifamily mortgage whether a rental unit in a property denominator and excluded from the purchased by a GSE counts toward securing a single-family mortgage numerator; achievement of the Low- and Moderate- purchased by a GSE counts toward (B) For home purchase mortgages and Income Housing Goal or the Special achievement of the Low- and Moderate- for refinance mortgages separately, Affordable Housing Goal because Income Housing Goal or the Special multiplying the number of owner- neither the income of prospective or Affordable Housing Goal because occupied units with missing borrower actual tenants, nor the actual or average neither the income of prospective or income information in properties rental data, are available, a GSE’s actual tenants, nor the actual or average securing mortgages purchased by the performance with respect to such unit rental data, are available, a GSE’s GSE in each census tract by the may be evaluated using estimated performance with respect to such unit percentage of all single-family owner- affordability information in accordance may be evaluated using estimated occupied mortgage originations in the with one of the following methods: affordability information in accordance respective tracts that would count (1) Multiplying the number of rental with one of the following methods: toward achievement of each goal, as units with missing affordability (1) Excluding rental units in 1-to 4- determined by HUD based on the most information in properties securing unit properties with missing recent HMDA data available; or multifamily mortgages purchased by the affordability information from the (C) Such other data source and GSE in each census tract by the denominator as well as the numerator in methodology as may be approved by percentage of all rental dwelling units in calculating performance under those HUD. the respective tracts that would count goals;
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00063 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63642 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
(2) Multiplying the number of rental §§ 81.12 through 81.14. However, (14) Seller dissolution option. (i) units with missing affordability performance under the subgoals shall be Mortgages acquired through transactions information in properties securing counted using a fraction that is involving seller dissolution options single family mortgages purchased by converted into a percentage for each shall be treated as mortgage purchases, the GSE in each census tract by the subgoal and the numerator of the and receive credit toward the percentage of all rental dwelling units in fraction for each subgoal shall be the achievement of the housing goals, only the respective tracts that would count number of home purchase mortgages in when: toward achievement of each goal, as metropolitan areas financed by each (A) The terms of the transaction determined by HUD based on the most GSE’s mortgage purchases in a provide for a lockout period that recent decennial census. For units with particular year that count towards prohibits the exercise of the dissolution missing affordability information in achievement of the applicable housing option for at least one year from the date tracts for which such methodology is goal. The denominator of each fraction on which the transaction was entered not possible, such units will be shall be the total number of home into by the GSE and the seller of the excluded from the denominator as well purchase mortgages in metropolitan mortgages; and as the numerator in calculating areas financed by each GSE’s mortgage (B) The transaction is not dissolved performance under the respective purchases in a particular year. For during the one-year minimum lockout housing goal(s); or purposes of each subgoal, the procedure period. (3) Such other data source and for addressing missing data or (ii) The Secretary may grant an methodology as may be approved by information, as set forth in paragraph (d) exception to the one-year minimum HUD. of this section, shall be implemented lockout period described in paragraph (B) In any calendar year, a GSE may using numbers of home purchase (c)(14)(i)(A) and (B) of this section, in use only one of the methods specified mortgages in metropolitan areas and not response to a written request from an in paragraph (e)(6)(ii)(A) of this section single-family owner-occupied dwelling enterprise, if the Secretary determines to estimate affordability information for units. that the transaction furthers the single-family rental units. (2) Special counting rule for purposes of FHEFSSA and the GSE’s (C) If a GSE chooses to use an mortgages with more than one owner- charter act; (iii) For purposes of this paragraph estimation methodology under occupied unit. For purposes of counting (c)(14), ‘‘seller dissolution option’’ paragraph (e)(6)(ii)(A)(2) or mortgages toward the Home Purchase means an option for a seller of (e)(6)(ii)(A)(3) of this section to Subgoals, where a single home purchase mortgages to the GSEs to dissolve or determine affordability for rental units mortgage finances the purchase of two otherwise cancel a mortgage purchase in properties securing single-family or more owner-occupied units in a agreement or loan sale. mortgage purchases eligible to be metropolitan area, the mortgage shall counted toward the respective housing count once toward each subgoal that * * * * * goal, then that methodology may be applies to the GSE’s mortgage purchase. I 8. Revise § 81.102 to read as follows: used up to nationwide maximums of I 7. In § 81.16, revise paragraphs (c)(6) and (c)(7), remove and reserve § 81.102 Verification and enforcement to five percent of the total number of rental ensure GSE data integrity. units in properties securing non- paragraphs (c)(10) and (c)(11), and add a paragraph (c)(14), to read as follows: (a) Independent verification authority. seasoned single-family mortgage The Secretary may independently verify purchases by the GSE in the current § 81.16 Special counting requirements. the accuracy and completeness of the year and 20 percent of the total number * * * * * data, information, and reports provided of rental units in properties securing (c) * * * by each GSE, including conducting on- seasoned single-family mortgage (6) Seasoned mortgages. A GSE’s site verification, when such steps are purchases by the GSE in the current purchase of a seasoned mortgage shall reasonably related to determining year. If either or both of these be treated as a mortgage purchase for whether a GSE is complying with 12 maximums are exceeded, the estimated purposes of these goals and shall be U.S.C. 4541–4589 and the GSE’s Charter number of goal-qualifying units will be included in the numerator, as Act. adjusted by the ratio of the applicable appropriate, and the denominator in (b) Certification. (1) The senior officer nationwide maximum number of units calculating the GSE’s performance of each GSE who is responsible for for which affordability information may under the housing goals, except where: submitting to HUD the fourth quarter be estimated to the total number of (i) The GSE has already counted the Annual Mortgage Report and the AHAR single-family rental units with missing mortgage under a housing goal under sections 309(m) and (n) of the affordability information in properties applicable to 1993 or any subsequent Fannie Mae Charter Act or sections securing seasoned or unseasoned year; or 307(e) and (f) of the Freddie Mac Act, mortgages purchased by the GSE, as (ii) HUD determines, based upon a as applicable, or for submitting to the applicable. Single-family rental units in written request by a GSE, that a Secretary such other report(s), data, or excess of the maximums set forth in this seasoned mortgage or class of such information for which certification is paragraph (e)(6)(ii)(C), and any units for mortgages should be excluded from the requested in writing by the Secretary, which estimation information is not numerator and the denominator in order shall certify such report(s), data or available, shall be removed from the to further the purposes of the Special information. denominator of the respective goal Affordable Housing Goal. (2) The certification shall state as calculation. (7) Purchase of refinanced mortgages. follows: ‘‘To the best of my knowledge * * * * * Except as otherwise provided in this and belief, the information provided (i) Counting mortgages toward the part, the purchase of a refinanced herein is true, correct and complete.’’ Home Purchase Subgoals. (1) General. mortgage by a GSE is a mortgage (3) If the Secretary determines that a The requirements of this section, except purchase and shall count toward GSE has failed to provide the for paragraphs (b) and (e) of this section, achievement of the housing goals to the certification required by paragraphs shall apply to counting mortgages extent the mortgage qualifies. (b)(1) and (b)(2) of this section, or that toward the Home Purchase Subgoals at * * * * * a GSE has provided the certification
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00064 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63643
required by paragraph (b) in connection and any extension period, the Secretary under sections 309(m) and (n) of the with data, information or report(s) that will notify the GSE in writing and seek Fannie Mae Charter Act or sections the Secretary later determines are not clarification or additional information to 307(e) and (f) of the Freddie Mac Act, true, correct and complete, the Secretary correct the error, omission or as applicable. An error, omission or may pursue the enforcement remedies discrepancy. The GSE shall have 10 discrepancy is material if it results in an under paragraph (e) of this section. For working days (or such longer period as overstatement of credit for a housing data, information or report(s) subject to the Secretary may establish, not to goal or Special Affordable subgoal and, paragraphs (c) or (d) of this section, the exceed 30 working days) from the date without such overstatement, the GSE Secretary may pursue the enforcement of the Secretary’s written notice to would have failed to meet such housing remedies described in paragraph (e) respond in writing to the notice. If the goal or Special Affordable subgoal for only in connection with material errors, GSE fails to submit a written response the prior year. A ‘‘prior year’’ for omissions or discrepancies as those to the Secretary within this period, or if purposes of this section is any one of terms are defined in § 81.102(c) or (d). the Secretary determines that the GSE’s the two years immediately preceding (c) Verification procedure and written response fails to correct or the latest year for which data on adjustment to correct errors, omissions otherwise resolve each error, omission housing goals performance was reported or discrepancies in AHAR data for the or discrepancy in its reported year-end to HUD. immediately preceding year. (1) This data to the Secretary’s satisfaction, the (2) Procedural requirements. In the paragraph (c) pertains to the GSEs’ Secretary will determine the appropriate event the Secretary determines that a submission of year-end data. For adjustments to the numerator and the GSE’s prior year’s fourth quarter Annual purposes of this paragraph, ‘‘year-end denominator of the applicable housing Mortgage Report or AHAR contain a data’’ means data that HUD receives goal(s) and Special Affordable material error, omission or discrepancy, from the GSEs related to housing goals subgoal(s) due to the GSE’s failure to the Secretary will provide the GSE with performance in the immediately provide the Secretary with accurate an initial letter containing written preceding year and covering data submissions of data. findings and determinations within 24 reported in the fourth quarter Annual (4) The Secretary, or his or her months of the end of the relevant GSE Mortgage Report and the GSE’s AHAR. designee, shall inform a GSE in writing, reporting year. The GSE shall have an An ‘‘error’’ means a technical mistake, at least five working days prior to HUD’s opportunity, not to exceed 30 days from such as a mistake in coding or release of its official performance figures the date of receipt of the Secretary’s calculating data. An ‘‘omission’’ means to the public, of HUD’s determination of initial letter, to respond in writing with a GSE’s failure to count units in the official goals performance figures, supporting documentation, to contest denominator. A ‘‘discrepancy’’ means including any adjustments. During the the Secretary’s initial determination that any difference between HUD’s analysis five working days prior to such public there was a material error, omission or of data and the analysis contained in a release, a GSE may request, in writing, discrepancy in a prior year’s data. The GSE’s submission of data, including a a reconsideration of HUD’s final Secretary shall then issue a final discrepancy in goal or Special determination of its performance and determination letter within 60 days of Affordable subgoal performance. must provide the basis for requesting the date of HUD’s receipt of the GSE’s (2) If HUD finds errors, omissions or the reconsideration. If the request is written response or, if no response is discrepancies in a GSE’s year-end data granted, the Secretary will consider the received, within 90 days of the date of submissions relative to HUD’s GSE’s request for reconsideration of its the GSE’s receipt of the Secretary’s regulations, HUD will first notify the determination of goals performance and initial letter. The Secretary may extend GSE by telephone or e-mail make a final determination regarding the period for issuing a final transmission of each such error, the GSE’s performance, within 10 determination letter by an additional 30 omission or discrepancy. The GSE must working days of the Secretary’s granting days and may grant the GSE an respond within five working days of of the GSE’s written request for opportunity, for a period not to exceed each such notification. HUD may, in its reconsideration. 10 working days from the date of the discretion or upon a request by a GSE (5) Should the Secretary determine GSE’s receipt of the determination letter within the five working day period, that additional enforcement action to request that the determination be extend the response period for up to an against the GSE is warranted for reconsidered. additional 20 working days. Information material errors, omissions or (3) If the Secretary determines that a exchanges during the five working day discrepancies with regard to a housing GSE’s prior year’s fourth quarter Annual period following initial notification, and goal or Special Affordable subgoal, it Mortgage Report or AHAR contained a any subsequent extensions of time that may pursue additional remedies under material error, omission or discrepancy, may be granted, may be by electronic paragraph (e) of this section. An error, the Secretary may direct the GSE to mail. Any person with delegated omission or discrepancy is material if it correct the overstatement by purchasing authority from the Secretary, or the results in an overstatement of credit for mortgages to finance the number of Director of HUD’s Financial Institution a housing goal or Special Affordable units that HUD has determined were Regulation Division, or his or her subgoal, and, without such overstated in the prior year’s goal designee, shall be responsible for overstatement, the GSE would have performance (or, for the Special issuing initial notifications regarding failed to meet such housing goal or Affordable subgoal, the number or errors, omissions, or discrepancies; Special Affordable subgoal for the dollar amount, as applicable, of making determinations on the adequacy immediately preceding year. mortgage purchases that HUD has of responses received; approving any (d) Adjustment to correct prior year determined were overstated), or that extensions of time permitted under this reporting errors, omissions or equal the percentage of the provision; and managing the data discrepancies. (1) General. The overstatement in the prior year’s goal or verification process. Secretary may require a GSE to correct Special Affordable subgoal performance (3) If each error, omission or a material error, omission or as applied to the most current year-end discrepancy is not resolved to HUD’s discrepancy in a GSE’s prior year’s data performance, whichever is less. Units or satisfaction during the initial five reported in the fourth quarter Annual mortgages purchased to remedy an working day period from notification, Mortgage Report and the GSE’s AHAR overstatement in the housing goals or
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00065 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63644 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
the Special Affordable subgoal must be with material errors, omissions or Appendix A eligible to qualify under the same goal discrepancies arising under paragraph Departmental Considerations To Establish or Special Affordable subgoal that HUD (d) of this section if the GSE has failed The Low- and Moderate-Income Housing has determined were overstated in the to purchase a sufficient amount or type Goal prior year. of mortgages, as provided in paragraphs A. Introduction (4) If a GSE does not purchase a (d)(3) and (d)(4) of this section. sufficient amount or type of mortgages Sections 1 and 2 provide a basic (2) Remedies. (i) Submissions to meet the requirements set forth in description of the rule process. Section 3 required under the GSE’s charter acts. paragraph (d)(3) of this section as discusses HUD’s conclusions based on After the Secretary makes a directed by the Secretary by no later consideration of the factors. than the end of the calendar year determination under paragraph (e)(1) of 1. Establishment of Low- and Moderate- immediately following the year in this section that any of the Income Housing Goal which the Secretary notifies the GSE of circumstances described in paragraphs In establishing the Low- and Moderate- such overstatement (unless, upon (e)(1)(i) or (ii) has occurred with respect Income Housing Goals for the Federal written request from the GSE, the to data, information, or report(s) National Mortgage Association (Fannie Mae) Secretary, in his or her discretion, required by sections 309(m) or (n) of the and the Federal Home Loan Mortgage Fannie Mae Charter Act, or by sections Corporation (Freddie Mac), collectively determines that a grant of additional referred to as the Government-Sponsored time is appropriate to correct or 307(e) or (f) of the Freddie Mac Act, the Secretary may pursue any or all of the Enterprises (GSEs), section 1332 of the compensate for the overstatement) the Federal Housing Enterprises Financial Safety Department may pursue any or all of the following remedies in accordance with and Soundness Act of 1992 (12 U.S.C. 4562) following remedies: paragraph (e)(3), or applicable law, as (FHEFSSA) requires the Secretary to (i) Issue a notice that the GSE has appropriate: consider: failed a housing goal or Special (A) A cease-and-desist order against 1. National housing needs; Affordable subgoal in the prior year; the GSE for failing to submit the 2. Economic, housing, and demographic (ii) Seek additional enforcement required data, information or report(s) conditions; remedies under paragraph (e) of this 3. The performance and effort of the in accordance with this section; enterprises toward achieving the Low- and section; Moderate-Income Housing Goal in previous (iii) Pursue any other civil or (B) Civil money penalties against the GSE for failing to submit the required years; administrative remedies as are available 4. The size of the conventional mortgage to it. data, information or report(s) in accordance with this section; market serving low- and moderate-income (e) Additional enforcement options. families relative to the size of the overall (1) General. In the event the Secretary (C) Any other civil or administrative conventional mortgage market; determines, either as a result of his or remedies or penalties against the GSE 5. The ability of the enterprises to lead the her independent verification authority that may be available to the Secretary by industry in making mortgage credit available described in paragraph (a) of this virtue of the GSE’s failing to submit or for low- and moderate-income families; and section, or by the authority set forth in certify the required data, information or 6. The need to maintain the sound financial condition of the enterprises. paragraphs (b), (c) or (d) of this section, report(s) in accordance with this The Secretary also considered these factors that any of the following circumstances section. in establishing a low- and moderate-income has occurred with respect to data, (ii) Submissions required under subgoal for home purchase loans on single- information or report(s) required by subpart E of this part. After the family-owner properties in metropolitan sections 309(m) or (n) of the Fannie Mae Secretary makes a determination under areas. Charter Act, sections 307(e) or (f) of the paragraph (e)(1) of this section that any 2. Underlying Data Freddie Mac Act, or subpart E of this of the circumstances described in In considering the statutory factors in part, the Secretary may regard this as a paragraphs (e)(1)(i) or (ii) has occurred establishing these goals, HUD relied on data GSE’s failure to submit such data, with respect to data, information or from the 2001 American Housing Survey, the information or report(s) and, report(s) required under subpart E of 2000 Censuses of Population and Housing, accordingly, the Secretary may take the this part (but that are not required by the 2001 Residential Finance Survey (RFS), additional enforcement actions the 1995 Property Owners and Managers sections 309(m) or (n) of the Fannie Mae Survey (POMS), other government reports, authorized by paragraph (e)(2) of this Charter Act or by sections 307(e) or (f) section: reports submitted in accordance with the of the Freddie Mac Act), the Secretary Home Mortgage Disclosure Act (HMDA), and (i) A GSE fails to submit the may pursue any civil or administrative the GSEs. In order to measure performance certification required by paragraphs remedies or penalties against the GSE toward achieving the Low- and Moderate- (b)(1) and (b)(2) of this section in that may be available to the Secretary. Income Housing Goal in previous years, HUD connection with such data, information The Secretary shall pursue such analyzed the loan-level data on all mortgages or report(s); or remedies under applicable law. purchased by the GSEs for 1993–2003 in (ii) A GSE submits the certification accordance with the goal counting provisions required by paragraph (b) of this section, (3) Procedures. The Secretary shall established by the Department in the but the Secretary later determines that comply with the procedures set forth in December 1995 and October 2000 rules (24 the data, information or report(s) are not subpart G of this part in connection CFR part 81). true, correct and complete. For data, with any enforcement action that he or 3. Conclusions Based on Consideration of the information or report(s) subject to she may initiate against a GSE under Factors paragraphs (c) or (d) of this section, the paragraph (e) of this section. The discussion of the first two factors covers a range of topics on housing needs Secretary may pursue the additional Dated: October 22, 2004. and economic and demographic trends that enforcement remedies under paragraph John C. Weicher, (e)(2) only in connection with material are important for understanding mortgage markets. Information is provided which errors, omissions or discrepancies, as Assistant Secretary for Housing—Federal Housing Commissioner. describes the market environment in which those terms are defined in § 81.102(c) or the GSEs must operate (for example, trends (d). In addition, the Secretary may only Note: The Appendices will not appear in in refinancing activity). In addition, the pursue such remedies in connection the Code of Federal Regulations. severe housing problems faced by lower-
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00066 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63645
income families are discussed, as are the discrimination in housing markets and would have been 50.6 percent in 2000, 47.7 barriers that minorities face when attempting mortgage lending. percent in 2001, 46.1 percent in 2002, and to become homeowners. This discussion • A HUD-published study of 45.0 percent in 2003. Thus, both Fannie Mae serves to provide useful background discrimination in the rental and owner and Freddie Mac would have surpassed the information for the discussion of the markets found that while differential 50 percent goal in 2000 and fallen short in Geographically Targeted and Special treatment between minority and white home 2001, 2002, and 2003. Affordable Housing Goals in Appendixes B seekers had declined over the past ten years, • This Appendix includes a and C, as well as for the Low- and Moderate- it continued at an unacceptable level in the comprehensive analysis of each GSE’s Income Housing Goal in this Appendix. year 2000. In addition, disparities in performance in funding home purchase The third factor (past performance) and the mortgage lending continued across the nation mortgages for borrowers and neighborhoods fifth factor (ability of the GSEs to lead the in 2002, when the loan denial rate was 7.8 covered by the three housing goals—special industry) are also discussed in some detail in percent for white mortgage applicants, but affordable and low- and moderate-income this Appendix. With respect to home 20.1 percent for African Americans and 15.5 borrowers and underserved. The GSEs’ purchase mortgages, the past performance of percent for Hispanics.1 performance in funding first-time home the GSEs and their ability to lead the • Americans with the lowest incomes face buyers is also examined. industry are examined for all three housing persistent housing problems. Recent HUD • While Freddie Mac has improved its goals; that analysis provides the basis for analysis reveals that in 2001, 5.1 million affordable lending performance in recent establishing the three subgoals for the GSEs’ households had ‘‘worst case’’ housing needs, years, it has consistently lagged the acquisitions of home loans on single-family- defined as housing costs greater than 50 conventional conforming market in funding owner properties. percent of household income or severely affordable home purchase loans for special The fourth factor (size of the market) and inadequate housing among unassisted very- affordable and low-moderate-income the sixth factor (need to maintain the GSEs’ low-income renter households. Among these borrowers and underserved neighborhoods sound financial condition) are mentioned households, 90 percent had a severe rent targeted by the housing goals.2 In 2003, its only briefly in this Appendix. Detailed burden, 6 percent lived in severely performance on the underserved areas goal analyses of the fourth factor and the sixth inadequate housing, and 4 percent suffered was particularly low relative to both the factor are contained in Appendix D and in from both problems. performances of Fannie Mae and the market; the economic analysis of this rule, • Over the past ten years, there has been in that year, underserved area loans respectively. a ‘‘revolution in affordable lending’’ that has accounted for only 24.0 percent of Freddie The factors are discussed in sections B extended homeownership opportunities to Mac’s purchases compared with 26.8 percent through H of this appendix. Section I historically underserved households. Fannie of Fannie Mae’s purchases and 27.6 percent summarizes the findings and presents the Mae and Freddie Mac have been a substantial of market originations. (These underserved Department’s conclusions concerning the part of this ‘‘revolution in affordable area data are based on 1990 Census Low- and Moderate-Income Housing Goal. geography.) lending’’. During the mid-to-late 1990s, they • Section I also gives the rationale for a low- added flexibility to their underwriting In general, Fannie Mae’s affordable and moderate-income subgoal for home guidelines, introduced new low-down- lending performance has been better than purchase loans. payment products, and worked to expand the Freddie Mac’s. But like Freddie Mac, Fannie The consideration of the factors in this use of automated underwriting in evaluating Mae’s average performance during past Appendix has led the Secretary to the the creditworthiness of loan applicants. periods (e.g., 1993–2003, 1996–2003, 1999– following conclusions: HMDA data suggest that the industry and 2003) has been below market levels. • Changing population demographics will GSE initiatives are increasing the flow of However, it is encouraging that Fannie Mae markedly improved its affordable lending result in a need for primary and secondary credit to underserved borrowers. Between performance relative to the market during mortgage markets to meet nontraditional 1993 and 2003, conventional loans to low- 2001, 2002, and 2003, the first three years credit needs, respond to diverse housing income and minority families increased at under the higher housing goal targets that preferences, and overcome information and much faster rates than loans to upper-income HUD established in the GSE Final Rule dated other barriers that many immigrants and and non-minority families. October 2000. Over this three-year period, minorities face. Growing housing demand • The Low- and Moderate-Income Goal Fannie Mae led the primary market in from immigrants (both those who are already was set at 50 percent beginning in 2001. funding special affordable and low-mod here and those projected to come) and non- Effective on January 1, 2001, several changes loans but lagged the market in funding traditional homebuyers will help to offset in counting requirements came into effect, declines in the demand for housing caused underserved areas loans. In 2003, Fannie including (1) ‘‘bonus points’’ (double credit) Mae’s increased performance placed it by the aging of the population. Immigrants for purchases of mortgages on small (5–50 and other minorities—who accounted for significantly above the special affordable unit) multifamily properties and, above a market (a 17.1 percent share for Fannie Mae more than a third of household growth since threshold level, mortgages on 2–4 unit the 1990s—will be responsible for almost compared with a 15.9 percent share for the owner-occupied properties; and (2) a market) and the low-mod market (a 47.0 two-thirds of the growth in the number of ‘‘temporary adjustment factor’’ (1.35 units new households over the next ten years. As percent share for Fannie Mae compared with credit) for Freddie Mac’s purchases of a 44.6 percent share for the market). these demographic factors play out, the mortgages on large (>50 unit) multifamily overall effect on housing demand will likely However, Fannie Mae continued to lag the properties. With these two counting rules, underserved areas market in 2003 (a 26.8 be sustained growth and an increasingly Fannie Mae’s performance was 51.5 percent diverse household population from which to percent share for Fannie Mae compared with in 2001, 51.8 percent in 2002 and 52.3 a 27.6 percent share for the market). In this draw new renters and homeowners. percent in 2003, and Freddie Mac’s • case, which is referred to in the text as the Despite the record national performance was 53.2 percent in 2001, 50.5 homeownership rate of 68.3 percent in 2003, ‘‘purchase year’’ approach, Fannie Mae’s percent in 2002, and 51.2 percent in 2003; performance is based on comparing its much lower rates prevailed for minorities, thus, both GSEs surpassed this higher goal in especially for African-American households purchases of all loans (both seasoned loans all three years. and newly-originated mortgages) during a (48.4 percent) and Hispanics (47.4 percent), • The bonuses and temporary adjustment and these lower rates are only partly particular year with loans originated in the factor expired at the end of 2003. Without market in that year. When Fannie Mae’s accounted for by differences in income, age, these rules, Fannie Mae’s performance would and other socioeconomic factors. have been 51.3 percent in 2000, 49.2 percent • 2 The ‘‘affordable lending performance’’ of Fannie In addition to low incomes, barriers to in 2001, 49.0 percent in 2002, and 48.7 Mae and Freddie Mae refers to the performance of homeownership that disproportionately percent in 2003. Freddie Mac’s performance affect minorities and immigrants include lack the GSEs in funding loans for low-income and underserved borrowers through their purchase (or of capital for down payments and closing 1 Mortgage denial rates are based on 2002 HMDA guarantee) of loans originated by primary lenders. costs, poor credit history, lack of access to data for home purchase loans; manufactured It does not, of course, imply that the GSEs mainstream lenders, little understanding of housing lenders are excluded from these themselves are lenders originating loans in the the homebuying process, and continued comparisons. primary market.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00067 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63646 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
performance is measured on an ‘‘origination markets. Government action—through land for minority households was lower—for year’’ basis (that is, allocating Fannie Mae’s use regulation, building codes, and example, just 48.4 percent of African- purchases in a particular year to the year that occupancy standards—are major contributors American households and 47.4 percent of the purchased loan was originated), Fannie to those high costs. Hispanic households owned a home.3 Mae also led the 2003 market in funding • The market for financing multifamily Differences in income and age between special affordable and low- and moderate- apartments has grown to record volumes. minorities and whites do not fully explain income loans, and lagged the market in Fannie Mae and Freddie Mac have been these gaps. The Joint Center for Housing funding underserved area loans. among those boosting volumes and Studies estimated that if minorities owned • Both Fannie Mae and Freddie Mac lag introducing new programs to serve the homes at the same rates as whites of similar the conventional conforming market in multifamily market. Fannie Mae’s age and income, a homeownership gap of 10 funding first-time homebuyers, and by a multifamily purchases jumped from about percentage points would still exist.4 rather wide margin. Between 1999 and 2001, $10 billion in 1999 and 2000 to $18.7 billion first-time homebuyers accounted for 27 in 2001, $18.3 billion in 2002, and $33.3 a. Importance of Homeownership percent of each GSE’s purchases of home billion in 2003—the last three years were Homeownership is one of the most loans, compared with 38 percent for home characterized by heavy refinancing activity. common forms of property ownership as well loans originated in the conventional • Freddie Mac has re-entered the as savings.5 Historically, home equity has conforming market. multifamily market, after withdrawing for a been the largest source of wealth for most • The GSEs have accounted for a time in the early 1990s. Concerns regarding Americans, and wealth gains in housing have significant share of the total (government as Freddie Mac’s multifamily capabilities no been more widely distributed among the well as conventional) market for home longer constrain its performance with regard population than gains in the stock market.6 purchase loans, but their market share for to the housing goals. Freddie Mac’s With stocks appreciating faster than home each of the affordable lending categories (e.g., multifamily purchases increased from a prices over the past decade, home equity as low-income borrowers and census tracts) has relatively low $3 billion in 1997 to a share of all family assets fell from 38 been less than their share of the overall approximately $7 billion during the next percent in 1989 to 33 percent in 1998 and 32 market. three years (1998 to 2000), before rising percent in 2001.7 However, many of the gains • The GSEs also account for a very small further to $11.9 billion in 2001, $13.3 billion in the stock market were erased after 1999 share of the market for important groups such in 2002, and $21.6 billion in 2003. and housing was once again a more as minority first-time homebuyers. • The overall presence of both GSEs in the significant asset in the household balance Considering the total mortgage market (both rental mortgage market falls short of their sheet than stocks in 2001.8 Even with a bull government and conventional loans), it is involvement in the single-family owner market through most of the 1990s, 59 percent estimated that the GSEs purchased only 14 market. Between 1999 and 2002, the GSEs’ of all homeowners in 1998 held more than percent of loans originated between 1999 and purchases totaled for 61 percent of the owner half of their net wealth in the form of home 2001 for African-American and Hispanic market, but only 37 percent of the single- equity.9 From 2001 to 2003, homes prices first-time homebuyers, or one-third of their family rental and multifamily rental market. appreciated an average of 23 percent which share (42 percent) of all home purchase loans Certainly there is room for expansion of the meant $30,900 in housing equity originated during that period. Considering GSEs in supporting the nation’s rental accumulation for a typical homeowner.10 the conventional conforming market and the markets, and that expansion is needed if the Moreover, unlike stock wealth, aggregate same time period, it is estimated that the GSEs are to make significant progress in home equity has steadily increased over the GSEs purchased only 31 percent of loans closing the gaps between the affordability of past 40 years with only occasional small originated for African-American and their mortgage purchases and that of the dips.11 Hispanic first-time homebuyers, or overall conventional conforming market. Among low-income homeowners • approximately one-half of their share (57 Considering both owner and rental (household income less than $20,000), home percent) of all home purchase loans in that properties, the GSEs’ presence in the goals- equity accounted for about 72 percent of market. The GSEs’ small share of the first- qualifying market has been significantly less household wealth, and approximately 55 time homebuyer market could be due to the than their presence in the overall percent for homeowners with incomes preponderance of high (over 20 percent) conventional conforming mortgage market. between $20,000 and $50,000. Median net downpayment loans in their mortgage Specifically, HUD estimates that the GSEs wealth for low-income homeowners under 65 purchases. accounted for 55 percent of all owner and • rental units financed in the primary market This Appendix discusses the dynamic 3 Joint Center for Housing Studies of Harvard nature of the single-family mortgage market between 1999 and 2002, but only 48 percent of units qualifying for the low-mod goal, 48 University, State of the Nation’s Housing 2004, p. and the numerous changes that this market 35. percent of units qualifying for the has undergone over the past few years. Some 4 Joint Center for Housing Studies of Harvard important trends that will likely factor into underserved areas goal, and 41 percent of University, State of the Nation’s Housing 2003, p. the GSEs’ performance in meeting the needs units qualifying for special affordable goal. 16. of underserved borrowers include the growth B. Factor 1: National Housing Needs 5 According to the National Association of of the subprime market, the increasing use of Realtors, Housing Market Will Change in New This section reviews the general housing automated underwriting systems, and the Millennium as Population Shifts, November 7, needs of lower-income families that exist introduction of risk-based pricing into the 1998. Forty-five percent of U.S. household wealth today and are expected to continue in the was in the form of home equity in 1998. Since 1968, market. near future. Affordability problems that home prices have increased each year, on average, • The long run outlook for the multifamily lower-income families face in both the rental at the rate of inflation plus two percentage points. rental market is sustained, moderate growth, and owner markets are examined. The 6 Todd Buchholz, ‘‘Safe At Home: The New Role based on favorable demographics. The section also describes racial disparities in of Housing in the U.S. Economy,’’ a paper minority population, especially Hispanics, homeownership and the causes of these commissioned by the Homeownership Alliance, 2002. provides a growing source of demand for disparities. It also notes some special affordable rental housing. ‘‘Lifestyle renters’’ 7 Federal Reserve Board, ‘‘Recent Changes in U.S. problems, such as the need to rehabilitate our Family Finances: Results from the 1998 and 2001 (older, middle-income households) are also a older urban housing stock, that are discussed fast-growing segment of the rental Survey of Consumer Finances,’’ January 2003, p. 16. throughout this appendix. 8 population. Provision of affordable housing, Mark Zandi, ‘‘Housing’s Rising Contribution,’’ 1. Homeownership Gaps June 2002, p. 5. however, will continue to challenge 9 suppliers of multifamily rental housing and Joint Center for Housing Studies of Harvard Despite recent record homeownership University, State of the Nation’s Housing 1998. policy makers at all levels of government. rates, many Americans, including 10 Lawrence Yun, ‘‘The Forecast,’’ National Low incomes combined with high housing disproportionate numbers of racial and Association of Realtors Real Estate Outlook, costs define a difficult situation for millions ethnic minorities, are shut out of February 2004, p. 4. of renter households. Housing cost homeownership opportunities. Although the 11 Joint Center for Housing Studies of Harvard reductions are constrained by high land national homeownership rate for all University, State of the Nation’s Housing 2004, p. prices and construction costs in many Americans was 68.3 percent in 2003, the rate 15.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00068 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63647
was twelve times that of a similar renter.12 homeownership for younger families. As obstacle to buying a home versus 32 percent Thus a homeownership gap continues to home prices skyrocketed during the late of all Americans.22 A study by Gyourko, translate directly into a wealth gap. For this 1970s and early 1980s, real incomes also Linneman, and Wachter found significant reason, President Bush issued the stagnated, with earnings growth particularly racial differences in homeownership rates in ‘‘Homeownership Challenge’’ in June 2002 to slow for blue collar and less educated ‘‘wealth-constrained’’ households while increase minority homeownership by 5.5 workers. Through most of the 1980s, the finding no racial differences in million by the end of the decade. By combination of slow income growth and homeownership rates among households December of 2003, the Census estimated that increasing rents made saving for home with wealth sufficient to meet down payment the number of minority homeowners had purchase more difficult, and relatively high and closing costs.23 Minorities and increased by 1.53 million. Meaning that in interest rates required large fractions of immigrants are much less likely to receive the fourth quarter of 2003, for the first time family income for home mortgage payments. gifts and inheritances from their parents to ever, the majority of minority households are Thus, during that period, fewer households assist them in becoming a homeowner. homeowners.13 had the financial resources to meet down (ii) Poor Credit History. Poor credit history High rates of homeownership support payment requirements, closing costs, and also differentially affects minority economic stability within housing and monthly mortgage payments. households. In the same Fannie Mae survey, related industries, sectors that contributed Economic expansion and lower mortgage nearly a third of African-American nearly one-third of the total gain in real GDP rates substantially improved homeownership respondents said their credit rating would be since the beginning of the decade.14 In affordability during the 1990s. Many young, an obstacle to buying a home versus 23 addition, more than half of the refinancing low-income, and minority families who were percent of all Americans.24 Because African- mortgages in the first two years of the decade closed out of the housing market during the American and Hispanic borrowers are more were cash-out, defined as refinancing 1980s re-entered the housing market during likely than others to have little traditional procedures by which the mortgage balance is the last decade. Even with an economic credit history or a poorer credit history, they increased by more than five percent in order slowdown in 2000–2001 and climbing house face increased difficulties in being accepted to tap into home equity. Cash-outs injected appreciation in 2002–2003, after-tax for mortgage credit. This is because credit more than $300 billion into the economy mortgage payments fell in 2003 for buyers of history scores (such as a FICO score) are a between 2000 and 2002 and were responsible median priced homes because of historically major component of the new automated for one-fifth of real GDP growth since during low interest rates.20 However, many mortgage scoring systems. These systems are that period.15 In addition to economic households still lack the earning power to more likely to refer minority borrowers for benefits such as jobs and residential take advantage of today’s home buying more intensive manual underwriting, rather investment, studies show that the better opportunities. Several trends have than to automatically accept them for the less living environment associated with owning a contributed to the reduction in the real costly, expedited processing. In these home has positive impacts on children, in earnings of young adults without college situations, there is the additional concern terms of lower rates of teenage pregnancy and education over the last 15 years, including that ‘‘referred’’ borrowers may not always higher reading other test scores. The current technological changes that favor white-collar receive a manual underwriting for the loan literature substantiates that the benefits of employment, losses of unionized that they initially applied for, but rather be homeownership extend beyond individual manufacturing jobs, and wage pressures directed to a high-cost subprime loan homeowners and their families to society at exerted by globalization. Over 42 percent of product. large. Homeownership promotes social and the nation’s population between the ages of (iii) Lack of Access to Mainstream Lenders. community stability by increasing the 25 and 34 had no advanced education in Minorities face heightened barriers in 21 number of stakeholders and reducing 2000 and were therefore at risk of being accessing credit because of their often limited disparities in the distributions of wealth and unable to afford homeownership. African access to mainstream lenders. Access to income. The empirical literature is generally Americans and Hispanics, who have lower lenders becomes difficult when mainstream supportive of a relationship between average levels of educational attainment than financial institutions are not located in whites, are especially disadvantaged by the homeownership and greater investment in neighborhoods where minorities live. The erosion in wages among less educated property.16 Homeownership is also growth in subprime lending over the last workers. associated with neighborhood stability (lower several years has benefited credit-impaired Immigrants and other minorities, who mobility), greater participation in voluntary borrowers—those who may have blemishes accounted for nearly 40 percent of the growth and political activities,17 and links to in their credit record, insufficient credit in the homeownership rate over the past five entrepreneurship.18 history, or non-traditional credit sources. years, will be responsible for two-thirds of Subprime lenders have allowed these 19 b. Barriers to Homeownership the growth in new households over the next borrowers to access credit that they could not Insufficient income, high debt burdens, ten years. These groups have unique housing otherwise obtain in the prime credit market. and limited savings are obstacles to needs and face numerous hurdles in However, studies by HUD, The Woodstock becoming homeowners. In addition to low Institute and others have shown that 12 U.S. Department of Housing and Urban income, barriers to homeownership that subprime lending is disproportionately Development, ‘‘Economic Benefits of Increasing disproportionately affect minorities and concentrated in low-income and minority Minority Homeownership,’’ p. 7. immigrants include: neighborhoods.25 While these studies 13 http://www.whitehouse.gov/infocus/ (1) Lack of capital for down payment and homeownership/. Accessed July 28, 2004. closing costs; 22 Fannie Mae, Fannie Mae National Housing 14 Homeownership Alliance, ‘‘The Economic (2) Poor credit history; Survey, 2002, p. 11. Contribution of the Mortgage Refinancing Boom,’’ (3) Lack of access to mainstream lenders; 23 Joseph Gyourko, Peter Linneman, and Susan December 2002, p. 2. (4) Complexity and fear of the home buying Wachter. ‘‘Analyzing the Relationships among Race, 15 Homeownership Alliance, ‘‘The Economic process; and, Wealth, and Home Ownership in America,’’ Journal Contribution of the Mortgage Refinancing Boom,’’ (5) Continued discrimination in housing of Housing Economics 8 (2), p. 63–89, as discussed December 2002, p. 4–5. markets and mortgage lending. in Thomas P. Boehm and Alan M. Schlottmann. 16 Robert Dietz and Donald Haurin, ‘‘The Social (i) Lack of Cash for Down Payment. In the ‘‘Housing and Wealth Accumulation: and Private Consequences of Homeownership,’’ 2002 Fannie Mae National Housing Survey, Intergenerational Impacts,’’ in Low-Income May 2001, p. 51. 40 percent of Hispanics reported not having Homeownership: Examining the Unexamined Goal, 17 William M. Rohe, George McCarthy, and Brookings Institution Press (2002), p. 408. enough money for a down payment as an Shannon Van Zandt, ‘‘The Social Benefits and Costs 24 Fannie Mae, Fannie Mae National Housing of Homeownereship,’’ May 2000, p. 31. Survey, 2002, p. 11. 18 U.S. Deparmtent of Housing and Urban Unexamined Goal, Washington, DC: Brookings 25 See Dan Immergluck, Stark Differences: The Development, ‘‘Economic Beneifts of Increasing Institution Press, 2002. Explosion of the Subprime Industry and Racial Minority Homeownership,’’ p. 8–9. 20 Joint Center for Housing Studies of Harvard Hypersegmentation in Home Equity Lending, 19 For a dicusssion of the causes of existing University, State of the Nation’s Housing 2004, p. Woodstock Institute, October 2000; and Daniel disparities in homeownership, see the various 15. Immergluck and Marti Wiles, Two Steps Back: The articles in Nicolas P. Retsinas and Eric S. Belsky 21 U.S. Census Bureau, Current Population Dual Mortgage Market, Predatory Lending, and the (Eds), Low-Income Homeowernsip: Examining the Survey, March 2000. Continued
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00069 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63648 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
recognize that differences in credit behavior Study (HDS) in the sale and rental of differential treatment discrimination at the explain some of the disparities in subprime housing. The study, entitled Discrimination pre-application level occurred at significant lending across neighborhoods, they argue in Metropolitan Housing Markets: National levels in at least some cities.31 Minorities that the absence of mainstream lenders has Results from Phase I of The Housing were less likely to receive information about also contributed to the concentration of Discrimination Study was conducted by the loan products, received less time and subprime lending in low-income and Urban Institute.29 This results of this HDS information from loan officers, and were minority neighborhoods. More competition were based on 4,600 paired tests of minority quoted higher interest rates in most of the by prime lenders in inner city neighborhoods and non-minority home seekers conducted cities where tests were conducted. A second could lower the borrowing costs of families during 2000 in 23 metropolitan areas HUD-sponsored study by the Urban Institute who currently have only the option of a high- nationwide. The report showed large used the paired testing methodology in Los cost subprime loan. This issue of the lack of decreases between 1989 and 2000 in the level Angeles and Chicago and found similar mainstream lenders in inner city of discrimination experienced by Hispanics results. African Americans and Hispanics neighborhoods is discussed further in and African Americans seeking to buy a faced a significant risk of unequal treatment subsection 2, below, in connection with home. There has also been a modest decrease when they visited mainstream mortgage disparities between neighborhoods. in discrimination toward African Americans lending institutions to make pre-application (iv) Complexity and Fear of Homebuying seeking to rent a unit. This downward trend, inquiries.32 Process. An additional barrier to however, has not been seen for Hispanic Several possible explanations for these homeownership is fear and a lack of renters, who now are more likely to lending disparities have been suggested. A understanding about the buying process and experience discrimination in their housing study by the Boston Federal Reserve Bank the risks of ownership. Many Americans search than do African-American renters. But found that racial disparities cannot be could become homeowners if provided with while generally down since 1989, the report explained by reported differences in information to correct myths, found that housing discrimination still exists creditworthiness.33 In other words, misinformation, and concerns about the at unacceptable levels. The greatest share of minorities are more likely to be denied than mortgage process. Some potential discrimination for Hispanic and African- whites with similar credit characteristics, homeowners, particularly minorities, are American home seekers can still be attributed which suggests lender discrimination. In unaware that they may already qualify for a to being told units are unavailable when they addition, loan officers, who may believe that mortgage they can afford. The 2002 Fannie are available to non-Hispanic whites, and race is correlated with credit risk, may use Mae survey revealed that 30 percent of being shown and told about fewer units than race as a screening device to save time, rather Americans believe erroneously that they comparable non-minority home seekers. than devote effort to distinguishing the need to pay 20 percent of the cost of a home Although discrimination is down on most creditworthiness of the individual up-front. In addition, Fannie Mae reported areas for African-American and Hispanic applicant.34 This violates the Fair Housing that half of Americans are only ‘‘somewhat’’ homebuyers, there remain worrisome upward Act. or ‘‘not at all’’ comfortable with mortgage trends of discrimination in the areas of Underwriting rigidities may fail to terms.26 Freddie Mac reports that six of 10 geographic steering for African Americans accommodate creditworthy low-income or Hispanics are uncomfortable with home and, relative to non-Hispanic whites, the minority applicants. For example, under buying terminology, and think they need amount of help agents provide to Hispanics traditional underwriting procedures, ‘‘perfect credit’’ to buy; and less than four in with obtaining financing. On the rental side, applicants who have conscientiously paid 10 are aware that lenders are not required by Hispanics were more likely in 2000 than in rent and utility bills on time but have never law to give them the lowest interest rate 1989 to be quoted a higher rent than their used consumer credit would be penalized for possible.27 A study using focus groups with white counterpart for the same unit. having no credit record. Applicants who renters found that even among those whose Another HUD-sponsored study asked have remained steadily employed, but have financial status would make them capable of respondents to a nationwide survey if they changed jobs frequently, would also be homeownership, many felt that the buying ‘‘thought’’ they had ever been discriminated penalized. As discussed in Section C below, process was insurmountable because they against when trying to buy or rent a house lenders, private mortgage insurers, and the feared rejection by the lender or being taken or an apartment.30 While the responses were GSEs have been adjusting their underwriting advantage of.28 subjective, they are consistent with the guidelines to take into account these special (v) Discrimination in the Housing and findings of the HDS. African Americans and circumstances of lower-income families. Mortgage Markets. Finally, differential Hispanics were considerably more likely Many of the changes recently undertaken by treatment of minorities in the sales and rental than whites to say they have suffered the industry focused on finding alternative markets and in the mortgage lending market discrimination—24 percent of African underwriting guidelines to establish has been well documented. The continued Americans and 22 percent of Hispanics creditworthiness that do not disadvantage discrimination in these markets is discussed perceived discrimination, compared to only creditworthy minority or low-income in the next section. 13 percent of whites. applicants. However, because of the Mortgage Lending Market. Research based 2. Disparities in Housing and Mortgage enhanced roles of credit scoring and on Home Mortgage Disclosure Act (HMDA) Markets automated underwriting in the mortgage data suggests pervasive and widespread origination process, it is unclear to what Sales and Rental Markets. In 2002, HUD disparities in mortgage lending across the released its third Housing Discrimination Nation. For 2001, the mortgage denial rate for 31 Margery Austin Turner, John Yinger, Stephen white mortgage applicants was 23 percent, Ross, Kenneth Temkin, Diane Levy, David Levine, Undoing of Community Development, Woodstock while 36 percent of African-American and 35 Robin Ross Smith, and Michelle deLair, What We Institute, Chicago, IL, November 1999. For a percent of Hispanic applicants were denied. Know About Mortgage Lending Discrimination, The national analyses, see the HUD report Unequal Two recent HUD-sponsored studies of Urban Institute, contract report for the Department Burden: Income and Racial Disparities in Subprime paired-testing at the mortgage pre-application of Housing and Urban Development, December Lending in America, April 2000; and Randall M. stage also points to discrimination by 1998. Scheessele, Black and White Disparities in 32 mortgage lenders. Based on its review of pair Margery Austin Turner, All Other Things Being Subprime Mortgage Refinance Lending, Housing Equal: A Paired Testing Study of Mortgage Lending Finance Working Paper No. HF–114, Office of tests conducted by the National Fair Housing Institutions, The Urban Institute Press, April 2002. Policy Development and Research, U.S. Department Alliance, the Urban Institute concluded that 33 Alicia H. Munnell, Geoffrey M.B. Tootell, Lynn of Housing and Urban Development, April 2002. E. Browne, and James McEneaney, ‘‘Mortgage 26 Fannie Mae, Fannie Mae National Housing 29 Margery Austin Turner, Stephen L. Ross, Lending in Boston: Interpreting HMDA Data,’’ Survey, 2002, p. 9. George Galster, and John Yinger, ‘‘Discrimination in American Economic Review, 86, March 1996. 27 See ‘‘Immigration Changes Won’t Hurt Metropolitan Housing Markets,’’ The Urban 34 See Charles W. Calomeris, Charles M. Kahn and Housing,’’ in National Mortgage News, January 27, Institute Press, November 2002. Stanley D. Longhofer, ‘‘Housing Finance 2003, p. 8. 30 Martin D. Abravanel and Mary K. Cunningham, Intervention and Private Incentives: Helping 28 Donald S. Bradley and Peter Zorn, ‘‘Fear of How Much Do We Know? Public Awareness of the Minorities and the Poor,’’ Journal of Money, Credit Homebuying: Why Financially Able Households Nation’s Fair Housing Laws. A report prepared for and Banking, 26, August 1994, pp. 634–74, for more May Avoid Ownership,’’ Secondary Mortgage HUD by the Urban Institute, Washington, DC, April discussion of this phenomenon, which is called Markets, 1996. 2002. ‘‘statistical discrimination.’’
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00070 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63649
degree the reduced rigidity in industry 1. Cost or rent burdens where housing Joint Center for Housing Studies also standards will benefit borrowers who have costs or rent exceed 50 percent of income (a attributes this to the growing gap between been adversely impacted by the traditional ‘‘severe burden’’) or range from 31 percent to housing costs and the incomes of the nation’s guidelines as discussed in section C.7, some 50 percent of income (a ‘‘moderate burden’’); poorest households.40 As a result of the industry observers have expressed a concern 2. The presence of physical problems increased incidence of severe and moderate that the greater flexibility in the industry’s involving plumbing, heating, maintenance, cost burdens, the share of owners reporting written underwriting guidelines may not be hallway, or the electrical system, which may no problems fell from 84 percent in 1978 to reflected in the numerical credit and lead to a classification of a residence as 78 percent in 1989 and 75 percent in 1999. mortgage scores which play a major role in ‘‘severely inadequate’’ or ‘‘moderately Between 1999 and 2001, the number of low the automated underwriting systems that the inadequate;’’ and, income owners with severe cost burdens GSEs and others have developed. 3. Crowded housing, where there is more (meaning those with incomes below 120 Disparities Between Neighborhoods. than one person per room in a residence. percent of AMI and spending more than half Mortgage credit also appears to be less The study reveals that in 2001, 5.1 million of their reported income on housing) shot up accessible in low-income and high-minority very low income renter households had by one million. This increase proved to be neighborhoods. As discussed in Appendix B, ‘‘worst case’’ housing needs, defined as the main cause of a highly significant nine 2001 HMDA data show that mortgage denial housing costs greater than 50 percent of percent jump in the overall number of low rates are nearly twice as high in census tracts household income or severely inadequate and moderate income owners and renters with low-income and/or high-minority housing among unassisted very-low-income with critical housing needs. Part of this could composition, as in other tracts (16.8 percent renter households.37 Among the 5.1 million be due to the heavy home equity borrowing versus 8.7 percent). Numerous studies have worst case needs renters, 4.8 million (94 that has characterized the housing market found that mortgage denial rates are higher percent) had a severe rent burden and 10 from the late 1990s to the present day, as in low-income census tracts, even accounting percent of renters lived in housing that was well as the fact that increases in house prices for other loan and borrower characteristics.35 severely inadequate. have outpaced increases in household These geographical disparities can be the a. Problems Faced by Owners income. As a corollary, subprime lending, result of cost factors, such as the difficulty of especially in minority communities, rose by appraising houses in these areas because of Of the 68.8 million owner households in 1999, 5.8 million (8 percent) confronted a about ten percentage points from the early the paucity of previous sales of comparable 1990s to 2001.41 homes. Sales of comparable homes may also severe cost burden and another 8.7 million be difficult to find due to the diversity of (12.7 percent) faced a moderate cost burden. b. Problems Faced by Renters central city neighborhoods. The small loans There were 870,000 households with severe Problems of all three types listed above are prevalent in low-income areas are less physical problems, 2 million with moderate more common among renters than among profitable to lenders because up-front fees to physical problems and 905,000 that were homeowners. In 1999 there were 6.3 million loan originators are frequently based on a overcrowded. The report found that 25 renter households (19 percent of all renters) percentage of the loan amount, although the percent of American homeowners faced at who paid more than 50 percent of their costs incurred are relatively fixed. As noted least one severe or moderate problem. income for rent.42 Another 7.1 million faced above, racial disparities in mortgage access Not surprisingly, problems were most a moderate rent burden. Thus in total 40 38 may be due to the fact that mainstream common among very low-income owners. percent of renters paid more than 30 percent lenders are not doing business in certain Almost a third of these households (31 of their income for rent. inner city neighborhoods. There is evidence percent) faced a severe cost burden, and an Among very-low-income renters, 71 additional 22 percent faced a moderate cost that mainstream lenders active in white and percent faced an affordability problem, burden. And 8 percent of these families lived upper-income neighborhoods are much less including 40 percent who paid more than in severely or moderately inadequate active in low-income and minority half of their income in rent. Almost one-third housing, while 2 percent faced overcrowding. neighborhoods—often leaving these (31 percent) of renters with incomes between Only 42 percent of very-low-income owners neighborhoods to unregulated subprime 51 percent and 80 percent of area median lenders. Geographical disparities in mortgage reported no problems. Over time the percentage of owners faced family income also paid more than 30 lending are discussed further in Section C.8 percent of their income for rent. below (which examines subprime lending) with severe or moderate physical problems has decreased, as has the portion living in Affordability problems have increased over and in Appendix B (which examines the time among renters. The shares of renters Geographically Targeted Goal). overcrowded conditions. However, affordability problems have become more with severe or moderate rent burdens rose 3. Affordability Problems and Worst Case common—the shares facing severe from 32 percent in 1978 to 36 percent in 1989 Housing Needs (moderate) cost burdens were only 3 percent and 40 percent in 1999. The severe affordability problems faced by (5 percent) in 1978, but rose to 5 percent (11 The share of households living in low-income homeowners and renters are percent) in 1989 and 8 percent (13 percent) inadequate housing in 1999 was higher for documented in HUD’s ‘‘Worst Case Housing in 1999. The increase in affordability renters (11 percent) than for owners (4 Needs’’ reports. These reports, which are problems apparently reflects a rise in percent), as was the share living in prepared biennially for Congress, are based mortgage debt in the late 1980s and early overcrowded housing (5 percent for renters, on the American Housing Survey (AHS), 1990s, from 21 percent of homeowners’ but only 1 percent for owners). Crowding and conducted every two years by the Census equity in 1983 to 36 percent in 1995.39 The inadequate housing were more common Bureau for HUD. The latest detailed report among lower-income renters, but among even analyzes data from the 1999 AHS. Although 37 This does not constitute a significant difference the lowest income group, affordability was it focuses on the housing problems faced by from the 1999 figure of 4.9 million households. the dominant problem. The prevalence of very-low-income renters, it also presents However, when the focus is narrowed to renters inadequate and crowded rental housing basic data on families and households in with incomes below 50 percent of AMI, a diminished over time until 1995, while owner-occupied housing.36 statistically significant change emerges; there were affordability problems grew. The ‘‘Worst Case’’ report measures three 4 percent fewer units affordable to this group in Other problems faced by renters discussed 2001 than there were in 1999. in the most recent detailed ‘‘Worst Case’’ types of problems faced by homeowners and 38 Very-low-income households are defined as report include a sharp decline (of 2.3 million, renters: those whose income, adjusted for household size, does not exceed 50 percent of HUD-adjusted area 35 Robert B. Avery, Patricia E. Beeson and Mark median income. This differs from the definition 40 Joint Center for Housing Studies of Harvard E. Sniderman, Understanding Mortgage Markets: adopted by Congress in the GSE Act of 1992, which University, State of the Nation’s Housing: 2000, p. Evidence from HMDA, Working Paper Series 94–21, uses a cutoff of 60 percent and which does not 24. Federal Reserve Bank of Cleveland, December 1994. adjust income for family size for owner-occupied 41 Joint Center for Housing Studies of Harvard 36 HUD has published an update on ‘‘worst case dwelling units. University, State of the Nation’s Housing 2004. p. housing needs,’’ which found that the number of 39 Edward N. Wolff, ‘‘Recent Trends in the Size 1–2, 4. such households rose from 4.86 million in 1999 to Distribution of Household Wealth,’’ The Journal of 42 Rent is measured in this report as gross rent, 5.07 million in 2001. However, detailed tables for Economic Perspectives, 12(3), (Summer 1998), p. defined as contract rent plus the cost of any utilities 2001 have not been published. 137. that are not included in contract rent.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00071 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63650 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
or 14 percent) between 1991 and 1999 in the associated with producing affordable in 2002 by almost 9 percent, and setting an number of rental units affordable to very-low- multifamily housing—just to name a few. all-time high in the 35-year history of the income families, and a worsening of the series. Combined new and existing home C. Factor 2: Economic, Housing, and national shortage of units affordable and sales set a national record of 6.2 million in Demographic Conditions: Single-Family available to extremely-low-income families 2002 and a record of almost 7.2 million in Mortgage Market (those with incomes below 30 percent of area 2003. median income). In 2001, the shortage for This section discusses economic, housing, One of the strongest sectors of the housing extremely-low-income families was and demographic conditions that affect the market in past years had been manufactured approximately 5 million units, not single-family mortgage market. After a review homes, but that sector has declined recently. statistically different from the 1999 number. of housing trends and underlying Between 1991 and 1996, manufactured home However, between 1999 and 2001, the demographic conditions that influence shipments more than doubled, peaking in number of units available to renters with homeownership, the discussion focuses on 1998 at 373,000. However, shipments fell incomes below 50 percent of AMI dropped specific issues related to the single-family more than 20 percent in both 2000 and 2001. from 78 units to 76 units per 100 renters, in owner mortgage market. This subsection In 2002, the industry shipped 169,000 new part because more of the units affordable to includes descriptions of recent market manufactured homes, down 12.4 percent this group of renters were occupied by interest rate trends, refinance and home from 2001. This was the lowest number of higher-income renters. Shortages of units purchase activity, homebuyer characteristics, manufactured home shipments since 1963. In affordable and available to extremely-low- and the state of affordable lending. Other 2003, the number of new manufactured income households were most pressing in the special topics examined include the growth homes shipped plummeted to 131,000, down West and Northeast, especially in in subprime lending, the increased use of 22.5 percent from 2002. Repossession has metropolitan areas in those regions. In 2001, automated underwriting, and the remaining been cited as a cause for the sales drop-off, the West was the only region to experience homeownership potential among existing as has the popularity of conventional stick- a significant decline in number of units renters. Section D follows with a discussion built housing. of the economic, housing, and demographic affordable to renters with incomes below 50 Homeownership Rate. In 1980, 65.6 conditions affecting the mortgage market for percent of AMI. This decline occurred even percent of Americans owned their own multifamily rental properties. in the wake of an increase in affordable units home, but due to the unsettled economic in the West during the 1990s. 1. Recent Trends in the Housing Market conditions of the 1980s, this share fell to 63.8 4. Rehabilitation and Other National Housing While most other sectors of the economy percent by 1989. But since 1994, gains in the Needs were weak or declining during 2001 and homeownership rate have occurred in each 2002, the housing sector showed remarkable year, with the rate reaching another record In addition to the broad housing needs strength. Again in 2003, the housing market mark of 68.3 percent in 2003. discussed above, there are additional needs enjoyed an outstanding year. The numbers of Gains in homeownership have been confronting specific sectors of the housing 44 single-family permits, starts, completions, widespread over the last eight years. As a and mortgage markets. One example of these new home sales, and existing home sales result, the homeownership rate rose from: specific needs concerns the rehabilitation of were record-breaking. Home ownership was • 42.0 percent in 1993 to 48.8 percent in the nation’s older housing stock. A major also at an all-time high, and mortgage interest 2003 for African American households, problem facing lower-income households is rates continued to stay under six percent on • 39.4 percent in 1993 to 46.7 percent in that low-cost housing units continue to average. In addition, the prosperity of the 2003 for Hispanic households, disappear from the existing housing stock. market stimulated GDP, contributing 0.37 • 73.7 percent in 1993 to 79.1 percent in Older properties are in need of upgrading percent to its overall growth rate of 3.1 2003 for married couples with children, and rehabilitation. These aging properties are percent. Although the multifamily sector • 65.1 percent in 1993 to 68.4 percent in concentrated in central cities and older inner experienced high vacancies and low lease-up 2003 for household heads aged 35–44, and suburbs, and they include not only detached rates, the vitality of the single family market • 48.9 percent in 1993 to 52.3 percent in single-family homes, but also small was strong enough to result in a spectacular 2003 for central city residents. multifamily properties that have begun to peak in total permits and starts as well as However, as these figures demonstrate, deteriorate. But obtaining the funds to fix up builders’ attitudes and housing sizable gaps in homeownership remain. older properties can be difficult. The owners affordability.43 Economy/Housing Market Prospects. Job of small rental properties in need of Single-Family Permits, Starts, and growth has been less robust in the recent rehabilitation may be unsophisticated in Completions. Builders took out 1,440,400 recovery than some previous recoveries. obtaining financing. The properties are often single-family permits in 2003, up 6 percent However, the economy grew at a rate of 2.2 occupied, and this can complicate the from 2002. The 2003 level was the highest percent in 2002 and even faster in 2003.45 rehabilitation process. Lenders may be number of single-family permits ever Although the Federal Reserve has recently reluctant to extend credit because of a reported in the 44-year history of this series. begun raising short term interest rates, sometimes-inaccurate perception of high Single-family starts totaled 1,498,500 housing mortgage interest rates remain low, credit risk involved in such loans. The GSEs units, up 10 percent from 2002, a new single- supporting housing affordability. and other market participants have recently family record. Construction was completed Fannie Mae expects existing home sales to begun to pay more attention to these needs on 1,386,200 single-family housing units, up reach 5.7 million in 2004 and 2005.46 for financing of affordable rental housing 5 percent from 2002. Projected at 1.84 million in 2003, the rehabilitation. However, extra effort is Sales of New and Existing Homes. After National Association of Home Builders required, due to the complexities of leveling out in 2000, housing sales have expects housing starts to decline to 1.77 rehabilitation financing, as there is still a boomed in the past three years, reaching million in 2004 and 1.71 million in 2005.47 need to do more. record highs in 2001, 2002, and again in The Mortgage Bankers Association forecasts The rehabilitation of our aging housing 2003. New single family home sales, which that 2004 housing starts will total 1.73 stock is but one example of the housing and increased an average 6.3 percent per year million units and the 30-year fixed mortgage mortgage issues that need to be addressed. between 1992 and 2002, reached a record Several other examples will be provided high of 1,085,000 units in 2003, an increase throughout the following sections on the of 12 percent over 2002 sales. The market for 44 Homeownership rates prior to 1993 are not economic, housing, and demographic new homes has been strong in the Mid strictly comparable with those beginning in 1993 conditions in the single-family and because of a change in weights from the 1980 Atlantic, Midwest and Great Plains. Census to the 1990 Census. multifamily markets, as well as in The National Association of Realtors 45 National Association of Realtors, ‘‘Near Record Appendices B–D. The discussion will cover reported that 6.1 million existing homes were Homes Sales Projected for 2003,’’ December 3, 2002. a wide range of topics, such as subprime sold in 2003, overturning the old record set 46 Fannie Mae, ‘‘Berson’s Economic and Mortgage lending, predatory lending, automated Market Development Outlook,’’ December 2003. underwriting systems, manufactured 43 US Housing Market Conditions, 4th Quarter, http://www.fanniemae.com/media/pdf/berson/ housing, the special needs of the single- 2003. HUD Office of Policy Development and monthly/2003/121203.pdf. family rental market, and challenges Research. 47 http://www.nahb.org.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00072 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63651
rate will average 6.1 percent.48 After more peak in the next two decades. However, reach peak home buying age later in the first than doubling from a relative trough in 2000 because homeownership rates do not peak decade of the millennium. to an estimated $2.6 trillion in 2002, Fannie until population groups reach 65 to 74 years Immigrant Homebuyers. Past, present, and Mae projected in December 2003 that of age, this age cohort will continue to future immigration will also contribute to mortgage originations will drop to $1.8 provide housing demand. According to gains in the homeownership rate. During the trillion in 2004 and $1.5 trillion in 2005.49 Riche, the increasing presence of older 1990s, 9.8 million legal immigrants entered 2. Underlying Demographic Conditions households should increase the proportion of the United States, as compared to 6.3 million the population that owns, rather than rents entering in the 1980s and 4.2 million during Between 2000 and 2025, the U.S. housing.55 the 1970s. Overall, the increase in the population is expected to grow by an average immigrant population directly accounted for 50 Growing housing demand from immigrants of 2.5 million per year. This will likely and non-traditional homebuyers will help to 35 percent of the nation’s rise in population result in at least 1.1 million new households 59 offset declines in the demand for housing in the 1990s. As a result, the foreign-born per year.51 Recently revised increases in caused by the aging of the population. population of the United States more than population projections by the Census Bureau Riche’s study estimates that minorities will tripled from 9.6 million in 1970 to 31.1 push population figures higher with the Joint account for two-thirds of the growth in U.S. million in 2000. Immigrants who become Center estimating new household growth at households over the next 25 years,56 and by citizens buy homes at rates nearly as high as 13.3 million from 2005 to 2015.52 This 2025, non-family households will make up a their same-aged native-born counterparts and section discusses important demographic third of all households. The ‘‘echo baby- for those aged 25 to 34, the gap is virtually trends behind these overall household 60 boom’’ (that is, children of the baby-boomers) nonexistent. Moreover, U.S.-born children numbers that will likely affect housing will also add to housing demand in the of immigrants often have higher demand in the future. These demographic current and next decades. Finally, the homeownership rates than the same-age forces include the baby-boom, baby-bust and 61 growing income inequality between people children of native-born parents. However, echo baby-boom cycles; immigration trends; with and without a post-secondary education there are concerns about the assimilation into non-traditional and single households; will continue to affect the housing market. homeownership of recent Hispanic ‘‘trade-up buyers;’’ and the growing income The Baby-Boom Effect. The demand for immigrants who are less educated than inequality between people with different housing during the 1980s and 1990s was earlier cohorts of immigrants. Many levels of education. HUD’s Office of Policy driven, in large part, by the coming of home immigrants also locate in high-priced Development and Research funded a study, buying age of the baby-boom generation, housing markets, which makes it more Issue Papers on Demographic Trends those born between 1945 and 1964. difficult for them to achieve homeownership. Important to Housing, which analyzes effects Homeownership rates for the oldest of the Although net foreign immigration is of demographic conditions on the housing baby-boom generation, those born in the projected to decline in the current decade market. The findings are presented after 2002, high levels of immigration in the 1940s, rival those of the generation born in throughout the sections that follow.53 late 1980s and throughout the 1990s will the 1930s. Due to significant house price As explained below, the role of traditional have lasting positive effects on housing appreciation in the late-1970s and 1980s, first-time homebuyers, 25-to-34-year-old demand. New immigration in the current and older baby-boomers have seen significant married couples, in the housing market will next decades is projected to create 6.9 gains in their home equity and subsequently be smaller in the current decade due to the million net new households, but the majority have been able to afford larger, more aging of the population. For the first time in of household growth in the period (16.9 expensive homes. Circumstances were not so history, the population will have roughly million) will come from people already favorable for the middle baby-boomers. equal numbers of people in every age group. resident in the U.S. including the foreign- Housing was not very affordable during the Between 2000 and 2025, the Census Bureau born population.62 While immigrants tend to 1980s, their peak home buying age period. As projects that the largest growth in households rent their first homes upon arriving in the a result, the homeownership rate, as well as will occur among householders 65 and United States, homeownership rates are wealth accumulation, for the group of people over.54 Thus, an increasing percentage of the substantial for those that have lived here for born in the 1950s lags that of the generations population will be past their home buying 57 at least 6 years. In 1996, the homeownership before them. rate for recent immigrants was 14.7 percent As the youngest of the baby-boomers (those 48 Mortgage Bankers Association of America, while it was 66.9 percent for foreign-born born in the 1960s) reached their peak home naturalized citizens after six years.63 Higher- Mortgage Finance Forecast, December 17, 2003. buying years in the 1990s, housing became http://www.mbaa.org/marketdata/forecasts/ than-average foreign-born fertility rates and mffore1103.pdf. more affordable. While this cohort has high rates of homeownership for immigrants 49 Fannie Mae, ‘‘Berson’s Economic and Mortgage achieved a homeownership rate equal to the living in the country for several years and Market Development Outlook,’’ December 2003. middle baby-boomers, they live in larger, among the children of immigrants suggest 50 U.S. Census Bureau, Population Projections more expensive homes. As the baby-boom that past immigration will continue to create Table NP–T1. generation ages, demand for housing from housing demand. 58 51 Martha Farnsworth Riche, ‘‘How Changes in this group is expected to wind down. Past and future immigration will lead to the Nation’s Age and Household Structure Will The baby-boom generation was followed by increasing racial and ethnic diversity, Reshape Housing Demand in the 21st Century,’’ in the baby-bust generation, from 1965 through especially among the young adult Issue Papers on Demographic Trends Important to 1977. Since this population cohort is smaller Housing, Urban Institute Final Report to the Office than that of the baby boom generation, it 59 Federation for American Immigration Reform, of Policy Development and Research, U.S. reduced housing demand in the preceding Department of Housing and Urban Development,
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00073 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63652 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
population. As immigrant minorities account distribution was virtually unchanged from 2001.70 Originations then jumped to $3.8 for a growing share of first-time homebuyers 1977 to 2001.67 trillion in 2003, with refinances accounting in many markets, HUD and others will have The increase in income inequality over for 66 percent (or $2.5 trillion) of this total. to intensify their focus on removing past decades has been especially significant This boom in lending over the past three discrimination from the housing and between those with and those without post- years can be attributed to low mortgage mortgage finance systems. The need to meet secondary education. The Census Bureau interest rates and a record number of nontraditional credit needs, respond to reports that by 1999, the annual earnings of refinances. Approximately 40 percent of diverse housing preferences, and overcome workers with a bachelor’s degree were 1.8 mortgage debt outstanding, or $2.5 trillion, the information barriers that many times the annual earnings of workers with a was refinanced during the 2001–02 refinance immigrants face will take on added high school education.68 The inflation- boom. Freddie Mac calculates total home importance. In order to address these needs, adjusted median earnings of high school equity cashed out in 2002 at 105.4 billion the mortgage industry must offer innovative graduates were at the same level in 2001 as and estimates that number will increase to products and improve outreach efforts to in 1991 while the earnings of bachelor 138.8 billion in 2003.71 This section focuses attract minority homebuyers. degree-holders rose nearly 9 percent over the on recent interest rate trends, the refinance 69 Nontraditional and Single Homebuyers. same period. market, the home purchase market, and first- While overall growth in new households has So, while homeownership is highly time homebuyers. The section concludes by slowed down, nontraditional households affordable, those without post-secondary examining the GSEs’ acquisitions as a share have become more important in the education often lack the financial resources of the primary single-family mortgage market, homebuyer market. As the population ages to take advantage of the opportunity. As and provides mortgage market prospects. both relatively and absolutely, the nation’s discussed earlier, the days of the well-paying a. Mortgage Characteristics unionized factory job have passed. They have households will become smaller and more given way to technological change that favors Interest Rate Trends and Volatility. diverse. Riche notes that in 2000, traditional white-collar jobs requiring college degrees, Historically low mortgage interest rates in the family households represented fewer than and wages in the manufacturing jobs that late 1990s and 2001–2003 helped maintain one in four households and were surpassed remain are experiencing downward pressures consumer confidence in the housing sector as by both single-person households and from economic globalization. The effect of the economy emerged from its first recession married couples without children. With later this is that workers without the benefit of a in almost a decade. After high and marriages and more divorces, single-parent post-secondary education find their demand fluctuating mortgage rates in the 1980s and and single-person households have increased for housing constrained. This is especially early 1990s, recent years have seen a period rapidly. In fact, single-parent households problematic for recent immigrants who are of lower and more stable rates. The 1980s grew from 4 percent of family households in more likely to have limited educational began with interest rates on mortgages for 1950 to 12 percent in 2000. Single-person attainment and English language proficiency. new homes above 12 percent but quickly rose households are now the nation’s second most Summary. Over the next two-and-a-half to more than 15 percent.72 By 1987–88, rates numerous household type, accounting for decades, the number of U.S. households is dipped into single digits but were rising over 25 percent of all households. In the projected to increase by nearly 27 million. Of again by 1989–90. Rates declined in the early future, longer life expectancies and the these new households, non-Hispanic white 1990s, reaching a low of 6.8 percent in late continuing preference for one or two children and traditional households will contribute 1993. An upturn in rates in 1994 and 1995 will make households without children even only one-third and one-tenth of the growth, peaked at 8.3 percent in early 1995. By 1998, more numerous. Projected to compose 80 respectively. As the baby-boomers aged out 30-year fixed conventional mortgages percent of all households by 2025, of their peak home buying stage and the averaged 6.95 percent, the lowest level since nontraditional family households will play baby-bust generation aged into their peak 1968 but saw a rise in 1999 to 7.44 percent. an increasingly important role in the home home buying stage in the late 1980s, demand Mortgage rates then continued to rise in buying market.64 for housing was dampened by demographic 2000, averaging 8.05 percent for the year, Trade-up Buyers. Due to weak house price factors during the 1990s. (Of course, other before falling to a low of 6.62 percent in appreciation, traditional ‘‘trade-up buyers’’ factors such as low interest rates propelled October 2001 and averaging 6.97 percent for stayed out of the market during the early the housing market to record levels during 2001 as a whole.73 Rates averaged 5.83 1990s. Their absence may explain, in part, this period.) As the echo baby-boomers begin percent during 2003 74, reaching a low of 5.23 the large representation of nontraditional to enter their peak home buying age, housing in June. Rates in 2004 have averaged 5.83 homebuyers during that period. However, demand should pick up again through the through August, reaching a low of 5.45 in since 1995 home prices have increased more remainder of the current decade and into the March. 75 than 30 percent.65 The greater equity next. As these demographic factors play out, Other Loan Terms. When mortgage rates resulting from recent increases in home the overall effect on housing demand will are low, most homebuyers prefer to lock in prices should lead to a larger role for ‘‘trade- likely be sustained growth and an a fixed-rate mortgage (FRM). Adjustable-rate up buyers’’ in the housing market during the increasingly diverse household population mortgages (ARMs) are more attractive when next 10 to 15 years. In addition, the growing from which to draw new homeowners. There number of higher-income, mid-life are continuing concerns about the increasing 70 ‘‘Mortgage Originations Hit Record-Busting households will increase households’ income inequality of our population and $2.5 Trillion in 2002, IMF Numbers Reveal,’’ Inside potential to ‘‘trade up’’ to more expensive those recent immigrants and other persons Mortgage Finance, January 24, 2003, p. 3. housing.66 who have limited education. 71 Freddie Mac ‘‘Cash-Out Refi Report.’’ Growing Income Inequality. The Census 72 Interest rates in this section are effective rates 3. Basic Trends in the Single-Family paid on conventional home purchase mortgages on Bureau reported that the top 5 percent of Mortgage Market American households received 22.4 percent new homes, based on the Monthly Interest Rate of aggregate household income in 2001, up Mortgage lending in the nation is growing Survey (MIRS) conducted by the Federal Housing Finance Board and published by the Council of from 21.4 percent in 1998 and up sharply at unprecedented levels. Residential mortgage originations soared to $2.5 trillion Economic Advisers annually in the Economic from 16.1 percent in 1977. The share Report of the President and monthly in Economic accruing to the lowest 80 percent of in 2002, a 22 percent increase over the Indicators. These are average rates for all loan types, households fell from 56.5 percent in 1977 to previous record of $2.06 trillion set in encompassing 30-year and 15-year fixed-rate 50.8 percent in 1998 and again to 49.8 mortgages and adjustable rate mortgages. percent in 2001. The share of aggregate 67 All data in this paragraph are from the U.S. 73 U.S. Housing Market Conditions, 2nd Quarter income accruing to households between the Census Bureau’s Historical Income Table H2. 2002, August 2002, Table 14. 80th and 95th percentiles of the income 68 Jennifer Cheeseman Day and Eric C. 74 U.S. Housing Market Conditions, 4th Quarter Newburger, The Big Payoff: Educational Attainment 2003, February 2004, pg.1. and Synthetic Estimates of Work-Life Earnings, U.S. 75 Mortgage Bankers Association website. MBA 64 Riche, 2002, p. 1. Bureau of the Census, Current Population Reports Weekly Survey of Mortgage Applications, Monthly 65 Average new-home price: U.S. Census Bureau, P23–210, July 2002, p. 3. Average Interest Rates On 30-Year Fixed-Rate
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00074 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63653
rates are high, because they carry lower rates b. Refinance Mortgages Historically low interest rates and than FRMs and because buyers may hope to Over the past ten years, refinance booms declining mortgage transaction costs have refinance to an FRM when mortgage rates occurred three times, during 1992–93, 1998, driven the latest refinancing boom. Given decline. The Federal Housing Finance Board and 2001–03. Refinancing has fueled the these conditions, the after-tax cost saving on (FHFB) reports that the ARM share of the growth in total mortgage originations, which a new, lower-rate loan is much greater than market fell from 20 percent in 1993 to a were $638 billion in 1995 (a period of low the transaction costs of refinancing. In record low of 12 percent in 1998, before refinance activity), but topped $2.5 trillion in addition, the appreciation of housing prices rising back to 21 percent in 1999. The ARM 2002 (a period of heavy refinance activity). has also contributed to the increase in share continued to rise to 24 percent in 2000, The refinance share of total mortgage refinancing. Over the past five years, the but then fell dramatically to a low of 12 originations rose to 50 percent in 1998, then value of housing rose by approximately $5 percent in 2001 as mortgage rates decreased. decreased to 19 percent in 2000 before However, in 2002 and 2003, there was a trillion, and the rise in value has enabled jumping to 57 percent in 2001, and 59 lenders to service refinancing homeowners rebound in the ARM share of the market. percent in 2002. During the 2001–02 because of greater confidence in the Though it still is nowhere near the size it was refinance boom, approximately 40 percent of creditworthiness of borrowers.84 in the mid to late 1990s, the past two years the $2.5 trillion in mortgage debt outstanding have seen the share climb to 17 and 19 was refinanced. In 2003, the refinance share Over the past few years, homeowners have percent, respectively.76 of total mortgage originations hit 66 percent, become more willing to draw on the rising In 2003, the term-to-maturity was 30 years though late 2003 saw a steep drop-off from equity in their homes. According to Fannie for 80 percent of conventional home a 68 percent share in the third quarter to a Mae’s 2002 National Housing Survey, purchase mortgages, continuing to decline 49 percent share in the fourth.81 homeowners that refinanced during 2001 after steadily climbing to a high of 90 percent In 1989–90 interest rates exceeded 10 withdrew about $110 billion in accumulated in 2000. The other major term of maturity in percent, and refinancings accounted for less home equity wealth.85 Freddie Mac estimates 2003 was 15 years (16 percent).77 than 25 percent of total mortgage that more than one-half of all refinance Low- and no-point mortgages continue to originations.82 The subsequent sharp decline mortgages in the past two years involved be a popular option for mortgage purchases. in mortgage rates drove the refinance share cash-out refinancing.86 FHFB reports that average initial fees and over 50 percent in 1992 and 1993 and The refinancing boom contributed to an charges (‘‘points’’) have decreased from 2.5 propelled total single-family originations to estimated one-fifth of the national economy’s percent of loan balance in the mid-1980s to more than $1 trillion in 1993—twice the level real GDP growth since late 2000.87 During 2 percent in the late-1980s, 1.5 percent in the attained just three years earlier. 2001 and 2002, roughly $270 billion was early 1990s, and less than 1 percent in 1995– The refinance wave subsided after 1993, 97. The downward trend continued because most homeowners who found it raised in cash-out refinancing. throughout the late 1990s with the average beneficial to refinance had already done so Approximately one-half of cash from cash- initial fees and charges reaching a low of one- and because mortgage rates rose once again.83 out refinancing has enabled consumers to half percent in 2001, staying there in 2002, Total single-family mortgage originations finance more spending for expenses such as and dipping even further down in 2003. bottomed out at $638 billion in 1995, when home improvements, medical payments, Coupled with declining interest rates, these the refinance share was only 21 percent. education, and vehicles during a weakened lower transactions costs have increased the Total originations, driven by the volume of economy. Roughly one-third of the cash from propensity of homeowners to refinance their refinancings, amounted to $1.507 trillion in cash-out refinancing has allowed consumers mortgages.78 1998, nearly 50 percent higher than the to repay other debt.88 The remaining cash Another major change in the conventional previous record level of $1.02 trillion from cash-out refinancing has enabled home mortgage market has been the attained in 1993. consumers to invest in other assets. proliferation and then diminution of high The refinance wave from late 1997 through Refinancing households save approximately loan-to-value ratio (LTV) mortgages. early 1999 reflected other factors besides $10 billion in their annual interest payments According to data from the Federal Housing interest rates, including greater borrower on their mortgage and consumer installment Finance Board, loans with LTVs greater than awareness of the benefits of refinancing, a liabilities. highly competitive mortgage market, and the 90 percent (that is, down payments of less The refinancing boom will have lingering enhanced ability of the mortgage industry, than 10 percent) made up less than 10 effects. Mortgage borrowers that were able to utilizing automated underwriting and percent of the market in 1989–91, but 25 secure low long-term interest rates through percent of the market in 1994–97, gradually mortgage origination systems to handle an unprecedented volume of originations. The fixed rate mortgages will have more of their decreasing to an average of 20 percent of the budgets to spend on other items. Meanwhile, market in 2003. Loans with LTVs less than refinance share decreased to 19 percent in 2000 before jumping to a record 57 percent cash-out borrowers, who are just receiving or equal to 80 percent fell from three-quarters in 2001. their money, will spend this year. It must be of the market in 1989–91 to an average of 56 noted there is some concern regarding the percent of the market in 1994–97, but then the primary market than the Finance Board’s potential for increased credit risk stemming rose to an average of 63 percent of mortgages from mortgage debt from cash out borrowers. originated in 1998–2001, and rose again to an survey. However, the Chicago Title survey does not separate FHA-insured loans from conventional According to a 2002 Regional Finance average of 70 percent of mortgages originated mortgages. In addition, the statistics cited above Review article, the mortgage liabilities of in 2002–2003.79 As a result, the average LTV pertain only to home purchase mortgages. households have been growing at a rate more rose from 75 percent in 1989–91 to nearly 80 Refinance mortgages generally have shorter terms percent in 1994–97, and then declined to and lower loan-to-value ratios than home purchase than double the growth in household 76.2 percent in 2001, 75.1 percent in 2002, mortgages. incomes. However, this potential credit risk and 73.5 percent in 2003.80 81 The source for the refinance share and total is moderated by the strong growth in housing mortgage originations is the Mortgage Bankers values. The ratio of mortgage debt to housing Association (http://www.mortgagebankers.org/ 76 _ http://www.fhfb.gov/mirs/mirs t25.xls. marketdata/forecasts/mffore1203.pdf, http:// 84 77 http://www.fhfb.gov/mirs/mirstbl5.xls; data for www.mortgagebankers.org/marketdata/forecasts/ Economy.com, ‘‘The Economic Contribution of 2003 is average of May through December data. ffJUNE2004.pdf). the Mortgage Refinancing Boom,’’ December 2002, 78 p. 4. This is discussed in more detail in Paul 82 Refinancing data is taken from Freddie Mac’s 85 Bennett, Richard Peach, and Stavros Peristani, monthly Primary Mortgage Market Survey. Fannie Mae, 2002 Fannie Mae National Housing Survey.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00075 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63654 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
values, the aggregate loan-to-value ratio, has As discussed earlier, barriers are 1996 and 2001. In addition, first-time remained fairly stable for a decade.89 preventing many potential homeowners from homebuyers comprised approximately 60 c. Home Purchase Mortgages becoming homeowners, thus reducing the percent of all minority home purchases possible amount of home purchase loans. during the 1990s, compared with about 35 The volume of home purchase mortgages While the strong housing sector has provided percent of all home purchases by non- was $505 billion in 1995, rose to $848 billion financial security for many Americans, a Hispanic white families. in 1999, and remained in the $829–$873 2002 Fannie Mae survey found that In comparison to repeat homebuyers, first- billion range between 1999–2001 before ‘‘information barriers still keep many time homebuyers are more likely to be jumping to $1.02 trillion in 2002 and $1.30 financially qualified families-particularly younger, have lower incomes, and purchase trillion in 2003. The Mortgage Bankers minority Americans from becoming less expensive houses. According to the AHS, Association (MBA) forecasts that the home homeowners or obtaining the lowest-cost more than one-half or first-time homebuyers purchase volume will be $1.52 trillion in financing available to them.’’ 93 were below the age of 35, compared with less 2004 as the home purchase share rises to 57 These homeownership barriers pose a than one-quarter of repeat buyers in the percent of all originations.90 The home serious problem for many Americans who 1990s. Thirty-nine percent of first-time purchase share of total mortgage originations view homeownership as a smart, safe, long- buyers had incomes below 80 percent of the was 79 percent in 1995, declined to 50 term investment, rating homeownership as a median compared to 30 percent of repeat percent in 1998, rose to 81 in 2000, and better investment than the stock market. buyers. Fifty-four percent of first-time buyers sharply fell to 43 percent in 2001, 41 in 2002, Home equity is the single most important purchased homes priced below $100,000, and 34 percent in 2003, as refinance asset for approximately two-thirds of compared to 37 percent of repeat buyers. mortgage volume grew. This section American households that are homeowners. Minorities comprise a higher proportion of discusses the important issue of housing affordability and then examines the value of Considering that half of all homeowners held first-time buyers (32 percent) compared to homeownership as an investment. at least 50 percent of their net wealth in repeat buyers (14 percent). Compared to The National Association of Realtors (NAR) home equity in 1998, increasing housing repeat buyers, first-time homebuyers are has developed a housing affordability index, affordability is important for many more likely to purchase a home in the central Americans.94 city and more likely to be a female-headed calculated as the ratio of median household 98 income to the income needed to qualify for First-time Homebuyers. First-time household. a median price home (the latter income is homebuyers are a driving force in the The National Association of Realtors called the ‘‘qualifying income’’). In 1993, nation’s mortgage market. The recent low reports that the average first-time homebuyer NAR’s affordability index was 133, which interest rates have made it an opportune time in the first quarter of 2003 was 32 years old meant that the median family income of for first-time homebuyers, which are with a household income of $54,800, $37,000 was 33 percent higher than that typically people in the 25–34 year-old age compared to an average age of 46 years and income needed to qualify for the median group that purchase modestly priced houses. average household income of $74,600 for priced home. Housing affordability remained As the post-World War II baby boom repeat buyers. The average first-time at about 130 for 1994–97, with home price generation ages, the percentage of Americans homebuyers made a downpayment of 6 in this age group decreased from 28.3 percent percent on a home that cost $136,000 while increases and somewhat higher mortgage 95 rates being offset by gains in median family in 1980 to 25.4 percent in 1992. Even the average repeat buyer made a income.91 Falling interest rates and higher though this cohort is smaller, first-time downpayment of 23 percent on a home income led to an increase in affordability to homebuyers increased their share of home costing $189,000. In the NAR survey, 37 143 in 1998, reflecting the most affordable sales. According to Chicago Title data for percent of first-time homebuyers were single 99 housing in 25 years. Affordability remained major metropolitan areas, the first-time buyer compared to 28 percent of repeat buyers. high in 1999, despite the increase in share of the homebuyer market increased Many African Americans and Hispanics mortgage rates. NAR’s affordability index from roughly 40 percent in the beginning of are likely to purchase homes in the coming the 1990s to 45–47 percent during the-mid years, contributing to the number of first-time declined from 140 in 1999 to 129 in 2000 as 96 mortgage rates increased. The index turned and late 1990s. Since the late 1990s, home-buyers fueling growth in the housing upward to 136 in 2001 as mortgage rates fell industry survey data suggest that the first- sector. The number of homeowners will rise and maintained this average in 2002, before time homebuyer percentage has decreased by an average of 1.1 million annually over rising further to 140 in 2003.92 slightly. In the first quarter of 2003, the share the next two decades. The sizeable rise in the Although the share of home purchase loans of all home purchases by first-time foreign-born population since the 1970’s for lower-income households and/or homebuyers was 40 percent compared to 42 coupled with the increase in Latin American 97 households living in lower-income percent in 2001. and Asian immigration will also contribute communities increased over the past decade, In the 1990s, lenders developed special much to this growth.100 affordability still remains a challenge for programs targeted to first-time homebuyers d. GSEs’ Acquisitions as a Share of the many. The median sales price of existing and revised their underwriting standards to Primary Single-Family Mortgage Market single-family homes in the United States enhance homeownership opportunities for Purchases by the GSEs of single-family continues to rise, reaching $158,100 in 2002 low-income families with special mortgages amounted to $519 billion during and $170,000 in 2003. The production of circumstances. The disproportionate growth the heavy refinancing year of 1993, stood at affordable housing and low interest rates in the number of first-time homebuyers and $215 billion in 1995, and were at $618 billion could offset the negative impact of rising minority homebuyers largely drove the rising during the heavy refinancing year of 1998. house prices, which undermine housing trend in total home purchases. Analysis of Purchases then fell to $395 billion in 2000 affordability for many Americans, the American Housing Survey (AHS) before reaching record levels during the particularly in several high-cost markets on indicates there were 1.3 million new first- the east and west coasts. time homebuyers during 1991, in comparison heavy refinancing years of 2001 ($961 with over two million in each year between billion) and 2002 ($1,090 billion). Purchases by Fannie Mae decreased from $316 billion 89 Economy.com, ‘‘The Economic Contribution of in 1999 to $227 billion in 2000, before rising the Mortgage Refinancing Boom,’’ December 2002, 93 Fannie Mae, September 4, 2002, p.2. p. 9. 94 Ibid. to $568 billion in 2001, $800 billion in 2002, 90 Mortgage Bankers Association, ‘‘Mortgage 95 U.S. Department of Commerce, Bureau of the and $1.3 trillion in 2003. Freddie Mac’s Finance Forecast’’, September 17, 2004. http:// Census, Money Income of Households, Families, www.mortgagebankers.org/marketdata/forecasts/ and Persons in the United States: 1992, Special 98 U.S. Housing Market Conditions, 3rd Quarter mffore1203.pdf. Studies Series P–60, No. 184, Table B–25, October 2001, November 2001, Table 4. 91 Housing affordability varies markedly between 1993. 99 National Association of Realtors. ‘‘New NAR regions, ranging in January 2004 from 194 in the 96 Chicago Title and Trust Family of Insurers, Survey of Home Buyers and Sellers Shows Growing Midwest to 107 in the West, with the South and Who’s Buying Homes in America, 1998. Web Use in a Dynamic Housing Market.’’ http:// Northeast falling in between. 97 National Association of Realtors. ‘‘New NAR www.realtor.org. 92 National Association of REALTORS. Housing Survey of Home Buyers and Sellers Shows Growing 100 Joint Center for Housing Studies of Harvard Affordability Index,
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00076 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63655
single-family mortgage purchases followed a Fannie Mae and Freddie Mac have been a In 2000, Fannie Mae launched the similar trend, falling from $233 billion in part of this ‘‘revolution in affordable ‘‘MyCommunityMortgage’’ suite of products, 1999 to $168 billion in 2000, and then rising lending’’. During the mid-to-late 1990s, they which provides high loan-to-value product to $393 billion in 2001 and $475 billion in added flexibility to their purchase guidelines, options for low- and moderate-income 2002.101 they introduced new low-down-payment borrowers. In 2003, Fannie Mae purchased or The Office of Federal Housing Enterprise products, and they worked to expand the use securitized more than $2.27 billion of Oversight (OFHEO) estimates that the GSEs’ of credit scores and automated underwriting MyCommunityMortgage products, which share of total originations in the conventional in evaluating the creditworthiness of loan helped provide affordable housing solutions single-family mortgage market, measured in applicants. These major trends reflect for 20,400 households. In addition, Fannie dollars, declined from 37 percent in 1996 to changes in the GSEs’ underwriting that have Mae enhanced the MyCommunityMortgage to 32 percent in 1997—well below the peak of impacted affordable lending. Through these help lenders further expand affordable 51 percent attained in 1993. OFHEO trends, Fannie Mae and Freddie Mac have financing to underserved families. Examples attributes the 1997 downturn in the GSEs’ attempted to increase their capacity to serve of these enhancements included adding role to increased holdings of mortgages in low- and moderate-income homebuyers. MyCommunityMortgage to Desktop portfolio by depository institutions and to This section summarizes recent initiatives Underwriter in order to provide lenders increased competition with Fannie Mae and undertaken by the GSEs and others in the easier access to customized CRA-targeted Freddie Mac by private label issuers. industry to expand affordable housing. The loan products, adding new credit and income However, OFHEO estimates that the GSEs’ end of this section will present evidence that flexibilities for borrowers purchasing single share of the conventional market rebounded these new industry initiatives are working, as family homes, Community HomeChoice sharply in 1998–99, to 43–42 percent. The increased mortgage credit has been flowing to which offers more flexible requirements for GSEs’ share then decreased to approximately low-income and minority families. The persons with disabilities, Community 2–4 30 percent of the single-family conventional following section will continue the affordable FamilyTM to help make the purchase of 2– mortgages originated in 2000, and then lending theme by examining the performance 4 unit homes more affordable for first time increased sharply to 40 percent in 2001. of different market sectors (e.g., depositories, homebuyers, and Community RenovationTM Total GSE purchases, including loans 1–4 Family Pilot to help borrowers with originated in prior years, amounted to 46 GSEs, etc.) in funding loans for low-income and minority families. That section will also home improvement and housing preservation percent of conventional originations in costs.106 Additionally, in 2003, Fannie Mae 2001102 and approximately 38 percent of discuss the important role that FHA plays in making affordable housing available to enhanced Community 2–4 Family and family home mortgage originations in Community Renovation 1–4 Family pilots. 2002.103 historically underserved groups as well as the continuing concern that participants in This product provides lower down payments e. Mortgage Market Prospects the conventional market could be doing even and flexible parameters for owner-occupants 107 The Mortgage Bankers Association (MBA) more to help underserved families. of 1–4 unit properties. reports that mortgage originations in 2001 Fannie Mae also expanded its ‘‘Flexible’’ a. Lowering Down Payments and Up-Front product line with the ‘‘Flexible 100’’ product, were $2.0 trillion, which is almost twice the Costs volume of originations in 2000. Mortgage which eliminates the requirement for a down originations then increased to record levels of Numerous studies have concluded that payment by providing 100 percent loan-to- $2.5 trillion in 2002 and $3.8 trillion in 2003, saving enough cash for a down payment and value financing. The borrower is required to with refinancings representing 66 percent of for up-front closing costs is the greatest make either a minimum of 3% (of the lesser originations and the purchase volume barrier that low-income and minority of the sales price or appraised value) from amounting to $1.3 trillion. Estimates indicate families face when considering approved flexible sources or making a that ARMs accounted for 19 percent of total homeownership.105 To assist in overcoming minimum contribution of $500 from their mortgage originations in 2003.104 In its this barrier, the industry (including lenders, own funds. The 3% may come from a variety September 17, 2004 forecast, MBA predicts private mortgage insurers and the GSEs) on sources such as gifts, grants, or unsecured that single-family mortgage originations will began offering in 1994 mortgage products loans from relatives, employers, public amount to $2.7 trillion in 2004 and $1.8 that required down payments of only 3 agencies, or nonprofits. In 2003, Fannie Mae trillion in 2005, with refinancings percent, plus points and closing costs. Other purchased $13.7 billion in Flexible loans that representing 43 percent and 25 percent of industry efforts to reduce borrowers’ up-front benefited 100,866 households.108 originations respectively. costs included zero-point-interest-rate Fannie Mae has also developed products specifically geared toward populations with 4. Affordable Lending in the Mortgage mortgages and monthly insurance premiums unique needs such as seniors, Native Market: New Products and Outreach with no up front component. These new plans eliminated large up-front points and Americans and families living near public Extending homeownership opportunities premiums normally required at closing. transit routes. Examples of these targeted to historically underserved households has During 1998, Fannie Mae introduced its products include the Home Equity been a growing concern for conventional ‘‘Flexible 97’’ and Freddie Mac introduced its Conversion Mortgage (HECM) which allows lenders, private mortgage insurers and the ‘‘Alt 97’’ low down payment lending seniors to convert the equity in their homes GSEs. The industry has responded in what programs. Under these programs, borrowers to receive cash. In 2003, Fannie Mae some have called a ‘‘revolution in affordable were required to put down only 3 percent of purchased 27,644 HECM’s for a total value of lending’’. The industry has offered more TM the purchase price. The down payment, as $1.87 billion. PaymentPower allows customized mortgage products, more flexible borrowers with strong credit to skip their underwriting, and expanded outreach so that well as closing costs, could be obtained from a variety of sources, including gifts, grants or regularly scheduled monthly payment up to the benefits of the mortgage market can be two times during a twelve-month period and extended to those who have not been loans from a family member, the government, a non-profit agency and loans secured by life up to ten times during the life of the loan. adequately served through traditional This pilot was launched in July 2002 and by products, underwriting, and marketing. insurance policies, retirement accounts or other assets. Fannie Mae continues to offer year-end 2003, Fannie Mae purchased 963 PaymentPowerTM mortgages totaling $126 101 Office of Federal Housing Enterprise Oversight the ‘‘Flexible’’ line of products, and Freddie Mac continues to list ‘‘Alt 97.’’ million. Navajo Community Guaranty (OFHEO), Report to Congress, 2004, Tables 1 and Initiative allows Navajo families to contribute 11. 102 Office of Federal Housing Enterprise 105 See Charles, K. K. and E. Hurst (2002). ‘‘The Oversight. ‘‘Mortgage Markets and The Enterprises Transition to Home Ownership and the Black-White 106 Fannie Mae, 2003 Annual Housing Activities in 2001,’’ August 2002, p. 13. Wealth Gap.’’ The Review of Economics and Report, 2004, pp. 8–9. 103 http://www.financialservicesfacts.org/ Statistics, 84(2): 281–297; Mayer, C. and G. 107 Fannie Mae, ‘‘Fannie Mae’s Comments on financial2/mortgage/mortgages/ Engelhardt (1996). ‘‘Gift Down Payments and HUD’s Proposed Housing Goals for Fannie Mae and ?table_sort_734796=4. Housing Affordability.’’ Journal of Housing Freddie Mac for the years 2005–2008 and 104 Mortgage market projections from the MBA’s Research, 7(1): 59–77; and Quercia, R. G., G. W. Amendments to HUD’s Regulation of Fannie Mae MBA Mortgage Finance Forecast, December 17, McCarthy, et al. (2003). ‘‘The Impacts of Affordable and Freddie Mac,’’ July 16, 2004, p. I–58. 2003. 2000 and 2001 numbers from the MBA’s MBA Lending Efforts on Homeownership Rates.’’ Journal 108 Fannie Mae, 2003 Annual Housing Activities Mortgage Finance Forecast, January 10, 2002. of Housing Economics, 12(1): 29–59. Report, 2004, p. 6.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00077 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63656 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
a minimum of $500 or 1% of the purchase consists of funds that have been matched Financial Institution investment; the Native price, whichever is lower. This initiative, through an Individual Development Account American Homeownership Initiative, which announced in 2003, will provide $3 million homebuyer savings program. And in 2003, has committed to invest at least $350 million in home financing to help 60 families Freddie Mac provided increased liquidity for to support homeownership strategies for currently living on a reservation. The Smart affordable housing through a series of 4,600 Native American families and to work CommuteTM Initiative, which targets targeted investments in Mortgage Revenue with 100 tribes; the Minority- and Women- borrowers purchasing homes near a public Bonds containing state and local housing Owned Lenders Initiative, to reach transit route, recognizes that homebuyers finance agency mortgages.114 underserved communities and to develop will save commuting expenses and therefore b. Partnerships—Fannie Mae innovative solutions for increasing business have more disposable income to pay housing opportunities for these lenders; The expenses. In 2003 Fannie Mae purchased In addition to developing new affordable products, lenders and the GSEs have been Employer-Assisted Housing Initiative, approximately $5 million in Smart designed to assist employers in developing a CommuteTM Initiative loans.109 entering into partnerships with local governments and nonprofit organizations to company benefit that helps employees meet In 2000, Freddie Mac introduced its their housing needs; and the Initiative to ‘‘Freddie Mac 100’’ product, which is increase mortgage access to underserved Reduce Barriers to Affordable Housing, designed to assist borrowers who have good borrowers. Fannie Mae operates 55 which has established local partnerships in credit but lack the ability to provide a large partnership offices throughout the country, seven new states and localities in 2003. down payment. ‘‘Freddie Mac 100’’ allows a including the West Virginia Partnership 100 percent loan-to-value ratio with the Office, which opened in 2003. These offices Additionally, Fannie Mae conducts various condition that the borrower has the funds for coordinate Fannie Mae’s programs with local underwriting experiments aimed at closing costs. In 2003, a refinance option was governments, lenders, public officials, eliminating obstacles faced by prospective added to Freddie Mac 100 and the cost of the housing organizations, community homebuyers across the country. In 2003, loan was reduced through lower mortgage nonprofits, real estate professionals, and Fannie Mae approved $222 million worth of 115 insurance coverage and a lower fee for the other local stakeholders. Housing and Community Development place- product. These changes have made the Fannie Mae continues to reach out to based commitments for a total of 55 Freddie Mac 100 available to borrowers who national groups and work with local affiliates experiments.117 may not have been able to take advantage of to expand homeownership. Fannie Mae has Fannie Mae’s American Dream the refinance boom as a result of low or no established multi-year partnerships to Commitment is part of its National Minority equity in their homes.110 increase affordable housing opportunities Homeownership Initiative which has pledged Another Freddie Mac product, Affordable with organizations such as: The Enterprise to contribute at least $700 billion in private Gold 97 permits borrowers to make 3% Foundation, The Neighborhood capital to serve 4.6 million families towards down payments from personal cash and to Reinvestment Corporation, ACORN Housing President George W. Bush’s goal of use other sources to cover their closing costs, Corporation, The National Council of La expanding homeownership to 5.5 million and offers flexible ratio and reserves Raza, and many others engaged in promoting new minority Americans by the end of the guidelines. In 2003 this product was affordable housing. In 2003, Fannie Mae decade. Towards this goal, in 2003, Fannie financed $1.3 billion of mortgages with these enhanced with a refinance option allowing Mae executed 17 new Housing and national partners and participating lenders, more borrowers to take advantage of the low Community Development lender which resulted in 9,597 loans. For example, rates in the market. The Affordable Gold partnerships which seek to provide $394 100 provides 100 percent financing to low- Fannie Mae maintains a partnership with the National Urban League (NUL) and the JP billion in affordable housing lending to and moderate-income borrowers for the 118 Morgan Chase Bank to increase NUL’s minority families. purchase price of a home in California. Under the American Dream Commitment, Affordable Gold 100 combines mortgage homeownership counseling capacity by providing the necessary technology and tools Fannie Mae has committed to establishing insurance benefits provided by a state 250 faith-based homeownership partnerships insurance fund, the secondary mortgage to support the effort, and to purchase $50 million in mortgage products over five years in communities across the country by the end market, and a team of the nation’s leading of the current decade. The objective of this mortgage lenders.111 that are specifically targeted to increase initiative is to build strong partnerships with Additional Freddie Mac products include homeownership among minorities. In 2003, national faith-based organizations in order to the Alt 97SM for borrowers who have good approximately $6 million in loans were reach potential new homeowners, work with credit but limited cash for a down payment. originated through this initiative. Another faith-based and nonprofit partners to help In 2003, this product was enhanced with a example is Fannie Mae’s partnership with refinance option and reduced fees. The Two- the AFL-CIO Housing Investment Trust (HIT) increase access to homeownership Family 95 Percent LTV Program offers low and Countrywide Home Loans, which information and education, partner with down payment loans to purchasers of two- launched ‘‘HIT HOME’’ in 2001. HIT HOME lenders to increase access to mortgage family properties when the borrowers occupy is an affordable home mortgage initiative that financing, and provide faith-based one of the units as their primary residence.112 targets 13 million union members in 35 cities organizations with the tools, training, and Other initiatives include policies aimed at throughout the nation to provide union resources needed to advance their improving the homeownership rate among members with a variety of affordable community development efforts. Fannie immigrant families and the Section 8 Rental mortgage choices that enable them to qualify Mae’s work under the Faith-Based Initiative to Homeownership program, which allows for competitively priced loans with new re- in 2003 resulted in $125 million in mortgage people currently receiving Section 8 rental payment terms. In 2003, over $132 million financing to underserved families across the 119 subsidies to use them toward mortgage worth of mortgages were originated through country. Additionally, Fannie Mae this partnership.116 payments. 113 Freddie Mac purchases loans attended more than 12 faith-based in which the borrower’s down payment In order to meet the needs of underserved symposiums providing training and technical and low- and moderate-income populations, assistance to over 2,000 symposium Fannie Mae has targeted specific populations 120 109 attendees. Fannie Mae, 2003 Annual Housing Activities for initiatives. These include the Section 8 Report, 2004, pp. 9–10. Homeownership Initiative, which purchased 110 Freddie Mac, Opening Doors for America’s 117 Fannie Mae, 2003 Annual Housing Activities Families: Freddie Mac’s Annual Housing Activities 81 Section 8 loans and funded an additional Report, 2004, pp. 17–22. Report for 2003, March 15, 2004, p. 62. 55 loans through a Community Development 118 Fannie Mae, 2003 Annual Housing Activities 111 Freddie Mac, Opening Doors for America’s Report, 2004, pp. 16. Families: Freddie Mac’s Annual Housing Activities 114 Freddie Mac, Opening Doors for America’s 119 Fannie Mae, 2003 Annual Housing Activities Report for 2003, March 15, 2004, p. 62. Families: Freddie Mac’s Annual Housing Activities Report, 2004, pp. 17–18. 112 Freddie Mac, Opening Doors for America’s Report for 2003, March 15, 2004, pp. 62–64. 120 Fannie Mae, ‘‘Fannie Mae’s Comments on Families: Freddie Mac’s Annual Housing Activities 115 Fannie Mae, 2003 Annual Housing Activities HUD’s Proposed Housing Goals for Fannie Mae and Report for 2003, March 15, 2004, p. 62–64. Report, 2004, pp. 22–24. Freddie Mac for the years 2005–2008 and 113 Freddie Mac Public Comment Letter on HUD’s 116 Fannie Mae, 2003 Annual Housing Activities Amendments to HUD’s Regulation of Fannie Mae Proposed Goals, July 2004, p. 2. Report, 2004, pp. 13–16. and Freddie Mac,’’ July 16, 2004, p. I–60.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00078 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63657
c. Partnerships—Freddie Mac workshops for their partners and their local The programs mentioned above are partners resulting in 326 trainers who are examples of the partnership efforts Freddie Mac does not have a partnership office structure similar to Fannie Mae’s, but authorized to teach the CreditSmart undertaken by the GSEs. There are more it has undertaken a number of initiatives in Espan˜ ol curriculum. Thus far 503 adults have partnership programs than can be adequately specific metropolitan areas.121 Freddie Mac been trained in the CreditSmart Espan˜ ol described here. Fuller descriptions of these works with affordable housing lenders to financial literacy program.126 The programs are provided in their Annual design creative solutions to meet CreditSmart /Homeownership Development Housing Activity Reports. homeownership needs of specific Initiative with the National Urban League has d. Underwriting and GSE Purchase populations in targeted areas; explore nine affiliates located in Birmingham, AL; Guidelines efficient use of public subsidies to make Charlotte, NC; Louisville, KY; Greenville, SC; Lenders, mortgage insurers, and the GSEs homeownership more affordable and develop Oklahoma City, OK; Springfield, IL; and have also been modifying their mortgage homebuyer education/counseling and debt Washington, DC; with Orlando, FL and underwriting standards to address the needs management assistance programs.122 In 2001, Knoxville, TN added in 2003. Since the of families who have historically found it Freddie Mac joined the Congressional Black initiative’s launch in early 2002, 41 difficult to qualify under traditional Caucus to launch a new initiative, ‘‘With CreditSmart financial literacy workshops guidelines. In addition to the changes in Ownership Wealth,’’ designed to increase have been presented to more than 600 underwriting standards, the use of automated African-American homeownership with one minority participants. Those participants are underwriting has dramatically transformed million new families by 2005.123 Freddie proceeding to the next steps to achieving Mac has partnered with the National Council homeownership, and in 2003 313 loans have the mortgage application process. This of La Raza (NCLR), 20 community based closed as a direct result.127 section focuses on changes to traditional NCLR affiliated housing counseling In 2002 and 2003, Freddie Mac joined with underwriting standards and recent GSE organizations, the National Association of the American Community Bankers, the Credit initiatives for credit-impaired borrowers. Hispanic Real Estate Professionals Union National Association, and the Subsequent sections will provide more (NAHREP), EMT Applications and Independent Community Bankers of America details on the impact of automated participating Freddie Mac Seller/Servicers in strategic alliances to better enable member underwriting. including Bank of America, U.S. Bank and banks and credit unions access to the The GSEs modified their underwriting Wells Fargo Home Mortgage on the ‘‘En Su secondary market.128 standards to address the needs of families Casa’’ initiative. This $200 million In June 2002, President George W. Bush who find qualifying under traditional guidelines difficult. The goal of these homeownership initiative combines challenged the nation’s housing industry to underwriting changes is not to loosen technology tools with flexible mortgage invest more than $1 trillion to make underwriting standards, but rather to identify products to meet the needs of Hispanic homeownership a reality for 5.5 million more creditworthiness by alternative means that borrowers. Mortgage products include low minority households for the decade. Freddie more appropriately measures the unique down payments, flexible credit underwriting Mac responded to the challenge with Catch circumstances of low-income, immigrant, and debt-to-income ratios, and streamlined the Dream which is a comprehensive set of and minority households. Examples of processing for resident alien borrowers.124 25 high impact initiatives aimed at changes that the GSEs and others in the In 2002, Freddie Mac joined with the City accelerating the growth in minority industry have made to their underwriting of Boston and the U.S. Conference of Mayors homeownership. The initiatives range from standards include the following: to make available the ‘‘Don’t Borrow homebuyer education and outreach, to new • Trouble’’ predatory lending educational Using a stable income standard rather technologies with innovative mortgage than a stable job standard (or a minimum campaign to approximately 1,100 cities. As of products. Freddie Mac has committed to the end of 2003, the campaign has been period of employment). This particularly purchase $400 billion in mortgages made to benefits low-skilled applicants who have launched in more than 30 localities. minority families by the end of the decade.129 Additionally, in late 2003, Freddie Mac successfully remained employed, even with Catch the Dream represents a collaborative frequent job changes. sponsored a national Don’t Borrow Trouble effort with lenders, nonprofit housing and • summit. Attorneys, community activists and Using an applicant’s history of rent and community-based organizations, and other utility payments as a measure of local leaders from 23 cities convened to share industry participants to expand campaign experiences and to learn about creditworthiness. This measure benefits homeownership opportunities for America’s lower-income applicants who have not emerging predatory lending trends from some minorities.130 In 2003 initiatives were of the nation’s leading community lending established a credit history. implemented in Birmingham, Charlotte, • Allowing pooling of funds for experts.125 Atlanta, DeKalb County (GA), Lansing, and In addition, Freddie Mac joined with qualification purposes. This change benefits San Antonio. In 2003, single-family owner applicants with extended family members. Rainbow/PUSH and the National Urban occupied mortgage purchases financed Freddie Mac, for example, allows income League to promote the CreditSmart homes for almost 700,000 minority families, financial educational curriculum that helps from relatives who live together to pool their including mortgages for 133,000 African- funds to cover downpayment and closing consumers understand, obtain and maintain American and 250,000 Hispanic families good credit, thereby preparing them for costs and to combine their incomes for use (this comprised 16% of Freddie Mac’s single- in calculating the borrower’s stable monthly homeownership and other personal financial family, owner-occupied mortgage purchases goals. Rainbow/PUSH has organized income. and 22.6% of their first-time homebuyer These underwriting changes have been CreditSmart classes with more than 80 mortgage purchases).131 churches across the nation, reaching more accompanied by homeownership counseling than 2,500 congregants. Bilingual curriculum to ensure homeowners are ready for the 126 was launched for this program in December Freddie Mac, Opening Doors for America’s responsibilities of homeownership. In Families: Freddie Mac’s Annual Housing Activities addition, the industry has engaged in 2002, and during 2003 CreditSmart Espan˜ ol Report for 2003, March 15, 2004, pp. 38–39. conducted a total of 23 Train-the-Trainer intensive loss mitigation to control risks. 127 Freddie Mac, Opening Doors for America’s In 1999, HUD commissioned a study by the Families: Freddie Mac’s Annual Housing Activities Urban Institute to examine the underwriting 121 Freddie Mac, News Release, January 15, 1999. Report for 2003, March 15, 2004, pp. 39–40. 122 128 criteria that the GSEs use when purchasing Freddie Mac Public Comment Letter on HUD’s Freddie Mac, Opening Doors for America’s 132 Proposed Goals, July 2004, p. 3. Families: Freddie Mac’s Annual Housing Activities mortgages from primary lenders. 123 Freddie Mac, Opening Doors for America’s Report for 2003, March 15, 2004, pp. 42–43. According to the study, while the GSEs had Families: Freddie Mac’s Annual Housing Activities 129 Freddie Mac Public Comment Letter on HUD’s improved their ability to serve low- and Report for 2003, March 15, 2004, pp. 67. Proposed Goals, July 2004, p. 4. moderate-income borrowers, it did not 124 Freddie Mac, Opening Doors for America’s 130 Freddie Mac, Opening Doors for America’s Families: Freddie Mac’s Annual Housing Activities Families: Freddie Mac’s Annual Housing Activities 132 Kenneth Temkin, Roberto Quercia, George Report for 2003, March 15, 2004, pp. 66–67. Report for 2003, March 15, 2004, pp. 29–30. Galster, and Sheila O’Leary, A Study of the GSEs’ 125 Freddie Mac, Opening Doors for America’s 131 Freddie Mac, Opening Doors for America’s Single Family Underwriting Guidelines: Final Families: Freddie Mac’s Annual Housing Activities Families: Freddie Mac’s Annual Housing Activities Report. Washington DC: U.S. Department of Report for 2003, March 15, 2004, pp. 37–38. Report for 2003, March 15, 2004, pp. 30–34. Housing and Urban Development, April 1999.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00079 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63658 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
appear at that time that they had gone as far established a credit history. During 2003, factors that Loan Prospector uses to analyze as some primary lenders to serve these Freddie Mac entered into several new loans, and put the list on the Freddie Mac borrowers. From the Urban Institute’s markets under the LeasePurchase Plus website.137 discussion with lenders, it was found that Initiative and purchased more than $16 In 2003, Fannie Mae released two versions primary lenders were originating mortgages million in loans.134 of its automated underwriting service, to lower-income borrowers using According to Freddie Mac, its automated ‘‘Desktop Underwriter’’ (DU), to expand its underwriting guidelines that allow lower underwriting system, ‘‘Loan Prospector’’ has mortgage product offerings and to update down payments, higher debt-to-income ratios reduced costs, made approving mortgages underwriting guidelines. Desktop and poorer credit histories than allowed by easier and faster, and increased the Underwriter 5.3 outlined new eligibility the GSEs’ guidelines. consistency of the application of objective From this and other evidence, the Urban underwriting criteria. In addition, Freddie requirements for mortgages secured by Institute concluded that the GSEs were Mac states that ‘‘Loan Prospector’’ extends manufactured homes. It also expanded the lagging the market in servicing low- and the benefits of the mortgage finance system InterestFirstTM mortgage product line to offer moderate-income and minority borrowers. to borrowers with less traditional credit borrowers greater purchasing power by Furthermore, the Urban Institute found ‘‘that profiles and limited savings by more allowing lower initial monthly payments the GSEs’ efforts to increase underwriting accurately measuring risk. Since its than those available with traditional loan flexibility and outreach has been noticed and introduction in 1995, Freddie Mac reports products. Desktop Underwriter 5.3.1 is applauded by lenders and community that they have doubled their share of enhanced the Flexible 100 mortgage to allow advocates. Despite the GSEs’ efforts in recent mortgage purchases with loan-to-value borrowers to contribute as little as $500 of years to review and revise their underwriting rations of 95 percent or above.135 In 2003, their own funds to the transaction. The criteria, however, they could do more to lenders and brokers used Loan Prospector to remainder of the funds can come from serve low- and moderate-income borrowers evaluate 9.5 million loan applications and flexible sources of funds and interested party and to minimize disproportionate effects on Loan Prospector has evaluated more than 35 contributions subject to Fannie Mae’s minorities.’’ 133 Since the Urban Institute million mortgage applications since its standard contribution limit.138 In addition, study, Freddie Mac and Fannie Mae have introduction in 1995.136 Freddie Mac reports Fannie Mae added MyCommunityMortgage been playing a larger role in financing low- that its automated underwriting system, Loan income and minority borrowers. (See Section Prospector, has resulted in higher approval to Desktop Underwriter in 2003, providing E.2.) rates for minority borrowers than under lenders easier access to customized CRA- In addition to offering low-down-payment traditional manual underwriting because of targeted loan products.139 Automated programs, the GSEs’ recent efforts have also improved predictive powers. As mentioned mortgage scoring and the potential for centered around their automated in Section C.7, the 2000 version of LP disparate impacts on borrowers will be underwriting systems and their treatment of approved 87.1 percent of loans generated further discussed in a later section. borrowers with blemished credit, the latter through affordable housing programs, 5. Affordable Single-family Lending: Data being perhaps the most controversial compared to 51.6 percent approved by Trends underwriting issue over the past few years. manual underwriting. The Freddie Mac study Freddie Mac has a variety of products and found automated mortgage scoring less a. 1993–2003 Lending Trends initiatives aimed at providing borrowers with discriminatory and more accurate in HMDA data suggest that the industry and impaired credit more mortgage choices. predicting risk. However, as noted below in GSE initiatives are increasing the flow of These products include: CreditWorksSM the automated mortgage scoring section, credit to underserved borrowers. Between which helps borrowers with excessive debt there are concerns that the codification of and impaired credit to become eligible for a certain underwriting guidelines could result 1993 and 2003, conventional loans to low- prime market rate mortgage faster than would in unintentional discrimination or disparate income and minority families increased at otherwise be possible, Affordable Merit treatment across groups. In response to the much faster rates than loans to higher income RateSM Mortgage which permits borrowers to potential disparate impact of automated and non-minority families. As shown below, qualify at an initial interest rate that in many underwriting, Freddie Mac have launched conventional home purchase originations to cases is lower than the usual subprime rate, initiatives to make the mortgage process more African Americans more than doubled and LeasePurchase Plus Initiative, which transparent by disclosing both credit and between 1993 and 2003 and those to provides closing cost and down payment non-credit factors that Loan Prospector Hispanic borrowers more than tripled. Home assistance in addition to extensive consider when evaluating a loan application. loans to low-income borrowers and to low- counseling for borrowers who have had In 2000, Freddie Mac has launched an income and high-minority census tracts also credit issues in the past or who have never initiative that published a list of all of the more than doubled during this period.
1993–2003 1993–2003 Growth rate: Growth rate: all conventional home loans home loans (percent) (percent)
African-American Borrowers ...... 106 206 Hispanic Borrowers ...... 235 357 White Borrowers ...... 44 64 Low-Income Borrower (Less than 80% of AMI) ...... 101 150 Upper-Income Borrower (More than 120% of AMI) ...... 88 108 Low-Income Census Tract (only 1993–2002) ...... 99 143 Upper-Income Census Tract (only 1993–2002) ...... 64 78 High-Minority Tract (only 1993–2002) (50% or more minority) ...... 113 167 Predominantly-White Tract (only 1993–2002) (Less than 10% minority) ...... 53 64
133 Temkin, et al. 1999, p. 28. 136 Freddie Mac, Opening Doors for America’s 139 Fannie Mae, ‘‘Fannie Mae’s Comments on 134 Freddie Mac, Opening Doors for America’s Families: Freddie Mac’s Annual Housing Activities HUD’s Proposed Housing Goals for Fannie Mae and Families: Freddie Mac’s Annual Housing Activities Report for 2003, March 15, 2004, p. 19. Freddie Mac for the years 2005–2008 and 137 Report for 2003, March 15, 2004, pp. 36–37. Freddie Mac, 2002 Annual Housing Activities Amendments to HUD’s Regulation of Fannie Mae Report, 2003, p. 57. 135 Freddie Mac Public Comment Letter on HUD’s and Freddie Mac,’’ July 16, 2004, p. I–57. 138 Fannie Mae, 2003 Annual Housing Activities Proposed Goals, July 2004, p. 5. Report, 2004, pp. 11–12.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00080 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63659
GSE purchases showed similar trends, as b. Affordable Lending Shares by Major insured by FHA, purchased by the GSEs, indicated by the following 1993-to-2003 Market Sector originated by depository institutions (mainly percentage point increases for metropolitan Section E below compares the GSEs’ banks and thrift), and originated in the areas: African-American borrowers (199 performance with the performance of conventional conforming market and in the percent), Hispanic borrowers (259 percent), primary lenders in the conventional total market for owner-occupied properties in and low-income borrowers (212 percent). conforming market. To provide a useful 140 While their annual purchases of all home metropolitan areas. In this case, the ‘‘total’’ context for that analysis, this section loans increased by 60 percent between 1993 market consists of both the conventional and 2003, their purchases of mortgages that examines the role of the conventional conforming market and the government qualify for the three housing goals increased conforming market in funding low-income (mainly FHA and VA loans) market; ‘‘jumbo’’ as follows: special affordable by 287 percent; and minority families and their loans above the conventional conforming neighborhoods. Information on the mortgage low- and moderate-income by 156 percent; loan limit are excluded from this analysis.141 market’s funding of homes purchased by and underserved areas by 121 percent. BILLING CODE 4210–27–P While low interest rates and economic first-time homebuyers is also provided. In expansion certainly played an important role addition, this section compares the GSEs in the substantial increase in conventional with other sectors of the mortgage market. affordable lending in recent years, most The important role of FHA in the affordable 140 Table A.3 also provides the same average observers believe that the efforts of lenders, lending market is highlighted and questions (1999 to 2003) information as Tables A.1 and A.2 private mortgage insurers, and the GSEs were are raised about whether the conventional but for total (both home purchase and refinance) also important contributors. In addition, conforming market could be doing a better loans. Thus, it provides a complete picture of many observers believe that government job helping low-income and minority overall mortgage activity. initiatives such as the GSE housing goals and borrowers obtain access to mortgage credit. 141 The ‘‘Total Market’’ is defined as all loans the Community Reinvestment Act have also Table A.1 reports borrower characteristics (including both government and conventional) played a role in the growth of affordable and Table A.2 reports neighborhood below the conforming loan limit of $240,000 in lending over the past 10 years. characteristics for home purchase mortgages 1999, $252,700 in 2000, $275,000 in 2001, $300,700 in 2002 and $322,700 in 2003.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00081 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63660 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00082 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02no04.010 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63661
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00083 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02no04.011 63662 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00084 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02no04.012 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63663
HMDA is the source of the FHA, targeted group and that sector must be of the conventional conforming sector are low depository, and market data, while the GSEs some size. The discussion below will focus for minority borrowers, particularly African- provide their own data. Low-income, on the degree to which different mortgage American and Hispanic borrowers. These African-American, Hispanic, and minority sectors concentrate on targeted groups, while borrowers accounted for only 15.2 percent of borrowers are covered in Table A.1. Table Section E will also provide estimates of all conventional conforming loans originated A.2 provides information on four types of market shares. between 1999 and 2003, compared with 34.4 neighborhoods—low-income census tracts, The main insights from the ‘‘distribution of percent of FHA-insured loans and 19.2 tracts where minorities (or African business’’ percentages in Tables A.1 and A.2 percent of all loans originated in the total Americans) account for more than 30 percent pertain to four topics. (government and conventional conforming) of the census tract population, and (i) FHA-Insured Loans. FHA has market. Not surprisingly, the minority underserved areas as defined by HUD. The traditionally been the mechanism used by lending performance of conventional lenders average data reported in Tables A.1 and A.2 borrowers who face difficulty obtaining has been subject to much criticism. Recent for the years 1999 to 2003 offer a good mortgage financing in the private studies contend that primary lenders in the summary of recent lending to low-income conventional market. FHA has long been conventional market are not doing their fair and minority borrowers and their recognized as the major source of funding for share of minority lending which forces communities.142 Individual year data are also first-time, low-income and minority minorities, particularly African-American provided. homebuyers who are not often able to raise and Hispanic borrowers, to rely on more The focus of different market sectors on cash for large downpayments.143 Tables A.1 costly FHA and subprime loans.145 Thus, it affordable lending is summarized by the and A.2 show that FHA places much more appears that conventional lenders could be percentages reported in Tables A.1 and A.2. emphasis on affordable lending than the doing a better job helping minority borrowers These percentages show each sector’s other market sectors. Between 1999 and obtain access to mortgage credit. ‘‘distribution of business,’’ defined as the 2003, low-income borrowers accounted for • The GSEs’ funding of minority loans can share of loans originated (or, for the GSEs, 51.2 percent of FHA-insured loans, compared be compared with mortgages originated for purchased) that had a particular borrower or with 27.8 percent of the home loans minority borrowers in the conventional neighborhood characteristic. The purchased by the GSEs, 29.1 percent of home conforming market, although the latter may interpretation of the ‘‘distribution of loans originated by depositories, and 29.2 be a poor benchmark, as discussed above. business’’ percentages can be illustrated percent of all originations in the Between 1999 and 2003, home purchase using the FHA percentage for low-income conventional conforming market (see Table loans to African-American and Hispanic borrowers: Between 1999 and 2003, 51.2 A.1). Likewise, 40.7 percent of FHA-insured borrowers accounted for 10.4 percent of percent of all FHA-insured home purchase loans were originated in underserved census Freddie Mac’s purchases, 14.0 percent of loans in metropolitan areas were originated tracts, while only 24.1 percent of the GSE- Fannie Mae’s purchases, and 15.2 percent of for borrowers with an income less than 80 purchased loans, 25.9 percent of home loans loans originated in the conventional percent of the local area median income. originated by depositories, and 26.9 percent conforming market (or 14.3 percent if B&C These percentages are to be contrasted with of conventional conforming loans were loans are excluded from the market 144 ‘‘market share’’ percentages, which are originated in these tracts (see Table A.2). definition). Thus, since 1999, the African- presented below in Section E. A ‘‘market As discussed in Section E, FHA’s share of the American and Hispanic share of the GSEs’ share’’ percentage is the share of loans with minority lending market is particularly high. purchases has been lower than the a particular borrower or neighborhood While FHA insured only 16 percent of all corresponding share for the conventional home purchase mortgages originated below characteristic that was funded by a particular conforming market.146 the conforming loan limit in metropolitan market sector (e.g., FHA-insured, GSEs, • As the above comparisons show, Fannie areas between 1999 and 2003, it is estimated depositories). As will discussed below, Mae has had a much better record than that FHA insured 29 percent of all home FHA’s ‘‘market share’’ for low-income Freddie Mac in funding loans for minority loans originated for African-American and borrowers during the 1999-to-2003 period families. And Fannie Mae significantly Hispanic borrowers. was estimated to be 24 percent which is increased its purchases of loans for African- (ii) Conventional and GSE Minority interpreted as follows: Of all home purchase Lending. The affordable lending shares for American and Hispanic borrowers during loans originated for low-income borrowers in 2001, raising the share of its purchases to metropolitan areas between 1999 and 2003, market levels—13.7 percent for both Fannie 143 24 percent were FHA-insured loans. Thus, in Almost two-thirds of the borrowers with an Mae and the conforming market (without FHA-insured home purchase loan make a this example, the ‘‘distribution of business’’ B&C loans). In 2002, Fannie Mae surpassed percentage measures the importance (or downpayment less than five percent, and over 80 percent are first-time home buyers. For discussions concentration) of low-income borrowers in of the role of FHA in the mortgage market, see (a) 145 See Green and Associates, Fair Lending in FHA’s overall business while the ‘‘market Harold L. Bunce, Charles A. Capone, Sue G. Neal, Montgomery County: A Home Mortgage Lending share’’ percentage measures the importance William J. Reeder, Randall M. Scheessele, and Study, a report prepared for the Montgomery of FHA to the market’s overall funding of Edward J. Szymanoski, An Analysis of FHA’s County Human Relations Commission, March 1998; loans for low-income borrowers. Both Single-Family Insurance Program, Office of Policy and Calvin Bradford, Crisis in De´ja` vu: A Profile of concepts are important for evaluating Development and Research, U.S. Department of the Racial Patterns in Home Purchase Lending in performance—for an industry sector such as Housing and Urban Development, 1995; and (b) the Baltimore Market. Report for The Public Justice FHA or the GSEs to have a significant impact Office of Policy Development and Research, ‘‘FHA’s Center, May 2000; and The Patterns of GSE Impact on Homeownership Opportunities for Low- Participation in Minority and Racially Changing on lending to a targeted group, that sector’s Income and Minority Families During the 1990s’’ Markets Reviewed from the Context of Levels of business must be concentrated on the Issue Brief IV, U.S. Department of Housing and Distress Associated with High Levels of FHA Urban Development, December 2000. For data on Lending, GSE Study No. 11, U.S. Department of 142 The affordable market shares reported in Table the credit characteristics of FHA borrowers, see Housing and Urban Development, September 2000. A.1 for the ‘‘Conventional Conforming Market Harold L. Bunce, William J. Reeder and Randall For analysis suggesting some minorities receiving W/O B&C’’ were derived by excluding the estimated Scheessele, ‘‘Understanding Consumer Credit and FHA loans could qualify for conventional loans, see number of B&C loans from the market data reported Mortgage Scoring: A Work in Progress at HUD’’, Anthony Pennington-Cross, Anthony Yezer, and by HMDA. Because B&C lenders operate mainly in U.S. Department of Housing and Urban Joseph Nichols, Credit Risk and Mortgage Lending: the refinance sector, excluding these loans from the Development, Unpublished Paper, 1999. Who Uses Subprime and Why? Working Paper No. conforming market has little impact on the home 144 FHA, which focuses on low downpayment 00–03. Research Institute for Housing America, purchase percentages reported in Table A.1. It loans and also accepts borrowers with credit 2000. Also see the series of recent studies should be recognized that there exists some blemishes, experiences higher mortgage defaults concerning the lack of mainstream lenders in uncertainty regarding the number of B&C loans in than conventional lenders and the GSEs. Still, the minority neighborhoods. the HMDA data. The adjustment assumes that the FHA system is actuarially sound because it charges 146 For a comprehensive analysis of the GSEs’ B&C loans represent one-half of the subprime an insurance premium that covers the higher purchases of minority loans through 1999, see market. The adjustment for home purchase loans is default costs. For the results of FHA’s actuarial Harold L. Bunce, An Analysis of GSE Purchases of small because supbrime (B&C) loans are mainly analysis, see Deloitte & Touche, Actuarial Review of Mortgages for African-American Borrowers and refinance loans. The method for excluding B&C MMI Fund as of FY 2000, report for the U.S. their Neighborhoods, Housing Finance Working loans is explained in Section E below and Department of Housing and Urban Development, Paper No. 11, Office of Policy Development and Appendix D. January 2001. Research, HUD, December 2000.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00085 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63664 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
the conventional conforming market in again below the market in low-income census 2001,150 and for 37.6 percent of home loans funding African-American and Hispanic tracts (a 10.3 share for Freddie Mac and a originated in the conventional conforming borrowers (a 15.8 percent share for Fannie 11.5 percent share for the market). A more market. The AHS defines a first-time Mae and a 15.0 share for the market), but in complete analysis of the GSEs’ recent homebuyer as someone who has never 2003 fell slightly behind the market (a 16.6 improvements in purchasing home loans that owned a home. Using a more liberal percent share for Fannie Mae and a 16.9 qualify for the housing goals is provided definition of a first-time homebuyer percent share for the market). When all below in Section E. (someone who has not owned a home in the minority borrowers are considered, Fannie (iv) Depositories. Within the conventional past three years), FHA reports that first-time Mae has purchased mortgages for minority conforming market, depository institutions homebuyers accounted for 80.5 percent of all borrowers at a higher rate (years 2001, 2002 (mainly banks and thrifts) are important home loans that it insured between 1999 and and 2003) than these loans were originated providers of affordable lending for lower- 2001 and the GSEs report that first-time by primary lenders in the conventional income families and their neighborhoods.147 homebuyers accounted for 26.5 percent of the conforming market (without B&C loans). Between 1999 and 2003, underserved areas home loans purchased by each GSE during Freddie Mac, on the other hand, lagged accounted for 26.9 percent of loans held in that same period. Given FHA’s low behind both the market and Fannie Mae in depository portfolios, which compares downpayment requirements, it is not funding loans for minority borrowers during favorably with the underserved areas surprising that FHA focuses on first-time 2001–2003, as well as during the entire 1999- percentage (26.2 percent) for the overall homebuyers. The GSEs, on the other hand, to-2003 period. The share of Freddie Mac’s conventional conforming market.148 fall at the other end of the continuum, with purchases for African-American and Depository lenders have extensive knowledge their first-time homebuyer share (26.5 Hispanic borrowers declined from 10.9 of their communities and direct interactions percent) falling far short of the first-time percent in both 2000 and 2001 to 10.1 with their borrowers, which may enable them homebuyer share (37.6 percent) of the percent in 2002 before rising slightly to 10.7 to introduce flexibility into their conventional conforming market. Section E percent in 2003. will include a more detailed comparison of • underwriting standards without unduly Considering the minority census tract increasing their credit risk. The Community the GSEs and the conventional conforming data reported in Table A.2, Fannie Mae Reinvestment Act provides an incentive for market in serving first-time homebuyers. In lagged behind the conforming market banks and thrifts to initiate affordable addition, Section E will conduct a market (without B&C loans) in high-minority lending programs with underwriting share analysis that examines the funding of neighborhoods and in high-African-American flexibility and to reach out to lower income minority first-time homebuyers. Consistent neighborhoods during the 1999-to-2003 families and their communities.149 Many of with the earlier discussion, that analysis period. However, Fannie Mae improved its the CRA loans are held in portfolio by suggests that conventional lenders and the mortgage purchases in African-American lenders, rather than sold to Fannie Mae or GSEs have played a relatively small role in neighborhoods after 2001 and essentially Freddie Mac. the market for minority first-time matched the market in 2001–2003. And (v) First-time Homebuyers. As explained in homebuyers. One analysis reported in during 2001, 2002 and 2003, Fannie Mae also Section E, market information on first-time Section E estimates that mortgage purchases purchased loans in high-minority census homebuyers is not as readily available as the by the GSEs between 1999 and 2001 totaled tracts at a higher rate than loans were HMDA data reported in Tables A.1 and A.2 41.5 percent of all home loans originated, but originated by conventional lenders in these on the income and racial characteristics of they accounted for only 14.3 percent of home tracts. While Freddie Mac has generally borrowers and census tracts served by the loans originated for first-time African- lagged the primary market in funding mortgage market. However, the limited American and Hispanic homebuyers. minority neighborhoods, note in Table A.2 market data that are available from the c. Community Reinvestment Act that high African-American tracts increased American Housing Survey, combined with The Community Reinvestment Act (CRA) from 3.9 percent of Freddie Mac’s purchases the first-time homebuyer data reported by in 2001 to 5.3 percent in 2002, placing requires depository institutions to help meet FHA and the GSEs, indicate a rather large the credit needs of their communities.151 Freddie Mac above the conventional variation in the funding of first-time conforming market level (4.6 percent) in CRA loans are typically made to low-income homebuyers across the different sectors of the borrowers earning less than 80 percent of 2002. However, in 2003, Freddie Mac fell mortgage market. Based on the American behind the market. area median income, and in moderate- Housing Survey (AHS), it is estimated that income neighborhoods. CRA provides an (iii) Low-Income Lending by the GSEs. first-time homebuyers accounted for 42.3 Information is also provided on the GSEs’ incentive for lenders to initiate affordable percent of all home purchase loans originated lending programs with underwriting purchases of home loans for low-income throughout the market between 1999 and borrowers (A.1) and for families living in flexibility. CRA loans are usually smaller low-income neighborhoods (A.2). than typical conventional mortgages and also 147 Historically, the GSEs have lagged behind the Tables A.1, A.2, and A.3 include data for all are more likely to have a higher LTV, higher conventional conforming market in funding home loans originated by depositories as well as for debt-to-income ratios and no payment the subset of loans originated but not sold, the latter reserves, and may not be carrying private affordable loans for these groups. During the being a proxy for loans held in depository 1999-to-2003 period, low-income borrowers mortgage insurance (PMI). Generally, at the portfolios. (See the notes to Table A.1 for time CRA loans are originated, many do not (census tracts) accounted for 27.4 (9.7) definitions of the depository data.) meet the underwriting guidelines required in percent of Freddie Mac’s purchases, 28.1 148 However, as shown in Table A.1 , depository (10.1) percent of Fannie Mae’s purchases, institutions resemble other conventional lenders in order for them to be purchased by one of the 29.1 (11.2) percent of loans originated by their relatively low level of originating loans for GSEs. Therefore, many of the CRA loans are depositories, and 29.1 (11.1) percent of home African-American, Hispanic and minority held in portfolio by lenders, rather than sold loans originated by conventional conforming borrowers. Within the conventional conforming to Fannie Mae or Freddie Mac. Evidence is lenders (without B&C loans). By the end of market, Fannie Mae has done a better job than growing that CRA-type lending to low- depositories in funding minority borrowers, income families can be profitable, this period, Fannie Mae had significantly particularly Hispanic borrowers and minority improved its performance relative to the particularly when combined with intensive borrowers as a group. During the last three years, loss mitigation efforts to control credit risk. market. In 2003, low-income borrowers Fannie Mae has also funded African-American accounted for 31.0 percent of Fannie Mae’s borrowers at a higher rate than have depository In a recent survey conducted by the Federal purchases, compared with 29.2 percent for institutions. the conforming market. It is also interesting 149 CRA loans are typically made to low-income 150 In this case, the market includes all that even though Freddie Mac lagged the borrowers earning less than 80 percent of area government and conventional loans, including median income, and in moderate-income jumbo loans. market in funding home loans for low- 151 income borrowers during 2002 (28.6 percent neighborhoods. For a comprehensive analysis of For a comprehensive analysis of CRA and its CRA and its impact on affordable lending, see impact on affordable lending, see Robert E. Litan, versus 29.1 percent), it surpassed the market Robert E. Litan, Nicolas P. Retsinas, Eric S. Belsky Nicolas P. Retsinas, Eric S. Belsky and Susan White in financing properties in low-income census and Susan White Haag, The Community Haag, The Community Reinvestment Act After tracts (11.3 percent versus 11.1 percent). Reinvestment Act After Financial Modernization: A Financial Modernization: A Baseline Report, U.S. During 2003, Freddie Mac’s performance was Baseline Report, U.S. Department of Treasury, 2000. Department of Treasury, 2000.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00086 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63665
Reserve, lenders reported that most CRA is offered by over 300 lender partners until they have been in this country for loans are profitable although not as profitable nationwide, and marries targeted pricing eleven years. As noted by Fannie Mae staff, as the lenders’ standard products.152 with affordability features, such as 100 ‘‘there are enough immigrants already in this Some anticipate that the big growth market percent loan-to-value ratios with only $500 country to keep housing strong for at least six over the next decade for CRA-type lending from the borrower’s own funds.155 In 2003, and perhaps even 10 more years’’.159 As will be urban areas. There has been some Fannie Mae purchased or securitized more these demographic factors play out, the movement of population back to cities, than $2.27 billion of MyCommunityMortgage overall effect on housing demand will likely consisting of aging Baby Boomers (so-called products, which helped provide affordable be sustained growth and an increasingly ‘‘empty nesters’’), the children of Baby housing solutions for 20,400 households.156 diverse household population from which to Boomers (the Echo Boomers aged 18–25), and In addition, Freddie Mac is also purchasing draw new homeowners. immigrants, particularly Hispanics but also seasoned affordable mortgage portfolios Surveys indicate that these demographic Asians.153 The current low homeownership originated by depositories to help meet their trends will be reinforced by the fact that most in inner cities (compared with the suburbs) CRA objectives. In 2003, Freddie Mac Americans desire, and plan, to become also suggests that urban areas may be a developed credit enhancements that enable homeowners. According to the 2002 Fannie potential growth market for lenders. Lenders depositories to profitably sell their loans to Mae Foundation annual National Housing are beginning to recognize that urban Freddie Mac—these transactions facilitate Survey, Americans rate homeownership as borrowers are different from suburban targeted affordable lending activity by the best investment they can make, far ahead borrowers. A new or recent immigrant may providing immediate liquidity. Freddie Mac of 401Ks, retirement accounts, and stocks. have no credit history or, more likely, a loan- also increased its ability to purchase smaller The percentage of Americans who said it was worthy credit history that can’t be portfolios opening this option to many a good time to buy a home was at its highest 154 substantiated by the usual methods. community banks that otherwise would not level since 1994 at 75 percent, a jump of 21 Products for duplexes and four-plexes are not have an outlet for their portfolios.157 The percentage points since May 2001.160 In the same as a mortgage for a subdivision billions of dollars worth of CRA loans that addition, the survey found that 27 percent of house in the suburbs. Programs are being will be originated, as well as the CRA loans Americans report they are likely to buy in the implemented to meet the unique needs of being held in bank and thrift portfolios, offer next three years, and 23 percent of those have urban borrowers. One program emphasizing both GSEs an opportunity to improve their started to save or have saved enough money urban areas was initiated by the American performance in the single-family area. for a down payment.161 Community Bankers (ACB). Under the ACB 6. Potential Homebuyers Further increases in the homeownership program, which made $16.2 billion in loans rate depend on whether or not recent gains in 2002, lenders originated a variety of While the growth in affordable lending and in the home owning share(s) of specific affordable products for first-time homebuyers homeownership has been strong in recent groups are maintained. Minorities accounted and non-traditional borrowers that are then years, attaining this Nation’s homeownership for 17 percent of owner households in 2001, sold to Fannie Mae, Freddie Mac, goals will not be possible without tapping but the Joint Center for Housing Studies Countrywide, or other investors that are into the vast pool of potential homebuyers. reports that minorities were responsible for partnering with the ACB. It is reported that Due to record low interest rates, expanded more than 40 percent (a total of 5.2 million) some lenders are making these non- homeownership outreach, and new flexible of the net growth in homeowners between traditional loans for the first time. mortgage products, the homeownership rate 1993 and 2002.162 As reported by the Fannie For banks and thrifts, selling their CRA reached an annual record of 67.9 percent in Mae survey, 42 percent of African-American loans will free up capital to make new CRA 2002, reaching 68.6 percent in the fourth families reported that they were ‘‘very or loans. As a result, the CRA market segment 158 quarter of 2003. This section discusses the fairly likely’’ to buy a home in the next three provides an opportunity for Fannie Mae and potential for further increases beyond those years, up from 38 percent in 1998 and 25 Freddie Mac to expand their affordable resulting from current demographic trends. lending programs. Section E.3c below percent in 1997. Among Hispanics and The potential homeowner population over Hispanic immigrants, the numbers reached presents data showing that purchasing the next decade will be highly diverse, as targeted seasoned loans has been one strategy 37 percent and 34 percent respectively. The growing housing demand from immigrants 2002 survey also reports that more than half that Fannie Mae has chosen to improve its (both those who are already here and those goals performance. Fannie Mae has been of Hispanic renters cite homeownership as projected to come) and non-traditional being ‘‘one of their top priorities’’. In offering CRA programs since mid-1997, when homebuyers will help to offset declines in it launched a pilot program, ‘‘Community addition, nearly a third (31 percent) of baby the demand for housing caused by the aging boomers said they are ‘‘very or fairly likely’’ Reinvestment Act Portfolio Initiative,’’ for of the population. As noted in the above purchasing seasoned CRA loans in bulk to buy a home in the next three years. discussion of CRA, many of these potential In spite of these trends, potential minority transactions, taking into account track record homeowners will be located in urban areas. as opposed to relying just on underwriting and immigrant homebuyers see more As noted in the above discussion of obstacles to buying a home, compared with guidelines. Fannie Mae also started another underlying demographic conditions (section pilot program in 1998, involving purchases of the general public. These barriers to C.2.), immigrants and other minorities—who homeownership are discussed in detail in CRA loans on a flow basis, as they are accounted for nearly 40 percent of the growth originated. As part of the American Dream section B.1.b above and include: lack of in the nation’s homeownership rate over the capital for down payment and closing costs; Commitment, Fannie Mae has committed to past five years—will be responsible for investing $20 billion in CRA-targeted poor credit history; lack of access to almost two-thirds of the growth in the mainstream lenders; complexity and fear of business, and funding $530 billion in CRA- number of new households over the next ten eligible investments. One CRA-eligible the homebuying process; and, continued years. This trend does not depend on the discrimination in housing markets and product in 2003 included the future inflow of new immigrants, as MyCommunityMortgage TM suite, which mortgage lending. To address the needs of immigrants don’t enter the housing market provides flexible product options for low- to the new group of potential homeowners, the moderate-income families, including mortgage industry will have to address these 155 minorities, immigrants, first-time Fannie Mae, ‘‘Fannie Mae’s Comments on needs on several fronts, such as expanding HUD’s Proposed Housing Goals for Fannie Mae and homebuyers, and underserved borrowers education and outreach efforts, introducing Freddie Mac for the years 2005–2008 and new products, and adjusting current living in rural areas. MyCommunityMortgage Amendments to HUD’s Regulation of Fannie Mae and Freddie Mac,’’ July 16, 2004, p. I–59. underwriting standards to better reflect the 152 Board of Governors of the Federal Reserve 156 Fannie Mae, 2003 Annual Housing Activities System. The Performance and Profitability of CRA- Report, 2004, pp. 8–9. 159 Ibid. Related Lending. Washington, DC, 2000. 157 Freddie Mac, Opening Doors for America’s 160 Fannie Mae, Fannie Mae National Housing 153 This discussion of urban lending draws from Families: Freddie Mac’s Annual Housing Activities Survey, 2002, p. 6. Jeff Siegel, ‘‘Urban Lending Helps Increase Volume Report for 2003, March 15, 2004, p. 64. 161 Ibid. p. 8. and Meet CRA Requirements,’’ Secondary 158 U.S. Department of HUD, Office of Policy 162 Joint Center for Housing Studies of Harvard Marketing Executive, February 2003, pp. 21–23. Development and Research, U.S. Housing Market University, State of the Nation’s Housing 2003, p. 154 Ibid. Conditions, May 2004, p. 81. 15.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00087 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63666 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
special circumstances of these new of an applicant’s creditworthiness or credit that Loan Prospector reduces the average households. history. Mortgage scorecards are statistically time lenders spend underwriting most loans Thus, the new group of potential estimated regression-type equations, based and reduces origination costs by about an homeowners will have unique needs. To tap on historical relationships between mortgage average of $650 or more per loan.168 In this potential homeowner population, the foreclosures (or defaults) and the addition, Freddie Mac analyzed about 1,000 mortgage industry will have to address these underwriting variables. The level of down loans originated in 1993 and 1994. Of the needs on several fronts, such as expanding payment and credit history indicators, such loans, manual underwriters rated 52 percent education and outreach efforts, introducing as a FICO score, are typically the most accept, compared to a Loan Prospector accept new products, and adjusting current important predictors of default in mortgage rate of 87 percent.169 In total, Freddie Mac underwriting standards to better reflect the scoring systems. reports that innovations in the originations special circumstances of these new For example, HUD has developed FHA process, including automated underwriting, households. TOTAL Scorecard to evaluate the credit risk have reduced mortgage transaction costs by The Bush administration has outlined a of FHA loans submitted to an automated more than 70 percent between 1990 and 2003 plan to expand minority homeownership by underwriting system. The Scorecard works from 1.87 points to 0.46 points—a decline of 5.5 million families by the end of the decade. with Fannie Mae’s Desktop Underwriter to $1,410 per $100,000 borrowed.170 The Joint Center for Housing Studies has provide a recommended level of As explained above, automated mortgage stated that if favorable economic and housing underwriting and documentation for FHA scoring allows tradeoffs between risk factors market trends continue, and if additional loans and to determine a loan’s eligibility for to be quantified more precisely, providing efforts to target mortgage lending to low- insurance with FHA. In 2003, Fannie Mae the industry more confidence in ‘‘pushing income and minority households are made, conducted a market test of the Scorecard the envelope’’ of acceptable expected default the overall homeownership rate could reach with 18 FHA approved Desktop rates. The GSEs’ willingness to offer low- 70 percent by 2010.163 Underwriter lenders. Over 3,000 loans were down-payment programs was based on their 7. Automated Underwriting Systems and submitted to the Total Scorecard through belief that their scoring models could Mortgage Scorecards Desktop Underwriter during the market test identify the more creditworthy of the cash- period.166 constrained applicants. The GSEs’ new This, and the following two sections, This increased accuracy in risk assessment ‘‘timely reward’’ products for subprime discuss special topics that have impacted the of mortgage scorecards has allowed risk borrowers (discussed later) are integrated primary and secondary mortgage markets in managers to set more lenient risk standards, with their mortgage scoring systems. recent years. They are automated mortgage and thus originate more loans to marginal Automated mortgage scoring presents the scoring, subprime loans, and risk-based applicants. Applicants who would otherwise opportunity to remove discrimination from pricing. The GSEs’ use of automated be rejected by manual underwriting are being mortgage underwriting, to accept all underwriting and mortgage scoring systems applicants, and to bring fair, objective, was briefly discussed in the earlier section on qualified for mortgages with automated mortgage scoring in part because the statistically based competitive pricing, underwriting standards. This section greatly reducing costs for all risk groups. expands on issues related to automated scorecard allows an applicant’s weaker areas to be offset by stronger characteristics. Some institutions have sought to better underwriting, a process that has spread model and automate marginal and higher-risk throughout the mortgage landscape over the Typically, applicants whose projected monthly debt payment (mortgage payment loans, which have tended to be more costly past five years, due mainly to the efforts of to underwrite and more difficult to Fannie Mae and Freddie Mac. plus credit card payment plus automobile loan payment and so on) comprise a high automate.171 Automated mortgage scoring was Along with the promise of benefits, developed as a high-tech tool with the percentage of their monthly income would be turned down by a traditional underwriting however, automated mortgage scoring has purpose of identifying credit risks in a more raised concerns. These concerns are related efficient manner. Automated mortgage system that relied on fixed debt-to-income ratios (such as 36 percent). In a mortgage to the possibility of disparate impact and the scoring has grown as competition and proprietary nature of the mortgage score decreased profit margins have created scoring system, these same applicants might be automatically accepted for a loan due to inputs. The first concern is that low-income demands to reduce loan origination costs. As and minority homebuyers will not score well a result, automated mortgage scoring has their stellar credit record or to their ability to raise more cash for a down payment. The enough to be accepted by the automated become the predominant (around 60 to 70 underwriting system, resulting in their 164 entity funding or insuring the mortgage (i.e., percent) mortgage underwriting method. getting fewer loans. African-American and According to Freddie Mac economists, a lender, private mortgage insurer, or a GSE) allows these positive characteristics to offset Hispanic borrowers, for example, tend to automated mortgage scoring has enabled have a poorer credit history record than other lenders to expand homeownership the negative characteristics because its borrowers, which means they are more likely opportunities, particularly for underserved confidence in the ability of the empirically- to be referred (rather than automatically populations.165 There is growing evidence based mortgage scorecard to accurately accepted) by automated mortgage scoring that automated mortgage scoring is more identify those applicants who are more likely systems that rely heavily on credit history accurate than manual underwriting in or less likely to eventually default on their measures such as a FICO score. There is also predicting borrower risks. Mortgage loan. The mortgage score is in essence a a significant statistical relationship between scorecards express the probability that an recommendation to the lender to accept the credit history scores and the minority applicant will default as a function of several application, or to refer it for further review composition of an area, after controlling for underwriting variables such as the level of through manual underwriting. Accepted other locational characteristics.172 down payment, monthly-payment-to-income loans benefit from reduced document ratios, cash reserves, and various indicators requirements and expedited processing. The second concern relates to the ‘‘black In 2003, Fannie Mae conducted a study of box’’ nature of the scoring algorithm. The scoring algorithm is proprietary and therefore 163 Joint Center for Housing Studies of Harvard automated underwriting systems and University, State of the Nation’s Housing 1998, p. concluded that the production cost per loan 20. decreased significantly as lenders moved 168 Freddie Mac, Opening Doors for America’s 164 John W. Straka, ‘‘A Shift in the Mortgage automated underwriting closer to the point of Families: Freddie Mac’s Annual Housing Activities Landscape: The 1990s Move to Automated Credit sale. Specifically, retail lenders using an Report for 2003, March 15, 2004, p. 55. 169 Evaluations,’’ Journal of Housing Research, 2000, integrated automated underwriting system at Freddie Mac, Opening Doors for America’s Families: Freddie Mac’s Annual Housing Activities (11)2: p. 207. the point of sale reported originations savings 165 Peter M. Zorn, Susan Gates, and Vanessa Report for 2003, March 15, 2004, p. 54. of more than $1,000 over manual 170 Freddie Mac Public Comment Letter on HUD’s Perry, ‘‘Automated Underwriting and Lending underwriting.167 Freddie Mac also reported Outcomes: The Effect of Improved Mortgage Risk Proposed Goals, July 2004, p. 5. Assessment on Under-Served Populations. Program 171 Ibid. pp. 208–217. on Housing and Urban Policy,’’ Conference Paper 166 Fannie Mae, 2003 Annual Housing Activities 172 Robert B. Avery, Raphael W. Bostic, Paul S. Series, Fisher Center for Real Estate and Urban Report, 2004, p. 12. Calem, and Glenn B. Canner, Credit Scoring: Issues Economics. University of California Berkeley, 2001, 167 Fannie Mae, 2003 Annual Housing Activities and Evidence from Credit Bureau Files, mimeo, p. 5. Report, 2004, p. 36. 1998, p. 24.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00088 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63667
it is difficult for applicants to know the disadvantages for less educated and lower- by manual underwriting to qualify for reasons for their scores. However, it should income consumers. In addition to the digital mortgages. be noted that the GSEs have taken steps to divide, the lack of financial literacy in the Analysis of the first group of loans showed make their automated underwriting systems United States may also result in a disparate that loans rated as ‘‘caution’’ were four times more transparent. Both Fannie Mae and impact on low-income and minority more likely to default than the average for all Freddie Mac have published the factors used borrowers.176 loans. Minority borrowers whose loans were to make loan purchase decisions in Desktop 2002 Urban Institute Study. The Urban rated as ‘‘caution’’ were five times more Underwriter and Loan Prospector, Institute submitted a report to HUD in 2002 likely to default, and low-income borrowers respectively. In response to criticisms aimed on subprime markets, the role of GSEs, and whose loans were rated as ‘‘caution’’ were at using FICO scores in mortgage risk-based pricing.177 The study took a four times more likely to default than the underwriting, Fannie Mae’s new versions of preliminary look at the use of automated average for all loans. The 2000 version of LP Desktop Underwriter (DU) 5.3 and 5.3.1 [the underwriting systems for a small sample of approved 87.1 percent of loans generated newest versions are 5.3 and 5.3.1—they lenders. After conducting interviews with through affordable housing programs, probably keep the following practices, but both subprime and prime lenders, the report compared to a 51.6 percent approval rate add no substantive underwriting practices, noted that all of the lenders in the study had when the same loans were assessed using but rather lower downpayment options] implemented some type of automated manual underwriting procedures. Further, replaces credit scores with specific credit underwriting system. These lenders stated the study found LP more accurate than characteristics and provides expanded that automated underwriting raised their approval product offerings for borrowers who manual underwriting at predicting default have blemished credit. The specific credit business volume and streamlined their risk even with a higher approval rate. The characteristics include variables such as past approval process. In addition, the lenders study also demonstrated that Freddie Mac’s delinquencies; credit records, foreclosures, reported they were able to direct more year 2000 version of LP was more accurate and accounts in collection; credit card line underwriting resources to borderline in predicting risk than its 1995 version. and use; age of accounts; and number of applications despite an increase in business Concluding Observations. Automated credit inquiries.173 volume. underwriting has enabled lenders to reach With automated mortgage scoring replacing Even with the use of automated mortgage new markets and expand homeownership traditional manual underwriting comes the scoring, the lenders in the study continued opportunities, as illustrated by the 2001 fear that the loss of individual attention to conduct at least a cursory review to Freddie Mac study. Increased accuracy with poses a problem for people who have validate the application material. The automated mortgage scoring has led to the inaccuracies on their credit report or for majority of the lenders still used manual development of new mortgage products that members of cultural groups or recent underwriting to originate loans not would have been previously considered too immigrants who do not use traditional credit recommended for approval with automated risky. For example, Freddie Mac uses Loan and do not have a credit score. Some mortgage scoring. The lenders reported they Prospector to approve Alt A loans, which subprime lenders and underwriters have formulated their policies and procedures to tend to have nontraditional documentation; claimed that their manual underwriting of make certain that borrowers receive the best A-minus loans, which pose a higher risk of high-risk borrowers cannot be automated mortgage, according to product eligibility. default; and other higher-risk mortgages, like with mortgage scoring. Although automated This study will be further referenced in a 100 percent LTV loans. Both GSEs have and mortgage scoring has greatly reduced the cost following section regarding subprime continue to add new products to develop of many lower-risk loans that are easier to markets. their automated underwriting systems to rate, the cost of manually underwriting gray- 2001 Freddie Mac Study. According to a reach more marginal borrowers. area and higher-risk applicants still remains Freddie Mac study published by the Fisher Despite the gains in automated mortgage high.174 There is also the fear that applicants Center for Real Estate and Urban Economics scoring and other innovations, minorities are who are referred by the automated system at University of California at Berkeley, still less likely to be approved for a loan. The will not be given the full manual underserved populations have benefited from difference in minority and non-minority underwriting for the product that they automated mortgage scoring because of the accept rates may reflect greater social initially applied for—rather they might be increased ability to distinguish between a inequities in financial capacity and credit, pushed off to higher priced products such as range of credit risks. In this paper, Freddie which are integral variables in both manual a subprime or FHA loan. In this case, the Mac economists compared the manual and and automated underwriting. In the future, applicant may have had special automated mortgage scoring approval rates of the accuracy of automated mortgage scoring circumstances that would have been clarified a sample of minority loans originated in will hinge on updating the models and by the traditional manual underwriting, thus 1993–94 and purchased by Freddie Mac. making them more predictive while reducing enabling the applicant to receive a prime While manual underwriters rated 51 percent the disparate impact on low-income and loan consistent with his or her of the minority loans in the sample as accept, minority borrowers.180 The fairness of creditworthiness. automated mortgage scoring would have automated scoring systems will also depend Banking regulators and legal analysts 178 rated 79 percent of the loans as accept. importantly on whether referred applicants acknowledge the value of automated In comparison to manual underwriting, mortgage scoring, although some skeptics receive a traditional manual underwriting for this study found automated mortgage scoring the loan that they initially applied for, rather have noted concerns regarding fair lending, not only less discriminatory but also more than being immediately offered a higher potential fraud, privacy issues, and the accurate in predicting risk. Two versions of priced loan that does not recognize their true ability of models to withstand changing Freddie Mac’s automated underwriting creditworthiness. economic conditions.175 With the rise of system, Loan Prospector (LP), were used to In addition to using automated automated mortgage scoring, the great review three groups of mortgage loans underwriting systems as a tool to help difference in Internet usage known as the purchased by Freddie Mac.179 The study determine whether a mortgage application ‘‘digital divide’’ could result in informational found that LP was a highly accurate predictor should be approved, the GSEs’ automated of mortgage default. The resulting improved underwriting systems are being further 173 Fannie Mae, September 4, 2002, p. 33. accuracy translates into benefits for adapted to facilitate risk-based pricing. With 174 Kenneth Temkin, Jennifer E.H. Johnson, and borrowers, who would otherwise be rejected Diane Levy, Subprime Markets, The Role of GSEs, risk-based pricing, mortgage lenders can offer and Risk-Based Pricing, Washington: The Urban each borrower an individual rate based on Institute. Report Prepared for the U.S. Department 176 Zorn et al., 2001, pp. 19–20. his or her risk. The division between the of Housing and Urban Development, 2002. 177 Kenneth Temkin, Jennifer E.H. Johnson, and subprime and the prime mortgage market 175 Allen J. Fishbein, ‘‘Is Credit Scoring a Winner Diane Levy, Subprime Markets, The Role of GSEs, will begin to fade with the rise of risk-based for Everyone?’’ Stone Soup, 2000, 14(3): pp. 14–15. and Risk-Based Pricing, Washington: The Urban pricing, which is discussed in the next Institute. Report Prepared for the U.S. Department See also Fitch IBCA, Inc., Residential Mortgage section on the subprime market. Credit Scoring, New York, 1995 and Jim Kunkel, of Housing and Urban Development, 2002. ‘‘The Risk of Mortgage Automation,’’ in Mortgage 178 Zorn, et al., 2001, pp. 14–15. Banking, 1995, 57(8): pp. 69–76. 179 Ibid. p. 5. 180 Ibid. pp. 18–19.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00089 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63668 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
8. Subprime Lending more than three out of four loans in the income and minority neighborhoods have The subprime mortgage market provides subprime market. raised concerns about whether mainstream mortgage financing to credit-impaired The subprime market is divided into lenders are adequately serving these borrowers—those who may have blemishes different risk categories, ranging from least neighborhoods. A study of subprime lending in their credit record, insufficient credit risky to most risky: A-minus, B, C, and D. in Chicago by The Woodstock Institute history, or non-traditional credit sources. While there are no clear industry standards concluded that a dual, hyper-segmented This section examines several topics related for defining the subprime risk categories, mortgage market existed in Chicago, as to subprime lending including (a) the growth Inside Mortgage Finance defines them in mainstream lenders active in white and and characteristics of subprime loans, (b) the terms of FICO scores—580–620 for A-minus, upper-income neighborhoods were much less neighborhood concentration of subprime 560–580 for B, 540–560 for C, and less than active in low-income and minority lending, (c) predatory lending, and (d) 540 for D. The A-minus share of the neighborhoods–effectively leaving these purchases of subprime mortgages by the subprime market rose from 61.6 percent in neighborhoods to unregulated subprime 185 GSEs. Section C.9 follows with a discussion 2000 to 70.7 percent in 2001. For the first lenders.189 As part of the HUD-Treasury Task of risk-based pricing. nine months of 2002, the A-minus share Force on Predatory Lending, HUD’s Office of accounted for 74 percent of the market, while Policy Development and Research released a a. The Growth and Characteristics of the B share accounted for 11 percent, the C national level study—titled Unequal Burden: Subprime Loans share accounted for 7.2 percent, and the D Income and Racial Disparities in Subprime The subprime market has grown rapidly share accounted for 7.9 percent of the Lending in America—that showed families over the past several years, increasing from market.186 living in low-income and African-American an estimated $35 billion in 1994 to $160 Delinquency rates by type of subprime loan neighborhoods in 1998 relied billion in 1999 and $173.3 billion in 2001, are as follows: 3.36 percent for A-minus disproportionately on subprime refinance before rising to $213 billion in 2002. The loans, 6.67 percent for B, 9.22 percent for C, lending, even after controlling for subprime share of total market originations and 21.03 percent for D, according to the neighborhood income. An update of that rose from 4.6 percent in 1994 to a high of 15 Mortgage Information Corporation.187 analysis for the year 2000 yields the percent in 1999, and then fell to 8.5 percent Because of their higher risk of default, following trends: 190 in both 2001 and 2002.181 Various factors subprime loans typically carry much higher • In 2000, 36 percent of refinance have led to the rapid growth in the subprime mortgage rates than prime mortgages. Recent mortgages in low-income neighborhoods market: Federal legislation preempting state quotes for a 30-year Fixed Rate Mortgage were subprime, compared with only 16 restrictions on allowable rates and loan were 8.85 percent for A-minus (with an 85 percent in upper-income neighborhoods. features, the tax reform act of 1986 which percent LTV), 9.10 percent for B credit (with • Subprime lending accounted for 50 encouraged tax-exempt home equity an 80 percent LTV), and 10.35 percent for C percent of refinance loans in majority African 188 financing of consumer debt, increased credit (with a 75 percent LTV). As the low American neighborhoods—compared with demand for and availability of consumer loan-to-value (LTV) ratios indicate, one loss only 21 percent in predominantly white areas debt, a substantial increase in homeowner mitigation technique used by subprime (less than 30 percent of population is African equity due to house price appreciation, and lenders is a high down payment requirement. American). a ready supply of available funds through Some housing advocates have expressed • The most dramatic view of the disparity Wall Street securitization.182 It is important concern that the perceptions about the risk of in subprime lending comes from comparing to note that subprime lending grew in the subprime loans may not always be accurate, homeowners in upper-income African- 1990s mostly without the assistance of for example, creditworthy borrowers in inner American and white neighborhoods. Among Fannie Mae and Freddie Mac. city neighborhoods may be forced to use homeowners living in the upper-income subprime lenders because mainstream Generally, there are three different types of white neighborhoods, only 16 percent turned lenders are not doing business in their products available for subprime borrowers. to subprime lenders in 2000. But 42 percent neighborhoods (see below). These include: Home purchase and refinance of homeowners living in upper-income Subprime borrowers are much more likely mortgages designed for borrowers with poor African-American neighborhoods relied upon to be low income and be a minority than credit histories; ‘‘Alt A’’ mortgages that are subprime refinancing which is substantially other borrowers. Between 1999 and 2001, usually originated for borrowers who are more than the rate (30 percent) for 43.1 percent of subprime loans in the unable to document all of the underwriting homeowners living in low-income white conventional conforming market went to information but who may have solid credit neighborhoods. low-income borrowers, compared with 29.5 records; and high loan-to-value mortgages • Similar results are obtained when the percent of conventional conforming loans. originated to borrowers with fairly good analysis is conducted for borrowers instead During that same period, 19.9 percent of credit. Fannie Mae and Freddie Mac are more of neighborhoods. Upper-income African- subprime loans were for African-American likely to serve the first two types of subprime American borrowers are twice as likely as 183 borrowers, compared with 6.5 percent of all borrowers. low-income white borrowers to have conventional conforming loans. However, Borrowers use subprime loans for various subprime loans. Over one-half (54 percent) of purposes, which include debt consolidation, what distinguishes subprime loans from other loans is their concentration in African- home improvements, and an alternative 189 American neighborhoods. Daniel Immergluck, The Predatory Lending source of consumer credit. Between 1999 and Crisis in Chicago: The Dual Mortgage Market and 2001, about two-thirds of subprime loans b. The Neighborhood Concentration of Local Policy, testimony before the Chicago City were refinance loans. It has been estimated Subprime Lending Council, April 5, 2000. Immergluck found that subprime lenders received 74 percent of refinance that 59 percent of refinance loans were ‘‘cash The growth in subprime lending over the 184 applications in predominantly black tracts out’’ loans. According to a joint HUD- last several years has benefited credit- Treasury report, first liens accounted for compared to 21 percent in predominantly white impaired borrowers as well as those tracts in 1998. According to Immergluck, these borrowers who choose to provide little racial disparities provide evidence that the 181 Subprime origination data are from Inside documentation for underwriting. However, residential finance market in Chicago is Mortgage Finance. For the 2002 estimates, see studies showing that subprime lending is hypersegmented, resulting in the increased ‘‘Subprime Origination Market Shows Strong disproportionately concentrated in low- likelihood that minorities receive mortgage credit Growth in 2002,’’ Inside B&C Lending, published by from a subprime, rather than a prime, lender in Inside Mortgage Finance, February 3, 2003, page 1. Chicago. Also see Daniel Immergluck, Stark 182 Temkin et. al., 2002, p. 1. 185 ‘‘Wholesale Dominates Subprime Market Differences: The Explosion of the Subprime 183 Kenneth Temkin, Jennifer E.H. Johnson, Diane Through 3rd Quarter ’02,’’ Inside B&C Lending, Industry and Racial Hypersegmentation in Home Levy, Subprime Markets, The Role of GSEs, and published by Inside Mortgage Finance, December Equity Lending, Woodstock Institute, October 2000 Risk Based Pricing, Washington: The Urban 16, 2002, pp. 1–2. 190 See Randall M. Scheessele, Black and White Institute. Report Prepared for the Department of 186 Inside B&C Lending, November 16, 2002, p.2. Disparities in Subprime Mortgage Refinance Housing and Urban Development, 2002, p. 4. 187 Mortgage Information Corporation, The Lending, Housing Finance Working Paper HF–014, 184 U.S. Department of Housing and Urban Market Pulse, Winter 2001, pp. 4–6. Office of Policy Development and Research, U.S. Development/U.S. Department of the Treasury, 188 Inside B&C Lending, published by Inside Department of Housing and Urban Development, Curbing Predatory Lending Report, 2000, p.31. Mortgage Finance, February 17, 2003, page 13. April 2002.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00090 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63669
low-income African-American borrowers concluded that the subprime loans had an The term ‘‘predatory lending’’ is a short turn to subprime lenders, as does over one- unexplained interest rate premium of 100 hand term that is used to encompass a wide third (35 percent) of upper-income African- basis points on average.193 range of abuses. While there is broad public American borrowers. By comparison, only 24 Banking regulators have recognized the agreement that predatory lending should percent of low-income white borrowers and link between the growth in subprime lending have no place in the mortgage market, there 12 percent of upper-income white borrowers, and the absence of mainstream lenders and are differing views about the magnitude of rely upon subprime lenders for their have urged banks and thrifts that lending in the problem, or even how to define practices refinance loans.191 these neighborhoods not only demonstrates that make a loan predatory. The joint HUD- It does not seem likely that these high responsible corporate citizenship but also Treasury report, Curbing Predatory Home market shares by subprime lenders in low- profitable lending. Ellen Seidman, former Mortgage Lending, concluded that a loan can income and African-American Director of the Office of Thrift Supervision, be predatory when lenders or brokers: charge neighborhoods can be justified by a heavier stated that, ‘‘Many of those served by the borrowers excessive, often hidden fees concentration of households with poor credit subprime market are creditworthy borrowers (called ‘‘packing fees’’); successively in these neighborhoods. Rather, it appears who are simply stuck with subprime loans or refinance loans at no benefit to the borrower that subprime lenders may have attained subprime lenders because they live in (called ‘‘loan flipping’’); make loans without such high market shares by serving areas neighborhoods that have too few credit or regard to a borrower’s ability to repay; and, where prime lenders do not have a banking opportunities.’’ engage in high-pressure sales tactics or significant presence. The above finding that With respect to the question of whether outright fraud and deception. These practices upper-income black borrowers rely more borrowers in the subprime market are are often combined with loan terms that, heavily on the subprime market than low- sufficiently creditworthy to qualify for more alone or in combination, are abusive or make traditional loans, Freddie Mac has said that income white borrowers suggests that a the borrower more vulnerable to abusive one of the promises of automated portion of subprime lending is occurring practices. Vulnerable populations, including underwriting is that it might be better able to with borrowers whose credit would qualify the elderly and low-income individuals, and identify borrowers who are unnecessarily them for lower cost conventional prime low-income or minority neighborhoods, assigned to the high-cost subprime market. loans. A lack of competition from prime appeared to be especially targeted by Freddie Mac has estimated that 10–30 unscrupulous lenders. lenders in low-income and minority percent of borrowers who obtain mortgages neighborhoods has increased the chances One consequence of predatory lending is in the subprime market could qualify for a that borrowers are stripped of the equity in that borrowers in these communities are conventional prime loan through Loan their homes, which places them at an paying a high cost for credit. As explained Prospector, Freddie Mac’s automated increased risk of foreclosure. In fact, high next, there is also evidence that the higher underwriting system.194 Fannie Mae has foreclosure rates for subprime loans provide interest rates charged by subprime lenders stated that half of all mortgage borrowers the most concrete evidence that many cannot be fully explained solely as a function steered to the high-cost subprime market are subprime borrowers are entering into of the additional risks they bear. Thus, a in the A-minus category, and therefore are mortgage loans that they simply cannot greater presence by mainstream lenders 195 prime candidates for Fannie Mae. afford. The high rate of foreclosures in the could possibly reduce the high up-front fees c. Predatory Lending subprime market has been documented by and interest rates being paid by residents of HUD and others in recent research studies.196 low-income and minority neighborhoods. Predatory lending has been a disturbing part of the growth in the subprime market. These studies have found that foreclosures by The Freddie Mac study presented evidence subprime lenders grew rapidly during the that subprime loans bear interest rates that Although questions remain about its magnitude, predatory lending has turned 1990s and now exceed the subprime lenders’ are higher than necessary to offset the higher share of originations. In addition, the studies 192 homeownership into a nightmare for far too credit risks of these loans. The study indicate that foreclosures of subprime loans compared (a) the interest rate on subprime many households. The growing incidence of abusive practices has been stripping occur much more quickly than foreclosures loans rated A-minus by the lenders on prime loans, and that they are originating these loans with (b) the interest borrowers of their home equity, threatening families with foreclosure, and destabilizing concentrated in low-income and African- rates on prime loans purchased by Freddie American neighborhoods. Of course, given Mac and rated A-minus by a Freddie Mac neighborhoods. Also, in some cities, there are indications that unscrupulous realtors, the riskier nature of these loans, a higher underwriting model. Despite the fact that foreclosure rate would be expected. With the both loan groups were rated A-minus, on mortgage brokers, appraisers, and lenders are duping some FHA borrowers into purchasing information available it is not possible to average the subprime loans bore interest rates evaluate whether the disparities in that were 215 basis points higher. Even homes at an inflated price or with significant undisclosed repairs. The problems associated foreclosure rates are within the range of what assuming that the credit risk of the subprime with home equity fraud and other mortgage would be expected for loans prudently loans was in fact higher than the prime loans, abuses are not new ones, but the extent of originated within this risk class. But findings the study could not account for such a large this activity seems to be increasing. The from these studies about the high rate of discrepancy in interest rates. Assuming that expansion of predatory lending practices mortgage foreclosure associated with default rates might be three to four times along with subprime lending is especially subprime lending reinforce the concern that higher for the subprime loans would account troubling since subprime lending is predatory lending can potentially have for a 90 basis point interest rate differential. disproportionately concentrated in low- and devastating effects for individual families Assuming that servicing the subprime loans very-low income neighborhoods, and in and their neighborhoods. would be more costly would justify an African-American neighborhoods. At this time, there are open questions additional 25 basis point differential. But about the effectiveness of the different even after allowing for these possible 193 It should also be noted that higher interest approaches being proposed for eradicating differences, the Freddie Mac researchers rates are only one component of the higher cost of subprime loans since borrowers also often face 196 For an overview of these studies, see Harold 191 For an update to 2001, see The Association of higher origination points. The Freddie Mac study L. Bunce, Debbie Gruenstein, Christopher E. Community Organizers for Reform Now (ACORN), did not find a large differential between prime and Herbert, Randall M. Scheessele, Subprime Separate and Unequal Predatory Lending in subprime loans in points paid, but the study notes Foreclosures: The Smoking Gun of Predatory America, 2002. In 2001, subprime lenders that subprime loans often have points rolled into Lending, 2000. Also see Abt Associates Inc., originated 27.8 percent of all conventional the loan principal, which cannot be identified with Analyzing Trends in Subprime Originations and refinance loans for African-Americans, 13.6 percent their data. Foreclosures: A Case Study of the Atlanta Metro for Hispanic homeowners, and just 6.3 percent for 194 Freddie Mac, We Open Doors for America’s Area, February 2000 and Analyzing Trends in white homeowners. Overall, African-Americans Families, Freddie Mac’s Annual Housing Activities Subprime Originations and Foreclosures: A Case were 4.4 times more likely to use a subprime lender Report for 1997, March 16, 1998, p. 23. Study of the Boston Metro Area, September 2000; than whites, and Hispanics were 2.2 times more 195 Rommy Fernandez, ‘‘Fannie Mae Eyes Half of National Training and Information Center, Preying likely to do so. the Subprime Market,’’ in The American Banker, on Neighborhoods: Subprime Mortgage Lenders and 192 Howard Lax, Michael Manti, Paul Raca, and March 1, 2002. Also see ‘‘Fannie Mae Vows More Chicagoland Foreclosures, 2000; and the HUD Peter Zorn, ‘‘Subprime Lending: An Investigation of Minority Lending,’’ Washington Post, March 16, study, Unequal Burden in Baltimore: Income and Economic Efficiency,’’ February 25, 2000. 2000, p. EO1. Racial Disparities in Subprime Lending, May 2000.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00091 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63670 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
predatory lending and the appropriate roles classes of senior/subordinated securities. The In recent years, Freddie Mac has instituted of different governmental agencies—more two GSEs also purchase subprime loans on measures designed to protect consumers from legislation versus increased enforcement of a flow basis. Fannie Mae began purchasing predatory lending. For example, Freddie Mac existing laws, long-run financial education subprime loans through its Timely Payment has announced that, effective August 1, 2004, versus mortgage counseling, Federal versus Reward Mortgage program in June 1999, and they will no longer invest in subprime state and local actions. In its recent issuance Freddie Mac rolled out a similar product, mortgages originated after that date that of predatory lending standards for national Affordable Merit Rate, in May 2000 contain mandatory arbitration clauses. Since banks, the Office of the Comptroller of the (described below). In addition to purchasing 2000, Freddie Mac has prohibited purchases Currency (OCC) cited the efforts of Fannie subprime loans for borrowers with blemished of mortgages that impose a prepayment Mae and Freddie Mac in reducing predatory credit, the GSEs also purchase another non- premium for a term of more than five years, lending.197 The OCC advised banks against conforming loan called an Alternative-A or and in March 2002, this prohibition was abusive practices, such as rolling single- ‘‘Alt-A’’ mortgage. These mortgages are made reduced to no more than three years. Freddie premium life insurance into a loan. The to prime borrowers who do not want to Mac does not purchase high-rate or high-fee agency cited guidelines developed by Fannie provide full documentation for loans. The loans that are covered by the Home Mae and Freddie Mac as a ‘‘useful reference’’ GSEs’ interest in the subprime market has Ownership and Equity Protection Act of 1994 or starting point for national banks. coincided with a maturation of their (HOEPA); and they do not purchase Following publication of HUD’s proposed traditional market (the conforming mortgages containing a prepaid single- 2000 Rule inviting comments on disallowing conventional mortgage market), and their premium credit life, credit disability, credit goals credit for high cost mortgage loans, development of mortgage scoring systems, unemployment or credit property insurance Fannie Mae and Freddie Mac told lenders which they believe allows them to accurately policy. Freddie Mac also requires all lenders they would no longer purchase loans with model credit risk. Although the GSEs account servicing their loans to report monthly certain abusive practices, such as excessive for only a modest share of the subprime borrower mortgage payments to all four major fees and failing to consider a borrower’s market today, some market analysts estimate credit repositories, and conducts onsite ability to repay the debt. that they could purchase as much as half of reviews of their customers and holds them It is important to re-emphasize that the overall subprime market in the next few accountable if their business practices do not predatory lending generally occurs in years.198 meet Freddie Mac standards.202 neighborhoods where borrowers have limited Precise information on the GSEs’ purchases Fannie Mae initiated its Timely Payments access to mainstream lenders. While of subprime loans is not readily available. product in September 1999, under which predatory lending can occur in the prime Data can be pieced together from various borrowers with slightly damaged credit can market, it is ordinarily deterred in that sources, but this can be a confusing exercise qualify for a mortgage with a higher interest market by competition among lenders, because of the different types of non- rate than prime borrowers. Under this greater homogeneity in loan terms and conforming loans (Alt-A and subprime) and product, a borrower’s interest rate will be greater financial information among the different channels through which the reduced by 100 basis points if the borrower borrowers. Thus, one solution to address this GSEs purchase these loans (through makes 24 consecutive monthly payments problem would be to encourage more securitizations and through their ‘‘flow- without a delinquency. Fannie Mae has mainstream lenders to do business in our based’’ product offerings). Freddie Mac, revamped its automated underwriting system inner city neighborhoods. which has been the more aggressive GSE in (Desktop Underwriter) so loans that were Certain commenters urged the Department the subprime market, purchased traditionally referred for manual to adopt predatory lending safeguards in the approximately $12 billion in subprime loans underwriting are now given four risk final rule that would prohibit the GSEs from during 1999—$7 billion of A-minus and classifications, three of which identify counting loans that included mandatory alternative-A loans through its standard flow potential subprime (A-minus) loans.203 arbitration clauses or loans with prepayment programs and $5 billion through structured Fannie purchased about $600 million of penalties beyond three years towards the transactions.199 In 2000, Freddie Mac subprime loans on a flow basis in 2000.204 goals. In the 2000 rulemaking, the purchased $18.6 billion of subprime loans on Fannie Mae securitized around $0.6 billion Department determined that the GSEs should a flow basis in addition to another $7.7 in subprime mortgages in 2000, before not receive goals credit for purchasing high billion of subprime loans through structured increasing to $5.0 billion in 2001 and 7.3 cost mortgages including mortgages with transactions.200 Freddie Mac securitized $9 billion in 2002.205 In terms of total subprime unacceptable features as explained in the billion in subprime and Alt-A product in activity (both flow and securitization preamble. The Department is aware that 2001 and $11.1 billion in 2002. activities), Fannie Mae purchased $9.2 certain practices that were not enumerated in Fannie Mae’s anti-predatory lending billion in 2001 and over $15 billion in 2002, the regulations adopted in 2000, such as strategy includes eight major components. the latter figure representing about 10 percent loans with prepayment penalties after three These components include: establishing of the market, according to Fannie Mae years and loans with mandatory arbitration business guidelines that ensure that liquidity staff.206 clauses, often lock borrowers into is provided for only responsible lenders; A greater GSE role in the subprime lending disadvantageous loan products. The expanding the application of conventional market will most likely have a significant Department will rely on existing regulatory conforming mortgage practices to more impact on the subprime market. Currently, authorities to monitor the GSEs’ performance borrowers; advancing the Mortgage the majority of subprime loans are not in this area. Should the Department later Consumer Bill of Rights Agenda; offering a purchased by GSEs, and the numbers of determine that there is a need to specifically broad range of alternative responsible lenders originating subprime loans typically enumerate additional prohibited predatory products; leveraging technology and the do not issue a large amount of prime loans. practices, it will address such practices in a Internet to expand markets and reduce costs Partly in response to higher affordable future rulemaking. for consumers; working with partners to keep borrowers in their homes; supporting the d. Purchases of Subprime Mortgages by the Amendments to HUD’s Regulation of Fannie Mae GSEs home-buyer education industry to empower and Freddie Mac,’’ July 16, 2004, p. I–59. educators to reach more consumers; and Fannie Mae and Freddie Mac have shown 202 Freddie Mac Public Comment Letter on HUD’s supporting the Fannie Mae Foundation in Proposed Goals, July 2004, p. 6. increasing interest in the subprime market 201 consumer education and outreach. 203 since the latter half of the 1990s. The GSEs See Lederman, et al., Op cit. entered this market by purchasing securities 204 Kenneth Temkin, Jennifer E. H. Johnson, and 198 Temkin et al., 2002, p. 1. backed by non-conforming loans. Freddie Diane K. Levy, ‘‘Subprime Markets, the Role of 199 David A. Andrukonis, ‘‘Entering the Subprime GSEs, and Risk-Based Pricing,’’ Urban Institute, Mac, in particular, increased its subprime Arena,’’ Mortgage Banking, May 2000, pp. 57–60. August 2001, p. 1. business through structured transactions, 200 Subprime Lenders Mixed on Issue of GSE 205 Inside Mortgage Finance’s, ‘‘Inside MBS & with Freddie Mac guaranteeing the senior Mission Creep,’’ Inside B and C Lending, March 19, ABS,’’ December 15, 2000 and March 8, 2002. 2001. 206 Statement by Mercy Jimenez of Fannie Mae in 197 ‘‘OCC Cites Fannie, Freddie Predatory Lending 201 Fannie Mae, ‘‘Fannie Mae’s Comments on ‘‘Fannie Mae: Forges Ahead in Subprime,’’ Rules As Model,’’ Dow Jones Business News, HUD’s Proposed Housing Goals for Fannie Mae and Secondary Marketing Executive, February 2003, February 25, 2003. Freddie Mac for the years 2005–2008 and p.15.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00092 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63671
housing goals set by HUD in its new rule set penalty terms that exceed three years. insurance premiums.213 Rating agencies vary in 2000, the GSEs are increasing their Freddie Mac indicated that the increased subordination requirements based on the business in the subprime market. In the 2000 goals would limit its ability to influence credit qualify of the underlying collateral. GSE Rule, HUD identified subprime subprime lending practices. Many believe there is cross-subsidization borrowers as a market that can assist Fannie Several commenters suggest that if the within the crude risk categories used in Mae and Freddie Mac in reaching their GSEs are pushed to serve more of the today’s market. For example, some of the higher affordable housing goals while also subprime market, they will skim a significant better quality subprime borrowers in the A- helping establish more standardization in the portion of the lower-risk borrowers from that minus category may be inappropriately subprime market. According to an Urban market. The resulting smaller subprime assigned to the subprime market. The GSEs Institute study in 2002, many subprime market would be comprised of the neediest and others are attempting to learn more about lenders believe that successful companies borrowers. Concerned was raised by the subprime market, and their initial efforts serving high-risk borrowers need to have commenters that these higher risk borrowers suggest that there will be an increase in the specialized expertise in outreach, servicing, would pay more based on three factors. First use of risk-based pricing within this market, and underwriting, which is lacking among lower risk borrowers would not be present to although it is recognized that the expansion most prime lenders.207 These lenders do not subsidize them. Second, the market’s high of risk-based pricing depends importantly on believe the more standardized approaches of fixed costs would be distributed across fewer these parties gaining a better understanding prime lenders and the GSEs will work with borrowers. Finally, a significantly smaller of the subprime borrower and the ability of subprime borrowers, who require the more subprime market for private lenders would their mortgage scoring systems to predict risk customized and intensive origination and drive some lenders out of business within this market. It must be noted that the loan servicing processes currently offered by translating into less competition. power of the underlying algorithm in experienced subprime lenders. 9. Risk-Based Pricing automated underwriting systems determines As noted above, both Fannie Mae and The expanded use of automated the ability of these systems to accurately Freddie Mac make the claim that the underwriting and the initial uses of risk- predict risk and set prices. subprime market is inefficient, pointing to based pricing are changing the mortgage If prime lenders adopted risk-based evidence indicating that subprime borrowers lending environment, often blurring the pricing, many would be willing to lend to pay interest rates, points, and fees in excess distinctions between the prime and subprime riskier subprime borrowers because their risk of the increased costs associated with serving market. Prime lenders are now using would now be offset with an increase in riskier borrowers in the subprime market.208 automated underwriting systems that are price. In theory, the mortgage market should A recent Freddie Mac study found automated being adapted to facilitate risk-based pricing. expand because all mortgages will be mortgage scoring less discriminatory and For some time, the majority of prime approved at a price commensurate with risk, more accurate in predicting risk than manual mortgage borrowers have received loan rates rather than setting a risk floor and approving systems such as those currently used by based on average cost pricing. Generally, no one beneath the floor. Risk-based pricing subprime lenders.209 According to Fannie borrowers receive roughly the same Annual could also expand the prime lenders’ market Mae, although a high proportion of borrowers Percentage Rate 211 (APR), regardless of the by enabling them to reach a new group of in the subprime market could qualify for less risk of loss to the lender. The risk of all underserved customers.214 Taking advantage costly prime mortgages, it remains unclear borrowers is averaged together, and the price of GSEs’ lower cost of capital, GSEs may be why these borrowers end up in the subprime is determined by the average risk. able to offer borrowers who could not afford market.210 Fannie Mae and Freddie Mac In contrast, risk-based pricing enables a rate in the subprime market a rate they can believe they can bring more efficiency to the mortgage lenders to offer each borrower an afford resulting from risk-based pricing. subprime market by creating standardized individualized interest rate based on his or Risk-based pricing also poses challenges on underwriting and pricing guidelines in the her risk. Or, more broadly, to offer interest the mortgage market because some of the subprime market. An expanded GSE rates based on whether or not the borrower more risky borrowers (who are currently presence in the subprime market could be of falls into a certain category of risk, such as cross-subsidized by less risky borrowers) may significant benefit to lower-income and specific loan-to-value and FICO score not be able to afford their higher, risk-based minority families if it attracted more combination or specified mortgage score interest rate. Also, the adoption of an mainstream lenders and competition to those range. Lenders could also set the interest rate automated risk-based pricing system may inner-city neighborhoods that are currently based on various factors including the have an uncertain effect on minority groups, served mainly by subprime lenders. probability of prepayment and characteristics who tend to have lower credit scores, as Several commenters indicated that to of the underlying collateral, as well as the discussed earlier. On the other hand, if obtain the higher housing goals the GSEs default risk of the borrower. Borrowers that minorities are eligible for prime financing, would increase their purchasing of subprime pose a lower risk of loss to the lender would the cost of financing minorities may fall as loans. While some industry commenters then be charged a comparatively lower rate will the potential for subprime lenders to welcome the entrance of the GSEs into the than those borrowers with greater risk. Rather draw minorities to their higher-priced subprime market because their presence than lower risk borrowers cross-subsidizing products. brings stability and standardizes business higher risk borrowers like in average cost As the GSEs become more comfortable practices, they are concerned that pricing, lower risk borrowers pay a relatively with subprime lending, the line between unrealistically high goals could force the lower rate. what today is considered a subprime loan GSEs to jump into the market in a manner In response to the expanded use of versus a prime loan will likely deteriorate, that negatively distorts underwriting and automated underwriting and pressures from making expansion by the GSEs look more pricing. These commenters report that the the GSEs, other purchasers of loans, mortgage like an increase in the prime market. This GSEs can bring capital and standards but insurers, and rating firms, lenders are melding of markets could occur even if many must gradually and carefully enter the increasing their use of risk-based pricing.212 of the underlying characteristics of subprime subprime market in order to have a positive In today’s markets, some form of differential borrowers and the market’s evaluation of the effect. pricing exists for the various subprime risks posed by these borrowers remain In the past, Fannie Mae and Freddie Mac categories, for new products targeted at unchanged. Increased involvement by the have voluntarily decided not to purchases credit-impaired borrowers (such as Fannie GSEs in the subprime market will result in subprime loans with features such as single- Mae’s Timely Payments product), and for more standardized underwriting guidelines premium life, HOEPA loans, and prepayment private mortgage insurance across all credit and the increased participation of traditional ranges. For example, private mortgage lenders. In fact, there are indications that 207 Temkin et al., 2002, p. 1. insurers use FICO scores and ‘‘Accept’’ mainstream players are already increasing 208 See Lax et al., 2000. determinations from the GSEs’’ automated 209 Zorn, et al., 2001, p. 5. underwriting systems to make adjustments to 213 For example, see Radian’s product offerings at 210 Fannie Mae, Remarks Prepared for Delivery by http://www.radiangroupinc.com. Franklin Raines, Chairman and CEO of Fannie Mae 211 Annual Percentage Rate takes into account 214 Vanessa Bush, ‘‘Risk-Based Pricing Trend to the National Community Reinvestment Coalition. points, fees, and the periodic interest rate. Could Make Mortgage Lending More Efficient,’’ Washington, D.C. March 20, 2000. 212 Temkin et al., 2002, p.29. America’s Community Banker, October 1, 1998.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00093 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63672 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
their activity in this market. According to Rule and the HUD-sponsored research report, slowdown in economic activity meant fewer staff from Moody’s Investors Service, the Study of Multifamily Underwriting and the apartment customers, with less money, than growing role of large mortgage aggregators in GSEs’ Role in the Multifamily Market (Abt if the economy were vigorously expanding. the subprime market has been a key factor in Associates, 2001). Persistent low interest rates have also enticed the improving credit qualify on deals issued 2. The Multifamily Rental Housing Market: renters into the home purchase market as 215 in 2002. According to a representative 2000–2003 evidenced by the U.S. homeownership rate, from Washington Mutual, subprime credit which grew to 68.4 percent in 2003, further qualify has also improved as lenders carve The definition of ‘‘good’’ market conditions contributing to a weakness in rental demand. out new loan categories that fall somewhere in multifamily rental housing depends on The reduced demand is most evident in the between the large Alt A market and one’s perspective. Investors and lenders like national statistics on employment. Job traditional subprime business.216 As the low vacancies, steady rent increases, and growth began decelerating in late 2000 and subprime market becomes more rising property values. Developers like strong throughout 2001, turning negative late that standardized, market efficiencies will reduce demand for new construction and favorable year. The largest year-over-year job loss of the borrowing costs. Lending to credit-impaired terms on construction financing. Consumers, economic downturn occurred in February in contrast, prefer low rents and a wide borrowers will, in turn, increasingly make 2002, and year-over-year losses have selection of available apartments. good business sense for the mortgage market. continued through October 2003.221 The mid- to late-1990s were among the Apartment demand seems particularly most successful of recent history, in that D. Factor 2: Economic, Housing, and sensitive to labor market conditions, given apartment market conditions were generally Demographic Conditions: Multifamily the importance of rental housing to mobile good for all of these interest groups. Mortgage Market individuals and families accepting new jobs Investment returns were favorable, 1. Introduction or transfers. Reis, Inc., a real estate market construction volumes were steady at research firm, estimates that the total number At the time of the previous GSE sustainable levels, and many consumers had of occupied apartments (in properties with rulemaking in 2000, the multifamily rental income gains in excess of their rent increases. housing market was coming off several years Market conditions for multifamily rental 40+ units) actually declined in both 2001 and of generally positive performance. Vacancies 2002 in the large markets nationwide that are housing began to weaken toward the end of 222 were low in most markets and rent increases 2000. Early warnings came from the publicly monitored by the company. Job numbers were matching or exceeding economy-wide traded apartment companies, some of which showed some rebound in the subsequent inflation. A key to this strong performance reported easing in demand growth in the first period. was the volume of new multifamily months of 2001, coinciding with a slowdown Households, not jobs, fill apartments, and construction, which was at a level consistent in job growth to its lowest level since 1992. for this reason household formations are a with demand growth. Job growth and income By 2003, rental units were experiencing preferable indicator of demand for gains helped many renters pay the higher record high vacancy rates and newly apartments as well as other types of housing. rents without undue burden. As always, completed apartments faced record low The Census Bureau estimates that the total conditions varied from region to region, and absorption or ease-up rates. The rental sector number of renter households nationwide has across market segments, but the overall tone vacancy rate averaged 9.8 percent in 2003, up been essentially unchanged at approximately of the apartment market was quite healthy. 0.8 percent from 2002, and the highest 34.8 million since 1996. Yet during the late Much has changed in the subsequent years. annual vacancy rate I the more than 40-year 1990s apartment demand was expanding, An economic slowdown reduced apartment history of the measure.217 and apartments were apparently picking up demand, and with new multifamily Apartments—especially those serving the market share from other rental housing. The construction about unchanged, vacancies top end of the rental market—appear to have past two or three years may have seen a rose and rents softened. Provision of decent performed worse than other rental housing in reversal of that trend in share. housing affordable to households of moderate the past four years, after several years of rent Long-term demographic trends are expected to be favorable for rental housing or low incomes is a challenge even in strong growth and occupancies surpassing the rental 223 economic times, and with the unemployment market averages. The multifamily (5+ units in demand. The maturing of the ‘‘Baby Boom rate rising above 6 percent before falling to structure) vacancy rate has increased more Echo’’ generation will increase the number of persons under age 25 who will seek rental about 5 and a half percent, affordability than the overall rental market vacancy rate in housing, immigration is expected to continue problems increased for many, despite the each of the years 2000, 2001, 2002, and 2003. to fuel demand for rental housing, and softness in rents. For example, the Census Bureau’s estimate of minority populations, while increasing their Despite the recent weakness in the a 0.9 percentage point increase in vacancies homeownership rates, are growing and will apartment property market, the market for for multi-family apartments in 2003 exceeds contribute to higher absolute demand for financing of apartments has grown to record the overall rental vacancy rate of 0.6 rental housing. Thus demographic trends volumes. The favorable long-term prospects percent.218 Similarly, while rent growth has support an improvement in the long-run for apartment investments, combined with decelerated slightly for all rental housing demand for rental demand, which is likely to record low interest rates, has kept investor according to the CPI, industry surveys of include higher multifamily rental demand. demand for apartments strong and supported apartment rents show year-over-year declines Supply growth has been maintained, even property prices. Refinancings too have in rents in many local markets.219 In 2003, though the current reduced multifamily grown, and credit quality has remained very asking rents remained flat nationally, as demand warrants less new construction. high. Fannie Mae and Freddie Mac have been multifamily completions declined 5 220 Total multifamily starts (2+ units) have been among those boosting volumes and percent. running approximately 325-to-350 thousand introducing new programs to serve the a. Apartment Demand and Supply annually for the past six years, according to multifamily market. The primary reason for the softening in the Census Bureau statistics, adding about 1 This section will review these market multifamily rental market has been a percent annually to the total multifamily developments, interpret the performance of reduction in the growth of consumer demand stock. Most of these new units are built for Fannie and Freddie within that market for apartment housing. The general rental use, with only about 20 percent in context, and discuss future prospects for the multifamily rental market, its financing, and 217 the GSE role. The intention here is only to U.S. Department of HUD, Office of Policy 221 U.S. Department of Labor, Bureau of Labor Development and Research, U.S. Housing Market update the discussion from 2000. For general Statistics, ‘‘Bureau of Labor Statistics Data,’’ Conditions: 4th Quarter 2003, February 2004, p. 3. Accessed July 31, 2004, http://data.bls.gov/servlet/ background information on the multifamily 218 U.S. Department of HUD, Office of Policy SurveyOutputServlet?data_tool mortgage market and the GSEs, see the 2000 Development and Research, U.S. Housing Market =latest_numbers&series_id=LNS14000000. Conditions: 4th Quarter 2003, February 2004, p. 84. 222 ‘‘Apartment Landlords Gather to Dreary 215 ‘‘Improving Credit Quality, Maturing Business 219 See, for example, Marcus & Millichap Outlook for Sector,’’ Wall Street Journal, January 15, Stoke Confidence in Subprime MBS Market,’’ Inside Research Services, National Apartment Report, 2003, Section B. MBS & ABS, published by Inside Mortgage Finance, January 2003. 223 Mortgage Bankers Association of America, February 21, 2003. 220 Marcus & Millichap Research Services, ‘‘MBA News Link: Rental Market Demographics 216 Ibid. National Apartment Report, January 2004. ‘‘Favorable,’’ Report Says,’’ January 2003.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00094 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63673
recent years reported as being built as for-sale growth of 1997–1999.225 Other work noted other subsidies, most of the multifamily condominium units. that rent to income ratios for the lowest rental construction in recent years has been The reduced short-term demand has shown income quintile of renters rose during the targeted on the upper end of the market, through in absorption speeds for new late 1990s even as these ratios were stable or often the only segment for which apartments. The percentage of newly declining for other renters.226 Harvard’s State unsubsidized new construction is completed unfurnished apartments rented of the Nation’s Housing report for 2002 economically feasible. The median asking within three months of completion fell from highlighted the variability of the affordability rent on new unfurnished apartments 72 percent during 2000 to 63 percent during problem from place to place.227 completed in 2001 was $877, up 11 percent 2001 and to 59 percent during 2002, the Little research is available on affordability over the previous two years. In 2002 median lowest level in the 33-year history of the data trends since 1999. However, tabulations from asking rent for these properties was $905. Of series, according to the Census Bureau. This the 2001 American Housing Survey indicate those units brought to market in 2002, 45 percentage rose slightly to 60 percent in that income growth between 1999 and 2001 percent were at rents at or above $950. 2003.224 in the lowest quintile of renter households 3. Multifamily Financing Trends b. Performance by Market Segments continued to lag that of higher income In contrast to the softening observed in the renters, and fell short of the average rent Some segments of the multifamily rental demand/supply balance for multifamily, increases during this period. Together, these market have been more affected than others mortgage financing of these properties has by the recent softening. As mentioned earlier, statistics suggest that affordability has been at a record pace in the past three years. deteriorated early this decade among at least the top end of the apartment market seems a. Lending Volume especially hard hit, as measured by rising this group of very low-income renters. Other Total multifamily mortgage debt vacancies and reduced rent growth. This work using the AHS found that the number outstanding increased 11 percent in 1999, 8.7 segment is particularly dependent on job of low-to moderate-income working families percent in 2000, 11.2 percent in 2001, 9.6 growth and transfers for new customers, and with severe rental cost burdens increased 24 228 percent in 2002, and 11.2 percent in 2003 is particularly vulnerable to losses of percent between 1999 and 2001. according to the Federal Reserve’s flow of residents and prospective customers to home The low-income housing tax credit funds accounts. The dollar volume for 2003, purchase. According to reports by apartment (LIHTC) continues to finance much of the $544.2 billion, is above those of any previous REITs and other investors, these top-end newly built multifamily rental housing that is affordable to households with moderate year. The pace seems to have slowed for properties have not been getting the job- 2004, with the first quarter indicating an related in-movers, but have still been losing income. Restricted to households with incomes no greater than 60 percent of the annualized growth of 4.9 percent. a lot of customers to home purchase. These Furthermore, a 2003 survey by the Mortgage properties generally have annual resident local median, this program financed approximately 75,000 units in 2001, Bankers Association of America show that of turnover rates of above 50 percent, and thus 48 member firms surveyed, representing all are particularly quickly influenced by according to the National Council of State Housing Agencies, after running in the mid- large mortgage banking firms an a cross changes in demand. Furthermore, this is the section of smaller mortgage companies, segment of the apartment market into which to high-60 thousand range the previous three years. About 70 percent of these units are multifamily origination volume increased most of the new construction is built. 21.5 percent in 2003—from $41 billion in Performance has varied geographically as newly built, and the rest are renovations of existing units. 2002 to $49.8 billion in 2003. well. Some of the coastal markets, especially The apparent inconsistency between Expenditures for improvements to existing in Northern California, saw the double-digit current market fundamentals and financing rental apartments have grown in recent years. rent increases of the late 1990s replaced by can be explained by low interest rates. The In 2001 the total of $11.3 billion was nearly double-digit declines, before stabilizing more same financial forces that lowered the twice the figure of three years earlier, recently. ‘‘Supply constrained markets’’ had mortgage rates for home purchasers to record been preferred by apartment investors during according to the Census Bureau, and more lows by 2002 also reduced the financing the 1990s, but recent market performance has than a third as large as construction spending costs of multifamily properties. The ten year reminded investors and analysts that all for newly built multifamily structures, Treasury yield, a common benchmark for markets have their day. For example, including owner-occupied condos. Many of multifamily loan pricing, fell to a 45-year low Houston posted the biggest year-over-year these improvements are to older properties in of 3.3 percent in June 2003 from 6.3 percent rent increase of any major apartment market high-demand neighborhoods. Improvements as recently as the end of 1999. in 2001, despite a long-run history of to the physical structures have external Another feature boosting investor demand moderate rent growth and few barriers to new benefits. But often the renovations are in for apartment properties and the resulting apartment construction. Rent changes in the connection with re-positionings that move demand for debt to finance those purchases 27 metro markets for which estimates are the apartments into a higher rent range and has been the lack of attractive returns on available from the CPI ranged from a low of bring changes in the demographic many financial assets and other alternative ¥0.3 percent to a high of 6.7 percent in the composition of the resident base. investments. Despite the current weak first half of 2003 relative to a year earlier. In 2002, expenditures on total performance of apartments, investors And across the 75 metro areas for which improvements to existing apartments apparently are looking through to the long- rental vacancy rates (apartments plus other declined to $9.8 billion, while new run outlook for these assets, which is rentals combined) are available, rates for the construction spending increased $2 billion. generally thought to be favorable, as year 2002 ranged from 2.4 percent to 15.4 This shift further suggests a re-positioning to indicated most recently by investor surveys percent, according to the Census Bureau. In apartments with a higher rent range. fielded by the Urban Land Institute and by a historical context, this variation is Excluding units financed with tax credits or Lend Lease and PriceWaterhouseCoopers.229 moderate, although up somewhat since the The net change in mortgage debt late 1990s. 225 Center for Housing Policy/National Housing outstanding is defined as loan originations Conditions in the ‘‘affordable’’ segment of Conference, ‘‘Housing America’s Working Families: less repayments and charge offs. As the apartment market are harder to track than A Further Exploration,’’ New Century Housing, Vol. discussed in Appendix D, net change is a in the high-end segment because of lesser 3, No. 1, March 2002; Mark Obrinsky and Jill lower bound on originations. By all accounts, investor interest and analyst coverage. Data Meron, ‘‘Housing Affordability: The Apartment originations—for which no single source of for the late 1990s analyzed by the National Universe,’’ National Multi Housing Council, 2002. estimates is available—are much higher than Housing Conference saw affordability 226 ‘‘Housing Affordability in the United States: net change in most years. High levels of problems continuing, although a study of Trends, Interpretations, and Outlook,’’ a report refinancings of existing multifamily prepared for the Millennial Housing Commission by apartment renters by the National Multi J. Goodman, November, 2001. mortgages in recent years has been a factor Housing Council saw some improvement in 227 Joint Center for Housing Studies of Harvard in originations exceeding the net change in affordability during the strong economic University, State of the Nation’s Housing, 2002. debt outstanding. 228 Center for Housing Policy/National Housing 224 U.S. Department of HUD, Office of Policy Conference, ‘‘America’s Working Families and the 229 Urban Land Institute, The ULI Forecast, 2002; Development and Research, U.S. Housing Market Housing Landscape 1997–2001,’’ New Century Lendlease and PriceWaterhouseCoopers, Emerging Conditions: 4th Quarter 2003, February 2004, p. 70. Housing, Vol. 3, No. 2, November 2002 Trends in Real Estate, 2003.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00095 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63674 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Most mortgage lending is in the No source of apartment property sales refinancing of FHA multifamily mortgages, ‘‘conventional’’ market. Multifamily loan statistics matches the comprehensive experienced an increase in origination programs of the Federal Housing national coverage of the single-family market volume of 133 percent in Fiscal Year 2003 Administration accounted about $7 billion in provided by the National Association of and 181 percent in Fiscal Year 2002. ($1.73 new insured mortgages in fiscal year 2003— Realtors’ monthly estimates. But surveys by billion in FY 2003, $0.74 billion in FY 2002, up from $6 billion in fiscal year 2002 and $5 the National Multi Housing Council and and $0.26 billion in FY 2001). Similarly, the billion in fiscal 2001. Despite the recent other apartment industry reports indicate GSEs increased their combined volume of increase in FHA originations, and the likely that transactions volume dipped during 2001 refinances by 83 percent from 1999–2000 to continued strong performance for FHA but since then have grown appreciably in 2001–2002, from $17.6 billion to $32.1 multifamily programs in the foreseeable both number of sales and aggregate dollar billion. Refinancings, especially when future, 230 FHA remains but a small portion value. of the total multifamily mortgage market. Mortgage lending volumes have recently motivated by a desire to lower interest Outstanding FHA-insured multifamily been boosted by shifts in property expense rather than to extract equity, do not mortgage debt was $55 billion at the end of ownership. Publicly traded real estate add as much to debt outstanding as do the first quarter of 2003—only about 11 investment trusts had been the big gainers purchase loans, which often are much larger percent of all multifamily mortgage debt during most of the 1990s, and by 1999 owned than the seller’s existing mortgage that is outstanding. nearly 6 percent of all apartments nationwide repaid at the time of sale. Nonetheless, Multifamily lending has been spurred by and a considerably larger share of all big refinancings represent a significant part of all new apartment construction, property sales, (100+ unit) properties. But beginning in 1999 multifamily mortgage lending. and refinancings. New multifamily capital market developments made private c. Sources of Financing and Credit Quality construction was valued at $34.1 billion in buyers more competitive. Since then the 2003, according to the Census Bureau, up 21 number of apartments owned by large REITs The sources of funding of multifamily percent from 2000.231 The number of new has declined about 5 percent, with diverse mortgages shifted somewhat in the past few multifamily units completed over this period private interests apparently picking up years, judging from the Flow of Funds actually declined 12 percent, and the market share. accounts. As shown in Table A.4, four increased expenditures reflect higher costs Private investors are able to use more categories of lenders have dominated per unit. The increase in asking rents leverage—greater debt—in financing their multifamily mortgage lending since the mid- described earlier suggests higher property transactions than the market permits the 1990s. Of those four, commercial banks have values and greater debt carrying capacity. public REITs. As a result, the very low played a lesser, although still substantial, b. Property Sales and Refinancings mortgage rates recently have given them an role in recent years, providing 20 percent of Sales of existing apartment properties tend advantage in bidding on properties. In the $86 billion in net additional funding of to be procyclical. Increasing asset values addition, equity funding costs of REITs rose multifamily mortgages during 2000 and 2001. bring buyers to the market and tempt sellers as their stock prices flattened or moved down The portfolio holdings of the GSEs, by to realize their capital gains. In soft markets, as part of the broader equity market contrast, have been much more important in contrast, the bid-ask spread generally correction. than during the last half of the 1990s. widens and the volume of sales declines, as Refinancings have, by all accounts, also Mortgage backed securities, both from the been strong. Despite the lockout provisions sellers perceive current offers as beneath the GSEs and especially from other issuers, and yield maintenance agreements that property’s long run value and buyers are accounted for proportionally less of the constrain early refinancings of many reluctant to pay for past performance or the growth in 2000–01 than in 1995–99, but hope of future gains. Sales tend to increase multifamily loans, lenders reported very between them still accounted for nearly half mortgage debt, because the loan originated to strong refinancing activity in 2001 and of all the net credit extensions. Some slight finance the purchaser’s acquisition is continuing into 2002. Although refinancing typically considerably larger than the volume data for the entire market are not broadening of the base of multifamily lending mortgage retired by the seller. available, the trends in refinance volume for in the past two years, as these four lender FHA and the GSEs show very strong groups accounted for only 85 percent of the net credit extended in 2000 and 2001, 230 Merrill Lynch, A New Look at FHA increases in refinance activity during 2002 Prepayments and Defaults, September 2002. and 2003. For example, FHA’s Section compared to all of it in the previous five-year 231 Eight percent inflation adjusted. 223(a)(7) program, which is limited to period.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00096 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63675
The market values of apartment properties loan officers surveyed quarterly by the 1995. By this combined measure of portfolio have generally held up well, although the Federal Reserve have reported tightening holdings and MBS outstanding, at year-end most recent indicators suggest some their terms on commercial mortgages, and 2002 Fannie Mae had nearly twice ($65 flattening. Properties in the portfolios of that shift likely has occurred in their billion versus $37 billion) the multifamily pension funds continued to appreciate into multifamily lending as well. Perhaps the best business of Freddie Mac, although Freddie the second half of 2002, according to the indicator of discipline in multifamily lending was growing its multifamily business more National Council of Real Estate Investment is the fact that, despite the strong apartment rapidly (67 percent increase between 2000 Fiduciaries, although at a reduced annual demand during the last half of the 1990s, and 2002, compared to 46 percent increase rate of less than 2 percent. And the sales construction never rose above its long-run for Fannie Mae). In 2003, Freddie Mac’s price per square foot of ‘‘Class A’’ properties sustainable level, unlike the rampant multifamily business activities totaled monitored by Global Real Analytics rose overbuilding that plagued the industry in the $21.587 billion ($14.894 billion of mortgage until turning down in early 2002, posting a mid- and late-1980s. purchases and $6.693 billion in investment 1.6 percent year over year decline in the activities). These activities financed rental second quarter. 4. Recent GSE Involvement in Multifamily housing for 549,083 families. Nearly 92 The continuing value of collateral has Finance percent of these units were affordable to low- helped keep loan quality high on multifamily As the multifamily mortgage market has and moderate-income renters. Since 1993, mortgages. Delinquency rates from all major expanded since 1999, Fannie Mae and Freddie Mac has purchased $75.5 billion in reporters are at or near record lows, and well Freddie Mac have increased their lending, multifamily mortgages, financing housing for below the rates reported for single-family picked up market share, introduced new more than 2.2 million families.232 mortgages and commercial properties. At programs, and enhanced others. Measures that focus on new multifamily commercial banks, the FDIC reports a 0.38 Beginning with their whole loans, the GSEs activity, specifically gross mortgage purchase non-current loan percentage in the second added 34 percent to their combined holdings volumes and new security issuance, vary quarter of 2002. In life insurance company of multifamily loans in 2001, and another 26 across recent years and between the GSEs. portfolios the only 0.05 percent of residential percent in 2002 (see Table A.6 below). The For the GSEs combined, these measures of mortgages were overdue at the end of 2002, growth in multifamily MBS volume was current business activity show sharp gains of and as of the third quarter of 2002 the GSEs nearly as dramatic, increasing 26 percent in over 70 percent in 2001, following small were both reporting similarly miniscule 2001 and another 14 percent in 2002. The decreases in activity in 2000. In 2002, the delinquency rates of below 0.1 percent; all of gains resulted in the GSEs increasing their GSEs combined posted small declines for these rates are below those of a year earlier. share (whole loans and securities combined) both measures. Measures of multifamily gross Multifamily lenders have remained of all multifamily debt outstanding to 22.8 mortgage purchases and new security cautious in their underwriting and, together percent by the third quarter of 2003, up from with their regulators; have avoided repeating 19 percent at year-end 2001, 15 percent at 232 Freddie Mac Public Comment Letter on HUD’s the mistakes of the 1980s. Many of the senior year-end 1999 and 11 percent at the end of Proposed Goals, July 2004, p.3.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00097 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.013 63676 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
issuance diverged for the two GSEs in 2002. market remains well below their dominance heavily influenced by the volume of Fannie Mae experienced declines in these in single-family mortgage finance. At the end refinancings in the single-family market, balance sheet and new business indicators in of 2002, the GSEs’ market share of single which spiked in 1998 and again in 2001 and 2002 while Freddie Mac experienced gains, family debt outstanding was 44 percent, 2002 in response to the big decline in particularly in new security issuance. As twice the share of multifamily debt held or mortgage rates in those years. Because of discussed earlier, the credit quality of GSE securitized by these two companies, lock-out agreements and other loan multifamily loans has remained very high according to Federal Reserve statistics. even with the large gains in loan volume. Furthermore, the multifamily share of all covenants, multifamily loans are not as prone Despite the substantial pickup in GSE housing units financed by the GSEs to rate-induced refinancings as are single- multifamily activity, the position of these combined has declined from its 1997 level family mortgages. companies in the multifamily mortgage (Table A.5), although the annual statistics are BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00098 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63677
a. Contrasting Business Models distinct business models in achieving that relied more on loans purchased and held in While both Fannie Mae and Freddie Mac growth. As shown in Table A.6, most of its portfolio. At the end of 2002, Fannie Mae have significantly increased their multifamily Fannie Mae’s multifamily growth has come had almost four dollars of outstanding MBSs activities in recent years, they have pursued in MBS products, whereas Freddie Mac has for every dollar of portfolio holdings. Freddie
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00099 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.014 63678 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Mac, on the other hand, more than three times as much volume in portfolio as it had in MBS outstanding.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00100 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.015
The differing emphasis on portfolio assumes the entire credit risk on loans it Owing to this high propensity to qualify as holdings and securities issuance is related to purchases, some commercial banks and other affordable lending, financing of multifamily the GSEs’ contrasting approaches to credit financial institutions desiring to remove rental housing is especially important for the underwriting.233 Fannie Mae has long had multifamily loans and all related liabilities GSEs attainment of their affordable housing risk-sharing arrangements with its from their books find Freddie Mac’s program goals. Less than 8 percent of the units multifamily loan originators, and currently preferable. financed by the GSEs in 2002 were has over 25 Delegated Underwriters and b. Affordable Multifamily Lending multifamily rentals, as described above. Yet Servicers who are authorized to originate 15 percent of the units qualifying as low- and Because most of the GSEs’ multifamily loans meeting Fannie Mae’s requirements for moderate-income purchases were lending is on properties affordable to sale to the GSE without prior approval of households with low-or moderate incomes, multifamily, according to Table 1 of the individual transactions. These ‘‘DUS’’ financing of affordable multifamily housing GSEs’ activity reports for 2002. lenders retain part of the credit risk on the by the GSEs has increased almost as much as The GSEs increased the volume of their loans sold to Fannie. their total multifamily lending. affordable multifamily lending dramatically Freddie Mac has taken a different approach Approximately 87 percent of Fannie Mae’s in 2001, the first year of the new, higher to credit underwriting. In the wake of large multifamily lending volume in 2003 affordable housing goals set for the GSEs. As credit losses on its multifamily business in qualified as affordable to low-or moderate measured by number of units financed, the the late 1980s and 1990, Freddie Mac income households, according to Fannie total affordable lending (shown in the ‘‘low- essentially withdrew from the market. When Mae’s annual Housing Activity Report, as did mod total’’ rows of Table A.7) more than it re-entered in late 1993, the company 92 percent of Freddie Mac’s multifamily doubled from a year earlier, especially after elected to retain all underwriting in-house units financed. For the entire multifamily application of the upward adjustment factor and not delegate this function to the loan rental market, HUD estimates that 90 percent authorized for Freddie Mac in the 2000 Rule. originators participating in Freddie Mac’s of all housing units qualify as affordable to In 2003 the GSEs maintained a high volume Program Plus network. Because Freddie Mac families at or below 100 percent of the area of affordable multifamily lending.234 median income, the standard upon which the 233 ‘‘No Mistaking GSEs for Twins in low- and moderate-income housing goal is 234 This change was a percentage decrease but a Multifamily,’’ American Banker, October 2, 2002. defined. volume increase.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00101 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63680 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00102 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.016 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63681
The figures in Table A.7 are exclusive of financed over 809,703 multifamily units. Of GSEs face strong competition in this market the ‘‘Temporary Adjustment Factor (TAF)’’ this total, over 87% were affordable to from small banks and other depository granted to Freddie Mac as part of the 2000 families at or below the median income of institutions that prefer to hold these loans in Rule. The TAF was a response to Freddie their communities.238 Freddie Mac their own portfolios.244 Mac’s limited opportunities for refinancing multifamily business activities totaled a In 2003, Freddie Mac continued to test business because of its minimal involvement record $21.587 billion which financed rental initiatives through pilots, and implement in the multifamily market in the early and housing for 549,083 families. Nearly 92 enhancements to existing multifamily mid-1990s.235 The TAF, which expired at the percent of these apartment units were mortgage products which cover a broad array end of 2003, provided a 20 percent upward affordable to low- and moderate income of eligible mortgage products. Freddie Mac’s adjustment to multifamily units in properties renters.239 tax-exempt bond credit enhancements with with 50 or more units, for purposes of the Programs introduced or enhanced by the synthetic fixed-rate financing continued to be affordable housing goals. GSEs in the past two years have contributed popular. Freddie Mac’s innovations to certain Multifamily financing made major to these striking numerical results. Delegated cash products including various contributions not only to the GSEs’ Underwriting and Servicing (DUS) is Fannie combinations of fixed-rate, adjustable-rate attainment of the overall goal for affordable Mae’s principle product line for purchasing and interest-only mortgages have been lending in 2002, but also to the ‘‘underserved individual multifamily loans. This product adopted by others in the industry. For areas’’ goal and ‘‘special affordable’’ goal. As line is offered through 26 lenders with example, the Fixed-to-Float execution shown in Table A.7, the 2001 increases in expertise in financing multifamily properties. provides borrowers with a reduced fixed lending in each of these categories were In 2003, 91% of the DUS loan activity served interest rate and a one-year extension of the substantial at both Fannie Mae and Freddie affordable housing needs, 42% of DUS loans mortgage term at a floating rate. In 2003, Mac, again leveling off for both in 2002. The in underserved markets, and 52% addressed borrowers used Fixed-to-Float option for $4.0 GSEs also met the special multifamily ‘‘special affordable’’ needs.240 Believing that billion in mortgages.245 affordable subgoal set in the 2000 Rule in small multifamily properties are a vital part In 2003, Freddie Mac purchased $6.6 both 2001 and 2002. of the country’s affordable housing stock, billion in mortgages to finance more than c. Multifamily Initiatives of the GSEs Fannie Mae has focused efforts on providing 181,000 apartment units in 5-to 50-unit Fannie Mae and Freddie Mac have taken a financing for these projects through the properties. Freddie Mac committed to invest number of steps since 2000 to expand their development of the MFlex Loan Product, the $958 million to Low Income Housing Tax multifamily lending and to respond 3MaxExpress Streamlined Mortgage Loan Credits (LIHTC). Altogether, the LIHTC specifically to the goals established in the Product and the Affordable Alliances Loan investments made by Freddie Mac are 2000 Rule. These initiatives are summarized Product. The MFlex Loan Product was approaching the $3.6 billion mark and have in the annual activity reports filed by the established in 2000 to target lending partners constructed or rehabbed more than 216,000 GSEs.236 that serve small property borrowers and rental units for very-low and low income One focus of the 2000 Rule was on lending increase Fannie Mae’s participation in the 5– families in close to 3,000 projects. In 2003, to small (5-to-50 units) multifamily 50 unit property market. By 2003, Fannie Freddie purchased $412 million in newly properties, which the Rule identified as an Mae had seven MFlex lending partners and issued multifamily mortgage revenue bonds. underserved market. HUD-sponsored had purchased $1.6 billion of these loans. These bonds, issued by state, county or city research has found that the supply of Fannie Mae markets its specialized government agencies, finance the acquisition mortgage credit to small properties was 3MaxExpress Streamlined Mortgage Loan and rehabilitation of nonprofit borrowers or impeded by the substantial fixed costs of Product line for loans worth less than or property owners who agree to keep rents at multifamily loan originations, by owners’ equal to $3 million. In 2003, Fannie Mae affordable levels. These multifamily bond insufficient documentation of property provided $1 billion in financing, which purchases will finance 6,100 estimated units income and expense, and by the limited assisted over 34,000 families living in small of affordable housing with an estimate that opportunities for fees for underwriting and multifamily properties. The Affordable 58 percent of those units will be affordable servicing small loans.237 As a result, many Alliances Loan Product is responsible for to very low income families. In 2003, Freddie multifamily lenders focus on larger debt investments in rental housing targeted issued a record $7.7 billion of securities properties, which were found to have more to persons of low- and moderate-income and backed by multifamily mortgages through loan products available to them and to pay to rental markets that are underserved. negotiated transactions. More than 85 percent lower interest rates than did small properties. During 2003, these financing initiatives of these securities financed mortgages for 246 In an attempt to promote the supply of provided affordable housing for 3,850 affordable housing. credit to small properties, the 2000 Rule families. 241 Fannie Mae additionally has The 2000 Rule discussed other ways in provided incentives for the GSEs to step up federal Low-Income Housing Tax Credit which the GSEs might help promote their involvement in this segment of the (LIHTC) programs and special financing financing of affordable multifamily housing. multifamily mortgage market. The incentives projects for special use properties such as Two of those were lending for property likely contributed to the huge increases in Seniors Housing. In 2003, Fannie Mae rehabilitation and leadership in establishing small property lending posted by both Fannie committed over $1.6 billion in LIHTC equity standards for affordable multifamily lending. Mae and Freddie Mac in 2001 and continuing properties to help make affordable rental Many affordable properties are old and in into 2002 (Table A.7). The combined total of housing possible for over 30,000 families.242 need of capital improvements if they are to these units financed in 2001 and 2002 was During 2003, Freddie Mac used innovative remain in the housing stock. Rehabilitation almost 8 times those financed in the previous financing structures combined with prudent, lending is a specialized field, and one in two years. This lifted the percentage of all flexible multifamily lending practices, which which the GSEs for a variety of reasons have GSE multifamily lending that was on small enabled them to reach a record level of not been major players. Less than 1 percent properties to their highest levels ever. multifamily mortgage purchases.243 The of all GSE multifamily lending in 2002 was During 2003, multifamily business activity for property rehabilitation. In 2002, Fannie at Fannie Mae topped $33 billion which 238 Fannie Mae, 2003 Annual Housing Activities Mae hosted its first ever Preservation Report, March 15, 2004, p. 26. Advisory Meeting with leaders in the 239 housing and real estate finance industry to 235 For background information on the Freddie Freddie Mac, Opening Doors for America’s Mac TAF, see pages 65054 and 65067–65068 of the Families: Freddie Mac’s Annual Housing Activities identify best practices and formulate real 2000 Rule. Report for 2003, March 15, 2004, p. 44. 236 Fannie Mae’s 2002 Annual Housing Activities 240 Fannie Mae, 2003 Annual Housing Activities 244 ‘‘Fannie Courting Multifamily Sellers; Small Report, pages 24–27; Freddie Mac’s Annual Report, March 15, 2004, p. 27. Banks Balking,’’ American Banker, January 13, Housing Activities Report for 2002, pages 41–47. 241 Fannie Mae, 2003 Annual Housing Activities 2003. 237 Abt Associates Inc., An Assessment of the Report, March 15, 2004, p. 28. 245 Freddie Mac, Opening Doors for America’s Availability and Cost of Financing for Small 242 Fannie Mae, 2003 Annual Housing Activities Families: Freddie Mac’s Annual Housing Activities Multifamily Properties, a report prepared for the Report, March 15, 2004, p. 29. Repoert for 2003, March 15, 2004, p. 47 & 49. U.S. Department of Housing and Urban 243 Freddie Mac, Opening Doors for America’s 246 Freddie Mac, Opening Doors for America’s Development, Office of Policy Development and Families: Freddie Mac’s Annual Housing Activities Families: Freddie Mac’s Annual Housing Activities Research, August 2001. Report for 2003, March 15, 2004, p. 47. Report for 2003, March 15, 2004, p. 50–52.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00103 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63682 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
world solutions to this critical policy strong growth in single-family housing unlikely to change until voter attitudes issue.247 demand, some of that coming from apartment change. Setting standards for affordable renters seeking more space. With single- b. The Future Role of the GSEs multifamily lending was identified in the family activity already near record highs, 2000 Rule as another area where the GSEs boosted by historically low mortgage Regarding the mortgage financing of could provide greater leadership. It was also interests rates and despite the recently soft multifamily rental apartments, it is hard to noted, based on HUD-sponsored research economy, it is uncertain how much higher anticipate events that might disrupt the flow underway at that time,248 that market single-family demand—and the or alter the sources of mortgage credit to participants believe the GSEs to be accompanying losses of apartment customers apartments. In the past, certain events have conservative in their approaches to affordable to homeownership—can go. triggered such changes—notably the savings property lending and underwriting. Actions A stronger economy will put the and loan debacle of the 1980s and Freddie Mac’s withdrawal from the market following described in the GSEs’ annual activity reports multifamily rental market back onto a long- large losses in the early 1990s—but these are, for 2001, 2002 and 2003 indicate attempts by run path that appears to promise sustained, by definition, surprises. The current structure the GSEs to promote market standards that moderate growth. As discussed in the 2000 and performance of the multifamily mortgage will reduce the transactions costs of Rule, the demographic outlook is favorable market provide some comfort that the risks multifamily lending while also providing for apartment demand. Even if the are slight. The lender base is not overly programs that have the flexibility needed to homeownership rate increases further and dependent on any one institution or lender deal with unique circumstances. the total number of renter households grows type for either loan originations or funding. only slowly, as described in the discussion 5. Future Prospects Lending discipline appears to have been of the single-family housing market earlier in The outlook for the multifamily rental maintained, given the low mortgage this Rule, apartment demand can be expected housing market is marked by near-term risks delinquency rates even during the weak and longer-run optimism, according to most to increase more rapidly than that for other economy of the past two years. The near term observers. The prospects for the next few rental housing, owing to the likely changes outlook of most market participants is for quarters are dominated by the in age composition and reductions in average ample supply of mortgage financing at macroeconomy. In particular, job growth, household size. One estimate projects the historically low interest rates.253 Yet with its implications for formations of annual growth in apartment households to be complacency would be a mistake. 251 households, will be a key for the resumption one percent. Responding to both market incentives and of growth in apartment demand. Many a. The Outlook for Multifamily Housing their public charters, Fannie Mae and forecasters would ascribe to the Federal Supply Freddie Mac can be expected to build on Reserve’s forecast of a slight increase in GDP Regarding supply, one of the secrets of the their recent records of increased multifamily 249 growth to 4.3 percent in 2004 , while also success of the multifamily sector during the lending and continue to be leaders in agreeing with the Fed’s warning that ‘‘An 1990s was that production never rose above financing volumes, in program innovations, unusual degree of uncertainty attends the its long-run sustainable level. The discipline and in standards setting. Certainly there is economic outlook at present, in large of developers, investors, and their lenders room for expansion of the GSEs’ share of the measure, but not exclusively, because of that brought that result needs to be continued multifamily mortgage market, which, as 250 potential geopolitical developments.’’ if the apartment market is to maintain mentioned earlier, is by the measure of dollar When consumer demand does pick up, stability. volume outstanding currently only about half recovery should be reasonably fast. While the Multifamily housing may benefit in the the market share enjoyed by the GSEs in recent production levels have outpaced future from more favorable public attitudes single-family lending. And from the demand, they have been near the middle of and local land use regulation. Higher density perspective of units financed, the statistics the long run historical range and very close housing is a potentially powerful tool for from Table A.5 combined with data from the to the average of the last half of the 1990s. preserving open space, reducing sprawl, and 2001 American Housing Survey indicate that, Judging from the firm tone to rents and while the GSEs financed 7.2 percent of all the promoting transportation alternatives to the vacancies during that period, total nation’s year-round housing units that year, automobile. The recently heightened multifamily completions production of the percentage of multifamily rental units attention to these issues may increase the 275,000 to 350,000 units is a sustainable (that is renter-occupied units and vacant acceptance of multifamily rental construction level of annual production—that is, the level rental units in structures with at least five to both potential customers and their consistent with long run demographic trends units) was only 5.7 percent. prospective neighbors. and replacement of units lost from the stock. The sharp gains since 2000 in small Provision of affordable housing will Because new construction has remained property lending by Fannie Mae and Freddie continue to challenge suppliers of moderate, there is no massive overhang of Mac demonstrate that it is feasible for this multifamily rental housing and policy product that will need to be absorbed. With important segment of the affordable housing makers at all levels of governments. Low increased demand, vacancies should fall and market to be served by the GSEs. Building on rents firm reasonably promptly. A key incomes combined with high housing costs the expertise and market contacts gained in assumption behind this forecast for vacancies define a difficult situation for millions of the past three years, the GSEs should be able and rents is that new apartment construction renter households. Housing cost reductions to make even greater in-roads in small will not rise appreciably from its current are constrained by high land prices and property lending, although the challenges level. construction costs in many markets. noted earlier will continue. Recovery in the apartment market may Government action—through land use The GSEs’ size and market position also, perversely, be promoted by the recent regulation, building codes, and occupancy between loan originators and mortgage unprecedented strength of the single-family standards—are major contributors to those investors makes them the logical institutions market. Typically, economic recoveries bring high costs, as is widely recognized by market to identify and promote needed innovations participants, including the leaders of the and to establish standards that will improve GSEs.252 Reflecting the preferences of the 247 Fannie Mae, 2002 Annual Housing Activities market efficiency. As their presence in the Report, 2003, p. 27. electorate, these regulated constraints are multifamily market continues to grow, the 248 Abt Associates, ‘‘Study of Multifamily GSEs will have both the knowledge and the Underwriting and the GSEs’’ Role in the 251 Jack Goodman, ‘‘The Changing Demography of ‘‘clout’’ to push simultaneously for market Multifamily Market,’’ Final Report to the U.S. Multifamily Rental Housing,’’ Housing Policy standardization and for programmatic Department of Housing and Urban Development, Debate, Winter 1999. flexibility to meet special needs and Office of Policy Development and Research, August 252 Remarks by Franklin D. Raines, Chairman and circumstances, with the ultimate goal of 2001. CEO, Fannie Mae, to the Executive Committee of increasing the availability and reducing the 249 Federal Reserve, Survey of Professional the National Association of Home Builders, January Forecasters, November 2003. 18, 2003. See also Edward Glaeser and Joseph 250 Board of Governors of the Federal Reserve Gyourko, ‘‘The Impact of Zoning on Housing 253 ‘‘Capital Markets Outlook 2003,’’ Apartment System, Monetary Policy Report to the Congress, Affordability,’’ Working Paper 8835, National Finance Today, Vol. 7, No. 1 (January/February February 11, 2003, page 4. Bureau of Economic Research, March 2002. 2003).
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00104 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63683
cost of financing for affordable and other • In the October 2000 rule, the low- and borrowers and for first-time homebuyers. A multifamily rental properties. moderate-income goal was set at 50 percent summary of the main findings from that for 2001–03. As of January 1, 2001, several analysis is given in Section E.6. Section E.13 E. Factor 3: Performance and Effort of the changes in counting provisions took effect for then summarizes some recent studies on the GSEs Toward Achieving the Low- and the low- and moderate-income goal, as GSEs’ market role and section E.14 discusses Moderate-Income Housing Goal in Previous follows: ‘‘bonus points’’ (double credit) for the GSEs’ role in the financing of single- Years purchases of goal-qualifying mortgages on family rental properties. small (5–50 unit) multifamily properties and, This section first discusses each GSE’s 1. Performance on the Low- and Moderate- above a threshold level, mortgages on 2–4 performance under the Low- and Moderate- Income Housing Goal in 1996–2003 Income Housing Goal over the 1996–2003 unit owner-occupied properties; a period.254 The data presented are ‘‘official ‘‘temporary adjustment factor’’ (1.20 units HUD’s December 1995 rule specified that results—i.e., they are based on HUD’s credit, subsequently increased by Congress to in 1996 at least 40 percent of the number of analysis of the loan-level data submitted to 1.35 units credit) for Freddie Mac’s units financed by each of the GSEs that were the Department by the GSEs and the counting purchases of goal-qualifying mortgages on eligible to count toward the Low- and provisions contained in HUD’s regulations in large (more than 50 units) multifamily Moderate-Income Goal should qualify as low- properties; changes in the treatment of or moderate-income, and at least 42 percent 24 CFR part 81, subpart B. As explained missing data; a procedure for the use of of such units should qualify in 1997–2000. below, in some cases these ‘‘official results’’ imputed or proxy rents for determining goal HUD’s October 2000 rule made various differ from goal performance reported by the credit for multifamily mortgages; and changes in the goal counting rules, as GSEs in the Annual Housing Activities eligibility of purchases of certain qualifying discussed below, and increased the Low- and Reports (AHARs) that they submit to the government-backed loans to receive goal Moderate-Income Goal to 50 percent for Department. credit. These changes are explained below. 2001–03. The main finding of this section Fannie Mae’s low-mod goal performance was Table A.8 shows low-mod goal concerning the overall housing goals is that 51.5 percent in 2001, 51.8 percent in 2002, performance over the 1996–2003 period, both Fannie Mae and Freddie Mac surpassed and 52.3 percent in 2003; Freddie Mac’s based on HUD’s analysis. The table shows the Department’s Low- and Moderate-Income performance was 53.2 percent in 2001, 50.5 that Fannie Mae surpassed the goals by 5.6 Housing Goals for each of the eight years percent in 2002, and 51.2 percent in 2003, percentage points and 3.7 percentage points during this period. Specifically: thus both GSEs surpassed this higher goal in in 1996 and 1997, respectively, while • The goal was set at 40 percent for 1996; all three years. This section discusses the Freddie Mac surpassed the goals by narrower Fannie Mae’s performance was 45.6 percent October 2000 counting rule changes in detail margins, 1.1 and 0.6 percentage points. and Freddie Mac’s performance was 41.1 below, and provides data on what goal During the heavy refinance year of 1998, percent. performance would have been in 2001–03 Fannie Mae’s performance fell by 1.6 255 • The goal was set at 42 percent for 1997– without these changes. percentage points, while Freddie Mac’s 2000. Fannie Mae’s performance was 45.7 After the discussion of the overall housing performance rose slightly, by 0.3 percentage percent in 1997, 44.1 percent in 1998, 45.9 goals in Sections E.1 to E.5, Sections E.6 to point. Freddie Mac showed a gain in E.12 examine the role of the GSEs in funding percent in 1999, and 49.5 percent in 2000; performance to 46.1 percent in 1999, home purchase loans for lower-income and Freddie Mac’s performance was 42.6 exceeding its previous high by 3.2 percentage percent in 1997, 42.9 percent in 1998, 46.1 points. Fannie Mae’s performance in 1999 percent in 1999, and 49.9 percent in 2000. 255 To separate out the effects of changes in was 45.9 percent, which, for the first time, counting rules that took effect in 2001, this section slightly lagged Freddie Mac’s performance in also compares performance in 2001 to estimated 254 Performance for the 1993–95 period was performance in 2000 if the 2001 counting rules had that year. discussed in the October 2000 rule. been in effect in that year. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00105 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63684 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00106 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.017 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63685
Both GSEs exhibited sharp gains in goal reflected a difference in the counting rules income goal and the special affordable goal performance in 2000—Fannie Mae’s applicable to the two GSEs that was enacted that took effect in 2001, the GSEs are allowed performance increased by 3.6 percentage by Congress; if the same counting rules were to exclude loans with missing borrower points, to a record level of 49.5 percent, applied to both GSEs (that is, Freddie Mac income from the denominator if the property while Freddie Mac’s performance increased did not receive the 1.35 Temporary is located in a below-median income census even more, by 3.8 percent percentage points, Adjustment Factor), Fannie Mae’s tract. This exclusion is subject to a ceiling of which also led to a record level of 49.9 performance would have exceeded Freddie 1 percent of total owner-occupied units percent. Fannie Mae’s performance was 51.5 Mac’s performance, by 51.5 percent to 50.5 financed. The enterprises are also allowed to percent in 2001, 51.8 percent in 2002, and percent. exclude single-family rental units with 52.3 percent in 2003; Freddie Mac’s In 2002, Freddie Mac’s performance on the missing rental information from the performance was 53.2 percent in 2001, 50.5 low mod-goal (50.5 percent) fell short of denominator in calculating performance for percent in 2002, and 51.2 percent in 2003. Fannie Mae’s performance (51.8 percent), these two goals; there is no ceiling or However, as discussed below, using even though Freddie Mac had the advantage restriction to properties located in below- consistent accounting rules for 2000–03, each of the Temporary Adjustment Factor. The gap median income census tracts for this GSE’s performance in 2001–03 was below its would have been wider without this factor, exclusion of single-family rental units. No performance in 2000. and in fact Freddie Mac’s performance would single-family loans can be excluded from the The official figures for low-mod goal have been short of the goal, at 49.2 percent. denominator in calculating performance on performance presented above differ from the This same pattern prevailed in 2003, when the underserved areas goal—that is, if a GSE corresponding figures presented by Fannie Freddie Mac’s performance on this goal (51.2 does not have sufficient information to Mae and Freddie Mac in their Annual percent) was significantly below Fannie determine whether or not a property is Housing Activity Reports to HUD by 0.2–0.3 Mae’s performance (52.3 percent), even located in an underserved area, all units in percentage point in both 1996 and 1997, though Fannie Mae did not have the such a property are included in the reflecting minor differences in the advantage of the Temporary Adjustment denominator, but not in the numerator, in application of counting rules. These Factor. The gap in performance between the calculating performance on this goal. differences also persisted for Freddie Mac for GSEs would have been much wider without • Missing data and proxy rents for 1998–2000, but the goal percentages shown this factor, as Freddie Mac’s performance multifamily properties. In the past, if a GSE above for Fannie Mae for these three years would again have fallen short of the goal, at lacked data on rent for rental units in are the same as the results reported by Fannie 48.4 percent. multifamily properties whose mortgages it Mae to the Department. Fannie Mae reported 2. Changes in the Goal Counting Rules for purchased, such units were included in the its performance in 2001 as 51.6 percent and 2001–03 denominator, but not in the numerator, in Freddie Mac reported its performance as 53.6 A number of changes in the counting rules calculating goal performance. Since some of percent—both were slightly above the underlying the calculation of low- and these units likely would have qualified for corresponding official figures of 51.5 percent moderate-income goal performance took one or more of the housing goals, this rule and 53.4 percent, respectively. For 2002, effect beginning in 2001, as follows: lowered goal performance. Under the new Fannie Mae’s reported performance was the • Bonus points for multifamily and single- counting rules that took effect in 2001, if rent same as reported by HUD (51.8 percent), family rental properties. During the 2001–03 is missing for multifamily units, a GSE may while Freddie Mac’s reported performance period the Department awarded ‘‘bonus estimate ‘‘proxy rents,’’ and, up to a ceiling was 51.3 percent, slightly above HUD’s points’’ (double credit in the numerator) for of 5 percent of total multifamily units official figure of 50.5 percent. For 2003, goal-qualifying units in small (5–50 unit) financed, may apply these proxy rents in Fannie Mae’s reported performance on this multifamily properties and, above a determining whether such units qualify for goal was 51.8 percent, somewhat below threshold, 2–4 unit owner-occupied the low- and moderate income goal and HUD’s official figure of 52.3 percent, while properties whose loans were purchased by special affordable goal. If such proxy rents Freddie Mac’s reported performance (51.1 the GSEs. By letters dated December 24, cannot be estimated, these multifamily units percent) was essentially the same as HUD’s 2003, the Department notified the GSEs that are excluded from the denominator in official figure of 51.2 percent. these bonus points would not be in effect calculating performance under these goals. Fannie Mae’s performance on the Low- and after December 31, 2003. No multifamily loans can be excluded from Moderate-Income Goal was in the range • Freddie Mac’s Temporary Adjustment the denominator in calculating performance between 44 percent and 46 percent between Factor. As part of the Consolidated on the underserved areas goal—that is, if a 1996 and 1999, but jumped sharply in just Appropriations Act of 2000, Congress GSE does not have sufficient information to one year, from 45.9 percent in 1999 to 49.5 required the Department to award 1.35 units determine whether or not a property is percent in 2000. Freddie Mac’s performance of credit for each unit financed in ‘‘large’’ located in an underserved area, all units in was in the range between 41 percent and 43 multifamily properties (i.e., those with 51 or such a property are included in the percent between 1996 and 1998, and then more units) in the numerator in calculating denominator, but not in the numerator, in rose to 46.1 percent in 1999 and 49.9 percent performance on the housing goals for Freddie calculating performance on this goal. in 2000. As discussed above, official Mac for 2001–03.256 This ‘‘temporary • Purchases of certain government-backed performance rose for both GSEs in 2001–02, adjustment factor’’ (TAF) did not apply to loans. Prior to 2001, purchases of but this was due to one-time changes in the goal performance for Fannie Mae during this government-backed loans were not taken into counting rules—abstracting from counting period. By letters dated December 24, 2003, account in determining performance on the rule changes, performance fell for both GSEs. the Department notified Freddie Mac that GSEs’ low- and moderate-income and Fannie Mae’s performance on the Low- and this factor would not be in effect after underserved area housing goals. That is, all Moderate-Income Goal surpassed Freddie December 31, 2003. such loans were excluded from both the Mac’s in every year through 1998. This • Missing data for single-family properties. numerator and the denominator in pattern was reversed in 1999, as Freddie Mac In the past, if a GSE lacked data on rent for calculating goal performance on these two surpassed Fannie Mae in goal performance rental units or on borrower income for goals, and in accordance with Section for the first time, though by only 0.2 owner-occupied units in single-family 1333(b)(1)(A) of the Federal Housing percentage point. This improved relative properties whose mortgages it purchased, Enterprises Financial Safety and Soundness performance of Freddie Mac was due to its such units were included in the Act of 1992, purchases of only certain increased purchases of multifamily loans, as denominator, but not in the numerator, in government-backed loans were included in it re-entered that market, and to increases in calculating goal performance. Since some of determining performance on the GSEs’ the goal-qualifying shares of its single-family these units likely would have qualified for special affordable goals. In October 2000 the mortgage purchases. Freddie Mac’s one or more of the housing goals, this rule Department took steps to encourage the performance also slightly exceeded Fannie lowered goal performance. Under the new enterprises to play more of a role in the Mae’s performance in 2000, 49.9 percent to counting rules for the low- and moderate- secondary market for several types of 49.5 percent. Freddie Mac’s official government-backed loans where it appeared performance also exceeded Fannie Mae’s 256 See Congressional Record, December 15, 2000, that greater GSE involvement could increase official performance in 2001, but this pp. H12295–96. the liquidity of such mortgages. Home equity
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00107 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63686 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
conversion mortgages (HECMs) were under the Rural Housing Service’s Single and 2001–03 are somewhat of an ‘‘apples-to- developed in the late-1980s by the Federal Family Housing Guaranteed Loan Program oranges comparison.’’ For this reason, the Housing Administration (FHA); these may also count toward all of the housing Department has calculated what performance mortgages allow senior citizens to draw on goals.257 would have been in 2000 under the 2001–03 the equity in their homes to obtain monthly 3. Effects of Changes in the Counting Rules rules; this may be compared with official payments to supplement their incomes. Thus on Goal Performance in 2001–03 performance in 2001–03—an ‘‘apples-to- purchases of FHA-insured HECMs now count apples comparison.’’ HUD has also calculated toward the low- and moderate-income Because of the changes in the low- and moderate-income goal counting rules that what performance would have been in 2001– housing goals if the mortgagor’s income is 03 under the 1996–2000 rules; this may be less than median income for the area. took effect in 2001, direct comparisons compared with official performance in Similarly, purchases of mortgages on between official goal performance in 2000 properties on tribal lands insured under 2000—an ‘‘oranges-to-oranges comparison.’’ These comparisons are presented in Table FHA’s Section 248 program or HUD’s Section 257 Prior to the October 2000 rule, purchases of A.9. 184 program may qualify for the GSEs’ these government-backed mortgages were only housing goals. And purchases of mortgages eligible for credit under the special affordable goal. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00108 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63687
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00109 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.018 63688 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Specifically, Table A.9 shows performance This section breaks down the effects of these renters. Bonus points did not apply to under the low- and moderate-income goal in changes on goal performance for both GSEs; purchases of mortgages for owner-occupied three ways. Baseline A represents results are shown in Table A.9. 1-unit properties, for investor-owned 1–4 performance under the counting rules in • Freddie Mac. The largest impact of the unit properties, and for large (more than 50 effect in 1996–2000. Baseline B incorporates counting rule changes on Freddie Mac’s goal units) multifamily properties, although as the technical changes in counting rules— performance was due to the application of also discussed above, a ‘‘temporary changes in the treatment of missing data the temporary adjustment factor for adjustment factor’’ applied to Freddie Mac’s (including use of proxy rents), and eligibility purchases of mortgages on large multifamily purchases of qualifying mortgages on large for the goals of certain government-backed properties, as enacted by Congress; this multifamily properties. loans. Baseline C incorporates in addition to added 2.7 percentage points to goal Bonus points for small multifamily the technical changes the bonus points and, performance in 2001, as shown in Table A.9. properties. Each unit financed in a small for Freddie Mac, the temporary adjustment Bonus points for purchases of mortgages on multifamily property that qualified for any of factor. Baseline B corresponds to the small multifamily properties added 1.5 the housing goals was counted as two units counting approach proposed in this rule to percentage points to performance, and bonus in the numerator (and one unit in the take effect in 2005. Boldface figures under points for purchase of mortgages on owner- denominator) in calculating goal performance Baseline A for 1999–2000 and under Baseline occupied 2–4 unit rental properties added 1.4 for that goal. For example, if a GSE financed C for 2001–03 indicate official goal percentage points to performance. The a mortgage on a 40-unit property in which 10 performance, based on the counting rules in remaining impact (0.5 percentage point) was of the units qualified for the low- and effect in those years—e.g., for Fannie Mae, due to technical changes in counting rules— moderate-income goal, 20 units would be 45.9 percent in 1999, 49.5 percent in 2000, primarily, the exclusion of single-family entered in the numerator and 40 units in the 51.5 percent in 2001, 51.8 percent in 2002, units with missing information from the denominator for this property in calculating and 52.3 percent in 2003. denominator in calculating goal performance. goal performance. • Performance on the Low- and Moderate- Credit for purchases of qualifying Small multifamily bonus points thus Income Goal under 1996–2000 Counting government-backed loans played a minor role encouraged the GSEs to play a larger role in Rules Plus Technical Changes. If the in determining Freddie Mac’s goal this market, and also to purchase mortgages ‘‘Baseline B’’ counting approach had been in performance. These same patterns also on such properties in which large shares of effect in 2000–03 and the GSEs’ had appeared in 2002. But in 2003, bonus points the units qualified for the housing goals. purchased the same mortgages that they for purchases of low-mod mortgages on Some evidence may be gleaned from the data actually did purchase in those years, both single-family rental properties had a larger provided to HUD by the GSEs for 2001–03. Fannie Mae and Freddie Mac would have impact on Freddie Mae’s low-mod goal Fannie Mae financed 37,403 units in small surpassed the low- and moderate-income performance than Freddie Mac’s temporary multifamily properties in 2001 that were goal in 2000 and fallen short in 2001, 2002, adjustment factor. eligible for the low- and moderate-income and 2003. Specifically, Fannie Mae’s • Fannie Mae. The temporary adjustment goal, 58,277 such units in 2002, and 214,619 performance would have been 51.3 percent factor applies to Freddie Mac’s goal such units in 2003, as compared with only in 2000, 49.2 percent in 2001, 49.0 percent performance, but not to Fannie Mae’s 7,196 such units financed in 2000. Small in 2002, and 48.7 percent in 2003. Freddie performance, thus counting rule changes had multifamily properties also accounted for a Mac’s performance would have been 50.6 less impact on its performance than on greater share of Fannie Mae’s multifamily percent in 2000, 47.7 percent in 2001, 46.1 Freddie Mac’s performance in 2001. The business in 2001–03—7.4 percent of total percent in 2002, and 45.0 percent in 2003. largest impact of the counting rule changes multifamily units financed in 2001, 13.2 • Performance on the Low- and Moderate- on Fannie Mae’s goal performance was due percent in 2002, and 28.6 percent in 2003, up Income Goal under 2001–2003 Counting to the application of bonus points for from 2.5 percent in 2000. However, HUD’s Rules. If the 2001–03 counting rules had also purchases of mortgages on owner-occupied 2000 rule reported information from the 1991 been in effect in 2000 and the GSEs’ had 2–4 unit rental properties, which added 1.6 Residential Finance Survey that small purchased the same mortgages that they percentage points to performance, and for multifamily properties accounted for 37 actually did purchase in those years (i.e., purchases of mortgages on small multifamily percent of all multifamily units, thus Fannie abstracting from any behavioral effects of properties, which added 0.7 percentage point Mae was still less active in this market than ‘‘bonus points,’’ for example), both GSEs to performance. The remaining impact (1.3 in the market for large multifamily would have substantially surpassed the low- percentage points) was due to technical properties.259 and moderate-income goal in all four years, changes—primarily, the exclusion of single- Within the small multifamily market, there but both GSEs’ performance figures would family units with missing information from was no evidence that Fannie Mae targeted have deteriorated somewhat from 2000 to the denominator in calculating goal affordable properties to a greater extent in 258 2001, and, for Freddie Mac, from 2001 to performance. Credit for purchases of 2001–03 than in 2000. That is, 87 percent of 2002 and 2003. Specifically, Fannie Mae’s qualifying government-backed loans and the Fannie Mae’s small multifamily units ‘‘Baseline C’’ performance would have been use of proxy rent for multifamily properties qualified for the low- and moderate-income 52.5 percent in 2000, 51.5 percent in 2001, played a minor role in determining Fannie goal in 2000; this fell to 75 percent in 2001, 51.8 percent in 2002, and 52.3 percent in Mae’s goal performance. These same patterns rose to 89 percent in 2002, and then declined 2003. Freddie Mac’s performance would have also appeared in 2002 for Fannie Mae, but for to 82 percent in 2003. been 55.1 percent in 2000, surpassing its 2003 bonus points for purchases of low-mod Freddie Mac financed 50,299 units in small official performance level of 53.2 percent in mortgages on small multifamily properties multifamily properties in 2001 that were 2001, 50.5 percent in 2002, and 51.2 percent had more impact on performance than bonus eligible for the low- and moderate-income points for single-family rental properties. in 2003. Measured on this consistent basis, goal, 22,255 such units in 2002, and 177,561 then, Fannie Mae’s performance fell by 1.0 4. Bonus Points for the Low- and Moderate- such units in 2003, as compared with only percentage point in 2001, and Freddie Mac’s Income Goal such units financed in 2000. Small by 1.9 percentage points in 2001 and an As discussed above, the Department multifamily properties also accounted for a additional 2.0 percentage points in 2002–03. established ‘‘bonus points’’ to encourage the significantly greater share of Freddie Mac’s These reductions were primarily due to GSEs to step up their activity in 2001–03 in multifamily business in 2001–2003—16.1 2001–03 being years of heavy refinance two segments of the mortgage market—the percent of total multifamily units financed in activity. small (5–50 unit) multifamily mortgage 2001, 7.5 percent in 2002, and 25.4 percent Details of Effects of Changes in Counting market, and the market for mortgages on 2– in 2003, up from 1.8 percent in 2000. Rules on Goal Performance in 2001–03. As 4 unit properties where 1 unit is owner- Within the small multifamily market, there discussed above, counting rule changes that occupied and 1–3 units are occupied by was some evidence that Freddie Mac targeted took effect in 2001 had significant positive affordable properties to a greater extent in impacts on the performance of both GSEs on 258 Exclusion of loans with missing information 2001–2002 than in 2000. That is, 87 percent the low- and moderate-income goal in that had a greater impact on Fannie Mae’s goal year—3.8 percentage points for Fannie Mae, performance than on Freddie Mac’s goal 259 Federal Register, October 31, 2000, Footnote and 6.0 percentage points for Freddie Mac. performance. 145, p. 65141.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00110 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63689
of Freddie Mac’s small multifamily units faster rate than its OO24 business in 2003, For scoring loans purchased by the GSEs qualified for the low- and moderate-income thus the share of its business accounted for year-by-year from 1993 through 2002, area goal in 2000; this rose to 96 percent in 2001, by OO24s declined to 3 percent last year. median income estimates produced by HUD’s but declined back to 87 percent in 2002 and As for Fannie Mae, within the OO24 Economic and Market Analysis Division were 2003. market there was no evidence that Freddie used. An example will illustrate the In summary, then, there is strong evidence Mac targeted affordable properties to a estimation procedure. To generate the area that bonus points for small multifamily greater extent in 2001–03 than in 2000. That median income estimates that were used to properties had an impact on Fannie Mae’s is, 68–69 percent of Fannie Mae’s OO24 units score GSE loans in 2002, data from the 1990 role in this market in 2001–2003 and an even qualified for the low- and moderate-income census on 1989 area median incomes were larger impact on Freddie Mac’s role in this goal in each year from 2000 through 2002; adjusted to 2002 using Bureau of Labor market. In addition, Fannie Mae has this decreased to 64 percent in 2003. Statistics survey data on rates of cha nge in announced a program to increase its role in 5. Effects of 2000 Census on Scoring of Loans average incomes for MSAs and counties this market further in future years.260 Toward the Low- and Moderate-Income between 1989 and 1999, data from the Bonus points for single-family rental Housing Goal Census Bureau’s Current Population Survey properties. Above a threshold, each unit on rates of change in median family incomes financed in a 2–4 unit property with at least Background. Scoring of housing units under the Low- and Moderate-Income for the nine Census Divisions between 1989 one owner-occupied unit (referred to as Housing Goal is based on data for mortgagors’ and 2000, and an assumed 4.0 percent per ‘‘OO24s’’ below) that qualified for any of the incomes for owner-occupied units, rents for year inflation factor between 2000 and housing goals was counted as two units in rental units, and area median incomes, as 2002.262 263 the numerator (and one unit in the follows: 2005 Procedure. Relative to the above denominator) in calculating goal performance For single-family owner-occupied units: procedure, scoring of loans purchased by the for that goal in 2001–2003. The threshold The mortgagors’ income at the time of GSEs in and after 2005 will be affected by was equal to 60 percent of the average mortgage origination. two factors. First, the Economic and Market number of such qualifying units over the The median income of an area specified as Analysis Division has begun to incorporate previous five years. For example, Fannie Mae follows: (i) For properties located in data from the 2000 census into its procedure financed an average of 50,030 low- and Metropolitan Statistical Areas (MSAs), the for estimating annual area median incomes moderate-income units in these types of area is the MSA; and (ii) for properties and American Community Survey data are properties between 1996 and 2000, and located outside of MSAs, the area is the becoming available at increasingly finer 101,423 such units in 2001. Thus Fannie Mae county or the non-metropolitan portion of the levels of geographical detail for use in annual received 71,405 bonus points in this area in State in which the property is located, updating. Beginning in 2005 Bureau of Labor 2001—that is, 101,423 minus 60 percent of whichever has the larger median income, as Statistics data on rates of inflation in average 50,030. So 172,828 units were entered in the of the year of mortgage origination (which wages will not be used. For 2005, the numerator for these properties in calculating may be for the current year or a prior year). procedure for estimating area median low- and moderate-income goal performance. For rental units in single-family properties incomes will be to adjust 2000 census data Single-family rental bonus points thus with rent data are available (assuming no on 1999 area median incomes to 2003 using encouraged the GSEs to play a larger role in income data available for actual or data from the Census Bureau’s American this market, and also to purchase mortgages prospective tenants): Community Survey (ACS) on rates of change on such properties in which large shares of The unit rent (or average rent for units of in average incomes for States between 1999 the units qualified for the housing goals. As the same type) at the time of mortgage for small multifamily bonus points, again origination. and 2003, with a further adjustment to 2005 some evidence may be gleaned from the data The area median income as specified for based on an appropriate annual inflation 264 provided to HUD by the GSEs for 2001–03. single-family owner-occupied units. factor. Increasingly more detailed ACS Fannie Mae financed 175,103 units in For rental units in multifamily properties data will be available and will be used in OO24s in 2001 that were eligible for the low- where rent data are available: subsequent years, as ACS estimates for and moderate-income goal, 229,632 such The unit rent (or the average rent for units metropolitan and micropolitan areas and units in 2002, and 355,994 such units in of the same type) at the time of mortgage counties become available. 2003, well above the 77,930 units financed in acquisition by the GSE. The second factor is the Office of 2000. However, with the refinance boom, The area median income as specified for Management and Budget’s June, 2003, re- Fannie Mae’s total single-family business single-family owner-occupied units, but as of increased at approximately the same rate as the year the GSE acquired the mortgage. 262 The procedure is explained in detail in annual its OO24 business in 2001–03, thus the share For rental units in multifamily properties releases entitled ‘‘HUD Methodology for Estimating of its business accounted for by OO24s was where rent data are not available, the GSE FY [year] Median Family Incomes’’ for years 1993 the same in 2001–03 as in 2000—4 percent. may apply HUD-estimated rents which are through 2002, issued by the Economic and Market Analysis Division, Office of Economic Affairs, Within the OO24 market, there was no based on the following area data: The median rent in the census tract where PD&R, U.S. Department of Housing and Urban evidence that Fannie Mae targeted affordable Development. the property is located, as of the most recent properties to a greater extent in 2001–03 than 263 The procedure applicable to the decennial in 2000. That is, approximately 55–60 decennial census. census data used to generate estimated rents is percent of Fannie Mae’s OO24 units qualified The area median income as specified for explained in connection with data used to define for the low- and moderate-income goal in single-family owner-occupied units, but as of Underserved Areas in Appendix B. each of these three years. the most recent decennial census. 264 Transition from the 2002 methodology to the Freddie Mac financed 96,050 units in Thus, scoring loans under the Low- and 2005 methodology is occurring in stages in 2003 OO24s in 2001 that were eligible for the low- Moderate-Income Goal requires a data series and 2004. To generate the area median income and moderate-income goal, 146,222 such showing annual median incomes for MSAs, estimates used to score GSE loans in 2003, data from the 2000 census on 1999 area median incomes units in 2002, and 154,535 such units in non-metropolitan counties, and the non- metropolitan portions of states; and were adjusted to 2001 using Bureau of Labor 2003, as compared with the 49,993 units Statistics survey data on rates of change in average financed in 2000. However, Freddie Mac’s decennial census data on median incomes for 261 incomes for MSAs and counties between 1999 and total single-family business increased at census tracts. 2000, data on rates of change in median incomes approximately the same rate as its OO24 for the United States and individual States between business in 2001–02, thus the share of its 261 In New England, MSAs were defined through 1999 and 2001 from Census Bureau’s Current business accounted for by OO24s was the mid-2003 in terms of Towns rather than Counties, Population Survey and American Communities same in 2002 as in 2000—4 percent. And its and the portion of a New England county outside Survey, and an assumed 3.5 percent per year of any MSA was regarded as equivalent to a county inflation factor between 2001 and 2003. (See ‘‘HUD total single-family business increased at a in establishing the metropolitan or non- Methodology for Estimating FY 2003 Median metropolitan location of a property. The MSA Family Incomes,’’ issued by the Economic and 260 ‘‘Fannie Courting Multifamily Sellers; Small definitions established by the Office of Management Market Analysis Division, op cit.) A similar Banks Balking,’’ American Banker, January 13, and Budget (OMB) in June, 2003 defined MSAs in procedure has been used to generate area median 2003, p. 1. New England in terms of counties. income estimates for scoring GSE loans in 2004.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00111 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63690 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
specification of MSA boundaries based on estimates were 1999 census medians trended were used further in estimating the share of analysis of 2000 census data.265 by three-fourths of the 4.0 percent annual loans originated in metropolitan areas that Analysis. For purposes of specifying the trending factor (to adjust the figures from would be eligible to score toward the Low- level of the Low- and Moderate-Income mid-1999 to April 1, 2000). For 2001, the and Moderate-Income Housing Goal, from Housing Goal, HUD developed a estimates were based on one-and-three- HMDA data. The results of the retrospective methodology for scoring loans purchased by fourths years of trending, since no data GSE analysis are provided in Table A.10. The the GSEs in past years through 2002 as would have been available to use for results of the GSE-HMDA comparative though the re-benchmarking of area median updating. The 2002 estimates would have analysis are presented in the next section. income estimates to the 2000 census and the used one year of data and 1.75 years of Table A.10 shows three sets of estimates 2003 re-designation of MSAs had been in trending. The 2003 estimates would have for each GSE, based respectively on the effect and HUD had been using an ACS-based used two years of data plus 1.75 years of counting rules in place in 2001–2002 (but estimation procedure at the time the disregarding the bonus points and Temporary estimates for these years were prepared. For trending. Area median incomes from 1989 to 1999 were estimated based on trend-lines Adjustment Factor), on the addition of 2000 this purpose, HUD created a series of annual census re-benchmarking, and finally on the estimates of median incomes for MSAs, non- between 1989 and 1999 census data. The addition of both 2000 census re- metropolitan counties, and the non- 2003 OMB MSA designations were applied. benchmarking and 2003 MSA specification. metropolitan portions of states. For 2000, the The resulting estimates of area median incomes for MSAs, non-metropolitan Re-benchmarking occurred to adjust for some differences between Census 1990 and Census 265 counties, and the non-metropolitan parts of HUD has deferred application of the 2003 2000 tracts. MSA specification to 2005, pending completion of States, were used to re-score loans purchased the present rulemaking process. by the GSEs between 1999 and 2002, and BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00112 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63691
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00113 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.019 63692 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
6. GSEs Compared With the Primary Mac’s purchases compared with 26.8 percent percent of loans originated for African- Conventional Conforming Mortgage Market of Fannie Mae’s purchases and 27.6 percent American and Hispanic first-time This section and the next five sections of market originations. homebuyers, or about one-half of their share (Sections E.7 to E.12) provide a detailed 2. In general, Fannie Mae’s affordable (57 percent) of all home purchase loans in analysis of the extent to which the GSEs’ loan lending performance has been better than that market. purchases mirror or depart from the patterns Freddie Mac’s. But like Freddie Mac, Fannie 6. The GSEs’ small share of the first-time found in the primary mortgage market. As in Mae’s average performance during past homebuyer market could be due to the Section C.5, the GSEs’ affordable lending periods (e.g., 1993–2003, 1996–2003, 1999– preponderance of high (over 20 percent) performance is also compared with the 2003) has been below market levels. downpayment loans in their mortgage performance of depository lenders such as However, it is encouraging that Fannie Mae purchases. commercial banks and thrift institutions. markedly improved its affordable lending b. Specific Findings on GSE Performance in Dimensions of lending considered include performance relative to the market during the Single-family Market 2001, 2002, and 2003, the first three years the three ‘‘goals-qualifying’’ categories— This section presents 17 specific findings under the higher housing goal targets that special affordable borrowers, less-than- from the analyses reported in Sections E.7 HUD established in the GSE Final Rule dated median income borrowers, and underserved through 12; they are grouped under the October 2000. areas. The special affordable category following five topic-headings: Over this three-year period, Fannie Mae consists mainly of very-low-income (b.1) Longer-term Performance of the GSEs; led the primary market in funding special borrowers, or borrowers who have an annual (b.2) Performance of the GSEs During affordable and low-mod loans but lagged the income less than 60 percent of area median Recent Years; market in funding underserved areas loans. income. Because this category is more (b.3) The GSEs’ Funding First-time In 2003, Fannie Mae’s increased performance targeted than the broadly-defined less-than- Homebuyer Loans; median-income (or low-mod) category, the placed it significantly above the special (b.4) Performance of the GSEs Based on discussion below will often focus on the affordable market (a 17.1 percent share for Total (Home Purchase and Refinance) Loans; special affordable category as well as the Fannie Mae compared with a 15.9 percent (b.5) GSE Market Shares; and, underserved areas category which adds a share for the market) and the low-mod (b.6) Additional Findings. market (a 47.0 percent share for Fannie Mae neighborhood dimension (low-income and (b.1) Longer-Term Performance of the GSEs high-minority census tracts) to the analysis. compared with a 44.6 percent share for the This section will also compare the market). However, Fannie Mae continued to The longer-run performance of the GSEs is performance of Fannie Mae and Freddie Mac lag the underserved areas market in 2003 (a examined between 1993 and 2003 (which in funding first-time homebuyers with that of 26.8 percent share for Fannie Mae compared covers the period since the housing goals primary lenders in the conventional with a 27.6 percent share for the market). In were put into effect) and between 1996 and conforming market. this case, which is referred to in the text as 2003 (which covers the period under the The remainder of this introductory section the ‘‘purchase year’’ approach, Fannie Mae’s current definitions of the housing goals). Of E.6 provides a list of the major and specific performance is based on comparing its the two borrower-income goals, the analysis findings which are presented in detail in the purchases of all loans (both seasoned loans below will typically focus on the special following Sections E.7 through 12. Sections and newly-originated mortgages) during a affordable category, which is a more targeted 7 and 8 define the primary mortgage market particular year with loans originated in the category than the rather broadly defined low- and discuss some technical issues related to market in that year. When Fannie Mae’s and moderate-income category. the use of the GSE and HMDA data. Sections performance is measured on an ‘‘origination (1) Since the early nineties, the mortgage 8 and 9 compare the GSEs’ performance with year’’ basis (that is, allocating Fannie Mae’s industry has introduced new affordable market performance for home purchase and purchases in a particular year to the year that lending programs and has allowed greater first-time homebuyer loans, while Section 10 the purchased loan was originated), Fannie flexibility in underwriting lower-income does the same for total single family loans Mae also led the 2003 market in funding loans. There is evidence that these programs are paying off in terms of more mortgages for (that is, refinance loans and home purchase special affordable and low- and moderate- low-income and minority borrowers. As loans). Section 11 examines GSE purchases income loans, and lagged the market in noted earlier, Fannie Mae and Freddie Mac in individual metropolitan areas. Following funding underserved area loans. have played an active role in this upsurge of these analyses, Section 12 examines the 3. Both Fannie Mae and Freddie lag the affordable lending, as indicated by the high overall market share of the GSEs in important conventional conforming market in funding growth rates of their goals-qualifying submarkets such as first-time homebuyers. first-time homebuyers, and by a rather wide business. margin. Between 1999 and 2001, first-time a. Main Findings on GSEs’ Performance in • Between 1993 and 2003, the GSEs’ homebuyers accounted for 27 percent of each the Single-family Market purchases of home loans in metropolitan GSE’s purchases of home loans, compared There are six main findings from this areas increased by 60 percent.267 Their with 38 percent for home loans originated in purchases of home loans for the three analysis concerning the GSEs’ purchases of the conventional conforming market. single-family-owner mortgages: housing goals increased at much higher 4. The GSEs have accounted for a rates—287 percent for special affordable 1. While Freddie Mac has improved its significant share of the total (government as affordable lending performance in recent loans, 156 percent for low- and moderate- well as conventional) market for home income loans, and 121 percent for loans in years, it has consistently lagged the purchase loans, but their market share for conventional conforming market in funding underserved census tracts. each of the affordable lending categories (e.g., (2) Both Fannie Mae and Freddie Mac have affordable home purchase loans for special low-income borrowers and census tracts) has affordable and low-moderate-income improved their purchases of affordable loans been less than their share of the overall since the housing goals were put in place, as borrowers and underserved neighborhoods market. 266 indicated by the increasing share of their targeted by the housing goals. In 2003, its 5. The GSEs also account for a very small performance on the underserved areas goal business going to the three goals-qualifying share of the market for important groups such categories. (See Table A.15 in Section E.9.) was particularly low relative to both the as minority first-time homebuyers. performances of Fannie Mae and the market; Considering the total mortgage market (both 267 Throughout this analysis, the terms ‘‘home in that year, underserved area loans government and conventional loans), it is accounted for only 24.0 percent of Freddie loan’’ and ‘‘home mortgage’’ will refer to a ‘‘home estimated that the GSEs purchased only 14 purchase loan,’’ as opposed to a ‘‘refinance loan.’’ percent of loans originated between 1999 and As noted earlier, the mortgage data reported in this 266 The ‘‘affordable lending performance’’ of 2001 for African-American and Hispanic paper are for metropolitan areas, unless stated Fannie Mae and Freddie Mac refers to the first-time homebuyers, or less than half of otherwise. Restricting the GSE data to metropolitan performance of the GSEs in funding loans for low- their share (42 percent) of all home purchase areas is necessary to make it comparable with the income and underserved borrowers through their HMDA-reported conventional primary market data, purchase (or guarantee) of loans originated by loans originated during that period. which is more reliable for metropolitan areas. The primary lenders. It does not, of course, imply that Considering the conventional conforming analysis of first-time homebuyers in Sections E.9 the GSEs themselves are lenders originating loans market and the same time period, it is and E.12 cover both metropolitan and non- in the primary market. estimated that the GSEs purchased only 31 metropolitan areas.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00114 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63693
• Between 1992 and 2003, the special the ‘‘Freddie-Mac-to-market’’ ratio was also Mae in purchasing affordable home loans affordable share of Fannie Mae’s business 0.91. (although only by a very slight margin). almost tripled, rising from 6.3 percent to 17.1 • During the same period, underserved • In 2000, Freddie Mac’s sharper increases percent, while the underserved areas share area loans accounted for 24.7 percent of in special affordable and low-mod purchases increased more modestly, from 18.3 percent Fannie Mae’s purchases, 23.1 percent of further reduced the ‘‘Fannie-Mae-to-Freddie- to 26.8 percent. The figures for Freddie Mac Freddie Mac’s purchases, and 26.2 percent of Mac’’ ratios for these two categories to 0.90 are similar. The special affordable share of loans originated in the market; the ‘‘Fannie- and 0.96, respectively. Fannie Mae’s sharper Freddie Mac’s business rose from 6.5 percent Mae-to-market’’ ratio was 0.94 and the increase in underserved areas funding to 15.6 percent, while the underserved areas ‘‘Freddie-Mac-to-market’’ ratio was only resulted in the ‘‘Fannie-Mae-to-Freddie-Mac’’ share also increased but more modestly, from 0.88.269 ratio rising from slightly below one (0.98) in 18.6 percent to 24.0 percent. (5) After experiencing declines from 1997 1999 to 1.06 in 2000. (3) While both GSEs improved their to 1999, Fannie Mae’s affordable lending • Fannie Mae’s stronger performance performance, they have lagged the primary performance improved between 2000 and during 2001–2003 returned the ‘‘Fannie-Mae- market in providing affordable loans to low- 2003. to-Freddie-Mac’’ ratios for special affordable income borrowers and underserved • After declining from 23.0 percent in and low-mod loans to above one (1.10 and neighborhoods. Freddie Mac’s average 1997 to 20.4 percent in 1999, the share of 1.09 respectively), indicating better performance, in particular, fell far short of Fannie Mae’s purchases financing properties performance for Fannie Mae in 2003. The market performance during the 1990s. Fannie in underserved areas jumped by three ‘‘Fannie-Mae-to-Freddie-Mac’’ ratio for the Mae’s average performance was better than percentage points to 23.4 percent in 2000, underserved area category increased to 1.12 Freddie Mac’s during the 1993–2003 period and then increased further to 26.7 percent in by 2003. as well as during the 1996–2003 period, 2002 and 26.8 percent in 2003. (8) While Freddie Mac has consistently which covers the period under HUD’s • After declining from 13.2 percent in improved its performance relative to the currently-defined housing goals. 1998 to 12.5 percent in 1999, the share of market, it continued to lag the market in • Between 1993 and 2003, 12.2 percent of Fannie Mae’s purchases going to special funding affordable home loans during 2001– Freddie Mac’s mortgage purchases were for affordable loans rebounded to 13.3 percent in 2003. • special affordable borrowers, compared with 2000, 14.9 percent in 2001, 16.3 percent in Unlike Fannie Mae, Freddie Mac had not 13.3 percent of Fannie Mae’s purchases, 15.4 2002 and 17.1 percent in 2003. made any progress through 1997 in closing percent of loans originated by depositories, (6) Freddie Mac’s performance on the two its gap with the market. The ‘‘Freddie Mac- and 15.5 percent of loans originated in the borrower-income categories improved to-market’’ ratio for the special affordable conventional conforming market (without between 2000 and 2002, but not as much as category actually declined from 0.63 in 1992 estimated B&C loans).268 Fannie Mae’s performance. Freddie Mac’s to 0.59 in 1996. But Freddie Mac’s sharp • Considering the underserved areas performance on the underserved areas improvement in special affordable purchases category for the 1996–2003 period, 22.0 category increased substantially between resulted in the ‘‘Freddie-Mac-to-market’’ ratio percent of Freddie Mac’s purchases financed 2001 and 2002, but then declined between rising to 0.89 by 2000. After declining from properties in underserved neighborhoods, 2002 and 2003. 0.84 in 1992 to 0.79 in 1997, the ‘‘Freddie- compared with 24.0 percent of Fannie Mae’s • The share of Freddie Mac’s single- Mac-to-market’’ ratio for underserved areas purchases, 25.1 percent of loans originated family-owner business going to special had risen only modestly to 0.84 by the year by depositories, and 25.7 percent of loans affordable home loans increased from 9.2 in 2000. Thus, Freddie Mac’s improvements originated in the conventional conforming 1997 to 14.7 percent in 2000 before falling to prior to 2001 allowed it to close its gap with market. 14.4 percent in 2001 and rising to 15.8 the market, mainly for the special affordable percent in 2002 and 15.6 percent in 2003. category where its gap had been the widest. (b.2) Performance of the GSEs During Recent • • During 2001, 2002 and 2003, Freddie Years Freddie Mac’s purchases of underserved area loans increased at a modest rate from Mac continued to close its gap with the The recent performance of the GSEs is 19.7 percent in 1997 to 22.3 percent in 2001, market on the special affordable and low- examined for the four-year period between before jumping to 25.8 percent in 2002 and mod categories. By 2003, these ‘‘Freddie- 1999 and 2003 and then for 2001, 2002 and then dropping to 24.0 percent in 2003. Mac-to-market’’ ratios were higher than in 2003, which were the first three years that (7) The long-standing pattern of Fannie 2000, although they both continued to fall the GSEs operated under the higher goal Mae outperforming Freddie Mac was below one: at 0.98 for both categories. targets established by HUD in the 2000 Rule. reversed during 1999 and 2000. But that Between 2002 and 2003, Freddie Mac’s As explained below, the most interesting pattern returned in 2001–2003 when Fannie market ratio for underserved areas fell from recent trend concerned Fannie Mae, which Mae outperformed Freddie Mac on all three 0.98 to 0.87 (24.0 percent for Freddie Mac improved its performance during 2001–2003, goals-qualifying categories. and 27.6 percent for Fannie Mae). Thus, at a time when the conventional conforming • Fannie Mae and Freddie Mac had during 2003, Freddie Mac lagged the market market was showing little change in practically the same performance in 1992 on on all three goals-qualifying categories. affordable lending. the three housing goal categories—special (9) Through 1998, Fannie Mae had (4) During the recent 1999-to-2003 period, affordable loans accounted for 6.3 percent of significantly improved its performance both Fannie Mae and Freddie Mac fell Fannie Mae’s purchases and 6.5 percent of relative to the market. But as a result of shifts significantly below the market in funding Freddie Mac’s purchases, for a ‘‘Fannie-Mae- in its purchases of affordable loans, Fannie affordable loans. Mae lagged the market even further in 2000 • to-Freddie-Mac’’ ratio of 0.97. The 1992 ratio Between 1999 and 2003, special for underserved areas was also 0.98 and that than it had in some earlier years. During affordable loans accounted for 15.1 percent of for low-mod, 1.02. Reflecting Fannie Mae’s 2001–2003, Fannie Mae again improved its Fannie Mae’s purchases, 14.7 percent of much better performance, the special performance relative to the market and, in Freddie Mac’s purchases, and 16.2 percent of affordable ‘‘Fannie-Mae-to-Freddie-Mac’’ 2003, Fannie Mae led the special affordable loans originated in the market; thus, the ratio had risen to 1.27 by 1997, the and low-mod markets but lagged the ‘‘Fannie-Mae-to-market’’ ratio was 0.93 and underserved area ratio to 1.17, and the low- underserved areas market. • mod ratio to 1.10. The above analysis and the data reported 268 Unless otherwise noted, the conventional • However, in 1999, the ‘‘Fannie-Mae-to- under this specific finding (9) are based on conforming market data reported in this section Freddie-Mac’’ ratio for each of the three the ‘‘purchase year’’ approach for measuring exclude an estimate of B&C loans; the less-risky A- goals-qualifying categories fell to slightly GSE activity. The purchase year approach minus portion of the subprime market is included below one. 1999 was the first year since 1992 assigns GSE purchases of both prior-year in the market definition. See Section E.7 and that Freddie Mac had outperformed Fannie (seasoned) and newly-originated mortgages to Appendix D for a discussion of primary market the calendar year in which they were definitions and the uncertainty surrounding estimates of the number of B&C loans in HMDA 269 Fannie Mae had a particularly poor year purchased by the GSE; this results in an data. As noted there, B&C loans are much more during 1999. Therefore, the text also reports inconsistency with the HMDA-reported likely to be refinance loans rather than home averages for 2000–2003, dropping the year 1999 (see market data, which covers only newly- purchase loans. Table A.13 in Section E.9). originated mortgages. Sections E.9 and E.10
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00115 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63694 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
also report the results of an alternative • The choice of which approach to follow • 2002 and 2003. During 2002, the first of ‘‘origination year’’ approach that assigns GSE particularly affected conclusions about these two years of heavy refinancing, Fannie purchases to their year of origination, placing Fannie Mae’s performance relative to the Mae’s performance was slightly above the them on a more consistent basis with the market in 2002 (but not in 2001). Under the market on the low-mod category and slightly HMDA-reported market data. The findings origination-year approach, Fannie Mae below market performance on the special from the origination-year approach are lagged the market on all three housing goal affordable and underserved areas categories; discussed under specific finding (10). categories during 2001 and on the essentially, Fannie Mae matched the market • Fannie Mae’s decline in performance underserved area category during 2002. In on all three categories in 2002. In 2003, during 1999 resulted in the ‘‘Fannie-Mae-to- 2002, Fannie Mae matched the market on the Fannie Mae led the market on the special market’’ ratio falling sharply to 0.74 for special affordable category and led the affordable and low-mod categories and special affordable, to 0.81 for underserved market on the low-mod category (45.5 lagged the market on the underserved areas areas and to 0.89 for low-mod. In 2000, percent for Fannie Mae compared with 44.6 category. The 2003 ‘‘Fannie-Mae-to-market’’ Fannie Mae improved and reversed its percent of the market). ratios were 1.02 for special affordable loans, declining trend, as the ‘‘Fannie-Mae-to- • During 2003, the origination year 1.03 for low-mod loans, and 0.97 for market’’ ratios increased to 0.80 for special approach gives the similar results as the underserved area loans. In 2003, the affordable purchases, to 0.89 for underserved purchase year approach—Fannie Mae led the ‘‘Freddie-Mac-to-market’’ ratios were much area purchases, and to 0.93 for low-mod special affordable and low-mod markets and lower: 0.86 for special affordable loans, 0.90 purchases. lagged the underserved areas market. for low-mod loans, and 0.82 for underserved • During 2001, Fannie Mae increased its (b.3) The GSEs’ Funding of First-time area loans. special affordable percentage by 1.6 Homebuyer Loans (b.5) GSE Market Shares percentage points to 14.9 percent, which was only 0.7 percentage point below the market’s (11) The GSEs’ funding of first-time This analysis includes an expanded performance of 15.6 percent. Fannie Mae homebuyers has been compared to that of ‘‘market share’’ analysis that documents the increased its low-mod percentage from 40.8 primary lenders in the conventional GSEs’’ contribution to important segments of percent to 42.9 percent at the same time that conforming market. Both Fannie Mae and the home purchase and first-time homebuyer the low-mod share of the primary market was Freddie lag the market in funding first-time markets. homebuyers, and by a rather wide margin. (13) The GSEs account for a significant falling from 43.9 percent to 42.9 percent, • placing Fannie Mae at the market’s First-time homebuyers account for 27 share of the total (government as well as performance. Similarly, Fannie Mae percent of each GSE’s purchases of home conventional conforming) market for home increased its underserved area percentage loans, compared with 38 percent for home purchase loans. However, the GSEs’ market from 23.4 percent in 2000 to 24.4 percent in loans originated in the conventional share for each of the affordable lending categories is much less than their share of the 2001 while the underserved area share of the conforming market. overall market. primary market was falling from 26.2 percent (b.4) Performance of the GSEs Based on Total • The GSEs’ purchases were estimated to to 25.2 percent, placing Fannie Mae at 0.8 (Home Purchase and Refinance) Loans be 46 percent of all home loans originated in percentage point from the market’s (12) The GSEs’ acquisitions of total loans metropolitan areas between 1999 and 2003 performance. (including refinance loans as well as home but only 30 percent of loans originated for • During 2002, Fannie Mae continued to purchase loans) were also examined. The African-American and Hispanic borrowers, improve its performance on all three goals main results indicate (a) Freddie Mac has 38 percent of loans originated for low-income categories. Using the purchase-year approach improved its performance but has borrowers, and 37 percent for properties in to measure GSE performance, Fannie Mae consistently lagged the market in funding underserved areas. The GSEs’ market share slightly led the market on the special loans (home purchase and refinance) that for the various affordable lending categories affordable category (16.3 percent for Fannie qualify for the housing goals; and (b) Fannie increased during 2001–2003, but the above- Mae and 16.1 percent for the market), led the Mae has not only improved its performance mentioned pattern remained. market on the low-mod category (45.3 but matched the low-mod market in 2001 and • A study by staff from the Federal Reserve percent for Fannie Mae compared with 44.6 2002 and led both the special affordable and Board suggests that the GSEs have a much percent for the market), and led the market low-mod markets in 2003. Fannie Mae, more limited role in the affordable lending on the underserved area category (26.7 however, lagged the primary market in market than is suggested by the data percent for Fannie Mae versus 26.3 percent funding underserved areas during 2003. (See presented above.271 The Fed study, which for the market). Table A.20 of Section E.10, which is based combined market share, downpayment, and • During 2003, Fannie Mae’s further on the purchase-year approach for measuring default data, concluded that the GSEs play a improvement resulted in Fannie Mae leading GSE activity.) very minimal role in providing credit support the special affordable market (17.1 percent • 1999–2003. During the recent 1999-to- and assuming credit risk for low-income and for Fannie Mae compared with 15.9 percent 2003 period, both Fannie Mae and Freddie minority borrowers; for example, the study for the market) and continuing to lead the Mac fell significantly below the market in concluded that in 1995 the GSEs provided low-mod market (47.0 percent for Fannie funding affordable total (home purchase and only four percent of the credit support going Mae compared with 44.6 percent for the refinance) loans. Between 1999 and 2003, to African-Americans and Hispanic market). During 2003, Fannie Mae lagged special affordable loans accounted for 14.0 borrowers. behind the underserved areas market (26.8 percent of Fannie Mae’s purchases, 13.2 • Section V of this study begins to percent for Fannie Mae compared with 27.6 percent of Freddie Mac’s purchases, and 15.6 reconcile these different results by examining percent for the market). percent of loans originated in the market; the role of the GSEs in the first-time (10) This analysis addresses several thus, the ‘‘Fannie-Mae-to-market’’ ratio was homebuyer market and the downpayment technical issues involved in measuring GSE 0.93 and the ‘‘Freddie-Mac-to-market’’ ratio characteristics of mortgages purchased by the performance. The above analysis was based was 0.88 during this period. GSEs. on the ‘‘purchase year’’ approach, as defined • During the same period, underserved (14) The market role of the GSEs appears in (9) above. An alternative ‘‘origination area loans accounted for 23.8 percent of to be particularly low in important market year’’ approach has also been utilized, which Fannie Mae’s purchases, 22.1 percent of segments such as minority first-time assigns GSE purchases to their year of Freddie Mac’s purchases, and 25.2 percent of homebuyers. origination, placing them on a more loans originated in the market; thus, the consistent basis with the HMDA-reported ‘‘Fannie-Mae-to-market’’ ratio was 0.94 and market data. While the average results (e.g., loans. The effects of excluding B&C loans from the the ‘‘Freddie-Mac-to-market’’ ratio was total market can be seen by comparing the third and 1999–2003 GSE performance) are similar 270 0.88. sixth columns of data in Table A.19 in Section E.10. under the two reporting approaches, GSE 271 See Glenn B. Canner, Wayne Passmore, and performance in any particular year can be 270 As explained in Section E.9, deducting B&C Brian J. Surette, ‘‘Distribution of Credit Risk Among affected, depending on the extent to which loans from the market totals has more impact on the Providers of Mortgages to Lower-Income and the GSE has purchased goals-qualifying market percentages for total (both home purchase Minority Homebuyers’ in Federal Reserve Bulletin, seasoned loans in that particular year. and refinance) loans than for only home purchase 82(12): 1077–1102, December, 1996.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00116 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63695
• Recent analysis has estimated that the of Fannie Mae’s purchases of home loans, 5.8 Manufactured Housing Loans. Both GSEs GSEs’ share of the market for first-time percent of Freddie Mac’s purchases, and 10.6 have raised questions about whether loans on African-American and Hispanic homebuyers percent of home loans originated in the manufactured housing should be excluded was only 14.3 percent between 1999 and conventional conforming market. For this when comparing the primary market with the 2001, or about one-third of their share (41.5 subgroup, Fannie Mae’s performance is 62 GSEs. The GSEs purchase these loans, but percent) of all home purchases during that percent of market performance, while they have not played a significant role in the period. This analysis includes the total Freddie Mac’s performance is 55 percent of manufactured housing loan market. As market, including government and market performance. emphasized by HUD in its 2000 GSE Rule, conventional loans. (17) Some studies have concluded that manufactured housing is an important source • A similar market share analysis was HMDA data overstate the share of market of home financing for low-income families conducted for the conventional conforming loans going to low-income borrowers and and for that reason, should be included in market. Between 1999 and 2001, the GSEs’ underserved areas. This analysis does not any analysis of affordable lending. However, purchases accounted for 56.6 percent of all support that conclusion. for comparison purposes, data are also home loans originated in the conventional • This compares the low-income and presented for the primary market defined conforming market of both metropolitan underserved areas characteristics of the without manufactured housing loans. areas and non-metropolitan areas. Their GSEs’ purchases of newly-originated Because this analysis focuses on purchases of first-time homebuyer loans, on (‘‘current-year’’) loans as reported both by the metropolitan areas, it does not include the the other hand, accounted for only 39.8 GSEs’’ own data and by HMDA data.272 For substantial number of manufactured housing percent of all first-time homebuyer loans recent years, HMDA data on loans sold to the loans originated in non-metropolitan areas. originated in that market. GSEs do not always have higher percentages Subprime Loans. Both GSEs also raised • The GSEs have funded an even lower of low-income and underserved areas loans questions about whether subprime loans share of the minority first-time homebuyer than the GSEs’ own data on their purchases should be excluded when comparing the market in the conventional conforming of newly-originated mortgages. For example, primary market with their performance. In its market. Between 1999 and 2001, the GSEs from 1996–2003, both HMDA and Fannie final 2000 GSE Rule, HUD argued that purchases of African-American and Hispanic Mae reported that special affordable loans borrowers in the A-minus portion of the first-time homebuyer loans represented 30.9 accounted for about 13 percent of Fannie subprime market could benefit from the percent of the conventional conforming Mae’s purchases of newly-originated loans. standardization and lower interest rates that market for these loans. Thus, while the GSEs HMDA reported a 22.6 underserved areas typically accompany an active secondary have accounted for 56.6 percent of all home percentage for Fannie Mae, which was rather market effort by the GSEs. A-minus loans are loans in the conventional conforming market, similar to the underserved areas percentage not nearly as risky as B&C loans and the they have accounted for only 30.9 percent of (23.1 percent) reported by Fannie Mae itself. GSEs have already started purchasing A- loans originated in that market for African- Given that similar patterns were observed for minus loans (and likely the lower ‘‘B’’ grade American and Hispanic first-time Freddie Mac’s mortgage purchases, it appears subprime loans as well). The GSEs homebuyers. that there is no upward bias in the HMDA- themselves have mentioned that a large (15) A noticeable pattern among the lower- based market benchmarks used in this study. portion of borrowers in the subprime market income-borrower loans purchased by the 7. Definition of Primary Market could qualify as ‘‘A credit.’’ This analysis GSEs is the predominance of loans with high Conventional Conforming Market. The includes the A-minus portion of the downpayments. This pattern of purchasing subprime market, or conversely, excludes the mainly high downpayment loans is one market analysis section is based mainly on HMDA data for mortgages originated in the B&C portion of that market. factor explaining why the Fed study found Unfortunately, HMDA does not identify such a small market role for the GSEs. It may conventional conforming market of metropolitan areas during the years 1992 to subprime loans, much less separate them into be the explanation for the small role of their A-minus and B&C components.274 Fannie Mae and Freddie Mac in the first-time 2003. Only conventional loans with a principal balance less than or equal to the Randall M. Scheessele at HUD has identified homebuyer market. Further study of this approximately 200 HMDA reporters that issue is needed. conforming loan limit are included; the • conforming loan limit was $322,700 in primarily originate subprime loans and During 2001 and 2002, approximately 50 account for about 60–70 percent of the percent of Fannie Mae’s special affordable, 2003—these are called ‘‘conventional subprime market.275 To adjust HMDA data low-mod, and underserved areas loans had conforming loans.’’ The GSEs’ purchases of for B&C loans, this analysis follows HUD’s downpayments of at least 20 percent, a FHA-insured, VA-guaranteed, and Rural 2000 Rule which assumed that the B&C percentage only slightly smaller that the Housing Service loans are excluded from this portion of the subprime market accounted for corresponding percentage (53 percent) for all analysis. The conventional conforming one-half of the loans originated by the Fannie Mae’s home loan purchases. Similar market is used as the benchmark against subprime lenders included in Scheessele’s patterns of above-20-percent downpayments which to evaluate the GSEs because that is list.276 As shown below, the effects of on goals-qualifying loans were evident in the market definition Congress requires that Freddie Mac’s 2001, 2002, and 2003 HUD consider when setting the affordable purchases, as well as in prior years for both housing goals. However, as discussed in earlier in Tables A.1 and A.2, which included all GSEs. During 2003, Fannie Mae’s high Section II, some have questioned whether home loans below the conforming loan limit, that lenders in the conventional market are doing is, government loans as well as conventional downpayment share of their special conforming loans. The market share analysis affordable purchases dropped to 45 percent an adequate job meeting the credit needs of reported in Section E.12 also examine the GSEs’ while the patterns for Fannie Mae’s low-mod minority borrowers, which suggests that this role in the overall market. 273 and underserved area purchases did not market provides a low benchmark. 274 And there is some evidence that many change, remaining about 50 percent. subprime loans are not even reported to HMDA, (b.6) Additional Findings 272 In this comparison, a higher special affordable although there is nothing conclusive on this issue. percentage for HMDA-reported mortgage See Fair Lending/CRA Compass, June 1999, p. 3. This analysis examines two additional originations that lenders report as also being sold 275 The list of subprime lenders as well as topics related to minority first-time to the GSEs—as compared with the special Scheessele’s list of manufactured housing lenders homebuyers and the use of HMDA data for affordable percentage for newly-originated are available at http://www.huduser.org/ measuring the characteristics of loans mortgages that the GSEs report as being actually publications/hsgfin.html. originated in the conventional conforming purchased by them—would suggest that HMDA 276 The one-half estimate is conservative as some market. market data are biased; that is, in this situation, the observers estimate that B&C loans account for only (16) The share of the GSEs’ purchases for special affordable percentage for all mortgage 30–40 percent of the subprime market. However, originations reported in HMDA would likely be minority first-time homebuyers was much varying the B&C share from 50 percent to 30 percent larger than the special affordable percentage for all does not significantly change the following analysis less than the share of newly-originated new mortgage originations, including those not of home purchase loans because subprime loans are mortgages in the conventional conforming reported in HMDA as well as those reported in mainly for refinance purposes. Overstating the market for those homebuyers. HMDA. share of B&C loans in this manner also allows for • Between 1999 and 2001, minority first- 273 The market definition in this section is any differences in HMDA reporting of different time homebuyers accounted for 6.6 percent narrower than the ‘‘Total Market’’ data presented Continued
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00117 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63696 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
adjusting the various market percentages for for the FHA market (rather than the the GSEs. Since about one-fourth of GSE B&C loans are minor mostly because the conventional market), they suggest that purchases in any particular year involve analysis in this section focuses on home including HMDA-reported purchased loans loans originated in prior years, HMDA data purchase loans, which historically have in the market definition would overstate will not provide an accurate measure of the accounted for less than one quarter of the mortgage origination totals. Third, Abt goals-qualifying characteristics of the GSEs’ mortgages originated by subprime lenders— Associates surveyed nine wholesale lenders total purchases when the characteristics of the subprime market is mainly a refinance and questioned them concerning their prior-year loans differ from those of newly- market.277 guidelines for reporting in HMDA loans originated, current-year loans. Lender-Purchased Loans in HMDA. When purchased from brokers. Most of these A related issue concerns the appropriate analyzing HMDA data, Fannie Mae includes lenders said brokered loans were reported as definition of the GSE data when making in its market totals those HMDA loans originations if they [the wholesale lender] annual comparisons of GSE performance identified as having been purchased by the make the credit decision; this policy is with the market. On the one hand, the GSE reporting lender, above and beyond loans consistent with the Fed’s guidelines for annual data can be expressed on a purchase- that were originated by the reporting HMDA reporting. Abt Associates concluded year basis, which means that all GSE lender.278 Fannie Mae contends that there are that ‘‘brokered loans do seem more likely to purchases in a particular year would be a subset of loans originated by brokers and be reported as originations * * *.’’ 281 assigned to that particular year. subsequently purchased by wholesale Finally, it should be noted that including Alternatively, the GSE annual data can be lenders that are neither reported by the purchased loans in the market definition expressed on an origination-year basis, which brokers nor the wholesale lenders as does not significantly change the goals- means that GSE purchases in a particular originations but are reported by the qualifying shares of the market, mostly year would be assigned to the calendar year wholesale lenders as purchased loans. because borrower income data are missing for that the GSE-purchased mortgage was According to Fannie Mae, these HMDA- the majority of purchased loans. In addition, originated; for example, a GSE’s purchase reported purchased loans should be added to the low-income and underserved area shares during 2001 of a loan originated in 1999 HMDA-reported originated loans to arrive at for purchased and originated loans are rather would be assigned to 1999, the year the loan an estimate of total mortgage originations. similar. In 2001, the following differences in was originated. These two approaches are This rule’s market definition includes only shares for the conventional conforming home discussed further below. HMDA-reported originations; purchased purchase market were obtained for purchased A final technical issue concerns the loans are excluded from the market and originated loans: Low-income (25.8 reliability of HMDA for measuring the definition. While some purchased loans may percent for purchased loans, 28.3 percent for percentage of goals-qualifying loans in the not be reported as originations in HMDA (the market originations), low-mod income (41.3 primary market. Both GSEs refer to findings Fannie Mae argument), there are several percent, 43.2 percent), and underserved areas from a study by Jim Berkovec and Peter Zorn reasons for assuming that most HMDA- (24.2 percent, 25.8 percent). The comparisons concerning potential bias in HMDA data.284 reported purchased loans are also reported in were also similar for 2002.282 Based on a comparison of the borrower and HMDA as market originations. First, Fed staff 8. Technical Issues: Using HMDA Data To census tract characteristics between Freddie- have told HUD that including purchased Measure the Characteristics of GSE Purchases Mac-purchased loans (from Freddie Mac’s loans would result in double counting and Mortgage Market Originations 283 own data) and loans identified in 1993 279 Second, mortgage originations. This section discusses important technical HMDA data as sold to Freddie Mac, Berkovec comparisons of HMDA-reported FHA data issues concerning the use of HMDA data for and Zorn conclude that HMDA data overstate with data reported by FHA supports the measuring the GSEs’ performance relative to the percentage of conventional conforming Fed’s conclusion. For instance, FHA’s own the characteristics of mortgages originated in loans originated for lower-income borrowers data indicate that during 2001 FHA insured the primary market. The first issue concerns and for properties located in underserved 752,319 home purchase loans in the reliability of HMDA data for measuring census tracts. If HMDA data overstate the metropolitan areas; the sum of HMDA- the borrower income and census tract percentage of goals-qualifying loans, then reported purchased home loans and HMDA- characteristics of loans sold to the GSEs. HUD’s market benchmarks (which are based reported originated home loans in Fannie Mae, in particular, has contended that on HMDA data) will also be overstated. The metropolitan areas alone yields a much HMDA data understates the percentages of its analysis below does not support the Berkovec higher figure of 845,176 FHA-insured loans business that qualify for the three housing and Zorn findings—it appears that HMDA during 2001.280 While these calculations are goals. In its comments on the proposed 2000 data do not overstate the share of goals- Rule, Fannie Mae questioned HUD’s reliance qualifying loans in the market. The types of loans—for example, if B&C loans account on HMDA data for measuring its discussion below of the GSEs’ purchases of for 35 percent of all subprime loans, then assuming performance. As discussed below, HMDA prior-year and current-year loans also that they account for 50 percent is equivalent to highlights the strategy of purchasing assuming that B&C loans are reported in HMDA at data on loans sold to the GSEs do not include prior-year (seasoned) loans that are sold to seasoned loans that qualify for the housing 70 percent of the rate of other loans. goals. The implications of this strategy for 277 The reductions in the market shares are more understanding recent shifts in the relative significant for total loans, which include refinance originated loans. See Harold L. Bunce, The GSEs’ performance of Fannie Mae and Freddie Mac as well as home purchase loans; for data on total Funding of Affordable Loans: A 2000 Update, loans, see Table A.19 in Section 10. Subprime Working Paper HF–013, Office of Policy are discussed below in Section E.9. lenders have been focusing more on home purchase Development and Research, HUD, April 2002, for a. GSEs’ Purchases of ‘‘Prior-Year’’ and loans recently. The home purchase share of loans an alternative analysis showing that a market ‘‘Current-Year’’ Mortgages originated by the subprime lenders in Scheessele’s estimate based on adding HMDA-reported list increased from 26 percent in 1999 to 36 percent purchased loans to HMDA-reported originations There are two sources of loan-level in 2000 before dropping to about 30 percent during would substantially overstate the volume of FHA information about the characteristics of the heavy refinancing years of 2001 and 2002. mortgage originations in metropolitan areas. mortgages purchased by the GSEs—the GSEs 278 In 2001 (2002), lenders reported in HMDA that 281 See Chapter III, ‘‘Reporting of Brokered and themselves and HMDA data. The GSEs they purchased 851,735 (906,684) conventional Correspondent Loans under HMDA’’, in Exploratory provide detailed data on their mortgage conforming, home purchase loans in metropolitan Study of the Accuracy of HMDA Data, by Abt purchases to HUD on an annual basis. As areas; this compares with 2,763,230 (2,929,197) Associates Inc. under contract for the Office of part of their annual HMDA reporting loans that these same lenders reported that they Policy Development and Research, HUD, February responsibilities, lenders are required to originated in metropolitan areas. 12, 1999, page 18. indicate whether their new mortgage 279 282 See Randall M. Scheessele, HMDA Coverage of The percentage shares for purchased loans are originations or the loans that they purchase the Mortgage Market, Housing Finance Working obtained after eliminating purchased loans without Paper No. HF–007. Office of Policy Development data and purchased loans that overlap with (from affiliates and other institutions) are and Research, U.S. Department of Housing and originated loans. The calculations included 138,536 sold to Fannie Mae, Freddie Mac or some Urban Development, July, 1998. purchased loans for 2001 and 182,290 purchased 280 In this example, HMDA-reported purchased loans for 2002. 284 See Jim Berkovec and Peter Zorn, ‘‘How loans insured by FHA have been reduced from 283 Readers not interested in these technical Complete is HMDA? HMDA Coverage of Freddie 411,930 to 100,251 by a procedure that accounts for issues may want to proceed to Section E.9, which Mac Purchases,’’ The Journal of Real Estate missing data and overlapping purchased and compares GSE performance to the primary market. Research, Vol. II, No. 1, Nov. 1, 1996.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00118 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63697
other entity. There have been numerous Fannie Mae has argued that HMDA data the home loans purchased by each GSE.287 studies by HUD staff and other researchers understate its past performance, where HMDA data provide information mainly on that use HMDA data to compare the borrower performance is defined as the percentage of newly-originated mortgages that are sold to and neighborhood characteristics of loans Fannie Mae’s mortgage purchases accounted the GSEs—that is, HMDA data on loans sold sold to the GSEs with the characteristics of for by one of the goal-qualifying categories. to the GSEs will not include many of their all loans originated in the market. One As explained below, over the past six years, purchases of prior-year loans. The question is whether HMDA data, which is implications of this for measuring GSE HMDA has provided rather reliable national- widely available to the public, provides an performance can be seen in Table A.11, level information on the goals-qualifying accurate measure of GSE performance, as which provides annual data on the borrower compared with the GSEs’ own data.285 percentages for the GSEs’ purchases of and census tract characteristics of GSE ‘‘current-year’’ (i.e., newly-originated) loans, purchases, as measured by HMDA data and 285 For another discussion of this issue, see but not for their purchases of ‘‘prior-year’’ by the GSEs’ own data. Table A.11 divides Randall M. Scheessele, HMDA Coverage of the loans.286 each of the GSEs’ goals-qualifying Mortgage Market, Housing Finance Working Paper In any given calendar year, the GSEs can percentages for a particular acquisition year HF–007, Office of Policy Development and purchase mortgages originated in that into two components, the percentage for Research, Department of Housing and Urban calendar year or mortgages originated in a ‘‘prior-year’’ loans and the percentage for Development, July 1998. Scheessele reports that ‘‘current-year’’ loans. HMDA data covered 81.6 percent of the loans prior calendar year. In 2001 and 2002, for acquired by Fannie Mae and Freddie Mac in 1996. example, purchases of prior-year mortgages BILLING CODE 4210–27–P The main reason for the under-reporting of GSE accounted for approximately 20 percent of acquisitions is a few large lenders failed to report 287 The ‘‘prior-year’’ share dropped to 16 percent the sale of a significant portion of their loan during the heavy refinancing year of 2003. During 286 Between 1993 and 1996, the GSEs’ purchases originations to the GSEs. Also see the analysis of the 1990s, the GSEs increased their purchases of of prior-year loans were not as targeted as they were HMDA coverage by Jim Berkovec and Peter Zorn. seasoned loans; see Paul B. Manchester, Goal after 1996; thus, during this period, HMDA ‘‘Measuring the Market: Easier Said than Done,’’ Performance and Characteristics of Mortgages Secondary Mortgage Markets. McLean VA: Freddie provided reasonable estimates of the goals- Purchased by Fannie Mae and Freddie Mac, 1998– Mac, Winter 1996, pp. 18–21; as well as the qualifying percentages of the GSEs’ purchases of all Berkovec and Zorn study cited in the above (both current-year and prior-year) loans, with a few 2000, Housing Finance Working Paper No. HF–015, footnote. exceptions (see Table A.11). Office of Policy Development and Research, HUD, May 2001.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00119 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63698 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00120 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.020
Consider Fannie Mae’s special affordable HMDA figure is similar in concept to the purchases (with the exception of Freddie purchases in 2002. According to Fannie current-year percentage from the GSEs’ own Mac’s underserved area loans, as discussed Mae’s own data, 16.3 percent of its purchases data. And the HMDA figure and the GSE below). For example, Fannie Mae reported during 2002 were special affordable loans. current-year figure are practically the same in that 13.7 percent of the current-year loans it According to HMDA data, only 15.5 percent this case (15.5 versus 15.8 percent). Thus, the purchased between 1996 and 2003 were for of loans sold to Fannie Mae fell into the relatively large share of special affordable special affordable borrowers. In their HMDA special affordable category. In this case, mortgages in Fannie Mae’s purchases of submissions, lenders reported a nearly HMDA data underestimate the special prior-year mortgages explains why Fannie identical figure of 13.4 percent for the special Mae’s own data show an overall (both prior- affordable share of Fannie Mae’s purchases affordable share of loans that they sold to year and current-year) percentage of special during 2002. What explains these different Fannie Mae. The corresponding numbers for patterns in the GSE and HMDA data? The affordable loans that is higher than that reported for Fannie Mae in HMDA data. Freddie Mac were 12.8 percent reported by reason that HMDA data underestimate the them and 12.1 percent reported by HMDA. special affordable percentage of Fannie Mae’s b. Reliability of HMDA Data During the same period, both Fannie Mae 2002 purchases can be seen by disaggregating With the above explanation of the basic and HMDA reported that approximately 23 Fannie Mae’s purchases during 2002 into differences between GSE-reported and percent of current-year loans purchased by their prior-year and current-year HMDA-reported loan information, issues Fannie Mae financed properties in components. Table A.11 shows that the related to the reliability of HMDA data can underserved areas. However, Freddie Mac overall figure of 16.3 percent for special now be discussed. Table A.12 presents the reported that 21.3 percent of the current-year affordable purchases is a weighted average of same information as Table A.11, except that loans it purchased between 1996 and 2003 18.8 percent for Fannie Mae’s purchases the data are aggregated for the years 1993–5, financed properties in underserved areas, a during 2002 of prior-year mortgages and 15.8 1996–2003, and 1999–2003. Comparing figure somewhat higher than the 19.6 percent percent for its purchases of current-year HMDA-reported data on GSE purchases with that HMDA reported as underserved area purchases. The HMDA-reported figure of 15.5 GSE-reported current-year data suggests that, percent is based mainly on newly-mortgaged on average, HMDA data have provided loans sold to Freddie Mac during that 288 (current-year) loans that lenders reported as reasonable estimates of the goals-qualifying period. being sold to Fannie Mae during 2002. The percentages for the GSEs’ current-year BILLING CODE 4210–27–P
288 Freddie Mac’s underserved area figures for 2002 and 2003 showed particularly large discrepancies. As shown in Table A.11, Freddie Mac reported that 25.0 (23.4) percent of the current- year loans it purchased during 2002 (2003) financed properties in underserved areas, a figure much higher than the 21.4 (20.3) percent that HMDA reported as underserved area loans sold to Freddie Mac during 2002. These discrepancies are the largest in Table A.11, and it is not clear what explains them. This downward bias for HMDA data, is the opposite of that suggested by Berkovec and Zorn, who argued that affordability percentages from HMDA data are biased upward.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00121 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63700 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00122 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.021 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63701
The facts that the Fannie Mae and HMDA credit for their purchase activity in the year 9. Affordable Lending by the GSEs: Home figures for special affordable, low-mod and that the purchase actually takes place; this Purchase Loans underserved area loans are similar, and that approach is also consistent with the statutory This section compares the GSEs’ affordable the Freddie Mac discrepancies are the result requirement for measuring GSE performance lending performance with the primary of Freddie Mac reporting higher percentages under the housing goals. However, this market for the years 1993–2003. The analysis than HMDA, suggest that the Berkovec and approach results in an obvious ‘‘apples to in this section begins by presenting the GSE Zorn conclusions about HMDA being upward oranges’’ problem with respect to the HMDA- data on a purchase-year basis. As discussed biased are wrong.289 For the 1996-to-2003 based market data, which include only above, the GSE data that are reported to HUD period, the discrepancies reported in Table newly-originated mortgages (i.e., current-year A.11 as well as Table A.12 are mostly mortgages). To place the GSE and market include their purchases of mortgages consistent with HMDA being biased in a data on an ‘‘apples to apples’’ basis, HUD has originated in prior years as well as their downward direction, not an upward also used an alternative approach that purchases of mortgages originated during the direction as Berkovec and Zorn contend.290 expresses the GSE annual data on an current year. The market data reported by In particular, the Freddie-Mac-reported origination-year basis. In this case, all HMDA include only mortgages originated in underserved area percentage (as well as its purchases by a GSE in any particular year the current year. This means that the GSE- special affordable percentage) being larger would be fully reported but they would be versus-market comparisons are defined than the HMDA-reported underserved area allocated to the year that they were somewhat inconsistently for any particular percentage suggests a downward bias in originated, rather than to the year they were calendar year. Each year, the GSEs have HMDA. The more recent and complete purchased. Under this approach, a GSE’s data newly-originated loans available for (Fannie Mae data as well as Freddie Mac for the year 2000 would not only include that purchase, but they can also purchase loans data) analysis does not support the Berkovec GSE’s purchases during 2000 of newly- from a large stock of seasoned (prior-year) and Zorn finding that HMDA overstates the originated mortgages but also any year-2000- loans currently being held in the portfolios goals-qualifying percentages of the market.291 originations purchased in later years (i.e., of depository lenders. One method for during 2001, 2002 and 2003 in this analysis). making the purchase-year data more c. Purchase-Year Versus Origination-Year This approach places the GSE and the market consistent is to aggregate the data over Reporting of GSE Data data on a consistent, current-year basis. In several years, instead of focusing on annual In comparing the GSEs’ performance to the the above example, the market data would data. This provides a clearer picture of the primary market, HUD has typically expressed present the income and underserved area types of loans that have been originated and the GSEs’ annual performance on a purchase- characteristics of mortgages originated in are available for purchase by the GSEs. This year basis. That is, all mortgages (including 2000, and the GSE data would present the approach is taken in Tables A.14 and A.15, both current-year mortgages and prior-year same characteristics of all year-2000- which are discussed below. Another method mortgages) purchased by a GSE in a mortgages that the GSE has purchased to date for making the GSE and market data particular year are assigned to the year of (i.e., through year 2003).292 consistent is to express the GSE data on an GSE purchase. The approach of including a Below, results will be presented for both origination-year basis; that approach is taken GSE’s purchases of both ‘‘current-year’’ and the purchase-year and origination-year in Table A.16, which is discussed after ‘‘prior-year’’ mortgages gives the GSE full approaches. Following past HUD studies that presenting the annual results on a purchase- have compared GSE performance with the year basis. 289 primary market, most of the analysis in this The data in Table A.12 that support Berkovec a. Longer-Term Performance, 1993–2003 and and Zorn are the 1993–95 special affordable and section reports the GSE data on a purchase- 1996–2003 low-mod data (particularly for Freddie Mac) that year basis; however, the main results are show HMDA over reporting percentages by more repeated with the GSE data reported on an Table A.13 summarizes the funding of than a half percentage point. Otherwise, the data in origination-year basis. This allows the reader goals-qualifying mortgages by the GSEs, Table A.12, as well as Table A.11, do not present to compare any differences in findings about depositories and the conforming market for a picture of HMDA’s having an upward bias in the ten-year period between 1993 and 2003. reporting targeted loans. In fact, the recent years’ how well the GSEs have been doing relative data suggest a downward bias in HMDA’s reporting to the market. Data are also presented for two important of targeted loans. sub-periods: 1993–95 (for showing how 290 Of course, on an individual year basis, the 292 Under the origination-year approach, GSE much the GSEs have improved their GSEs’ current-year data can differ significantly from performance for any specific origination year (say performance since the early-to-mid 1990s); the HMDA-reported data on GSE purchases. The year 2000) at the end of a particular GSE purchase and 1996–2003 (for analyzing their other annual data reported in Table A.11 show a year (say year 2003) is subject to change in the performance since the current definitions of mixture of results—in some cases the HMDA future years. Table A.16 (in Section E.9 below) the housing goals were put into effect). Given percentage is larger than the GSE ‘‘current year’’ reports that 13.7 percent of year-2000 mortgage the importance of the GSEs for expanding percentage (e.g., Fannie Mae’s special affordable originations that Fannie Mae purchased through homeownership, this section focuses on purchases in 2000) while in other cases the HMDA year 2002 qualify as special affordable; the special percentage is smaller than the GSE current year affordable share for the market was 16.6 percent in home purchase mortgages, and the next percentage (e.g., Freddie Mac’s special affordable 2000, which indicates that, to date, Fannie Mae has section will examine first-time homebuyer purchases in recent years). As noted in the text, the lagged the primary market in funding special loans. Section IV below will briefly discuss differential is typically in the opposite direction to affordable mortgages originated during 2000. the GSEs’ overall performance, including that predicted by Berkovec and Zorn, particularly However, Fannie Mae’s special affordable refinance and home purchase loans. Several on the underserved areas category. performance could change in the future as Fannie points stand out concerning the affordable 291 Table A.12 also includes aggregates for the Mae continues to purchase year-2000 originations lending performance of Freddie Mac and more recent period, 1999–2003. The ratios of during 2004 and the following years. Of course, HMDA-reported-to-GSE-reported averages for this whether Fannie Mae’s future purchases result in it Fannie Mae over the two longer-term periods, sub-period are similar to those reported for 1996– ever leading the 2000-year market is not known at 1993–2003 and 1996–2003. 2003. this time. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00123 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63702 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00124 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.022 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63703
Freddie Mac lagged both Fannie Mae and Freddie Mac’s purchases were for low- and Mac-to-market’’ ratios for 1996–2003 were the primary market in funding affordable moderate-income borrowers, compared with lower—0.83 for special affordable, 0.92 for home loans in metropolitan areas between 42.2 percent of Fannie Mae’s purchases, 43.1 low-mod, and 0.86 for underserved areas. 1993 and 2003. During that period, 12.2 percent of loans originated by depositories, The above analysis has defined the market percent of Freddie Mac’s mortgage purchases and 43.6 percent of loans originated in the to exclude B&C loans, which HUD believes were for special affordable (mainly very-low- conventional conforming market. Over the is the appropriate market definition. income) borrowers, compared with 13.3 same period, 22.0 percent of Freddie Mac’s However, to gauge the sensitivity of the percent of Fannie Mae’s purchases, 15.4 purchases financed properties in results to how the market is defined, Table percent of loans originated by underserved neighborhoods, compared with A.14 shows the effects on the market depositories,293 and 15.5 percent of loans 24.0 percent of Fannie Mae’s purchases, 25.1 percentages for different definitions of the originated in the conforming market without percent of depository originations, and 25.7 conventional conforming market, such as percent of loans originated in the primary B&C loans.294 excluding manufactured housing loans, small market. Although Freddie Mac consistently loans, and all subprime loans (i.e., the A- improved its performance during the 1990s, Fannie Mae’s affordable lending performance was better than Freddie Mac’s minus portion of the subprime market as well a similar pattern characterized the 1996–2003 as the B&C portion). For example, the average period. During that period, 40.3 percent of over the 1993 to 2003 period as well as during the 1996 to 2003 period. However, special affordable (underserved area) market Fannie Mae lagged behind depositories and percentage for 1996–2003 would fall by about 293 As shown in Table A.13, the depository the overall market in funding affordable 1.6 (1.2) percentage points if both small loans percentage is higher (16.8 percent) if the analysis loans during both of these periods (see above (less than $15,000) and manufactured loans is restricted to those newly-originated loans that paragraph). Between 1996 and 2003, the in metropolitan areas were also dropped from depositories do not sell (the latter being a proxy for ‘‘Fannie-Mae-to-market’’ ratio was only 0.89 the market definition (see right-hand-side loans held in depositories’ portfolios). Note that column in Table A.14). Except for Fannie during the recent, 1999-to-2003 period (also on the special affordable category, obtained reported in Table A.13), there is less difference by dividing Fannie Mae’s performance of Mae’s relative performance on the low-mod between the two depository figures. 14.1 percent by the market’s performance of category, the above findings with respect to 294 Unless stated otherwise, the market in this 15.9 percent. Fannie Mae’s market ratio was the GSEs’ longer-term performance are not section is defined as the conventional conforming 0.97 on the low-mod category and 0.93 on the much affected by the choice of market market without estimated B&C loans. underserved area category. The ‘‘Freddie- definition.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00125 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63704 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00126 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.023 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63705
b. Recent Performance, 1999–2003 loans at a lower rate than Fannie Mae (24.7 averages if the year 1999 is dropped—over This and the next subsection focus on the percent) and the primary market (26.2 the 2000–2003 period, Fannie Mae’s average data for 1999–2003 in Table A.13 and percent). As these figures indicate, both performance on the underserved area the annual data reported in Table A.14. As Fannie Mae and Freddie Mac continued to category is practically at market levels under explained below, the annual data are useful lag the market during this recent four-year the above alternative definition of the market, for showing shifts in the relative positions of period. The GSEs’ market ratios were 0.91– and its performance on the special affordable Fannie Mae and Freddie Mac that began in 0.93 for special affordable loans and 0.97– and low-mod categories is above market 1999, and for highlighting the improvements 0.99 for low-mod loans. Although less than levels. one (where one indicates equal performance made by Fannie Mae during 2001–2003 Finally, Tables A.13 and A.14 report GSE with the market), the ‘‘Fannie-Mae-to- (which were the first three years under and market data for the even more recent HUD’s higher goal levels) and by Freddie market’’ ratio (0.94) for the underserved area category was much higher than the ‘‘Freddie- period, 2001–2003, which represents the first Mac during 2002. Between 1993 and 1998, Mac-to-market’’ ratio (0.88). three years under the current housing goal Freddie Mac’s performance fell below Fannie Fannie Mae’s performance in 1999 was targets (put in place by HUD in its Final Rule Mae’s, but a sharp improvement in Freddie significantly below its long run trend. Thus, dated October 30, 2000). These data show Mac’s performance during 1999 pushed it averages for 2000–2003 are also presented in that Freddie Mac’s average performance pass Fannie Mae on all three goals-qualifying Table A.13, dropping 1999. These data show during this period was below the market on categories. In 2000, Fannie Mae improved its an increase in Fannie Mae’s performance each of the three housing goals (with market underserved areas performance enough to relative to the market. Between 2000 and ratios of 0.96 for special affordable, 0.98 for surpass Freddie Mac on that category, while 2003, special affordable (underserved area) low-mod, and 0.91 for underserved areas and Freddie Mac continued to out-perform loans accounted for 15.6 percent (25.5 that Fannie Mae’s average performance was Fannie Mae on the borrower-income percent) of Fannie Mae’s purchases, above the market on the special affordable categories (special affordable and low-mod). compared with 16.0 percent (26.4 percent) By 2002, Fannie Mae had improved its and low-mod categories (with a market ratio for the market. During this 2000–2003 period, of 1.02 on each category) but below the performance enough to surpass Freddie Mac Fannie Mae slightly led the low-mod market on all three goals-qualifying categories and to market on the underserved areas category (44.4 percent for Fannie Mae and 44.1 (with a market ratio of 0.98). lead the special affordable and low-mod percent for the primary market). markets, while lagging the underserved areas Table A.14 shows the effects on the market c. GSEs’ Performance—Annual Data market. percentages for 1999–2003 (as well as 2000– Freddie Mac’s Annual Performance. As Consider first the average data for 1999– 2003) of different definitions of the shown by the annual data reported in Table 2003 reported in Table A.13. During this conventional conforming market. Excluding recent period, Freddie Mac’s average A.15, Freddie Mac significantly improved its both small loans and manufactured housing purchases of goals-qualifying loans during performance was similar to Fannie Mae’s loans (as well as B&C loans) in metropolitan performance for the special affordable the 1990s. Its purchases of loans for special areas would reduce the 1999–2003 market affordable borrowers increased from 6.5 category. Between 1999 and 2003, 14.7 percentage for special affordable loans from percent of its business in 1992 to 9.2 percent percent of Freddie Mac’s purchases and 15.1 16.2 percent to 14.9 percent, which would percent Fannie Mae’s mortgage purchases place Fannie Mae slightly above the market in 1997, and then jumped to 14.7 percent in consisted of special affordable loans, and Freddie Mac close to the market. 2000 before falling slightly to 14.4 percent in compared with a market average of 16.2 Similarly, excluding these loans would 2001 and rising again to almost 16 percent in percent. During this period, Freddie Mac reduce the 1999–2003 market percentage for 2002 and 2003. The underserved areas share purchased low-mod loans lower than the rate underserved areas from 26.2 percent to 25.2 of Freddie Mac’s purchases increased at a of Fannie Mae—42.6 percent for Freddie percent, which would raise Fannie Mae’s more modest rate, rising from 18.6 percent in Mac, 43.6 percent for Fannie Mae, and 44.1 market ratio from 0.94 to 0.98 and Freddie 1992 to 22.3 percent by 2001; it then jumped percent for the market. Freddie Mac (23.1 Mac’s, from 0.88 to 0.92. As shown in Table to 25.8 percent in 2002 but fell to 24.0 percent) also purchased underserved area A.14, Fannie Mae is even closer to the market percent in 2003.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00127 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63706 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00128 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.024 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63707
With its improved performance, Freddie of market originations, for a ‘‘Fannie Mae-to- 1.6 percentage points, from 13.3 percent in Mac closed its gap with the market in market’’ ratio of 0.82. By 1998, underserved 2000 to 14.9 percent in 2001, and then funding goals-qualifying loans. In 2003, areas accounted for 22.7 percent of Fannie increased it further to 16.3 percent in 2002, special affordable loans accounted for 15.6 Mae’s purchases and 24.2 percent of market the latter being slightly above the market’s percent of Freddie Mac’s purchases and 15.9 originations, for a higher ‘‘Fannie Mae-to- performance of 16.1 percent. The ‘‘Fannie- percent of loans originated in the market’’ ratio of 0.94.297 Mae-to-market’’ ratio for special affordable conventional conforming market, which The year 1999 saw a shift in the above loans jumped from 0.80 in 2000 to 1.01 in produces a ‘‘Freddie-Mac-to-market’’ ratio of patterns, with Fannie Mae declining in 2002. In 2003, Fannie Mae’s special 0.98 (15.6 divided by 15.9). Table A.15 shows overall performance while the share of goals- affordable performance jumped to 17.1 the trend in the ‘‘Freddie-Mac-to-market’’ qualifying loans in the market increased. percent while the market declined slightly to ratio from 1992 to 2003 for each of the goals- Between 1998 and 1999, the special 15.9 percent, increasing Fannie Mae’s market qualifying categories. For the special affordable share of Fannie Mae’s business ratio to 1.08. affordable and low-mod categories, Freddie declined from 13.2 percent to 12.5 percent Between 2000 and 2001, Fannie Mae Mac’s performance relative to the market while this type of lending in the market increased its low-mod percentage from 40.8 remained flat (at approximately 0.60 and increased from 15.4 percent to 17.0 percent. percent to 42.9 percent at the same time that 0.80, respectively) through 1997; by 2003, the For this reason, the ‘‘Fannie-Mae-to-market’’ the low-mod share of the primary market was ‘‘Freddie-Mac-to-market’’ ratios had risen to ratio for special affordable loans declined falling from 43.9 percent to 42.9 percent, 0.98 for both the special affordable and low- sharply from 0.86 in 1998 to 0.74 in 1999. placing Fannie Mae at the market’s mod categories. The share of Fannie Mae’s purchases in performance in 2001. During 2002, the low- Surprisingly, Freddie Mac did not make underserved areas also declined, from 22.7 mod share of Fannie Mae’s purchases of much progress during the 1990s closing its percent in 1998 to 20.4 percent in 1999, home loans increased further to 45.3 percent, gap with the market on the underserved areas which lowered the ‘‘Fannie-Mae-to-market’’ placing Fannie Mae 0.7 percentage points category. The ‘‘Freddie-Mac-to-market’’ ratio ratio from 0.94 to 0.81. above the market performance of 44.6 for underserved areas was the same in 2000 After declining in 1999, Fannie Mae’s percent. Between 2002 and 2003 Fannie (0.84) as it was in 1992 (0.84). While it rose performance rebounded in 2000, particularly Mae’s performance jumped to 47.0 percent, to 0.88 in 2001, that was due more to a on the underserved areas category. Fannie while the primary market remained at 44.6 decline in the market level than to an Mae’s underserved areas percentage jumped percent, giving Fannie Mae a market ratio of improvement in Freddie Mac’s performance. by three percentage points from 20.4 percent 1.05 in 2003. However, due to a substantial increase in in 1999 to 23.4 percent in 2000. The 2000 Fannie Mae increased its underserved area Freddie Mac’s underserved area percentage figure was similar to its level in 1997 but percentage from 23.4 percent in 2000 to 24.2 from 22.3 percent in 2001 to 25.8 percent in below Fannie Mae’s peak performances of percent in 2001 while the underserved area 2002, Freddie Mac’s performance approached 24–25 percent during 1994 and 1995. share of the primary market was falling from market performance (26.3 percent) during Between 1999 and 2000, the ‘‘Fannie-Mae-to- 26.4 percent to 25.2 percent, placing Fannie 2002. 295 In the ten years under the housing market’’ ratio for underserved areas increased Mae at less than one percentage point from goals, the year 2002 represented the first time from 0.81 to 0.89. Fannie Mae improved its the market’s performance. The ‘‘Fannie-Mae- that Freddie Mac’s performance in performance on the special affordable goal at to-market’’ ratio for underserved area loans purchasing home loans in underserved areas a more modest rate. Fannie Mae’s special was 0.97 in 2001. During 2002, the had ever been within two percentage points affordable percentage increased by 0.8 underserved area share of Fannie Mae’s of the market’s performance.296 But, as noted percentage points from 12.5 percent in 1999 purchases of home loans increased further to above, Freddie Mac’s performance on the to 13.3 percent in 2000. The 2000 figure was 26.7 percent, placing Fannie Mae slightly underserved areas goal fell to 24.0 percent in similar to its previous peak level (13.2 ahead of market performance (26.3 percent). 2003, leaving it with a ‘‘Freddie Mac-to- percent) in 1998. The ‘‘Fannie-Mae-to- However, between 2002 and 2003, Fannie Market’’ ratio of 0.87. market’’ ratio for special affordable loans Mae showed little improvement (rising to Fannie Mae’s Annual Performance. With increased from 0.74 in 1999 to 0.80 in 2000, 26.8 percent) while the market increased to respect to purchasing affordable loans, with the latter figure remaining below Fannie 27.6 percent, leaving Fannie Mae with a Fannie Mae followed a different path than Mae’s peak market ratio (0.86) in 1998. market ratio of 0.97. Freddie Mac. Fannie Mae improved its Fannie Mae continued its improvement in As noted earlier, Tables A.13 and A.14 performance between 1992 and 1998 and purchasing targeted home loans during 2001, summarize Fannie Mae’s average made much more progress than Freddie Mac at a time when the conventional conforming performance over the 2001–2003 period. in closing its gap with the market. In fact, by market was experiencing a decline in During these first three years under the 1998, Fannie Mae’s performance was close to affordable lending; and again in 2002, at a current housing goal targets, Fannie Mae led that of the primary market for some time when the conventional conforming the special affordable market (average important components of affordable lending. market was increasing enough to return performance of 16.2 percent versus 15.9 In 1992, special affordable loans accounted approximately to its year-2000 level. Thus, percent for the market) and the low-mod for 6.3 percent of Fannie Mae’s purchases during the 2000-to-2003 period, Fannie Mae market (average performance of 45.2 percent and 10.4 percent of all loans originated in the significantly improved its targeted versus 44.1 percent for the market) but lagged conforming market, giving a ‘‘Fannie Mae-to- purchasing performance while the primary the underserved areas market (average market’’ ratio of 0.61. By 1998, this ratio had market originated targeted home loans at performance of 26.0 percent versus 26.4 risen to 0.86, as special affordable loans had about the same rate in 2002 as it did in 2000. percent for the market). Table A.14 also increased to 13.2 percent of Fannie Mae’s As a result, Fannie Mae’s performance during reports Fannie Mae’s 2001–2003 performance purchases and to 15.4 percent of market 2001 approached the market on the special under alternative definitions of the primary originations. A similar trend in market ratios affordable and underserved area categories market. As shown there, the above findings can be observed for Fannie Mae on the and matched the market on the low-mod of Fannie Mae’s improvement relative to the underserved areas category. In 1992, category. In 2002, Fannie Mae outperformed market during 2001–2003 are further underserved areas accounted for 18.3 percent the market on all three areas categories. reinforced when lower market percentages of Fannie Mae’s purchases and 22.2 percent As shown in Table A.15, Fannie Mae are used. For example, Fannie Mae increased its special affordable percentage by essentially matches the underserved areas 295 Table A.14 reports annual market percentages market if manufactured housing loans in that exclude the effects of manufactured housing, 297 Freddie Mac, on the other hand, fell further metropolitan areas (in addition to B&C loans) small loans, and subprime loans. Freddie Mac’s behind the market during this period. In 1992, are excluded from the market definition (a performance is closer to the market average under Freddie Mac had a slightly higher underserved Fannie Mae share of 26.0 percent and a the alternative market definitions, particularly areas percentage (18.6 percent) than Fannie Mae market share of 26.1 percent). during 2001 and 2002. (18.3 percent). However, Freddie Mac’s 296 Prior to 2002, Freddie Mac’s performance on underserved areas percentage had only increased to Changes in the ‘‘Fannie-Mae-to-Freddie- the underserved areas category had not approached 19.8 percent by 1998 (versus 22.7 percent for Mac’’ Performance Ratio. The above the market even under the alternative market Fannie Mae). Thus, the ‘‘Freddie Mac-to-market’’ discussion documents shifts in the relative definitions reported in Table A.14. ratio fell from 0.84 in 1992 to 0.82 in 1998. performance of Fannie Mae and Freddie Mac
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00129 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63708 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
over the past few years. To highlight these 2002 and 2003) have seen a substantial when Fannie Mae improved its overall changing patterns, Table A.15 reports the improvement in Fannie Mae’s performance affordable lending performance. For example, ratio of Fannie Mae’s performance to Freddie on all three goals-qualifying categories. consider Fannie Mae’s underserved area Mac’s performance for each goals category for Fannie Mae ended the 1990s with a decline performance of 24.4 percent during 2001, the years 1992 to 2003. As shown there, the in affordable lending performance at the which was helped by its purchases of ‘‘Fannie-Mae-to-Freddie-Mac’’ ratio for the same time that Freddie Mac was improving seasoned mortgages on properties located in special affordable category increased from and the share of goals-qualifying loans was underserved neighborhoods. The approximately one in 1992 (indicating equal increasing in the market. Both GSEs’ underserved area percentage for Fannie performance) to over 1.3 during the 1994–97 performance during 2000 was encouraging— Mae’s purchases of newly-originated period, indicating that Fannie Mae clearly Freddie Mac continued to improve, (current-year) mortgages was only 23.3 out-performed Freddie Mac during this particularly with respect to the borrower- percent in 2001, or 1.9 percentage points period. Between 1997 and 2000, Freddie Mac income categories, while Fannie Mae below the market average of 25.2 percent. substantially increased its special affordable reversed its declining performance, Fannie Mae obtained its higher overall share (from 9.2 percent to 14.7 percent), particularly with respect to underserved percentage (24.4 percent) by purchasing causing the ‘‘Fannie-Mae-to-Freddie-Mac’’ areas. During 2000, Freddie Mac seasoned loans with a particularly high ratio to fall from 1.27 in 1997 to 0.90 in 2000 outperformed Fannie Mae on the special concentration (28.3 percent) in underserved (indicating Freddie Mac surpassed Fannie affordable and low-mod categories, while areas. Similarly, during 2001, the special Mae). But Fannie Mae’s stronger performance Fannie Mae purchased a higher percentage of during 2001–2003 returned the ratio to above loans in underserved areas. During 2001, affordable share of Fannie Mae’s purchases of one (1.03 in 2001 and 2002 and 1.10 in 2003), Fannie Mae continued to improve its newly-originated mortgages was only 14.2 indicating better performance for Fannie Mae performance while Freddie Mac’s percent, or 1.4 percentage points below the (e.g., 17.1 percent in 2002 versus 15.6 percent performance remained about the same and market average of 15.6 percent. Again, Fannie for Freddie Mac). The ‘‘Fannie-Mae-to- the market’s originations of affordable loans Mae improved its overall performance by Freddie-Mac’’ performance ratio for low-mod declined somewhat. The result was that purchasing seasoned loans with a high loans followed a similar pattern, standing at during 2001 Fannie Mae outperformed percentage (18.1 percent) of special 1.07 in 2003 (47.0 percent for Fannie Mae Freddie Mac on all three goals-qualifying affordable loans, enabling Fannie Mae to versus 43.8 percent for Freddie Mac). categories, and even matched the market on reduce its gap with the market to 0.7 Prior to 2000, the ‘‘Fannie-Mae-to-Freddie- the low-mod category. During 2002, both percentage points—14.9 percent versus 15.6 Mac’’ ratio for underserved areas had also Fannie Mae and Freddie Mac again improved percent. followed a pattern similar to that outlined their performance; Fannie Mae continued to As shown in Table A.11, Freddie Mac also above for special affordable loans, but at a outperform Freddie Mac and outperformed followed a strategy of purchasing seasoned lower overall level—rising from about one in the market on all three goals-qualifying special affordable loans mainly after 1999. 1992 (indicating equal performance) to categories. While Freddie Mac lagged the Prior to 2000, Freddie Mac had not pursued approximately 1.2 during the 1994–97 market on all three goals-qualifying such a strategy, or at least not to the same period, before dropping to slightly below one categories during 2002, it had significantly degree as Fannie Mae. During the 1997–99 (0.98) in 1999. However, Fannie Mae closed its gap by the end of 2002, particularly period, Freddie Mac’s purchases of prior-year increased its underserved areas percentage on the underserved area category. During mortgages and newly-originated mortgages from 20.4 percent in 1999 to 24.4 percent in 2003, Fannie Mae made significant had similar percentages of special affordable 2001 while Freddie Mac only increased its improvement in the special affordable and (and low-mod) borrowers. Over time, there percentage from 20.9 percent to 22.3 percent. low-mod categories, allowing it to lead the have been small differentials between This resulted in the ‘‘Fannie-Mae-to-Freddie- primary market. Freddie Mac, on the other Freddie Mac’s prior-year and newly- Mac’’ ratio rising from 0.98 in 1999 to 1.09 hand, simply maintained its 2002 originated mortgages for the underserved in 2001. But during 2002, Freddie Mac’s performance in these two categories, which areas category but they have been smaller underserved area percentage jumped by 3.5 meant it lagged further behind Fannie Mae. than the differentials for Fannie Mae (see percentage points to 25.8 percent, while On the underserved area category, Fannie Table A.11). Fannie Mae’s increased at a more modest rate Mae maintained its 2002 performance during d. GSEs’ Annual Purchases of Home Loans— (by 2.3 percentage points) to 26.7 percent, 2003 while Freddie Mac significantly Origination-Year Basis with the result being that the ‘‘Fannie-Mae- reduced its performance, leaving both GSEs, to-Freddie-Mac’’ ratio for underserved area but particularly Freddie Mac, behind the Table A.16 reports GSE purchase data for loans fell from 1.09 in 2001 to 1.03 in 2002. primary market on this category. 1996 to 2003 on an origination-year basis. During 2003, Fannie Mae essentially GSE Purchases of Seasoned Loans. When Recall that in this case, mortgages purchased maintained its performance (26.8 percent), the GSE data are expressed on a purchase- by a GSE in any particular calendar year are while Freddie Mac reduced its performance year basis (as in the above analysis), one allocated to the year that the mortgage was by 1.8 percentage points to 24.0 percent. This factor which affects each GSE’s performance originated, rather than to the year that the increased the 2003 ‘‘Fannie Mae-to-Freddie concerns their purchases of seasoned (prior- mortgage was purchased (as in the above). Mac’’ ratio for underserved areas to 1.12. year) loans. As shown in Table A.11, Fannie This approach places the GSE and the market To conclude, while Freddie Mac ended the Mae followed a strategy of purchasing data on a consistent, current-year basis, as 1990s on a more encouraging note than targeted seasoned loans between 1996 and explained earlier. Fannie Mae, the past four years (2000, 2001, 1998, and again during 2000–2002—all years BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00130 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63709
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00131 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.025
In general, the comparisons of Freddie during 2003, which was 1.4 percentage Table A.17a compares the first-time Mac’s and the market’s performance are points below market performance. homebuyer share of GSE purchases with similar to those discussed in Sections E.9.a- e. GSEs’ Purchases of First-Time Homebuyer corresponding share of home loans originated c above, except for some differences on the Mortgages—1999 to 2001 in the conventional conforming market. Readers are referred to recent work by Bunce special affordable category. The ‘‘Freddie While not a specific housing goal category, and Gardner 300 for the derivation of the Mac to market’’ ratios in Table A.16 show mortgages for first-time homebuyers are an that Freddie Mac has improved its important component of the overall home estimates of first-time homebuyer market performance but has also consistently lagged loan market. Making financing available for shares reported in Table A.17a. Between the primary market in funding mortgages first-time homebuyers is one approach for 1999 and 2001, first-time homebuyers covered by the housing goals. helping young families enter the accounted for 26.5 percent of Fannie Mae’s The ‘‘Fannie Mae to market’’ ratios in homeownership market. Therefore, this purchases of home loans, 26.5 percent of Table A.16 show that Fannie Mae has section briefly compares the GSEs’ funding of Freddie Mac’s, and 37.6 percent of home improved its performance, has generally first-time homebuyer loans with that of loans originated in the conventional outperformed Freddie Mac, and led the primary lenders in the conventional conforming market. Thus, both Fannie Mae market during 2003 on both the special conforming market. and Freddie Mac fell substantially short of affordable and low-mod goals. Under the During the past few years, the GSEs have the primary market in financing first-time origination-year approach, Fannie Mae increased their purchases of first-time homebuyers during this time period. The lagged the market on all three housing goal homebuyer loans. For example, Fannie Mae’s GSEs’ performance was only 70.5 percent of categories during 2001 and on the annual purchases of first-time homebuyer market performance (26.5 percent divided by loans increased from approximately 287,000 underserved area category during 2002. 37.6 percent). in 1999 to 423,485 in 2003.298 However, Fannie Mae matched the market in funding since 1999, the first-time homebuyer share of it then dropped to 30 percent in 1998 and to 26 special affordable loans during 2002 and led the GSEs’ purchases of home loans has the market in funding low-mod loans. During percent in 1999. The first-time homebuyer share for remained relatively flat, varying within the Freddie Mac was approximately 29 percent in 1996 2003, Fannie Mae led the primary market on 299 25–28 percent range. and 1997 before dropping to about 25 percent in both the special affordable and low-mod 1998 and 1999. categories but lagged the market on the 298 These figures include estimates of first-time 300 See Harold L. Bunce and John L. Gardner, underserved area category. For instance in homebuyer loans for those home purchase loans ‘‘First-time Homebuyers in the Conventional 2003, low- and moderate-income loans with a missing first-time homebuyer indicator; the Conforming Market: The Role of the GSEs’’ accounted for 47.0 percent of Fannie Mae’s estimates were obtained by multiplying the GSE’s (unpublished paper), January 2004. An update of purchases and 44.6 percent of the market first-time homebuyer share (based only on data with this work to include data for 2002 and 2003 shows a first-time homebuyer indicator) by the number of originations, placing Fannie Mae 2.4 similar patterns as those reported in the text for loans with a missing first-time homebuyer 1999–2001. See Harold L. Bunce and John L. percentage points above the market. On the indicator. Gardner, ‘‘First-time Homebuyers in the other hand, underserved areas accounted for 299 The first-time homebuyer share for Fannie Conventional Conforming Market: The Role of 26.3 percent of Fannie Mae’s purchases Mae was almost 35 percent between 1996 and 1998; GSEs: An Update’’ (October, 2004).
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00132 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63711
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00133 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.026
Table A.17a also reports first-time opportunities for lower-income homebuyers its 2002–3 performance while Fannie Mae homebuyer shares for African Americans and who are expected to enter the housing market would have to maintain its 2003 performance Hispanics and for all minorities. Between over the next few years. As detailed in of 47 percent. 1999 and 2001, African-American and Section I, there are four specific reasons for As explained earlier, HUD will be re- Hispanic first-time homebuyers accounted establishing this subgoal: (1) The GSEs have benchmarking its median incomes for for 4.0 percent of Fannie Mae’s purchases of the expertise, resources, and ability to lead metropolitan areas and non-metropolitan home loans, 3.4 percent of Freddie Mac’s the single-family-owner market, which is counties based on 2000 Census median purchases, and 6.9 percent of home loans their ‘‘bread and butter’’ business; (2) except incomes, and will be incorporating the effects for the recent performance of Fannie Mae, the originated in the conventional conforming of the new OMB metropolitan area GSEs have historically lagged the primary market. for this subgroup, Fannie Mae’s definitions. As shown in Table 17b, HUD performance is 58 percent of market market for low- and moderate-income loans, projected the effects of these two changes on performance, while Freddie Mac’s not led it; (3) the GSEs can improve their the low- and moderate-income shares of the performance is 49 percent of market funding of first-time homebuyers and help single-family-owner market for the years performance. The group of all minority first- reduce troublesome disparities in time homebuyers accounted for 6.6 percent homeownership and access to mortgage 1999–2003. These estimates will be referred of Fannie Mae’s purchases of home loans, 5.8 credit; and (4) there are ample opportunities to as ‘‘projected data’’ while the 1990-based percent of Freddie Mac’s purchases, and 10.6 for the GSEs to expand their purchases in data reported in the various tables will be percent of home loans originated in the important and growing market segments such referred to as ‘‘historical data.’’ With the conventional conforming market. In this case, as the market for minority first-time historical data, the average low-mod share of Fannie Mae’s performance is 62 percent of homebuyers. Sections E.9 and G of this the conventional conforming market (without market performance, while Freddie Mac’s appendix provide additional information on B&C loans) was 44.2 percent for home performance is 55 percent of market opportunities for an enhanced GSE role in purchase loans (weighted average of 1999– performance. the home purchase market and on the ability 2003 percentages in Table A.13); the Section E.12 below will continue this of the GSEs to lead that market. corresponding average with the projected examination of first-time homebuyers by As shown in Tables A.13 and A.15, low- data was 43.5 percent, a differential of 0.7 presenting market share analysis that and moderate-income families accounted for percentage points. However, note that in estimates the GSEs’ overall importance in the an average of 44.1 percent of home purchase 2003, the projected data for both GSEs and funding of first-time homebuyers. loans originated in the conventional the market exhibit higher low-mod shares conforming market of metropolitan areas than the corresponding historical data. For f. Low- and Moderate-Income Subgoal for between 1999 and 2003; the figure is 43.6 2003, the low-mod shares for the projected Home Purchase Loans percent if the average is computed for the The Department is proposing to years between 1996 and 2003 or 44.1 percent and historical data are as follows: Fannie establishing a subgoal of 45 percent for each if the average is computed for the more Mae (47.5 percent for the projected data GSE’s purchases of home purchase loans for recent 2001–2003 period. Loans in the B&C versus 47.0 percent for the historical data), low- and moderate-income families in the portion of the subprime market are excluded Freddie Mac (44.2 percent versus 43.8 single-family-owner market of metropolitan from these market averages. To reach the 45- percent), and the market (45.6 percent versus areas for 2005, with the subgoal rising to 46 percent subgoal for 2005, Freddie Mac would 44.6 percent). Thus, based on 2003 percent for 2006 and 47 percent for 2007 and have to improve its performance by one experience, it appears that the low-mod share 2008. If the GSEs meet this subgoal, they will percentage point over its approximately 44 for single-family-owners in the conventional be leading the primary market by percent low-mod performance during 2002 conforming market actually increase based approximately one percentage point in 2005 and 2003, while Fannie Mae would have to on the re-benchmarking of area median and by three percentage points in 2007–08, maintain its performance of 45–47 percent incomes and the new OMB definitions of based on historical data (see below). This over these two years. To reach the 47 percent metropolitan areas. Thus, based on 2003 home purchase subgoal will encourage the subgoal in 2007–08, Freddie Mac would have data, the 47-percent subgoal for 2007 is 2.4 GSEs to expand homeownership to improve by three percentage points over percentage points above the 2003 market.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00134 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63713
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00135 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.027
In terms of projected data, Fannie Mae available for metropolitan areas. HMDA data is appropriate given the importance of the could meet both the 2005 and 2007 subgoals for non-metropolitan areas are not reliable GSEs for expanding homeownership by maintaining its projected 2003 low-mod enough to serve as a market benchmark. The opportunities. To provide a complete picture performance of 47.5 percent. Freddie Mac’s Department is also setting home purchase of the GSEs’ mortgage purchases in projected low-mod performance for 2003 was subgoals for the other two goals-qualifying metropolitan areas, Tables A.18, A.19, A.20, 44.2 percent, about 0.4 percentage points categories, as explained in Appendices B and and A.21 report the GSEs’ purchases of all above its 2003 performance of 43.8 percent C. single-family-owner mortgages, including based on historical data. Thus, to reach the It should be noted that the findings in sub- 45-percent subgoal for 2005, Freddie Mac sections 9.a–e above concerning the both home purchase loans and refinance would have to increase its 2003 projected performance of the GSEs relative to the home loans.301 performance by 0.8 percentage point, and to purchase market do not change when reach the 47-percent 2007 subgoal, Freddie projected, rather than historical data, are Mac would have to increase its performance used. by 2.8 percentage points over its projected 10. GSEs Purchases of Total (Home Purchase performance of 44.2 percent for 2003. 301 The GSE total (home purchase and refinance) The subgoal applies only to the GSEs’ and Refinance) Loans data in Tables A.18–A.20 are presented on a purchases in metropolitan areas because the Section E.9 examined the GSEs’ purchase-year basis; Table A.21 presents similar HMDA-based market benchmark is only acquisitions of home purchase loans, which data on an origination-year basis.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00136 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63715
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00137 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.028 63716 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00138 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 Er02no04.029 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63717
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00139 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.030 63718 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00140 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.031 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63719
Table A.18 provides a long-run perspective performance between 1999 and 2000, but Freddie Mac significantly lagged the on the GSEs’ overall performance. Between they each fell back a little during the heavy single-family (home purchase and refinance 1993 and 2003, as well as during the 1996– refinancing year of 2001. But the primary loans combined) market during 2001–2003. 2003 period, the GSEs’ performance was 81– market (without B&C loans) experienced a In 2003, the ‘‘Freddie-Mac-to-market’’ ratios 91 percent of market performance for the much larger decline in affordable lending were 0.86 for special affordable loans, 0.98 special affordable category, 91–97 percent of during the refinancing wave than did either for low-mod loans, and 0.82 for underserved market performance for the low-mod of the GSEs. Fannie Mae stood out in 2001 area loans. category, and 87–93 percent of market because of its particularly small decline in Subprime Loans. Table A.14 in Section E.9 performance for the underserved areas affordable lending. Between 2000 and 2001, showed that the goals-qualifying shares of the category. For example, between 1996 and Fannie Mae’s special affordable lending fell home purchase market did not change much 2003, underserved areas accounted for 23.4 by only 0.6 percentage points while Freddie when originations by subprime lenders are percent of Fannie Mae’s purchases and 21.8 Mac’s fell by 2.8 percentage points and the excluded from the analysis; the reason is that percent of Freddie Mac’s purchases, market’s fell by 3.6 percentage points. The subprime lenders operate primarily in the compared with 25.2 percent for the corresponding percentage point declines for refinance market. Therefore, in this section’s conventional conforming market (without the underserved areas category were 1.0 for analysis of the total market (including B&C loans). Similarly, for special affordable Fannie Mae, 1.9 for Freddie Mac, and 3.8 for refinance loans), one would expect the loans, both GSEs lagged the market during the market. By the end of 2001, Fannie Mae treatment of subprime lenders to significantly the 1996–2003 period—Fannie Mae and led Freddie Mac in all three goals-qualifying affect the market estimates and, indeed, this Freddie Mac averaged approximately 13.0 categories, and had erased its gap with the is the case. For the year 2001, excluding percent while the market was over two low-mod market, but continued to lag the subprime loans reduced the goal-qualifying percentage points higher at 14.8 percent. market on the special affordable and shares of the total market as follows: special Similar to the patterns discussed for home underserved areas categories. affordable, from 15.0 to 13.9 percent; low- purchase loans, Fannie Mae has tended to During the refinancing wave of 2002, mod, from 42.3 to 40.9 percent; and outperform Freddie Mac. This can be seen by Fannie Mae improved slightly on the special underserved areas, from 25.7 to 23.9 percent. examining the various ‘‘Fannie-Mae-to- affordable and low-mod categories and (See Table A.19.) Similar declines take place Freddie-Mac’’ ratios in Table A.18, which are declined slightly on the underserved area in 2002 and 2003. all equal to or greater than one. Over the category. Freddie Mac showed slight As explained earlier, the comparisons in recent 1999–2003 period, Fannie Mae and improvement on the special affordable and this appendix have defined the market to Freddie Mac continued to lag the overall underserved area categories and remained exclude the B&C portion of the subprime market on all three goals-qualifying about the same on the low-mod category. The market. Industry observers estimate that A- categories. Special affordable (underserved result of these changes can be seen by minus loans account for about two-thirds of considering the market ratios in Table A.20. area) loans averaged 14.0 (23.8) percent of all subprime loans while the more risky B&C In 2002, special affordable loans accounted Fannie Mae’s purchases, 13.2 (22.1) percent loans account for the remaining one-third. As for 14.3 percent of Fannie Mae’s purchases of Freddie Mac’s purchases, and 15.0 (25.2) explained earlier, this analysis reduces the and 14.4 percent of loans originated in the percent of market originations. For Fannie goal-qualifying percentages from the HMDA non-B&C portion of the conventional Mae, the market ratio was 0.93 for special data by half the differentials between (a) the conforming market, yielding a ‘‘Fannie-Mae- affordable loans, 0.98 for low-mod loans, and to-market’’ ratio of 0.99. Since Fannie Mae’s market (unadjusted) and (b) the market 0.94 for underserved area loans. As with market ratio for the special affordable without the specialized subprime lenders home purchase loans, dropping the year 1999 category stood at 0.80 in 2000, Fannie Mae identified by Scheessele. As shown in Table and characterizing recent performance by the substantially closed its gap with the market A.19, accounting for B&C loans in this 2000–2003 period improves the performance during 2001 and 2002. During this period, manner reduces the year 2001 HMDA- of both GSEs relative to the market, but Fannie Mae also mostly eliminated its market reported goal-qualifying shares of the total particularly Fannie Mae. Over the 2000–2003 gap for the other two goals-qualifying (home purchase and refinance) conforming period, the ‘‘Fannie-Mae-to-market’’ ratio was categories. In 2002, underserved area loans market as follows: special affordable, from 0.97 for special affordable loans, 1.00 for low- accounted for 24.0 percent of Fannie Mae’s 15.0 to 14.5 percent; low-mod, from 42.3 to mod loans, and 0.96 for underserved area purchases and 24.2 percent of loans 41.6 percent; and underserved areas, from loans. Over the last three years (2001–2003), originated in the non-B&C portion of the 25.7 to 24.9 percent. Obviously, the GSEs’ the ‘‘Fannie-Mae-to-market’’ ratios are even conventional conforming market, yielding a performance relative to the market will higher—1.00 for special affordable loans, ‘‘Fannie-Mae-to-market’’ ratio of 0.99, or depend on which market definition is used 1.01 for low-mod loans, and 0.98 for approximately one. During 2002, low-mod (much as it did with the earlier examples of underserved area loans. In other words, loans accounted for 42.2 percent of Fannie excluding manufactured housing loans in during the first three years under the current Mae’s purchases and 42.0 percent of loans metropolitan areas from the market housing goal targets, Fannie Mae matched the originated in the market, yielding a ‘‘Fannie- definition). For example, defining the special affordable market, led the low-mod Mae-to-market’’ ratio of 1.00 (also note that conventional conforming market to exclude market, and lagged the underserved areas Fannie Mae slightly outperformed the low- subprime loans, rather than only B&C loans, market. mod market during 2001). Thus, during 2002, would increase Fannie Mae’s 2001 special The above analysis has defined the market Fannie Mae essentially matched the market affordable (underserved area) market ratio to exclude B&C loans. Table A.19 shows the on each of the three goals-qualifying from 0.96 to 1.00 (0.97 to 1.01). Similarly, it effects on the market percentages of different categories. would increase Freddie Mac’s special definitions of the conventional conforming In 2003, Fannie Mae’s continued to affordable (underserved area) market ratio market. For example, the average 1999–2003 improve its performance on the special from 0.92 to 0.96 (0.90 to 0.94). For the market share for special affordable affordable and low-mod categories. In 2003, broader-defined low-mod category, (underserved areas) loans would fall to 14.4 special affordable loans accounted for 14.3 redefining the 2001 market to exclude (24.8) percent if small loans and percent of Fannie Mae’s purchases and 14.0 subprime loans, rather than only B&C loans, manufactured housing loans in metropolitan percent of loans originated in the market, would increase Fannie Mae’s (Freddie Mac’s) areas were excluded from the market yielding a ‘‘Fannie-Mae-to-market’’ ratio of market ratio from 1.00 to 1.02 (0.97 to 0.98). definition along with B&C loans. In this case, 1.02. During that year, low-mod loans Table A.21 reports GSE purchase data for the market ratio for Fannie Mae (Freddie accounted for 42.3 percent of Fannie Mae’s total (home purchase and refinance) loans on Mac) would be was 0.97 (0.92) for special purchases and 41.2 percent of total (home an origination-year basis. The ‘‘Freddie Mac- affordable loans, 1.00 (0.95) for low-mod purchase and refinance) loans originated in to-market’’ ratios in Table A.21 show that loans, and 0.96 (0.89) for underserved area the market, yielding a ‘‘Fannie-Mae-to- Freddie Mac has lagged the primary market loans. market’’ ratio of 1.03. On the underserved in funding mortgages covered by the housing Shifts in performance occurred during areas category, Fannie Mae continued to lag goals. The ‘‘Fannie Mae-to-market’’ ratios in 2001–2003, the first three years under HUD’s behind the market (a 23.7 percent share for Table A.21 show that Fannie Mae has always higher housing goal targets. Table A.20 Fannie Mae and a 24.5 percent share for the lagged the primary market in funding home shows that both GSEs improved their overall market). purchase and refinance mortgages for
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00141 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63720 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
properties in underserved areas but, in 2002 are computed for each GSE’s purchases for 2001, 2002 and 2003 originations of home and 2003, led the low-mod market, and in each year and for each MSA. These two sets purchase loans. A GSE’s performance is 2003 led the special affordable market. of percentages are the same as those used in determined to be lagging the market if the 11. GSE Mortgage Purchases in Individual the aggregate analysis discussed in the above ratio of the GSE housing goal loan purchases Metropolitan Areas sections. Third, the ‘‘GSE-to-market’’ ratio is to their overall purchases is less than 99 While the above analyses, as well as earlier then calculated by dividing each GSE percent of that same ratio for the market. studies, concentrate on national-level data, it percentage by the corresponding market (The analysis was conducted where the ‘‘lag’’ is also instructive to compare the GSEs’ percentage. For example, if it is calculated determination is made at 98 percent instead purchases of mortgages in individual that one of the GSEs’ purchases of low- and of 99 percent and the results showed little metropolitan areas (MSAs). In this section, moderate-income loans in a particular MSA change.) In the example given in the above the GSEs’ purchases of single-family owner- is 40 percent of their overall purchases in occupied home purchase loans are compared that MSA, while 44 percent of all home loans paragraph, that GSE would be considered to the market in individual MSAs. There are (excluding B&C loans) in that MSA qualify as lagging the market. Tables A.22 (2001), A.23 three steps. First, goals-qualifying low-mod, then the GSE-to-market ratio is 40/ (2002) and A.24 (2003) report the number of percentages for conventional conforming 44 (or 0.91). The goals-qualifying ratios for MSAs in which each GSE under-performs the mortgage originations (without B&C loans) Fannie Mae and Freddie Mac can be market with respect to each of the three are computed for each year and for each compared for each MSA in a similar manner. housing goal categories. The following points MSA, based on HMDA data. Second, Tables A.22, A.23, and A.24 summarize the can be made from this data: corresponding goals-qualifying percentages performance of the GSEs within MSAs for BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00142 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63721
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00143 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.032 63722 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00144 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.033 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63723
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00145 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.034 63724 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Fannie Mae’s improvement between 2001 percentage) but also on the size, or overall loans purchased by the GSEs, a potentially and 2003 shows up in these tables. In 2001, mortgage volume, of the sector. If an industry important determinant of the GSEs’ ability to Fannie Mae lagged the market in 264 (80 sector has a large ‘‘market share’’ for a reach first-time homebuyers. percent) of the 331 MSAs in the purchase of targeted group, then that sector is making an a. GSEs’ Share of Home Purchase Lending underserved area loans; this number important contribution to meeting the credit decreased to 236 (71 percent) MSAs in 2002 needs of the group. Both ‘‘distribution of Table A.25 reports market share estimates and to 243 (73 percent) MSAs in 2003. business’’ and ‘‘market share’’ data are derived by combining HMDA market data Fannie Mae’s improvement was even greater important for evaluating the GSEs’ with GSE and FHA loan-level data. To for special affordable and low-mod loans; in performance. In fact, given the large size of understand these estimates, consider the GSE the latter case, Fannie Mae lagged the market the GSEs, one would expect that a ‘‘market market share percentage of 46 percent for in 51 (15 percent) MSAs in 2003, compared share’’ analysis would highlight their ‘‘All Home Purchase Loans’’ at the bottom of with 194 (59 percent) MSAs in 2001. importance to the affordable lending market. the first column in the table. That market Freddie Mac’s improvement between 2001 The second difference is that this section share percentage is interpreted as follows: and 2003 was greater for underserved area also examines the role of the GSEs in the It is estimated that home loans acquired by loans. In 2001, Freddie Mac lagged the total market for home loans, as well as in the Fannie Mae and Freddie Mac during the market in 261 (79 percent) of the 331 MSAs conventional conforming market. Such an years, 1999 to 2003, totaled 46 percent of all in the purchase of underserved area loans; approach provides a useful context for home loans originated in metropolitan areas this number decreased to 168 (51 percent) commenting on the contribution of Fannie during that period. Mae and Freddie Mac to overall affordable MSAs in 2002 before rising to 222 (67 It should be noted that ‘‘all home loans’’ lending, particularly given evidence that percent) MSAs in 2003. Freddie Mac’s made refers to all government (FHA and VA) loans conventional lenders have done a relatively less improvement on the special affordable plus all conventional loans less than the poor job providing credit access to and low-mod categories; in the former case, conforming loan limit; in other words, only Freddie Mac lagged the market in 234 (71 disadvantaged families, which renders the ‘‘jumbo loans’’ are excluded from this percent) MSAs in 2003, compared with 279 conventional market a poor benchmark for analysis.302 The analysis is restricted to (84 percent) MSAs in 2001. evaluating GSE performance. The analysis of first-time homebuyers conducts the market metropolitan areas because HMDA data (the 12. GSE Market Shares: Home Purchase and share analysis in terms of both the total source of the market estimates) are reliable First-Time Homebuyer Loans market (Section E.12.b) and the conventional only for metropolitan areas. B&C originations This section examines the role that the conforming market (Section E.12.c). are included in the market data, since the GSEs have played in the overall affordable While the GSEs have accounted for a large purpose here is to gauge the GSEs’ role in the lending market for home loans. There are two share of the overall market for home overall mortgage market. As discussed in differences from the above analyses in purchase loans, they have accounted for a Section E.9, excluding B&C loans, or even all Sections E.9 and E.10. The first difference is very small share of the market for important subprime loans, would not materially affect that this section focuses on ‘‘market share’’ groups such as minority first-time this analysis of the home loan market since percentages rather than ‘‘distribution of homebuyers. But as this section documents, subprime loans are mainly for refinance business’’ percentages. A ‘‘market share’’ the GSEs have been increasing their share of purposes. The analysis below frequently percentage measures the share of loans with the low-income and minority market, which combines purchases by Fannie Mae and a particular borrower or neighborhood provides an optimistic note on which to go Freddie Mac since previous sections have characteristic that is funded by a particular forward. already compared their performance relative market sector (such as FHA or the GSEs). In Section E.12.a uses HMDA and GSE data to each other. other words, a ‘‘market share’’ percentage to estimate the GSEs’ share of home loans measures a sector’s share of all home loans originated for low-income and minority 302 Following the purchase-year approach used in originated for a particular targeted group. The borrowers and their neighborhoods. Sections Sections E.9 and E.10, the GSE purchase data ‘‘market share’’ of a sector depends not only E.12.b and E.12.c summarize recent research include their acquisitions of ‘‘prior-year’’ as well as on the degree to which that sector on the role of the GSEs in the first-time ‘‘current-year’’ mortgages, while the market data concentrates its business on a targeted group homebuyer market. Section E.12.d examines include only newly-originated (or ‘‘current year’’) (i.e., its ‘‘distribution of business’’ the downpayment characteristics of home mortgages.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00146 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63725
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00147 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.035 63726 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
The GSE market share percentage for particularly high for African-American and (18 percent) of all home purchases.306 These ‘‘Low-Income Borrowers’’ at the top of the Hispanic borrowers, as FHA insured 29 data highlight the small role that the GSEs first column of Table A.25 has a similar percent of all home loans originated for these have played in the important market for interpretation: borrowers between 1999 and 2003—a figure minority first-time homebuyers. It is estimated that home loans for low- only one percentage point higher than the Bunce, Neal and Vandenbroucke (BNV) 304 income borrowers acquired by Fannie Mae GSEs’ share of 30 percent. Thus, during recently updated through 2001 the study by and Freddie Mac between 1999 and 2003 the 1999–2003 period, FHA’s overall market Bunce. In addition, BNV developed an share (16.0 percent) was about one-third of totaled 38 percent of all home loans improved methodology that combined the GSEs’ market share (45.6 percent), but its originated for low-income borrowers in industry, HMDA and AHS data to estimate metropolitan areas. share of the market for loans to African- the number of first-time homebuyers (by race According to the data in Table A.25, the Americans and Hispanics was almost equal GSEs account for a major portion of the to the GSEs’ share of that market. and ethnicity) in the mortgage market during 307 market for targeted groups. For example, The data for the two recent years (2002 and the years 1996 to 2001. BNV’s analysis purchases by Fannie Mae and Freddie Mac 2003) indicate a larger market role for Fannie includes the total mortgage market, that is, represented 38 percent of the low-income- Mae and Freddie Mac relative to FHA. While the government, conventional conforming, borrower market and 36–38 percent of the the GSEs continued to have a much larger and jumbo sectors of the mortgage market. markets in low-income, high-minority, and share of the overall market than FHA (47–49 Table A.26 presents the key market shares underserved census tracts. Thus, access to percent for the GSEs versus 11–14 percent for estimated by BNV for the GSEs and FHA. The credit in these historically underserved FHA), their share of home loans for African- first figure (40.7) in Table A.26 is interpreted markets depends importantly on the Americans and Hispanics jumped to 34 as follows: purchases of home loans by purchase activities of Fannie Mae and percent during 2002 and 2003, which was Fannie Mae and Freddie Mac totaled 40.7 Freddie Mac. However, the data in Table higher than the percentage share for FHA percent of all home loans financed between A.25 show that the GSEs’ role in low-income (17–25 percent). The differentials in market 1996 and 2001. Going down the first column and minority markets is significantly less share between FHA and the GSEs on the shows that the GSEs’ share of the first-time than their role in the overall home loan other affordable lending categories listed in homebuyer market was 24.5 percent during Table A.25 were also higher in 2002 and market. Fannie Mae and Freddie Mac the 1996-to-2001—a market share 2003 than in earlier years. accounted for 46 percent of all home loans significantly lower than their overall market but only 37 percent of the loans financing b. The GSEs’ Share of the Total First-Time share of 40.7 percent. properties in underserved neighborhoods. Homebuyer Market BILLING CODE 4210–27–P Their market share was even lower for loans This section summarizes two recent to African-American and Hispanic analyses of mortgage lending to first-time 306 Bunce explains numerous assumptions and borrowers—30 percent, or 16 percentage homebuyers; these two studies examine the points less than the GSEs’ overall market caveats related to combining American Housing total mortgage market, including both Survey data on homebuyers with FHA and GSE share of 46 percent. government and conventional loans data on mortgages. For example, the American An encouraging finding is that the GSEs originated throughout the U.S. (i.e., in both Housing Survey (AHS) data used by Bunce have increased their presence in the metropolitan areas and non-metropolitan included both financed home purchases and homes affordable lending market during 2002 and areas). Section E.12.c will summarize a third purchased with cash. If only financed home 2003, when they accounted for 40–44 percent study of first-time homebuyers that focuses purchases were used, the market shares of both of the loans financing properties in low- on the conventional conforming market. All FHA and the GSEs would have been slightly higher (although the various patterns would have income, high-minority, and underserved three studies are market share studies that neighborhoods and for 34 percent of loans for remained the same). The AHS defines first-time examine the GSEs’ role in the first-time homebuyers as buyers who have never owned a African-American and Hispanic borrowers. homebuyer market. home, while FHA and the GSEs define a first-time These market share figures for the GSEs are First, a study by Bunce concluded that the homebuyer more expansively as buyers who have generally higher than their performance GSEs have played a particularly small role in not owned a home in the past three years. If it were during the two earlier years, 2000 and 2001. funding minority first-time homebuyers.305 possible to re-define the FHA and GSE data to be To provide additional perspective, Table Because HMDA does not require lenders to consistent with the AHS data, the FHA and GSE A.25 also reports market share estimates for report information on first-time homebuyers, first-time homebuyer shares would be lower (to an FHA.303 During the 1999–2003 period, FHA’s Bunce used data from the American Housing unknown degree). For additional caveats with the overall market share was less than half of the AHS data, also see David A. Vandenbroucke, Sue Survey to estimate the number of first-time G. Neal, and Harold L. Bunce, ‘‘First-Time GSEs’ market share, as FHA insured only 16 homebuyers in the market. Using American Homebuyers: Trends from the American Housing percent of all home mortgages originated in Housing Survey data on home purchases Survey’’, November 2001, U.S. Housing Market metropolitan areas. However, FHA’s shares of from 1997 to 1999, Bunce estimated that the Condition, a quarterly publication of the Office of the underserved segments of the market were GSEs’ share of the market for first-time Policy Development and Research at HUD. In some much higher than its overall market share. African-American and Hispanic homebuyers years, home purchases as measured by the AHS For instance, between 1999 and 2003, FHA was only 10–11 percent, or less than one- declined while home purchases as measured by insured 24 percent of all mortgages third of their share (36 percent) of all home other data sources (e.g., HMDA) increased. In originated in low-income census tracts, even purchases during that period. FHA’s share of addition, the AHS home purchase data for separate minority groups (e.g., African-Americans, though it insured only 16 percent of all home this market was 36 percent, or twice its share Hispanics) sometimes exhibited shifts inconsistent loans. FHA’s share of the market was with other sources. 304 As explained in the notes to Table A.25, 307 BNV’s methodology for estimating first-time 303 As explained in Section E.7, the GSEs’ HMDA data are the source of the market figures. It borrowers consists of three steps: (1) estimate the affordable lending performance is evaluated relative is assumed that HMDA data cover 85 percent of all total number of home purchase loans originated to the conventional conforming market, as required mortgage originations in metropolitan areas. If during a particular year using a mortgage market by Congress in the 1992 GSE Act that established HMDA data covered higher (lower) percentages of model that they develop; (2) disaggregate the home the housing goals. However, it is insightful to market loans, then the market shares for both the purchase loans in step (1) into racial and ethnic examine their overall role in the mortgage market GSEs and FHA would be lower (higher). groups using HMDA data for metropolitan areas; and to contrast them with other major sectors of the 305 See Harold L. Bunce, The GSEs’ Funding of and (3) for each racial and ethnic group in step (2), market such as FHA. There is no intention here to Affordable Loans: A 2000 Update, Housing Finance estimate the number of first-time homebuyers using imply that the GSEs should purchase the same Working Paper No. HF–013, Office of Policy mortgage and first-time homebuyer information types of loans that FHA insures. Development and Research, HUD, April 2002. from the American Housing Survey.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00148 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63727
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00149 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.036
FHA’s greater focus on first-time two-fifths of the GSEs’ share of the overall year 2001, compared with the two previous homebuyers is also reflected in the market home purchase market (16.4 percent versus years, 1999 and 2000 (not reported). The share data reported in Table A.26. While 41.5 percent), FHA’s market share was over GSEs’ share of the market for first-time FHA insured only 16.6 percent of all home three times the GSEs’ share of the market for African-American and Hispanic homebuyers loans originated between 1996 and 2001, it first-time African-American and Hispanic jumped from about 11–12 percent during insured 30.9 percent of all first-time- homebuyers (46.5 percent versus 14.3 1999 and 2000 to 19.7 percent in 2001. homebuyer loans during that period. The percent). This finding that the GSEs have Fannie Mae’s share of this market almost GSEs, on the other hand, accounted for a played a relatively minor role in the first- doubled during this period, rising from 7.0 larger share (40.7 percent) of the overall time minority market is similar to the percent in 1999 to 12.6 percent in 2001. home purchase market but a smaller share conclusion reached by the Fed researchers Thus, while the GSEs continue to play a (24.5 percent) of the first-time homebuyer (see below) and Bunce (2002) that the GSEs relatively small role in the minority first-time market. have provided little credit support to this homebuyer market, during 2001 they Table A.26 also reports home purchase and underserved borrower group. improved their performance in this area.309 first-time homebuyer information for The results reported in Table A.26 for the c. The GSEs’ Share of the Conventional minorities. During the more recent 1999-to- year 2001 suggest some optimism concerning Conforming, First-time Homebuyer Market 2001 period, the GSEs’ loan purchases the GSEs’ role in the first-time homebuyer represented 41.5 percent of all home market. As explained in earlier sections, both Bunce and Gardner (2004) recently mortgages but only 24.3 percent of home GSEs, but particularly Fannie Mae, improved conducted an analysis of first-time loans for African-American and Hispanic their affordable lending performance during homebuyers for the conventional conforming families, and just 14.3 percent of home loans 2001, at a time when the overall market’s market. The Bunce and Gardner analysis for African-American and Hispanic first-time performance was slightly declining. This used a similar methodology to the study by homebuyers. During this period, the GSEs’ improvement is reflected in the higher first- Bunce, Neal, and Vandenbroucke of first-time role in the market for first-time African- time market shares for the GSEs during the homebuyers in the total mortgage market. American and Hispanic homebuyers was Bunce and Gardner restricted their analysis only one-third of their role in the overall the 1999–2001 market shares for FHA and the to the funding of first-time homebuyers in the home loan market (14.3 percent versus 41.5 conventional conforming market in metropolitan conventional conforming market, which is percent). areas calculated using the same methodology as the market where Fannie Mae and Freddie Table A.25 with (b) the 1999–2001 market share Mac operate. Their market share results are FHA, on the other hand, accounted for a estimates reported in Table A.25 for the entire much larger share of the minority first-time mortgage market (including jumbo loans and summarized in Table A.27. homebuyer market than it did of the overall covering non-metropolitan areas as well as homebuyer market. Between 1999 and 2001, metropolitan areas). The results are strikingly 309 For other analyses of the GSEs’ market role, FHA insured 46.5 percent of all loans for consistent. For the 1999-to-2001 period, the FHA see the following study by economists at the African-American and Hispanic first-time share of the overall (African American and Federal Reserve Board: Glenn B. Canner, Wayne homebuyers—a market share that was almost Hispanic) home loan market is estimated to be 19.0 Passmore, and Brian J. Surette, ‘‘Distribution of three times its overall market share of 16.4 percent (35.8 percent) under (a) versus 16.4 percent Credit Risk among Providers of Mortgages to Lower- 308 (31.2 percent) under (b). Lower percentage shares Income and Minority Homebuyers’’ in Federal percent. While FHA’s market share was are expected for (b) because (b) includes jumbo Reserve Bulletin, 82(12): 1077–1102, December, loans. For the same period, the GSE share of the 1996. This study considered several characteristics 308 See Bunce, Neal, and Vandenbroucke, op. cit., overall (African American and Hispanic) home loan of the GSEs’ loan purchases (such as amount of for comparisons of various estimates of the market market is estimated to be 46.0 percent (25–28 downpayment) and concluded that the GSEs have shares for FHA and the GSEs using different data percent) under (a) versus 41.5 percent (24.3 percent) played a minimal role in providing credit support bases and estimation methods. One can compare (a) under (b). for underserved borrowers.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00150 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63729
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00151 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.037
Between 1999 and 2001, the GSEs’ for 56.6 percent of all home loans in the accounted for 6.9 percent of Fannie Mae’s purchases accounted for 56.6 percent of all conventional conforming market, they have purchases of home loans between 1999 and home loans originated in the conventional accounted for only 30.9 percent of loans 2001, a figure higher than Freddie Mac conforming market of both metropolitan originated in that market for African- percentage of 5.3 percent. Loans for African- areas and non-metropolitan areas. In other American and Hispanic first-time American and Hispanic first-time words, Fannie Mae and Freddie Mac funded homebuyers. homebuyers accounted for 10.2 percent of all almost three out of every five homebuyers The market share data in Table A.27 show home loans originated in the conventional entering the conventional conforming market some slight differences between the Freddie conforming market. between 1999 and 2001. Their purchases of Mac and Fannie Mae in serving minority d. Downpayments on Loans Purchased by the first-time homebuyer loans, on the other first-time homebuyers. During the 1999-to- GSEs hand, accounted for only 39.8 percent of all 2001 period, Freddie Mac’s share (11.9 first-time homebuyer loans originated in that percent) of the African-American and The level of downpayment can be an market. Thus, while the GSEs funded Hispanic first-time homebuyer market was important obstacle to young families seeking approximately two out of every five first-time only one-half of its share (24.0 percent) of the their first homes. Examining the homebuyers entering the conventional home loan market. On the other hand, Fannie downpayment characteristics of the conforming market, their market share (39.8 Mae’s share (19.0 percent) of the African- mortgages purchased by the GSEs might help percent) for first-time homebuyers was only American and Hispanic first-time homebuyer explain why they have played a rather 70 percent of their market share (56.6 market was almost 60 percent of its share limited role in the first-time homebuyer percent) for all homebuyers. (32.5 percent) of the home loan market. Thus, market As shown in Table A.27, the GSEs have while Fannie Mae performance in serving Table A.28 reports the loan-to-value (LTV) funded an even lower share of the minority minority first-time homebuyers has been distribution of home purchase mortgages first-time homebuyer market. Between 1999 poor, it has been better than Freddie Mac’s. acquired by the GSEs between 1997 and and 2001, the GSEs purchases of African- This difference in performance between 2003. In Table A.29, LTV data are provided American and Hispanic first-time homebuyer Fannie Mae and Freddie Mac was also seen for the GSEs’ purchases of home loans that loans represented 30.9 percent of the in the portfolio percentages reported earlier qualify for the three housing goals—special conventional conforming market for these in Table A.17a. Loans for African-American affordable, low-mod, and underserved areas. loans. Thus, while the GSEs have accounted and Hispanic first-time homebuyers Three points stand out.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00152 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63731
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00153 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.038 63732 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00154 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.039 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63733
First, the GSEs (and particularly Fannie predominance of loans with high purchase mortgages appeared to boost prices Mae) have recently increased their purchases downpayments. For example, 54.3 percent of one and two years later, no consistent of home loans with low downpayments. special affordable home loans purchased by impacts of purchasing rates on sales prices After remaining about 4 percent of Fannie Freddie Mac during 2003 had a could be observed. In addition, there was no Mae’s purchases between 1997 and 2000, downpayment of at least 20 percent, a robust evidence that GSE purchasing rates over-95-percent-LTV loans (or less-than-five- percentage not much lower than the high- were positively associated with single-family percent downpayment loans) jumped to 7.1 downpayment share (59.5 percent) of all home transactions volumes or sales prices percent during 2001, 7.7 percent in 2002 and Freddie Mac’s home loan purchases. during any periods. 11.5 percent in 2003. It is interesting that this Similarly, 49.8 percent of the home loans Urban Institute Rural Markets Study.312 A jump in less-than-five-percent downpayment purchased by Fannie Mae in underserved study by Jeanette Bradley, Noah Sawyer, and loans occurred in the same years that Fannie areas during 2003 had a twenty percent or Kenneth Temkin uses both quantitative and Mae improved its purchases of loans for low- higher downpayment, compared with 54.6 qualitative data to explore the issue of GSE income homebuyers, as discussed in earlier percent of all home loans purchased by service to rural areas. The study first sections. As a share of Freddie Mac’s Fannie Mae. summarizes the existing research on rural purchases, over-95-percent-LTV loans Thus, the data in Tables A.28 and A.29 lending and GSE service to rural areas. It increased from 1.1 percent in 1997 to 5.9 show a preponderance of high downpayment then reviews the current underwriting percent in 2000, before falling to 4.3 percent loans, even among lower-income borrowers guidelines of Fannie Mae, Freddie Mac, the in 2001, 4.8 percent in 2002 and 4.7 percent who qualify for the housing goals. The past USDA Rural Housing Service, and Farmer in 2003. If the low-downpayment definition focus of the GSEs on high-downpayment Mac, focusing on issues relevant to rural is expanded to ten percent (i.e., over-90- loans provides some insight into a study by underwriting. The GSE public-use database is percent-LTV loans), Freddie Mac had about staff at the Federal Reserve Board who found used to analyze GSE non-metro loan the same percentage (25 percent) of low- that the GSEs have offered little credit purchasing patterns from 1993–2000. Finally, downpayment loans during 2001 as Fannie support to the lower end of the mortgage the study presents the results of a series of Mae. In fact, under the more expansive market.310 The fact that approximately half of discussions conducted with key national definition, Freddie Mac had the same share the goals-qualifying loans purchased by the industry and lender experts and local experts of over-90-percent-LTV loans in 2001 as it GSEs have a downpayment of over twenty in three rural sites in south-central Indiana, did in 1997 (about 25 percent), while Fannie percent is also consistent with findings southwestern New Mexico and southern New Mae exhibited only a modest increase in the reported earlier concerning the GSEs’ Hampshire chosen for the diversity of their share of its purchases with low minimal service to first-time homebuyers, region, population, economic structures, and downpayments (from 23.2 percent in 1997 to who experience the most problems raising housing markets. 25.4 percent in 2001). The share of over-90- cash for a downpayment. On the other hand, The authors of the study conclude that percent-LTV loans in Freddie Mac’s the recent experience of Fannie Mae suggests while Fannie Mae and Freddie Mac have purchases fell sharply from 25.0 percent in that purchasing low-downpayment loans increased their lending to rural areas since 2001 to 21.9 percent in 2002 and 19.9 percent may be one technique for reaching out and 1993, their non-metro loan purchases still lag in 2003, while the share in Fannie Mae’s funding low-income and minority families behind their role in metro loan purchases, purchases fell more modestly from 25.4 who are seeking to buy their first home. particularly in regard to the percentage of percent in 2001 to 24.2 percent in 2002 affordable loans. From the discussions with before rebounding to 25.3 percent in 2003. 13. Other Studies of the GSEs’ Performance Relative to the Market experts, the authors of the study make the Second, loans that qualify for the housing following policy recommendations: goals have lower downpayments than non- This section summarizes briefly the main Underserved populations and rural areas qualifying loans. In 2001 and 2002, over-95- findings from other studies of the GSEs’ should be specifically targeted at the census- percent-LTV loans accounted for about 15 affordable housing performance. These tract level; HUD should set manufactured percent of Fannie Mae’s purchases of special include studies by the HUD and the GSEs as housing goals; HUD should consider affordable loans, 13 percent of low-mod well as studies by academics and research implementing a survey of small rural lenders loans, and 12 percent of underserved area organizations. or setting up an advisory group of small rural 311 loans, compared with about 7.5 percent of Freeman and Galster Study. A recent lenders in order to determine their Fannie Mae’s purchases of all home loans. study by Lance Freeman and George Galster suggestions for creating stronger (See Table A.29.) In 2003 these percentages uses econometric analysis to test whether the relationships between the GSEs and rural increased to 23, 19 and 19 percent for special Government-Sponsored Enterprises (GSEs) lenders with the goal of increasing GSE non- affordable, low-mod and underserved areas Fannie Mae and Freddie Mac purchases of metro purchase rates. respectively. These low-downpayment shares home mortgages in neighborhoods Urban Institute GSE Impacts Study.313 A for 2001, 2002 and 2003 were double those traditionally underserved by financial report by Thomas Thibodeau, Brent for 2000 when over-95-percent-LTV loans institutions stimulate housing market activity Ambrose, and Kenneth Temkin analyzes the accounted for 8.4 percent of Fannie Mae’s in those neighborhoods. Specifically, this extent to which the GSEs’ responses to the purchases of special affordable loans and study analyzes data of single-family home Federal Housing Enterprises Financial Safety about 7 percent of its purchases of low-mod sales volumes and prices of mortgages and Soundness Act’s (FHEFSSA) affordable and underserved area loans. Fannie Mae’s originated from 1993–1999 in Cleveland, OH. housing goals have had their intended effect low-downpayment shares during 2001 were The study concludes that aggressive of making low- and moderate-income higher than Freddie Mac’s shares of 12.3 secondary market purchasing behavior by families better off. Specifically the report percent for special affordable loans and about non-GSE entities stimulated sales volumes examines several methodologies determining 9 percent for low-mod and underserved area and prices of homes in low-income and that the conceptual model created by Van loans. Between 2001 and 2003, Freddie Mac’s predominantly minority-occupied Order in 1996 314 provided the most complete over-95-percent-LTV shares fell sharply to 3– neighborhoods of Cleveland. The study description of how the primary and 4 percent for the three housing goal results also showed a positive relationship secondary markets interact. This model was categories, while Fannie Mae’s shares between home transaction activity and the then applied in a narrow scope to capital increased to the 13–23 percent range. Under actions of the secondary mortgage market, market outcomes which included GSE the more expansive, over-90-percent-LTV and concludes that the secondary mortgage market shares and effective borrowing costs, definition, almost one-third of Fannie Mae’s market (and the non-GSE sector in particular) goals-qualifying purchases during 2001 purchases of mortgages had a positive effect 312 would be considered low downpayment, as on the number of sales transactions one year GSE Service to Rural Areas, 2002. 313 would a slightly smaller percentage of later. However, the study also concludes that An Analysis of the Effects of the GSE Affordable Goals on Low- and Moderate-Income Freddie Mac’s purchases. However, during although non-GSE purchases of non-home Families, 2001. 2003, Freddie Mac’s over-90-percent-LTV 314 Van Order, Robert. 1996. ‘‘Discrimination and shares for the goals-qualifying loans fell to 310 Canner, et al., op. cit. the Secondary Mortgage Market.’’ In John Goering 20–22 percent. 311 The Impact of Secondary Mortgage Market and and Ronald Wienk, eds. Mortgage Discrimination, Third, a noticeable pattern among goals- GSE Purchases on Underserved Neighborhood Race, and Federal Policy. The Urban Institute Press, qualifying loans purchased by the GSEs is the Housing Markets: Final Report to HUD. July 2002. Washington, DC: 335–363.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00155 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63734 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
and housing market outcomes that include the Public Use Data Base at HUD for characteristics across properties and low- and moderate-income homeownership compilations of GSE data sets for the entire individuality of owners. This makes it rates. Finally, metropolitan American nation (GSE PUDB File B) to conduct difficult for lenders to properly evaluate the Housing Survey (AHS) data for eight cities descriptive and multivariate analyses of probability of default and severity of loss for were used to conduct empirical analyses of nationwide lending between 1993 and 2000. these properties. the two categories of outcomes. These cities Additionally, separate analyses are Single-family rental properties could be included areas surveyed in 1992, the year conducted that include and exclude loans important for the GSEs housing goals, before HUD adopted the affordable housing from subprime and manufactured housing especially for meeting the needs of lower- goals, to provide the baseline for the analysis. lenders. income families. Between 1999 and 2002, 87 Four metropolitan areas were surveyed in The study concludes that the GSEs are not percent of the GSEs’ single-family rental 1992 and again in 1996: Cleveland, leading: They do not purchase relatively units qualified for the Low- and Moderate- Indianapolis, Memphis and Oklahoma City. more underserved market loans than the Income Goal, compared with 40 percent of Four cities were surveyed in 1992 and again primary market makes nor do they purchase one-family owner-occupied properties. (See in 1998: Birmingham, Norfolk, Providence as many of these loans as their secondary Table A.30.) This heavy focus on lower- and Salt Lake City. market competitors. Additionally, the study income families meant that single-family The study’s empirical analysis suggests concludes that the disparities between the rental properties accounted for 14 percent of that the GSE affordable goals have helped to GSEs and the primary market are even greater the units qualifying for the Low- and make homeownership more attainable for once the growing role of subprime and Moderate-Income Goal, even though they target families. The assessment of the effects manufactured housing is considered. The accounted for 7.6 percent of the total units of the affordable goals on capital markets authors admit that there have been signs of (single-family and multifamily) financed by showed that the GSE share of the progress, particularly in 1999 and 2000 when the GSEs. conventional conforming market has primary market lending to underserved Given the large size of this market, the high increased, especially for lower income markets increased and GSE purchases of percentage of these units which qualify for borrowers and neighborhoods. The study also underserved market loans increased even the GSEs’ housing goals, and the weakness of concludes that the affordable housing goals faster. Regardless, the study concludes that the secondary market for mortgages on these have an impact on the purchase decisions of there continues to be significant racial, properties, an enhanced presence by Fannie Fannie Mae and Freddie Mac. The study also economic, and geographic disparities in the Mae and Freddie Mac in the single-family finds that interest rates are lower in markets way that the benefits of GSE activities are rental mortgage market would seem in which Fannie Mae and Freddie Mac distributed and that the benefits of GSE warranted.316 Single-family rental housing is purchase a higher proportion of conventional activities still go disproportionately to an important part of the housing stock loans. Finally, the study’s analysis shows members of served rather than underserved because it is an important source of housing that overall lending volume in a metropolitan markets. for lower-income households. area increases when the GSEs purchase 14. The GSEs’ Support of the Mortgage Despite the size and importance of single- seasoned loans. Market for Single-Family Rental Properties family rental properties for low-income Specifically, that homeownership rates people, HUD received several comments increased at a faster rate for low-income The 1996 Property Owners and Managers advocating exclusion of single-family rentals families when compared to all families, and Survey reported that 49 percent of rental from goals consideration. These commenters that in a subset of MSAs, minority units are found in the ‘‘mom and pop shops’’ pointed out that single-family owner- homeownership rates also grew faster when of the rental market—‘‘single-family’’ rental occupiers often maintain their properties compared to overall homeownership changes properties, containing 1–4 units. These small more effectively than single-family absentee in those MSAs. properties are largely individually-owned landlords or their tenants. HUD was asked to Finally, the affordable housing goal effects and managed, and in many cases the owner- exclude single-family investor owned were examined for 80 MSAs in relation to the managers live in one of the units in the properties to reduce these neighborhood homeownership rate changes between 1991 property. They include many properties in effects. and 1997. The study found that the GSEs, by older cities, in need of financing for Community associations raise an important purchasing loans originated to low-income rehabilitation. Single-family rental units play issue for neighborhood development. families, helped to reduce the disparity an especially important role in lower-income However, they do not address the question of between homeownership rates for lower and housing, over half of such units are effective goals promotion for all segments of higher income families, suggesting that the affordable to very low-income families. the housing market. They compare liquidity created when the GSEs purchase There is not, however, a strong secondary maintenance by owner-occupiers to loans originated to low-income families is market for single-family rental mortgages. maintenance by investors in the single-family recycled into more lending targeted to lower While single-family rental properties market. This does not address the housing income homebuyers. comprise a large segment of the rental stock outcomes for tenants with access to single- The authors of the study qualify their for lower-income families, they make up a family rental compared to tenants in results by stating that they are based on small portion of the GSEs’ business. Between multifamily rental. With nearly half of rental available data that does not provide the level 1999 and 2002, single-family rental units in older cities composed of smaller of detail necessary to conduct a fully properties accounted for only 7.6 percent of single-family units, denial of goals eligibility controlled national assessment. total (both single-family and multifamily) for single-family investors would exclude a Williams and Bond Study.315 Richard units financed by the GSEs during this substantial proportion of housing units Williams and Carolyn Bond examine GSE period. It follows that since single-family available to low income people. leadership of the mortgage finance industry rentals make up such a small part of the GSEs Furthermore, single-family investors in making credit available for low- and business, they have not penetrated the single- provide additional market benefits to the moderate-income families. Specifically, it family rental market to the same degree that housing system. The whole structure of the asks if the GSEs are doing relatively more of they have penetrated the owner-occupant GSEs provides liquidity to the housing their business with underserved markets than market. Table A.30 below shows that market by allowing investors additional other financial institutions, and whether the between 1999 and 2002, the GSEs financed channels to fund mortgages. The question is GSEs’ leadership helps to narrow the gap in 61 percent of owner-occupied dwelling units not always between single-family investors home mortgage lending that exists between in the conventional conforming market, but and single-family owner-occupiers. served and underserved markets. The study only 40 percent of single-family rental units. Sometimes, the question is between a single- uses HMDA data for metropolitan areas and There are a number of factors that have limited the development of the secondary 316 market for single-family rental property A detailed discussion of the GSEs’ activities in 315 Are the GSEs Leading, and if So Do They Have this area is contained in Theresa R. Diventi, The Any Followers? An Analysis of the GSEs’ Impact on mortgages thus explaining the lack of GSEs’ Purchases of Single-Family Rental Property Home Purchase Lending to Underserved Markets penetration by the GSEs. Little is collectively Mortgages, Housing Finance Working Paper No. During the 1990s. University of Notre Dame known about these properties as a result of HF–004, Office of Policy Development and Working Paper and Technical Series Number 2003– the wide spatial dispersion of properties and Research, Department of Housing and Urban 2. 2002. owners, as well as a wide diversity of Development, (March 1998).
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00156 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63735
family investor and a property unable to be lenders in funding first-time homebuyers, A.31b and A.31c repeat this information but sold or even abandoned. Although the goals lower-income borrowers and underserved for lower multifamily shares of the mortgage strongly support home ownership for low- communities, although Fannie Mae’s recent market.) HUD estimates that there were income neighborhoods, investors in single- performance has placed it ahead of the 47,551,039 owner and rental units financed family properties also play an important role. special affordable and low-mod markets for by new conventional conforming mortgages single-family-owner loans. As required by between 1999 and 2002. Fannie Mae’s and F. Factor 4: Size of the Conventional FHEFSSA, the Department has produced Conforming Mortgage Market Serving Low- Freddie Mac’s mortgage purchases financed estimates of the portion of the total (single- 26,118,927 of these dwelling units, or 55 and Moderate-Income Families Relative to family and multifamily) mortgage market that the Overall Conventional Conforming Market percent of all dwelling units financed. As qualifies for each of the three housing goals shown in Table A.30, the GSEs have played The Department estimates that dwelling (see Appendix D). Congress intended that the a smaller role in the goals-qualifying markets units serving low- and moderate-income Department use these market estimates as than they have played in the overall market. families will account for 51–56 percent of one factor in setting the percentage target for Between 1999 and 2002, new mortgages were total units financed in the overall each of the housing goals. The Department’s originated for 26,051,771 dwelling units that conventional conforming mortgage market estimate for the size of the Low- and qualified for the Low- and Moderate-Income during 2005–2008, the period for which the Moderate-Income market is 51–56 percent, Low- and Moderate-Income Housing Goal which is higher than the GSEs’ performance Goal; the GSEs low-mod purchases financed will be effective. The market estimates on that goal. 12,608,215 dwelling units, or 48 percent of exclude B&C loans and allow for much more This section provides another perspective the low-mod market. Similarly, the GSEs’ adverse economic and market affordability on the GSEs’ performance by examining the purchases accounted for 48 percent of the conditions than have existed recently. The share of the total conventional conforming underserved areas market, but only 41 detailed analyses underlying these estimates mortgage market and the share of the goal- percent of the special affordable market. are presented in Appendix D. qualifying markets (low-mod, special Obviously, the GSEs did not lead the affordable, and underserved areas) accounted industry during this period in financing units G. Factor 5: GSEs’ Ability to Lead the for by the GSEs’ purchases. This analysis, that qualify for the three housing goals. They Industry which is conducted by product type (single- need to improve their performance and it FHEFSSA requires the Secretary, in family owner, single-family rental, and appears that there is ample room in the non- determining the Low- and Moderate-Income multifamily), shows the relative importance GSE portions of the goals-qualifying markets Housing Goal, to consider the GSEs’ ability of the GSEs in each of the goal-qualifying for them to do so. For instance, the GSEs to ‘‘lead the industry in making mortgage markets. were not involved in three-fifths of the credit available for low- and moderate- 1. GSEs’ Role in Major Sectors of the special affordable market during the 1999-to- income families.’’ Congress indicated that Mortgage Market 2002 period. this goal should ‘‘steer the enterprises toward BILLING CODE 4210–27–P the development of an increased capacity Tables A.30 and A.31a compare GSE and commitment to serve this segment of the mortgage purchases with HUD’s estimates of housing market’’ and that it ‘‘fully expect[ed] the numbers of units financed in the GSEs’ purchases of a particular origination year conventional conforming market. Table A.30 cohort through 2003. This approach contrasts with [that] the enterprises will need to stretch the approach that examines GSE purchases on a their efforts to achieve [these goals].’’317 presents aggregate data for 1999–2002 while Table A.31a presents more summary market ‘‘backward looking basis by purchase year’’, for The Department and independent example, GSE purchases during 2000 of both new researchers have published numerous studies share data for individual years 2000, 2001 318 2000 originations and originations during previous examining whether or not the GSEs have and 2002. (As explained below, Tables years (the latter called ‘‘prior-year’’ or seasoned been leading the single-family market in loans). Either approach is a valid method for terms of their affordable lending 318 Tables A.30 and A.31 examine GSE purchases examining GSE purchases; in fact, when analyzing performance. This research, which is on a ‘‘going forward basis by origination year’’. aggregated data such as the combined 1999–2002 summarized in Section E, concludes that the Specifically, it considers GSE purchases of: (a) 2000 data in Table A.30, the two approaches yield mortgage originations during 2000, 2001, 2002 and somewhat similar results. HUD’s methodology for GSEs have generally lagged behind primary 2003; (b) 2001 originations during 2001, 2002 and deriving the market estimates is explained in 2003; and (c) 2002 originations during 2002 and Appendix D. B&C loans have been excluded from 317 Senate Report 102–282, May 15, 1992, p. 35. 2003. In other words, this analysis looks at the the market estimates in Tables A.30 and A.31.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00157 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63736 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00158 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.040 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63737
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00159 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.041 63738 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00160 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.042 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63739
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00161 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.043 63740 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
While the GSEs are free to meet the and, thus, more are sold to the GSEs. As a single-family-owner loans accounted for 83 Department’s goals in any manner that they result, during low interest rate periods, the percent of all dwelling units financed by the deem appropriate, it is useful to consider GSE share of mortgages increases. GSEs during this period, versus 75 percent of their performance relative to the industry by Single-Family Rental Market. Single-family all units financed in the conventional property type. The GSEs accounted for 61 rental housing is a major source of low- conforming market. percent of the single-family owner market but income housing. As discussed in Appendix While it is recognized that the GSEs have only 35 percent of the multifamily market D, data on the size of the primary market for been increasing their multifamily purchases, and 40 percent of the single-family rental mortgages on these properties is limited, but a further enlargement of their role in the market (or a combined 37 percent share of available information indicate that the GSEs multifamily market seems feasible and the rental market). are much less active in this market than in appropriate, particularly in the affordable Single-Family Owner Market. As stated in the single-family owner market. HUD (lower rent) end of the market. As noted in the 2000 Rule, the single-family-owner estimates that GSE purchases between 1999 Section D.3, market participants believe that market is the bread-and-butter of the GSEs’ and 2002 totaled only 40 percent of all the GSEs have been conservative in their business, and based on the financial and newly-mortgaged single-family rental units approaches to affordable multifamily lending other factors discussed below, the GSEs that were affordable to low- and moderate- and underwriting.320 Certainly the GSEs face clearly have the ability to lead the primary income families. a number of challenges in better meeting the market in providing credit for low- and As explained in the 2000 Rule, many of needs of the affordable multifamily market. moderate-income owners of single-family these properties are ‘‘mom-and-pop’’ For example, thrifts and other depository properties. However, the GSEs have operations, which may not follow financing institutions may sometimes retain their best historically lagged behind the market in procedures consistent with the GSEs’ loans in portfolio, and the resulting funding single-family-owner loans that guidelines. Much of the financing needed in information asymmetries may act as an qualify for the housing goals and, as this area is for rehabilitation loans on 2–4 impediment to expanded secondary market discussed in Section E, they have played a unit properties in older areas, a market in transaction volume.321 However, the GSEs rather small role in funding minority first- which the GSEs’ have not played a major have demonstrated that they have the depth time homebuyers. The market share data role. However, this sector could certainly of expertise and the financial resources to reported in Table A.30 for the single-family- benefit from an enhanced role by the GSEs, devise innovative solutions to problems in owner market tell the same story. The GSEs’ and the data in Table A.30 indicate that there the multifamily market. The GSEs can build purchases of single-family-owner loans is room for such an enhanced role, as on their recent records of increased represented 61 percent of all single-family- approximately two-thirds of this market multifamily lending and innovative products owner loans originated between 1999 and remains for the GSEs to enter. to make further in-roads into the affordable 2002, compared with 57 percent of the low- Once again, Table A.31a shows a large market. As explained in Section D.3, the mod loans that were originated, 55 percent of increase in the GSE share of newly- GSEs have the expertise and market presence underserved area loans, and 52 percent of the mortgaged units financed in 2002 compared to push simultaneously for market special affordable loans. to those financed in 2000 and 2001. As standardization and for programmatic The data in Table A.31a indicate the GSEs’ described above for the single-family owner flexibility to meet the special needs and growing market share during the heavy market, this large increase is due to the large circumstances of the lower-income portion of refinance years of 2001 and 2002. For share of fixed-rate mortgages, compared to the multifamily market. example, the GSEs accounted for 74 percent adjustable rate mortgages, originated during As discussed in Appendix D, the GSEs of the overall single-family-owner market in 2002. questioned HUD’s historical estimates of the 2002, and 67–69 percent of the markets Multifamily Market. Fannie Mae is the multifamily market as too high. Section C of covered by the three housing goal categories. largest single source of multifamily finance Appendix D discusses these comments and While this improvement is an encouraging in the United States, and Freddie Mac has responds. As indicated in Table A.30, trend, there are ample opportunities for the made a solid reentry into this market over the multifamily loans accounted for 14.8 percent GSEs to continue their improvement. Almost last nine years. However, there are a number of all financed units in the market, excluding one-third of the goals-qualifying loans of measures by which the GSEs lag the B&C loans. As reported in Appendix D, HUD originated during 2002 remained available to multifamily market. For example, the share also conducted sensitivity analyses that the GSEs to purchase; there are clearly of GSE resources committed to the reduced its 1999–2002 multifamily shares for affordable loans being originated that the multifamily purchases falls short of the the market by approximately two percentage GSEs can purchase. Furthermore, the GSEs’ multifamily proportion prevailing in the points. The results for these lower purchases under the housing goals are not overall mortgage market. HUD estimates that multifamily market shares are reported in limited to new mortgages that are originated newly-mortgaged units in multifamily Table A.31b (1999–2002 aggregate results) in the current calendar year. The GSEs can properties represented almost 15 percent of and Table A.31c (2000–2002 individual year purchase loans from the substantial, existing all (single-family and multifamily) dwelling results). In this case, 1999–2002 multifamily stock of affordable loans held in lenders’ units financed between 1999 and 2002.319 As units decreased from 7,018,044 units to portfolios, after these loans have seasoned shown in Table A.30, multifamily 5,991,036 units (reducing the multifamily and the GSEs have had the opportunity to acquisitions represented 9.5 percent of share from 14.8 percent to 12.9 percent). observe their payment performance. In fact, dwelling units financed by the GSEs between With these reduced multifamily market based on Fannie Mae’s recent experience, the 1999 and 2002. numbers, the GSEs’ share of the multifamily purchase of seasoned loans appears to be one The GSEs’ role in the multifamily market market increased from 35 percent to 41 effective strategy for purchasing goals- is significantly smaller than in single-family. percent. The GSEs also accounted for higher shares of the goals-qualifying multifamily qualifying loans. As shown in Table A.30, GSE purchases have market: 42 percent for low-mod units, 34 The data in Table A.31a show a strong accounted for 35 percent of newly financed percent for underserved area units, and 37 upward trend from 2000 and 2001 to 2002 in multifamily units between 1999 and 2002— percent for special affordable units. In this the GSE share of the single-family-owner a market share much lower than their 61 case, the GSEs’ shares of the overall goals- market. Their share of 2000 financed units in percent share of the single-family-owner qualifying markets increased as follows: low- the conventional conforming market totaled market. Stated in terms of portfolio shares, mod—from 48 percent (see right column of 48 percent. This increased to 55 percent in Table A.30) to 50 percent (see right column 2001 then to 74 percent in 2002. The large 319 Based on Table A.30, multifamily properties increase in 2002 can be attributed to the represented 14.8 percent of total units financed 320 relatively low interest rates and heavy between 1999 and 2002 (obtained by dividing Abt Associates, op. cit. (August 2002). 321 refinancing activity in 2003. During such a 7,018,044 multifamily units by 47,551,039 ‘‘Total The problem of secondary market ‘‘adverse period, the share of fixed rate mortgage Market’’ units). Increasing the single-family-owner selection’’ is described in James R. Follain and number in Table A.30 by 2,648,757 to account for Edward J. Szymanoski. ‘‘A Framework for originations increases relative to adjustable excluded B&C mortgages increases the ‘‘Total Evaluating Government’s Evolving Role in rate mortgages. Due to the higher risk Market’’ number to 50,199,796, which produces a Multifamily Mortgage Markets,’’ Cityscape: A associated with fixed rate mortgages, less multifamily share of 14.0 percent. See Appendix D Journal of Policy Development and Research 1(2), thrift institutions are willing to hold them, for discussion of the B&C market. 1995.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00162 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63741
of Table A.31b); underserved areas—from 48 b. Underwriting Standards for the Primary income borrowers and the ‘‘black box’’ nature percent to 49 percent; and special Mortgage Market of the score algorithm. affordable—from 41 percent to 43 percent. The GSEs’ underwriting guidelines are The GSEs are using their state-of-the-art Conclusions. While HUD recognizes that followed by virtually all originators of prime technology in certain ways to help expand some segments of the market may be more mortgages, including lenders who do not sell homeownership opportunities. For example, challenging for the GSEs than others, the data many of their mortgages to Fannie Mae or Fannie Mae has developed Fannie Mae reported in Table A.30 and Tables A.31a–c Freddie Mac. The guidelines are also Property GeoCoder a computerized mapping show that the GSEs have ample opportunities commonly followed in underwriting service offered to lenders, nonprofit to purchase goals-qualifying mortgages. ‘‘jumbo’’ mortgages, which exceed the organizations, and state and local Furthermore, if a GSE makes a business governments to help them determine whether decision to not pursue certain types of goals- maximum principal amount which can be purchased by the GSEs (the conforming loan a property is located in an area that qualifies qualifying loans in one segment of the for Fannie Mae’s community lending market, they are free to pursue goals- limit)—such mortgages eventually might be sold to the GSEs, as the principal balance is products designed to increase qualifying owner and rental property homeownership and revitalization in mortgages in other segments of the market. amortized or when the conforming loan limit traditionally underserved areas. In addition, As market leaders, the GSEs should be is otherwise increased. Changes that the eFannieMae.com is Fannie Mae’s business- looking for innovative ways to pursue this GSEs have made to their underwriting to-business Web site where lenders can business. Furthermore, there is evidence that standards in order to address the unique the GSEs can earn reasonable returns on their needs of low-income families were discussed access product information and important goals business. The Regulatory Analysis that in Section C.4 of this Appendix. The GSEs’ technology tools, view upcoming events, and accompanies this final rule provides market influence is one reason these new, receive news about training opportunities. evidence that the GSEs can earn financial more flexible underwriting standards have This site receives on average 80,000 visitors returns on their purchases of goals-qualifying spread throughout the market. Because the per week.323 Freddie Mac has introduced in loans that are only slightly below their return GSEs’ guidelines set the credit standards recent years Internet-based debt auctions, on equity from their normal business. against which the mortgage applications of debt repurchase operations, and debt exchanges. These mechanisms benefit 2. Qualitative Dimensions of the GSEs’ lower-income families are judged, the investors by providing more uniform pricing, Ability to Lead the Industry enterprises have a profound influence on the rate at which mortgage funds flow to low- greater transparency and faster price This section discusses several qualitative and moderate-income borrowers and discovery—all of which makes Freddie Mac factors that are indicators of the GSEs’ ability underserved neighborhoods. debt more attractive to investors and reduces to lead the industry in affordable lending. It As discussed below, the GSEs’ new the cost of funding mortgages.324 In addition, discusses the GSEs’ role in the mortgage automated underwriting systems are widely Freddie Mac has provided automated tools market; their ability, through their used to originate mortgages in today’s for lenders to identify and work with underwriting standards, new programs, and market. As discussed in Sections C.7 and C.8, borrowers most likely to encounter problems innovative products, to influence the types of the GSEs have started adapting their making their mortgage payments. loans made by private lenders; their underwriting systems for subprime loans and EarlyIndicator has become the industry development and utilization of state-of-the- other loans that have not met their traditional standard for default management technology. art technology; the competence, expertise underwriting standards. It can reduce the consequences of mortgage and training of their staffs; and their financial resources. c. State-of-the-Art Technology delinquency for borrowers, servicers and investors.325 a. Role in the Mortgage Market Both GSEs are in the forefront of new The GSEs are also expanding The GSEs have played a dominant role in developments in mortgage industry homeownership opportunities through the the single-family mortgage market. As technology. Automated underwriting and use of the Internet in processing mortgage reported in Section C.3, mortgage purchases online mortgage processing are a couple of originations. New online mortgage by the GSEs reached extraordinary levels in the new technologies that have impacted the originations reached $267.6 billion in the 2001–2003. Purchases by Fannie Mae stood mortgage market, expanding homeownership first half of 2002, compared with $97 billion at $568 billion in 2001 and $848 billion in opportunities. This section provides an for the first six months of 2001. The 2002 six- 2002. Freddie Mac’s single-family mortgage overview of these new technologies and the month volume comprised 26.5 percent of the purchases were $393 billion in 2001 and extent of their use. estimated $1.01 trillion in total mortgage $475 billion in 2002. The Office of Federal Each enterprise released an automated originations for the same time period.326 Housing Enterprise Oversight (OFHEO) underwriting system in 1995—Freddie Mac’s Freddie Mac made Loan Prospector on the estimates that the GSEs purchased 40 percent ‘‘Loan Prospector’’ (LP) and Fannie Mae’s Internet service available to lenders for their of newly-originated conventional mortgages ‘‘Desktop Underwriter’’ (DU). During 2001 retail operations. Freddie Mac also adopted in 2001. Total GSE purchases, including and 2002, roughly 60 percent of all newly- the mortgage industry’s XML (extensible loans originated in prior years, amounted to originated mortgages the GSEs purchased markup language) data standard, which is 46 percent of conventional originations in were processed through these systems. integral to streamlining and simplifying 2001. Lenders and brokers used LP to evaluate 7.3 Internet-based transactions. In addition, The dominant position of the GSEs in the million loan applications in 2001, 8.2 million Congress enacted legislation that allows the 322 mortgage market is reinforced by their in 2002, and 9.5 million in 2003. use of electronic signature in contracts in relationships with other market institutions. Similarly, DU was used to evaluate 8 million 2001, making a completely electronic Commercial banks, mutual savings banks, loans in 2001, over 10 million in 2002, and mortgage transaction possible. With the use and savings and loans are their competitors 14.8 million loans in 2003. The GSEs’ of electronic signatures, electronic mortgages as well as their customers—they compete to systems have also been adapted for FHA and are expected to improve the mortgage the extent they hold mortgages in portfolio, jumbo loans. Automated underwriting process, further reducing origination and but at the same time they sell mortgages to systems are being further adapted to facilitate servicing costs. In October 2000, Freddie Mac the GSEs. They also buy mortgage-backed risk-based pricing, which enables mortgage securities, as well as the debt securities used lenders to offer each borrower an individual 323 to finance the GSEs’ portfolios. Mortgage rate based on his or her risk. As discussed Fannie Mae, 2002 Annual Housing Activities earlier, concerns about the use of automated Report, 2003, pp. 10–11. bankers sell virtually all of their prime 324 conventional conforming loans to the GSEs. underwriting and risk-based pricing include Freddie Mac, Opening Doors for America’s Families: Freddie Mac’s Annual Housing Activities Private mortgage insurers are closely linked the disparate impact on minorities and low- Report for 2003, March 15, 2004, p. 14. to the GSEs, because mortgages purchased by 325 Freddie Mac, 2002 Annual Housing Activities the enterprises that have loan-to-value ratios 322 This section is based heavily on ‘‘DU and LP Report, 2003, p. 52. in excess of 80 percent are normally required Usage Continues to Rise,’’ in Inside Mortgage 326 Inside Mortgage Finance, ‘‘Online Volume to be covered by private mortgage insurance, Technology published by Inside Mortgage Finance, Comprises One-Fourth of Total Originations in First in accordance with the GSEs’ charter acts. January 27, 2003, page 1–2. Half ’02,’’ September 20, 2002, p. 8.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00163 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63742 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
purchased its first electronic mortgage under e. Financial Strength Other Indicators. Additional indicators of the new law. Fannie Mae. The benefits that accrue to the the strength of the GSEs are provided by The GSEs also offer a variety of other GSEs because of their GSE status, as well as various rankings of American corporations. online tools and applications that have the their solid management, have made them two Business Week has reported that among potential to make the mortgage loan process of the nation’s most profitable businesses. Standard & Poor’s performance ranking of more cost effective and efficient for lenders. Fannie Mae’s net income was $3.9 billion in 500 companies in 2004, Fannie Mae was Freddie Mac, for example, has launched 1999, $4.4 billion in 2000, $5.9 billion in ranked 117, down from 91 in 2003 and dontborrowtrouble.com, which contains 2001, $4.6 billion in 2002,332 and $7.9 billion Freddie Mac was listed as ‘‘INC’’ for 2004 information on local anti-predatory lending in 2003.333 Fannie Mae’s return on equity and 16th for 2003. Additionally, Fannie Mae campaigns, consumer tips on avoiding averaged 24.0 percent over the 1995–99 was ranked as 29th in overall market value, predatory lending, and information on how period—far above the rates achieved by most 17th in sales and 9th in profits, and Freddie to start a local campaign and obtain financial corporations. Fannie Mae’s return Mac was ranked 59th in market value and additional resources.327 Fannie Mae offers on equity was 26.0 percent in 2003, while ‘‘NR’’ in sales and profits.338 According to ‘‘HomeBuyer Funds Finder,’’ a one-stop this represented no change from 2002, it was Fortune’s annual listing of the 500 largest online resource designed for lenders and an increase of 3 percent over 2001.334 In U.S. Corporations, Fannie Mae was ranked other housing professionals, enables users to 2003, Fannie Mae’s total stockholders’ equity 20th in 2003, down from 16th in 2002, and access a database of local housing subsidy increased by 37% to $22.373 million, core Freddie Mac was ‘‘displaced’’ from the programs available for low- and moderate- business earnings grew by 14 percent ($7.3 ranking in 2003, but ranked 32nd in 2002. income borrowers. In 2002, the HomeBuyer billion), credit losses increased by $42 Additionally, Fannie Mae ranked 11th for Funds Finder Web site received over 24,500 million to $111 million with the resulting most profitable companies, 3rd for revenues per employee, and in the ‘‘Diversified hits.328 ‘‘Home Counselor Online’’ provides credit loss ratio at .006% (represents credit Financials’’ category, they ranked 2nd out of homeownership counselors with the losses divided by average single family 12 companies.339 And, according to Fortune’s necessary tools to help consumers financially mortgage credit book of business) and taxable 335 Global 500 listing of the world’s largest prepare to purchase a home. In 2003, 641 equivalent revenues grew by 24 percent. Fannie Mae’s basic net earnings per corporations, Fannie Mae ranked 56th in counselors representing over 2,000 common share increased from $3.75 in 1999 2003, (ranking 17th in highest profits) down organizations used Home Counselor to $7.93 in 2003, dividends per common from 45th in 2002, and Freddie Mac ranked 329 ‘‘True Cost Calculator 2.0’’ is Online. share have increased from $.96 in 1998 to 104th in 2003, down from 90th in 2002.340 designed to help homebuyers make informed $1.68 in 2003, a 27% increase over 2002, and f. Conclusion About Leading the Industry home purchase decisions by helping them operating earnings per diluted common share compare loan products and prices. Over 60 increased from 2002 to 2003 by 71% to In light of these considerations, the Fannie Mae partners officer the True Cost $7.72.336 Secretary has determined that the GSEs have Calculator through their Web sites and a Freddie Mac. Freddie Mac has shown the ability to lead the industry in making Spanish version is also available on similar trends. Freddie Mac’s net income was mortgage credit available for low- and Univision.com.330 A more complete list of $3.158 billion in 2001, $10.090 billion in moderate-income families. Fannie Mae’s online tool and applications 2002, and $4.891 billion in 2003, and total H. Factor 6: The Need to Maintain the Sound can be found in its Annual Housing stockholder’s equity increased by 10% over Financial Condition of the GSEs Activities Report. In 2002, Fannie Mae’s total 2002 to $31.562 billion. Freddie Mac’s return eBusiness volume was $1.1 trillion, up from on equity averaged 23.4 percent over the HUD has undertaken a separate, detailed $800 billion in 2000.331 1995–1999 period, also well above the rates economic analysis of this final rule, which includes consideration of (a) the financial d. Staff Resources achieved by most financial corporations. Credit losses increased by $8 million to $82 returns that the GSEs earn on low- and Both Fannie Mae and Freddie Mac are million with the resulting credit loss ratio at moderate-income loans and (b) the financial well-known throughout the mortgage 0.7 (represents annualized credit losses safety and soundness implications of the industry for the expertise of their staffs in divided by average total mortgage portfolio). housing goals. Based on this economic carrying out their current programs, Basic earnings per common share (after analysis and the Office of Federal Housing conducting basic and applied research cumulative effect of change in accounting Enterprise Oversight review, HUD concludes regarding mortgage markets, developing principles, net of taxes) was $4.25 in 2001, that the goals raise minimal, if any, safety innovative new programs, and undertaking $14.23 in 2002 and $6.80 in 2003. Dividends and soundness concerns. sophisticated analyses that may lead to new per common share have increased from 0.80 programs in the future. The leaders of these I. Determination of the Low- and Moderate- in 2001 to $1.04 in 2003, an 18% increase Income Housing Goals corporations frequently testify before over 2002, and operating earnings per diluted Congressional committees on a wide range of common share (after cumulative effect of The annual goal for each GSE’s purchases housing issues, and both GSEs have change in accounting principles, net of taxes) of mortgages financing housing for low- and developed extensive working relationships decreased from 2002 to 2003 by $7.39 to moderate-income families is being with a broad spectrum of mortgage market $6.79.337 established at 52 percent of eligible units participants, including various nonprofit financed in each of calendar years 2005, 53 groups, academics, and government housing 332 The 22% decrease in Fannie Mae’s 2002 net percent in 2006, 55 percent in 2007, and 56 authorities. Federal agencies and foreign income resulted primarily from a $4.508 billion percent in 2008. This goal will remain in governments and businesses seek them out increase in purchased options expense, which effect thereafter, unless changed by the for advice and consultation because of their occurred due to an increase in the notional amount Secretary prior to that time. In addition, a expertise. The role that the GSEs have played of purchased options outstanding and the declining low- and moderate-income subgoal of 45 in spreading the use of technology interest rate environment. Recorded purchased percent in 2005, 46 percent in 2006, and 47 throughout the mortgage market reflects the options expense for 2001 was only $37 million by percent in both 2007 and 2008 is being enormous expertise of their staff. comparison. Fannie Mae 2002 Annual Report, 2003, established for the GSEs’ acquisitions of p. 23. single-family-owner home purchase loans in 333 Fannie Mae, 2003 Annual Report, ‘‘Financial 327 Freddie Mac, Opening Doors for America’s Highlights.’’ metropolitan areas. This subgoal is designed Families: Freddie Mac’s Annual Housing Activities 334 Fannie Mae, 2003 Annual Report, ‘‘Financial to encourage the GSEs to lead the primary Report for 2003, March 15, 2004, p. 38. Highlights.’’ market in offering homeownership 328 Fannie Mae, 2002 Annual Housing Activities 335 Fannie Mae, 2003 Annual Report, ‘‘Financial Report, 2003, p. 12. Highlights’’ and United States Securities and 338 ‘‘The Standard and Poor’s Five Hundred: 329 Fannie Mae, 2003 Annual Housing Activities Exchange Commission form 10–K, p. 108. Performance Ranking S&P 500’’, Business Week, Report, 2004, p. 13. 336 Fannie Mae, 2003 Annual Report to April 5, 2004, p. 127. 330 Fannie Mae, 2003 Annual Housing Activities Shareholders, Financial Highlights and Financial 339 ‘‘Fortune 500 Largest U.S. Corporations,’’ Report, 2004, p. 13. Information. Fortune, April 5, 2004, p. F–1. 331 Fannie Mae, 2002 Annual Housing Activities 337 Freddie Mac, Consolidated Statements of 340 ‘‘Fortune 500 Largest U.S. Corporations,’’ Report, 2003, p. 10. Income 2003 and Freddie Mac Core Tables 2003. Fortune, July 26, 2004, p. 159.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00164 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63743
opportunities to low- and moderate-income demand will average from 1.84 to 2.19 purchase at record levels as well. As noted families. The Secretary’s consideration of the million units over the next decade.341 above, the potential homeowner population six statutory factors that led to the choice of Growth in Single-Family Affordable over the next decade will be highly diverse, these goals is summarized in this section. Lending. Many younger, minority and lower- as growing demand from immigrants and income families did not become homeowners minorities are expected to sustain the home 1. Housing Needs and Demographic during the 1980s due to the slow growth of purchase market, as our population ages. Conditions earnings, high real interest rates, and Single-family housing starts are expected to Affordability Problems. Data from the 2000 continued house price increases. Over the continue in the 1.65–1.70 million range over Census and the American Housing Surveys past ten years, economic expansion, the next few years. Refinancing of existing demonstrate that there are substantial accompanied by low interest rates and mortgages, which accounted for about 60 housing needs among low- and moderate- increased outreach on the part of the percent of originations during 2001–2003 is income families. Many of these households mortgage industry, has improved expected to return to more normal levels. As are burdened by high homeownership costs affordability conditions for these families. As this Appendix has explained, the GSEs will or rent payments and will likely continue to this appendix has explained, there has been continue to play a dominant role in the face serious housing problems. There is a ‘‘revolution in affordable lending’’ that has single-family market and will both impact evidence of persistent housing problems for extended homeownership opportunities to and be affected by major market Americans with the lowest incomes. Since historically underserved households. The developments such as the growth in 1977, the percentage of U.S. households with mortgage industry has offered more subprime lending and the increasing use worst case needs has hovered around five customized mortgage products, more flexible automated underwriting. percent, with the worst year being 1983 (6.03 underwriting, and expanded outreach to low- Multifamily Mortgage Market. The market percent) and the best year being 1999 (4.72 income and minority borrowers. Fannie Mae for financing of multifamily apartments has percent). The proportion in 2001 was 4.77 and Freddie Mac have been a big part of this grown to record volumes. The favorable long- term prospects for apartments, combined percent, which is not significantly different ‘‘revolution in affordable lending’’. HMDA with record low interest rates, have kept from the 1999 figure. HUD’s analysis of data suggest that the industry and GSE investor demand for apartments strong and American Housing Survey data reveals that, initiatives are increasing the flow of credit to underserved borrowers. Between 1993 and supported property prices. As explained in 2001, 5.1 million unassisted very-low above, Fannie Mae and Freddie Mac have income renter households had ‘‘worst-case’’ 2003, conventional loans to low-income and minority families increased at much faster been among those boosting volumes and housing needs, defined as housing costs introducing new programs to serve the rates than loans to upper-income and non- greater than 50 percent of household income multifamily market. The long run outlook for minority families. Thus, the 1990s and the or severely inadequate housing. Among these the multifamily rental market is sustained, early part of the current decade have seen the households, 90 percent had a severe rent moderate growth, based on favorable development of a strong affordable lending burden, 6 percent lived in severely demographics. The minority population, inadequate housing, and 4 percent suffered market. especially Hispanics, provides a growing from both problems. Among the 34 million Disparities in Housing and Mortgage source of demand for affordable rental renters in all income categories, 6.3 million Markets. Despite this strong growth in housing. ‘‘Lifestyle renters’’ (older, middle- (19 percent) had a severe rent burden and affordable lending, serious disparities in the income households) are also a fast growing over one million renters (3 percent) lived in nation’s housing and mortgage markets segment of the rental population. However, housing that was severely inadequate. remain. The homeownership rate for African- provision of affordable housing will continue Demographic Trends. Changing population American and Hispanic households is about to challenge suppliers of multifamily rental demographics will result in a need for the 25 percentage points below that of white housing and policy makers at all levels of primary and secondary mortgage markets to households. In addition to low income, government. Low incomes combined with meet nontraditional credit needs, respond to barriers to homeownership that high housing costs define a difficult situation diverse housing preferences and overcome disproportionately affect minorities and for millions of renter households. Housing information and other barriers that many immigrants include: lack of capital for down cost reductions are constrained by high land immigrants and minorities face. It is payment and closing costs; poor credit prices and construction costs in many projected that there will be 1.2 million new history; lack of access to mainstream lenders; markets. Government action—through land households each year over the next decade. little understanding of the homebuying use regulation, building codes, and The aging of the baby-boom generation and process; and continued discrimination in occupancy standards—are major contributors the entry of the baby-bust generation into housing markets and mortgage lending. With to those high costs. In addition to fewer respect to the latter, a recent HUD-sponsored prime home buying age will have a regulatory barriers and costs, multifamily study of discrimination in the rental and dampening effect on housing demand. housing would benefit from more favorable owner markets found that while differential public attitudes. Higher density housing is a However, the continued influx of immigrants treatment between minority and white home potentially powerful tool for preserving open will increase the demand for rental housing, seekers had declined over the past ten years, space, reducing sprawl, and promoting while those who immigrated during the it continued at an unacceptable level in the transportation alternatives to the automobile. 1980s and 1990s will be in the market for year 2000. In addition, disparities in The recently heightened attention to these owner-occupied housing. Immigrants and mortgage lending continued across the nation issues may increase the acceptance of other minorities—who accounted for nearly in 2003, when the loan denial rate for multifamily rental construction to both 40 percent of the growth in the nation’s African-American applicants was almost potential customers and their prospective homeownership rate over the past five three times that for white applicants, even neighbors. years—will be responsible for almost two- after controlling for income of the applicant. 2. Past Performance of the GSEs thirds of the growth in the number of new HUD studies also show that African- households over the next ten years. Non- This section reviews the low- and Americans and Hispanics are subject to moderate-income performance of Fannie Mae traditional households have become more discriminatory treatment during the pre- important, as overall household formation and Freddie Mac. It first reviews the GSEs’ qualification process of applying for a performance on the Low- and Moderate- rates have slowed. With later marriages, mortgage. divorce, and non-traditional living Income Goal, then reviews findings from Single-Family Mortgage Market. Heavy Section E.2 regarding the GSEs’ purchases of arrangements, the fastest growing household refinancing due to low interest rates groups have been single-parent and single- home loans for historically underserved increased single-family mortgage originations families and their communities. Finally, it person households. As these demographic to record levels during 2001–2003. factors play out, the overall effect on housing reviews findings from Section G concerning Demographic forces, industry outreach, and the GSEs’ presence in owner and rental demand will likely be sustained growth and low interest rates also kept lending for home an increasingly diverse household markets. population from which to draw new renters a. Housing Goals Performance 341 National Association of Home Builder, 2004 and homeowners. According to the National Spring Construction Forecast Conference, April 21, In the October 2000 rule, the low- and Association of Homebuilders, annual housing 2004. moderate-income goal was set at 50 percent
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00165 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63744 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
for 2001–03. Effective on January 1, 2001, rents for determining goal credit for purchased the same mortgages that they several changes in counting requirements multifamily mortgages. Fannie Mae’s actually did purchase in both years, then came into effect for the low- and moderate- performance was 51.5 percent in 2001, 51.8 Fannie Mae’s performance would have been income goal, as follows: (a) ‘‘Bonus points’’ percent in 2002, and 52.3 percent in 2003; 51.3 percent in 2000, 49.2 percent in 2001, (double credit) for purchases of mortgages on Freddie Mac’s performance was 53.2 percent 49.0 percent in 2002, and 48.7 percent in small (5–50 unit) multifamily properties and, in 2001, 50.5 percent in 2002, and 51.2 2003. Freddie Mac’s performance would have above a threshold level, mortgages on 2–4 percent in 2003—thus both GSEs surpassed been 50.6 percent in 2000, 47.7 percent in unit owner-occupied properties; (b) a this higher goal in all three years. 2001, 46.1 percent in 2002, and 44.6 percent ‘‘temporary adjustment factor’’ (1.35 units Counting requirements (a) and (b) expired in 2003. Thus, both Fannie Mae and Freddie credit) for Freddie Mac’s purchases of at the end of 2003, while (c) and (d) will Mac would have surpassed the low- and mortgages on large (more than 50 units) remain in effect after that. If this counting moderate-income goal of 50 percent in 2000 multifamily properties; (c) changes in the approach—without the bonus points and the and fallen short in 2001 through 2003. (See treatment of missing data; and (d) a ‘‘temporary adjustment factor’’—had been in Figure A.1.) procedure for the use of imputed or proxy effect in 2000 and 2001, and the GSEs had BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00166 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63745
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00167 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.044
b. Single-Family Affordable Lending Market entering new and challenging fields of periods, such as 1993–2003, 1996–2003, and The GSEs have played a major role in the mortgage finance, such as purchasing 1999–2003, they have lagged the overall single-family mortgage market over the past subprime mortgages. conventional conforming market in providing ten years. Their purchases of single-family- Despite these efforts and the overall gains affordable loans to lower-income borrowers owner mortgages accounted for 61 percent of in goal performance, the Department remains and underserved areas. This finding is based all mortgages originated in the single-family concerned about the GSEs’ support of home on HUD’s analysis of GSE and HMDA data conventional conforming market between lending for the lower-income end of the and on numerous studies by academics and 1999 and 2002. Their underwriting and market and for first-time homebuyers. The research organizations. purchase guidelines are market standards, shares of the GSEs’ purchases are too low, • The GSEs have shown different patterns used in all segments of the mortgage market. particularly for underserved areas and groups of mortgage purchases. Except for two years The GSEs have worked to improve their such as minority first-time homebuyers. (1999 and 2000), Fannie Mae has performed affordable lending record—they have This appendix included a comprehensive better than Freddie Mac since 1993 on all introduced new low-downpayment products analysis of the GSEs’ performance in funding three goals-qualifying categories—low-mod, targeted at lower-income families; they have home purchase mortgages for families and special affordable, and underserved areas. As customized their underwriting standards to communities that historically have not been a result, the percentage of Freddie Mac’s recognize the unique needs of immigrant and well served by the mortgage market. The purchases benefiting historically underserved minority families; and, they have entered following findings are offered with respect to families and their neighborhoods has been into numerous partnerships with lenders and the GSEs’ acquisitions of home purchase less than the corresponding shares of total non-profit groups to reach out to underserved loans that qualify for the three housing goals market originations, while Fannie Mae’s populations. The enterprises’ role in the (special affordable and underserved areas as purchases have been somewhat closer to the mortgage market is also reflected in their use well as low- and moderate-income) and their patterns of originations in the primary of cutting edge technology, such as the acquisitions of first-time homebuyer loans: market. development of Loan Prospector and Desktop • Fannie Mae and Freddie Mac have both • The above patterns can be seen by the Underwriter, the automated underwriting improved their support for the single-family following percentage shares of home systems developed by Freddie Mac and affordable lending market over the past purchase loans that qualified for the three Fannie Mae, respectively. Both GSEs are also eleven years, but historically over past housing goals between 1996 and 2003:
Under- Special Low-Mod served affordable (percent) areas (percent) (percent)
Freddie Mac ...... 13.2 40.3 22.0 Fannie Mae ...... 14.1 42.2 24.0 Market (w/o B&C) ...... 15.9 43.6 25.7
• During 2001–2003, Fannie Mae underserved areas market. During 2001– low- and moderate-income shares for Fannie improved its performance enough to lead the 2003, Freddie Mac lagged the conventional Mae, Freddie Mac and the market. special affordable and low-moderate income conforming market on all three goals- markets, although it continued to lag the qualifying categories; see Figure A.2 for the
VerDate jul<14>2003 19:29 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00168 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63747
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00169 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.045
• Both Fannie Mae and Freddie lag the performance on the special affordable and income subgoal is discussed in Section I.3 conventional conforming market in funding low-moderate categories, the GSEs have not below. first-time homebuyers, and by a rather wide been leading the single-family-owner market margin. Between 1999 and 2001, first-time in purchasing goals-qualifying and first-time c. Overall Market Shares homebuyers accounted for 27 percent of each homebuyer loans. Freddie Mac, in This appendix also included an analysis of GSE’s purchases of home loans, compared participation, continues to lag the market on the GSEs’ role in the overall (owner and with 38 percent for home loans originated in all categories considered. There is room for rental) conventional conforming mortgage the conventional conforming market. Freddie Mac, as well as Fannie Mae, to market. While GSE mortgage purchases • The GSEs also account for a very small further improve their performance in represented 55 percent of total dwelling units share of the market for important groups such purchasing affordable loans in the as minority first-time homebuyers. underserved portion of the market, financed between 1999 and 2002, they Considering the total mortgage market (both particularly in the minority first-time represented smaller shares of the three goals- government and conventional loans), it is homebuyer market. Evidence suggests that qualifying markets: 48 percent of housing estimated that the GSEs purchased only 14 there is a significant population of potential units financed for both low- and moderate- percent of loans originated between 1999 and homebuyers who might respond well to income families and properties located in 2001 for African-American and Hispanic aggressive outreach by the GSEs—immigrants underserved areas; and 41 percent of units first-time homebuyers, or one-third of their and minorities, in particular, are expected to financed for the very-low-income and other share (42 percent) of all home purchase loans be a major source of future homebuyers. families that qualify as special affordable. originated during that period. Considering Furthermore, studies indicate the existence (See Figure A.3.) In other words, the GSEs the conventional conforming market and the of a large untapped pool of potential same time period, it is estimated that the homeowners among the rental population. accounted for approximately 50 percent or GSEs purchased only 31 percent of loans Indeed, the GSEs’ recent experience with less of the single-family and multifamily originated for African-American and new outreach and affordable housing units financed in the goals-qualifying Hispanic first-time homebuyers, or initiatives is important confirmation of this markets. This market share analysis suggests approximately one-half of their share (57 potential. To move the GSEs into a that there is room for the GSEs to increase percent) of all home purchase loans in that leadership position, the Department is their purchases in these goals-qualifying market. establishing three subgoals for home markets. To summarize, the Department’s analysis purchase loans that qualify for the three BILLING CODE 4210–27–C suggests that, except for Fannie Mae’s recent housing goals. The low- and moderate-
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00170 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63749
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00171 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.046 63750 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
The market analysis also examined the mortgage industry has not been fully data for non-metropolitan areas are not GSEs’ presence in the owner-occupied home developed. As reported earlier in Tables A.30 available from HMDA. The Department is purchase mortgage and rental property and A.31b, the GSEs’ purchases between also setting home purchase subgoals for the sectors of the mortgage market: single-family 1999 and 2002 accounted for 35–41 percent other two goals-qualifying categories, as owner (a 61 percent share for the GSEs of the multifamily units that received follows: 17–18 percent for special affordable between 1999 and 2002) and single-family financing during this period. Certainly there loans and 32–34 percent for underserved area rental and multifamily rental (a combined are ample opportunities and room for loans (also called Geographically Targeted rental share of 37 percent). The GSEs, and expansion of the GSEs’ share of the loans). particularly Freddie Mac, have historically multifamily mortgage market. The GSEs’ size The Department considered the following played a smaller role in the market financing and market position between loan originators factors when setting the subgoal for low- and rental properties, as compared with their role and mortgage investors makes them the moderate-income loans. in the owner market. Fannie Mae and logical institutions to identify and promote (a) The GSEs have the ability to lead the Freddie Mac have recently increased their needed innovations and to establish market. The GSEs have the ability to lead the purchases of these mortgages, but their standards that will improve market primary market for single-family-owner purchases totaled only 37 percent of the efficiency. As their role in the multifamily loans, which is the ‘‘bread-and-butter’’ of rental units that received financing between market continues to grow, the GSEs will have their business. They both have substantial 1999 and 2002.342 A further increased the knowledge and market presence to push experience in this market, which means there presence by Fannie Mae and Freddie Mac simultaneously for standardization and for are no issues as whether or not the GSEs have would bring lower interest rates and liquidity programmatic flexibility to meet special yet penetrated the market, as there are with to this market, as well as improve their needs and circumstances, with the ultimate the single-family rental and multifamily housing goals performance. goal of increasing the availability and markets. Both GSEs have not only been d. The GSEs’ Purchases of Multifamily reducing the cost of financing for affordable operating in the owner market for years, they Mortgages and other multifamily rental properties. have been the dominant players in that 3. Ability to Lead the Single-Family-Owner market, funding 61 percent of the single- Fannie Mae and, especially, Freddie Mac family-owner mortgages financed between have rapidly expanded their presence in the Market: A Low- and Moderate-Income Subgoal 1999 and 2002. As discussed in Section G, multifamily mortgage market in the period their underwriting guidelines are industry As discussed in Section E, the Department since the passage of FHEFSSA. The Senate standards and their automated mortgage is proposing to establish a subgoal of 45 report on this legislation in 1992 referred to systems are widely used throughout the percent for each GSE’s purchases of home the GSEs’ activities in the multifamily arena mortgage industry. Through their new purchase loans for low- and moderate- as ‘‘troubling,’’ citing Freddie Mac’s downpayment and subprime products, and income families in the single-family-owner September 1990 suspension of its purchases their various partnership initiatives, the GSEs of new multifamily mortgages and criticism market of metropolitan areas for 2005, with have shown that they have the capacity to of Fannie Mae for ‘‘creaming’’ the market.343 the subgoal rising to 46 percent in 2006 and reach out to lower-income families seeking to Freddie Mac has successfully rebuilt its 47 percent in 2007 and 2008. The purpose of buy a home. Both Fannie Mae and Freddie multifamily acquisition program, as shown this subgoal is to encourage the GSEs to Mac have the staff expertise and financial by the increase in its purchases of improve their acquisitions of home purchase resources to make the extra effort to lead the multifamily mortgages: from $27 million in loans for lower-income families and first- primary market in funding single-family- 1992 to $3 billion in 1997 and then to time homebuyers who are expected to enter owner mortgages for low- and moderate- approximately $7 billion during the next the homeownership market over the next few income mortgages, as well for special three years (1998 to 2000), before rising years. If the GSEs meet this goal, they will affordable and undeserved area mortgages. further to $11.9 billion in 2001, $13.3 billion be leading the primary market by in 2002, and $21.6 billion in 2003. approximately one percentage point in 2005 (b) GSEs’ Performance Relative to the Multifamily properties accounted for 10.3 and by three percentage points in 2007 and Market. Even though the GSEs have had the percent of all dwelling units (both owner and 2008, based on the income characteristics of ability to lead the home purchase market, rental) financed by Freddie Mac during 2003. home purchase loans reported in HMDA. their past average performance (1993–2003, Concerns regarding Freddie Mac’s Between 2002 and 2003, HMDA data show 1996–2003, and 1999–2003) has been below multifamily capabilities no longer constrain that low- and moderate-income families market levels. During 2002 and 2003, Fannie their performance with regard to low- and accounted for an (unweighted) average of Mae improved its performance enough to moderate-income families. 44.1 percent of single-family-owner loans lead the low-mod market for home purchase Fannie Mae never withdrew from the originated in the conventional conforming loans, but Freddie Mac, although it also multifamily market, but it has also stepped market of metropolitan areas. (The market improved its performance during this recent up its activities in this area substantially, and GSE data reported in this paragraph are period, continues to lag behind the primary with multifamily purchases rising from $3.0 based on ‘‘projected’’ data that account for market. The subgoals will ensure that Fannie billion in 1992 to $9.4 billion in 1999, $18.7 new Census geography and the new OMB Mae maintains and further improves its billion in 2001, $18.3 billion in 2002, and metropolitan area definitions; see Table above-market performance and that Freddie $33.3 billion in 2003. Multifamily units as a A.17b.) Loans in the B&C portion of the Mac not only erases its current gap with the share of all dwelling units (both owner and subprime market are not included in these market but also takes a leadership position as rental) financed by Fannie Mae varied in the averages. To reach the 45-percent (47- well. With respect to the GSEs’ historical 10–13 percent range between 1999 and 2001, percent) subgoal, Freddie Mac would have to performance, low- and moderate-income before falling to 7.3 percent during heavy improve its performance by 0.8 (2.8) mortgages accounted for 40.3 (42.6) percent refinancing year of 2002 and 8 percent in percentage points over its 2003 performance. of Freddie Mac’s purchases during 1996– 2003. Fannie Mae would have to keep up its high 2003 (1999–2003), for 42.2 (43.6) percent of The increased purchases of multifamily level (47.5 percent) of performance during Fannie Mae’s purchases, and for 43.6 (44.1) mortgages by Fannie Mae and Freddie Mac 2003. The approach taken is for the GSEs to percent of primary market originations have major implications for the Low- and obtain their leadership position by staged (excluding B&C loans). The type of Moderate-Income Housing Goal, since a very increases in the low-mod subgoal; this will improvement needed for Freddie Mac to high percentage of multifamily units have enable the GSEs to take new initiatives in a meet this new low-mod subgoal was rents which are affordable to low- and correspondingly staged manner to achieve demonstrated by Fannie Mae during 2001– moderate-income families. However, the the new subgoal each year. Thus, the 2003, as Fannie Mae increased its low-mod potential of the GSEs to lead the multifamily increases in the low-mod subgoal are purchases from 40.8 percent of its single- sequenced so that the GSEs can gain family-owner business in 2000 to 45.3 342 As shown in Table A.31b, the GSEs’ share of experience as they improve and move toward percent in 2002 and47.0 percent in 2003. (As the rental market increases to 41 percent when a the new higher subgoal targets. noted above, Fannie Mae’s 2003 performance lower multifamily share is assumed in the market As explained in Section E.9, the subgoal was slightly higher at 47.5 percent when analyses. applies only to the GSEs’ purchases in measured based on the new 2000 Census 343 Senate Report 102–282, May 15, 1992, p. 36. metropolitan areas because reliable market geography and new OMB definitions.)
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00172 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63751
(c) Disparities in Homeownership and The GSEs can purchase loans from the percentage points over its previous peak Credit Access Remain. There remain substantial, existing stock of affordable loans performance in 2000. Freddie Mac’s troublesome disparities in our housing and held in lenders’ portfolios, after these loans performance is estimated to have been 46.0 mortgage markets, even after the ‘‘revolution have seasoned and the GSEs have had the percent in 1999, 50.2 percent in 2000, 47.0 in affordable lending’’ and the growth in opportunity to observe their payment percent in 2001, 44.6 percent in 2002, and homeownership that has taken place since performance. In fact, based on Fannie Mae’s 45.3 percent in 2003—for 2005, Freddie Mac the mid-1990s. The homeownership rate for recent experience, the purchase of seasoned would have to increase its performance by African-American and Hispanic households loans appears to be one useful strategy for 5.3 percentage points over its average remains 25 percentage points below that of purchasing goals-qualifying loans. (unweighted) performance of 46.7 percent white households. Minority families face For the reasons given above, the Secretary over these last five years, or by 1.8 percentage many barriers in the mortgage market, such believes that the GSEs can do more to raise points over its previous peak performance as lack of capital for down payment and lack the low- and moderate-income shares of their (50.2 percent in 2000). By 2008, Freddie of access to mainstream lenders (see above). mortgages on these properties. This can be Mac’s performance would have to increase by Immigrants and minorities are projected to accomplished by building on various 9.3 percentage points over average 1999– account for almost two-thirds of the growth programs that the enterprises have already 2003 performance, and by 5.8 percentage in the number of new households over the started, including (1) their partnership and points over its previous peak performance. next ten years. As emphasized throughout outreach efforts, (2) their incorporation of However, the low- and moderate-income this Appendix, changing population greater flexibility into their underwriting market is estimated to be 51–56 percent. demographics will result in a need for the guidelines, (3) their purchases of CRA loans, Thus, the GSEs should be able to improve primary and secondary mortgage markets to and (4) their targeting of important markets their performance enough to meet these goals meet nontraditional credit needs, respond to where they have had only a limited presence of 52–56 percent. diverse housing preferences and overcome in the past, such as the market for minority The objective of the Low- and Moderate- information and other barriers that many first-time homebuyers. A wide variety of Income Goal is to bring the GSEs’ immigrants and minorities face. The GSEs quantitative and qualitative indicators performance to the upper end of HUD’s have to increase their efforts in helping these indicate that the GSEs’ have the resources market range estimate for this goal (51–56 families because so far they have played a and financial strength to improve their percent), consistent with the statutory surprisingly small role in serving minority affordable lending performance enough to criterion that HUD should consider the GSEs’ first-time homebuyers. It is estimated that the lead the market for low- and moderate- ability to lead the market for each Goal. To GSEs accounted for 46.5 percent of all (both income families. The recent experience of enable the GSEs to achieve this leadership, government and conventional) home loans Fannie Mae indicates that the GSEs can lead the Department is proposing modest originated between 1999 and 2001; however, the low- and moderate-income market. increases in the Low- and Moderate-Income they accounted for only 14.3 percent of home 4. Size of the Mortgage Market for Low- and Goal for 2005 which will increase further, loans originated for African-American and Moderate-Income Families year-by-year through 2008, to achieve the Hispanic first-time homebuyers. Within the ultimate objective for the GSEs to lead the As detailed in Appendix D, the low- and conventional conforming market, it is market under a range of foreseeable economic moderate-income mortgage market accounts estimated that the GSEs purchased only 20 circumstances by 2008. Such a program of for 51 to 56 percent of dwelling units percent of loans originated for African- staged increases is consistent with the financed by conventional conforming American and Hispanic first-time statutory requirement that HUD consider the mortgages. In estimating the size of the homebuyers, even though they accounted for past performance of the GSEs in setting the market, HUD excluded the effects of the B&C 57 percent of all home purchase loans in that Goals. Staged annual increases in the Low- market. HUD also used alternative market. A subgoal for home purchase loans and Moderate-Income Goal will provide the assumptions about future economic and should increase the GSEs’ efforts in enterprises with opportunity to adjust their market affordability conditions that were less business models and prudently try out important sub-markets such as the one for favorable than those that existed over the last minority first-time homebuyers. business strategies, so as to meet the required five years. HUD is well aware of the volatility 2008 level without compromising other (d) There are ample opportunities for the of mortgage markets and the possible impacts GSEs to improve their performance. Low- business objectives and requirements. of changes in economic conditions on the Figure A.3 summarizes many of the points and moderate-income loans are available for GSEs’ ability to meet the housing goals. the GSEs to purchase, which means they can made in this section regarding opportunities Should conditions change such that the goals for Fannie Mae and Freddie Mac to improve improve their performance and lead the are no longer reasonable or feasible, the primary market in purchasing loans for their overall performance on the Low- and Department has the authority to revise the Moderate-Income Goal. The GSEs’ purchases borrowers with less-than-median income. goals. Three indicators of this have already been provided financing for 26,118,927 (or 55 discussed. First, Sections B and C of this 5. The Low- and Moderate-Income Housing percent) of the 47,551,039 single-family and appendix and Appendix D explain that the Goal for 2005–2008. multifamily units that were financed in the affordable lending market has shown an The Low- and Moderate-Income Housing conventional conforming market between 1999 and 2002. However, in the low- and underlying strength over the past few years Goal is 52 percent of eligible units for 2005, moderate-income part of the market, the that are unlikely to vanish (without a 53 percent for 2006, 55 percent for 2007, and 12,608,215 units that were financed by GSE significant increase in interest rates or a 56 percent for 2008. The market for the Low- purchases represented only 48 percent of the decline in the economy). The low-mod share and Moderate-Income Goal is estimated to be 26,051,771 dwelling units that were financed of the home purchase market has averaged 51–56 percent. Under the new counting rules in the market. Thus, there appears to be 43.6 percent since 1996 and annually has (i.e., 2000-Census income re-benchmarking ample room for the GSEs to increase their ranged from 42.1 percent to 44.8 percent. and the new OMB metropolitan area purchases of loans that qualify for the Low- Second, the market share data reported in definitions), Fannie Mae’s low- and and Moderate-Income Goal. Examples of Table A.30 of Section G demonstrate that moderate-income performance is estimated to specific market segments that would there are newly-originated low- and have been 46.3 percent in 1999, 51.2 percent particularly benefit from a more active moderate-income loans available each year in 2000, 48.7 percent in 2001, 47.9 percent secondary market have been provided for the GSEs to purchase. As noted above, the in 2002, and 49.5 percent in 2003—for 2005, throughout this appendix. GSEs have only a minimal presence in Fannie Mae would have to increase its special sub-markets such the minority first- performance by 3.3 percentage points over its 6. Conclusions time homebuyer market, which suggests average (unweighted) performance of 48.7 Having considered the projected mortgage there are ample opportunities available for percent over these last five years, or by 0.8 market serving low- and moderate-income the GSEs to increase their purchases of loans percentage point over its previous peak families, economic, housing and for low- and moderate-income families. performance (51.2 percent in 2000). By 2008, demographic conditions for 2005–08, and the Finally, the GSEs’ purchases under the Fannie Mae’s performance would have to GSEs’ recent performance in purchasing subgoal are not limited to new mortgages that increase by 6.3 percentage points over mortgages for low- and moderate-income are originated in the current calendar year. average 1999–2003 performance, and by 5.8 families, the Secretary has determined that
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00173 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63752 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
the goals of 52 percent of eligible units 6. The need to maintain the sound In this Rule, the underserved census tracts financed in 2005, 53 percent in 2006, 55 financial condition of the enterprises. are defined in terms of the 2000 Census percent in 2007, and 56 percent in 2008 are Organization of Appendix. The remainder rather than the 1990 Census. As shown in feasible. The Secretary is also establishing a of Section A first defines the Underserved Table B.1a, switching to 2000 Census data subgoal of 45 percent for the GSEs’ purchases Areas Housing Goal for both metropolitan and re-specified MSA boundaries as of June of single-family-owner home purchase areas and nonmetropolitan areas. Sections B 2003, increases the proportions of mortgages in metropolitan areas in 2005, and C address the first two factors listed underserved census tracts, population, increasing to 46 percent in 2006 and 47 above, focusing on findings from the owner-occupied housing units, and percent in 2007 and 2008. The Secretary has literature on access to mortgage credit in population below the poverty line in considered the GSEs’ ability to lead the metropolitan areas (Section B) and in metropolitan areas. The definition now industry as well as the GSEs’ financial nonmetropolitan areas (Section C). Separate covers 26,959 (51.3 percent) of the 52,585 condition. The Secretary has determined that discussions are provided for metropolitan census tracts in metropolitan areas, which the proposed goals and the proposed and nonmetropolitan (rural) areas because of subgoals are necessary and appropriate. include 48.7 percent of the population and differences in the underlying markets and the 38.0 percent of the owner-occupied housing Appendix B—Departmental data available to measure them. Section D units in metropolitan areas.1 The 1990-based Considerations To Establish the Central discusses the past performance of the GSEs definition covered 21,587 (47.5 percent) of Cities, Rural Areas, and Other on the Underserved Areas Housing Goal (the the 45,406 census tracts in metropolitan Underserved Areas Goal third factor) and Sections E-G report the areas, which included 44.3 percent of the Secretary’s findings for the remaining factors. population and 33.7 percent of the owner- A. Introduction Section H presents the Department’s rules occupied units in metropolitan areas. 1. Establishment of Goal relating to the definition of underserved areas The census tracts included in HUD’s in nonmetropolitan areas. Section I The Federal Housing Enterprises Financial definition of underserved areas exhibit low summarizes the Secretary’s rationale for Safety and Soundness Act of 1992 rates of mortgage access and distressed establishing a subgoal for single-family- (FHEFSSA) requires the Secretary to socioeconomic conditions. Between 1999 and establish an annual goal for the purchase of owner home purchase mortgages and for 2002, the unweighted average mortgage mortgages on housing located in central setting the level for the Underserved Areas denial rate in these tracts was 17.5 percent, cities, rural areas, and other underserved Housing Goal. almost double the average denial rate (9.3 areas (the ‘‘Underserved Areas Housing 2. HUD’s Underserved Areas Housing Goal percent) in excluded tracts. The underserved tracts include 75.3 percent of the number of Goal’’). HUD’s definition of the geographic areas persons below the poverty line in In establishing this annual housing goal, targeted by this goal is basically the same as metropolitan areas. Section 1334 of FHEFSSA requires the that used during 1996–2003. It is divided Secretary to consider: BILLING CODE 4210–27–P into a metropolitan component and a 1. Urban and rural housing needs and the nonmetropolitan component. However, as housing needs of underserved areas; explained below, switching to 2000 Census 1 This analysis excludes Puerto Rico. In addition, 2. Economic, housing, and demographic geography increases the number of census tracts are excluded if median income is suppressed conditions; in the underlying census data. There are 379 such 3. The performance and effort of the tracts defined as underserved, and this necessitates an adjustment of the goal level. tracts. When reporting analysis of mortgage loan enterprises toward achieving the denial, origination, and application rates later in Underserved Areas Housing Goal in previous Metropolitan Areas. This rule provides that this appendix, tracts are excluded if there are no years; within metropolitan areas, mortgage purchase or refinance applications. Tracts are also 4. The size of the conventional mortgage purchases will count toward the goal when excluded if: (1) Group quarters constitute more than market for central cities, rural areas, and those mortgages finance properties that are 50 percent of housing units or (2) there are less than other underserved areas relative to the size of located in census tracts where (1) median 15 home purchase applications in the tract and the the overall conventional mortgage market; income of families in the tract does not tract denial rates equal 0 or 100 percent. Excluded exceed 90 percent of area (MSA) median tracts account for a small percentage of mortgage 5. The ability of the enterprises to lead the loan applications (1.4 percent). These tracts are not industry in making mortgage credit available income or (2) minorities comprise 30 percent excluded from HUD’s underserved areas if they throughout the United States, including or more of the residents and median income meet the income and minority thresholds. Rather, central cities, rural areas, and other of families in the tract does not exceed 120 the tracts are excluded to remove the effects of underserved areas; and percent of area median income. outliers from the analysis.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00174 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63753
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00175 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.047 63754 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
HUD’s establishment of this definition is nonmetropolitan median income, or (2) proportions of counties, population, owner- based on a substantial number of studies of minorities comprise 30 percent or more of occupied housing units, and poverty mortgage lending and mortgage credit flows the residents and median income of families population in non-metropolitan areas. In conducted by academic researchers, in the county does not exceed 120 percent of terms of the 2000 Census geography and June community groups, the GSEs, HUD and other the greater of (a) state nonmetropolitan 2003 metropolitan area specification, the government agencies. As explained in the median income or (b) nationwide 2000 Rule, one finding stands out from the nonmetropolitan median income. definition covers 1,260 (61.4 percent) of the existing research literature on mortgage In 1995, two important factors influenced 2,052 counties in nonmetropolitan areas, access for different types of neighborhoods: HUD’s definition of nonmetropolitan which include 51.0 percent of the High-minority and low-income underserved areas—lack of available data for population, 50.7 percent of the owner- neighborhoods continue to have higher measuring mortgage availability in rural areas occupied housing units, and 64.3 percent of mortgage denial rates and lower mortgage and lenders’ difficulty in operating mortgage the population below the poverty level in origination rates than other neighborhoods. programs at the census tract level in rural non-metropolitan areas. The 1990-based A neighborhood’s minority composition and areas. Because of these factors, the 1995 Rule definition covered 1,514 (65.5 percent) of the its level of income are highly correlated with (as well as the 2000 Rule) used a more 2,311 counties in non-metropolitan areas, access to mortgage credit. inclusive, county-based approach to Nonmetropolitan Areas. In designating underserved portions of rural which included 54.6 percent of the nonmetropolitan areas, mortgage purchases areas. As discussed in a later section, HUD population, 53.4 percent of the owner- count toward the Underserved Areas Housing is now replacing the county-based definition occupied units, and 67.9 percent of the poor Goal for properties which are located in with a tract-based definition. in non-metropolitan areas.2 counties where (1) median income of families As shown in Table B.1b, switching from in the county does not exceed 95 percent of 1990 to 2000 Census data and incorporating 2 Kalawao County, Hawaii, which has a very the greater of (a) state nonmetropolitan the June, 2003 specification of metropolitan small population, is excluded from the analysis for median income or (b) nationwide areas causes a slight decrease in underserved 1990 but included for 2000.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00176 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63755
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00177 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.048
Data comparable to that in Table B.1b is contain 52.5 percent of the nonmetropolitan families most in need, as shown, for example, presented in Table B.1c based on census population (comparable to the 51.0 percent by the fact that it includes 68.9 percent of the tracts, rather than counties, in using a county-based definition) and 50.4 population in poverty, exceeding the nonmetropolitan areas. As indicated, the percent of owner-occupied housing units corresponding figure of 64.3 percent under tract-based definition includes 6,782 (54.9 (close to the corresponding figure of 50.7 the county-based definition of percent) of the 12,359 nonmetropolitan percent under the county-based approach). nonmetropolitan underserved areas. census tracts in the country. These tracts But the tract-based approach better targets
GSE Performance. Table B.1d shows the for Fannie Mae (adjusted to be comparable performance; and 31.6 percent and 27.7 increases in the GSEs’ overall goals with the 2000 figures) are 35.7 percent and percent for 2003 performance. (The 2001–03 performance under the more expansive 30.4 percent for 2001; 35.0 percent and 30.2 housing goals percentages in the table are geography of the 2000 Census. During 2000, percent for 2002; and 34.1 percent and 29.2 adjusted to exclude the effects of the bonus Fannie Mae’s performance would have been percent for 2003. The corresponding figures points and Freddie Mac’s Temporary an estimated 37.5 percent if underserved for Freddie Mac are 34.1 percent and 29.2 Adjustment Factor, which became applicable areas were defined in terms of 2000 Census percent for 2000 performance; 32.5 percent geography, compared with 31.0 percent and 28.2 percent for 2001 performance; 32.4 in 2001 for scoring of loans toward the under 1990 Census geography. These results percent and 28.0 percent for 2002 housing goals.)
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00178 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.049 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63757
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00179 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.050 63758 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Goal and Subgoal Levels. The Department access to mortgage credit, with higher lender bias against certain kinds of borrowers establishes the Underserved Areas Housing mortgage denial rates and lower origination relative to the degree to which they reflect Goal as 37 percent of eligible units financed rates for mortgages. Thus, the income and the credit quality of potential borrowers (as for 2005, 38 percent for 2006 and 2007, and minority composition of an area is a good indicated by applicants’ available assets, 39 percent for 2008. measure of whether that area is being credit rating, employment history, etc.). HUD is establishing a subgoal of 32 percent underserved by the mortgage market. Without fully accounting for the for the share of each GSE’s total single- • Research supports a targeted creditworthiness of the borrower, racial family-owner mortgage purchases that neighborhood-based definition of differences in denial rates cannot be finance single-family-owner properties underservice. Studies conclude that attributed to lender bias. Some studies of located in underserved census tracts of characteristics of mortgage loan applicants credit disparities have attempted to control metropolitan areas for 2005, with this and the neighborhood where the property is for credit risk factors that might influence a subgoal rising to 33 percent for 2006 and located are the major determinants of lender’s decision to approve a loan. 2007 and 34 percent in 2008. In this case, mortgage denial rates and origination rates. Boston Fed Study. The best example of subgoal performance for a particular calendar Once these characteristics are accounted accounting for credit risk is the study of year would be calculated for each GSE by for, other influences, such as location in a mortgage denial rates by researchers at the dividing (a) the number of mortgages central city, play only a minor role in Federal Reserve Bank of Boston.5 This purchased by the GSE that finance single- explaining disparities in mortgage lending.3 landmark study found that racial differentials family-owner properties located in 1. Discrimination in the Mortgage and in mortgage denial rates cannot be fully underserved areas (i.e., census tracts) of Housing Markets—An Overview explained by differences in credit risk. To control for credit risk, the Boston Fed metropolitan areas by (b) the number of The nation’s housing and mortgage markets mortgages purchased by the GSE that finance researchers included 38 borrower and loan are highly efficient systems, where most variables indicated by lenders to be critical single-family-owner properties located in homebuyers can put down relatively small metropolitan areas. As explained in Section to loan decisions. For example, the Boston amounts of cash and obtain long-term Fed study included a measure of the H, the purpose of this subgoal is to encourage funding at relatively small spreads above the the GSEs to lead the primary market in borrower’s credit history, which is a variable lender’s borrowing costs, even though not included in other studies. The Boston funding mortgages in underserved census transactions costs are still too high and too Fed study found that minorities’ higher tracts. bundled. Unfortunately, this highly efficient denial rates could not be explained fully by B. Consideration of Factors 1 and 2 in financing system does not work everywhere income and credit risk factors. The denial Metropolitan Areas: The Housing Needs of or for everyone. Studies have shown that rate for African Americans and Hispanics Underserved Urban Areas and Housing, access to credit often depends on improper was 17 percent, compared with 11 percent Economic, and Demographic Conditions in evaluation of characteristics of the mortgage for Whites with similar characteristics. That Underserved Urban Areas applicant and the neighborhood in which the is, African Americans and Hispanics were applicant wishes to buy. In addition, though about 60 percent more likely to be denied This section discusses differential access to racial discrimination has become less blatant credit than Whites, even after controlling for mortgage funding in urban areas and in the home purchase market, studies have credit risk characteristics such as credit summarizes available evidence on shown that it is still widespread in more history, employment stability, liquid assets, identifying those neighborhoods that have subtle forms. Partly as a result of these self-employment, age, and family status and historically experienced problems gaining factors, the homeownership rate for composition. Although almost all highly- access to credit. Section B.1 provides an minorities is substantially below that of qualified applicants were approved, overview of the problem of unequal access to whites. Appendix A provided an overview of differential treatment was observed among mortgage funding, focusing on discrimination the homeownership gaps and lending borrowers with more marginal qualifications. and other housing problems faced by disparities faced by minorities. This section That is, highly-qualified borrowers of all minority families and the communities briefly reviews evidence on lending races seemed to be treated equally, but in where they live. Section B.2 examines discrimination as well as a recent HUD- cases where there was some flaw in the mortgage access at the neighborhood level sponsored study of discrimination in the application, white applicants seemed to be and discusses in some detail the rationale for housing market. given the benefit of the doubt more the Underserved Areas Housing Goal in Mortgage Denial Rates. A quick look at frequently than minority applicants. A metropolitan areas. The most thorough mortgage denial rates reported by Home subsequent refinement of the data used by studies available provide strong evidence Mortgage Disclosure Act (HMDA) data the Federal Reserve Bank of Boston that low-income and high-minority census reveals that in 2002 minority denial rates confirmed the findings of that study.6 tracts are underserved by the mortgage were higher than those for white loan The Boston Fed study, as well as market. Section B.3 presents recent statistics applicants. For lower-income borrowers, the reassessments of that study by other on the credit characteristics and denial rate for African Americans applying researchers, concluded that the effect of socioeconomic characteristics of underserved for conventional loans was 2.1 times the borrower race on mortgage rejections persists areas under HUD’s definition. Readers are denial rate for white borrowers, while for even after controlling for legitimate referred to the expansive literature on this higher-income borrowers, the denial rate for determinants of lenders’ credit decisions.7 issue, which is reviewed in some detail in African Americans was 2.7 times the rate for Appendix B of HUD’s 2000 Rule. This white borrowers.4 5 Alicia H. Munnell, Lynn E. Browne, James section focuses on some of the main studies Differentials in denial rates, such as those McEneaney, and Geoffrey M.B. Tootell, ‘‘Mortgage and their findings. reported above, are frequently used to Lending in Boston: interpreting HMDA Data,’’ Three main points are made in this section: demonstrate the problems that minorities American Economic Review, March 1996. • Both borrowers and neighborhoods can face obtaining access to mortgage credit. 6 William C. Hunter, ‘‘The Cultural Affinity be identified as currently being underserved However, an important question is the degree Hypothesis and Mortgage Lending Decisions,’’ WP– to which variations in denial rates reflect 95–8, Federal Reserve Bank of Chicago, 1995. by the nation’s housing and mortgage Hunter confirmed that race was a factor in denial markets. Appendix A provided evidence of rates of marginal applicants. While denial rates racial disparities in the sale and rental of 3 In this appendix, the term ‘‘central city’’ is used were comparable for borrowers of all races with housing and in the provision of mortgage to mean ‘‘OMB-designated central city.’’ ‘‘good’’ credit ratings, among those with ‘‘bad’’ credit. Partly as a result of this, the 4 The actual denial rates were as follows: 23.6 credit ratings or high debt ratios, minorities were homeownership rate for minorities is percent for low-income (80% AMI or less) African significantly more likely to be denied than substantially below that for whites. Americans, 15.5 percent for upper-income (120% similarly-situated whites. The study concluded that • The existence of substantial AMI or more) African Americans. 11.4 percent for the racial differences in denial rates were consistent with a cultural gap between white loan officers and neighborhood disparities in mortgage credit low-income Whites, and 5.6 percent for upper- income Whites. The overall denial rate in the minority applicants, and conversely, a cultural is well documented for metropolitan areas. conventional conforming home purchase market affinity with white applicants. Research has demonstrated that census tracts was 9.7 percent in 2002. The data exclude 7 For a reassessment of the Boston Fed study, see with lower incomes and higher shares of applications to lenders that specialize in Stephen Ross and John Yinger, The Color of Credit, minority population consistently have poorer manufactured home lending. MIT Press 2002, and other studies cited there.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00180 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63759
Thus, these studies imply that variations in in discrimination toward African Americans Appendix A cited several studies showing mortgage denial rates, such as those reported seeking to rent a unit. This downward trend, that these inner city neighborhoods are often above, are not determined entirely by however, has not been seen for Hispanic served mainly by subprime lenders. In borrower risk, but reflect discrimination in renters, who now are more likely to addition, there is evidence that denial rates the housing finance system. However, the experience discrimination in their housing are higher in minority neighborhoods independent race effect identified in these search than are African American renters. regardless of the race of the applicant. The studies is still difficult to interpret. In But while generally down since 1989, the next section explores the issue of credit addition to lender bias, access to credit can report found that housing discrimination still availability in neighborhoods in more detail. be limited by loan characteristics that reduce exists at unacceptable levels. The greatest 2. Evidence About Access to Credit in Urban 8 profitability and by underwriting standards share of discrimination for Hispanic and Neighborhoods—An Overview that have disparate effects on minority and African American home seekers can still be lower-income borrowers and their attributed to being told units are unavailable HUD’s Underserved Areas Housing Goal neighborhoods.9 when they are available to non-Hispanic focuses on low-income and high-minority Paired-Testing Studies. As discussed in whites and being shown and told about fewer neighborhoods that are characterized by high Appendix A, paired testing studies of the units than a comparable non-minority. loan application denial rates and low loan pre-qualification process have supported the Although discrimination is down on most origination rates. As explained in Section B.3 findings of the Boston Fed study. Based on areas for African American and Hispanic below, the mortgage denial rate during 2001 a review of paired tests conducted by the homebuyers, there remain worrisome upward in census tracts defined as underserved by National Fair Housing Alliance, The Urban trends of discrimination in the areas of HUD was twice the denial rate in excluded Institute concluded that differential geographic steering for African Americans (or ‘‘served’’) tracts. In addition to such treatment discrimination at the pre- and, relative to non-Hispanic whites, the simple denial rate comparisons, there is a application level occurred at significant amount of help agents provide to Hispanics substantial economics literature justifying the levels in at least some cities. Minorities were with obtaining financing. On the rental side, targeted neighborhood definition that HUD less likely to receive information about loan Hispanics are more likely in 2000 than in has used to define underserved areas. products, received less time and information 1989 to be quoted a higher rent than their Appendix B of the 1995 and 2000 GSE Rules from loan officers, and were quoted higher white counterpart for the same unit. reviewed that literature in some detail; thus, interest rates in most of the cities where tests Another HUD-sponsored study asked this section simply provides an overview of were conducted.10 Another Urban Institute respondents to a nationwide survey if they the main studies supporting the need to study used the paired testing methodology to ‘‘thought’’ they had ever been discriminated improve credit access to low-income and examine the pre-application process in Los against when trying to buy or rent a house high-minority neighborhoods. Readers not Angeles and Chicago. African Americans and or an apartment.13 While the responses were interested in this overview may want to Hispanics faced a significant risk of unequal subjective, they are consistent with the proceed to Section B.3, which examines the treatment when they visited mainstream findings of the HDS. African Americans and credit and socioeconomic characterizes of the mortgage lending institutions to make pre- Hispanics were considerably more likely census tracts included in HUD’s underserved application inquiries.11 than whites to say they have suffered area definition. Sales and Rental Markets. In 2002, HUD discrimination—24 percent of African As explained in HUD’s 2000 Rule, the released its third Housing Discrimination Americans and 22 percent of Hispanics viability of neighborhoods—whether urban, Study (HDS) in the sale and rental of perceived discrimination, compared to only rural, or suburban—depends on the access of housing. The study, entitled Discrimination 13 percent of whites. their residents to mortgage capital to in Metropolitan Housing Markets: National Segregation in Urban Areas. purchase and improve their homes. While Results from Phase I of the Housing Discrimination, while not the only cause, neighborhood problems are caused by a wide Discrimination Study (HDS), was conducted contributes to the pervasive level of range of factors, including substantial by the Urban Institute.12 The results of this segregation that persists between African inequalities in the distribution of the nation’s HDS were based on 4,600 paired tests of Americans and Whites in our urban areas. income and wealth, there is increasing minority and non-minority home seekers The Census Bureau recently released one of agreement that imperfections in the nation’s conducted during 2000 in 23 metropolitan the most exhaustive studies of residential housing and mortgage markets are hastening areas nationwide. The report showed large segregation ever undertaken, entitled Racial the decline of distressed neighborhoods. decreases between 1989 and 2000 in the level and Ethnic Residential Segregation in the Disparate denial of credit based on of discrimination experienced by Hispanics United States: 1980–2000.14 The Census geographic criteria can lead to disinvestment and African Americans seeking to buy a Bureau found that the United States was still and neighborhood decline. Discrimination home. There has also been a modest decrease very much racially divided. While African and other factors, such as inflexible and Americans have made modest strides, they restrictive underwriting guidelines, limit 8 Since upfront loan fees are frequently remain the most highly segregated racial access to mortgage credit and leave potential determined as a percentage of the loan amount, group. The authors said that residential borrowers in certain areas underserved. lenders are discouraged from making smaller loans segregation likely results from a variety of Data on mortgage credit flows are far from in older neighborhoods, because such loans factors, including choices people make about perfect, and issues regarding the generate lower revenue and are less profitable to where they want to live, restrictions on their identification of areas with inadequate access lenders. choices, or lack of information. The fact that to credit are both complex and controversial. 9 Traditional underwriting practices may have many mainstream lenders do not operate in For this reason, it is essential to define excluded some lower income families that are, in segregated areas makes it even more difficult ‘‘underserved areas’’ as accurately as possible fact, creditworthy. Such families tend to pay cash, leaving them without a credit history. In addition, for minorities to obtain access to reasonable- based on existing data and evidence. There 15 the usual front-end and back-end ratios applied to priced mortgage credit. Section C.8 of are three sets of studies that provide the applicants’ housing expenditures and other on- rationale for the Department’s definition of going costs may be too stringent for lower income 13 How Much Do We Know? Public Awareness of underserved areas: (1) Studies examining households, who typically pay larger shares of their the Nation’s Fair Housing Laws, prepared for HUD racial discrimination against individual income for housing (including rent and utilities) by Martin D. Abravanel and Mary K. Cunningham mortgage applicants; (2) studies that test than higher income households. of the Urban Institute, April 2002. whether mortgage redlining exists at the 10 Margery A. Turner and Felicity Skidmore, eds., 14 U.S. Bureau of the Census, August 2002. The neighborhood level; and (3) studies that Mortgage Lending Discrimination: A Review of co-authors of the study were John Iceland and support HUD’s targeted approach to Existing Evidence. The Urban Institute: Washington, Daniel H. Wienberg. For a summary of the study, measuring areas that are underserved by the DC, June 1999. see ‘‘Residential Segregation Still Prevalent’’, mortgage market. In combination, these 11 Margery Austin Turner, All Other Things Being National Mortgage News, January 6, 2003, page 1. studies provide strong support for the Equal: A Paired Testing Study of Mortgage Lending 15 See Randall M. Scheessele, Black and White Institutions, The Urban Institute Press, April 2002. Disparities in Subprime Mortgage Refinance definition of underserved areas chosen by 12 Margery Austin Turner, Stephen L. Ross, Lending, Housing Finance Working Paper No. HF– HUD. The main studies of discrimination George Galster, and John Yinger, Discrimination in 14, Office of Policy Development and Research, against individuals have already been Metropolitan Housing Markets, The Urban Institute U.S. Department of Housing and Urban summarized in Section B.1 above. Thus, this Press, November 2002. Development, April 2002. section focuses on the neighborhood-based
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00181 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63760 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
studies in (2) and (3). As noted above, this More Comprehensive Tests of the Redlining that act as proxies for neighborhood risk, the brief overview of these studies draws from Hypothesis. Recent statistical studies have results do not reveal a pattern of redlining.’’22 Appendix B of the 1995 GSE Rule; readers sought to test the redlining hypothesis by Other Redlining Studies. To highlight the are referred there for a more detailed more completely controlling for differences methodological problems of single-equation treatment of earlier studies of the issues in neighborhood risk and demand. In these studies of mortgage redlining, Fred Phillips- discussed below. studies, the explanatory power of Patrick and Clifford Rossi developed a a. Controlling for Neighborhood Risk and neighborhood race is reduced to the extent simultaneous equation model of the demand Tests of the Redlining Hypothesis that the effects of neighborhood risk and and supply of mortgages, which they demand are accounted for; thus, they do not estimated for the Washington, DC In its deliberations leading up to support claims of racially induced mortgage metropolitan area.23 Phillips-Patrick and FHEFSSA, Congress was concerned about redlining. Many of these studies find that the Rossi found that the supply of mortgages is geographic redlining—the refusal of lenders race of the individual borrower is more negatively associated with the racial to make loans in certain neighborhoods important than the racial composition of the composition of the neighborhood, which led regardless of the creditworthiness of neighborhood. However, these studies cannot them to conclude that the results of single- individual applicants. During the 1980s and reach definitive conclusions about redlining equation models (such as the one estimated early 1990s, a number of studies using because segregation in inner cities makes it by Holmes and Horvitz) are not reliable HMDA data (such as that reported in Tables difficult to distinguish the impacts of indicators of redlining or its absence. B.2 and B.3, below) attempted to test for the geographic redlining from the effects of However, Phillips-Patrick and Rossi noted existence of mortgage redlining. Consistent individual discrimination. The following are that even their simultaneous equations model with the redlining hypothesis, these studies two good examples of these studies. does not provide definitive evidence of found lower volumes of loans going to low- Holmes and Horvitz examined variations redlining because important underwriting income and high-minority neighborhoods.16 in conventional mortgage originations across variables (such as credit history), which are However, such analyses were criticized census tracts in Houston.20 Their model omitted from their model, may be correlated because they did not distinguish between explaining census-tract variations in with neighborhood race. demand, risk, and supply effects 17—that is, mortgage originations included the following A few studies of neighborhood redlining they did not determine whether loan volume types of explanatory variables: (a) The have attempted to control for the credit was low because families in high-minority economic viability of the loan, (b) history of the borrower, which is the main and low-income areas were unable to afford characteristics of properties in and residents omitted variable in the redlining studies homeownership and therefore were not of the tract (e.g., house value, income, age reviewed so far. Samuel Myers, Jr. and Tsze applying for mortgage loans, or because distribution and education level), (c) Chan, who studied mortgage rejections in the borrowers in these areas were more likely to measures of demand (e.g., recent movers into state of New Jersey in 1990, developed a default on their mortgage obligations, or the tract and change in owner-occupied units proxy for bad credit based on the reasons that because lenders refused to make loans to between 1980 and 1990), (d) measures of lenders give in their HMDA reports for creditworthy borrowers in these areas.18, 19 credit risk (defaults on government-insured denying a loan.24 They found that 70 percent loans and change in tract house values of the gap in rejection rates could not be 16 These studies, which were conducted at the between 1980 and 1990), and (e) the racial explained by differences in Black and white census tract level, typically involved regressing the composition of the tract, as a test for the borrower characteristics, loan characteristics, number of mortgage originations (relative to the existence of racial redlining. Most of the neighborhoods or bad credit. Myers and Chan number of properties in the census tract) on neighborhood risk and demand variables concluded that the unexplained Black-white characteristics of the census tract including its were significant determinants of the flow of gap in rejection rates is a result of minority composition. A negative coefficient discrimination. With respect to the racial estimate for the minority composition variable was conventional loans in Houston. The often interpreted as suggesting redlining. For a coefficients of the racial composition composition of the census tract, they found discussion of these models, see Eugene Perle, variables were insignificant, which led that Blacks are more likely to be denied loans Kathryn Lynch, and Jeffrey Horner, ‘‘Model Holmes and Horvitz to conclude that in racially integrated or predominantly-white Specification and Local Mortgage Market allegations of redlining in the Houston neighborhoods than in predominantly-Black Behavior,’’ Journal of Housing Research, Volume 4, market could not be supported. neighborhoods. They concluded that middle- Issue 2, 1993, pp. 225–243. Schill and Wachter include several class Blacks seeking to move out of the inner 17 For critiques of the early HMDA studies, see individual borrower and neighborhood city would face problems of discrimination Andrew Holmes and Paul Horvitz, ‘‘Mortgage characteristics to explain mortgage in the suburbs.25 Redlining: Race, Risk, and Demand,’’ The Journal of acceptance rates in Philadelphia and Geoffrey Tootell has authored two papers Finance, Volume 49, No. 1, March 1994, pp. 81–99; 21 on neighborhood redlining based on the and Michael H. Schill and Susan M. Wachter, ‘‘A Boston. They found that the applicant race Tale of Two cities: Racial and Ethnic Geographic variables—whether the applicant was African Disparities in Home Mortgage Lending in Boston American or Hispanic—showed significant 22 Schill and Wachter, page 271. Munnell, et al. and Philadelphia,’’ Journal of Housing Research, negative effects on the probability that a loan reached similar conclusions in their study of Volume 4, Issue 2, 1993, pp. 245–276. would be accepted. Schill and Wachter stated Boston. They found that the race of the individual 18 Like early HMDA studies, an analysis of deed that this finding does not provide evidence mattered, but that once individual characteristics transfer data in Boston found lower rates of of individual race discrimination because were controlled, racial composition of the mortgage activity in minority neihborhoods. The applicant race is most likely serving as a neighborhood was insignificant. 23 discrepancies held even after controlling for proxy for credit risk variables omitted from Fred J. Phillips-Patrick and Clifford V. Rossi, ‘‘Statistical Evidence of Mortgage Redlining? A income, house values and other economic and non- their model (e.g., credit history, wealth and racial factors that might explain differences Cautionary Tale’’, The Journal of Real Estate differences in demand and housing market activity. liquid assets). Schill and Wachter find that Research, Volume 11, Number 1, 1996, pp.13–23. The study concluded that ‘‘the housing market and when their neighborhood risk proxies are 24 Samuel L. Myers, Jr. and Tsze Chan, ‘‘Racial the credit market together are functioning in a way included in the model along with the Discrimination in Housing Markets: Accounting for that has hurt African American neighborhoods in individual loan variables, the percentage of Credit Risk’’, Social Science Quarterly, Volume 76, the city of Boston.’’ Katherine L. Bradbury, Karl E. the census tract that was African American Number 3, September 1995, pp. 543–561. Case, and Constance R. Dunham, ‘‘Geographic became insignificant. Thus, similarly to 25 For another study that uses HMDA data on Patterns of Mortgage Lending in Boston, 1982– Holmes and Horvitz, Schill and Wachter reasons for denial to construct a proxy for bad 1987,’’ New England Economic Review, September/ stated that ‘‘once the set of independent credit, see Steven R. Holloway, ‘‘Exploring the October 1989, pp. 3–30. variables is expanded to include measures Neighborhood Contingency of Race Discrimination 19 Using an analytical approach similar to that of in Mortgage Lending in Columbus, Ohio’’, Annals Bradbury, Case, and Dunham, Anne Shlay found of the Association of American Geographers, evidence of fewer mortgage loans originated in Community: Methods for Assessing Residential Volume 88, Number 2, 1998, pp. 252–276. black census tracts in Chicago and Baltimore. See Credit Disparities, Market Barriers, and Institutional Holloway finds that mortgage denial rates are Anne Shlay, ‘‘Not in That Neighborhood: The Reinvestment Performance in the Metropolis,’’ higher for black applicants (particularly those who Effects of Population and Housing on the Journal of Urban Affiars, Volume 11, No. 3, 1989, are making large loan requests) in all-white Distribution of Mortgage Finance within the pp. 201–223. neighborhoods than in minority neighborhoods, Chicago SMSA,’’ Social Sciene Research, Volume 20 Holmes and Horvitz, op. cit. while the reverse is true for white applicants 17, No. 2, 1988, pp. 137–163; and ‘‘Financing 21 Schill and Wachter, op. cit. making small loan requests.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00182 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63761
mortgage rejection data from the Boston Fed and the credit problems facing low-income acceptance rates in Dade County, Florida, study.26 Tootell’s studies are important and minority neighborhoods focus on the also consistent with the Lang and Nakamura because they include a direct measure of ‘‘thin’’ mortgage markets in these hypothesis.31 borrower credit history, as well as the other neighborhoods and the implications of Robert Avery, Patricia Beeson, and Mark underwriting, borrower, and neighborhood lenders not having enough information about Sniderman found significant evidence of characteristics that are included in the the collateral and other characteristics of economies associated with the scale of Boston Fed data base; thus, his work does not these neighborhoods. The low numbers of operation of individual lenders in a have the problem of omitted variables to the house sales and mortgages originated in low- neighborhood.32 They concluded that ‘‘The same extent as previous redlining studies.27 income and high-minority neighborhoods inability to exploit these economies of scale Tootell found that lenders in the Boston area result in individual lenders perceiving these is found to explain a substantial portion of did not appear to be redlining neighborhoods neighborhoods to be more risky. It is argued the higher denial rates observed in low- based on the racial composition of the census that lenders do not have enough historical income and minority neighborhoods, where tract or the average income in the tract. information to project the expected default the markets are generally thin.’’ Low-income Consistent with the Boston Fed and Schill performance of loans in low-income and and minority neighborhoods often suffer and Wachter studies, Tootell found that it is high-minority neighborhoods, which from low transactions volume, and low the race of the applicant that mostly affects increases their uncertainty about investing in transactions volume represents a barrier to the mortgage lending decision; the location of these areas. the availability of mortgage credit by making the applicant’s property appears to be far less This recent group of studies that focus on mortgage lenders more reluctant to approve relevant. However, he did find that the economies of scale in the collection of and originate mortgage loans in these areas. decision to require private mortgage information about neighborhood b. Geographic Dimensions of Underserved insurance (PMI) depends on the racial characteristics has implications for the Areas—Targeted Versus Broad Approaches composition of the neighborhood. Tootell identification of underserved areas and suggested that, rather than redline understanding the problems of mortgage HUD’s definition of metropolitan themselves, mortgage lenders may rely on access in low-income and minority underserved areas is a targeted neighborhood private mortgage insurers to screen neighborhoods. William Lang and Leonard definition, rather than a broad definition that applications from minority neighborhoods. Nakamura argue that individual home sale would encompass entire cities. It also focuses Tootell also noted that this indirect form of transactions generate information which on those neighborhoods experiencing the redlining would increase the price paid by reduce lenders’ uncertainty about property most severe credit problems, rather than applicants from minority areas that are values, resulting in greater availability of neighborhoods experiencing only moderate approved by private mortgage insurers. mortgage financing.29 Conversely, appraisals difficulty obtaining credit. During the In a 1999 paper, Stephen Ross and Geoffrey in neighborhoods where transactions occur regulatory process leading to the 1995 rule, Tootell used the Boston Fed data base to take infrequently will tend to be more imprecise, some argued that underserved areas under a closer look at both lender redlining and the resulting in greater uncertainty to lenders this goal should be defined to include all role of private mortgage insurance (PMI) in regarding collateral quality, and more parts of all central cities, as defined by OMB. neighborhood lending.28 They had two main reluctance by them in approving mortgage HUD concluded that such broad definitions findings. First, mortgage applications for loans in neighborhoods with thin markets. As were not a good proxy for mortgage credit properties in low-income neighborhoods a consequence, ‘‘prejudicial practices of the problems—to use them would allow the were more likely to be denied if the applicant past may lead to continued differentials in GSEs to focus on wealthier parts of cities, did not apply for PMI. Ross and Tootell lending behavior.’’ rather than on neighborhoods experiencing concluded that their study provides the first If low-income or minority tracts have credit problems. Appendix B of the 1995 and direct evidence based on complete experienced relatively few recent 2000 Rules reviewed findings from academic underwriting data that some mortgage transactions, the resulting lack of information researchers that support defining applications may have been denied based on available to lenders will result in higher underserved areas in terms of the minority neighborhood characteristics that legally denial rates and more difficulty in obtaining and/or income characteristics of census should not be considered in the underwriting mortgage financing, independently of the tracts, rather than in terms of a broad process. Second, mortgage applicants were level of credit risk in these neighborhoods. A definition such as all parts of all central often forced to apply for PMI when the number of empirical studies have found cities. This section briefly reviews two of the housing units were in low-income evidence consistent with the notion that studies. The targeted nature of HUD’s neighborhoods. Ross and Tootell concluded mortgage credit is more difficult to obtain in definition is also examined in Section B.3 that lenders appeared to be responding to areas with relatively few recent sales below, which describes the credit and CRA by favoring low-income tracts once PMI transactions. Some of these studies have also socioeconomic characteristics of underserved has been received, and this effect counteracts found that low transactions volume may census tracts. the high denial rates for applications without contribute to disparities in the availability of Shear, Berkovec, Dougherty, and Nothaft PMI in low-income tracts. mortgage credit by neighborhood income and conducted an analysis of mortgage flows and Studies of Information Externalities. minority composition. Paul Calem found application acceptance rates in 32 Another group of studies related to redlining that, in low-minority tracts, higher mortgage metropolitan areas that supports a targeted loan approval rates were associated with definition of underserved areas.33 They 26 See Geoffrey M. B. Tootell, ‘‘Redlining in recent sales transactions volume, consistent 30 Boston: Do Mortgage Lenders Discriminate Against with the Lang and Nakamura hypothesis. 31 David C. Ling and Susan M. Wachter, Neighborhoods?’’, Quarterly Journal of Economics, While this effect was not found in high- ‘‘Information Externalities and Home Mortgage 111, November, 1996, pp. 1049–1079; and minority tracts, he concludes that Underwriting,’’ Journal of Urban Economics, ‘‘Discrimination, Redlining, and Private Mortgage ‘‘informational returns to scale’’ contribute to Volume 44, 1998, pp. 317–332. Insurance’’, unpublished manuscript, October 1995. disparities in the availability of mortgage 32 Robert B. Avery, Patricia E. Beeson, and Mark 27 Tootell notes that both omitted variables and credit between low-minority and high- S. Sniderman, ‘‘Neighborhood Information and the strong correlation between borrower race and minority areas. Empirical research by David Home Mortgage Lending,’’ Journal of Urban neighborhood racial composition in segregated Ling and Susan Wachter found that recent Economics, Volume 45, 1999, pp. 287–310. cities have made it difficult for previous studies to tract-level sales transaction volume does 33 William Shear, James Berkovec, Ann distinguish the impacts of geographic redlining significantly contribute to mortgage loan Dougherty, and Frank Nothaft, ‘‘Unmet Housing from the effects of individual borrower Needs: The Role of Mortgage Markets,’’ Journal of discrimination. He can unravel these effects Housing Economics, Volume 4 , 1996, pp. 291–306. because he includes a direct measure of credit 29 William W. Lang and Leonard I. Nakamura, ‘‘A These researchers regressed the number of mortgage history and because over half of minority applicants Model of Redlining,’’ Journal of Urban Economics, originations per 100 properties in the census tract in the Boston Fed data base applied for mortgages Volume 33, 1993, pp. 223–234. on several independent variables that were in predominately white areas. 30 Paul S. Calem, ‘‘Mortgage Credit Availability in intended to account for some of the demand and 28 Stephen L. Ross and Geoffrey M. B. Tootell, Low- and Moderate-Income Minority supply (i.e., credit risk) influences at the census ‘‘Redlining, the Community Reinvestment Act, and Neighborhoods: Are Information Externalities tract level. See also Susan Wharton Gates, ‘‘Defining Private Mortgage Insurance’’, unpublished Critical?’’ Journal of Real Estate Finance and the Underserved,’’ Secondary Mortgage Markets, manuscript, March 1999. Economics, Volume 13, 1996, pp. 71–89. Continued
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00183 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63762 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
found: (a) Low-income census tracts and for purchase and refinance loans, even after are controlled for, other influences such as tracts with high concentrations of African other variables were accounted for. Avery, central city location play only a minor role American and Hispanic families had lower Beeson and Sniderman concluded that a in explaining disparities in mortgage lending. rates of mortgage applications, originations, tract-level definition is a more effective way HUD recognizes that the mortgage and acceptance rates; and (b) once census to define underserved areas than using the origination and denial rates forming the basis tract influences were accounted for, central list of OMB-designated central cities as a for the research mentioned in the preceding city location had only a minimal effect on proxy. paragraph, as well as for HUD’s definition of credit flows. These authors recognized that it c. Conclusions from the Economics Literature underserved areas, are the result of the is difficult to interpret their estimated about Urban Underserved Areas interaction of individual risk, demand and minority effects—the effects may indicate supply factors that analysts have yet to fully lender discrimination, supply and demand The implications of studies by HUD and others for defining underserved areas can be disentangle and interpret. The need effects not included in their model but continues for further research addressing this correlated with minority status, or some summarized briefly. First, the existence of problem. combination of these factors. Still, they large geographic disparities in mortgage conclude that income and minority status are credit is well documented. Low-income and 3. Characteristics of HUD’s Underserved better indicators of areas with special needs high-minority neighborhoods receive Areas than central city location. substantially less credit than other a. Credit Characteristics Avery, Beeson, and Sniderman of the neighborhoods and fit the definition of being Federal Reserve Bank of Cleveland underserved by the nation’s credit markets. HMDA data provide information on the specifically addressed the issue of Second, researchers are testing models that disposition of mortgage loan applications underserved areas in the context of the GSE more fully account for the various risk, (originated, approved but not accepted by the legislation.34 Their study examined demand, and supply factors that determine borrower, denied, withdrawn, or not variations in application rates and denial the flow of credit to urban neighborhoods. completed) in metropolitan areas. HMDA rates for all individuals and census tracts The studies by Holmes and Horvitz, Schill data include the census tract location of the included in the 1990 and 1991 HMDA data and Wachter, and Tootell are examples of property being financed and the race and base. These authors found that the individual this research. Their attempts to test the income of the loan applicant(s). Therefore, applicant’s race exerts a strong influence on redlining hypothesis show the analytical this is a rich data base for analyzing mortgage mortgage application and denial rates. insights that can be gained by more rigorous activity in urban neighborhoods. HUD’s African American applicants, in particular, modeling of this issue. However, the fact that analysis using HMDA data for 2003 shows had unexplainably high denial rates. Once urban areas are highly segregated means that that high-minority and low-income census individual applicant and other neighborhood the various loan, applicant, and tracts have both relatively high loan characteristics were controlled for, overall neighborhood characteristics currently being application denial rates and relatively low denial rates for purchase and refinance loans used to explain credit flows are often highly loan origination rates. were only slightly higher in minority census correlated with each other, which makes it Table B.2 presents mortgage denial and tracts than non-minority census tracts. For difficult to reach definitive conclusions about origination rates by the minority composition white applicants, on the other hand, denial the relative importance of any single variable and median income of census tracts in rates were significantly higher in minority such as neighborhood racial composition. metropolitan areas. Two patterns are clear: tracts. That is, minorities had higher denial Thus, their results are inconclusive, and the • Census tracts with higher percentages of rates wherever they attempted to borrow, but need continues for further research on the minority residents have higher mortgage whites faced higher denials when they underlying determinants of geographic denial rates and lower mortgage origination attempt to borrow in minority disparities in mortgage lending.35 rates than all-white or substantially-white neighborhoods. In addition, Avery et al. Finally, much research strongly supports a tracts. For example, in 2003 the denial rate found that home improvement loans had targeted definition of underserved areas. for census tracts that are over 90 percent significantly higher denial rates in minority Studies by Shear, et al. and Avery, Beeson, minority (20.6 percent) was 2.3 times that for neighborhoods. Given the very strong effect and Sniderman conclude that characteristics census tracts with less than 10 percent of the individual applicant’s race on denial of both the applicant and the neighborhood minority (9.0 percent). rates, the authors noted that since minorities where the property is located are the major • tend to live in segregated communities, a Census tracts with lower incomes have determinants of mortgage denials and higher denial rates and lower origination policy of targeting minority neighborhoods origination rates—once these characteristics may be warranted. They also found that the rates than higher income tracts. For example, median income of the census tract had strong in 2003 mortgage denial rates declined from 35 effects on both application and denial rates Methodological and econometric challenges 23.2 percent to 7.2 percent as tract income that researchers will have to deal with are discussed increased from less than 40 percent of area in Mitchell Rachlis and Anthony Yezer, ‘‘Serious median income to more than 150 percent of 1994 Mortgage Market Review Issue, 1995, pp. 34– Flaws in Statistical Tests for Discrimination in area median income. 48. Mortgage Markets,’’ Journal of Housing Research, 34 See Avery, et al. Volume 4, 1993, pp. 315–336. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00184 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63763
• Table B.3 illustrates the interaction minority), high-income (over 120 percent of (under 90 percent of area median) group had between tract minority composition and tract area median) group had a denial rate of 7.2 a denial rate of 19.3 percent and an income by aggregating the data in Table B.2 percent and an origination rate of 32.4 loans origination rate of only 17.8 loans per 100 into nine minority and income combinations. per 100 owner occupants in 2003. The high- owner occupants. The other groupings fall The low-minority (less than 30 percent minority (over 50 percent), low-income between these two extremes.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00185 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.051 63764 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00186 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.052
The advantages of HUD’s underserved area HUD’s definition does not include high- mortgage credit. The 16.8 percent denial rate definition can be seen by examining the income (over 120 percent of area median) in these neighborhoods in 2003 was almost minority-income combinations highlighted in census tracts even if they meet the minority twice the 8.9 percent denial rate in the Table B.3. The sharp differences in denial threshold. The average denial rate (10.3 remaining areas of central cities. A broad, rates and origination rates between the percent) for high-income tracts with a inclusive definition of ‘‘central city’’ that underserved and remaining served categories minority share of population over 30 percent includes all areas of all central cities would illustrate that HUD’s definition delineates is much less than the denial rate (15.9 areas that have significantly less success in percent) in underserved areas as defined by include these ‘‘remaining’’ portions of cities. receiving mortgage credit. In 2003 HUD. Figure B.1 shows that these areas, which underserved areas had over one and a three- Figure B.1 compares underserved and account for approximately 36 percent of the fourths times the average denial rate of served areas within central cities and population in central cities, appear to be well served areas (15.9 percent versus 8.9 percent) suburbs. First, Figure B.1 shows that HUD’s served by the mortgage market. As a whole, and two-thirds the average origination rate definition targets central city neighborhoods they are not experiencing problems obtaining per 100 owner occupants (20.1 versus 29.1). that are experiencing problems obtaining mortgage credit.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00187 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63766 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00188 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.053 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63767
Second, Figure B.1 shows that HUD’s which account for 34 percent of the suburban census tracts is 18.5 percent, or over three definition also targets underserved census population, are included in HUD’s definition times the poverty rate (5.7 percent) in served tracts in the suburbs as well as in central of other underserved areas. census tracts. The unemployment rate and cities. The average denial rate in underserved b. Socioeconomic Characteristics the high-school dropout rate are also higher suburban areas (14.8 percent) is 1.7 times The targeted nature of HUD’s definition in underserved areas. In addition, there are that in the remaining served areas of the can be seen from the data presented in Table nearly three times more female-headed suburbs (8.7 percent), and is almost as large B.4, which show that families living in tracts households with children in underserved as the average denial rate (16.8 percent) in within metropolitan areas that are areas (30.0 percent) than in served areas (13.2 underserved central city tracts. Low-income underserved based on HUD’s definition percent). Three-fourths of units in served and high-minority suburban tracts appear to experience much more economic and social areas are owner-occupied, while only one- have credit problems similar to their central distress than families living in served areas. half of units in underserved areas are owner- city counterparts. These suburban tracts, For example, the poverty rate in underserved occupied.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00189 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63768 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00190 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.054
C. Consideration of Factors 1 and 2 in is compared with the greater of state terms and in buying decent and affordable Nonmetropolitan Areas: The Housing Needs nonmetropolitan income and national housing, and it is for them that the of Underserved Rural Areas and the Housing, nonmetropolitan income. This is based on geographic goals were designed. The Economic, and Demographic Conditions in HUD’s analysis of 1990 census data, which geographic goals, then, are meant to target Underserved Rural Areas indicated that comparing county places where these ‘‘underserved’’ Based on discussions with rural lenders in nonmetropolitan income only to state populations live in order to stimulate local 1995, the definition of underserved rural nonmetropolitan income would lead to the mortgage lending and, it is hoped, the areas was established at the county level, exclusion of many lower-income low- availability of credit to those families who since such lenders usually do not make minority counties from the definition, reside there who, otherwise, will have distinctions on a census tract basis. A especially in Appalachia. Based on 1990 difficulty securing credit. This section census geography, underserved counties nonmetropolitan county is classified as an addresses the basic question of whether and account for 57 percent (8,091 of 14,419) of underserved area if median income of the extent to which HUD’s definition of the census tracts and 54 percent of the families in the county does not exceed 95 underservice in nonmetropolitan areas percent of the greater of state population in rural areas. By comparison, the definition of metropolitan underserved areas effectively targets areas that encompass large nonmetropolitan or national nonmetropolitan populations of socially and economically median income, or minorities comprise 30 encompassed 47 percent of metropolitan disadvantaged families. percent or more of the residents and the census tracts and 44 percent of metropolitan Table B.5 shows data on demographic and median income of families in the county does residents. not exceed 120 percent of the greater of state The purchasing of loans from underserved socioeconomic conditions of underserved nonmetropolitan or national nonmetropolitan areas by the GSEs is intended to induce and served nonmetropolitan areas based on median income. For nonmetropolitan areas greater homeownership among moderate, HUD’s definition applied at the county level the median income component of the low, very low income, and poor families and using Census 2000 data. (A later section underserved definition is broader than that minorities. For various reasons, including considers the effects of applying the used for metropolitan areas. While tract creditworthiness and lending discrimination, definition of the census tract level.) Several income is compared with area income for these groups experience greater difficulty in variables are used to describe area metropolitan areas, in rural counties income securing loans under fair and reasonable demographic and socioeconomic conditions.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00191 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63770 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00192 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.055 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63771
On the national level, a few key results median income is 1.83 times higher than • The goal was set at 24 percent for 1997– show that the 1995 definition of underservice income required to qualify to buy a home of 2000. Fannie Mae’s performance was 28.8 captures a potentially disadvantaged segment median value in underserved areas. The percent in 1997, 27.0 percent in 1998, 26.8 of the population. In examining the minority comparable factor for served areas is 1.78. percent in 1999, and 31.0 percent in 2000; composition, one can see that the percentage • Rental affordability is approximately the and Freddie Mac’s performance was 26.3 of African Americans, Hispanics/Latinos, and same in underserved and served areas. percent in 1997, 26.1 percent in 1998, 27.5 total minority population is higher in • While nearly all housing in served and percent in 1999, and 29.2 percent in 2000. underserved nonmetropolitan areas as underserved areas have complete plumbing • In the October 2000 rule, the compared to served nonmetropolitan areas. and kitchens, the percentage of units with underserved areas goal was set at 31 percent Overall, the minority population of incomplete facilities in underserved is twice for 2001–03. As of January 1, 2001, several underserved areas is 25.8 percent as the percentage in served areas. changes in counting requirements came into compared with 9.3 percent in served areas. • Crowded units are a small share of all effect for the undeserved areas goal, as Other supporting results include median housing in nonmetropolitan areas, but the follows: ‘‘bonus points’’ (double credit) for family income, poverty rate, unemployment rate is higher for underserved areas: 4.3 vs. purchases of goal-qualifying mortgages on rate, school dropout rate, and in-migration 2.3 percent. small (5–50 unit) multifamily properties and, rate. Specifically we find: Mikesell 37 found using the 1995 American above a threshold level, mortgages on 2–4 • Median income is approximately Housing Survey that while the rate of unit owner-occupied properties; a $10,000 less in underserved areas than in homeownership in nonmetropolitan areas is ‘‘temporary adjustment factor’’ (1.20 units served areas. This represents an average gap higher than metropolitan areas, the quality of credit, subsequently increased by Congress to of 25 percent. housing is lower as compared to 1.35 units credit) for Freddie Mac’s • Poverty in underserved areas is twice the metropolitan areas. Results based on the 2000 purchases of goal-qualifying mortgages on rate in served areas (14.5 vs. 7.5 percent). Census show that the homeownership rate large (more than 50-unit) multifamily • Unemployment is 7.3 percent in for nonmetropolitan areas was 74 percent (73 properties; and eligibility for purchases of underserved areas and 5.2 percent in served percent without manufactured homes), and certain qualifying government-backed loans areas. for metropolitan areas it was 64 percent, but to receive goal credit. These changes are • The school dropout rate is 28.1 percent both metropolitan and nonmetropolitan areas explained below. Fannie Mae’s performance in underserved areas and 18.7 percent in had approximately 97.5 percent of units with was 32.6 percent in 2001, 32.4 percent in served areas. complete plumbing and 99 percent with 2002, and 32.1 percent in 2003; and Freddie • Migration into underserved areas is complete kitchens. Mac’s performance was 31.7 percent in 2001, somewhat lower than in served areas: 7.4 vs. slightly less than 31 percent in 2002, and D. Factor 3: Previous Performance and Effort 32.7 percent in 2003, thus Fannie Mae 8.0 percent. of the GSEs in Connection With the Central Table B.5 also includes data on surpassed this higher goal in all three years Cities, Rural Areas and Other Underserved and Freddie Mac surpassed the goal in 2001 homeownership rates, housing affordability, Areas Goal housing quality, and overcrowding. On and 2003, but fell slightly short in 2002. This several of these dimensions, housing Section D.1 reports the past performance of section discusses the October 2000 counting each GSE with regard to the Underserved conditions and needs in underserved areas rule changes in detail below, and provides Areas Housing Goal. Section D.2 then are not substantially worse than in served data on what goal performance would have examines the role that the GSEs are playing 39 areas. Although housing quality and been in 2001–03 without these changes. in funding single-family mortgages in crowding appear to be marginally worse in a. Performance on the Underserved Areas underserved urban neighborhoods based on underserved areas, homeownership in the Housing Goal in 1996–2003 HUD’s analysis of GSE and HMDA data. That two areas is about the same and owning a section also discusses an underserved area HUD’s December 1995 rule specified that home actually appears to be more affordable subgoal for home purchase loans. Section D.3 in 1996 at least 21 percent of the number of in underserved areas than in served areas. concludes this section with an analysis of the units financed by each of the GSEs that were Specific findings include the following: eligible to count toward the Underserved • GSEs’ purchases in rural (nonmetropolitan) Homeownership is slightly higher in areas. Areas Goal should qualify as units in underserved than in served nonmetropolitan The increased coverage of the Underserved properties located in underserved areas, and counties: 74.3 percent vs. 73.7 percent. Areas Housing goal due to switching to 2000 at least 24 percent should qualify in 1997– Removing manufactured homes lowers census geography is discussed throughout 2000. HUD’s October 2000 rule made various ownership rates slightly, because ownership this section. changes in the goal counting rules, as of such homes is relatively high, but this discussed below, and increased the does not affect the basic result. 1. Past Performance of the GSEs Underserved Areas Goal to 31 percent for • Owner-occupied and rental vacancy This section discusses each GSE’s 2001–03. rates are both somewhat higher in performance under the Underserved Areas Table B.6 shows performance on the underserved areas. Housing Goal over the 1996–2003 period.38 underserved areas goal over the 1996–2003 • Median housing unit values are As explained in Appendix A, the data period, based on HUD’s analysis. The table significantly lower in underserved areas: presented are ‘‘official HUD results’’ which, shows that Fannie Mae surpassed the goals $67,358 vs. $88,099. in some cases, differ from goal performance by 7.1 percentage points and 4.8 percentage • The value of a housing affordability reported by the GSEs in the Annual Housing points in 1996 and 1997, respectively, while index for owner-occupied housing is slightly Activities Reports (AHARs) that they submit Freddie Mac surpassed the goals by narrower higher in underserved areas.36 On average, to the Department. margins, 4.0 and 2.3 percentage points. In The main finding of this section is that 1998 Fannie Mae’s performance fell by 1.8 36 The purchase affordability index assesses the Fannie Mae surpassed the Department’s percentage points, while Freddie Mac’s extent to which a family with the median income Underserved Areas Housing Goals for each of performance fell only slightly, by 0.2 of a given area would be able to afford a housing the seven years during this period. Freddie percentage point. Freddie Mac showed a gain unit that carries the median purchase price of that Mac surpassed the goal in six of the seven in performance to 27.5 percent in 1999, area. For example, a purchase affordability index years, falling slightly short in 2002. exceeding its previous high by 1.2 percentage number less than 100 means that a family with the Specifically: points. Fannie Mae’s performance in 1999 median income would not qualify for a mortgage on • The goal was set at 21 percent for 1996; was 26.8 percent, which, for the first time, a unit with the median value; a purchase Fannie Mae’s performance was 28.1 percent slightly lagged Freddie Mac’s performance in affordability index equal to 100 means that a family and Freddie Mac’s performance was 25.0 with the median income has exactly the level of that year. income needed to qualify for a mortgage on a unit percent. BILLING CODE 4210–27–P with the median value; and an index number greater than 100 means that a family with the 37 J.J. Mikesell, ‘‘Housing Problems across Types 39 To separate out the effects of changes in median income has 20 percent more than the level of Rural Households’’, Rural Conditions and counting rules that took effect in 2001, this section of income needed to qualify for a mortgage on a unit Trends, Volume 9, Number 2, pp. 97–101, 1999. also compares performance in 2001 to estimated with the median value. The rental affordability 38 Performance for the 1993–95 period was performance in 2000 if the 2001 counting rules had index is similarly constructed. discussed in the October 2000 rule. been in effect in that year.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00193 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63772 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00194 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.056
Both GSEs exhibited sharp gains in goal goal, at 31.4 percent in 2003, and fallen short in large multifamily properties in performance in 2000—Fannie Mae’s of the goal in 2001 and 2002. underserved areas whose mortgages were performance increased by 4.2 percentage b. Changes in the Goal Counting Rules for financed by Fannie Mae during this period. points, to a record level of 31.0 percent, 2001–03 Purchases of certain government-backed while Freddie Mac’s performance increased loans. Prior to 2001, purchases of somewhat less, by 1.7 percentage points, Several changes in the counting rules underlying the calculation of underserved government-backed loans were not taken into which also led to a record level of 29.2 account in determining performance on the percent. Fannie Mae’s performance was 32.6 areas goal performance took effect beginning GSEs’ low- and moderate-income and percent in 2001, 32.4 percent in 2002, and in 2001. These also applied to the low- and underserved area housing goals. As discussed 32.1 percent in 2003; Freddie Mac’s moderate-income goal and are discussed in Appendix A; only brief summaries of those in Appendix A, the 2000 rule established performance was 31.7 percent in 2001, 40 slightly less than 31 percent in 2002, and changes are given here: Bonus points for eligibility for FHA-insured home equity 32.7 percent in 2003. However, as discussed multifamily and single-family rental conversion mortgages (HECMs) for below, using consistent accounting rules for properties. Each qualifying unit in a small mortgagors in underserved areas, purchases 2000–03, under one method each GSE’s multifamily property counted as two units in of mortgages on properties on tribal lands performance in 2001–03 was below its the numerator in calculating performance on insured under FHA’s Section 248 program or performance in 2000. all of the goals for 2001–03. And, above a HUD’s Section 184 program, and purchases threshold equal to 60 percent of the average Fannie Mae’s performance on the of mortgages under the Rural Housing number of qualifying rental units financed in underserved areas goal surpassed Freddie Service’s Single Family Housing Guaranteed Mac’s in every year through 1998. This owner-occupied properties over the preceding five years, each unit in a 2–4 unit Loan Program to count toward the pattern was reversed in 1999, as Freddie Mac underserved area goal. surpassed Fannie Mae in goal performance owner-occupied property also counted as two for the first time, though by only 0.7 units in the numerator in calculating goal c. Effects of Changes in the Counting Rules percentage point. This improved relative performance. on Goal Performance performance of Freddie Mac was due to its Freddie Mac’s Temporary Adjustment Because of the changes in the underserved Factor. Freddie Mac received a ‘‘Temporary increased purchases of multifamily loans, as areas goal counting rules that took effect in Adjustment Factor’’ of 1.35 units of credit for it re-entered that market, and to increases in 2001, direct comparisons between official each qualifying unit financed in ‘‘large’’ the goal-qualifying shares of its single-family goal performance in 2000 and 2001–03 are mortgage purchases. However, Fannie Mae’s multifamily properties (i.e., those with 51 or more units) in the numerator in calculating somewhat of an ‘‘apples-to-oranges performance once again exceeded Freddie comparison.’’ For this reason, the Department Mac’s performance in 2000, 31.0 percent to its performance on the housing goals for 2001–03.41 This factor did not apply to units has calculated what performance would have 29.2 percent. Fannie Mae’s official been in 2000 under the 2001–03 rules; this performance also exceeded Freddie Mac’s may compared with official performance in official performance in 2001–02, despite the 40 Unlike the low- and moderate-income and 2001–03—an ‘‘apples-to-apples comparison.’’ fact that Freddie Mac benefited from a special affordable goals, there is no exclusion of difference in the counting rules applicable to units from the denominator for units with missing HUD has also calculated what performance information about the area in which a property is the two GSEs as enacted by Congress; if the would have been in 2001–03 under the 1996– located. That is, such units are counted in the 2000 rules; this may be compared with same counting rules were applied to both denominator, but not in the numerator, in GSEs, Fannie Mae’s performance would have determining underserved areas goal performance. official performance in 2000—an ‘‘oranges-to- exceeded Freddie Mac’s performance. In fact, 41 See Congressional Record, December 15, 2000, oranges comparison.’’ These comparisons are Freddie Mac would have just attained the pp. H12295–96. presented in Table B.7a.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00195 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63774 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00196 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.057 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63775
Specifically, Table B.7a shows Freddie Mac. The largest impact of the Residential Finance Survey finding that performance under the underserved areas counting rule changes on Freddie Mac’s goal small multifamily properties account for 37 goal in three ways. Baseline A represents the performance was due to bonus points for percent of total units in multifamily counting rules in effect in 1996–2000. purchases of mortgages on small multifamily mortgaged properties, thus Fannie Mae is Baseline B incorporates the one minor properties; this added 1.3 percentage points still somewhat less active in this market than technical change in counting rules pertaining to goal performance in 2001, 0.5 percentage in the market for large multifamily to the underserved areas goal—eligibility of point in 2002, and 2.9 percentage points in properties.42 certain government-backed loans for goals 2003, as shown in Table B.7a. The Within the small multifamily market, there credit. Baseline C incorporates in addition to application of the temporary adjustment was some evidence that Fannie Mae targeted that technical change the bonus points and, factor for purchases of mortgages on large properties in underserved areas to a greater for Freddie Mac, the temporary adjustment multifamily properties enacted by Congress extent in 2001 than in 2000. That is, 56 factor. Boldface figures under Baseline A for added 0.9 percentage points to goal percent of Fannie Mae’s small multifamily 1999–2000 and under Baseline C for 2001– performance in 2002 and 1.3 percentage units qualified for the underserved areas goal 02 indicate official goal percentages based on points in 2003. Bonus points for purchase of in 2000, but this rose to 64 percent in 2001. the counting rules in effect in those years— mortgages on owner-occupied 2–4 unit rental The share of Fannie Mae’s small multifamily e.g., for Freddie Mac, 27.5 percent in 1999, properties also added 1.1 percentage points units that qualified for the underserved areas 29.2 percent in 2000, 31.7 percent in 2001, to performance in 2001, 1.6 percentage points goal was 65 percent in 2002 and 50 percent slightly less than 31 percent in 2002, and in 2002, and 0.9 percentage point in 2003. in 2003. 32.7 percent in 2003. Credit for purchases of qualifying Freddie Mac financed 50,211 units in small Performance on the Underserved Areas government-backed loans played a minor role multifamily properties in 2001 that were Goal under 1996–2000 Counting Rules Plus in determining Freddie Mac’s goal eligible for the underserved areas goal, an Technical Changes. If the ‘‘Baseline B’’ performance. increase of more than 1500 percent from the counting approach had been in effect in Fannie Mae. The temporary adjustment small base of 2,985 units financed in 2000. 2000–03 and the GSEs’’ had purchased the factor which applied to Freddie Mac’s goal Financing of such units actually fell in 2002, same mortgages that they actually did performance did not apply to Fannie Mae, to 22,195 units, but rebounded to 181,126 purchase in those years, Fannie Mae would thus overall counting rule changes had less units in 2003. Small multifamily properties have just matched the underserved areas goal impact on its performance than on Freddie also accounted for a significantly greater in 2000 and fallen short in 2001–03, while Mac’s performance in 2001–03. The largest share of Freddie Mac’s multifamily business Freddie Mac would have fallen short of the impact of the counting rule changes on in 2001—16.1 percent of total multifamily goal in all four years, 2000–03. Specifically, Fannie Mae’s goal performance was due to units financed, up from 1.8 percent in 2000, Fannie Mae’s performance would have been the application of bonus points for purchases with this share amounting to 7.1 percent in 31.0 percent in 2000, 30.4 percent in 2001, of mortgages on owner-occupied 2–4 unit 2002 and 30.5 percent in 2003. 30.2 percent in 2002, and 29.2 percent in rental properties, which added 1.7 Within the small multifamily market, there 2003. Freddie Mac’s performance would have percentage points to performance in 2001, 1.8 was some evidence that Freddie Mac targeted been 29.2 percent in 2000, 28.2 percent in percentage points in 2002, and 1.7 percentage properties in underserved areas to a greater 2001, 28.0 percent in 2002, and 27.7 percent points in 2003, and for purchases of extent in 2001 than in 2000. That is, 61 in 2003. mortgages on small multifamily properties, percent of Freddie Mac’s small multifamily Performance on the Underserved Areas which added 0.5 percentage point to units qualified for the underserved areas goal Goal under 2001–2003 Counting Rules. If the performance in 2001, 0.8 percentage point in in 2000; this rose to 86 percent in 2001. The 2001–03 counting rules had been in effect in 2002, and 1.2 percentage points in 2003. share of Freddie Mac’s small multifamily 2000–02 and the GSEs had purchased the Credit for purchases of qualifying units that qualified for the underserved areas same mortgages that they actually did government-backed loans also played a goal was 88 percent in 2002 and 87 percent purchase in those years (i.e., abstracting from minor role in determining Fannie Mae’s goal in 2003. any behavioral effects of ‘‘bonus points,’’ for performance. Bonus points for single-family rental example), both GSEs would have surpassed d. Bonus Point Incentives for the GSEs’ properties. Above a threshold, each unit the underserved areas goal in all four years, Purchases in Underserved Areas financed in a 2–4 unit property with at least one owner-occupied unit (referred to as and both GSEs’ performance figures would The Department established ‘‘bonus ‘‘OO24s’’ below) that qualified for any of the have increased from 2000 to 2002. points’’ for 2001–03 to encourage the GSEs to housing goals was counted as two units in Specifically, Fannie Mae’s ‘‘Baseline C’’ step up their activity in two segments of the the denominator (and one unit in the performance would have been 32.3 percent mortgage market—the small (5–50 unit) numerator) in calculating goal performance in 2000, 32.6 percent in 2001, 32.4 percent multifamily mortgage market, and the market in 2002, and 32.1 percent in 2003. Freddie for mortgages on 2–4 unit properties where for that goal in 2001–03. The threshold was Mac’s performance would have been 31.4 1 unit is owner-occupied and 1–3 units are equal to 60 percent of the average number of percent in 2000, 31.7 percent in 2001, occupied by renters. such qualifying units over the previous five slightly less than 31.0 percent in 2002, and Bonus points for small multifamily years. For example, Fannie Mae financed an 32.7 percent in 2003. Measured on this properties. Each unit financed in a small average of 47,100 underserved area units in consistent basis, then, Fannie Mae’s multifamily property that qualified for any of these types of properties between 1996 and performance increased by 0.3 percentage the housing goals was counted as two units 2000, and 105,946 such units in 2001. Thus point in 2001, fell by 0.7 percentage points in the denominator (and one unit in the in 2001 Fannie Mae received 77,688 bonus in 2002, and increased by 1.3 percentage numerator) in calculating goal performance points in this area in 2001—that is, 105,946 points in 2003. These increases were the for that goal. minus 60 percent of 47,100. So 183,629 units effect of increased purchases of mortgages Fannie Mae financed 37,389 units in small were entered in the numerator for these eligible to receive bonus points between 2000 multifamily properties in 2001 that were properties in calculating underserved area and 2001–03. eligible for the underserved areas goal, an goal performance. Details of Effects of Changes in Counting increase of more than 400 percent from the Single-family rental bonus points thus Rules on Goal Performance in 2001. As 7,196 units financed in 2000. Further encouraged the GSEs to play a larger role in discussed above, counting rule changes that increases were recorded in 2002, to 77,382 this market, and also to purchase mortgages took effect in 2001 had significant impacts on units, and in 2003, to 230,290 units. As on such properties in which large shares of the performance of both GSEs on the explained in Appendix A, small multifamily the units qualify for the housing goals. As for underserved areas goal in that year—2.2 properties also accounted for a greater share small multifamily bonus points, some percentage points for Fannie Mae, and 3.5 of Fannie Mae’s multifamily business in evidence on the effects of such bonus points percentage points for Freddie Mac. This 2001—7.4 percent of total multifamily units on the GSEs’ operations may be gleaned from section breaks down the effects of these financed, up from 2.5 percent in 2000, with the data provided to HUD by the GSEs for changes on goal performance for both GSEs; this share rising to 16.8 percent in 2002 and 2001–2003. results are shown in Table B.7a along with 28.9 percent in 2003. However, HUD’s figures for other years. Housing Goals 2000 Final Rule cited a 42 65 FR 65141 & n. 145 (2000).
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00197 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63776 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Fannie Mae financed 177,872 units in MSA, and the percentage minority the basis of determination for non- OO24s in 2001 that were eligible for the population in the census tract where the metropolitan areas will be changed for underserved areas goal, an increase of 116 property is located. For properties located properties located outside of MSAs to: The percent from the 82,464 units financed in outside of MSAs scoring is based on the median income of the census tract where the 2000. Further increases were recorded in median income of the county, the median property is located; the median income of the 2002, to 231,581 units, and in 2003, to income of the non-metropolitan portion of non-metropolitan portion of the State in 353,916 units. However, as a result of the the State in which the property is located or which the property is located or of the non- refinance boom Fannie Mae’s total single- of the non-metropolitan portion of the United metropolitan portion of the United States, family business increased at approximately States, whichever has the larger median whichever is larger; and the percentage the same rate as its OO24 business in 2001– income, and the percentage minority minority population in the census tract 03, thus the share of its business accounted population in the county where the property where the property is located. for by OO24s was the same in 2001 as in is located. Thus, scoring loans under the Analysis. HUD used 2000 census data to 2000—4 percent, with this share also Underserved Areas Housing Goal requires generate underserved area designations for amounting to 4 percent in 2002 and 2003. decennial census data on median incomes for census tracts as defined for the 2000 census Within the OO24 market, there was no metropolitan census tracts, MSAs, non- evidence that Fannie Mae targeted affordable metropolitan counties, the non-metropolitan with 2003 MSA designations. Because Fannie properties to a greater extent in 2001 than in portions of States, and the non-metropolitan Mae and Freddie Mac geocoded the 2000. That is, approximately 60 percent of portion of the United States. The mortgages they purchased prior to 2003 Fannie Mae’s OO24 units qualified for the determination has been based on 1990 census based on census tract boundaries as underserved area goal in both 2000 and 2001. data through 2004, and beginning in 2005 established for the 1990 census, GSE The share of Fannie Mae’s OO24 units that will be based on 2000 census data.43, 44 Under mortgages purchased prior to 2003 can be qualified for the underserved areas goal was this rule, the basis for the determination directly identified as being from a served or 62 percent in 2002 and 60 percent in 2003. outside of MSAs will change from counties underserved area only where the property is Freddie Mac financed 96,983 units in to census tracts beginning in 2005. located in a 1990-defined census tract whose OO24s in 2001 that were eligible for the 2005 Procedure. Relative to the above area consists entirely of whole 2000-defined underserved areas goal, an increase of 91 procedure, Underserved Areas Housing Goals census tracts, or portions of such tracts, percent from the 50,868 units financed in performance percentages for loans purchased which are all designated either as served or 2000. Further increases were recorded in by the GSEs in and after 2005 will be affected as underserved. In the situation where the 2002, to 146,502 units, and in 2003, to by three factors. First, 2000 census data on area of a 1990-defined census tract includes 154,924 units. However, with the refinance median incomes and minority populations whole 2000-defined census tracts, or portions boom, Freddie Mac’s total single-family replace 1990 census data. Second, the Office of such tracts, some of which are served and business increased at approximately the same of Management and Budget in June, 2003, some underserved, HUD calculated an rate as its OO24 business in 2001–03, thus respecified MSA boundaries based on ‘‘underservice factor’’ defined as the the share of its business accounted for by analysis of 2000 census data. Third, the underserved percentage of the 1990-defined OO24s was the same in 2001 as in 2000—3 Department’s re-specification of the tract’s population, based on population data percent, with this share amounting to 3.7 Underserved Areas goal in terms of census from the 2000 census.46 These factors were percent in 2002 and 3.1 percent in 2003. tracts rather than counties in non- used in estimating underservice percentages As for Fannie Mae, within the OO24 metropolitan areas will come into effect.45 for aggregated GSE purchases in and before market there was no evidence that Freddie Thus, for properties located outside of MSAs 2003 based on the 2000 census. Mac targeted affordable properties to a The resulting underserved areas file was greater extent in 2001 than in 2000. That is, 43 In New England, MSAs were defined through used to re-score loans purchased by the GSEs 60 percent of Fannie Mae’s OO24 units mid-2003 in terms of Towns rather than Counties, between 1999 and 2003, and was used qualified for the underserved areas goal in and the portion of a New England county outside further in estimating the share of loans of any MSA is regarded as equivalent to a county both 2000 and 2001. The share of Freddie originated in metropolitan areas that would Mac’s OO24 units that qualified for the in establishing metropolitan or non-metropolitan location of a property. The MSA definitions be eligible to score toward the Underserved underserved areas goal was 61 percent in Areas Housing Goal, from HMDA data. The 2002 and 50 percent in 2003. established by the Office of Management and Budget (OMB) in June, 2003 defined MSAs in New results of the retrospective GSE analysis are e. Effects of 2000 Census on Scoring of Loans England in terms of counties. provided in Table B.7b The results of the Toward the Underserved Areas Housing Goal 44 The procedure used to generate estimated rents GSE–HMDA comparative analysis are Background. Scoring of housing units in connection with the Low- and Moderate Income presented in the next section. and Special Affordable Housing Goals, as BILLING CODE 4210–27–P under the Underserved Areas Housing Goal mentioned in Appendixes A and C, uses similar is based on decennial census data used to data series. identify underserved areas, as follows: For 45 HUD has deferred application of the 2000 46 8,717 tracts included both served and properties in MSAs scoring is based on the census data and 2003 MSA designations to 2005, underserved area, out of a total of 61,493 tracts that median income of the census tract where the pending completion of the present rulemaking could be classified as served or underserved or property is located, the median income of the process. assigned an underservice factor.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00198 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63777
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00199 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.058
Table B.7b shows four sets of estimates for GSEs’ performance in underserved properties located in underserved each GSE, based respectively on the counting neighborhoods is compared with the overall neighborhoods for the years 1993 to 2003. To rules in place in 2001–2003 (but disregarding market. This section therefore expands on the take advantage of historical data going back the bonus points and Temporary Adjustment discussion in Appendix A, which compared to 1993, these comparisons were first made Factor), on shifting from 1990 to 2000 census the GSEs’ funding of affordable loans with using 1990 Census tract geography. The data on median incomes and minority the overall conventional conforming market. findings with respect to the GSEs’ funding of concentrations, on the further addition 2003 A subgoal that the Department is establishing underserved neighborhoods are similar to MSA specification, and finally on shifting for each GSE’s acquisitions of home purchase those reported in Appendix A regarding the from counties to tracts as the basis for scoring loans financing properties in the underserved GSEs’ overall affordable lending performance loans in non-metropolitan areas. census tracts of metropolitan areas is also in the single-family-owner market. While both GSEs improved their performance, they 2. GSEs’ Mortgage Purchases in Metropolitan discussed subsection 2a. In subsection 2.b., historically lagged the conventional Neighborhoods the characteristics of the GSEs’ purchases within underserved areas are compared with conforming market in providing affordable Metropolitan areas accounted for about 85 those for their purchases in served areas. loans to underserved neighborhoods. The percent of total GSE purchases under the two GSEs themselves engaged in very Underserved Areas Housing Goal in 2001 and a. Comparisons With the Primary Market different patterns of funding—Freddie Mac 2002. This section uses HMDA and GSE data Market Comparisons Based on 1990 was less likely than Fannie Mae to fund for metropolitan areas to examine the Census Geography. Section E.8–10 in home loans in underserved neighborhoods, neighborhood characteristics of the GSEs’ Appendix A provided detailed information as the following percentage shares for home mortgage purchases. In subsection 2.a, the on the GSEs’ funding of mortgages for purchase loans indicate:
Market Year Freddie Mac Fannie Mae (w/o B&C) (percent) (percent) (percent)
1996–2003 ...... 22.0 24.0 25.7 1999–2003 ...... 23.1 24.7 26.2 2001–2003 ...... 24.1 26.0 26.4
Between 1996 and 2003, 22.0 percent of longer-term performance (since 1993 or 1996) allocate 1990-based GSE and HMDA data Freddie Mac’s purchases financed properties as well as its recent average performance into census tracts as defined by the 2000 in underserved neighborhoods, compared (1999 to 2003) has consistently been below Census. GSE and HMDA data for 2003 were with 24.0 percent of Fannie Mae’s purchases market levels. Still, it is encouraging that already expressed in terms of 2000 Census and 25.7 percent of home purchase loans Fannie Mae significantly improved its 2001– geography. originated in the conventional conforming 2003 performance and closed its gap with the The main results are provided in Table B.8, market (excluding B&C loans). Thus, Freddie market during the first three years of HUD’s which compares the GSEs to the market Mac performed at only 86 percent of the higher housing goal levels. using both the 1990 Census geography and market (22.0 divided by 25.7), while Fannie Market Comparisons Based on 2000 the 2000 Census geography. Switching to the Mae performed at 93 percent of the market. Census Geography. As explained in Section 2000-based tracts increases the underserved Freddie Mac’s recent performance has been A.2 of this appendix, HUD will be defining area share of market originations by about slightly closer to the market. Over the past underserved areas based on 2000 Census data five percentage points. Between 1999 and three years (2001 to 2003), Freddie Mac beginning in 2005. The number of census 2003, 31.4 percent of home purchase performed at 91 percent of the market (24.1 tracts in metropolitan areas covered by mortgages (without B&C loans) were percent for Freddie Mac compared at 26.4 HUD’s definition will increase from 21,587 originated in underserved tracts based on percent for the market). (See Tables A.13 to tracts (based on 1990 Census) to 26,959 tracts 2000 geography, compared with 26.2 percent A.16 in Appendix A for complete data going (based on 2000 Census and new OMB based on 1990 geography—a differential of back to 1993.) metropolitan area specifications). The 5.2 percentage points. As also shown in Fannie Mae has funded underserved areas increase in the number of tracts defined as Table B.8, the underserved areas share of at a higher level than Freddie Mac, as underserved means that both GSE Fannie Mae’s purchases rises by 5.3 indicated above. And during 2001 and 2003, performance and the market estimates will be percentage points, and the underserved areas Fannie Mae average performance was only higher than reported above. This section share of Freddie Mac’s purchases rises by 5.2 slightly below the market. In 2003, the share provides an analysis of the performance of percentage points. Thus, the conclusions of Fannie Mae’s purchases going to the GSEs in the single-family-owner market reported above and in Appendix A about the underserved areas was 26.8 percent, based on 2000 census tract geography. For GSEs’ performance relative to the market compared with a market level of 27.6 the years 1999, 2000, 2001, and 2002, HUD about remain the same when the analysis is percent. Like Freddie Mac, Fannie Mae’s used the apportionment technique to re- conducted based on 2000 Census geography.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00200 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63779
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00201 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.059
It is interesting to repeat the earlier 1990- geography. The following results are obtained based analysis of home purchase loans but for home purchase loans from Table B.8: this time based on the 2000 Census
Market Year Freddie Mac Fannie Mae (w/o B&C) (percent) (percent) (percent)
1999 ...... 25.6 25.3 30.2 2000 ...... 27.3 29.0 31.7 2001 ...... 27.3 29.8 30.7 2002 ...... 31.7 32.3 31.8 2003 ...... 29.0 32.0 32.5 1996–2003 (estimate) ...... 27.2 29.3 30.9 1999–2003 (average) ...... 28.3 30.0 31.4 2001–2003 (average) ...... 29.4 31.4 31.7
Between 1999 and 2003, 28.3 percent of the primary market by over two percentage previous peak performance of 32.3 percent in Freddie Mac’s purchases and 30.0 percent of points, based on historical data. This home 2003. To meet the 2008 subgoal, Freddie Mac Fannie Mae’s purchases financed properties purchase subgoal will encourage the GSEs to will have to improve by 2.6 percentage points in underserved neighborhoods, compared provide additional credit and capital to urban over its 2002–2003 average (unweighted) with 31.4 percent home purchase loans neighborhoods that historically have not performance of 30.4 percent, while Fannie originated in the conventional conforming been adequately served by the mortgage Mae will have to improve by 1.8 percentage market (excluding B&C loans). Thus, Freddie industry—but in the future may be the very points over its 2002–2003 average Mac performed at 90 percent of the market neighborhoods where the growing population performance of 32.2 percent. level, while Fannie Mae performed at 96 of immigrants and minorities choose to live. The subgoal applies only to the GSEs’ percent of the market level—both results As detailed in Section I.5 of this appendix, purchases in metropolitan areas because the similar to those reported above for there are four specific reasons for HMDA-based market benchmark is only underserved areas based on 1990 Census establishing this subgoal: (1) The GSEs have available for metropolitan areas. HMDA data geography. The 2000 Census data show that the expertise, resources, and ability to lead for non-metropolitan counties are not reliable the Fannie Mae has been much closer to the the single-family-owner market, which is enough to serve as a market benchmark. The market during the recent 2001–2003 period. their ‘‘bread and butter’’ business; (2) the Department is also setting home purchase The share of Fannie Mae’s purchases going GSEs have been lagging the primary market subgoals for the other two goals-qualifying to underserved areas was 31.4 during 2001– in underserved areas, not leading it; (3) the categories, as explained in Appendices A and 2003, which placed it close to the market GSEs can help reduce troublesome C. level of 31.7 percent. However, the 2000- neighborhood disparities in access to b. Characteristics of GSEs’ Purchases of based results show that, like Freddie Mac, mortgage credit; and (4) there are ample Mortgages on Properties in Metropolitan Fannie Mae’s longer-term performance (since opportunities for the GSEs to expand their Underserved Areas 1996) as well as its recent average purchases in low-income and high-minority performance (1999 to 2003) have consistently neighborhoods. Sections E.9 and G of Several characteristics of loans purchased been below market levels. (Note that the Appendix A provide additional information in 2003 by the GSEs in metropolitan 1996–2003 averages reported above are on the opportunities for an enhanced GSE underserved areas are presented in Table B.9. estimated by adding the following 2000- role in underserved area segment of the home As shown, borrowers in underserved areas Census versus 1990-Census differentials purchase market and on the ability of the are more likely than borrowers in served calculated for 1999–2003: 5.2 percentage GSEs to lead that market. areas to be first-time homebuyers, all female, points for Freddie Mac, 5.3 for Fannie Mae, As discussed above, underserved areas all male and younger than 40. And, as and 5.2 for the market.) accounted for an average of approximately expected, borrowers in underserved areas are Underserved Area Subgoal for Home 31.5 percent of home purchase loans more likely to have below-median income Purchase Loans. The Department is originated in the conventional conforming and to be members of minority groups. For establishing a subgoal of 32 percent for each market of metropolitan areas (computed over example, first-time homebuyers make up 6.7 GSE’s acquisitions of home purchase loans 1999–2003 or over 2001–2003). To reach the percent of the GSEs’ mortgage purchases in financing single-family-owner properties 34-percent subgoal for 2008, both GSEs will underserved areas and 4.2 percent of their located in the underserved census tracts of have to improve over their earlier peak business in served areas. In underserved metropolitan areas for 2005, with this performances—Freddie Mac by 2.3 areas, 53.7 percent of borrowers had incomes subgoal rising to 33 percent for 2006 and percentage points over its previous peak below the area median, compared with 36.4 2007, and to 34 percent in 2008. If the GSEs performance of 31.7 percent in 2002, and percent of borrowers in served areas. meet the 2008 subgoal, they will be leading Fannie Mae by 1.7 percentage points over its BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00202 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63781
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00203 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.060 63782 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Minorities’ share of the GSEs’ mortgage has a role in the accessibility of credit for lenders, and subprime lenders. Industry purchases in underserved areas (33.2 some people in nonmetropolitan areas (low representatives contacted by the Urban percent) was greater than two times their income, minority, and first-time Institute researchers assessed that the barriers share in served areas (13.9 percent). And the homebuyers). West North Central counties nonmetropolitan lenders faced were in the pattern was even more pronounced for (Minnesota, Missouri, South Dakota, Iowa, areas of availability of sales comparables, African Americans and Hispanics, who Kansas, Nebraska, and North Dakota) have technology, and the type and number of accounted for 23.1 percent of the GSEs’ much lower GSE activity than all other lenders in the area. They also believed that business in underserved areas, but only 7.0 geographic regions, suggesting that the 1995 for the GSEs’ market share to improve in percent of their purchases in served areas. definition of underservice does not capture underserved nonmetropolitan areas, the GSEs the specific characteristics of this region, Other similarities in Fannie Mae and would have to begin to build relationships leading to limited GSE activity. Freddie Mac purchases in served and with the community lenders and provide underserved areas include the following. The Additionally, The Urban Institute prepared a report for HUD that investigated the factors education/training on how to sell loans GSEs are slightly more likely to purchase directly to the GSEs rather than using refinance loans in served areas than in influencing GSE activity in nonmetropolitan 48 intermediaries. underserved areas; mortgage purchases with areas. The authors found that Fannie Mae loan-to-value ratios below 80 percent are and Freddie Mac have increased their a. Effects of 2000 Census Geography lending to nonmetropolitan areas since 1993; more likely to be in underserved than in In order to compare served and however, there are still weak areas in terms served areas; and seasoned mortgage of the percentage of affordable loans being underserved areas, either in terms of GSE purchases are more likely to be in offered.49 They also established that GSE performance or socioeconomic underserved than in served areas. underwriting criteria was not a major barrier characteristics, it is first necessary to update 3. GSE Mortgage Purchases in in nonmetropolitan areas. current geographic (county) designations, Nonmetropolitan Areas In nonmetropolitan areas, the financial which reflect 1990 census median income and minority population data, to reflect There are numerous studies that have market is often made up of locally owned newly available 2000 census data. Table B.10 evaluated the impact of the GSEs’ purchases community banks, manufactured home on metropolitan areas, but few address the shows the impact on 2000, 2001, and 2002 impact on nonmetropolitan areas; therefore, Development and Research, Volume 5, 2001, pp. GSE purchases. These are reported for total our understanding of the GSEs and the 219–264. GSE purchases and separately for Fannie Mae nonmetropolitan markets is very limited. 48 Jeanette Bradley, Noah Sawyer and Kenneth and Freddie Mac. As above, the results also A study of the GSE market share in Temkin, Factors Influencing GSE Service to Rural are shown separately for counties that change Areas, The Urban Institute, prepared for U.S. underserved counties 47 found that location classification and those that do not. This Department of Housing and Urban Development, analysis is limited to nonmetropolitan areas 2002. 47 Heather MacDonald, ‘‘Fannie Mae and Freddie 49 Affordable loans are defined as borrowers based on both the pre- and post-June, 2003 Mac in Nonmetropolitan Housing Markets: Does earning less than 80 percent the Area Median OMB metropolitan area designations. Space Matter?’’ Cityscape: A Journal of Policy Income. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00204 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63783
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00205 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.061
Applying 2000 census median income and for underserved (51%); 23.9 million for Unlike the underserved areas definition for minority population data results in a slight served (49%)) as shown in Table B.5. metropolitan areas, which is based on census drop in the proportion of counties that are Table B.10 shows that Fannie Mae’s tracts, the rural underserved areas definition classified as underserved. Out of a total of performance in 2002 (40.2 percent) was is based on counties. Rural lenders argued 2,493 counties, 1,514 (65.5 percent) are somewhat higher than Freddie Mac’s (36.3 that they identified mortgages by the counties underserved based on 1990 data, and 1,260 percent). This gap widens slightly (1.8 in which they were located rather than the (61.4 percent) based on 2000 data. This small percent) in applying 2000 census income and net change disguises a somewhat larger shift minority data and 2003 metropolitan area census tracts; and therefore, census tracts of counties, as about 11.2 percent of currently definitions. were not an operational concept in rural areas. Market data on trends in mortgage underserved counties are reclassified as b. Characteristics of GSEs’ Purchases of lending for metropolitan areas are provided served counties and 4.6 percent of currently Mortgages on Properties in Nonmetropolitan served counties are reclassified as Underserved Areas by HMDA; however, no comparable data underserved. source exists for rural mortgage markets. The Comparing underserved and served Nonmetropolitan mortgage purchases made absence of rural market data is a constraint up 12.6 percent of the GSEs’ total mortgage nonmetropolitan areas in Table B.10, it is for evaluating credit gaps in rural mortgage purchases in 2003. Mortgages in underserved apparent that underserved nonmetropolitan lending and for defining underserved areas. areas make up a larger percentage of counties made up 38.6 percent of the GSEs’ business in nonmetropolitan areas.50 One concern is whether the broad nonmetropolitan areas as a whole than do definition overlooks differences in borrower served nonmetropolitan areas, as shown by 50 characteristics in served and underserved the number of counties (1,260 for Underserved areas make up about 56 percent of underserved (61.4%); 792 for served the census tracts in nonmetropolitan areas and 47 counties that should be included. Table B.11 percent of the census tracts in metropolitan areas. (38.6%)). These relationships hold true also compares borrower and loan characteristics This is one reason why underserved areas comprise for the GSEs’ mortgage purchases in served for the number of households (9.5 million for a larger portion of the GSEs’ single-family underserved (50.5%); 9.3 million for served mortgages in nonmetropolitan areas (39 percent) and underserved areas. (49.5%)), and the population (24.9 million than in metropolitan areas (23 percent). BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00206 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63785
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00207 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.062 63786 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Fannie Mae is slightly more likely and goals. Based on this economic analysis and 9. Applying the current criteria for Freddie Mac is less likely to purchase loans reviewed by the Office of Federal Housing identifying underserved areas to tracts would for first-time homebuyers in underserved Enterprise Oversight, HUD concludes that the result in reclassifying approximately 23 areas than in served areas. Mortgages to first- goals raise minimal, if any, safety and percent of all tracts, with 28 percent of tracts time homebuyers accounted for 4.3 percent soundness concerns. in served counties being redesignated as of Fannie Mae’s mortgage purchases in H. Defining Nonmetropolitan Underserved underserved and 19 percent of tracts in served counties, compared with 4.6 percent Areas underserved counties being redesignated as of its purchases in underserved counties. For 1. Whether to Adopt a Tract-Based Definition served. Overall, roughly the same percentage Freddie Mac the corresponding figures are of families (and population) would be 3.4 percent in served counties and 3.3 of Underserved Areas reclassified. However, because underserved percent in underserved counties. The current county-based definition for tracts are somewhat less densely populated The GSEs are more likely to purchase targeting GSE purchases to underserved than served tracts, the corresponding mortgages for high-income borrowers in nonmetropolitan areas was adopted in 1995 proportions of families that shift from served underserved than in served counties. over alternative narrower definitions, such as Surprisingly, borrowers in served counties census tracts, despite the use of census tracts and underserved counties are closer: 25 vs. were more likely to have incomes below the in metropolitan areas. In the 1995 Final Rule, 21 percent. median than in underserved counties (39.6 HUD found the merits of a county-based a. Do Census Tracts Allow a Sharper percent compared to 35.4 percent). These system of targeting outweighed a tract-based Delineation of Served and Underserved findings lend some support to the claim that, system. Now, with seven years of experience Areas? in rural underserved counties, the GSEs under a county-based system, the release of This section compares the differences in purchase mortgages for borrowers that Census 2000 data, and improvements in housing need and economic, demographic, probably encounter few obstacles in information technology and systems, HUD obtaining mortgage credit. can reexamine whether to switch to census and housing conditions in served and The following similarities in Fannie Mae tracts for defining underserved underserved nonmetropolitan areas classified and Freddie Mac purchases in served and nonmetropolitan areas. This section on, respectively, counties and tracts. underserved counties in nonmetropolitan compares impacts of the potential shift in Additionally, the ‘‘efficiency’’ with which areas mirror those found for the GSEs in definition for both served and underserved counties and tracts cover the target served and underserved census tracts in populations as determined by tract-based and populations is compared. That is, does tract- metropolitan areas. The GSEs are slightly county-based definitions using a number of based targeting do a better job of capturing more likely to purchase refinance loans in common industry variables as focal points for lower income households and excluding served than in underserved counties; analysis. higher income households than county-based mortgage purchases with loan-to-value ratios The rationale for choosing counties in 1995 targeting? below 80 percent are more likely to be in rested primarily on perceived shortcomings Table B.12 presents several indicators of underserved than in served counties; and of census tracts.51 In particular, rural lenders socioeconomic and housing condition in seasoned mortgage purchases are more likely did not perceive their market areas in terms served and underserved areas under both a to be in underserved than in served counties. of census tracts, but rather, in terms of tract-based and a county-based definition. In counties. Another concern was a perceived E. Factor 4: Size of the Conventional addition, served and underserved counties lack of reliability in geocoding 1990 census are subdivided into their served and Conforming Mortgage Market for tracts. At the same time, HUD found merit in underserved tract components. This allows a Underserved Areas using a tract-based geography for closer examination of the population and HUD estimates that underserved areas nonmetropolitan areas. Because tracts housing characteristics of the tracts that are account for 35–39 percent of the encompass more homogeneous populations reclassified (i.e., served to underserved or conventional conforming mortgage market. than counties, they permit more precise The analysis underlying this estimate is targeting of underserved populations. In visa versa) under tract-based targeting. Thus, detailed in Appendix D. other words, more homogeneous geographic area characteristics of housing need and areas increase the potential for targeting the housing, economic, and demographic F. Factor 5: Ability to Lead the Industry GSE mortgage purchases into areas where conditions can be compared, for the This factor is the same as the fifth factor borrowers are more likely to face obstacles following four groups of tracts: (1) Tracts in considered under the goal for mortgage and other challenges in securing mortgage served counties that would remain ‘‘served’’ purchases on housing for low- and moderate- credit. classified as tracts; (2) tracts that remain income families. Accordingly, see Section G The criteria used for this analysis include ‘‘underserved’’; (3) tracts that shift from of Appendix A for a discussion of this factor, the following: served to underserved; and (4) tracts that as well as Section I.5 of this Appendix, 7. Do tracts provide a sharper delineation shift from underserved to served. In addition, which describes the home purchase subgoal of served and underserved areas? we provide counts of tracts falling into each which is designed to place the GSEs in a Specifically, are underserved of these groups. If a tract-based classification leadership role in the underserved market. nonmetropolitan populations more clearly of underserved areas improves geographic differentiated by adopting tracts vs. counties? targeting, the regrouping of tracts would be G. Factor 6: Need to Maintain the Sound Could service to the underserved more similar to one another than to the other Financial Condition of the Enterprises nonmetropolitan populations be more tracts in their respective counties: e.g., HUD has undertaken a separate, detailed comprehensive under tract-based definitions? formerly underserved areas that become economic analysis of this rule, which 8. What is the impact on GSE purchasing served should be more similar to tracts that includes consideration of (a) the financial patterns if underserved areas are defined by returns that the GSEs earn on loans in tract? were and remain served than to underserved underserved areas and (b) the financial safety (unchanged). and soundness implications of the housing 51 60 FR 61,925–58 (1995) (Appendix B). BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00208 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63787
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00209 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.063 63788 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Socioeconomic and Demographic indicators, measured need increases in who are not in need living in an underserved Conditions. Table B.12 shows that in underserved areas and the gap between area; a coverage efficiency index equal to ¥1 important socioeconomic and demographic served and underserved areas widens when implies that none of the families in need live characteristics, tract-based targeting would tracts are used to classify areas. Most notably, in an underserved area, or equivalently, all more effectively distinguish underserved the percent of owner-occupied housing units families in underserved areas are not in need. populations. Median family income, poverty, switches from being higher in underserved Comparing coverage efficiency for counties unemployment, school dropout rates, and than served counties to being significantly and tracts indicates that tracts do a better job; minority population all exhibit greater lower among underserved tracts. With a shift capturing a higher percentage of differences between served and underserved to tracts overall ownership drops in nonmetropolitan families whose income falls areas using tracts. For example, the underserved areas, from 74 to 72 percent, and below the applicable income threshold and difference in median income between served increases in served areas from 74 to 77 excluding more families whose income and underserved counties is $9,579, or percent. In contrast, the homeownership rate exceeds the threshold.52 Overall, the alternatively, between served and for tracts located in served counties that efficiency index rises from 0.22 to 0.274. underserved tracts, the difference is $12,744. would be deemed underserved if judged Given income thresholds that are not far Similarly, there is a 7-percentage point gap separately is only 65 percent. In fact, this rate away from median income in most places in poverty rates (7.5 vs. 14.5 percent poverty) is much lower even than underserved tracts and the degree of income variation even with using counties, which widens to 8.6 in underserved counties. Shifting these tracts census tract boundaries, it should not come percentage points (6.6 vs. 15.3 percent) using from served to underserved largely accounts as a great surprise that neither the levels of tracts. Minority population also is captured for the switching of homeownership rates. coverage efficiency (0.22–0.27) nor somewhat better with tracts, with the served/ Results for other indicators of housing improvement produced in applying tracts (5 underserved gap increasing from 16.5 to 17.3 need and conditions are less clear-cut. No percentage points) are not more dramatic. percentage points. In all cases, the levels of definitive patterns are apparent for two, Nevertheless, tracts do produce better the indicators for underserved areas move in admittedly weak, measures of housing tracking of lower income, very low income, a direction consistent with targeting lower quality—units with complete plumbing and income households and areas with higher units with complete kitchen facilities, as well and minority families. minority populations. as for crowding. Purchase affordability, as b. Does GSE Performance Vary Between The 4-way breakdown of served and measured by the ratio of median housing Served and Underserved Tracts Within underserved counties reveals some value to the income necessary to qualify for Underserved Counties? significant differences between the two a loan for the median valued unit, is higher A similar analytical approach is used to component groups. In most respects, in underserved areas than in served areas. examine how a shift to tracts would impact ‘‘underserved tracts’’ (i.e., those meeting the However, the measure of purchase GSE purchases. Having applied income and underserved criteria), whether located in an affordability presented here is influenced by minority thresholds from the 2000 census underserved or served county, are more alike many market and other economic factors, and updating census tract geography, Table than they are like served tracts. Using median some of which do not relate to housing need. B.13 compares, respectively, 2000, 2001, and income again to illustrate, the effect of For example, a low affordability ratio may 2002 GSE purchases for served and reclassifying areas by tract characteristics is reflect abundant supply, but it may also underserved counties and tracts and also for to put together two groups of underserved reflect low demand stemming from, e.g., the served and underserved tracts within tracts: Tracts that were in previously limited availability of credit or high interest county boundaries. On net there would be underserved counties and are not reclassified rates. somewhat more tracts classified as and tracts that were in served counties but Coverage Efficiency. The coverage underserved under a tract-based system than meet the underserved criteria. A new group efficiency index measures the effect of of served tracts is similarly formed. In both adopting tract-based targeting. This index can currently: 6,782 vs. 6,414. As noted above, cases, the difference in median incomes of be used to indicate how well underserved however, 23.1 percent of all tracts are the constituent groups is about $3,500. In areas encompass populations deemed to be reclassified. Moving to tracts also would have contrast, the served and underserved underserved (‘‘sensitivity’’) and to exclude a significant effect on the relative counties now encompass ‘‘served’’ and populations that are deemed to be served performance of the GSEs. In 2002, Fannie ‘‘underserved’’ groups of tracts whose (‘‘specificity’’). The index is computed for Mae’s performance would drop 2.1 respective median incomes differ by almost median income as the difference in two percentage points to 35.4 percent, while $11,000. Combined with the fact that a fairly percentages: (1) The proportion of all families Freddie Mac’s performance would increase large number of tracts are affected overall in nonmetropolitan areas that meet the by 0.9 percent to 32.7 percent. (i.e., switch), these results support an applicable income threshold who live in BILLING CODE 4210–27–P assessment that counties are relatively crude underserved tracts minus (2) the proportion for targeting underserved populations. of all families in nonmetropolitan areas that 52 In areas with 30 percent or greater minority Housing Needs and Conditions. Table B.12 do not meet the applicable underserved population, all families with income in excess of shows that tract-based targeting would income threshold who live in underserved 120 percent of the greater of State or national produce modest gains in focusing GSE areas. This difference can range from 1 median income are counted as qualifying as ‘‘in purchases on areas with relatively greater (perfect) to ¥1 (bad; perverse). For example, need’’ for these computations. Similarly, in areas housing needs and conditions as measured a coverage efficiency index equal to 1 implies with less than 30 percent minority, those minority by low owner-occupancy, higher vacancy that every family in need is living in an (headed) families with income between 95 and 120 rates, and crowding. For each of these underserved area while there are no families percent of the applicable median income are not classified as ‘‘in need.’’
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00210 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63789
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00211 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.064 63790 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Differences between qualifying purchases income threshold from 95 percent to 90 • Housing affordability would decline in of single-family and multifamily loans are percent to 80 percent. underserved areas, becoming nearly equal further increased when assessed at the tract • 30/120/95 vs. 30/110/95 vs. 30/110/80— with affordability in served areas at 80 level. Performance for single-family loans to examine the effect of lowering both the percent. drops 0.7 percentage points to 35.2, but for minority (from 120% to 110%) and general • Crowding would be higher in multifamily increases by 2.5 percentage income (from 95% to 80%) thresholds; and underserved areas, absolutely and relative to points to 46.8. These changes dramatically • 30/120/95 vs. 50/120/95—to examine the served areas. Thus, more narrowly defined compound the results observed in updating effect of increasing the minority population underserved areas would more strongly to 2000 census data, resulting in a widening threshold that must be attained before manifest conditions and needs associated of the single- and multifamily performance applying the minority income threshold. with underservice: lower income, higher difference from the current level of 7.0 For each alternative, indicators of poverty, higher minority populations, lower percentage points to 11.6 percentage points. socioeconomic and housing conditions are homeownership, lower affordability, more calculated for served and underserved areas 2. Alternative Definitions of Underservice crowding, etc. However, served areas would for each alternative and compare the results expand to encompass significant numbers of The current definition of underservice in to the current definition. Of particular these same underserved and target nonmetropolitan areas was established in interest is whether certain thresholds of populations. 1995 to be relatively broad, encompassing minority population and median income Use of the coverage efficiency index nearly twice as many underserved as served capture the differences in housing needs and highlights one of the tradeoffs between using counties and somewhat more than half of the conditions between served and underserved a low median income threshold versus a high total nonmetropolitan population. This was areas better than others. The ‘‘coverage median income threshold in redefining done primarily to ensure that certain areas efficiency’’ of each alternative relative to underservice. Coverage efficiency based on with low incomes and/or high minority households below the poverty line, below 50, all variables examined, including populations, which might not be considered 70, and 95 percent of area reference income, ‘‘underserved,’’ poor, very low income, low underserved in comparison to the rest of and below the alternative income level(s) income and even moderate income families, their State, would nevertheless be identified used to define underservice, is also declines sharply as the income threshold is as underserved from a national perspective. presented. GSE purchasing activity is also lowered from 95 to 80 percent, becoming This section summarizes a new analysis, examined for each alternative definition, negative for most groups. Coverage for the based on 2000 census data, to evaluate the specifically, the percentage of eligible loans ‘‘underserved’’ cohort declines from 22.0 to extent to which the current definition focuses that qualify towards the goal for underserved ¥1.0 percent, and for families with up to 95 GSE purchasing activity toward stimulating areas defined by different thresholds. Each percent of reference income, it declines from mortgage lending in areas with populations analysis is conducted both with counties and 17.2 to ¥10.0 percent. These changes result having greatest housing need. Alternative tracts as the geographic unit. from losing almost half of the families in definitions of underservice are considered as County Results. The main effect of target income ranges without any appreciable follows: (1) Variations of the current lowering the general income threshold from gain in specificity, i.e., shrinking the thresholds; (2) applying only the State 95 to 90 to 80 percent of the reference income proportion of people living in underserved median income level for qualifying is to roughly halve the number of counties counties with incomes above the respective underserved counties and tracts; and (3) and population residing in underserved target levels. Similar patterns are observed establishing different thresholds in areas. Under the current definition, 11.6 for families with below 70 percent of micropolitan and ‘‘outside of core’’ million people reside in underserved areas as reference income, below 50 percent of nonmetropolitan areas. In each case the opposed to fewer than 10 million in served reference income, and families in poverty. objective is to assess how redesignating areas. With a general income threshold of 80 served and underserved areas would affect percent, 5.7 million would be left in The second set of comparisons builds on relative conditions and needs and GSE underserved areas. A 90 percent threshold the first set by lowering the income threshold purchasing performance. In distinguishing would produce a shift of approximately half applicable to areas with relatively high micropolitan and ‘‘outside of core’’ areas, it this amount. minority populations (30 percent) from 120 is of interest to determine whether it would In terms of social, economic, demographic, to 110 percent in addition to the general be appropriate to establish different and housing characteristics, lowering the threshold. This change further shrinks, albeit, thresholds for underservice. The overarching income threshold from 95 to 80 percent only marginally, the size and population of criterion for evaluating and comparing would have the following notable underserved areas. Minority underserved definitions is their ability to serve very low- consequences: populations would be smaller and income, low-income and moderate-income • Minority population in underserved socioeconomic and housing conditions households, households in poverty, first-time areas would increase from 12.4 to 20.8 would be worse. Not surprisingly, coverage homebuyers, minorities, and households in percent with no significant change in served efficiencies and GSE purchase performance remote locations.53 areas. levels also would decline across the board, In the current definition, areas are • Median income would fall in both served although the marginal effects of reducing the classified as underserved if either the and underserved areas with the difference minority income threshold are quite small. minority population share is greater than 30 remaining nearly constant at $10,000. The 30/110/80 alternative is the narrowest percent and median income is less than 120 • Poverty, unemployment, school drop out definition examined and produces the biggest percent of the greater of State rates all would be higher in both served and losses in efficiency and GSE performance. nonmetropolitan or national nonmetropolitan underserved areas. The gap would increase The third variation of the current median income; or area median income is for each of these characteristics. definition is an increase in the minority less than or equal to 95 percent of the greater • Migration into underserved areas (from population threshold from 30 to 50 percent. of State nonmetropolitan or national other States) would be relatively lower than Thus, if an area does not qualify as nonmetropolitan median income. The greater into served areas with an 80 percent income underserved against the general income of State nonmetropolitan or national median threshold. threshold of 95 percent it could still qualify income is termed the ‘‘reference income.’’ • Indicators of homeownership would if its population is 50 percent minority and Denoting the current thresholds as ‘‘30/120/ decline somewhat in underserved areas median income is less than or equal to 120 95,’’ the following set of alternative relative to served areas. For all units, for percent of the reference income level. thresholds are evaluated: example, ownership would decline from 74.3 Relatively few counties qualify solely • 30/120/95 vs. 30/120/90 vs. 30/120/80— to 72.9 percent in underserved areas and under the current minority thresholds. to examine the effect of lowering the general increase from 73.5 to 74.3 percent in served Raising the population threshold would trim areas. this number by an additional 73 counties • (457 tracts). Not surprisingly, the percent 53 A more comprehensive presentation of this Median housing values would fall in analysis may be found in Economic Systems, Inc., both served and underserved areas with a minority in underserved areas would Indicators of Mortgage Market Underservice in Non- significant narrowing in the gap from decrease. However, the areas being Metropolitan Areas, Interim Report to HUD, March approximately $25,000 to $19,000 at an 80 redesignated as served are apparently 2003, Chapter 6. percent median income threshold. somewhat above average in terms of
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00212 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63791
socioeconomic and housing conditions in Census, the 2000 Census includes a larger tracts that are under 30 percent minority.56 underserved areas and below-average in number of census tracts that meet HUD’s Second, census tracts with lower incomes terms of conditions in served areas. Coverage definition of underserved area. The new 37 have higher denial rates and lower efficiencies for all cohorts would be lower percent-39 percent goals are commensurate origination rates than higher income tracts. than for the current definition of with recent market share estimates of 37–39 Tracts with income less than 90 percent of underservice and GSE performance overall percent for 1999–2002, presented in area median income have over twice the would be approximately 90 percent of the Appendix D. denial rate and almost half of the origination current level. In addition, an Underserved Areas Housing rate of tracts with income over 120 percent Using the State median income, alone, as Subgoal of 32 percent is established for the of area median income. the general reference income would reduce GSEs’ acquisitions of single-family-owner In both the 1995 and the 2000 GSE Rules, the number underserved counties relative to home purchase loans in metropolitan areas in HUD’s research determined that the current definition, and, although there 2005, with the subgoal rising to 33 percent ‘‘underserved areas’’ could best be would still be more underserved counties in 2006 and in 2007 and 34 percent in 2008. characterized in metropolitan areas as census (1,274 vs. 1,064), the underserved population The subgoal is designed to encourage the tracts where: (1) Median income of families actually would become smaller than the in the tract does not exceed 90 percent of GSEs to lead the primary market in providing served population. The effect of this area (MSA) median income or (2) minorities mortgage credit in underserved areas. alternative on differences in housing comprise 30 percent or more of the residents This section summarizes the Secretary’s conditions and needs between served and and median income of families in the tract underserved areas is generally small and consideration of the six statutory factors that does not exceed 120 percent of area median ambiguous, but overall, results in less led to the Underserved Area Housing Goal income. The earlier analysis was based on contrast. Consistent with the results for other and the subgoal for home purchase loans in 1990 Census data. HUD has now conducted alternatives, applying a State median income metropolitan areas. This section discusses the same analysis using 2000 Census data standard, alone, would result in lower the Secretary’s rationale for defining and has determined that the above definition coverage efficiency across all target groups. underserved areas and it compares the continues to be a good proxy for underserved Census Tract Results. As discussed above, characteristics of such areas and untargeted areas in metropolitan areas. The income and the adoption of a tract-based system would areas. The section draws heavily from earlier minority cutoffs produce sharp differentials result in greater coverage efficiency of sections which have reported findings from in denial and origination rates between underserved populations and sharper HUD’s analyses of mortgage credit needs as underserved areas and adequately served distinctions in the socioeconomic, well as findings from other research studies areas. For example, in 2003 the mortgage demographic and housing characteristics of investigating access to mortgage credit. denial rate in underserved areas (15.9 served and underserved areas. That is, tracts 1. Housing and Credit Disparities in percent) was over one and three-fourths more effectively carve out areas that exhibit Metropolitan Areas times that in adequately served areas (8.9 characteristics that are associated with percent). underservice, such as low income, large To identify areas underserved by the These minority population and income minority populations and low mortgage market, HUD focused on two thresholds apply in the suburbs as well as in homeownership. The converse is true for traditional measures used in a number of central cities. The average denial rate in served areas. In analysis at the tract level, studies based on HMDA data: Application underserved suburban areas (14.8 percent) is these patterns tend to be maintained quite denial rates and mortgage origination rates 1.7 times that in the remaining served areas consistently. A tract-based system would per 100 owner-occupied units. Tables B.2 of the suburbs (8.7 percent), and is almost as improve the power to differentiate and B.3 in Section B of this Appendix large as the average denial rate (16.8 percent) underserved and served populations. presented detailed data on denial and in underserved central city tracts. Low- According to virtually every indicator of origination rates by the racial composition income and high-minority suburban tracts socioeconomic, demographic, and housing and median income of census tracts for appear to have credit problems similar to conditions, applying State median income, metropolitan areas. Aggregating this data is their central city counterparts. Thus HUD alone, with a tract-based geography would useful in order to examine denial and uses the same definition of underserved areas produce superior differentiation to the origination rates for broader groupings of throughout metropolitan areas—there is no current county-based definition. In terms of census tracts: 55 need to define such areas differently in coverage efficiency, we again see central cities and in the suburbs. improvement with tracts, but not enough to Minority composition Denial Orig. This definition of metropolitan offset the loss of eliminating the national (percent) rate rate underserved areas based on 2000 Census median income threshold. For the (percent) geography includes 26,316 of the 51,040 underserved population, for example, census tracts in metropolitan areas, covering coverage efficiency would be 16.9 percent 0–30 ...... 9.6 26.7 49.2 percent of the metropolitan population with tracts, still below 22 percent under the 30–50 ...... 12.4 26.9 in 2000. (By contrast, the 1990-based current definition.54 50–100 ...... 17.2 20.8 definition included 21,587 of the 45,406 census tracts in metropolitan areas, covering I. Determination of the Underserved Areas 44.3 percent of the metropolitan population Housing Goal Denial Tract income rate Orig. in 1990.) The 2000-based definition includes The annual goal for each GSE’s purchases (percent) rate 75.7 percent of the population living in of mortgages financing housing for properties poverty in metropolitan areas. The located in geographically targeted areas Less than 90% of AMI .... 16.9 18.1 unemployment rate in underserved areas is (central cities, rural areas, and other 90–120% ...... 11.3 25.4 more than twice that in served areas, and underserved areas) is 37 percent of eligible Greater than 120% ...... 7.8 32.7 owner units comprise only 51.6 percent of units financed in 2005, 38 percent in 2006 total dwelling units in underserved tracts, and 2007, and 39 percent in 2008. The 2008 versus 75.9 percent of total units in served Two points stand out. First, high-minority goal will remain in effect in subsequent tracts. As shown in Table B.14, this census tracts have higher denial rates and years, unless changed by the Secretary prior definition covers most of the population in lower origination rates than low-minority to that time. The goal of 37 percent for 2005 several distressed central cities including tracts. Specifically, tracts that are over 50 is larger than the goal of 31 percent for 2001– Bridgeport (100 percent), Newark (99 03 mainly because, compared with the 1990 percent minority have nearly twice the denial rate and two-thirds the origination rate of 56 The differentials in denial rates are due, in part, 54 Note that, unlike the other panels in tables 6.3 to differing risk characteristics of the prospective and 6.8, ‘‘underserved population’’ is defined 55 Denial rates are computed for mortgage borrowers in different areas. However, use of denial according to the applicable definition. Thus, applications without manufactured housing loans. rates is supported by the findings in the Boston Fed eliminating the national median income test, Origination rates equal home purchase and study which found that denial rate differentials narrows the defined cohort of underserved families. refinance mortgages (without subprime loans) per persist, even after controlling for risk of the Despite this, coverage falls. 100 owner occupants in a census tract. borrower. See Section B for a review of that study.
VerDate jul<14>2003 19:33 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00213 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63792 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
percent), and Detroit (93 percent). The underserved areas: New York (68 percent), percent), Houston (73 percent), and Phoenix nation’s five largest cities also contain large Los Angeles (72 percent), Chicago (75 (50 percent). concentrations of their population in BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00214 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.065
2. Identifying Underserved Portions of greater housing needs and conditions as in areas with a minority population of 30 Nonmetropolitan Areas measured by owner-occupancy, vacancy percent or more, families with incomes Based on an exploration of alternative rates, and crowding. For each of these below 120 percent of the reference income) numerical criteria for identifying indicators, measured need increases in the use of more refined tract geography underserved nonmetropolitan areas using underserved areas and the gap between results in a 5 percentage point increase in the 2000 census data, HUD has concluded that served and underserved areas widens when coverage efficiency index, from 22 to 27 tracts are used to classify areas. Most notably, the current definition of underservice is percent. This reflects two improvements homeownership would be significantly lower broad but efficacious and that any narrower under a tract system: Underserved areas in underserved areas relative to served areas definition of underservice would not serve would capture more of the nonmetropolitan congressional intent under FHEFSSA. under a tract-based system. Currently, and contrary to expectations, homeownership ‘‘underserved’’ families (62 vs. 65 percent) Narrowing the definition of underservice and fewer ‘‘served’’ families (decreasing from potentially could promote more intense actually is slightly greater in underserved 40 to 37 percent of families in underserved purchasing in needier communities, but this areas. Driving this reversal is the fact that areas). seems unlikely. On the contrary, the greatest tracts in served counties that would be marginal impact of GSE purchasing could be reclassified as underserved tracts have an 3. Past Performance of the GSEs in the very areas that would be excluded ownership rate of just 65 percent, which is much lower even than in the underserved Goals Performance. In the October 2000 under the alternatives. rule, the underserved areas goal was set at 31 Research comparing a tract-based system tracts in underserved counties, where percent for 2001–03. Effective on January 1, for defining underserved areas with the ownership is 73 percent. Meanwhile, the current county-based system, using 2000 served tracts in served and underserved 2001, several changes in counting census data, indicates that a tract-based counties have the same ownership rate of 77 requirements came into effect for the system would result in more effective percent, which is significantly higher than in undeserved areas goal, as follows: (a) ‘‘Bonus geographic targeting of GSE purchases. underserved areas. points’’ (double credit) for purchases of Although the total number of tracts Two groups of measures of housing mortgages on small (5–50 unit) multifamily designated as served and underserved areas conditions—housing quality and properties and, above a threshold level, would change very little, 23 percent of all affordability—exhibit less clear-cut results mortgages on 2–4 unit owner-occupied tracts would be reclassified, reassigning from applying tracts. However, we conclude properties; (b) a ‘‘temporary adjustment approximately equal numbers of families that these results are consistent with the factor’’ (1.35 units credit) for Freddie Mac’s from served to underserved and from ambiguous patterns discussed in chapter 4 purchases of mortgages on large (more than underserved to served. above and do not undermine the overall 50 unit) multifamily properties; and (c) The main effect of the reclassification is to conclusion that basing geographic targeting eligibility for purchases of certain qualifying align tracts into more homogeneous and on tracts would more sharply define areas government-backed loans to receive goal with greater housing need and adverse distinct groups as measured by differences in credit. Under these counting rules, as shown key socioeconomic and demographic housing conditions. in Table B.7a and Figure B.2, Fannie Mae’s characteristics such as median family Not surprisingly, the results from analyzing income, poverty, unemployment, school housing, socioeconomic, and demographic performance was 32.6 percent in 2001, 32.4 dropouts, and minority population. As a characteristics are further reinforced in percent in 2002, and 32.1 percent in 2003, result of reclassification, underserved areas finding that a tract-based system would better while Freddie Mac’s performance was 31.7 stand out more as areas of lower income and capture underserved populations and percent in 2001, slightly less than 31 percent economic activity and somewhat larger exclude served populations from geographic in 2002, and 32.7 percent in 2003; thus minority populations. targeting. Defining underserved families as Fannie Mae surpassed the goal of 31 percent Tract-based targeting would potentially those in any area whose income was less in all thee years, while Freddie Mac fell focus GSE purchases in areas with relatively than 95 percent of the reference income (or slightly short of the goal in 2002.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00215 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63794 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C bonus points and the ‘‘temporary adjustment then Fannie Mae’s performance would have Counting requirements (a) and (b) expired factor’’—had been in effect in 2000–03, and been 31.0 percent in 2000, 30.4 percent in at the end of 2003, while (c) will remain in the GSEs’ had purchased the same mortgages 2001, 30.1 percent in 2002, and 29.2 percent effect. If this counting approach—without the that they actually did purchase in both years, in 2003. Freddie Mac’s performance would
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00216 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.066 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63795
have been 29.2 percent in 2000, 28.2 percent D.2 of this appendix compared the GSEs’ apportionment techniques to re-allocate in 2001, 28.0 percent in 2002, and 27.7 funding of home purchase loans in 1990-based GSE and HMDA data into census percent in 2003. Therefore, Fannie Mae underserved areas with originations by tracts as defined by the 2000 Census. The would have just matched the underserved lenders in primary market. To take advantage GSE and HMDA data for 2003 were already areas goal of 30 percent in 2000 and fallen of HMDA and GSE data going back to 1993, based on 2000 geography, so no short in 2001–03, while Freddie Mac would the analysis was conducted using 1990 apportionment was needed for that year. have fallen short of the goal in all four years, Census tract geography. While both GSEs Switching to the 2000-based tracts increases 2000–2003. have improved their performance since 1993, the underserved area share of market The above performance figures are for they have both lagged the conventional originations by 5.7 percentage points. underserved areas (census tracts in conforming market in providing affordable Between 1999 and 2002, 31.4 percent of metropolitan areas and counties in non- loans to underserved areas. The 1990-based mortgage originations (without B&C loans) metropolitan areas) defined in terms of 1990 analysis shows that the two GSEs have were originated in underserved tracts based Census geography. Switching to 2000 Census engaged in very different patterns of on 2000 geography, compared with 26.2 data increases the coverage of underserved funding—Freddie Mac has been much less areas, which increases the share of the GSEs’ likely than Fannie Mae to fund home loans percent based on 1990 geography. As shown purchases in underserved areas by in underserved neighborhoods. HUD will in Table B.8 of Section D.2, the underserved approximately 5 percentage points. Based on begin defining underserved areas based on areas share of each GSE’s purchases also rises 2000 Census geography, and excluding 2000 Census geography and new OMB by approximately five percentage points. counting requirements (a) and (b) then definitions of metropolitan areas in 2005, the Thus, conclusions about the GSEs’ Fannie Mae’s performance would have been first year of the rule. As noted above, the performance relative to the market are similar 38.1 percent in 2000, 36.6 percent in 2001, 2000-based definition of underserved areas whether the analysis is conducted in terms 35.9 percent in 2002, and 34.1 percent in includes 5,372 more census tracts in of 2000 Census geography or 1990 Census 2003, as shown in Table B.7b. Freddie Mac’s metropolitan areas than the 1990-based geography. performance would have been 35.1 percent definition, which means the GSE-market The analysis for home purchase loans in 2000, 33.5 percent in 2001, 33.3 percent comparisons need to be updated to based on 2000 Census geography will be in 2002, and 31.6 percent in 2003. incorporate tract designations from the 2000 summarized here (see Section D.2 of this Single-Family-Owner Home Purchase Census. Therefore, for the years 1999, 2000, appendix for a similar analysis using 1990- Mortgages. Sections E.9 of Appendix A and 2001, and 2002, HUD used various based geography):
Market Year Freddie Mac Fannie Mae (w/o B&C) (percent) (percent) (percent)
1999 ...... 25.6 25.3 30.2 2000 ...... 27.3 29.0 31.7 2001 ...... 27.3 29.8 30.7 2002 ...... 31.7 32.3 31.8 2003 ...... 29.0 32.0 32.5 1996–2003 (estimate) ...... 27.2 29.3 30.9 1999–2003 (average) ...... 28.3 30.0 31.4 2001–2003 (average) ...... 29.4 31.4 31.7
Between 1999 and 2003, 28.3 percent of 4. Ability To Lead the Single-Family-Owner points over its 2002–2003 average Freddie Mac’s purchases and 30.0 percent of Market: A Subgoal for Underserved Areas performance of 32.2 percent. Fannie Mae’s purchases financed properties The Secretary believes the GSEs can play The subgoal applies only to the GSEs’ in underserved neighborhoods, compared a leadership role in underserved markets. purchases in metropolitan areas because the with 31.4 percent home purchase loans Thus, as discussed in Section D.2, the HMDA-based market benchmark is only available for metropolitan areas. HMDA data originated in the conventional conforming Department is establishing a subgoal of 32 for nonmetropolitan counties are not reliable percent for each GSE’s acquisitions of home market (excluding B&C loans). Thus, Freddie enough to serve as a market benchmark. The purchase loans for single-family-owner Mac performed at 90 percent of the market Department is also setting home purchase level, while Fannie Mae performed at 96 properties located in the underserved census subgoals for the other two goals-qualifying percent of the market level—both results tracts of metropolitan areas in 2005, rising to categories, as explained in Appendices A and similar to those reported above for 33 percent in 2006 and 2007 and to 34 C. percent in 2008. If the GSEs meet the 2008 underserved areas based on 1990 Census The approach taken is for the GSEs to subgoal, they will be leading the primary geography. The 2000 Census data show that obtain their leadership position by staged market by over two percentage points. As increases in the underserved areas subgoal; Fannie Mae has been much closer to the discussed above, underserved areas market during the recent 2001–2003 period. this will enable the GSEs to take new accounted for an average of approximately initiatives in a correspondingly staged The share of Fannie Mae’s purchases going 31.5 percent of home purchase loans manner to achieve the new subgoal each to underserved areas was 31.4 percent during originated in the conventional conforming year. Thus, the increases in the underserved 2001–2003, which placed it closer to the market of metropolitan areas (computed over areas subgoal are sequenced so that the GSEs market level of 31.7 percent. However, the 1999–2003 or over 2001–2003). To reach the can gain experience as they improve and 2000-based results show that, like Freddie subgoal for 2008, both GSEs will have to move toward the new higher subgoal targets. Mac, Fannie Mae’s longer-term performance improve over their earlier peak Appendix A discusses in some detail the (since 1996) as well as its recent average performances—Freddie Mac by 2.3 factors that the Department considered when performance (1999 to 2003) have consistently percentage points over its previous peak setting the subgoal for low- and moderate- been below market levels. But, it is performance of 31.7 percent in 2002, and income loans. Several of the considerations Fannie Mae by 1.7 percentage points over its encouraging that Fannie Mae significantly were general in nature—for example, related previous peak performance of 32.3 percent in to the GSEs’ overall ability to lead the single- improved its performance relative to the 2003. To meet the 2008 subgoal, Freddie Mac family-owner market—while others were market during the first two years of HUD’s will have to improve by 2.6 percentage points specific to the low-mod subgoal. Because the higher housing goal levels. (See Section D.2 over its 2002–2003 average (unweighted) reader can refer to Appendix A, this for the method of estimating the 1996–2003 performance of 30.4 percent, while Fannie appendix provides a briefer discussion of the average results.) Mae will have to improve by 1.8 percentage more general factors. The specific
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00217 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63796 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
considerations that led to the subgoal for two-thirds of the growth in the number of reasonable or feasible, the Secretary has the underserved areas can be organized around new households over the next ten years. To authority to revise the goals. the following four topics: meet the diverse and unique needs of these 6. The Underserved Areas Housing Goal for (1) The GSEs have the ability to lead the families, the GSEs must continue adjusting 2005–2008 market. As discussed in Appendix A, the their underwriting guidelines and offering GSEs have the ability to lead the primary new products so that they can better serve The Underserved Areas Housing Goal for market for single-family-owner loans, which these areas and hopefully attract more 2005 is 37 percent of eligible purchases, is their core business. Both GSEs have been mainstream lenders into our inner city rising to 38 percent in 2006 and 2007 and 39 dominant players in the home purchase neighborhoods. percent in 2008. Five percent of the six market for years, funding 57 percent of the (4) There are ample opportunities for the percentage point increase in 2005 simply single-family-owner mortgages financed GSEs to improve their performance. reflects the expanded coverage of HUD’s between 1999 and 2002. Through their many Mortgages are available for the GSEs to definition in the 2000 Census tract data. The new product offerings and their various purchase in underserved areas. They can bonus points for small multifamily properties partnership initiatives, the GSEs have shown improve their performance and lead the and owner-occupied 2–4 units, as well as that they have the capacity to operate in primary market in purchasing loans in these Freddie Mac’s Temporary Adjustment Factor, underserved neighborhoods. They also have low-income and high-minority will no longer be in effect for goal counting the staff expertise and financial resources to neighborhoods. The underserved areas share purposes. It is recognized that neither GSE make the extra effort to lead the primary of the home purchase market has consistently would have met the 37 percent target for market in funding single-family-owner been around 31 percent since 1995 (and 32 2005 in the past three years, and only Fannie mortgages in underserved areas. percent in the last two years), which suggests Mae would have met this goal in 2000. (2) The GSEs have lagged the market. Even a degree of underlying strength in the market. Specifically, Fannie Mae’s performance is though they have the ability to lead the According to the market share data reported projected to have been 37.5 percent in 2000, market, they have not done so, as discussed in Table A.30 of Appendix A, the GSEs have 35.7 percent in 2001, 35.0 percent in 2002, above. Fannie Mae demonstrated the type of been purchasing 48 percent of new and 34.1 percent in 2003 under a 2000-based improvement needed to meet this new originations in underserved areas, which underserved area goal. On this basis, Freddie underserved area subgoal during 2001 and means there are plenty of purchase Mac’s performance is projected to have been 2002. During 2001, underserved area loans opportunities left for them in the non-GSE declined as a percentage of primary market portion of that market. In addition, the GSEs’ 34.1 percent in 2000, 32.5 percent in 2001, originations (from 31.7 to 30.7 percent), but purchases under the subgoal are not limited 32.4 percent in 2002, and 31.7 percent in they increased as a percentage of Fannie to new mortgages that are originated in the 2003. However, GSE goal performance in Mae’s purchases (from 29.0 to 29.8 percent); current calendar year. The GSEs can 2001–03 was reduced by the heavy refinance and during 2002, they increased further as a purchase loans from the substantial, existing wave of this period. percentage of Fannie Mae’s purchases (from stock of affordable loans held in lenders’ The objective of HUD’s Underserved Areas 29.8 to 32.3 percent), placing Fannie Mae at portfolios, after these loans have seasoned Housing Goal is to bring the GSEs’ the market level. and the GSEs have had the opportunity to performance to the upper end of HUD’s (3) There are disparities among observe their track record. In fact, both GSEs market range estimate for this goal (35–39 neighborhoods in access to mortgage credit. have often purchased seasoned loans that percent), consistent with the statutory There remain troublesome neighborhood were used to finance properties in criterion that HUD should consider the GSEs’ disparities in our mortgage markets, even underserved areas (see Table A.11 in ability to lead the market for each Goal. To after the substantial growth in conventional Appendix A). enable the GSEs to achieve this leadership, lending to low-income and minority To summarize, although single-family- the Department is modestly increasing the neighborhoods that accompanied the so- owner mortgages comprise the ‘‘bread-and- Underserved Areas Housing Goal for 2005 called ‘‘revolution in affordable lending’’. butter’’ of their business, the GSEs have which will increase further through 2008, to There is growing evidence that inner city lagged behind the primary market in achieve the ultimate objective for the GSEs to neighborhoods are not being adequately financing properties in underserved areas. lead the market under a range of foreseeable served by mainstream lenders. Some have For the reasons given above, the Secretary economic circumstances by 2008. Such a concluded that a dual mortgage market has believes that the GSEs can do more to raise program of staged increases is consistent developed in our nation’s financing system, the share of their home loan purchases in with the statutory requirement that HUD with conventional mainstream lenders underserved areas. This can be accomplished consider the past performance of the GSEs in serving white families living in the suburbs by building on efforts that the enterprises setting the Goals. Staged increases in the and FHA and subprime lenders serving have already started, including their new Underserved Areas Housing Goal will minority families concentrated in inner city affordable lending products, their many provide the enterprises with opportunity to neighborhoods.57 In addition to the partnership efforts, their outreach to inner adjust their business models and prudently unavailability of mainstream lenders, city neighborhoods, their incorporation of try out business strategies, so as to meet the families living in these often highly- greater flexibility into their underwriting required 2008 level without compromising segregated neighborhoods face many guidelines, and their purchases of CRA loans. other business objectives and requirements. additional hurdles, such as lack of cash for A wide variety of quantitative and qualitative The analysis of this section implies that a down payment, credit problems, and indicators indicate that the GSEs have the there are many opportunities for Fannie Mae discrimination. Immigrants and minorities, resources and financial strength to improve and Freddie Mac to improve their overall who disproportionately live in underserved their affordable lending performance enough performance on the Underserved Areas areas, are projected to account for almost to lead the market in underserved areas. Housing Goal. The GSEs provided financing 5. Size of the Mortgage Market for for 55 percent of the single-family and 57 See Dan Immergluck, Star Differences: The Underserved Areas multifamily units that were financed in the Explosion of the Subprime Industry and Racial As detailed in Appendix D, the market for conventional conforming market between Hypersegmentation in Home Equity Lending, 1999 and 2002. However, in the underserved Woodstock Institute, October 2000; and Daniel mortgages in underserved areas is projected Immergluck and Marti Wiles, Two Steps Back: The to account for 35–39 percent of dwelling areas portion of the market, the GSE Dual Mortgage Market, Predatory Lending, and the units financed by conventional conforming purchases represented only 48 percent of the Undoing of Community Development, Woodstock mortgages; in estimating the size of the dwelling units that were financed in the Institute, Chicago, IL, November 1999. For a market, HUD used alternative assumptions market. Thus, there appears to be ample national analyses, see the HUD report Unequal about future economic and market conditions room for the GSEs to increase their purchases Burden: Income and Racial Disparities in Subprime that were less favorable than those that of loans that qualify for the Underserved Lending in America, April 2000; and Randall M. Areas Housing Goal. In addition, there are Scheessele, Black and White Disparities in existed over the last five years. HUD is well Subprime Mortgage Refinance Lending, Housing aware of the volatility of mortgage markets several market segments that would benefit Finance Working Paper No. HF–114, Office of and the possible impacts on the GSEs’ ability from a greater secondary market role by the Policy Development and Research, U.S. Department to meet the housing goals. Should conditions GSEs, and many of these market segments are of Housing and Urban Development, April 2002. change such that the goals are no longer concentrated in underserved areas.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00218 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63797
7. Conclusions Multifamily Subgoal. HUD has established Goals for each of the seven years during this a special affordable subgoal for GSE period. Specifically: Having considered the projected mortgage • market serving low- and moderate-income purchases of multifamily mortgages. This The goal was set at 12 percent for 1996; families, economic, housing and subgoal is expressed in terms of a minimum Fannie Mae’s performance was 15.4 percent demographic conditions for 2005–08, and the annual dollar volume of multifamily and Freddie Mac’s performance was 14.0 mortgage purchases for units qualifying for percent. GSEs’ recent performance in purchasing • mortgages in underserved areas the Secretary the goal, rather than as a percentage of total The goal was set at 14 percent for 1997– has determined that the annual goal of 37 units financed, as for the three housing goals. 2000. Freddie Mac’s performance was 15.2 percent of eligible units financed in 2005, 38 Both GSEs have consistently surpassed the percent in 1997, 15.9 percent in 1998, 17.2 percent in 2006 and 2007, and 39 percent in multifamily subgoal since its establishment percent in 1999, and 20.7 percent in 2000; 2008 is feasible. The Secretary has also in 1996. The rule increases the subgoal such and Fannie Mae’s performance was 17.0 established a subgoal of 32 percent for the that, of the total Special Affordable mortgage percent in 1997, 14.3 percent in 1998, 17.6 purchases each year, each GSE must percent in 1999, and 19.2 percent in 2000. GSEs’ purchases of single-family-owner • mortgages in metropolitan areas for 2005, purchase special affordable multifamily In HUD’s Housing Goals 2000 Final Rule, rising to 33 percent in 2006 and 2007 and 34 mortgages in dollar amount equal to at least the special affordable goal was set at 20 percent in 2008. The Secretary has 1 percent of its combined (i.e., single-family percent for 2001–03. As of January 1, 2001, considered the GSEs’ ability to lead the and multifamily) annual average mortgage several changes in counting requirements industry as well as the GSEs’ financial purchases over the 2000–2002 period. The took effect for the special affordable goal, as condition. The Secretary has determined that level of this subgoal is $5.49 billion per year follows: ‘‘bonus points’’ (double credit) for the goals and subgoals are necessary and for Fannie Mae and $3.92 billion per year for purchases of goal-qualifying mortgages on appropriate. Freddie Mac. small (5–50 unit) multifamily properties and, Single-Family-Owner Home Purchase above a threshold level, mortgages on 2–4 Appendix C—Departmental Subgoal. The Department is establishing a unit owner-occupied properties; a Considerations To Establish the Special subgoal of 17 percent for the share of each ‘‘temporary adjustment factor’’ (1.20 units Affordable Housing Goal GSE’s purchases of single-family-owner credit, subsequently increased by Congress to home purchase mortgages that qualify as 1.35 units credit) for Freddie Mac’s A. Introduction special affordable and are originated in purchases of goal-qualifying mortgages on 1. Establishment of the Goal metropolitan areas in 2005 and 2006, with large (more than 50-unit) multifamily the subgoal rising to 18 percent in 2007 and The Federal Housing Enterprises Financial properties; changes in the treatment of 2008. Safety and Soundness Act of 1992 missing data; a procedure for the use of (FHEFSSA) requires the Secretary to B. Consideration of the Factors imputed or proxy rents for determining goal credit for multifamily mortgages; and establish a special annual goal designed to In considering the factors under FHEFSSA adjust the purchase by each GSE of mortgages changes regarding the ‘‘recycling’’ of funds to establish the Special Affordable Housing by loan originators. These changes are on rental and owner-occupied housing to Goal, HUD relied upon data gathered from meet the unaddressed needs of, and explained below. Fannie Mae’s performance the American Housing Survey through 2001, was 21.6 percent in 2001, 21.4 percent in affordable to, low-income families in low- the Census Bureau’s 1991 and 2001 income areas and very-low-income families 2002, and 21.2 percent in 2003. Freddie Residential Finance Surveys, the 1990 and Mac’s performance was 22.6 percent in 2001, (the Special Affordable Housing Goal). 2000 Censuses of Population and Housing, In establishing the Special Affordable 20.4 percent in 2002, and 21.4 percent in Home Mortgage Disclosure Act (HMDA) data 2003. Both GSEs surpassed this higher goal Housing Goal, FHEFSSA requires the for 1992 through 2003, and annual loan-level Secretary to consider: in all years. This section discusses the data from the GSEs on their mortgage October 2000 counting rule changes in detail 1. Data submitted to the Secretary in purchases through 2003. Appendix D connection with the Special Affordable and provides data on what goal performance discusses in detail how these data resources would have been in 2001–03 without these Housing Goal for previous years; were used and how the size of the 2. The performance and efforts of the GSEs changes.2 conventional conforming market for this goal In addition, HUD has established a special toward achieving the Special Affordable was estimated. affordable subgoal for GSE purchases of Housing Goal in previous years; The remainder of Section C discusses the multifamily mortgages. This subgoal is 3. National housing needs of targeted factors listed above, and Section D provides expressed in terms of a minimum annual families; the Secretary’s rationale for establishing the dollar volume of multifamily mortgage 4. The ability of the GSEs to lead the Special Affordable Housing Goal. industry in making mortgage credit available purchases for units qualifying for the goal, Factors 1 and 2. Data submitted to the for low-income and very-low-income rather than as a percentage of total units Secretary in connection with the Special families; and financed, as for the three housing goals. As Affordable Housing Goal for previous years, 5. The need to maintain the sound discussed below, both GSEs surpassed the and the performance and efforts of the financial condition of the enterprises. multifamily subgoal in each of these years. enterprises toward achieving the Special a. Performance on the Special Affordable 2. The Goal and Subgoals Affordable Housing Goal in previous years. Housing Goal in 1996–2003 Special Affordable Housing Goal. The rule The discussions of these two factors have HUD’s Housing Goals 1995 Final Rule provides that the Special Affordable Housing been combined because they overlap to a specified that in 1996 at least 12 percent of Goal will be 22 percent in 2005, 23 percent significant degree. the number of units financed by each of the in 2006, 25 percent in 2007, and 27 percent This section discusses each GSE’s in 2008. performance under the Special Affordable GSEs that were eligible to count toward the Units That Count Toward the Goal. Units Housing Goal over the 1996–2003 period.1 As Special Affordable Housing Goal should that count toward the Special Affordable explained in Appendix A, the data presented qualify for the goal (that is, be for very low- Housing Goal include units occupied by low- are ‘‘official HUD results’’ which, in some income families or low-income families in income owners and renters in low-income cases, differ from goal performance reported low-income areas), and at least 14 percent areas, and very low-income owners and by the GSEs in the Annual Housing Activities should qualify in 1997–2000. HUD’s October renters. Other low-income rental units in Reports (AHARs) that they submit to the multifamily properties count toward the goal Department. 2 To separate out the effects of changes in where at least 20 percent of the units in the The main finding of this section is that counting rules that took effect in 2001, this section property are affordable to families whose both Fannie Mae and Freddie Mac surpassed also compares performance in 2001 to estimated incomes are 50 percent of area median performance in 2000 if the 2001 counting rules had the Department’s Special Affordable Housing been in effect in that year. Freddie Mac’s goal income or less, or where at least 40 percent performance in 2002 has been revised due to coding of the units are affordable to families whose 1 Performance for the 1993–95 period was errors that were discovered in HUD’s review of incomes are 60 percent of area median discussed in HUD’s Housing Goals 2000 Finale 2002 data, as discussed in HUD’s press release No. income or less. Rule. 04–105, October 15, 2004.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00219 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63798 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
2000 rule made various changes in the goal $2.11 billion annually for Freddie Mac, or 1.0 and 1.2 percentage points. In 1998 Fannie counting rules, as discussed below, and percent of the average dollar volume of each Mae’s performance fell by 2.7 percentage increased the Special Affordable Housing GSE’s mortgage purchases over the 1997–99 points, while Freddie Mac’s performance Goal to 20 percent for 2001–03. period. continued to rise, by 0.7 percentage point, In the December 1995 rule, the minimum Table C.1 and Figure C.1 show thus for the first time Freddie Mac special affordable multifamily subgoals for performance on the special affordable goal 1996–2000 were set at 0.8 percent of the total and the special affordable multifamily outperformed Fannie Mae on this goal. dollar volume of each GSE’s mortgage subgoal over the 1996–2003 period, based on Freddie Mac showed a gain in performance purchases in 1994, or $1.29 billion annually HUD’s analysis. The table shows that Fannie to 17.2 percent in 1999, while Fannie Mae for Fannie Mae and $0.99 billion annually for Mae surpassed the goals by 3.4 percentage exhibited an even greater gain, to 17.6 Freddie Mac. These subgoals were increased points and 3.0 percentage points in 1996 and percent for 2001–03 in the October 2000 rule, to 1997, respectively, while Freddie Mac BILLING CODE 4210–27–P $2.85 billion annually for Fannie Mae and surpassed the goals by narrower margins, 2.0
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00220 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63799
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00221 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.067 63800 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00222 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.068 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63801
Both GSEs exhibited sharp gains in goal Special Affordable goal credit.3 This affordable goal performance took effect performance in 2000—Fannie Mae’s discrepancy did not persist in 2001–02 beginning in 2001. Most of these also applied performance increased by 1.6 percentage because of a change in counting to the low- and moderate-income goal and points, to a record level of 19.2 percent, requirements, described below. And for 2002, are discussed in Appendix A; only brief while Freddie Mac’s performance increased HUD’s official goal performance figure was summaries of those changes are given here: even more, by 3.5 percentage points, which 20.4 percent, somewhat below the figure of • Bonus points for multifamily and single- also led to a record level of 20.7 percent. 20.6 percent submitted to the Department by family rental properties. Each qualifying unit Fannie Mae’s performance was 21.6 percent Freddie Mac. For 2003, official performance in a small multifamily property counted as in 2001, 21.4 percent in 2002 and 21.3 on this goal for both GSEs was somewhat two units in the numerator in calculating percent in 2003; Freddie Mac’s performance greater than that reported by the GSEs— special affordable goal performance on all of was 22.6 percent in 2001, 20.4 percent in official performance was 21.2 percent for the goals for 2001–03. And, above a threshold 2002, and 21.4 percent in 2003. However, as Fannie Mae (as compared with 20.9 percent equal to 60 percent of the average number of discussed below, using consistent accounting reported by Fannie Mae to the Department) qualifying rental units financed in owner- and 21.4 percent for Freddie Mac (as rules for 2000–03, each GSE’s Special occupied properties over the preceding five compared with 20.3 percent reported by years, each qualifying unit in a 2–4 unit Affordable Housing Goal performance fell in Freddie Mac to the Department). owner-occupied property also counted as two every year from 2001 through 2003—in total, Fannie Mae’s performance on the Special units in the numerator in calculating goal by 2 percentage points for Fannie Mae and Affordable Housing Goal surpassed Freddie performance. 3.2 percentage points for Freddie Mac. Mac’s in 1996–97. This pattern was reversed • Freddie Mac’s Temporary Adjustment With regard to the special affordable in 1998, as Freddie Mac surpassed Fannie Factor. Freddie Mac received a ‘‘Temporary multifamily subgoal, Fannie Mae’s purchases Mae in goal performance for the first time, Adjustment Factor’’ of 1.35 units of credit for have exceeded the subgoal by wide margins though by only 0.2 percentage point. This each qualifying unit financed in ‘‘large’’ in all years, with performance ranging from improved relative performance of Freddie multifamily properties (i.e., those with 51 or 184 percent of the goal in 1996 to 315 percent Mac was due to its increased purchases of more units) in the numerator in calculating of the goal in 1999. Fannie Mae’s subgoal was multifamily loans, as it re-entered that special affordable goal performance for 2001– more than doubled in the October 2000 rule, market, and to increases in the goal- 03.4 This factor did not apply to special to a minimum of $2.85 billion in each year qualifying shares of its single-family affordable units in large multifamily from 2001 through 2003, but its qualifying mortgage purchases. However, Fannie Mae properties whose mortgages were financed by purchases amounted to $7.36 billion, or 258 again surpassed Freddie Mac in special Fannie Mae during this period. percent of the goal, in 2001, and $7.57 affordable goal performance in 1999, 17.6 • Missing data for single-family properties. billion, or 260 percent of the goal, in 2002; percent to 17.2 percent; Freddie Mac The GSEs may exclude loans with missing and $12.10 billion, or 425 percent of the regained the lead in 2000, 20.7 percent to borrower income from the denominator if the subgoal, in 2003. 19.2 percent. Freddie Mac’s official property is located in a below-median Freddie Mac has also exceeded its special performance also exceeded Fannie Mae’s income census tract, subject to a ceiling of 1 affordable multifamily subgoals in every official performance in 2001, but this percent of total owner-occupied units year, albeit by smaller margins than Fannie reflected a difference in the counting rules financed. The enterprises are also allowed to Mae. In 1996 Freddie Mac’s special applicable to the two GSEs that was enacted exclude single-family rental units with affordable multifamily mortgage purchases by Congress; if the same counting rules were missing rental information from the amounted to $1.06 billion, or 107 percent of applied to both GSEs, Fannie Mae’s denominator in calculating performance for the goal. This ratio rose to 122 percent in performance would have exceeded Freddie the special affordable goal. 1997, and exceeded 200 percent for each year Mac’s performance, by 21.6 percent to 21.1 • Missing data and proxy rents for from 1998 through 2000. Freddie Mac’s percent. multifamily properties. If rent is missing for subgoal was more than doubled in the In 2002, Freddie Mac’s performance on the multifamily units, the GSEs may apply special affordable goal was below Fannie October 2000 rule, to a minimum of $2.11 in ‘‘proxy rents,’’ up to a ceiling of 5 percent of Mae’s performance (21.4 percent), even each year from 2001 through 2003, but its total multifamily units financed, in though Freddie Mac had the advantage of the qualifying purchases amounted to $4.65 determining whether such units qualify for Temporary Adjustment Factor (TAF), which the special affordable goal. If such proxy billion, or 220 percent of the goal, in 2001; did not apply to performance by Fannie Mae. $5.22 billion, or 247 percent of the goal, in rents cannot be estimated, these multifamily Freddie Mac’s performance would have units are excluded from the denominator in 2002; and $8.79 billion, or 417 percent of the trailed Fannie Mae’s without this factor, and subgoal, in 2003. calculating performance under these goals. in fact Freddie Mac would have fallen short • Change in ‘‘recycling’’ requirements. The official figures for Freddie Mac’s of the goal, at 19.3 percent. In 2003, Freddie special affordable goal performance Under Section 1333(b)(1)(B) of FHEFSSA, if Mac’s performance (21.4 percent) slightly a GSE acquires a portfolio of mortgages presented above differ from the exceeded Fannie Mae’s performance (21.2 corresponding figures presented by Freddie originated in a previous year (that is, percent), but this resulted from application of seasoned mortgages) that qualify under the Mac in its Annual Housing Activity Reports the TAF to Freddie Mac’s performance— Special Affordable Housing goal, the seller to HUD by 0.1–0.2 percentage point for 1996– without this, Freddie Mac’s performance must be ‘‘engaged in a specific program to 2000, reflecting minor differences in the would have been 20.2 percent, barely in use the proceeds of such sales to originate application of counting rules. The official excess of the 20 percent goal. additional loans that meet such goal’’ and figures for special affordable goal b. Changes in the Goal Counting Rules for such purchases or refinancings must performance by both GSEs are the same as 2001–03 ‘‘support additional lending for housing that those submitted by the enterprises for both otherwise qualifies under such goal’’ in order GSEs for 2001, and for Fannie Mae for 2002. Several changes in the counting rules underlying the calculation of special to receive credit toward the goal. This has However, for 1996–2000, HUD’s official been referred to as the ‘‘recycling special affordable goal performance figures 3 During 1996–2000 Freddie Mac took steps to requirement.’’ The 2000 rule both clarified for Fannie Mae were approximately 1–3 the conditions under which HUD would percentage points lower than the acquire representations and warranties from lenders to attest that they were ‘‘recycling’’ the proceeds regard these statutory conditions to be corresponding figures reported by the from the sales of qualifying loans. Fannie Mae did satisfied and established certain categories of enterprise. This was due to differences not take such steps; rather, Fannie Mae excluded lenders that would be presumed to meet the between HUD and Fannie Mae in the such loans from the denominator in making its own recycling requirements. These included BIF- application of counting requirements calculations of its special affordable goal insured and SAIF-insured depository applicable to purchases of portfolios of performance. In 1996–2000 HUD counted all institutions that are regularly in the business eligible loans in the denominator, and, in the seasoned loans, based on a statutory of mortgage lending and which are subject to, requirement that the proceeds of such GSE absence of measures to verify ‘‘recycling’’ by Fannie Mae, did not award credit in the numerator of the purchases by the loan sellers should be special affordable goal for most of Fannie Mae’s 4 See Congressional Record, December 15, 2000, ‘‘recycled’’ in order for the GSE to receive seasoned mortgage purchases. pp. H12295–96.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00223 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63802 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
and have received at least a satisfactory has calculated what performance would have of missing data (including use of proxy Community Reinvestment Act performance been in 2000 under the 2001–03 rules; this rents), and changes in procedures related to evaluation rating under specified may be compared with official performance the ‘‘recycling’’ requirement. Baseline C conditions.5 in 2001–03—an ‘‘apples-to-apples incorporates in addition to the technical c. Effects of Changes in the Counting Rules comparison.’’ HUD has also calculated what changes the bonus points and, for Freddie on Goal Performance performance would have been in 2001–03 Mac, the temporary adjustment factor. Baseline B corresponds to the counting Because of the changes in special under the 1996–2000 rules; this may be approach used in this rule to take effect in affordable goal counting rules that took effect compared with official performance in 2005. Boldface figures under Baseline A for in 2001, direct comparisons between official 2000—an ‘‘oranges-to-oranges comparison.’’ 1999–2000 and under Baseline C for 2001– goal performance in 2000 and 2001–03 are These comparisons are presented in Table 03 indicate official goal performance based somewhat of an ‘‘apples-to-oranges C.2. comparison.’’ For this reason, the Department Specifically, Table C.2 shows performance on the counting rules in effect in those under the special affordable goal in three years—e.g., for Freddie Mac, 17.2 percent in 1999, 20.7 percent in 2000, 22.6 percent in 5 ways. Baseline A presents performance under The revised requirements are codified at 24 CFR 2001, 20.4 percent in 2002 and 21.4 percent 81.14(e)(4). The changes are discussed in detail in the counting rules in effect for 1996–2000. the rule preamble, 68 FR 65074–76 (October 31, Baseline B incorporates the technical changes in 2003. 2000). in counting rules—changes in the treatment BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00224 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63803
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00225 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.069 63804 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
• Performance on the Special Affordable Changes in the Department’s counting rules in 2003—this compares with only 7,196 such Housing Goal under 1996–2000 Counting related to ‘‘recycling’’ did not play a role in units financed in 2000. Small multifamily Rules Plus Technical Changes. If the Freddie Mac’s performance on the special properties also accounted for a greater share ‘‘Baseline B’’ counting approach had been in affordable goal. These same patterns also of Fannie Mae’s multifamily business in effect in 2000–03 and the GSEs’’ had generally appeared in 2002. But in 2003 2001–03—7.4 percent of total multifamily purchased the same mortgages that they bonus points for financing special affordable units financed in 2001, 13.2 percent in 2002, actually did purchase in those years, Fannie unit in small multifamily properties had a and 28.4 percent in 2003, up from 2.5 percent Mae would have surpassed the special greater impact on performance that the in 2000. However, HUD’s 2000 rule reported affordable goal in both 2000 and 2001, but temporary adjustment factor. information from the 1991 Residential not in 2002 or 2003, while Freddie Mac • Fannie Mae. The temporary adjustment Finance Survey that small multifamily would have surpassed the goal in 2000 but factor applied to Freddie Mac’s goal properties accounted for 37 percent of all fallen short in 2001–2003. Specifically, performance, but not to Fannie Mae’s multifamily units, thus Fannie Mae was still Fannie Mae’s performance would have been performance, thus counting rule changes had less active in this market than in the market 21.4 percent in 2000, 20.2 percent in 2001, less impact on its performance than on for large multifamily properties. Within the 19.9 percent in 2002, and 19.3 percent in Freddie Mac’s performance in 2001–03. The small multifamily market, there was no 2003. Freddie Mac’s performance would have largest impacts of the counting rule changes evidence that Fannie Mae targeted affordable been 21.0 percent in 2000, 19.3 percent in on Fannie Mae’s goal performance in 2001 properties to a greater extent in 2001–03 than 2001, 18.1 percent in 2002, and 17.8 percent were due to the application of bonus points in 2000. That is, 61 percent of Fannie Mae’s in 2003. for purchases of mortgages on owner- small multifamily units qualified for the • Performance on the Special Affordable occupied 2–4 unit rental properties, which special affordable goal in 2000; this fell to 46 Housing Goal under 2001–2003 Counting added 0.9 percentage point to performance; percent in 2001, 52 percent in 2002, and 42 Rules. If the 2001–03 counting rules had been bonus points for purchases of mortgages on percent in 2003. in effect in 2000–03 and the GSEs’ had small multifamily properties, which added Freddie Mac financed 50,299 units in small purchased the same mortgages that they 0.4 percentage point to performance; and multifamily properties in 2001 that were actually did purchase in that year (i.e., technical changes, which added 1.6 eligible for the special affordable goal, 22,255 abstracting from any behavioral effects of percentage points to performance—the latter such units in 2002, and 177,561 such units ‘‘bonus points,’’ for example), both GSEs included the change in the Department’s in 2003, as compared with only 2,996 such would have substantially surpassed the rules regarding ‘‘recycling’’ and the exclusion units financed in 2000. Small multifamily special affordable goal in all four years, but of single-family units with missing properties also accounted for a significantly both GSEs’ performance figures would have information from the denominator in greater share of Freddie Mac’s multifamily deteriorated somewhat between 2000 and calculating goal performance.6 The use of business in 2001–03—16.0 percent of total 2003. Specifically, Fannie Mae’s ‘‘Baseline proxy rents for multifamily properties played multifamily units financed in 2001, 7.5 C’’ performance would have been 22.2 a minor role in determining Fannie Mae’s percent in 2002, and 30.0 percent in 2003, up percent in 2000, 21.6 percent in 2001, 21.4 special affordable goal performance. These from 1.8 percent in 2000. percent in 2002, and 21.2 percent in 2003. same patterns also generally appeared in Within the small multifamily market, there Freddie Mac’s performance would have been 2002 and 2003. was some evidence that Freddie Mac targeted 23.4 percent in 2000, 22.6 percent in 2001, d. Bonus Points for the Special Affordable affordable properties to a greater extent in 20.4 percent in 2002 and 21.4 percent in Housing Goal 2001 than in 2000. That is, 55 percent of 2003. Measured on this consistent basis, As discussed above and in Appendix A, Freddie Mac’s small multifamily units then, Fannie Mae’s performance fell by 0.9 the Department established ‘‘bonus points’’ qualified for the special affordable goal in percentage point between 2000 and 2003. to encourage the GSEs to step up their 2000; this rose to 73 percent in 2001, but Freddie Mac’s ‘‘Baseline C’’ performance fell activity in 2001–03 in two segments of the declined to 60 percent in 2002 and 54 by 2.0 percentage points between 2000 and mortgage market—the small (5–50 unit) percent in 2003. 2003. These reductions were primarily due to multifamily mortgage market, and the market In summary, then, there is evidence that 2001–03 being years of heavy refinance for mortgages on 2–4 unit properties where bonus points for small multifamily properties activity. 1 unit is owner-occupied and 1–3 units are had an impact on Fannie Mae’s role in this Details of Effects of Changes in Counting occupied by renters. Bonus points did not market in 2001–03 and an even larger impact Rules on Goal Performance in 2001–03. As apply to purchases of mortgages for owner- on Freddie Mac’s role in this market. In discussed above, counting rule changes that occupied 1-unit properties, for investor- addition, Fannie Mae has announced a took effect in 2001 had significant impacts on owned 1–4 unit properties, and for large program to increase its role in this market the performance of both GSEs on the special (> 50-unit) properties, although as also further in future years.7 affordable goal in 2001—3.0 percentage discussed above, a ‘‘temporary adjustment Bonus points for single-family rental points for Fannie Mae and 3.5 percentage factor’’ applied to Freddie Mac’s purchases of properties. Above a threshold, each unit points for Freddie Mac. This section breaks goal-qualifying mortgages on large financed in a 2–4 unit property with at least down the effects of these changes on goal multifamily properties. one owner-occupied unit (referred to as performance for both GSEs; results are shown Bonus points for small multifamily ‘‘OO24s’’ below) that qualified for any of the in Table C.2. properties. Each unit financed in a small housing goals was counted as two units in • Freddie Mac. The largest impact of the multifamily property that qualified for any of the numerator (and one unit in the counting rule changes on Freddie Mac’s goal the housing goals was counted as two units denominator) in calculating goal performance performance was due to the application of in the numerator (and one unit in the for that goal in 2001–03. The threshold was the temporary adjustment factor for denominator) in calculating goal performance equal to 60 percent of the average number of purchases of mortgages on large multifamily for that goal. For example, if a GSE financed such qualifying units over the previous five properties, as enacted by Congress; this a mortgage on a 40-unit property in which 10 years. For example, Fannie Mae financed an added 1.4 percentage points to goal of the units qualified for the special average of 24,780 special affordable units in performance in 2001, as shown in Table C.2. affordable goal, 20 units would be entered in these types of properties between 1996 and Bonus points for purchases of mortgages on the numerator and 40 units in the 2000, and 55,118 such units in 2001. Thus small multifamily properties added 1.1 denominator for this property in calculating Fannie Mae received 40,250 bonus points in percentage points to performance, and bonus goal performance. this area in 2001—that is, 55,118 minus 60 points for purchase of mortgages on owner- Fannie Mae financed 37,449 units in small percent of 24,780. So 95,368 units were occupied 2–4 unit rental properties added 0.7 multifamily properties in 2001 that were entered in the numerator for these properties percentage point to performance. The eligible for the special affordable goal, 58,277 in calculating special affordable goal remaining impact (0.2 percentage point) was such units in 2002, and 214,619 such units performance. due to technical changes in counting rules— primarily, the exclusion of single-family 6 Exclusion of loans with missing information had 7 ‘‘Fannie Courting Multifamily Sellers; Small units with missing information from the a greater impact on Fannie Mae’s goal performance Banks Balking.’’ American Banker, January 13, denominator in calculating goal performance. than on Freddie Mac’s goal performance. 2003, p.1.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00226 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63805
Fannie Mae financed 176,369 units in the area where the property is located is in For scoring loans purchased by the GSEs OO24s that were eligible for the special the low-income range. Specifically, for year-by-year from 1993 through 2003, area affordable goal in 2001, 229,827 such units single-family owner-occupied units scoring is median income estimates produced by HUD’s in 2002, and 355,994 such units in 2003, as based on. Economic and Market Analysis Division were • compared with 77,985 such units financed in The mortgagors’ income at the time of used. The same median income data series 2000. However, Fannie Mae’s total single- mortgage origination • described in Appendix A for the Low- and family business increased at approximately The median income of an area specified Moderate-Income Goal was used. The in the same way as for the Low- and the same rate as its OO24 business over the determination of low-income areas was based Moderate-Income Housing Goal, that is: (i) 2001–03 period, thus the share of this on 1990 census data. business accounted for by OO24s was the For properties located in Metropolitan 2005 Procedure. Relative to the above same in 2001–03 as in 2000—4 percent. Statistical Areas (MSAs) the area is the MSA; Within the OO24 market, there was no and (ii) for properties located outside of procedure, scoring of loans purchased by the evidence that Fannie Mae targeted special MSAs, the area is the county or the non- GSEs in and after 2005 will be affected by affordable properties to a greater extent in metropolitan portion of the State in which two factors—first, re-benchmarking of area 2001–03 than in 2000. That is, approximately the property is located, whichever has the median incomes to the 2000 census as 30 percent of Fannie Mae’s OO24 units larger median income, as of the year of described in Appendix A, with a shift from qualified for the special affordable goal in mortgage origination (which may be for the 1990 to 2000 census data for identifying low- each of these years. current year or a prior year). income areas, and second, the Office of Freddie Mac financed 96,204 units in • Also, if the property is located in a Management and Budget’s June, 2003, re- OO24s that were eligible for the special Metropolitan Statistical Area (MSA), the specification of MSA boundaries based on affordable goal in 2001, 146,242 such units determination for purposes of the Special analysis of 2000 census data.9 in 2002, and 154,535 such units in 2003, as Affordable Housing Goal involves data on Analysis. For purposes of specifying the compared with 49,993 such units financed in median income of the MSA; or if the property level of the Special Affordable Housing Goal, is located elsewhere, the median income of 2000. However, Freddie Mac’s total single- the HUD estimates of area median incomes the county or the non-metropolitan portion of family business increased at approximately for MSAs, non-metropolitan counties, and the same rate as its OO24 business between the State in which the property is located, whichever is larger, as of the most recent the non-metropolitan parts of States, as 2000 and 2002, thus the share of this described in Appendix A, were used in business accounted for by OO24s was the decennial census. conjunction with the data identifying low- same in 2002 as in 2000—4 percent. And its Analogous specifications to those detailed in income areas based on the 2000 census, to re- overall single-family business increased more Appendix A for the Low- and Moderate- rapidly than its OO24 business in 2003, thus Income Housing Goal are applied in the case score loans purchased by the GSEs between OO24 units accounted for 3 percent of all of the Special Affordable Housing Goal for 1999 and 2003. The same data series were single-family units last year. rental units in single-family properties with used further in estimating the share of loans As for Fannie Mae, within the OO24 rent data available (assuming no income data originated in metropolitan areas that would market there was no evidence that Freddie available for actual or prospective tenants), be eligible to score toward the Special Mac targeted special affordable properties to for rental units in multifamily properties Affordable Housing Goal, from HMDA data. a greater extent in 2001–03 than in 2000. where rent data are available, and for rental The results of the retrospective GSE analysis That is, approximately 32–36 percent of units in multifamily properties where rent are provided in Table C.3. The results of the Freddie Mac’s OO24 units qualified for the data are not available. GSE–HMDA comparative analysis are special affordable goal in each of these four Thus, scoring loans under the Special presented in the next section. years. Affordable Housing Goal requires a data BILLING CODE 4210–27–P series showing annual median incomes for e. Effects of 2000 Census on Scoring of Loans MSAs, non-metropolitan counties, and the Toward the Special Affordable Housing Goal non-metropolitan portions of states; and the portion of a New England county outside Background. Scoring of housing units decennial census data on median incomes for of any MSA was regarded as equivalent to a county under the Special Affordable Housing Goal is census tracts; and decennial census data on in establishing the metropolitan or non- metropolitan location of a property. The MSA based on data for mortgagors’ incomes for median incomes for MSAs, non-metropolitan definitions established by the Office of Management owner-occupied units, rents for rental units, counties, and the non-metropolitan portions and Budget (OMB) in June 2003 defined MSAs in area median incomes, and, for units that are of States.8 New England in terms of counties. in the low-income but not the very low- 9 HUD has deferred application of the 2003 MSA income range, decennial census data used to 8 In New England, MSAs were defined through specification to 2005, pending completion of the determine whether the median income for mid-2003 in terms of Towns rather than Counties, present rulemaking process.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00227 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63806 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00228 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.070
Table C.3 shows three sets of estimates for the total dollar volume of single-family and was established as $2.85 billion per year and each GSE, based respectively on the counting multifamily mortgages purchased by the Freddie Mac’s as $2.11 billion per year. In rules in place in 2001–2003 (but disregarding respective GSE in 1994. Thus Fannie Mae’s 2001 Fannie Mae exceeded its subgoal by the bonus points and Temporary Adjustment subgoal was $1.29 billion per year and $4.51 billion and Freddie Mac exceeded its Factor), on the addition of 2000 census re- Freddie Mac’s subgoal was $988 million per subgoal by $2.54 billion. In 2002, Fannie Mae benchmarking and low-income areas, and year during that period. Fannie Mae exceeded its subgoal by $4.72 billion and finally on the further addition of 2003 MSA surpassed the subgoal by $1.08 billion, $1.90 Freddie Mac exceeded its subgoal by $3.11 specification. billion, $2.24 billion, $2.77 billion, and $2.50 billion. Both GSEs exceeded their subgoals in billion in 1996, 1997, 1998, 1999, and 2000 2003 by wide margins—Fannie Mae, with f. The GSEs’ Multifamily Special Affordable respectively, while Freddie Mac exceeded special affordable multifamily purchases of Purchases the subgoal by $18 million, $220 million, $12.11 billion (goal of $2.85 billion), and Since 1996 each GSE has been subject to $1.70 billion, $1.27 billion, and $1.41 billion. Freddie Mac, with purchases of $8.79 billion an annual dollar-based subgoal for Special The subgoal was established for 2001–03 as (goal of $2.11 billion.) Those subgoals are Affordable multifamily mortgage purchases, 1.0 percent of the average annual volume of also in effect for 2004. Table C.1 includes as discussed above. This subgoal was each GSE’s total mortgage purchases over the figures on subgoal performance, and they are established for 1996–2000 as 0.8 percent of 1997–99 period. Thus Fannie Mae’s subgoal depicted graphically in Figure C.2.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00229 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63808 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00230 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.071
g. Characteristics of the GSEs’ Special variable from year to year since 1996. Fannie purchase percentages for single-family owner Affordable Purchases Mae’s multifamily purchases were at 37.7 properties exhibited a similar variability The following analysis presents percent in 1996,28.8 percent in 2001, and through this entire period, as did their 20.0 percent in 2002, with a high of 44.0 information on the composition of the GSEs’ purchases of mortgages financing single- percent in 1997 and a low of 19.6 percent in family rental units from 1996 through 2003. Special Affordable purchases according to 2003. Freddie Mac’s multifamily purchases area income, unit affordability, tenure of unit represented 29.4 percent of all purchases Both GSEs’ high points for mortgages and property type (single-or multifamily). qualifying for the goal in 1996, 27.0 percent financing single-family rental units occurred Tables C.4 and C.5 show that each GSE’s in 2001, and 20.4 percent in 2002, with a in 2002: Fannie Mae’s purchase percentage reliance on multifamily housing units to high of 31.5 percent in 1997 and a low of was 20.0 percent while Freddie Mac’s was meet the special affordable goal has been 20.4 percent in 2002. The two GSEs’ 18.1 percent.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00231 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63810 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00232 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.072 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63811
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00233 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.073
Tables C.4 and C.5 also show the allocation were 38.3 percent in 1997 and 35.5 percent conforming market. (See Tables A.13 to A.16 of units qualifying for the goal as related to in 2001. Thus given the definition of special in Appendix A.). There were two main the family income and area median income affordable housing in terms of household and findings with respect to the special affordable criteria in the goal definition. Very-low- area income characteristics, both GSEs have category. First, Freddie Mac and Fannie Mae income families (shown in the two leftmost consistently relied substantially more on have historically lagged depositories and the columns in the tables) accounted for 83.4 low-income characteristics of households overall market in providing mortgage funds percent of Fannie Mae’s units qualifying than low-income characteristics of census for special affordable borrowers over periods, under the goal in 1997, rising to 85.2 percent tracts to meet this goal. such as 1993–2003, 1996–2003 and 1999– in 1999. For Freddie Mac, very-low-income h. The GSEs’ Performance Relative to the 2003. Between 1993 and 2003, 12.2 percent families accounted for 81.9 percent of units Market qualifying under the goal in 1997, rising to of Freddie Mac’s mortgage purchases were 84.9 percent in 1999. In contrast, mortgage Section E.9 in Appendix A uses HMDA for special affordable borrowers, 13.3 percent purchases from low-income areas (shown in data and GSE loan-level data for home of Fannie Mae’s purchases, 15.4 percent of the first and third columns in the tables) purchase mortgages on single-family-owner loans originated by depositories, and 15.5 accounted for 33.7 percent of Fannie Mae’s properties in metropolitan areas to compare percent of loans originated in the units qualifying under the goal in 1997, the GSEs’ performance in special affordable conventional conforming market (without compared to 35.5 percent in 2001. The lending to the performance of depositories estimated B&C loans). For the recent years, corresponding percentages for Freddie Mac and other lenders in the conventional the GSE-market comparisons are as follows:
Market Year Freddie Mac Fannie Mae (w/o B&C) (in percent) (in percent) (in percent) (in percent)
1999 ...... 12.8 12.5 17.0 2000 ...... 14.7 13.3 16.6 2001 ...... 14.4 14.9 15.6 2002 ...... 15.8 16.3 16.1 2003 ...... 15.6 17.1 15.9 1996–2003 ...... 13.2 14.1 15.9 1999–2003 ...... 14.7 15.1 16.2 2001–2003 ...... 15.2 16.2 15.9
During the period between 1999 and 2003, percent of all special affordable owner and complements Section C of Appendix A, the GSEs’ performance was slightly over 90 rental units that were financed in the which presents detailed analyses of housing percent of the market—special affordable conventional conforming market between problems and demographic trends for lower- loans accounted for 15.1 percent of Fannie 1999 and 2002. The GSEs’ 41-percent share income families which are relevant to the Mae’s purchases, 14.7 percent of Freddie of the special affordable market was three- issue addressed in this part of Appendix C. Mac’s purchases, and 16.2 percent of loans fourths of their 55-percent share of the Data from the American Housing Survey originated in the conforming market. overall market. Even in the owner market, demonstrate that housing problems and Second, while both GSEs have improved where the GSEs account for 61 percent of the needs for affordable housing continue to be their performance over the past few years, market, their share of the special affordable more pressing in the lowest-income Fannie Mae has been made more progress market was only 52 percent during this categories than among moderate-income than Freddie Mac in erasing its gap with the period. While the GSEs improved their families, as established in HUD’s analysis for market. During the first three years (2001, market shares during 2001–2003, this the 1995 and 2000 Final Rules. Table C.6 2002, and 2003) of HUD’s new housing goal analysis shows that there is room and ample displays figures on several types of housing targets, the average share of Fannie Mae’s opportunities for the GSEs, and particularly problems—high housing costs relative to purchases going to special affordable loans Freddie Mac, to improve their performance income, physical housing defects, and was 16.2 percent, which was above the in purchasing affordable loans at the lower- crowding–for both owners and renters. market average of 15.9 percent. The share of income end of the market. Section C.3 of this Figures are presented for households Freddie Mac’s purchases going to special appendix discusses a home purchase subgoal experiencing multiple (two or more) of these affordable loans was 15.2 percent during this designed to place the GSEs in such a problems as well as households experiencing period. leadership position in the special affordable a severe degree of either cost burden or Section G in Appendix A discusses the role single-family-owner market. physical problems. Housing problems in of the GSEs both in the overall special 2001 continued to be much more frequent for affordable market and in the different Factor 3. National Housing Needs of Low- the lowest-income groups.10 Incidence of segments (single-family owner, single-family Income Families in Low-Income Areas and problems is shown for households in the rental, and multifamily rental) of the special Very-Low-Income Families income range covered by the special affordable market. The GSEs’ special This discussion concentrates on very-low- affordable goal, as well as for higher income affordable purchases accounted for 41 income families with the greatest needs. It households.
10 Tabulations of the 2001 American Housing Research. The results in the table categorize renters reporting housing assistance as having no housing Survey by HUD’s Office of Policy Development and problems.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00234 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63813
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00235 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.074
This analysis shows that priority problems 1. Severe Housing Problems would have surpassed the goal in 2000 but of severe cost burden or severely inadequate The data presented in Section C.3 fallen short in 2001–03. housing are noticeably concentrated among demonstrate that housing problems and The above performance figures are for the renters and owners with incomes below 60 needs for affordable housing are much more special affordable goal defined in terms of percent of area median income: 30.5 percent pressing in the lowest-income categories than 1990 Census geography. Switching to 2000 of renter households and 34.9 percent of among moderate-income families. The high Census data slightly increases the coverage of owner households had priority problems. In incidence of severe problems among the special affordable goal, which increases the contrast, in the next higher income range, up lowest-income renters reflects severe special affordable share of the GSEs’ to 80 percent of area median income, 2.5 shortages of units affordable to those renters. purchases by up to one percentage point. percent of renter households and 7.3 percent At incomes below 60 percent of area median, of owner households had priority problems. Based on 2000 Census geography and adding 34.7 percent of renters and 21.6 percent of 2003 MSAs, and excluding counting The table demonstrates the significance of owners paid more than 50 percent of their affordability problems: Sixty-five percent of requirements (a) and (b), then Fannie Mae’s income for housing. In this same income performance would have been 21.7 percent very-low-income renter families had rent range, 65.6 percent of renters and 42.4 in 2000, 20.1 percent in 2001, 19.4 percent burden over 30 percent of income; 35 percent percent of owners paid more than 30 percent in 2002, and 20.8 percent in 2003. Freddie had rent burden over 50 percent of income. of their income for housing. In addition, 31.5 Thirteen percent had moderately or severely percent of renters and 23.8 percent of owners Mac’s performance would have been 20.8 inadequate housing; 6 percent lived in exhibited ‘‘priority problems’’, meaning percent in 2000, 19.1 percent in 2001, 17.3 crowded conditions, defined as more than housing costs over 50 percent of income or percent in 2002 and 19.0 percent in 2003. See one person per room. severely inadequate housing. Table C.3. Factor 4. The Ability of the Enterprises To Homeownership gaps and other disparities in b. Single-Family Market Comparisons in Lead the Industry in Making Mortgage Credit the housing and mortgage markets discussed Metropolitan Areas in Section H of Appendix A also apply to Available for Low-Income and Very-Low- The Special Affordable Housing Goal is Special Affordable housing and mortgages. Income Families designed, in part, to ensure that the GSEs The discussion of the ability of Fannie Mae 2. GSE Performance and the Market maintain a consistent focus on serving the and Freddie Mac to lead the industry in a. The GSEs’ Special Affordable Housing very low-income portion of the housing Section G of Appendix A is relevant to this Goals Performance market where housing needs are greatest. factor—the GSEs’ roles in the owner and Section C compared the GSEs’ performance In the October 2000 rule, the special rental markets, their role in establishing in special affordable lending to the widely-applied underwriting standards, their affordable goal was set at 20 percent for 2001–03. Effective on January 1, 2001, performance of depositories and other role in the development of new technology lenders in the conventional conforming for mortgage origination, their strong staff several changes in counting requirements market for single-family home loans. The resources, and their financial strength. came into effect for the special affordable analysis showed that while both GSEs have Additional analyses of the potential ability of goal, as follows: (a)‘‘bonus points’’ (double the enterprises to lead the industry in the credit) for purchases of mortgages on small been improved their performance, their past low- and very-low-income market appears (5–50 unit) multifamily properties and, above average performance (1993–2003, 1996–2003, below in Section D, which explains the a threshold level, mortgages on 2–4 unit and 1999–2003) has been below market Department’s rationale for the home purchase owner-occupied properties; (b) a ‘‘temporary levels. During 2002 and 2003, Fannie Mae subgoal for Special Affordable loans. adjustment factor’’ (1.35 unit credit) for improved its performance enough to lead the Freddie Mac’s purchases of mortgages on special affordable market for home purchase Factor 5. The Need to Maintain the Sound large (more than 50 unit) multifamily loans, but Freddie Mac, although it also Financial Condition of the GSEs properties; (c) changes in the treatment of improved its performance during this recent HUD has undertaken a separate, detailed missing data; (d) a procedure for the use of period, continues to lag behind the primary economic analysis of this final rule, which imputed or proxy rents for determining goal market. Between 1999 and 2003, special includes consideration of (a) the financial credit for multifamily mortgages; and (e) affordable borrowers accounted for 15.1 returns that the GSEs earn on special changes regarding the ‘‘recycling’’ of funds percent of the home loans purchased by affordable loans and (b) the financial safety by loan originators. Fannie Mae, 14.7 percent of Freddie Mac’s and soundness implications of the housing Counting requirements (a) and (b) expired purchases, 16.2 percent of home loans goals. Based on this economic analysis, HUD at the end of 2003 while (c)–(e) will remain originated by depositories, and 16.2 percent concludes that the housing goals in this final in effect after that. If this counting of all home loans originated in the approach—without the bonus points and the rule raise minimal, if any, safety and conventional conforming market (without ‘‘temporary adjustment factor’’—had been in soundness concerns. B&C loans). As noted above, while both GSEs effect in 2000–2003, and the GSEs’ had C. Determination of the Special Affordable purchased the same mortgages that they have improved their performance over the Housing Goal actually did purchase in both years, then past few years, Fannie Mae has made more Several considerations, many of which are Fannie Mae’s performance would have been progress than Freddie Mac in closing its gap reviewed in Appendixes A and B and in 21.4 percent in 2000, 20.2 percent in 2001, with the market. During 2003, the share of previous sections of this Appendix, led to the 19.9 percent in 2002, and 19.4 percent in Fannie Mae’s purchases going to special determination of the Special Affordable 2003. Freddie Mac’s performance would have affordable loans was 17.1 percent, which was Housing Goal, the multifamily special been 21.0 percent in 2000, 19.3 percent in 1.2 percentage points above the market affordable subgoal, and the special affordable 2001, 18.1 percent in 2002, and 17.8 percent average of 15.9 percent. The share of Freddie subgoal for home purchase loans on single- in 2003. Fannie Mae would have surpassed Mac’s purchases going to special affordable family-owner properties in metropolitan the special affordable goal in both 2000 and loans had improved to 15.6 percent by 2003. areas. 2001, but not in 2002 or 2003. Freddie Mac (See Figure C.3.)
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00236 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63815
BILLING CODE 4210—27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00237 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.075 63816 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
3. Ability To Lead the Single-Family Owner example, related to the GSEs’ overall ability years. As emphasized in Appendix A, Market: A Special Affordable Subgoal to lead the single-family-owner market— changing population demographics will The Secretary believes the GSEs can play while others were specific to the low-mod result in a need for the primary and a leadership role in the special affordable subgoal. Because the reader can refer to secondary mortgage markets to meet market. Thus, the Department is establishing Appendix A, this appendix provides a briefer nontraditional credit needs, respond to a subgoal of 17 percent for each GSE’s discussion of the more general factors. The diverse housing preferences, and overcome purchases of home purchase loans for special specific considerations that led to the subgoal information and other barriers that many affordable families in the single-family-owner for special affordable loans can be organized immigrants and minorities face. The GSEs market of metropolitan areas for 2005 and around the following four topics: have to increase their efforts in helping 2006, rising to 18 percent during 2007 and (1) The GSEs have the ability to lead the special affordable families—but so far they have played a surprisingly small role in 2008. The purpose of this subgoal is to market. As discussed in Appendix A, the serving minority first-time homebuyers. It is encourage the GSEs to improve their GSEs have the ability to lead the primary estimated that the GSEs accounted for 46.5 purchases of mortgages for very-low-income market for single-family-owner loans, which percent of all (both government and and minority first-time homebuyers who are is their ‘‘bread-and-butter’’ business. Both conventional) home loans originated between expected to enter the housing market over the GSEs have been dominant players in the 1999 and 2001; however, they accounted for next few years. If the GSEs meet the 18- home purchase market for years, funding 61 only 14.3 percent of home loans originated percent subgoal, they will be leading the percent of the single-family-owner mortgages for African-American and Hispanic first-time primary market by approximately two financed between 1999 and 2002. Through their many new product offerings and their homebuyers. A subgoal for special affordable percentage points, based on the income home purchase loans should increase the characteristics of home purchase loans various partnership initiatives, the GSEs have shown that they have the capacity to reach GSEs’ efforts in important sub-markets such reported in HMDA. HMDA data show that as the one for minority first-time special affordable families accounted for an out to very-low-income and other special affordable borrowers. They also have the staff homebuyers. average of 16.2 (15.9) percent of single- (4) There are ample opportunities for the family-owner loans originated in the expertise and financial resources to make the extra effort to lead the primary market in GSEs to improve their performance. Special conventional conforming market of affordable mortgages are available for the metropolitan areas between 1999 and 2003 funding single-family-owner mortgages for special affordable borrowers. GSEs to purchase, which means they can (2001 and 2003). Loans in the B&C portion improve their performance and lead the of the subprime market are not included in (b) GSEs’ Performance Relative to the Market. Even though the GSEs have had the primary market in purchasing loans for these these averages. As explained in Appendix D, very-low-income borrowers. Sections B, C, HUD also projected special affordable shares ability to lead the home purchase market, their past average performance (1993–2003, and I of Appendix A and Section H of for the market for 1999 to 2002 using the new Appendix D explain that the special 1996–2003, and 1999–2003) has been below 2000 Census geography and the new OMB affordable lending market has shown an market levels. During 2003, Fannie Mae specifications. For special affordable loans, underlying strength over the past few years improved its performance enough to lead the the 2000-based Census data resulted in that is unlikely to vanish (without a special affordable market for home purchase special affordable shares for the market and significant increase in interest rates or a loans, but Freddie Mac, although it also has the GSEs that were similar to the 1990-based decline in the economy). The special special affordable shares reported in Section improved its performance, continues to lag affordable share of the home purchase market C of this appendix. behind the primary market. The subgoals will has averaged approximately 16 percent since To reach the 18-percent subgoal for 2008, ensure that Fannie Mae maintains and 1996 and annually has been in the 15–17 Freddie Mac would have to improve its further improves its above-market percent range. Second, the market share data performance by 2.4 percentage points over its performance and that Freddie Mac not only reported in Table A.30 of Appendix A special affordable share of 15.6 percent in erases its current gap with the market but demonstrate that there are newly originated 2003. Fannie Mae would have to improve its also takes a leadership position as well. With loans available each year for the GSEs to performance by 0.9 percentage point over its respect to the GSEs’ historical performance, purchase. The GSEs’ purchases of single- market-leading special affordable share of special affordable mortgages accounted for family owner loans represented 61 percent of 17.1 percent in 2003. The approach taken is 13.2 (14.7) percent of Freddie Mac’s all single-family-owner loans originated for the GSEs to obtain their leadership purchases during 1996–2003 (1999–2003), for between 1999 and 2002, compared with 52 position by staged increases in the special 14.1 (15.1) percent of Fannie Mae’s percent of the special affordable loans that affordable subgoal; this will enable the GSEs purchases, and for 15.9 (16.2) percent of were originated during this period. Thus, half to take new initiatives in a correspondingly primary market originations (excluding B&C of the special affordable conforming market staged manner to achieve the new subgoal loans). The type of improvement needed for is not touched by the GSEs. As noted above, each year. Thus, the increases in the special Freddie Mac to meet this new special the situation is even more extreme for special affordable subgoal are sequenced so that the affordable subgoal was demonstrated by sub-markets such the minority first-time GSEs can gain experience as they improve Fannie Mae during 2001–2003, as Fannie homebuyer market where the GSEs have only and move toward the new higher subgoal Mae increased its special affordable a minimal presence. Between 1999 and 2001, targets. performance from 14.9 percent of its single- the GSEs purchased only 33 percent of The subgoal applies only to the GSEs’ family-owner business in 2001 to 16.3 conventional conforming loans originated for purchases in metropolitan areas because the percent in 2002 to 17.1 percent in 2003. minority first-time homebuyers, even though HMDA-based market benchmark is only (3) Disparities in Homeownership and they purchased 57 percent of all home loans available for metropolitan areas. HMDA data Credit Access Remain. There remain originated in the conventional conforming for non-metropolitan counties are not reliable troublesome disparities in our housing and market during that period. But also enough to serve as a market benchmark. The mortgage markets, even after the ‘‘revolution important, the GSEs’ purchases under the Department is also setting home purchase in affordable lending’’ and the growth in subgoal are not limited to new mortgages that subgoals for the other two goals-qualifying homeownership that has taken place since are originated in the current calendar year. categories, as explained in Appendices A and the mid-1990s. The homeownership rate for The GSEs can purchase loans from the B. Sections E.9 and G of Appendix A provide African-American and Hispanic households substantial, existing stock of special additional information on the opportunities remains 25 percentage points below that of affordable loans held in lenders’ portfolios, for an enhanced GSE role in the special white households. Minority families face after these loans have seasoned and the GSEs affordable segment of the home purchase many barriers in the mortgage market, such have had the opportunity to observe their market and on the ability of the GSEs to lead as lack of capital for down payment and lack payment performance. In fact, based on that market. of access to mainstream lenders (see above). Fannie Mae’s recent experience, the purchase The preamble and Appendix A discuss in Immigrants and minorities—many of whose of seasoned loans appears to be one useful some detail the factors that the Department very-low-income levels will qualify them as strategy for purchasing goals-qualifying considered when setting the subgoal for low- special affordable—are projected to account loans. and moderate-income loans. Several of the for almost two-thirds of the growth in the For the reasons given above, the Secretary considerations were general in nature—for number of new households over the next ten believes that the GSEs can do more to raise
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00238 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63817
the special affordable shares of the home upper end of HUD’s market range estimate have the ability to continue to provide loans they purchase on single-family-owner for this goal (23–27 percent), consistent with further support for the market. properties. This can be accomplished by the statutory criterion that HUD should Specifically, Fannie Mae’s total eligible building on efforts that the enterprises have consider the GSEs’ ability to lead the market multifamily mortgage purchase volume already started, including their new for each Goal. To enable the GSEs to achieve increased from $4.6 billion in 1993 to $12.5 affordable lending products aimed at special this leadership, the Department is billion in 1998, and then jumped sharply to groups such as first-time homebuyers, their establishing modest increases in the Special $18.7 billion in 2001 and $18.3 billion in many partnership efforts, their outreach to Affordable Goal for 2005, which will increase 2002, and $33.3 billion in 2003. Its special inner city neighborhoods, their incorporation year-by-year through 2008, to achieve the affordable multifamily mortgage purchases of greater flexibility into their underwriting ultimate objective for the GSEs to lead the followed a similar path, rising from $1.7 guidelines, and their purchases of seasoned market under a range of foreseeable economic billion in 1993 to $3.5 billion in 1998 and CRA loans. A wide variety of quantitative circumstances by 2008. Such a program of $4.1 billion in 1999, and also jumping and qualitative indicators indicate that the staged increases is consistent with the sharply to $7.4 billion in 2001 and $7.6 GSEs’ have the resources and financial statutory requirement that HUD consider the billion in 2002 and $12.2 billion in 2003. As strength to improve their special affordable past performance of the GSEs in setting the a result of its strong performance, Fannie performance enough to lead the market. Goals. Staged annual increases in the Special Mae’s purchases have been at least twice its 4. Size of the Overall Special Affordable Affordable Goal will provide the enterprises minimum subgoal in every year since 1997— Mortgage Market with opportunity to adjust their business 247 percent of the subgoal in that year, 274 models and prudently try out business percent in 1998, 315 percent in 1999, 294 As detailed in Appendix D, single-family strategies, so as to meet the required 2008 percent in 2000, and, under the new higher and multifamily special affordable mortgages level without compromising other business subgoal level, 258 percent in 2001, 266 are estimated to account for 23–27 percent of objectives and requirements. percent in 2002, and 426 percent in 2003. the dwelling units financed by conventional Section C compared the GSEs’ role in the Freddie Mac’s total eligible multifamily conforming mortgages; in estimating the size overall market with their role in the special mortgage purchase volume increased even of the market, HUD used alternative affordable market. The GSEs’ purchases more sharply, from $0.2 billion in 1993 to assumptions about future economic and provided financing for 26,118,927 dwelling $6.6 billion in 1998, and then jumped market affordability conditions that were less units, which represented 55 percent of the sharply in 2001 to $11.8 billion and $13.3 favorable than those that existed over the 47,551,039 single-family and multifamily billion in 2002, and $21.5 billion in 2003. Its past several years. HUD is well aware of the units that were financed in the conventional special affordable multifamily mortgage volatility of mortgage markets and the conforming market between 1999 and 2002. purchases followed a similar path, rising possible impacts on the GSEs’ ability to meet However, in the special affordable part of the from $0.1 billion in 1993 to $2.7 billion in the housing goals. Should conditions change market, the 5,103,186 units that were 1998, and also jumping sharply to $4.6 such that the goals are no longer reasonable financed by GSE purchases represented only billion in 2001 and $5.2 billion in 2002, and or feasible, the Secretary has the authority to 41 percent of the 12,413,759 dwelling units $8.8 billion in 2003. As a result of its strong revise the goals. that were financed in the market. Thus, there performance, Freddie Mac’s purchases have 5. The Special Affordable Housing Goal for appears to be ample room for the GSEs to also been at least twice its minimum subgoal 2005–2008 improve their performance in the special in every year since 1998—272 percent of the The Special Affordable Housing Goal for affordable market. In addition, there are subgoal in that year, 229 percent in 1999, 243 2005 is 22 percent of eligible purchases, a several market segments (e.g., first-time percent in 2000, and, under the new higher two percentage point increase over the homebuyers) that would benefit from a subgoal level, 220 percent in 2001, 247 current goal of 20 percent, with the goal greater secondary market role by the GSEs, percent in 2002, and 417 percent in 2003. rising to 23 percent in 2006, 25 percent in and special affordable borrowers are The Special Affordable Housing 2007, and 27 percent in 2008. The bonus concentrated in these markets. Multifamily Subgoals set forth in this rule are points for small multifamily properties and 6. Multifamily Special Affordable Subgoals reasonable and appropriate based on the owner-occupied 2–4 unit properties, as well Based on the GSEs’ past performance on Department’s analysis of this market. The as Freddie Mac’s Temporary Adjustment the special affordable multifamily subgoals, Department’s decision to retain the Factor, will no longer be in effect for goal and on the outlook for the multifamily multifamily subgoal is based on the fact that counting purposes. It is recognized that mortgage market, HUD is establishing that HUD’s analysis indicates that multifamily neither GSE would have met the 22-percent these subgoals be retained and increased for housing still serves the housing needs of target for 2005 in the past three years. Under the 2005–2008 period. Unlike the overall lower-income families and families in low- the new counting rules, Fannie Mae’s special goals, which are expressed in terms of income areas to a greater extent than single- affordable performance is estimated to have minimum goal-qualifying percentages of total family housing. By retaining the multifamily been 18.6 percent in 1999, 21.7 percent in units financed, these subgoals for 2001–03 subgoal, the Department ensures that the 2000, 20.1 percent in 2001, 19.4 percent in and in prior years have been expressed in GSEs continue their activity in this market, 2002, and 20.8 percent in 2003. Fannie Mae terms of minimum dollar volumes of goal- and that they achieve at least a minimum would have to increase its performance in qualifying multifamily mortgage purchases. level of special affordable multifamily 2005 by 1.9 percentage points over its Specifically, each GSE’s special affordable mortgage purchases that are affordable to average (unweighted) performance of 20.1 multifamily subgoal is currently equal to 1.0 lower-income families. The Department percent over these last five years. By 2008 percent of its average total (single-family plus establishes each GSE’s special affordable this increase relative to average 1999–2003 multifamily) mortgage volume over the 1997– multifamily subgoal as 1.0 percent of its performance would be 6.9 percentage points. 99 period. Under this formulation, in October average annual dollar volume of total (single- Freddie Mac’s performance is projected to 2000 the subgoals were set at $2.85 billion family and multifamily) mortgage purchases have been 17.4 percent in 1999, 20.8 percent per year for Fannie Mae and $2.11 billion per over the 2000–2002 period. In dollar terms, in 2000, 19.1 percent in 2001, 17.3 percent year for Freddie Mac, in each of calendar the Department’s subgoal is $5.49 billion per in 2002, and 19.0 percent in 2003. Freddie years 2001 through 2003. These represented year in special affordable multifamily Mac would have to increase its performance increases from the goals for 1996–2000, mortgage purchases for Fannie Mae, and in 2005 by 3.3 percentage points over its which were $1.29 billion annually for Fannie $3.92 billion per year in special affordable average (unweighted) performance of 18.7 Mae and $0.99 billion annually for Freddie multifamily mortgage purchases for Freddie percent over these last five years. By 2008 Mac. These subgoals are also in effect for Mac. These subgoals would be less than this increase relative to average 1999–2002 2004. actual special affordable multifamily performance would be 8.3 percentage points. HUD’s Determination. The multifamily mortgage purchase volume in 2001–2003 for However, GSE goal performance in 2001–03 mortgage market and both GSEs’ multifamily both GSEs; thus the Department believes that was reduced by the heavy refinance wave of transactions volume grew significantly over they would be feasible for the 2005–2008 this period. the 1993–2003 period, indicating that both period. The objective of HUD’s Special Affordable enterprises have provided increasing support Some commenters advocated increasing Goal is to bring the GSEs’ performance to the for the multifamily market, and that they the special affordable multifamily subgoals
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00239 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63818 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
from the levels in the rule. In light of the high G, and H report HUD’s estimates for the Low- However, in their comments on the levels of such purchases by both GSEs in and Moderate-Income Goal, the Underserved proposed rule, both GSEs criticized HUD’s 2003, HUD considered raising these subgoals, Areas Goal, and the Special Affordable implementation of its market methodology.4 but decided not to do so because HUD Housing Goal, respectively. As noted above, their major criticisms and believes that the overall special affordable HUD received numerous comments on the HUD’s responses to their criticisms can be goals established in this final rule will proposed rule relating to its market found throughout this appendix. HUD provide sufficient incentives for the GSEs to methodology and the size of its market ranges recognizes that there is no single, perfect data play a major role in the special affordable for each of the three goals. These comments, set for estimating the size of the affordable multifamily mortgage market, and that in all and HUD’s responses to them, are discussed lending market and that available data bases likelihood they will continue to exceed these throughout this appendix. on different sectors of the market must be subgoals by significant margins for 2005–08. In developing this final rule, HUD has combined in order to implement its market followed the same basic approach that it 7. Conclusion share model (as outlined in Section B below). followed in the last two GSE final rules and As this appendix will show, HUD has HUD has determined that the Special the recent GSE proposed rule. HUD has carefully combined various mortgage market Affordable Housing Goal in this rule carefully reviewed existing information on data bases in a manner which draws on the addresses national housing needs within the mortgage activity in order to understand the strength of each in order to implement its income categories specified for this goal, weakness of various data sources and has market methodology and to arrive at a while accounting for the GSEs’ past conducted sensitivity analyses to show the reasonable range of estimates for the three performance in purchasing mortgages effects of alternative parameter assumptions. goals-qualifying shares of the mortgage meeting the needs of very-low-income HUD is well aware of uncertainties with market. In this appendix, HUD demonstrates families and low-income families in low- some of the data and much of this appendix the robustness of its market estimates by income areas. HUD has also considered the is spent discussing the effects of alternative reporting the results of numerous sensitivity size of the conventional mortgage market assumptions about data parameters and analyses that examine a range of assumptions serving very-low-income families and low- presenting the results of an extensive set of about the relative importance of the rental income families in low-income areas. sensitivity analyses, many of the latter being and owner markets and the goals-qualifying Moreover, HUD has considered the GSEs’ directly related to comments received on the ability to lead the industry as well as their proposed rule. shares of the owner portion of the mortgage financial condition. HUD has determined In an earlier critique of HUD’s market share market. that a Special Affordable Housing Goal of 22 model, Blackley and Follain (1995, 1996) This appendix reviews in some detail percent in 2005, 23 percent in 2006, 25 concluded that conceptually HUD had HUD’s efforts to combine information from percent in 2007, and 27 percent in 2008 is chosen a reasonable approach to determining several mortgage market databases to obtain both necessary and achievable. HUD has also the size of the mortgage market that qualifies reasonable values for the model’s parameters. determined that a multifamily special for each of the three housing goals.1 Blackley The next section provides an overview of affordable subgoal for 2005–2008 set at 1.0 and Follain correctly note that the challenge HUD’s market share model. percent of the average of each GSE’s lies in getting accurate estimates of the B. Overview of HUD’s Market Share respective dollar volume of combined model’s parameters. In their comments on Methodology 5 (single-family and multifamily) 1999–2002 the 2000 Proposed GSE Rule, both Fannie mortgage purchases in is both necessary and Mae and Freddie Mac stated that HUD’s 1. Definition of Market Share achievable. Finally, HUD is establishing a market share model (outlined in Section B The size of the market for each housing subgoal of 17 percent for the GSEs’ purchases below) was a reasonable approach for goal is one of the factors that the Secretary of single-family-owner mortgages that qualify estimating the goals-qualifying (low-mod, is required to consider when setting the level for the special affordable goal and are special affordable, and underserved areas) of each housing goal.6 Using the Low- and originated in metropolitan areas, for 2005, shares of the mortgage market. Freddie Mac Moderate-Income Housing Goal as an with this subgoal remaining at 17 percent in stated: example, the market share in a particular 2006, then rising to 18 percent in both 2007 We believe the Department takes the correct year is defined as follows: and 2008. The Secretary has considered the approach in the Final rule by examining Low- and Moderate-Income Share of GSEs’ ability to lead the industry as well as several different data sets, using alternative Market: The number of dwelling units the GSEs’ financial condition. The Secretary methodologies, and conducting sensitivity financed by the primary mortgage market in has determined that the goals, the analysis. We applaud the Department’s a particular calendar year that are occupied multifamily subgoals, and the single-family- general approach for addressing the by (or affordable to, in the case of rental owner subgoals are necessary and empirical challenges.2 units) families with incomes equal to or less appropriate. *** than the area median income divided by the total number of dwelling units financed in Appendix D—Estimating the Size of the Similarly, Fannie Mae stated that ‘‘HUD has developed a reasonable model for assessing the conforming conventional primary Conventional Conforming Market for the size of the affordable housing market.’’ 3 mortgage market. Each Housing Goal There are three important aspects to this A. Introduction 1 Dixie M. Blackley and James R. Follain, ‘‘A definition. First, the market is defined in Critique of the Methodology Used to Determine terms of ‘‘dwelling units’’ rather than, for In establishing the three housing goals, the Affordable Housing Goals for the Government Secretary is required to assess, among a Sponsored Housing Enterprises,’’ unpublished 4 See Freddie Mac, ‘‘Comments of the Federal number of factors, the size of the report prepared for Office of Policy Development Home Loan Mortgage Corporation on HUD’s conventional market for each goal. This and Research, Department of Housing and Urban Proposed Housing Goals for the Federal National Development, October 1995; and ‘‘HUD’s Market appendix explains HUD’s methodology for Mortgage Association (Fannie Mae) and the Federal Share Methodology and its Housing Goals for the estimating the size of the conventional Home Loan Mortgage Corporation (Freddie Mac) for Government Sponsored Enterprises,’’ unpublished the Years 2005–2008 and Amendments to HUD’s market for each of the three housing goals. paper, March 1996. Following this overview, Section B 2 Regulation of Fannie Mae and Freddie Mac,’’ July See Freddie Mac, ‘‘Comments on Estimating the 16, 2004; and Fannie Mae, ‘‘Fannie Mae’s summarizes the main components of HUD’s Size of the Conventional Conforming Market for Comments on HUD’s Proposed Housing Goals for market-share model and identifies those Each Housing Goal: Appendix III to the Comments the Federal National Mortgage Association (Fannie parameters that have a large effect on the of the Federal Home Loan Mortgage Corporation on Mae) and the Federal Home Loan Mortgage HUD’s Regulation of the Federal National Mortgage relative market shares. Sections C and D Corporation (Freddie Mac) for the Years 2005–2008 Association (Fannie Mae) and the Federal Home discuss two particularly important market and Amendments to HUD’s Regulation of Fannie parameters, the size of the multifamily Loan Mortgage Corporation (Freddie Mac)’’, May 8, 2000, page 1. Mae and Freddie Mac,’’ July 16, 2004. market and the share of the single-family 3 See Fannie Mae, ‘‘Fannie Mae’s Comments on 5 Readers not interested in this overview may mortgage market accounted for by single- HUD’s Regulation of the Federal National Mortgage want to proceed to Section C, which begins the family rental properties. Section E provides Association (Fannie Mae) and the Federal Home market analysis by examining the size of the a more systematic presentation of the model’s Loan Mortgage Corporation (Freddie Mac)’’, May 8, multifamily market. equations and main assumptions. Sections F, 2000, page 53. 6 Sections 1332(b)(4), 1333(a)(2), and 1334(b)(4).
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00240 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63819
example, ‘‘value of mortgages’’ or ‘‘number of 2. Three-Step Procedure occupied properties’’ is the percentage of properties.’’ Second, the units are ‘‘financed’’ Ideally, computing the low- and moderate- those dwelling units financed by mortgages units rather than the entire stock of all income market share would be in a particular year that are occupied by mortgaged dwelling units; that is, the market- straightforward, consisting of three steps: households with incomes below the area share concept is based on the mortgage flow Step 1: Projecting the market shares of the median). Step 3: Multiplying the four percentages in in a particular year, which will be smaller four major property types included in the (2) by their corresponding market shares in than total outstanding mortgage debt. Third, conventional conforming mortgage market, (1), and summing the results to arrive at an the low- and moderate-income market is i.e.— (a) Single-family owner-occupied dwelling estimate of the overall share of dwelling units expressed relative to the overall conforming units (SF–O units); financed by mortgages that are occupied by conventional market, which is the relevant (b) Rental units in 2–4 unit properties low- and moderate-income families. 7 market for the GSEs. The low- and where the owner occupies one unit (SF 2–4 The four property types are analyzed moderate-income market is defined as a units); 8 separately because of their differences in percentage of the conforming market; this (c) Rental units in one-to-four unit low- and moderate-income occupancy. percentage approach maintains consistency investor-owned properties (SF Investor Rental properties have substantially higher with the method for computing each GSE’s units); and, percentages of low- and moderate-income performance under the Low- and Moderate- (d) Rental units in multifamily (5 or more occupants than owner-occupied properties. Income Goal (that is, the number of low- and units) properties (MF units).9 This can be seen in the top portion of Table D.1, which illustrates Step 3’s basic formula moderate-income dwelling units financed by Step 2: Projecting the ‘‘goal percentage’’ for for calculating the size of the low- and GSE mortgage purchases relative to the each of the above four property types (for example, the ‘‘Low- and Moderate-Income moderate-income market.10 In this example, overall number of dwelling units financed by Goal percentage for single-family owner- low- and moderate-income dwelling units are GSE mortgage purchases). estimated to account for 53.9 percent of the 8 The owner of the SF 2–4 property is counted in total number of dwelling units financed in 7 So-called ‘‘jumbo’’ mortgages, greater than (a). the conforming mortgage market. $333,700 in 2004 for 1-unit properties, are excluded 9 Property types (b), (c), and (d) consist of rental in defining the conforming market. There is some units. Property types (b) and (c) must sometimes be 10 The property shares and low-mod percentages overlap of loans eligible for purchase by the GSEs combined due to data limitations; in this case, they reported here are based on one set of model with loans insured by the FHA and guaranteed by are referred to as ‘‘single-family rental units’’ (SF– assumptions; other sets of assumptions are the Veterans Administration. R units). discussed in Section E.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00241 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63820 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
To examine the other housing goals, the data set for calculating either the property estimate the number of single-family units ‘‘goal percentages’’ in Step 2 would be shares or the housing goal percentages does financed in the conforming conventional changed and the new ‘‘goal percentages’’ not exist. However, there are several major market, HUD had to project certain market would be multiplied by Step 1’s property data bases that provide a wealth of useful parameters based on its judgment about the distribution, which remains constant. For information on the mortgage market. HUD reliability of different data sources. Sections example, the Underserved Areas Goal 11 combined information from the following D and E report HUD’s findings related to the would be derived as illustrated in the bottom sources: the Home Mortgage Disclosure Act single-family market. portion of Table D.1. In this example, units (HMDA) reports, the American Housing Total market originations are obtained by eligible under the Underserved Areas Goal Survey (AHS), HUD’s Survey of Mortgage adding multifamily originations to the single- are estimated to account for 31.4 percent of Lending Activity (SMLA), the Census family estimate. Because of the wide range of the total number of dwelling units financed Bureau’s AHS-based Property Owners and estimates available, the size of the in the conforming mortgage market.12 Managers Survey (POMS), and the Census multifamily mortgage market turned out to be Bureau’s recent 2001 Residential Finance one of the most controversial issues raised 3. Data Issues Survey (RFS). In addition, information on the during the initial rule-making process during Unfortunately, complete and consistent mortgage market was obtained from the 1995; this was also an issue that the GSEs mortgage data are not readily available for Mortgage Bankers Association, Fannie Mae, focused on in their comments on the 2000 carrying out the above three steps. A single Freddie Mac and other organizations. final rule and their comments on the 2004 Property Shares. To derive the property proposed GSE rule. Because most renters shares, HUD started with forecasts of single- qualify under the Low- and Moderate-Income 11 This goal will be referred to as the ‘‘Underserved Areas Goal’’. family mortgage originations (expressed in Goal, the chosen market size for multifamily dollars). These forecasts, which are available can have a substantial effect on the overall 12 The example in Table D.1 is based on 1990 Census tract geography. As explained in Section G, from the GSEs and industry groups such as estimate of the low- and moderate-income switching to 2000 Census tract geography the Mortgage Bankers Association, do not market (as well as on the estimate of the (scheduled for 2005) increases the underserved provide information on conforming special affordable market). Thus, it is areas market share by approximately five mortgages, on owner versus renter mortgages, important to consider estimates of the size of percentage points. or on the number of units financed. Thus, to the multifamily market in some detail, as
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00242 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.076 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63821
Section C does. In addition, given the market estimates (rather than precise point units financed were multifamily rental units. uncertainty surrounding estimates of the estimates) for each of the housing goals. Of the GSEs’ purchases qualifying as special multifamily mortgage market, it is important affordable mortgages during this period, 25 to consider a range of market estimates, as C. Size of the Conventional Multifamily 13 percent of the units financed were Sections F–H do. Mortgage Market multifamily rental units. Goal Percentages. To derive the goal This Section C differs from the version The methods used in the 2000 Rule for percentages for each property type, HUD published in the May 3, 2004, Proposed Rule estimating the size of the multifamily relied heavily on HMDA, AHS, POMS and in the following ways: The estimates from the mortgage market and related variables were RFS data. For single-family-owner ‘‘HUD New’’ and ‘‘Flow of Funds’’ methods the product of extensive research by HUD originations, HMDA provides comprehensive discussed below in parts 2 and 3 have been and review by interested parties. The information on borrower incomes and census updated through 2003, and responses to approach here is first to extend those tract locations for metropolitan areas. comments received on those methods have estimates through 2002 using the same Unfortunately, it provides no information on been added to those sections. The part titled methods as in the 2000 Rule, and then to the incomes of renters living in mortgaged ‘‘Most Likely Range’’ has been revised in present alternative methods, along with properties (either single-family or light of the 2003 estimates and comments commentary. multifamily) or on the rents (and therefore received. The discussion of ‘‘Loan Amount 1. Data Sources the affordability) of rental units in mortgaged per Unit,’’ part 5, has been revised in properties. The AHS, however, does provide response to comments and to newly available The data sources available for estimating a wealth of information on rents and the data from the GSEs and the 2003 American the size of the multifamily mortgage market affordability of the outstanding stock of Housing Survey. The multifamily mix are more limited in scope and timeliness single-family and multifamily rental discussion, part 6, has been revised in than was the case for the 2000 Rule. Among properties. An important issue here concerns accordance with other changes. Section C.7 the key sources described in detail in the whether rent data for the stock of rental has been added on the multifamily mix as 2000 Rule, the following are now less useful: properties can serve as a proxy for rents on estimated from the newly released 2001 Survey of Mortgage Lending Activity. This newly-mortgaged rental properties. During Residential Finance Survey (RFS). Lastly, survey has been discontinued; estimates are the 2000 rule-making process, POMS data Section C.8 discusses the multifamily mixes available only through 1997. were used to examine the rents of newly- that will be examined in HUD’s projection Residential Finance Survey: The 1991 mortgaged rental properties; thus, the POMS model for 2005–2008. Other than these Residential Finance Survey (RFS) is now 13 data supplements the AHS data. The recently changes and minor editorial corrections, the years out of date. (See Section C.7 for results released RFS provides information on text in this section is identical to that in the from the 2001 RFS.) property shares (e.g., the relative importance Proposed Rule published May 3, 2004. Urban Institute Statistical Model: This of rental versus owner properties) and several Changes to Tables D.2 through D.5 are noted model, developed in 1995 and calibrated other important parameters in HUD’s market in the text and table notes. The old Table D.5 using data from 1975–1990, is now even model. The data base issues as well as other is now D.5a and Tables D.5b and D.5c have further removed from its calibration period technical issues related to the goal been added. and probably captures current market percentages (such as the need to consider a This section provides estimates of (a) the conditions less accurately. Estimates from the GSEs: As part of their range of mortgage market environments) are annual dollar volume of conventional comments on the proposed 2000 Rule, discussed in Sections F, G, and H, which multifamily mortgage originations and (b) the Fannie Mae and Freddie Mac shared with present the market share estimates for the annual average loan amount per unit HUD their own estimates of the size of the Low- and Moderate-Income Goal, the financed. The estimates build on research multifamily mortgage market. Underserved Areas Goal, and the Special reported in the Final Rule on HUD’s Affordable Goal, respectively. Regulation of Fannie Mae and Freddie Mac Fortunately, several key sources are available 4. Conclusions as published in the Federal Register on with the timeliness and quality comparable October 31, 2000, especially in Appendix D. to the sources used during development of HUD is using the same basic methodology the 2000 Rule. These sources are: The Home for estimating market shares that it used in That material from the 2000 Rule will not be repeated here but will be referenced or Mortgage Disclosure Act (HMDA); activity its 1995 and 2000 final rules and its 2004 reports submitted to HUD and the Office of proposed rule. As demonstrated in the summarized where appropriate. This section uses the information on dollar Federal Enterprise Oversight (OFHEO) by remainder of this appendix, HUD has Fannie Mae and Freddie Mac; non-GSE attempted to reduce the range of uncertainty volume of multifamily originations and average loan amounts to estimate the number mortgage-backed security issuance from the around its market estimates by carefully Commercial Mortgage Alert database; and reviewing all known major mortgage data of multifamily units financed each year as a percentage share of the total (both single- multifamily mortgage activity by life sources, by considering comments on the insurance companies, as estimated by the 2004 proposed rule, and by conducting family and multifamily) number of dwelling units financed each year. This percentage American Council of Life Insurers (ACLI). numerous sensitivity analyses to show the For background information on each of these effects of alternative assumptions. Sections C, share, called the ‘‘multifamily mix’’, is an important parameter in HUD’s projection sources, readers are referred to Appendix D D, and E report findings related to the of the 2000 Rule. property share distributions called for in Step model of the mortgage market for 2005–08 1, while Sections F, G, and H report findings (see Section C.8 below) 2. Estimates Based on ‘‘HUD New’’ related to the goal-specific market parameters Estimating this ‘‘multifamily mix’’ is Methodology called for in Step 2. These latter sections also important because relative to its share of the In the 2000 Rule, HUD developed a new report the overall market estimates for each overall housing market, the multifamily methodology for estimating aggregate housing goal calculated in Step 3. rental sector has disproportionate importance multifamily conventional loan originations. In considering the levels of the goals, HUD for the housing goals established for Fannie The method, here labeled ‘‘HUD New’’, was carefully examined comments by the GSEs Mae and Freddie Mac. This is because most developed to make full use of the available and others on the methodology used to multifamily rental units are occupied by data, and in particular the four sources listed establish the market share for each of the households with low or moderate incomes. above, which encompass most of the goals. Based on that thorough evaluation, as Between 1999 and 2002, for example, the multifamily mortgage market. well as HUD’s additional analysis for this GSEs purchased mortgages on approximately The advantages of HUD New are that it final rule, HUD concludes that its basic 26.1 million housing units, of which only 9.5 provides reasonably complete coverage of the methodology is a reasonable and valid percent were multifamily rental units. market, produces those estimates within nine approach to estimating market shares. As in However, of the GSEs’ purchases qualifying months of the end of the year, generally the past, HUD recognizes the uncertainty as mortgages on low- and moderate-income includes only current originations and avoids regarding some of these estimates, which has housing during this period, 18 percent of the double counting. The main disadvantage of led the Department to undertake a number of HUD New is that it produces a lower bound sensitivity and other analyses to reduce this 13 This section is based on analysis by Jack estimate. Some loan originators are missed, uncertainty and also to provide a range of Goodman under contract with the Urban Institute. including pension funds, government entities
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00243 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63822 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
at the federal, state, and local levels, real sensitivity analysis can be performed to show The estimates from HUD New are estate investment trusts, and some mortgage the effects of different multifamily presented in Table D.2. This table is the bankers. Also, excluded are loans made by origination volumes on the goals qualifying counterpart of Table D.5 in the 2000 Rule. private individuals and partnerships. In market estimates (see Sections F–H). Due to The historical years have two columns each, addition to these exclusions, estimates from the reasonableness of the HUD New one for the estimates presented in the 2000 the covered lenders require some judgmental approach, the value of maintaining Rule and one for estimates independently adjustments to conform to the definitions and continuity in estimation methods, and the produced as part of this research. Footnotes time intervals of HUD New. fact that no data has become available in the Despite these limitations, HUD New is one past few years that would argue for to the table provide more complete sound way to estimate the size of the modifying HUD New, it is used here for the descriptions of the components. Additional multifamily conventional mortgage market. baseline estimate of the size of the background on the calculations is provided Although the method requires unavoidable conventional multifamily mortgage market in in the 2000 Rule (Appendix D, Section C). judgment calls on which analysts may differ, 2000, 2001, 2002 and 2003. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00244 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63823
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00245 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.077
The revisions to the historical estimates lending. Total conventional originations, of multifamily mortgage credit. The Flow of (i.e., those in the 2000 Final Rule) result from estimated at $89 billion, are up 32 percent Funds statistics refer to net changes in credit both revisions to some of the input data and from 2002, easily reaching a new record high. outstanding rather than gross originations. recalculations. For the years 1995 through A large increase was observed in each of the Specifically, balance sheet estimates of 1998, the revisions are small for the estimates five market segments listed in Table D.2. mortgage assets of lenders are used to of total originations. The only one of note is Several organizations commented on the produce estimated changes in holdings of a 5 percent upward revision to the estimate HUD New method. Fannie Mae says it mortgages over time. An alternative label for for 1995, prompted by a recalculation of the involves double counting of originations. the resulting time series is ‘‘net change in entry for life insurance companies. The However, the one example they offer— mortgage debt outstanding.’’ revision to 1999 is larger, and results mostly between life insurance company data and The historical relationship between gross from the substitution of the actual HMDA CMBS data—should not be subject to double originations and net change can be used to results for that year for the projected value counting because securitizations by life estimate recent origination volume. Separate used in the 2000 Rule. Surprisingly, the insurance companies are deleted from the information on FHA multifamily activity can revised estimate for 2000 based on complete CMBS totals, as noted in Table D.2 and in be used to convert the total originations to data for that year only varies slightly from the documentation included in the 2000 Rule. estimates of only conventional originations. projection made at the time of the 2000 Rule. Freddie Mac, through its contractor, uses an The Flow of Funds method that is described Most of the historical estimates produced in approach similar to HUD New but uses in this section will be called ‘‘FoF-based.’’ 2000 can be replicated or closely different data sources. Inadequate details are Flow of Funds estimates of mortgage debt approximated, including those for Fannie provided on the tabulations and judgments outstanding are based on data from sources and Freddie, CMBS, HMDA, and life applied to evaluate the method. Lastly, MBA of varying accuracy and timeliness. Bank and insurance companies. The replicability of the expresses a preference for the estimates thrift institution holdings, taken from CMBS figures is especially important, in light provided by HUD New and says, without regulatory filings, are by all accounts highly of all the selection criteria and hand providing detail, that estimates developed by accurate, as are those from the government calculations required to generate those their consultants are similar to those sponsored agencies and direct Federal estimates from the CMBS database. (In the presented in HUD New. government holdings. The private MBS data 2000 Rule, the estimates for Freddie Mac and The comments received fail to note the CMBS originations in 1997 appear to have and the life insurance company figures, both point made repeatedly in the proposed rule taken from Wall Street sources, are also been switched, and the revised estimates text that the HUD New estimates are lower make this correction.) thought to be reasonably accurate. Less bounds on the volumes of originations. While accurate are the estimates of loans made by The revised figures for 1999 and 2000 HUD New is characterized in the proposed indicate that total conventional originations private individuals and certain institutions, rule as providing ‘‘* * * the baseline for which comprehensive data on loans dropped 8 percent in 1999 from 1998’s very estimate of the size of the conventional strong level and another 13 percent in 2000. outstanding is provided only once every ten multifamily mortgage market * * *’’, other years, through the Residential Finance However, the HUD New estimate indicates language in the rule makes clear that Survey. Fortunately, the depository that total conventional originations then ‘‘baseline’’ is used in the sense of ‘‘starting institutions, GSEs, and mortgage-backed jumped 40 percent in 2001 and further point.’’ For example, the proposed rule also securities account for the bulk of all holdings increased 15 percent in 2002. Judging from states that ‘‘* * * unavoidable gaps in of mortgage debt (approximately 72 percent, Survey of Mortgage Lending Activity coverage make the resulting HUD New according to the Flow of Funds estimates for estimates since 1970, the 2002 number is a figures lower-bound estimates of actual year-end 2001). Thus, most of the Flow of new record high. For 2002, most of the originations rather than best ‘point’ Funds data are from highly accurate sources. increased volume is due to increases by estimates’’ (p. 24450). HMDA lenders and life insurance companies. The net change in mortgage debt One possible concern is that the significant 3. An Alternative Method outstanding in any year is the lower bound increase in the HMDA number in 2002 was The HUD New method makes use of all the on originations. This is because the net caused by the FFIEC relaxing its eligibility available sources of data on individual change is defined as originations less the sum requirements between 2001 and 2002. This origination sources in attempting to estimate of principal repayments and charge offs. concern turns out to be unfounded. The total conventional mortgage originations. Historically loan originations have exceeded FFIEC actually raised its eligibility However, as discussed in the 2000 Rule and the net change by a considerable margin in requirements. The level of assets required by summarized above, unavoidable gaps in both the multifamily and single-family FFIEC to be reported to HMDA increased coverage make the resulting HUD New markets. There are several reasons why the from $31 million in 2001 to $32 million in figures lower-bound estimates of actual relationship of originations to net change 2002. In addition, the number of HMDA originations rather than best ‘‘point’’ differs between the multifamily and single- reporters decreased from 7,771 in 2001 to estimates. In addition, even for those loans family sectors, but the basic principles apply 7,638 in 2002. that are available, certain assumptions must to both sectors. Compared with the version of Table D.2 in be made to convert the available data into Table D.3 presents the annual estimates the Proposed Rule of May 3, 2004, the estimates corresponding to the desired from the Flow of Funds. Also shown are the version here updates the estimates through definition and time periods. An alternative to estimates of multifamily conventional 2003 and revises the 2001 and 2002 estimates the bottom-up approach of HUD New avoids originations as published in Table D.10 from slightly in response to newly available data. some of the data problems. The Federal the 2000 rule, and FHA originations from The data for 2003 point to a large, broad- Reserve’s Flow of Funds accounts provide HUD administrative records. based increase in the volume of multifamily the most complete and timely set of estimates BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00246 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63825
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00247 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.078 63826 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
The ratio of mortgage originations to net the added refinancing incentive in 2001. By provided even stronger incentives for change should be positively correlated with the beginning of 2001, there were relatively refinancing. As shown in the last columns of the proportion of total originations that are few properties ‘‘at risk’’ of refinancing. Many Table D.3, ten-year Treasury yields in 2003 refinancings, for which the net change in presumably had refinanced in one of the averaged about 60 basis points below those mortgage debt would be expected to be low preceding years, and lock-out provisions, of 2002, and approximately 130 basis points relative to that on loans taken out in yield maintenance agreements, and other below the average of the previous five years. connection with a property acquisition. (This loan conditions may have kept these Comments on the Flow of Funds method is the pattern observed in the single-family properties from coming in for refinancings. for estimating multifamily originations mortgage market.) Refinancings, in turn, Also, there may have been some short-run focused on the approach to converting net would be expected to be prevalent relative to capacity problems in the multifamily loan change into loan originations. Fannie Mae purchase loans at times when interest rates origination industry in 2001 that further argued that it was preferable to convert by are low relative to their recent past. curtailed volume. applying a liquidation rate to the stock of The historical evidence generally supports Applying the 1.5 multiple to 2001’s net mortgage debt and deriving originations as this expectation regarding the relationship of change of $48 billion yields a total net change plus estimated liquidations. A originations to net lending. As shown in originations estimate of $71 billion. trade organization noted the historical Table D.3, total originations have been Subtracting FHA business results in a instability of the ratio of originations to net highest relative to net change when interest conventional originations estimate of $67 change and argued that the ‘‘HUD New’’ rates have been low relative to their recent billion, to which a subjective confidence approach to estimating originations was past. [Note: Columns A1, B1, C1, and D1 are band of at least ±$2 billion appears superior. Freddie Mac and its consultant, the figures appearing in the Proposed Rule warranted. while not commenting directly on the Flow version of this table. Owing to extensive As seen in Table D.3, the Flow of Funds of Funds method, expressed a preference for revisions to the input data, new columns methodology indicates that total a modified version of HUD New, as described with the revised inputs and calculated values conventional originations decreased 6.5% in the previous part of this section. have been added to facilitate comparisons. between 2001 and 2002. In 2002, the net The most recent data suggest that These revised figures appear in Columns A2, change in mortgage debt decreased slightly to originations may in fact have been higher B2, C2, and D2.] The ten-year Treasury yield, $44 billion. Using the 1.5 multiple for 2002’s than estimated in the Flow of Funds a common benchmark for pricing multifamily net change of $44.2 billion yields a total approach and that the 1.5 multiplier used to mortgages, has generally trended down since originations estimate of $67 billion. convert net change into originations is too 1990. The early 1990s were all marked by Subtracting $4.5 billion of FHA business low. The reason is that in both 2002 and high originations relative to net change, and results in a conventional originations 2003, the 1.5 multiplier results in estimated these were also years in which interest rates estimate of $62 billion. conventional originations that are less than were particularly low relative to their trailing This Flow of Funds estimate is over $5 those produced by the HUD New method. As five-year averages. In 1996 and 1997, by billion less than the estimate from HUD New. discussed earlier, HUD New provides a lower contrast, originations were less high relative This is surprising given that the HUD New bound estimate. Fannie Mae’s lower to net change, and these were years in which method is supposed to serve as a lower estimates of originations in recent years, interest rates were only slightly lower than boundary on the size of the multifamily relative to those in the proposed rule, result their five-year trailing averages. In estimating market, while the Flow of Funds method is from the liquidation rate used in the conventional originations for 1999–2002, the designed to produce a higher ‘‘point’’ calculation, which is that from Fannie Mae’s 1998 experience is a useful benchmark. That estimate of the actual size of the market. own portfolio. But Fannie Mae’s liquidation year, total originations exceeded the net Like the estimates for HUD New, those for rate would be expected to fall below the change by about 80 percent, as shown in the Flow of Funds method have been revised market wide average, because Fannie Mae’s Table D.3. There was also a big drop in and updated through 2003 to incorporate multifamily business has been growing more interest rates in 1998 relative to the recent new data. As with HUD New, the Flow of rapidly than the market overall, and as a past, providing an incentive for refinancings. Funds method suggests a large increase in result its loans presumably on average are As shown in the table, interest rates rose conventional mortgage lending in 2003. The ‘‘younger’’ and consequently less likely to slightly in 1999 and again in 2000, estimate for conventional originations in prepay or be retired than are the loans in the presumably diminishing the incentive to 2003 is $75 billion, up 29 percent from the market as a whole. Lastly, regarding the refinance. Nonetheless, the net change in revised estimate for 2002. In percentage historical instability of the ratio of mortgage debt was higher in 1999 and 2000 terms, the increase in 2003 almost matches originations to net change noted by a trade than it had been in 1998. that of the HUD New method’s estimates of organization, Table D.3 of the proposed rule Putting all this together, it seems that the Table D.2. also presented the annual difference between appropriate ratio of total originations to net The originations estimates for earlier years, originations and net change, which is change to apply to 1999 and 2000 would be and especially 2000–2002, have been revised considerably more stable. The differences below that of 1998 and of most other years downward in response to revisions by the corresponding to the 1.5 multiplier for the of the 1990s. Applying a ratio of 1.5 to the Federal Reserve to the Flow of Funds past several years are, as shown in D.3, below the historical averages. This is additional net change estimates in 1999 and 2000 accounts and by an update to HUD’s FHA evidence that the 1.5 multiplier is perhaps results in a total originations estimate of estimate for 2002. The downward revision too low. approximately $56 billion. Subtracting the $4 was largest for 2000, for which year the new billion in FHA originations results in figure of $44 billion of conventional 4. Most Likely Range estimates of $52 billion for conventional originations is $8 billion less than the earlier In the 2000 Rule, estimates of conventional originations in each year. A subjective estimate. multifamily loan originations from various confidence band around this point estimate The big increase in estimated originations sources and methods were evaluated in is at least +/¥$2 billion. in 2003 is largely the result of the Federal determining the most likely range of annual Turning to the estimate for 2001, the first Reserve’s estimate of a large increase that originations. Those estimates were thing to note is that net change in mortgage year in net change in mortgage debt summarized in Table D.10 in the 2000 Rule. debt jumped to $48 billion from $37 billion outstanding, shown in column A2 of Table Some of the estimates from that table are of the previous two years. The second thing D.3 The increase in 2003 in the Flow of reproduced below, in Table D.4, along with to note is that interest rates fell by nearly a Funds accounts is likely to be fairly accurate, updates and estimates from the Flow of percentage point in 2001 relative to their past because almost all of it is attributable to Funds method. average. For both of these reasons, total holder types for which the Fed has reliable Both HUD New (column #4 in Table D.4) originations in 2001 would be expected to statistics, specifically depository institutions and FoF-based (column #9) indicate a surge have been higher than in 1999 or 2000. How and GSE mortgage securities. As in 1999– in lending activity in 2001. Some much higher is a subjective judgment, but 1.5 2002, in 2003 the net change was converted corroboration of this jump is provided by would seem an appropriate multiple to apply into total originations by applying a other indicators, flawed though they may be. to the net change number in 2001. This is the multiplier of 1.5, under the assumption that HMDA has well-documented coverage same multiple as in 1999 and 2000, despite the continued decline in interest rates problems with multifamily loans, but it is
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00248 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63827
noteworthy that HMDA-estimated commitments, described in the 2000 Rule trend reflects changes in both market size conventional originations stayed in the same and updated here in column #5, also follows and GSE market share. FHA originations (not general range ($26 to $31 billion) in 1998– this basic path. Similarly, aggregate GSE shown) also rose substantially in 2001, but 2000 before jumping to $36 billion in 2001. multifamily purchases and securitizations this too may indicate more than just market The composite of 1.25 times HMDA stayed in the same general level in 1998– size trends. originations plus life insurance 2000, before jumping in 2001, although this BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00249 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.079
Column #11 of Table D.4 gives the likely significant amount of refinancing activity. In likely range, which the Department has set at ranges of originations for each of the years. such a case, there could be an increase in the $85 billion to $100 billion. As explained in These are based on the estimates from all size of the multifamily market without a Section C.6 below, HUD will conduct sources and interpretations of their strengths corresponding increase in net mortgage debt sensitivity analyses in Sections F–H showing and weaknesses. In 1999, the $4 billion outstanding. A higher multiple would need the effects of different multifamily mixes on upward revision to the HUD New estimate to be applied to the Flow of Funds model to the historical estimates of the goals- from the preliminary figure reported in the compensate for the increase in multifamily qualifying shares of the mortgage market. 2000 Rule, together with the higher estimate refinancings. produced by the FoF-based method, justify Due to data limitations, the above remains 5. Loan Amount per Unit an upward revision to the $45–$48 range a speculation. The largest increase in In determining the size of the conventional estimated in the 2000 Rule. The revised range multifamily volume came from HMDA multifamily mortgage market for purposes of is set at $50–54 billion. In 2000, HUD New reporting lenders. The HMDA data do not the GSE rules, the measure of market size is (revised and extended version) suggests that allow for the separation of multifamily the annual number of conventionally originations were somewhat lower than in purchase originations from refinancings. financed multifamily rental housing units. 1999, but FoF-based has originations holding Other data sources need to be explored to at $52 billion. Balancing these conflicting determine if an adjustment to the FoF-based The number of units is derived by dividing indicators, a range of $48–$52 billion is model is appropriate. the aggregate annual originations by an selected for 2000. Finally, all indicators point Both HUD New and the FoF-based method estimate of the average loan amount per to a substantial pickup in 2001, and the range indicate a large increase in conventional housing unit financed. For this reason, that seems to fit best with those indicators is multifamily loan originations in 2003. But accuracy in the estimate of loan amount per $65–$69 billion. the FoF estimates for each of the previous unit is as important as accuracy in the dollar In 2002, the various methods of estimation four years have been revised downward in estimate of aggregate conventional give a mixed picture. HUD New indicates a light of revised input data. According to originations. A 10 percent error in either will surge in lending activity in 2002, while the these updated and revised estimates, result in a 10 percent error in the estimate flow of funds method shows a decrease in conventional multifamily originations by of market size. lending activity. Other methods also show HUD New have exceeded the estimates of The 2000 Rule used estimates of loan divergent trends. The composite of 1.25 times FoF in two of the past five years, and in the amount per unit drawn from various sources. HMDA originations plus life insurance other three years FoF exceeded HUD New by As summarized in Table D.9 of the 2000 Rule commitments also shows a significant only narrow margins. Because HUD New and the accompanying text, the estimates for increase between 2001 and 2002. On the produces lower bound estimates of 1993–1998 were taken from the GSEs and for other hand, aggregate GSE multifamily originations, whereas FoF is intended to 1999 from CMBS data. ‘‘Unpaid Principal purchases and securitizations showed a provide best point estimates, the Department Balance’’ or UPB—a balance sheet measure slight decrease between 2001 and 2002. FHA concludes that the 1.5 multiplier applied in which for current year loan originations will originations (not shown) also decreased the FoF method is too low, and as a result differ little from the initial loan amount—is slightly in 2002. the FoF estimates understate originations in used to calculate aggregate originations of While this is a subjective judgment, 1.5 the past several years. In light of this loans bought or securitized by the GSEs or may not be the appropriate multiple to apply probable underestimate of the multiplier, and to net mortgage debt outstanding in the flow after consideration of comments received, the pooled into non-GSE mortgage-backed of funds model in 2002. The difference Department believes that the likely ranges of securities. The figures from Table D.9 of the between the flow of funds estimate and the conventional originations for 2002 and 2000 Rule are reproduced below in Table HUD estimate cannot be reconciled without earlier years as published in the May, 2004, D.5a, along with updated estimates from all adjusting the FoF multiple. Given the low Proposed Rule continue to be reasonable three sources for 2000, 2001 and 2002. The interest rates in 2002, and a refinancing boom estimates, although likely on the conservative estimates that are new since the 2000 Rule in the single-family mortgage market, it could side. As for 2003, the estimates from HUD appear in italics. be that the multifamily market also had a New and FoF indicate a substantially higher BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00250 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63829
BILLING CODE 4210–27–C
VerDate jul<14>2003 19:40 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00251 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.080 63830 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Several options are available for 2002 ...... 39,787 In the first 10 months of 2002, CMBS developing estimates for 2000, 2001 and properties showed a UPB/unit of $37,038, a 2002. The first is to use the UPB (unpaid The figure for 2002 is approximately 46 nearly 14 percent jump over the previous principal balance) per unit estimates from the percent higher than in 1997. Both Fannie year. Although slightly below the UPB/unit GSEs. These estimates, taken from the Fannie Mae and Freddie Mac’s portfolios generate for the GSEs, the CMBS numbers are closer Mae and Freddie Mac annual activity reports estimates of between $39,000 and $40,000 for to the GSE calculations than in previous to HUD, are as follows, computed as in the 2002. years. 2000 Rule as a unit-weighted average of the Several alternative approaches to Another approach is to move the 1999 unpaid principal balance (UPB) per estimating loan amount per unit are estimate of UPB/unit forward by some multifamily unit in Fannie Mae’s and available. The first is to base the estimate on justifiable index. The 2001 estimates use the Freddie Mac’s portfolios: CMBS data, as was done for 1999 in the 2000 change in average rent on multifamily rental 1997 ...... $27,266 Rulemaking. As shown in the last column of units from the American Housing Survey. 1998 ...... 31,041 Table D.5, the estimates of UPB/unit from Because AHS data are not available for 2002, 1999 ...... 35,038 this source are somewhat below those of the the 2002 estimate uses the consumer price 2000 ...... 37,208 GSEs and indicate less increase since the late index for rent of primary residence. Both 2001 ...... 37,258 1990s. AHS and CPI rent estimates are listed below:
Year Median Mean CPI
1999 ...... $550 $592 177.5 2001 ...... 590 647 192.1 2002 ...... N/A N/A 199.7
There is some variation between the two limited or negative recent growth in UPB/ originations and UPB per unit, the revisions measures. In the AHS, median rent rose 7.3 unit. The median and mean rents for 2003 would be at least partially offsetting, with percent over this two-year period, and mean that correspond to those in the table above little net effect on the historical estimates of rent increased 9.3 percent. Meanwhile, the are $609 and $671. Given the logic of this number of multifamily units financed. As for CPI showed an increase of 8.2 percent. In method as described in the proposed rule, it 2003, weighing all available information, the 2001, using the AHS produces an estimate of seemed most appropriate to use the percent Department has set the UPB/unit at $39,082, $34,000. The CPI yields a smaller estimate for increase in AHS rents from 1999 to 2003 to the weighted average of the GSEs’ actual 2001; applying the 8.2 percent increase from update the 1999 UPB/unit ($30,719) to a 2003 UPB/unit for that year. As explained in the the CPI results in a 2001 estimate of $33,200. figure. Using the 13.3 percent increase in next section, goals-qualifying estimates for Since the AHS data are unavailable in 2002, mean rent between 1999 and 2003 (the 1995–2002 are reported in Sections F–H that the CPI provides a 2002 estimate of increase in median was only 10.7 percent) include multifamily mixes approximately approximately $35,000. and moving the baseline UPB/unit from 1999 two-three percentage points lower that the In 2001, the rent-adjusted 1999 estimate forward to 2003 by this proportion brings the multifamily mixes suggested by the most was in between the estimates from the CMBS 2003 UPB/unit to $34,805. That is the likely range of multifamily dollar estimates and GSE data, and was a fair estimate of the number appearing in Table D.5a. For and the UPB/unit estimates. actual size of the market. In 2002, however, comparison, the CPI rent index rose 15.8 the rent-adjusted number is below both the percent between 1999 and 2003. 6. Multifamily Mix During the 1990s CMBS and GSE calculations. The rent- In commenting on HUD’s UPB/unit This section uses the information on dollar adjusted number could be underestimating estimates for 2000–2002, as published in the volume of multifamily originations (Table the 2002 UPB/unit. Either the CMBS or GSE May 2004 Proposed Rule, both Fannie Mae D.4) and average loan amounts (Table D.5a) calculations, or an average of the various and Freddie Mac expressed the view that the to estimate the number of multifamily units methods could be used. Sections F–H will estimates were too low. They cited both their financed each year as a percentage share of report the results of sensitivity analyses own experience and other evidence and the total (both single-family and multifamily) showing the effects of the different argued that HUD’s reliance on CMBS and number of dwelling units financed each year. multifamily mortgage estimates and different rent data, and switching of benchmark years, Because of the high goals-qualifying shares of per unit amounts on the goals-qualifying resulted in UPB/unit estimates that were multifamily housing, the multifamily mix is shares for the year 2002. Under the various substantially below the actual market an important parameter in HUD’s projection estimates, the multifamily mix (defined averages. model for the overall market; other things below) for 2002 varies from 9.5 percent–11 In reviewing the comments and in light of equal, a higher multifamily mix (or percent. these new data HUD has concluded that the conversely, a lower share of single-family Since the proposed rule was issued by the estimates in the proposed rule likely were too loans) leads to a higher estimate of goals- Department, data for 2003 have become low. The more difficult determination is qualifying loans in the overall mortgage available that permit updates of some of the where to set the estimates. The Department market. This percentage share, or sources of UPB/unit estimates. The GSEs’ has not revised its estimate of UPB/unit for ‘‘multifamily mix’’, is reported in the last two experience, shown in the bottom row of 2002 and earlier years, because of this columns of Table D.4 for the years 1991 to Table D.5a, was mixed. Fannie Mae’s UPB/ uncertainty. The situation is similar to that 2002.14 The ‘‘minimum’’ (‘‘maximum’’) unit increased about 4 percent from 2002, but discussed in the previous part of this section multifamily mix figure reflects the low Freddie Mac’s dropped 9 percent. The in discussing the likely range of conventional (upper) end of the ‘‘likely range’’ of volume-weighted average UPB/unit for the multifamily originations, where the new data multifamily dollar originations, also reported GSEs in 2003 was $39,082, off about 2 lead the Department to think the Flow of in Table D.4. Because they will be compared percent from the 2002 average of $39,787 Funds estimates may be too low, but no with other estimates of the MF mix, these shown in the text table above. adjustments were made to the likely range as ‘‘likely range’’ data are reproduced in the first The most recent rent estimates from the reported in Table D.4. If adjustments were two columns of Table D.5b. American Housing Survey also suggest made to the historical estimates of BILLING CODE 4210–27–P
14 1990 is excluded from this calculation because Also, the estimated multifamily mix from the HUD greater than the estimate from the Flow of Funds of the unusually high multifamily mix that year. New Method is also provided for 2002 since it was method.
VerDate jul<14>2003 19:40 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00252 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63831
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00253 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.081 63832 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Table D.5b includes several averages of the multifamily mixes less than the previous 2002, from 15.1 percent to 17.5 percent for MF mix for different time periods between lows of 11 percent in 1992 and 2002. recent home purchase years, and from 11.2 1991 and 2002. Based on the ‘‘likely range’’ As discussed earlier, several commented percent to 12.9 percent for the refinance of annual conventional multifamily that HUD had understated the UPB/unit, years of 1998, 2001, and 2002. origination volume, multifamily units have which caused HUD to overstate the share of The impact of the lower MF mix on the represented 15.3 percent (the average of the newly-mortgage multifamily dwelling units. UPB/unit assumption can be illustrated for ‘‘minimum’’ figures) to 16.6 percent (the Section C.5 explains that HUD’s UPB/unit average of the ‘‘maximum’’ figures) of units estimates for recent years are likely too low the case of 2001, which assumed a loan- financed each year between 1991 and 2002. but that could be offset by low estimates of amount-per-unit figure of $34,000. Reducing Considering the mid-points of the ‘‘likely originations. To allow for different views the MF mix from 13.5 percent to 12.0 percent range’’, the multifamily mix averaged 15.9 about the volume of mortgage originations is consistent with increasing the UPB/unit percent during this period. Notice that the and the UPB/unit, Sections F–H will conduct from $34,000 to $39,075 (holding constant multifamily mix is lower during years of sensitivity analyses with lower multifamily mortgage originations at $67 billion). Of heavy refinancing when single-family mixes than suggested by the mid-points of course, the lower MF mix of 12.0 percent is originations dominate the mortgage market; the likely ranges in Table D.5b. The third consistent with a lower volume of mortgage the multifamily mix was only 13–14 percent column of Table D.5b lists the ‘‘mid-point’’ originations if the initial UPB/unit of $34,000 during 1993, 1998, and 2001, and 11 percent MF mixes while the fourth column of Table is retained. (or less) during 2002.15 As discussed in D.5b lists the lower MF mixes used in Sections F–H, record single-family Sections F–H. Over the 1995–2002 period, Fannie Mae (op.cit., page I–29) developed originations ($3.8 trillion) during 2003 likely the average MF mix ranged from 13.9 percent three sets of UPB-per-unit figures for 1997 to resulted in that year having a lower (the lower MF mix approach) to 16.2 percent 2002; below Fannie Mae’s estimates are multifamily mix than any of the years (the mid-point MF mix approach).16 Over the compared with the UPB-per-unit figures that between 1991 and 2002. Sensitivity analyses more recent period, the averages have ranged result from HUD’s model that uses the lower are conducted to show the effects of from 12.6 percent to 14.5 percent for 1999– MF mixes.
Fannie Mae’s Estimates HUD’s Lower MF Mix High Low Baseline Model
1997 ...... $35,063 $28,488 $31,776 $33,582 1998 ...... 40,155 32,626 36,390 37,492 1999 ...... 42,430 33,992 38,211 36,260 2000 ...... 45,797 37,210 41,504 38,142 2001 ...... 48,363 39,295 43,829 39,075 2002 ...... 53,507 43,474 48,491 44,009 Average ...... 44,219 35,847 40,033 38,093
Three points stand out. First, there is a Lower averaged 14.5 percent and its lower MF mix rather large differential between Fannie Best bound estimates averaged 12.6 percent. estimates ICF also produces lower bound estimates Mae’s Low and High UPB-per-unit figures, (percent) estimates reflecting the lack of available data. Second, (percent) of the multifamily share of the market (see HUD’s UPB-per-unit estimates based on its above list for 1994 to 2002). ICF’s lower 1994 ...... 17.2 14.0 lower MF mix model are in between Fannie bound estimates for the multifamily mix 1995 ...... 16.5 14.0 averaged 11.3 percent between 1994 and Mae’s Low and Baseline estimates. Third, the 1996 ...... 13.7 11.5 2002. It is interesting that ICF’s lower bound differentials between HUD’s and Fannie 1997 ...... 14.4 12.3 estimates are in some cases either similar or Mae’s Baseline estimates are largest during 1998 ...... 11.3 9.9 less than the multifamily shares of Fannie the two heavy refinance years of 2001 and 1999 ...... 12.3 10.7 Mae’s business. The multifamily share of 2002. 2000 ...... 13.8 11.7 Fannie Mae’s business was 9.9 percent in HUD’s conducting its market share analysis 2001 ...... 10.8 9.0 1999 (versus ICF’s lower bound estimate for with the lower MF mixes (as well as with the 2002 ...... 10.2 8.5 the market of 10.7 percent), 13.3 percent in mid-point MF mixes) recognizes different 2000 (versus ICF’s lower bound of 11.7 views about the size of the mortgage market Various averages of ICF’s Best Estimates percent), and 10.9 percent in 2001 (versus and the UPB/unit. This does not mean that are calculated in Table D.4b. Over the 1995– ICF’s lower bound market estimate of 9.0 2002 period, ICF’s Best Estimates averaged the HUD’s range of MF mixes includes percent). Even though these Fannie Mae data 12.9 percent, while HUD’s mid-point include both their seasoned and current-year estimates as low as those suggested by ICF estimates averaged 16.2 percent and HUD’s purchases, it is surprising that ICF’s market (Freddie Mac’s contractor) and Fannie Mae. lower MF mix estimates averaged 13.9 estimates would be similar or less than ICF’s estimates of multifamily shares for percent. Thus, the average of ICF’s Best Fannie Mae’s multifamily shares, given that the 1994–2002 were lower than those that Estimates is slightly lower (one percentage Fannie Mae purchased practically no small HUD used (as reported in Table D.5b). ICF’s point) than the average of HUD’s lower MF (less-than-50-unit-property) multifamily Best Estimates and Lower Bound Estimates mixes. Over the more recent 1999–2002 loans during this period. were as follows:17 period, ICF’s Best Estimates averaged 11.8 In its comments, Fannie Mae also provided percent, while HUD’s mid-point estimates various historical estimates of the MF mix
15 The projection model for 2002 showed the 16 For purposes of sensitivity analysis, the lower 17 HUD estimated ICF’s MF mixes by including following multifamily mixes for 2002: 11.1 percent MF mixes were derived as follows: three percentage subprime loans in the data that ICF reported on for the HUD New multifamily estimate ($67.4 points were subtracted from the 1995–1997 mid- pages 58–60 of its Appendix (for the Best Estimate) billion); 10.5 percent for the top end ($64 billion) point MF mixes, which were in the high 18-to-21- and on pages 63–65 of its Appendix (for the Lower percent range; two percentage points were of the Flow of Funds multifamily range ($60–64 Bound Estimate). To the extent that ICF also subtracted from the 1998–2000 mid-point MF billion), 10.3 percent for the mid-point ($62 billion), mixes, which were in the 14-to-17-percent range; excluded other single-family loans (in addition to and 9.9 percent for the low end ($60 billion). In and 1.5 percentage points were subtracted from the subprime SF loans), the estimates reported in the Sections F–H, HUD will consider multifamily mixes 2001–2002 mid-point MF mixes, which were less text overstate ICF’s initial MF mixes. as low as 9.5 percent for 2002. than 13 percent.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00254 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63833
(see its Appendix I, pages I–29 and I–30). estimates of multifamily UPB/unit were too dates of RFS respondents) was only $457 First, without giving the details of its low, and that these two errors together billion. In other words, the RFS estimates a analysis, Fannie Mae asserts that ‘‘Fannie combined to produce an estimate multifamily stock of multifamily mortgage debt 32 Mae’s analysis shows an average multifamily market share that was one to four percentage percent larger than Federal Reserve. share of 10.2 percent for the 1997–2002 points too high. A trade organization reached As with debt outstanding, multifamily loan period, compared with HUD’s 14 to 15 similar conclusions in their comments on the originations in the RFS exceed most other percent range’’ (page I–30). Fannie Mae’s multifamily mix. estimates. Over the period 1998–2001, estimate of 10.2-percent is below ICF’s Best The Department has carefully considered annual originations averaged $66 billion Estimate (12.1 percent), HUD’s lower MF mix these comments and the analysis supporting according to the RFS, and conventional estimate (13.1 percent), and HUD’s mid-point them. But HUD’s conclusion is that the 15.0- originations (total less FHA insured) MF mix estimate (15.2 percent). (See Table percent baseline multifamily mix averaged $61 billion. HUD’s estimates of D.5b.) Fannie Mae’s estimate of 10.2 percent appropriately reflected the estimates and conventional multifamily originations for is practically the same as ICF’s Lower Bound analysis appearing in the May 2004 Proposed Estimate, which averaged 10.4 percent these years, as summarized in Table D.2 of Rule. The Department’s responses to the proposed rule, averaged only $56 billion. between 1997 and 2002; of course, this raises critiques of the individual components of the In commenting on the proposed rule, Fannie the same issue mentioned above with respect multifamily mix calculation appear earlier in Mae and Freddie Mac offered estimates of to ICF’s Lower Bound Estimates. this section. In addition, the Department’s market size considerably below these.18 Fannie Mae also provided various confidence that a 15 percent estimate for estimates of UPB per unit (see above) and multifamily’s share of conventionally The single-family mortgage estimates from applied its ‘‘Low UPB per Unit Assumption’’ financed is not too high is bolstered by data the 2001 RFS, like the multifamily estimates, and its ‘‘High UPB per Unit Assumption’’ to from the newly released 2001 Residential are at odds with those from some other HUD’s likely range of MF mortgage Finance Survey (RFS). As discussed in the sources. For example, total mortgage debt on originations (as reported in column 11 of next section, the RFS indicates a long-run 1-to-4 family residences, according to the Table D.4). For the period 1997–2002, Fannie market share for multifamily that is RFS, was $5.032 trillion, whereas the Flow Mae obtained: (A) a range of 12.7–13.8 considerably higher than 15 percent. After of Funds estimate for 2002Q1 was a much percent using its ‘‘Low UPB per Unit presenting the RFS results, Section C.8 will higher $6.546 billion. Assumption’’ and (B) a range of 10.5–11.5 return to the discussion of the baseline MF In summary, the RFS estimates a somewhat percent using its ‘‘High UPB per Unit mix used in HUD’s projection model. smaller residential mortgage market than the Assumption.’’ (See Fannie Mae’s Table I.6 on 7. Evidence on the Multifamily Mix from the Flow of Funds—19 percent smaller as page I–30.) Fannie Mae’s (A) results are 2001 Residential Finance Survey measured by total debt outstanding. similar to HUD’s lower MF mix estimates, Furthermore, multifamily debt is a much which averaged 13.1 percent over the 1997– Subsequent to the Department issuing the larger part of the total residential market in 2002 period; its (B) results are slightly higher proposed rule in May, 2004, the Census the RFS than in the Flow of Funds. than ICF’s Lower Bound Estimates, which Bureau released the 2001 Residential Finance The RFS also records the number of averaged 10.4 percent over the 1997–2002 Survey (RFS). The RFS provides new housing units at each surveyed property, information on the size and composition of period. providing an opportunity to measure directly the residential mortgage market. As noted by Finally, Fannie Mae notes that its baseline the number of housing units financed instead Fannie Mae, Freddie Mac, and other analysis shows that the multifamily share of relying on indirect methods. The RFS organizations commenting on the draft rule, dropped to 5.6 percent in 2003 and that estimates indicate that, as with debt HUD’s MF assumptions (e.g., 13.5 percent) the RFS is an important and unique data outstanding, the mix of mortgage lending by clearly overstate typical multifamily shares source of data, because it is designed to the measure of units financed is more heavily and therefore the likely market opportunity provide comprehensive, nationally multifamily than previously thought. This is for the GSEs (page I–30). HUD recognizes that representative estimates on the volume and the MF mix will be lower during heavy characteristics of single-family and shown in Table D.5c, where units financed refinance years such as 2003, making it more multifamily mortgage loans and the are presented for the loan origination years difficult for the GSEs to achieve the housing properties they finance. Some organizations 2000 and 2001. These are the years for which goals; HUD’s Advance Notice of Proposed urged that the Rule not be finalized until data the estimates are least likely to be biased by Rulemaking (described in the Preamble) from the RFS has been analyzed. refinancing between the loan origination date seeks proposals on how to treat heavy The RFS data suggest a mortgage market and the survey. The estimates for 2001 are refinance years in the goals determination somewhat different in size and composition incomplete, because approximately 10 process. The range of MF mixes (13.5–15.0– from that estimated by most analysts based percent of the survey respondents reported as 16.0 percent) in HUD’s projection model on partial data. Beginning with multifamily of dates prior to December 31, 2001 and loans apply to a home purchase environment, not lending, the multifamily mortgage market is subsequently originated on those properties a heavy refinance environment. considerably larger than most analysts have would not be included. This undercount As discussed in Section C.8 below, HUD thought, according to the RFS. For example, should affect single-family and multifamily will continue to use a 15 percent MF Mix as the RFS estimate of total mortgage debt reporting about proportionally, with little its baseline. In their comments on the outstanding on properties with five of more effect on the market share calculations. proposed rule, both Fannie Mae and Freddie housing units is $608 billion dollars. The BILLING CODE 4210–27–P Mac expressed the view that HUD’s 15 only other comprehensive estimate comes
percent baseline estimate of the multifamily from the Federal Reserve Board’s ‘‘Flow of 18 Funds’’ accounts, which draw on data from The multifamily origination data in this share of the conventional mortgage market paragraph reflect a recent release of the RFS; other was too high. As described earlier in this multiple sources and on judgments by the single-family and multifamily data in this section section, those organizations argued that Fed staff. The Flow of Funds estimate of draw from an earlier version of the RFS. HUD will HUD’s estimates of multifamily loan multifamily debt outstanding as of 2002Q2 continue its analysis of the RFS data as new originations were too high, that HUD’s (the quarter most comparable to reporting versions are released by the Census Bureau.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00255 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63834 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00256 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.082 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63835
By the housing goals’ metric of number of (both single-family and multifamily) number Based on its analysis of the multifamily conventionally financed, conforming housing of dwelling units financed each year. Because market, ICF, Freddie Mac’s contractor, units, the 2001 Residential Finance Survey of the high goals-qualifying shares of offered higher projections of the MF mix. indicates a multifamily market share multifamily housing, the multifamily mix is Specifically, ICF provided the following substantially above the pre-RFS estimates of an important parameter in HUD’s projection estimates of the multifamily mix during the HUD and GSEs. As detailed in Table D.5b, model for the overall market; other things projection period, 2005–08, as follows: the multifamily share estimated for 2001 is equal, a higher multifamily mix (or 0.197, or 19.7%, and the share for 2000 is a conversely, a lower share of single-family ICF MF Mix striking 0.254, or 25.4%. These high figures loans) leads to a higher estimate of goals- (percent) are particularly noteworthy because the year qualifying loans in the overall mortgage 2001 was marked by high levels of market. 2005 ...... 13.7 refinancings, which have been viewed as The multifamily share of the conforming 2006 ...... 14.5 boosting single-family lending proportionally conventional market (or ‘‘multifamily mix’’) 2007 ...... 14.7 more than multifamily. HUD’s estimate of the is utilized below as part of HUD’s analysis of 2008 ...... 13.9 multifamily share for 2000, for example, was the share of units financed each year meeting average ...... 14.2 only 13%–14%, as derived elsewhere in this each of the housing goals. The proposed rule rule. considered multifamily mixes of 13.5 Thus, ICF’s 14.2-percent average estimate There are several reasons for accepting the percent, 15.0 percent, and 16.5 percent, as is a little less than HUD’s baseline (15.0 RFS estimates as an accurate portrayal of the well as even lower multifamily mixes for percent), standing at the mid-point of HUD’s residential mortgage market. First, the heavy refinance environments such as 2001– 13.5 and 15.0 figures. For a discussion of estimates are generated from a national 03. The 15.0 percent level was considered as ICF’s methodology for estimating the representative sample of properties as drawn the baseline based on analysis of multifamily multifamily mix, and their actual use of their by experts at the U.S. Census Bureau. shares during home purchase environments estimated multifamily mixes in projecting Second, the survey forms were designed in of the 1990s. In the market sections below, overall market estimates for the three housing consultation with industry experts. Third, HUD continues to focus on the baseline 15.0 goal categories, see pages 126–140 of their participation in the survey was mandatory, percent but also considers a range of technical appendix, entitled ‘‘Analysis of the because it was conducted in conjunction estimates, including those provided by Proportion of the Mortgage Market that Meets with the 2000 Census. And fourth, data commenters on the proposed rule. Comments the GSEs’’ Affordable Housing Goals: Issues processing and editing at the Census Bureau by Fannie Mae and ICF are summarized of Variability and Uncertainty: Technical prior to public release of census and survey below. Appendix’’ (July 15, 2004). According to ICF, results is meticulous. In its projection model, Fannie Mae uses a they projected the number of multifamily Nonetheless, for the specific reasons noted, multifamily mix of 12.3 percent (see Table (MF) units based on the existing number of results from the RFS should be interpreted 1.6 on page 11). As noted in Section C.6 units likely to be refinanced (rollover) and cautiously. First, loan originations for any above, Fannie Mae estimated an average the expected number of MF units that would year will be understated, because the RFS multifamily mix of only 10.2 percent over the be added to the housing stock (new will record only those loans still outstanding 1997–2002 period. Fannie Mae notes that completions). The amount of rollover was as of the late 2001 or early 2002 survey date. HUD’s 13.5–16.5 range is ‘‘well above the estimated as the average of the number of Loans originated in, for example, 1998, will range of estimates suggested by an units financed 8, 9, and 10 years ago. ICF be recorded only if those loans have not been examination of all available data and is used these time periods because 10-year refinanced, repaid, or charged off prior to the inconsistent with the current weak balloon mortgages are the most common MF RFS survey date. For this reason, the RFS fundamentals in the multifamily market.’’ mortgages, and MF loans typically include a unit count and especially the market share (Fannie Mae, p. 15) Fannie Mae’s views yield maintenance period to limit estimates for 2001 are more reliable than about the future mortgage market were prepayments.19 In their basic report, they those for 2000 and earlier years. Second, discussed on pages I–14 to I–17 in its state that they view the above estimates from some of the results of the RFS are Appendix I (‘‘Comments on HUD’s Analysis their MF projection model as ‘‘our core, or substantially at odds with other evidence and of the Statutory Factors’’) to its comments. As our most likely forecast for 2005 through industry perceptions, as noted already. discussed earlier, Fannie Mae’s somewhat 2008’’ (ICF Report, p. 40). While they state Another example of a surprising RFS finding pessimistic views about the future market that ‘‘our [ICF] multifamily projections for is the time path of multifamily loan were driven by the current high vacancy rates 2005 through 2008 have a sound empirical originations. According to the RFS, for multifamily properties and the fact that basis owing to the nature of multifamily originations were roughly 50 percent greater the high-renter age group (the so-called ‘‘echo mortgages and new multifamily in 1998–1999 than in 2000–2001, whereas boom’’ aged 20–34) will not begin to increase construction,’’ ICF also reminds readers of most other evidence points to originations in until after 2007. Fannie Mae also emphasized the uncertainty of its MF projections when it 2000–2001 that at least equaled, and likely that the recent spike in multifamily states ‘‘while we believe the core range is the exceeded, the volume of 1998–1999. originations (beginning in 2001) means that best and most likely estimate of the future Lastly, in response to user feedback and its a large portion of today’s holders of market, we [ICF] recognize that it is possible own data checks, the Census Bureau has multifamily mortgages have already that the actual outcomes may be outside this revised the RFS estimates three times since refinanced and therefore will have only range, either higher or lower’’ (ICF Report, p. the initial data release in early July 2004. The limited ability and incentive to refinance 40). The ICF basic report is entitled possibility remains that additional errors will over the next several years, due to yield ‘‘Analysis of the Proportion of the Mortgage be found and that the resulting revisions to maintenance provisions on their existing Market that Meets the GSEs’’ Affordable the data will significantly change the RFS multifamily mortgages. According to Fannie Housing Goals: Issues of Variability and portrayal of the multifamily mortgage market. Mae, these loans will not begin to exit their Uncertainty: Technical Appendix’’ (July 15, HUD will continue its analysis of the RFS as yield maintenance periods until sometime 2004). Because the basic report and the new versions are released. between 2008 and 2010, with the result being appendix are paginated differently, they will On balance, the Department views the RFS that the 2005–2008 period appears to have be referenced separately—ICF’s basic report as providing strong additional evidence that relatively limited prospects for multifamily will be referred to as the ‘‘ICF Report’’, while the Department’s baseline multifamily mix refinancing. Fannie Mae notes that single- their appendix will be referred to as the ‘‘ICF percentage of 15% is not an overestimate. family lending is not subject to these Appendix’’. The RFS data, weighed alone, would have constraints and is more likely to undergo As discussed earlier, the 2001 RFS that percentage set much higher. modest refinance waves as a result of interest provides higher estimates of the MF mix for rate fluctuations. Based on its analysis, 8. Multifamily Mix in HUD’s Model—Further Fannie Mae concludes that a multifamily Discussion 19 Estimates of new MF units were created by share of 12.3 percent is ‘‘consistent with comparing the historical estimates of numbers of As noted above, the ‘‘multifamily mix’’ is reasonable estimates’’ of the multifamily units added by HUD and REIS, creating a ratio, and the number of multifamily units financed market (Fannie Mae Appendix, Table I.15, p. then applying that ratio to the REIS’ future each year as a percentage share of the total I–42). projections.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00257 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63836 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
1999–2001 than either Fannie Mae or ICF. combine mortgage originations for the three loans averaged 9.6 percent between 1993 and The RFS data suggest that 15.0 percent is a different types of single-family properties: 2003, as compared with a 6.8 percent average reasonable baseline, particularly for a home Owner-occupied, one-unit properties (SF-O); for refinance loans. purchase environment. Thus, the market 2–4 unit rental properties (SF 2–4); and 1– • The investor share for home purchase analysis of the housing goals in Sections F– 4 unit rental properties owned by investors loans recently increased, rising from slightly H will continue to use 15.0 percent as the (SF-Investor). The fact that the goal above 9.0 percent during 1999 to around 10.0 baseline MF mix. To reflect the uncertainty percentages are much higher for the two percent during 2000–2001 to 12.0–13.0 with the MF data, market projections will rental categories argues strongly for percent during 2002 and 2003. The average also be provided for alternative MF mixes of disaggregating single-family mortgage investor share for home purchase loans was 12.25 percent (approximating Fannie Mae’s originations by property type. This section 11.2 percent between 1999 and 2003. projection of 12.3 percent), 13.5 percent (the discusses available data for estimating the • In its comments, Fannie Mae noted that low-end projection for a home purchase relative size of the single-family rental HUD should deduct subprime loans from environment used in HUD’s 2004 proposed mortgage market. investor loans. As shown in the middle rule), 14.25 percent (approximating the 12.2 The Residential Finance Survey (RFS) and portion of Table D.6a, deducting investor percent average of ICF’s best projections of HMDA are the data sources for estimating the subprime loans reduces the overall investor MF mixes between 2005 and 2008), and 16.0 relative size of the single-family rental share by approximately one-half percentage percent (a half percentage point below the market. The 2001 RFS provides mortgage point (e.g., 1999–2003 average is reduced high-end projection for a home purchase origination estimates for each of the three from 8.3 percent to 7.7 percent).21 environment used in HUD’s 2004 proposed single-family property types, as it includes • HMDA data for metropolitan areas rule). Based on ICF’s best projection and mortgages originated during 2001, as well as (bottom portion of Table D.6.a) show a HUD’s analysis of the 2001 RFS, the bottom surviving mortgages that were originated in slightly lower investor share than HMDA end of the range probably should not go earlier years such as 1999 and 2000. HMDA data for both metropolitan and non- below 13.5 percent for a home purchase divides newly-originated single-family metropolitan areas (top portion of Table environment. However, results are provided 20 mortgages into two property types: D.6a). Between 1993 and 2003, the investor for the 12.25 percent in order to show the (1) Owner-occupied originations, which share in metropolitan areas averaged 7.5 sensitivity of the market sizing to the include both SF-O and SF 2–4. percent, as compared with 7.8 percent for the assumption made by Fannie Mae in its (2) Non-owner-occupied mortgage U.S. as a whole. During the more recent analysis. Of course, it is recognized that the originations, which include SF Investor. 1999–2003 period, the differential was multifamily mix will be significantly lower The percentage distributions of single- slightly higher, 7.8 percent versus 8.3 during heavy refinancing periods such as family mortgages from HMDA and the 2001 percent. 2001–2003. Therefore, additional sensitivity RFS are provided in Table D.6a and D.6b. analyses will be conducted to show the HMDA data will be discussed first. Because BILLING CODE 4210–27–P effects of even lower multifamily mixes. But HMDA combines the first two categories (SF- as explained in the Preamble of this Final O and SF 2–4), the comparisons between the 21 These data without subprime loans are Rule, in its goals scoring, HUD will reduce data bases must necessarily focus on the SF presented merely to provide a sense of the likely refinance loans so they account for not more investor category. The following points stand changes if one excludes subprime investor loans. than 40 percent of combined home purchase out from Table D.6.a: Three comments should be made about them. First, HUD’s procedure is to drop one-half of subprime and refinance loans. This addresses the • The investor share of all single-family problem of a low MF mix during a heavy loans as a proxy for B&C loans, which one reduce loans has ranged from 5.7 percent (1993) to the one-half percentage point differential refinancing period reducing the ability of the 9.1 percent (2000), with an average of 7.8 mentioned in the text to a one-quarter point GSEs to meet the new goal targets. percent. Over the more recent 1999–2003 percentage differential. Second, the comparisions in D. Single-Family Owner and Rental Mortgage period, the investor share has averaged 8.3 Table D.6a do not deduct single-family owner subrpime loans; doing that would raise the investor Market Shares percent. • The investor share is much higher for shares from those in middle portion of the table. 1. Available Data on Investor Share home purchase loans than for refinance Third, HUD’s model starts with investor and owner property shares that include subprime loans (such As explained later, HUD’s market model loans. The investor share of home purchase as those in the top portion of Table D.6a) and then will also use projections of mortgage excludes the subprime loans as part of the originations on single-family (1–4 unit) 20 The HMDA data reported in this section ignore derivations within the model. See Section F for an properties. Current mortgage origination data HMDA loans with ‘‘non-applicable’’ for owner type. explanation of this procedure.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00258 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63837
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00259 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.083 63838 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00260 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.084
Table D.6b provides information on similar for single-family mortgages originated mortgages purchased by the GSEs. Between investor loans from the 2001 RFS. During in earlier years that had also survived (i.e., 1999 and 2003, the investor share of Fannie 2001, investors accounted for 13.4 percent of not prepaid) until the time of the RFS survey Mae’s single-family mortgage purchases all new single-family mortgages. Similar to in 2001; for example, the investor share was ranged from 4.2 percent (1999) to 7.8 percent the pattern in HMDA, the RFS-reported 13.0 percent for surviving 1999 mortgages (2000). Freddie Mac’s investor share has been investor share of home purchase loans (15.7 and 14.0 percent for surviving year 2000 lower, ranging from 3.0 percent (2003) to 4.8 percent) was higher than the investor share mortgages. percent (2000). The low figure for 2003 was (9.0 percent) of refinance loans (see Table For comparison purposes, Table D.6c due to the heavy refinancing of owner loans D.6b). The RFS-based investor shares were provides investor shares of the single-family in that year.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00261 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63840 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00262 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.085 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63841
The RFS investor share of 13.4 percent in paper, they conclude that 12 percent is a the mix of units in the rental market and 2001 is substantially larger than the reasonable estimate of the investor share of overstates the size of the goals-eligible corresponding HMDA investor share of 7.8 single-family mortgage originations.23 portion of the rental market. percent. In their comments on the 2004 Blackley and Follain caution that uncertainty Similarly, Freddie Mac concluded that proposed rule, as well as in their comments exists around this estimate because of HUD overestimated the SF investor share of on HUD’s earlier 1995 and 2000 GSE rules, inadequate data. the market because it relied on the RFS rather than HMDA. Freddie Mac says investor- the GSEs have argued that HUD should use 3. GSE Comments on SF Rental Shares in the owners have an incentive to claim falsely the HMDA-reported SF investor share. In its Proposed Rule 1995 and 2000 rules and the 2004 proposed they are owner-occupants because of higher GSE rule, HUD’s baseline model assumed a Fannie Mae, Freddie Mac, and ICF thought underwriting standards and higher interest 10 percent share for the SF investor group— that the investor share should be lower than rates on investor-owner properties. only slightly higher than the HMDA-based the 10 percent used by HUD. While they According to Freddie Mac, these incentives estimates; alternative models assuming 8 agreed with HUD that the RFS provided the likely result in HMDA’s undercounting SF percent and 12 percent were also considered. most accurate estimate of the true investor investor loans relative to the more accurate At that time, HUD argued that its baseline share of the market, they emphasized that counts of investor loans from the RFS. projection of 10 percent was probably quite lender reporting of investor loans to the GSEs Freddie Mac concludes as follows: conservative; however, given the uncertainty was best proxied by HMDA data (which, of This undercounting [on the part of HMDA], around the data, it was difficult to draw firm course, are based on lender reports). That is, however, is exactly what is desired when conclusions about the size of the single- the actual opportunities available to the GSEs estimating the goal share available to the family investor market, which necessitated in the SF investor market are best measured GSEs. Because the GSEs’ information on their that HUD conduct sensitivity analyses using by data that lenders report based on loans has the same ‘‘bias’’ as does the HMDA investor shares (e.g., 8 percent) less than 10 information from actual loan applications. data. * * * The HMDA data, therefore, are percent. HUD’s argument that its 10 percent Based on this argument, they concluded that more appropriate to estimating the market for baseline work was probably conservative was HUD’s market sizing analysis should rely on goal setting than are the RFS data. (p.II–6) based on earlier work by Blackley and HMDA data, not RFS data. Essentially, Freddie Mac concludes that Follain. It is interesting to briefly review their For example, Fannie Mae argued that the HUD’s market estimates should measure work because they focused on the differences most valid measure of the single-family opportunities in the marketplace that are between RFS and HMDA data. rental market is the same measure (lender- actually available to the GSEs. Such reported data to HMDA) against which the opportunities are best measured by lender- 2. Blackley and Follain Analysis of Investor GSEs’ performance is measured. Fannie Mae reported HMDA data, not the more accurate Market Share points out that that two (10 percent and 12 RFS data. ICF reaches a similar conclusion, As mentioned, during the 1995 rule- percent) of the three scenarios that HUD uses as it states that ‘‘HMDA data, or its making, HUD asked the Urban Institute to exceed the highest investor share ever equivalent, are what the GSEs’ performance analyze the differences between the RFS and reported in HMDA. Fannie rejects HUD’s will be measured against and is therefore the HMDA investor shares and determine which justification (the 1991 RFS and the Blackley- appropriate metric for estimating market goal was the more reasonable. The Urban Follain analysis) for using the higher shares’’ (ICF Report, p.20). Institute’s analysis of this issue is contained scenarios because the lender reporting to the 4. SF Investor Shares in the Final Rule in reports by Dixie Blackley and James GSEs is closer to HMDA data than to the In this final Rule, HUD has switched to a Follain.22 Blackley and Follain provide reporting in the RFS. Fannie Mae argues that HMDA-based system and provides overall reasons why HMDA should be adjusted the 1995 Blackley and Follain analysis market share estimates for a range of single- upward as well as reasons why the RFS bolsters its case against the RFS measures. family investor shares. For each year between should be adjusted downward. They find that Fannie Mae notes that both HUD and 1993 and 2003, the top-right-hand-side HMDA may understate the investor share of Blackley and Follain conclude that there is portion of Table D.6a shows the projected single-family mortgages because of ‘‘hidden a reporting bias in the HMDA data that is not investor share in a ‘‘home purchase investors’’ who falsely claim that a property present in the RFS. The bias is in part due environment’’ assuming a refinance share of is owner-occupied in order to more easily to hidden investors. At the time of 35 percent, 40 percent, and 45 percent. obtain mortgage financing. RFS may overstate origination, the property may be owner- Refinance shares greater than 35 percent are the investor share of the market because units occupied or may be intended to be owner- included here because single-family investor that are temporarily rented while the owner occupied. In fact, the property may become loans typically have higher refinance shares seeks another buyer may be counted as rental rental shortly after origination. As a result, than single-family-owner loans. As shown in units in the RFS, even though rental status the RFS reports a more accurate higher Table D.6a, the average 1993–2003, HMDA- of such units may only be temporary. The percentage of rental housing because it is a based investor share would have been 8.5 RFS’s investor share should be adjusted snapshot of housing, not a collection of (8.4) percent if the investor refinance share downward in part because the RFS assigns information at mortgage origination. Fannie had been 40 (45) percent during this period. all vacant properties to the rental group, but Mae says HUD uses the RFS because it is the During the more recent 1999–2003 period, some of these are likely intended for the more accurate measure of the rental market which was characterized by particularly high owner market, especially among one-unit at any moment in time. However, Fannie Mae HMDA-reported investor shares for home properties. Blackley and Follain’s analysis of argues that the same bias in HMDA also purchase loans, the average investor share this issue suggests lowering the investor exists in its own reporting when it acquires would have been 9.4 (9.2) percent if the share from the 1991 RFS-reported investor mortgages. According to Fannie Mae, an investor refinance share had been 40 (45) share of 17.3 percent to about 14–15 percent. apples to apples comparison would make percent during this period. As noted earlier, Finally, Blackley and Follain note that a sure that the GSE goals contain the same the HMDA-reported investor shares for conservative estimate of the SF investor share biases that the GSE reports contain, rather metropolitan areas are slightly lower than is advisable because of the difficulty of than no bias. Finally, Fannie says that even those for the entire U.S. As shown in the measuring the magnitudes of the various HMDA overstates the investor share of the bottom-right-hand portion of Table D.6a, the effects that they analyzed. In their 1996 single-family market because of second average 1999–2003, HMDA-based investor homes. Second homes are reported in HMDA share for metropolitan areas would have been 22 Dixie M. Blackley and James R. Follain, ‘‘A as ‘‘not owner occupied’’ to determine 8.9 (8.7) percent if the investor refinance Critique of the Methodology Used to Determine investor status but are not goals eligible. share had been 40 (45) percent during this Affordable Housing Goals for the Government Therefore, according to Fannie Mae, HUD’s period. Sponsored Housing Enterprises,’’ report prepared use of HMDA would overestimate the goals- The above analysis suggests that the for Office of Policy Development and Research, eligible share of the single-family market. As HMDA-reported investor share of a future Department of Housing and Urban Development, a result of these data and methodology October 1995; and ‘‘HUD’s Market Share home purchase market will probably be Methodology and its Housing Goals for the issues, Fannie believes HUD miscalculates around 8.5–9.0 percent, or possibly higher if Government Sponsored Enterprises,’’ unpublished the recent figures for home purchase loans paper, March 1996. 23 Blackley and Follain (1996), p. 20. hold up (in this case, around 9.5 percent).
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00263 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63842 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Thus, HUD’s analysis of market shares in adjustments have to be made to the HMDA percent) and 89.92 percent for SF–O 1-Unit Sections F–H will report overall market data. mortgages (obtained by subtracting 1.58 estimates for a range of SF investor shares— First, the owner-occupied HMDA data have percent for the overall SF–O share of 91.5 8.0 percent, 8.5 percent, 9.0 percent, and 10.5 to be disaggregated between SF–O 1-Unit and percent). Thus, in this scenario, the percent. SF 2–4 mortgages. HUD’s 2004 proposed GSE distribution across SF mortgage types would 5. Single-Family Market in Terms of Unit rule assumed that SF 2–4 mortgages be as follows: (d) 89.92 percent for SF–O 1- Shares accounted for 2.0 percent of all single-family Unit mortgages; (b) 1.58 percent for SF–O 2– mortgages. Based on the 2001 RFS data, this 4 mortgages; and (c) 8.50 percent for SF- The market share estimates for the housing percentage is reduced to about 1.6 percent in Investor mortgages. Table D.6d shows the goals need to be expressed as percentages of this Final Rule. In 2001, the RFS shows the distribution of SF mortgages under the units rather than as percentages of mortgages. following distribution across the three single- various assumptions assumed in Sections F– Since a SF 2–4 and a SF-investor mortgage family mortgage types: (a) 85.1 percent for H. For comparison purposes, the SF–O 2–4 finances more than one dwelling unit, SF–O 1-Unit mortgages; (b) 1.5 percent for shares for the GSEs are reported in Table adjustments reflecting units-per-mortgage SF–O 2–4 mortgages; and (c) 13.4 percent for D.6c. The 1999 to 2003 shares for Fannie Mae have to be made in order to arrive at the SF-Investor mortgages (see Table D.6b). Thus, are approximately 2.0 percent while those for distribution of newly-financed single-family according to 2001 RFS data, SF 2–4 Freddie Mac are approximately 1.5 percent. dwelling units. From HMDA, one can obtain mortgages represent 1.73 percent of all Thus, the Fannie Mae shares are consistent the share of investor mortgages (those single-family-owner mortgages (obtained by with the 2.0 percent assumption used in the reported in Table D.6a) and the share of dividing (b) by the sum of (a) and (b)). In the 2004 proposed rule while the Freddie Mac owner mortgages (obtained by subtracting the market projection models, the SF-investor shares are consistent with the 1.6 percent share of investor mortgages from 100 mortgage share is assumed to be lower than assumption used in this Final Rule. percent). HMDA does not disaggregate the the RFS-reported figure of 13.4 percent. If the Sensitivity analyses in Sections F–G will SF-owner (SF–O) mortgage category into its SF-investor share is 8.5 percent, then the SF– show the effects of using the 2.0 percent two components: SF–O 1-Unit mortgages and O share is 91.5 percent, which is split as assumption (as compared with the 1.6 SF–O 2–4 mortgages. To arrive at shares of follows: 1.58 percent for SF–O 2–4 mortgages percent baseline). SF financed dwelling units, two sets of (obtained by multiplying 0.0173 by 91.5 BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00264 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63843
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00265 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.086
Second, the resulting mortgage-based the RFS. The corresponding 2001 figures 1.4) has little impact on the market sizing distributions have to be shifted to unit-based from the RFS were 2.1 and 1.4, respectively. results. distributions by applying the unit-per- As shown in Table D.6d, the GSE data has Based on these calculations, the percentage mortgage assumptions. The 2004 proposed consistently been around the figures in the distribution of newly-mortgaged single- GSE rule assumed the following: 2.25 units 2004 proposed GSE rule, which were 2.25 family dwelling units was derived for each of per SF 2–4 property and 1.35 units per SF and 1.35, respectively. Thus, it was decided the various estimates of the investor share of investor property. Based on RFS data, these to use the 1999–01 RFS averages which drop single-family mortgages. The results are numbers are reduced slightly to the each units-per-mortgage figure by 0.05. presented in Table D.6e for investor following: 2.2 units per SF 2–4 property and Sensitivity analysis shows that this issue percentage shares of 8.0, 8.5, 9.0, and 9.5. 1.3 units per SF investor property. These (whether to use the 1999–01 combination of Three points should be made about these figures are based on 1999–2001 averages from 2.2/1.3 or to use the 2001 combination of 2.1/ data.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00266 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63845
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00267 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.087
First, notice that the rental categories 2004 proposed GSE rule. If the single-family In its projections, Fannie Mae assumes 8.0 represent a larger share of the unit-based investor share is 9.0 (9.5) percent, then percent for the investor share of mortgages, market than they did of the mortgage-based single-family rental units account for account a figure Fannie Mae says is consistent with market reported earlier. For example, when for 13.0 (13.6) percent of all newly-mortgaged HMDA data (Fannie Mae Appendix I, Table the SF-investor category represents 8.5 SF units. 1.11, p. I–38). Under the 2001 RFS percent of all SF mortgages, it represents 10.6 ICF projected that SF rental units would assumptions (see above), this translates into percent of all SF units financed. This, of account for 12.0 percent of all single-family- a single-family rental share (on a units basis) course, follows directly from applying the financed units during the 2005–2008 loan-per-unit expansion factors. projection period (ICF Appendix, p.126). of 11.8 percent. Under the units-per-loan and Second, notice that the ‘‘All SF-Rental Under the units-per-mortgage and SF–O 2–4 SF–O 2–4 assumptions of the proposed rule, Units’’ column highlights the share of the share assumptions that ICF was using (2.25 this translates into a single-family rental single-family mortgage market accounted for for SF–O 2–4 and 1.35 for SF–Investor and share (expressed on a units basis) of 12.7 by all single-family rental units, for both SF– a 2.0 percent share for SF–O 2–4 mortgages), percent. O 2–4 properties and SF-Investor properties. ICF’s 12-percent assumption for single-family Third, if the investor mortgage share were For example, when the investor mortgage rental units translates back to an investor 13 percent (the 2001 figure from the RFS), share is 8.5 percent, single-family rental units mortgage share of 7.5 percent.24 (in SF 2–4 properties as well as in SF single-family rental units would account for investor properties) account for 12.4 percent over 17 percent of all newly-mortgaged 24 It should be mentioned that ICF’s 12.0 percent single-family units. of all newly-mortgaged SF units. This single- assumption for the SF rental share seems at odds family rental share compares with 15.1 with ICF’s Exhibit 6.4, which suggests that ICF’s The unit distributions reported for the percent under the baseline assumptions of 1994–2002 average SF rental share is 14.9 percent. GSEs in Table D.6f will be discussed in the the 2004 proposed GSE Rule; the 15.1 A 14.9 percent SF rental share would be consistent next section. percent figure is reported in Table D.6b of the with a 12 percent investor mortgage share.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00268 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63847
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00269 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.088 63848 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
E. HUD’s Market Share Model alternative assumptions will be Substituting these values into (2) yields an This section integrates findings from the examined, it must be emphasized that estimate of 8.5 million mortgages. previous two sections about the size of the the important concept for deriving the Third, the total number of single-family multifamily mortgage market and the relative goal-qualifying market shares is the mortgages is divided among the three single- distribution of single-family owner and rental relative importance of single-family family property types. Using the 89.9/1.6/8.5 mortgages into a single model of the mortgage versus multifamily mortgage originations percentage distribution for single-family market. The section provides the basic (the ‘‘multifamily mix’’ discussed in mortgages (see Section D), the following equations for HUD’s market share model and Section C) rather than the total dollar results are obtained: identifies the remaining parameters that must volume of single-family originations (3a) SF–OM# = 0.899 * CCSFM# = number be estimated. considered in isolation. of owner-occupied, one-unit mortgages = The output of this section is a unit-based Substituting these values into (1) yields an 7.642 million. estimate for the conventional conforming distribution for the four property types (3b) SF–2–4M# = 0.016 * CCSFM# = number market (CCSFM$) of $1,197 billion. discussed in Section B.25 Sections F–H will of owner-occupied, two-to-four unit Second, the number of conventional apply goal percentages to this property mortgages = 0.136 million. conforming single-family mortgages distribution in order to determine the size of (3c) SF–INVM# = 0.085 * CCSFM# = number (CCSFM#) is derived as follows: the mortgage market for each of the three of one-to-four unit investor mortgages = housing goals. (2) CCSFM# = (CCSFM$ * (1–REFI)/ 0.723 million. 1. Basic Equations for Determining Units PSFLOAN$) + (CCSFM$ * REFI)/ RSFLOAN$) Fourth, the number of dwelling units Financed in the Mortgage Market where financed for the three single-family property The model first estimates the number of REFI= the refinance rate, assumed to be 35 types is derived as follows: dwelling units financed by conventional percent for the baseline.28 (4a) SF–O = SF–OM# + SF–2–4M# = number conforming mortgage originations for each of PSFLOAN$ = the average conventional of owner-occupied dwelling units the four property types. It then determines conforming purchase mortgage amount financed = 7.778 million. each property type’s share of the total for single-family properties; estimated to (4b) SF 2–4 = 1.2 * SF–2–4M# = number of number of dwelling units financed. be $146,000.29 rental units in 2–4 properties where an a. Single-Family Units RSFLOAN$ = the average conventional owner occupies one of the units = 0.163 conforming refinance mortgage amount million.31 This section estimates the number of for single-family properties; estimated to (4c) SF–INVESTOR= 1.3 * SF–INVM# = single-family units that will be financed in be $131,000.30 number of single-family investor the conventional conforming market, where dwelling units financed = 0.940 million. single-family units (SF–UNITS) are defined as: $1,700 billion during the 2005–2008 period that the Fifth, summing equations 4a–4c gives the goals will be in effect. As recent experience shows, SF–UNITS = SF–O + SF 2–4 + SF– projected number of newly-mortgaged single- market projections often change. For example, in family units (SF–UNITS): INVESTOR January 2003, the MBAA projected $1,246 billion First, the dollar volume of conventional for 2003; of course, actual 2003 mortgage (5) SF–UNITS = SF–O + SF 2–4 + SF– conforming single-family mortgages originations were triple the latter amount. (See INVESTOR = 8.915 million http://www.MBAAa.org/marketdata/forecasts for b. Multifamily Units (CCSFM$) is derived as follows: January 2003 Mortgage Finance Forecasts.) While (1) CCSFM$ = CONV% * CONF% * SFORIG$ Sections F–H will report the effects on the market The number of multifamily dwelling units where estimates of alternative estimates of single-family (MF–UNITS) financed by conventional CONV% = conventional mortgage mortgage originations, it should be emphasized that conforming multifamily originations is originations as a percent of total the important parameter for the market sizing calculated by the following series of estimates is the share of single-family-owner units equations: mortgage originations; estimated to be relative to the share of single-family and 26 88%. multifamily rental units, not the absolute level of (5a) TOTAL = SF–UNITS + MF–UNITS CONF% = conforming mortgage originations single-family originations. (5b) MF–UNITS = MF–MIX * TOTAL = MF– (measured in dollars) as a percent of 28 The model requires an estimated refinance rate MIX * (SF–UNITS + MF–UNITS) = [MF– conventional single-family originations; because purchase and refinance loans can have MIX/(1 ¥ MF–MIX)] * SF–UNITS where forecasted to be 80% by industry. different shares of goals-qualifying units. In 2003, MF–MIX = the ‘‘multifamily mix’’, or the SFORIG$ = dollar volume of single-family the refinance rate was almost 70 percent. In its percentage of all newly-mortgaged one-to-four unit mortgages; $1,700 August 13, 2004 forecast, the MBAA projects 25 dwelling units that are multifamily; as billion is used here as a starting percent for 2005, as did Fannie Mae in its August 17, 2004 forecast. The baseline model uses a higher discussed in Section C, alternative assumption to reflect market conditions estimates of the multifamily market will 27 refinance rate of 35 percent because conforming during the years 2005–2008. While conventional loans tend to refinance at a higher rate be included in the analysis. As explained than the overall market. Sensitivity analyses for in Section C above, the baseline model 25 The property distribution reported in Table D.1 alternative refinance rates are presented in Sections assumes a multifamily mix of 15 percent; is an example of the output of the market share F–H. results are also presented in the basic model. Thus, this section completes Step 1 of the 29 The average 2002 purchase loan amount is market tables of Sections F–H for a three-step procedure outlined above in Section B. estimated at $135,060 for owner occupied units 26 higher (16.0 percent) multifamily mix According to estimates by the Mortgage using 2002 HMDA average loan amounts for single- and for lower (12.25 percent, 13.5 Bankers Association of America (MBAA), the family home purchase loans in metropolitan areas. conventional share of the 1–4 family market was A small adjustment is made to this figure to account percent and 14.25 percent) multifamily between 86 and 88 percent of the market from 1993 for a small number of two-to-four and investor mixes. In addition, further sensitivity to 1999, with a one-time low of 81 percent in 1994. properties (see Section D above). This produces an analyses are reported in those sections Calculated from ‘‘1–4 Family Mortgage average purchase loan size of $133,458 for 2002 for even lower multifamily mixes that Originations’’ tables (Table 1—Industry and Table which is then inflated 3 percent a year for three could occur during periods of heavy 2—Conventional Loans) from ‘‘MBAA Mortgage and years and then rounded to arrive at an estimated single-family refinancing activity. Market Data,’’ at www.MBAAa.org/marketkdata/ as $146,000 average loan size for home purchase loans of July 13, 2000. More recent unpublished estimates in 2005. Assuming a multifamily mix of 15 percent and solving (5b) yields the following: by MBAA are slightly higher. As discussed in the 30 The average refinance loan amount is estimated text, the market sizing shares are affected by by averaging the relationship between HMDA parameters other than this one, such as the average purchase and refinance loan amounts for amounts are used for both purchase and refinance multifamily share of newly-mortgaged dwelling 1999 and 2000, which were non-refinance loans. This relationship is consistent with the units. environments. Applying this average of 90 percent observed relationship in past refinance years such 27 In its August 17, 2004 forecast, Fannie Mae (refinance loan amount/purchase loan amount) to as 1998, 2001, and 2002. projected approximately $1.6 billion for 2005 and the $146,000 average loan amount for purchase 31 Based on the 2001 RFS, there is an average of 2006 while the MBAA projected $1.8 billion for loans gives a rounded estimate of $131,000 for 2.2 housing units per mortgage for 2–4 properties 2005 in its August 13, 2004 forecast. As discussed average refinance loan amounts. When refinance and 1.3 units per mortgage for single-family later, single-family originations could differ from environments are used, $146,000 average loan investor properties. See earlier discussion.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00270 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63849
(5c) MF–UNITS = [0.15/0.85] * SF–UNITS = % Share discussed in Section D, these investor 0.176 * SF–UNITS = 1.6 million. mortgage shares are lower than the range (8.0 c. Total Units Financed MF–UNITS ...... 15.0 percent, 10.0 percent, and 12.0 percent) considered in the 2004 proposed GSE rule. The total number of dwelling units Total ...... 100.0 The middle values (8.5 percent and 9.0 financed by the conventional conforming or mortgage market (TOTAL) can be expressed percent) are probably the ones that should be SF–O ...... 74.5 in three useful ways: considered as ‘‘baseline’’ projections; the SF–RENTER ...... 10.5 above example used a mortgage share of 8.5 (6a) TOTAL = SF–UNITS + MF–UNITS = MF–UNITS ...... 15.0 percent, but 9.0 percent could also have been 10.6 million (or more precisely, used to characterize a home purchase 10,632,145 units) Total ...... 100.0 (6b) TOTAL = SF–O + SF 2–4 + SF– environment. However, HUD recognizes the uncertainty of projecting origination volume INVESTOR + MF– UNITS Sections C and D discussed alternative (6c) TOTAL = SF–O + SF–RENTAL + MF– in markets such as single-family investor projections for the mix of multifamily properties; therefore, the analysis in Sections UNITS where SF–RENTAL equals SF–2– originations and the investor share of single- 4 plus SF–INVESTOR F–H considers market assumptions other family mortgages. This appendix will report than these baseline assumptions. 2. Dwelling Unit Distributions by Property results for multifamily mixes of 13.5 percent, Table D.7 reports the unit-based Type 15.0 percent, and 16.0 percent but sensitivity distributions produced by HUD’s market analyses for two other multifamily mix The next step is to express the number of share model for different combinations of assumptions (e.g., the 12.3 percent dwelling units financed for each property these projections. Unit-based distributions type as a percentage of the total number of assumption used by Fannie Mae and the 14.2 assumption used by ICF) will also be are reported for each combination of a units financed by conventional conforming multifamily mix (12.25, 13.5, 14.25, 15.0, and mortgage originations.32 reported. Under the baseline 15.0 percent multifamily mix, the newly-mortgaged unit 16.0) and investor mortgage share (8.0, 8.5, The projections used above in equations 9.0, and 9.5). The effects of the different (1)–(6) produce the following distributions of distribution would be 74.5 percent for Single- projections can best be seen by examining the financed units by property type: Family Owner, 10.5 percent for Single- Family Renter, and 15.5 percent for owner category which varies by 4.8 percentage points, from a low of 72.6 percent % Share Multifamily-Units. The analysis in sections F–H will focus on goals-qualifying market (multifamily mix of 16.0 percent coupled SF–O ...... 74.5 shares for this property distribution as well with an investor mortgage share of 9.5 SF 2–4 ...... 1.5 as the ones noted above. percent) to a high of 77.4 percent SF INVESTOR ...... 9.0 As discussed in Section D, the basic tables (multifamily mix of 12.25 percent coupled providing the goals-qualifying market with an investor mortgage share of 8.0 estimates in this appendix will report results percent). The overall rental share is also 32 The share of the mortgage market accounted for by owner occupants is (SF–O)/TOTAL; the share of for the following investor shares of single- highlighted in Table D.7, varying from 22.6 the market accounted for by all single-family rental family mortgages—8 percent, 8.5 percent, 9.0 percent to 27.4 percent. units is SF–RENTAL/TOTAL; and so on. percent, and 9.5 percent. For reasons BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00271 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63850 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00272 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.089
The baseline projection of newly- conforming mortgages, 68.3 percent were 3 of the three-step procedure discussed in mortgaged units in the 2004 proposed GSE owner-occupied units, 16.5 percent were Section B.2. rule was 72.2 percent for owner units, 12.8 single-family rental units, and 15.2 percent Technical issues and data adjustments percent for single-family rental units, and were multifamily rental units. Thus, the RFS related to the low- and moderate-income 15.0 percent for multifamily units. In this presents a much lower owner share than does percentages for owners and renters are Final Rule, the baseline projection is 74.5 HUD’s, ICF’s, or Fannie Mae’s models. See discussed in the first two subsections. Then, percent for owner units, 10.5 percent for Sections C and D for further discussion of the estimates of the size of the low- and single-family rental units, and 15.0 percent RFS. moderate-income market are presented along for multifamily units, if an investor mortgage Finally, it is interesting to compare the with several sensitivity analyses. Based on share of 8.5 percent is used. If an investor above market-based distributions of financed these analyses, HUD concludes that 51–56 share of 9.0 percent is used, then the baseline units with the distributions of units financed percent is a reasonable estimate of the projection is 74.0 percent for owner units, by mortgages purchased by Fannie Mae and mortgage market’s low- and moderate-income 11.0 percent for single-family rental units, Freddie Mac. As shown in Table D.6f, the share for the four years (2005–2008) when and 15.0 percent for multifamily units. Either 1993–2003 averages (unweighted) for Fannie the new goals will be in effect. way, compared with the 2004 proposed GSE Mae were 81.0 percent for owner units, 9.0 1. Low- and Moderate-Income Percentage for rule, the rental share of financed dwelling percent for single-family rental units, and Single-Family-Owner Mortgages units has dropped by approximately two 10.0 percent for multifamily units, which percentage points due to the lower HMDA- produces an overall rental share of 19.0 a. HMDA Data based investor shares used in the Final Rule. percent. During the year 2000, Fannie Mae’s The unit distribution in ICF’s projection The most important determinant of the overall rental share did reach a peak of 24.1 model for 2005–2008 averaged 75.5 percent low- and moderate-income share of the percent. Freddie Mac’s rental shares have for owner units, 10.3 percent for single- mortgage market is the income distribution of been markedly lower than Fannie Mae’s. The family rental units, and 14.2 percent for single-family borrowers. HMDA reports 1993–2003 averages (unweighted) for Freddie multifamily units, which produces an overall annual income data for families who live in rental share of 24.5 percent, a figure closed Mac were 84.3 percent for owner units, 6.3 metropolitan areas and purchase a home or to those reported above (25.5–26.0 percent). percent for single-family rental units, and 9.3 refinance their existing mortgage.35 The data The unit distribution used by Fannie Mae percent for multifamily units, which cover conventional mortgages below the produces an overall rental share of 15.7 was approximately 77.4 percent for owner conforming loan limit, which was $322,700 percent.34 Freddie Mac’s rental share did units, 10.4 percent for single-family rental in 2003. Table D.8a gives the percentage of peak at 17.5 percent in 2000. Still, it is clear units, and 12.3 percent for multifamily units, mortgages originated for low- and moderate- that the market-based distributions project which produces an overall rental share of income families for the years 1992–2003. 33 much higher rental shares than Freddie Mac 22.6 percent, a figure less than used by ICF Data are presented for home purchase, has been purchasing. For example, the HUD (24.5 percent) or HUD (25.0–26.0 percent). refinance, and all single-family-owner loans. projection of a 25-percent rental share is over Notice that Fannie Mae and ICF assume The discussion below will often focus on nine percentage points higher than Freddie similar single-family rental shares (about 10.3 home purchase loans because they typically Mac’s 1999–2003 average rental share (15.7 percent), but ICF assumes a larger account for the majority of all single-family- percent) and over seven percentage points multifamily mix than Fannie Mae (14.2 owner mortgages.36 For each year, a low- and higher than Freddie Mac’s peak rental share percent versus 12.3 percent). HUD’s single- moderate-income percentage is also reported (17.5 percent in 2000). The 31.7-percent family rental shares (10.5–11.0 percent) are for the conforming market without B&C rental share from the RFS is 16 percentage slightly higher than the shares (about 10.3 loans. points higher than Freddie Mac’s 1999–2003 percent) used by ICF and Fannie Mae. HUD’s Table D.8a also reports similar data for average rental share (15.7 percent) and over multifamily baseline share (15.0 percent) is very-low-income families (that is, families 14 percentage points higher than Freddie slightly higher than the average (14.2 with incomes less than 60 percent of area Mac’s peak rental share (17.5 percent in percent) of ICF’s best estimate, and median income). As discussed in Section H, 2000). significantly higher than Fannie Mae’s very-low-income families are the main assumed multifamily mix (12.3). F. Size of the Conventional Conforming component of the special affordable mortgage As discussed in Sections C and D, the Mortgage Market Serving Low- and Moderate- market. Residential Finance Survey is the only Income Families mortgage data source that provides unit- based property distributions directly This section estimates the size of the low- 35 HMDA data are expressed in terms of number comparable to those reported below. Based and moderate-income market by applying of loans rather than number of units. In addition, on RFS data for 2001, HUD estimated that, low- and moderate-income percentages to the HMDA data do not distinguish between owner- occupied one-unit properties and owner-occupied of total dwelling units in properties financed property shares given in Table D.7. This section essentially accomplishes Steps 2 and 2–4 properties. This is not a particular problem for by recently acquired conventional this section’s analysis of owner incomes. 36 Sensitivity analyses will focus on how the 33 Because of rounding, the two rental component 34 Because of rounding, the two rental component results change during a heavy refinancing shares do not add to the overall rental share. shares do not add to the overall rental share. environment.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00273 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63852 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00274 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.090 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63853
Two trends in the income data should be fact that lending to low-income families has 1997 and 1999, as well as in 1995 and 1996. mentioned—one related to the growth in the remained at a high level for ten years (See Table D.8a.) Even after excluding all market’s funding of low- and moderate- demonstrates that the market has changed in subprime loans from the market definition in income families during the 1990s (and fundamental ways from the mortgage market 1997 and 1999, the very-low-income and particularly the growth since 1998 which was of the early 1990s. The numerous innovative low-mod shares for refinance loans are only the last year analyzed in HUD’s 2000 GSE products and outreach programs that the slightly less (about one percentage point) Rule); and the other related to changes in the industry has developed to attract lower- than those for home purchase loans. borrower income distributions for refinance income families into the homeownership and The year 2000 stands out because of the and home purchase mortgages. Throughout mortgage markets appear to be working and extremely high lower-income shares for this appendix, ‘‘low- and moderate-income’’ there is no reason to believe that they will refinance loans. In that year, the low-mod will often be referred to as ‘‘low-mod’. not continue to assist in closing troubling (very-low-income) share of refinance loans Recent Trends in the Market Share for homeownership gaps that exist today. As was 7.0 (4.4) percentage points higher than Lower Income Borrowers. First, focus on the explained in Appendix A, the demand for the low-mod (very-low-income) share of percentages in Table D.8a for the total (both homeownership on the part of minorities, home purchase loans; this differential is home purchase and refinance) conforming immigrants and non-traditional borrowers reduced to 5.4 (3.3) percent if B&C loans are market. After averaging about 30 percent should help to maintain activity in the excluded from the market definition (see during 1992–93, the percentage of borrowers affordable portion of the mortgage market. Table D.8a). The differential for 2000 is with less than area median income jumped Thus, while economic recession or higher reduced further to 2.8 (1.5) percent if all to 41.0 percent in 1994, and remained above interest rates would likely reduce the low- subprime loans (both A-minus and B&C) are 40 percent through 2003. Over the ten-year and moderate-income share of mortgage excluded from the market definition (not period, 1994 to 2003, the low-mod share of originations, there is evidence that the low- reported). While the projection model the total market averaged 42.9 percent (or mod market might not return to the low (explained below) for years 2005–08 will 42.2 percent if B&C loans are excluded from levels of the early 1990s. There is also input low-mod percentages for the entire the market totals).37 The share of the market evidence that the affordable lending market conforming market, the model will exclude accounted for by very-low-income borrowers increased slightly since 1998, although it is the effects of B&C loans. Sensitivity analyses followed a similar trend, increasing from 6– recognized that this could be due to the will also be conducted showing the effects on 7 percent in 1992–93 to about 12 percent in recent period of historically low interest the overall market estimates of excluding all 1994 and averaging 13.2 percent during the rates. subprime loans as well as other loan 1994-to-2003 period (or 12.7 percent if B&C Refinance Mortgages. In the 2000 Rule, categories such as manufactured housing loans are excluded). HUD’s market projection model assumed that loans. Next, consider the percentages for home low-mod borrowers represented a smaller 2000 Census Data and New OMB purchase loans. The share of the home loan share of refinance mortgages than they do of Metropolitan Area Definitions. Going market accounted for by less-than-median- home purchase mortgages. However, as forward, HUD will be re-benchmarking its income borrowers increased from 34.4 shown in Table D.8a, the income median incomes for metropolitan areas and percent in 1992 to 44.7 percent in 2003. characteristics of borrowers refinancing non-metropolitan counties based on 2000 Within the 1994-to-2002 period, the low-mod mortgages seem to depend on the overall Census median incomes, and will be share of the home purchase market averaged level of refinancing in the market. During the incorporating the effects of the new OMB 44.4 percent between 1999 and 2003, refinancing wave of 1992 and 1993, metropolitan area definitions. Thus, under compared with 42.1 percent between 1994 refinancing borrowers had much higher the new housing goals, the GSEs’ and 1998. Similarly, the very-low-income incomes than borrowers purchasing homes. performance will be scored based on 2000 share of the home purchase market was also For example, during 1993 low- and Census data and new OMB definitions of higher during the 1999-to-2002 period than moderate-income borrowers accounted for metropolitan areas (labeled ‘‘CBSA during the 1994-to-1998 period (14.1 percent 29.3 percent of refinance mortgages, definitions’’). One issue concerns whether versus 12.6 percent). Note that within the compared to 38.9 percent of home purchase the new data and the new definitions will more recent period, the low-mod share for borrowers. While this same pattern was result in lower or higher low-mod home purchase loans was particularly high exhibited during the two recent refinancing percentages relative to historical low-mod during 1999 (45.2 percent) and 2000 (44.3 periods (1998 and 2001–2002–2003), the percentages based on the 1990 Census and percent) before falling slightly in 2001 (43.2 differentials were much smaller—during earlier OMB definitions of metropolitan areas percent), only to rebound again in 2002 (44.8 2001–2002–2003 (1998), low-mod borrowers (labeled ‘‘MSA definitions’’). HUD projected percent) and 2003 (44.7 percent). As shown accounted for 41.5 (39.7) percent of refinance the effects of these two changes on the low- in Table D.8a, the low-mod shares do not loans, compared with 44.2 (43.0) percent of and moderate-income shares of the single- change much when B&C home loans are home purchase loans. However, the refinance family-owner market for the years 1999– excluded from the market definition; this is effect was still evident, as can be seen by the 2003. The middle portion of Table D.8b because B&C loans are mainly refinance almost ten percentage point drop in the low- reports low-mod shares for single-family- loans. mod percentage for refinance loans between owner loans under the MSA and CBSA It appears that the affordable lending 2000 (a low refinance year) and 2001 (a high approaches for the years 1999–2003. Except market for home purchase loans is even refinance year). for 2003, the low-mod shares for both home stronger today than when HUD wrote the On the other hand, for recent years purchase and total SFO loans are lower 2000 Rule, which covered market data characterized by a low level of refinancing, under the new CBSA approach than under through 1998. The very-low-income and low- the low-mod share of refinance mortgages has the old MSA approach. Because the results mod percentages were higher during 1999 to been about the same or even greater than that for 1999–2002 differed from the results for 2003 than they were during the earlier of home purchase mortgages. As shown in 2003, these two periods are considered period. In addition, when HUD wrote the Table D.8a, there was little difference in the separately. Under the historical data, the 2000 Rule, there had been five years (1994– very-low-income and low-mod shares of average low-mod share of the conventional 98) of solid affordable lending for lower- refinance and home purchase loans during conforming market was 44.4 percent for income borrowers. Now, with five additional 1995 and 1996. In 1997, 1999, and 2000, the home purchase loans (unweighted average of years of data for 1999–2003, there have been two lower-income shares (i.e., very-low- 1999–2002 percentages in Table D.8a); the ten years of strong affordable lending. income and low-mod shares) of refinance corresponding average with the projected Of course, it is recognized that lending mortgages were significantly higher than the data was 43.2 percent, yielding a differential patterns could change with sharp changes in lower-income shares of home purchase loans. of 1.2 percentage points. For total (both home interest rates and the economy. However, the To a certain extent, this pattern was purchase and refinance) loans, the average influenced by the growth of subprime loans, low-mod share of the conventional 37 The annual averages of the goals-qualifying which are mainly refinance loans. If B&C conforming market based on historical data mortgages reported in this appendix are unweighted loans are excluded from the market was 44.6 percent (unweighted average of averages; for analyses using weighted averages see definition, the home purchase and refinance 1999–2002 percentages); the corresponding Appendix A. percentages are approximately the same in average with the projected data was 43.4
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00275 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63854 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
percent, again yielding a differential of 1.2 appears that the low-mod share for single- conforming market will be at least one percentage points, with the same pattern family-owners in the conventional percentage point less due to the re- exhibited for the annual differentials.38 It benchmarking of area median incomes and projected low-mod percentages was 1.1 percentage the new OMB definitions of metropolitan 38 Between 1999 and 2002, the average single- point for Fannie Mae and 1.3 percentage point for areas. family-owner differential between the historical and Freddie Mac. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00276 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63855
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00277 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.091 63856 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Based on the above analysis of 1999–2002 2001, to 172,000 in 2002, and to 135,000 in been readily available for rental properties. data, it would appear the low-mod share of 2003. Still, the affordability of manufactured Where such income information is not the conventional conforming market is about homes for lower-income families is available, FHEFSSA provides that the rent of one percentage point less when based on demonstrated by their average price of a unit can be used to determine the projected data, as compared with historical $48,800 in 2001, a fraction of the median affordability of that unit and whether it data. However, the data for 2003 suggest a price for new ($175,000) and existing qualifies for the Low- and Moderate-Income different picture. As shown in Table D.8b, ($147,800) homes. Many households live in Goal. A unit qualifies for the Low- and the 2003 CBSA-based low-mod share for manufactured housing because they simply Moderate-Income Goal if the rent does not home purchase loans is 45.8 percent, which cannot afford site-built homes, for which the exceed 30 percent of the local area median is 1.1 percentage points higher than the construction costs per square foot are much income (with appropriate adjustments for corresponding MSA-based percentage of 44.7 higher. family size as measured by the number of percent. Similarly, the CBSA-based Although manufactured home loans cannot bedrooms). Thus, the GSEs’ performance percentage is 1.1 percentage point higher be identified in the HMDA data, Randy under the housing goals is measured in terms when all owner loans are considered. Thus, Scheessele at HUD identified 21 lenders that of the affordability of the rental dwelling the more recent 2003 data suggest that the primarily originated manufactured home units that are financed by mortgages that the GSEs will be scored higher than they have loans in 2001 and likely account for most of GSEs purchase; the income of the occupants historically been scored. these loans in the HMDA data for of these rental units is not considered in the Table A.18 in Appendix A reported similar metropolitan areas.39 HMDA data on home calculation of goal performance. For this MSA and CBSA data for home purchase loans originated by these lenders indicate reason, it is appropriate to base estimates of loans acquired by Fannie Mae and Freddie that: 40 market size on rent affordability data rather Mac. Again, the low-mod shares for the GSEs’ • A very high percentage of these loans— than on renter income data. purchases of both home purchase and total 75 percent in 2001—would qualify for the A rental unit is considered to be SFO loans were lower under the new CBSA Low- and Moderate-Income Goal, ‘‘affordable’’ to low- and moderate-income approach than under the old MSA approach • A substantial percentage of these loans— families, and thus qualifies for the Low- and for 1999–2002, but not for 2003. The 42 percent in 2001—would qualify for the Moderate-Income Goal, if that unit’s rent is proposed GSE rule accounted for the 1999– Special Affordable Goal, and equal to or less than 30 percent of area 2002 discrepancy by reducing the overall • Almost half of these loans—47 percent in median income. Table D.14 of Appendix D in low-mod estimates by one percentage point. 2001—would qualify for the Underserved HUD’s 2000 Rule reported AHS data on the Given the 2003 results, which show higher Areas Goal (defined in terms of the 1990 affordability of the rental housing stock for low-mod shares under the new CBSA Census data).41 the survey years between 1985 and 1997. The approach, that procedure is questionable. Thus an enhanced presence in this market 1997 AHS showed that for 1–4 unit This Final Rule follows a different procedure. by the GSEs would benefit many lower- unsubsidized single-family rental properties, The actual CBSA-based low-mod shares for income families. It would also contribute to 94 percent of all units and of units owners (reported in Table D.8b) are their presence in underserved rural areas, constructed in the preceding three years had incorporated directly into the analysis. especially in the South. gross rent (contract rent plus the cost of all The projection model will initially assume utilities) less than or equal to 30 percent of that refinancing is 35 percent of the single- 2. Low- and Moderate-Income Percentage for area median income. For multifamily family mortgage market; this will be followed Renter Mortgages unsubsidized rental properties, the by projection models that reflect heavy Following the 2000 Rule, measures of the corresponding figure was 92 percent. The refinance environments. Given the volatility rent affordability of the single-family rental AHS data for the other survey years were of refinance rates from year to year, it is and the multifamily rental markets are similar to the 1997 data. important to conduct sensitivity tests using obtained from the American Housing Survey b. Property Owners and Managers Survey different refinance rates. However, as (AHS) and the Property Owners and (POMS) explained in the preamble, HUD has Managers Survey (POMS). As explained included a provision in this Final Rule that below, the AHS provides rent information for As discussed in the 2000 GSE Rule, there eliminates the negative effects of heavy the stock of rental properties while the POMS were concerns about using AHS data on rents refinancing periods on the GSEs’ goals provides rent information for flow of from the outstanding rental stock to proxy performance. mortgages financing that stock. As discussed rents for newly mortgaged rental units. HUD investigated that issue further using the b. Manufactured Housing Loans below, the AHS and POMS data provide very similar estimates of the low- and moderate- POMS. Because manufactured housing loans are income share of the rental market. POMS Methodology. The affordability of such an important source of affordable multifamily and single-family rental housing housing, they are included in the mortgage a. American Housing Survey Data backing mortgages originated in 1993–1995 market definition in this appendix—or at The American Housing Survey does not was calculated using internal Census Bureau least that portion of the manufactured include data on mortgages for rental files from the American Housing Survey- housing market located in metropolitan areas properties; rather, it includes data on the National Sample (AHS) from 1995 and the is included, as HMDA doesn’t adequately characteristics of the existing rental housing Property Owners and Managers Survey from cover non-metropolitan areas. The GSEs have stock and recently completed rental 1995–1996. The POMS survey was questioned HUD’s including these loans in properties. Current data on the income of conducted on the same units included in the its market estimates; therefore, following the prospective or actual tenants has also not AHS survey, and provides supplemental same procedure used in the 2000 Rule and information such as the origination year of the 2004 proposed GSE Rule, this Appendix 39 See Randall M. Scheessele, 1998 HMDA the mortgage loan, if any, recorded against will report the effects of excluding Highlights, op. cit. and ‘‘HUD Subprime and the property included in the AHS survey. manufactured home loans from the market Manufactured Home Lender List’’ at http:// Monthly housing cost data (including rent estimates. As explained later, the effect of www.huduseer.org/datasets/manu.html. and utilities), number of bedrooms, and manufactured housing on HUD’s 40 Since most HMDA data are for loans in metropolitan area (MSA) location data were metropolitan area market estimate for each of metropolitan areas and a substantial share of obtained from the AHS file. the three housing goals is approximately one manufactured homes are located outside In cases where units in the AHS were not percentage point or less. metropolitan areas, HMDA data may not accurately occupied, the AHS typically provides rents, As discussed in Appendix A, the state the goals-qualifying shares for loans on either by obtaining this information from manufactured housing market increased manufactured homes in all areas. property owners or through the use of 41 rapidly during the 1990s, as units placed in While many fewer manufactured home loans imputation techniques. Estimated monthly were identified in the 2002 and 2003 HMDA data, serviced increased from 174,000 in 1991 to the loans showed similar goals-qualifying shares: housing costs on vacant units were therefore 374,000 in 1998. However, due to various low-mod (77.6 percent and 75.4 percent, calculated as the sum of AHS rent and utility problems in the industry such as lax respectively), special affordable (45.0 percent and costs estimated using utility allowances underwriting and repossessions, volume has 47.1 percent, respectively), and underserved areas published by HUD as part of its regulation of declined in recent years, falling to 192,000 in (46.9 percent and 45.2 percent, respectively). the GSEs. Observations where neither
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00278 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63857
monthly housing cost nor monthly rent was property share for single-family rental recalculated after applying the new data and available were omitted, as were observations properties). new OMB definitions back to 1999. Similar where MSA could not be determined. Units Based on its analysis of the AHS (see low-mod results were obtained for both with no cash rent and subsidized housing Fannie Mae Appendix, I–31–I–32), Fannie single-family and multifamily rentals. Thus, units were also omitted. Because of the Mae concluded that the low-mod shares for the 2000 Census data and the new OMB shortage of observations with 1995 both single-family and multifamily properties metropolitan area definitions will have no originations, POMS data on year of mortgage had fallen from 90 percent in 1997 to 86 impact on the low-mod scoring of the GSEs’ origination were utilized to restrict the percent in 2001. In its analysis, Fannie Mae rental purchases. sample to properties mortgaged during 1993– provides a weight of 0.07 to the low-mod Most of ICF’s and the GSEs’ concerns about 1995. POMS weights were then applied to share (74.8 percent) of recently-constructed HUD’s estimates of the affordability of rental estimate population statistics. Affordability single-family rental units in the AHS, and the housing properties related to the sizing of the calculations were made using 1993–95 area residual 0.93 weight to the low-mod share special affordable market. Therefore, more median incomes calculated by HUD. (91.8 percent) of the remaining existing units POMS Results. The rent affordability in the AHS. While Fannie Mae appears to use detail treatment of these issues will be estimates from POMS of the affordability of a low-mod share of 86 percent for single- provided in Section H below. newly-mortgaged rental properties are quite family rentals in its market sizing models, 3. Size of the Low- and Moderate-Income consistent with the AHS data on the applying these weights to the 2001 AHS data Mortgage Market affordability of the rental stock (discussed (reported by Fannie Mae in Table I.7 on p. This section provides estimates of the size above). Ninety-six (96) percent of single- I–32) yields approximately 90 percent for the of the low- and moderate-income mortgage family rental properties with new mortgages low-mod share of single-family rental market. Subsection 3.a presents new between 1993 and 1995 were affordable to properties. Similarly, for multifamily estimates of the low-mod market while low- and moderate-income families, as were properties, Fannie Mae provides a weight of Subsection 3.b reports the sensitivity of the 96 percent of newly-mortgaged multifamily 0.11 to the low-mod share (75.3 percent) of properties. Thus, these percentages for recently-constructed multifamily rental units new estimates to changes in assumptions newly-mortgaged properties from the POMS in the AHS, and the residual 0.89 weight to about economic and mortgage market are similar to those from the AHS for the the low-mod share (91.3 percent) of the conditions. rental stock. remaining existing units in the AHS. Again, a. Estimates of the Low- and Moderate- Further Results and Comments. The while Fannie Mae appears to use a low-mod Income Market share of 86 percent for multifamily rentals in baseline projection from HUD’s market share This section provides HUD’s estimates for its market sizing models, applying the above model assumes that 90 percent of newly- the size of the low- and moderate-income weights to the 2001 AHS data also yields mortgaged, single-family rental and mortgage market that will serve as a proxy for multifamily units are affordable to low- and approximately 90 percent for the low-mod the four-year period (2005–2008) when the moderate-income families.42 As noted above, share of multifamily rental properties. Since new housing goals will be in effect. The the analysis of AHS and POMS data from the single-family and multifamily rental units estimates are compared with recent mid-1990s supports the use of a 90 percent combined account for about 25 percent of all low-mod figure, and also supports using financed units in the market sizing models, experience in the low-mod market since rental stock data from the AHS as a proxy for the effect on the overall low-mod share of 1999. As discussed in Sections C and D, the affordability characteristics of new using 86 percent instead of 90 percent would market estimates will be presented for mortgages financing rental properties. be about one percentage point. (the 4.0 different combinations of the investor Updating these results using the 2001 and percentage point low-mod differential mortgage share (8.8, 8.5, 9.0, and 9.5) and the 2003 AHS produced similar (over 90 percent) multiplied by the 0.25 property share for multifamily mix (12.25, 13.5, 14.25, 15.0, and low-mod estimates for both the single-family single-family and multifamily rental 16.0). This range reflects uncertainty about rental stock and the multifamily rental stock. properties).43 Fannie Mae expressed the data and future conditions in these rental For example, using ICF’s assumptions for an particular concern with HUD’s Case 3, which markets. As discussed in Section C, HUD AHS analysis (see ICF Appendix, p. 45), the assumed an even higher 95.0 percent low- continues to use a multifamily (MF) mix of 2003 AHS showed that 94 (93) percent of mod share for rental properties; HUD has 15.0 percent as its baseline for a home single-family (multifamily) rental units reduced this assumption to 92.5 percent in purchase environment; this is strongly would qualify as being affordable to low- and the Case 3 analysis below. HUD’s Case 2 will supported by RFS analysis. While results are moderate-income families. While ICF used 90 also consider a low-mod percentage of 87.5 reported for Fannie Mae’s MF mix of 12.3 percent for multifamily, ICF concluded that percent. percent, HUD does not believe the MF mix 87.5 percent should be used for single-family The low-mod characteristics of the GSEs’ will fall to that level in a home purchase rentals. HUD’s updated analysis of the AHS, own purchases can also be examined. environment; rather, the results are reported which is explained in more detail in Section Between 1999 and 2003, 86.4 percent of to gauge the effects on the market size of H below, does not support using ICF’s 87.5 Fannie Mae’s single-family rental purchases alternative assumptions supported by Fannie percent assumption, except for sensitivity qualified as low-mod, as did 87.3 percent of Mae. Three alternative sets of projections analysis. Since single-family rental units Freddie Mac’s purchases. During the same about rental property low- and moderate- account for approximately 10 percent all period, 90.7 percent of Fannie Mae’s income percentages are given in Table D.9. financed units in both ICF’s and HUD’s multifamily rental purchases qualified as Case 1 projections represent the baseline and market share models, the effect on the overall low-mod, as did 92.6 percent of Freddie intermediate case; for example, it assumes low-mod goal of using 87.5 percent instead Mac’s purchases. One issue discussed below that the low-mod share of rental loans is 90 of 90.0 percent would be only 0.25 concerns the impact on the GSEs’ low-mod percent. Case 1 will be the focus of the percentage point. (the 2.5 percentage point performance of switching to 2000 Census market analysis in this section. Case 2 low-mod differential multiplied by the 0.10 data and the new OMB metropolitan area assumes slightly lower goals-qualifying definitions. The above GSE percentages were shares (e.g., an 85 percent low-mod share) for 42 In 2002, 75 percent of GSE purchases of single- rental properties while Case 3 assumes family rental units and 89 percent of their 43 Applying Fannie Mae’s weights to data from slightly higher goals-qualifying shares (e.g., a purchases of multifamily units qualified under the the 2003 AHS produces low-mod shares of slightly 92.5 percent low-mod share). Low- and Moderate-Income Goal, excluding the over 90 percent for both single-family and effects of missing data. multifamily rental properties. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00279 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63858 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00280 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.092
Because single-family-owner units account 8.5–9.0 percent range. For simplicity, the home purchase or low-refinancing for about 75 percent of all newly mortgaged combination of a 15.0-percent MF mix and a environments. After presenting these results, dwelling units, the low- and moderate- 8.5-percent investor share will be labeled the market estimates reflecting heavy refinance income percentage for owners is the most baseline when presenting the results below. environments will be presented. Because of important determinant of the total market Including a 9.0-percent investor mortgage the increase in single-family mortgages, the estimate. Thus, Table D.10 provides market share as the baseline would increase the low- multifamily share of the mortgage market estimates for different low-mod percentages mod market estimate by about 0.2–0.3 for the owner market as well as for different percentage point. The low-mod market typically falls during a heavy refinance MF mix percentages and investor mortgage estimates in Table D.10 exclude B&C loans, environment; therefore, several sensitivity shares. In a home purchase environment, the as explained below. analyses using lower multifamily mixes are most likely MF mix is 15.0 percent and the Table D.10 assumes a refinance rate of 35 examined below. most likely investor mortgage share is in the percent, which means that the table reflects
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00281 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63860 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00282 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.093 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63861
In the 2000 Rule, HUD assumed that the percent combined with a higher low-mod Table D.8b, the CBSA-based low-mod shares low-mod share of refinance loans was three refinance percentage of 51.3 percent. Of of home purchase loans averaged almost 44 percentage points lower than the low-mod course, the 47 percent low-mod share for the percent between 1999 and 2003, suggesting share of borrowers purchasing a home. overall single-family-owner market could be an overall low-mod goal of 55.4 percent However, as discussed earlier, the low-mod consistent with other combinations of low- under the baseline, with a range from 54.8 share of refinance loans has equaled or been mod shares for home purchase and refinance percent to 56.1 percent. The CBSA-based greater than the low-mod share of home loans. In this case, a 47 percent assumption measures of the low-mod share varied from purchase loans during recent home purchase for the overall single-family-owner market approximately 42 percent (41.8 percent in environments such as 1995–97 or 1999–2000; produces an estimate of 57.8 percent for the 2001) to almost 46 percent (45.8 percent in thus, the assumption of a lower low-mod low-mod share of the overall (owner and 2003). Under baseline assumptions, an owner shares for refinance loans is initially dropped rental) market, excluding B&C loans. share of 42 percent translates into a 53.8 for this analysis but will be reintroduced While both approaches will be discussed percent overall low-mod share while a 46 during the sensitivity analysis and during the below, most of the discussion will focus on percent owner figure translates into a 57.0 discussion of heavy refinance environments. the first approach. It should be noted that percent low-mod share. There are two ways to view the single- several low-mod percentages of the owner Case 2 (see Table D.9) considered a smaller family-owner low-mod percentages reported market are given in Table D.10 to account for low- and moderate-income percentage (85 in the first column of Table D.10. A first different perceptions of that market. percent) for both SF and MF rental approach would be to view them as Essentially, HUD’s approach throughout this properties, as compared with the baseline representing low-mod percentages of only the appendix is to provide several sensitivity Case 1, which assumed 90 percent. home purchase market. For example, a low- analyses to illustrate the effects of different Incorporating the Case 2 assumption reduces mod percentage for home purchase loans of views about the goals-qualifying share of the the low-mod market shares by about 1.3 43 percent—combined with the assumption single-family-owner market. This approach percentage points. For example, if the SFO of an equal low-mod share for refinance loans recognizes that there is some uncertainty in home purchase share is 45 percent, the (i.e., also 43 percent) and with the other the data and that there can be different overall low-mod market estimate is 54.9 model assumptions (such as a multifamily viewpoints about the various market percent under Case 2, as compared with 56.2 mix of 15 percent and an investor share of definitions and other model parameters. percent under Case 1 (see Table D.10). ICF 8.5 percent)—produces an estimate of 54.6 Market Estimates. Considering a 15.0- considered a different option, as it reduced percent for the low-mod share of the overall percent MF mix and a 8.5-percent investor only the SF rental percentage from 90.0 (owner and rental) market, excluding B&C mortgage share, the low-mod market percent to 87.5 percent. Since SF rental units loans. Thus, the reader can view Table D.10 estimates reported in Table D.10 are: 55.7 account for about 10 percent of all financed as showing the overall low-mod market percent if the owner percentage is 44.4 units, this change reduces the overall low- estimate once the reader specifies his or her percent (average home purchase share for mod market estimates by about 0.25 views about the low-mod share of the single- 1999–2003); 56.2 percent if the owner percentage points. As discussed earlier, the family home purchase market (given the percentage is 45 percent (home purchase baseline Case 1 assumption of 90 percent is other model assumptions). In this case, if the share for 1999, 2002, and 2003); 55.4 percent a reasonable approach for estimating the low- reader believes that the low-mod share of if the owner percentage is 44 percent (home mod market shares. refinance loans should be lower than that for purchase share for 2000); 54.6 percent if the Multifamily Mix. The volume of home purchase loans, the reader simply has owner percentage is 43 percent (home multifamily activity is also an important to multiply the differential amount by 0.35 purchase share for 1998 and 2001); and 53.8 determinant of the size of the low- and (which is the refinance share of single- percent if the owner percentage is 42 percent moderate-income market. HUD is aware of family-owner loans) and 0.745 (which is the (home purchase average from 1994–97). the uncertainty surrounding projections of single-family-owner share of all dwelling Considering a range of 13.5–16.0 for the MF the multifamily market and consequently units in the model that assumes a 15 percent mix and a range of 8.5–9.0 for the investor recognizes the need to conduct sensitivity multifamily mix and 8.5 percent investor mortgage share, the low-mod market analyses to determine the effects on the mortgage share). For example, applying the estimates reported in Table D.10 are: 55.6– overall market estimate of different assumption in the 2000 Rule that the low- 57.1 percent if the owner percentage is 45 assumptions about the size of that market. mod share is three percentage points lower percent; 54.8–56.1 percent if the owner Section C of this appendix provided HUD’s for refinance loans would reduce the overall percentage is 44 percent; 54.0–55.3 percent if rationale for its baseline MF mix of 15.0 low-mod share of the market by 0.78 the owner percentage is 43 percent; and percent and for its 13.5–16.0 percent range of percentage points (3.0 times 0.35 times 53.1–54.5 percent if the owner percentage is MF mixes. Assuming a 13.5 percent 0.745); if the low-mod share of refinance 42 percent. If the low-mod percent is at its multifamily mix reduces the overall low-mod loans is one percentage point below that of 1999–2003 average (44.4 percent), the market market estimates by 0.6–0.7 percentage home purchase loans, then the overall low- range is 54.3–56.9 percent. If the low- and points compared with a 15 percent mix, and mod market estimate falls by 0.26 percentage moderate income percentage for home by 1.0–1.2 percentage points compared with point. In this manner, the reader can easily purchase loans fell to 38 percent—or five a 16.0 percent mix. For example, when the adjust the market estimates reported in Table percentage points from its 1994–2003 average low-mod share of the home purchase market D.10 to incorporate his or her own views level of 43 percent—then the overall market is at 44 percent (its CBSA-based average for about differences in the low-mod share of estimate would be about 51 percent. Thus, 51 1999–2003), the low-mod share of the overall home purchase and refinance loans. percent is consistent with a rather significant market is 54.8 percent assuming a 13.5 A second approach would be to view the decline in the low-mod share of the single- percent multifamily mix, compared with 55.4 low-mod percentages (in the first column of family home purchase market. Under the (56.8) percent assuming a 15 (16.0) percent Table D.10) as representing low-mod shares baseline projection, the home purchase multifamily mix. for the overall single-family-owner market, percentage can fall as low as 36 percent— As shown in Table D.10, ICF’s MF mix of including both home purchase and refinance about four-fifths of the 1994–2003 average— 14.2 percent produces results intermediate loans. This approach does not specify and the low- and moderate-income market between HUD’s 13.5 percent and 15.0 separate low-mod percentages for home share would still be 49 percent. percent. Estimates of the low-mod market purchase and refinance loans, but rather Table D.8b reported so-called ‘‘CBSA- based on a MF mix of 14.2 percent are only focuses on the overall single-family-owner based’’ low-mod shares for single-family 0.3–0.4 percentage points less than those environments. Thus, it allows for mortgage owner loans that reflect the new 2000 Census based on a MF mix of 15.0 percent. market environments where the low-mod data and the new OMB metropolitan area Fannie Mae’s model combined an even share of refinance loans is greater than the definitions. Since these differed slightly from lower MF mix of 12.3 percent with an low-mod share for home purchase loans. For the historical ‘‘MSA-based’’ low-mod shares, investor mortgage share of 8.0 percent. If the example, a low-mod percentage for single- it is useful to repeat the above analysis in low-mod share of home purchase loans is 44 family-owner loans of 47 percent would terms of these new data, which will serve as percent (the average for 1999–2003), then the reflect the year 2000 environment, which had the basis for scoring the GSEs’ performance estimate for the overall low-mod market is a low-mod home purchase percentage of 44.3 under the new housing goals. As shown in 54.0 percent based on Fannie Mae’s
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00283 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63862 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
assumptions. In contrast, HUD’s estimates 1997, with an average of 57.4 percent.44 1999 and 2000, were higher than those (55.0 (with a MF mix of 15.0 percent and 8.5–9.0 Given (a)–(c) in the previous paragraph and to 56.4) for the earlier home purchase years, percent investor share) are 55.4–55.7 the fact that the HMDA-reported mortgage 1995–1997. When the data are expressed on percent—about one and a half percentage investor share was approximately eight a CBSA basis, the average low-mod shares for points higher. If the low-mod share of home percent during these three years (instead of 1999 and 2000 decline slightly to 56.0 purchase loans is 45 percent (which is below the assumed 10 percent), these estimates percent (with a two-percentage point lower the CBSA-based percentage of 45.8 for 2003), should be reduced by about one percentage MF mix than the baseline) and to 56.8 then Fannie Mae’s assumptions result in a point, placing their average at 56.4. Allowing percent (with the baseline MF mix). market estimate of 54.8 percent while HUD’s for a multifamily mix of three percentage By comparison, ICF’s best (lower bound) assumptions (see previous sentence) result in points below the baseline estimates (similar estimates for these home purchase years were market estimates of 56.2–56.5 percent. to the approach used for 1999 and 2000 52 (49) percent for 1996, 55 (52–53) percent Investor Mortgage Share. As shown in above) would drop the 1995–1997 low-mod for 1997 and 1999, 56 (53) percent for 1995, Table D.10, increasing the investor mortgage estimates by approximately 1.4 percentage and 57 (54) percent for 2000 (ICF Appendix, share by one percentage point from 8.0 points.45 Thus, the 1995–1997 average would p. 66). Emphasizing the variability of these estimates, ICF also reported numerous other percent to 9.0 percent increases the low-mod range from about 55.0 percent (with a MF low-mod shares for these years, based on market estimate by approximately 0.5–0.6 mix of three percentage points below the various simulations and assumptions. Some percentage point. If the 10.0 percent baseline baseline estimate) to 56.4 percent (with the 46 seem rather strange, or suggested that their from the 2004 proposed GSE rule were used baseline MF mix). For 2000, the baseline model assumed a analysis simply reduced the various input in this analysis, the market estimates would parameters to show that low estimates of the be approximately 0.6 (0.4) percentage points multifamily mix of 17.2 percent and a mortgage investor share of 9.1 percent. Under low-mod market could be the output. For higher relative to the results reported in these assumptions, the 2000 low-mod market example, ICF reports an overall market share Table D.10 for a baseline of 8.5 (9.0) percent. is estimated to be 57.9 percent. A lower MF of 46.9 percent share for 2000 (p. 66), which Examples of Home Purchase Years. The mix—for example, 16.0 (15.0) percent instead is about the same as the HMDA-reported above projection results for a home purchase of 17.2 percent—would reduce the estimated single-family-owner percentage of 47.0 environment can be compared with actual 2000 low-mod market share to 57.4 (57.0) percent for 2000 (Table D.8a); it is difficult results for the two most recent home percent. The baseline 57.9 percent estimate to imagine what scenario would result in the purchase years, 1999 and 2000, as well as for 2000 is about one percentage point lower low-mod share of the rental market being in results from earlier home purchase years than the 59.1 percent share reported in Table the less-than-fifty-percent range (although it (1995–1997). According to the Mortgage D.9 of the proposed rule, mainly for the is recognized that ICF was probably using an Bankers Association of America, the reasons discussed in the previous paragraph. owner share less than 47 percent). ICF’s refinance rate was 21 percent in 1995, 29 The above market estimates for 1999 and report is full of such low estimates (e.g., 46.4 percent in 1996 and 1997, 34 percent in 2000 are slightly lower if the projected CBSA percent for 1996 on page 67, another 49.6 1999, and 29 percent in 2000. data are used instead of the historical 1990- percent for 2000 on page 61) without any For 1999, the baseline model assumed a based MSA data. The projected CBSA-based attempt to justify them, other than to argue multifamily mix of 16.0 percent (see Section low-mod estimate was 56.2 percent for 1999, that everything is variable and possible—an C) and a mortgage investor share of 8.2 or 0.7 percentage points lower than the 56.9 approach that is not very convincing if it percent (see Section D). Under these percent estimate based on the historical MSA produces a 46.9 percent low-mod share for assumptions, the 1999 market estimate is data. In this case, the low-mod estimate falls the year 2000. 56.9 percent; if the 1999 MF mix was lower— to 55.8 (55.4) percent if the MF mix is 15.0 Heavy Refinancing Environments. The low- for example, 15.0 (14.0) percent instead of (14.0) percent. Incorporating the CBSA data mod share of the market will decline during 16.0 percent—then the estimate of the 1999 lowered the estimate for 2000 by 0.5 a period of heavy refinancing due to (a) a low-mod market share would be 56.4 (55.9) percentage points to 57.4 percent, and to 56.9 decline in the low-mod share of single-family percent. (56.5) percent if the MF mix is 16.0 (15.0) refinance mortgages as middle- and upper- The 2004 proposed rule (Table D.9 in percent. income borrowers dominate the refinance Appendix D) reported a higher baseline To summarize, the historical MSA-based market; (b) a decline in the relative market estimate for 1999 of 58.2 percent, as low-mod share for all recent home purchase importance of the subprime market; and (c) compared with the 56.9 percent reported in environments (1995–97 and 1999–2000) a decline in the share of multifamily the previous paragraph. The difference is averaged from 55.6 percent (with a two- to mortgages. For example, during 2002, the largely due to the treatment of single-family three-percentage point lower MF mix than refinance share of low-mod loans was 41.8 rental mortgages. For example, using the the baseline) to 56.8 percent (with the percent (compared with 47–51 percent proposed rule’s 10-percent assumption for baseline MF mix). The averages (56.5 to 57.4) during the two home purchase years of 1999 the mortgage investor share (instead of the for the two most recent home purchase years, and 2000); the subprime share of the single- lower 8.2 percent HMDA-based mortgage family market was 8.6 percent (compared investor shares reported in the text) would 44 These three estimates were initially reported in with 13 percent during 1999 and 2000); and increase the 1999 estimate to 57.7 percent, HUD’s 2000 Final Rule, and repeated in Table D.9 the multifamily share of the market was 11 percent or less (compared with 16 percent or only 0.5 percentage points lower than the of Appendix D of the 2004 proposed GSE rule. 45 more during 1999 and 2000). Although there 58.2 percent reported in the proposed rule. Given that the midpoints of the multifamily mixes for 1995–1997 are in the high 18–20 percent is some uncertainty with the data, the Other minor changes that lower the market range (see Table D.5b), three percentage points were multifamily mix for 2003 could have been as estimate included: (a) Further reducing the dropped in the sensitivity analysis. low as 6 or 7 percent. SF mortgage investor share by excluding B&C 46 To provide some confirmation for these 1995– Table D.11 shows the impact on the low- investor loans from the HMDA data (see 1997 estimates, HUD went back and re-estimated mod market share under different Section C); (b) using 1.6 percent (instead of the model for 1997. As shown in Table D.9 of the assumptions about a refinancing 2.0 percent) for the mortgage share of single- 2004 GSE Proposed Rule (as well as in Table D.15 of the 2000 GSE Rule), HUD had earlier estimated environment. The table reports the results for family 2–4 property owners; and (c) using a 65 percent refinance environment, which slightly lower dwelling-units-per-mortgage a low-mod share of 57.5 percent for 1997 (which was about the same as the 57.3-percent low-mod has been characteristic of recent (2002 and assumptions for SF 2–4 properties (2.20 share estimated for 1995 and 1996). With a lower 2003) refinance waves. Refinancing instead of 2.25) and for SF investor mortgages investor share (8.4 percent instead of 10.0 percent) environments are characterized by lower MF (1.30 instead of 1.35). and other changes mentioned in the text, the new mixes because single owner properties The above changes also affect the 1995-to- estimate for the 1997 low-mod market was 56.4 dominate the market; therefore Table D.11 1997 estimates reported in Table D.9 of assuming a multifamily mix of 19.3 percent. If the considers MF mixes from 6 to 12 percent. multifamily mix is reduced to 17.3 (16.3) percent, Appendix D of the proposed rule for the Most likely, a MF mix of 12–13 percent three home purchase environments prior to the low-mod share of the 1997 market is 55.5 (55.0) percent. The 55.0–56.4 percent range for 1997 is the characterized 2001, 9–11 percent 1999. These estimates were 57.3 percent for same as the range reported in the text for 1995– characterized 2002, and less than 7 percent both 1995 and 1996 and 57.5 percent for 1997. characterized 2003; there is some uncertainty
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00284 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63863
with these estimates, as discussed in Section mod percentages for the last four refinance percentage point differential between home C of this appendix. In a refinancing wave, the years (1998, 2001, 2002, and 2003)—43 purchase loans and refinance loans in a low-mod percent is typically lower for percent for home purchase loans and 40 heavy refinancing environment. This analysis refinance loans than home purchase loans, as percent for refinance loans; and (B) Scenario assumed an investor mortgage share of 8.0 middle- and high-income borrowers take B represents the average low-mod percent (average for these refinancing years) advantage of reduced interest rates. With percentages for the two most recent refinance and a subprime market share of 8.5 percent respect to the low-mod characteristics of SF years (2002, and 2003)—44.5 percent for (instead of the 12-percent assumption in the owner loans, two scenarios were considered: home purchase loans and 40.5 percent for baseline model). (A) Scenario A represents the average low- refinance loans. Thus, there is a 3–4 BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00285 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63864 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00286 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.094 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63865
Under Scenario A, the low-mod shares of 11 percent—would reduce the estimated of them equal or exceed 51 percent. In the varied by approximately three percentage 2002 low-mod market share to 53.1 (52.5) home purchase environment, estimates points, from 51.6 percent with a 12 percent percent. below 51 percent would require the low-mod MF mix to 48.9 percent with a 6 percent MF Using the projected CBSA data (instead of share of the single-family-owner market for mix. Under Scenario B, the low-mod the historical 1990-based MSA data) lowered home purchase loans to drop to 38 percent, percentages are all 0.7 percent higher, and the 2001 and 2002 low-mod estimates by which would be five percentage points below the pattern is from 52.3 percent with a MF approximately one percentage point. The the 1994–2003 average of 43 percent. Thus, mix of 12 percent to 49.6 percent with a MF 2001 market estimates are reduced to 52.3 51 percent is consistent with a rather mix of 6 percent. Notice that under Scenario percent (13.5 MF mix), 51.8 percent (12.5 MF significant decline in the low-mod share of B, the low-mod share remains in the 50–51 mix), and 51.6 percent (12.0 MF mix). The the single-family home purchase market. percent range even if the MF mix falls to 6– 2002 market estimates are reduced to 52.1 Sensitivity analyses of different refinance 8 percent. These low-mod market shares are percent (11.1 MF mix), 52.0 (10.5 MF mix), environments and model estimates for 1998, 4–7 percentage points lower than the low- and 51.4 percent (9.5 MF mix). and 2001–2003 suggest that it would require mod shares reported in Table D.10 for HUD’s By comparison, ICF’s best estimates for a particularly heavy period of refinancing to baseline home purchase environment. In these refinancing years are one or two fall below a 51-percent low-mod market addition to higher-income borrowers percentage points lower than the above share. dominating the single-family market, the estimates: 49.7 percent for 1998, 51.1 percent b. Economic Conditions and the Feasibility of share of the ‘‘goals rich’’ rental market for 2001, and 50.9 percent for 2002; because the Low- and Moderate-Income Housing Goal declines in a refinancing wave, which tends of the unavailability of 2003 HMDA data, no to further reduce the low-mod share of estimate was provided by ICF for that year. Commenters expressed a general concern market activity. The right-hand column of (See ICF Appendix, p. 60). ICF’s lower bound that the market share estimates and the Table D.11 shows that the rental share falls estimates for these three years were in the housing goals failed to recognize the to the 17–22 percent range, or 4–9 percentage 47–48 percent range. But as noted earlier, ICF volatility of housing markets and the points less that the almost 26-percent rental also produces a number of even lower existence of macroeconomic cycles. There share in HUD’s baseline model. estimates without discussion of what was particular concern that the market shares Model estimates were also made for the circumstances might lead to them—examples and housing goals were based on a period of economic expansion accompanied by record recent refinancing years of 1998, 2001, 2002, include their 45.2 percent market estimate for low interest rates and high housing and 2003. The Mortgage Bankers Association 2001 when the SFO low-mod share was 42.3 affordability. This section continues the of America estimated that the refinance rate percent (see Table D.8a) and their 44.9 discussion of these issues, noting that the was 50 percent in 1998, 55 percent in 2001, percent estimate for 2002 when the SFO low- Secretary can consider shifts in economic 59 percent in 2002, and 66 percent in 2003. mod share was 42.7 percent. (See ICF conditions when evaluating the performance The year 2003 stands out not only for its high Appendix, p. 66.) of the GSEs on the goals, and noting further rate of refinancing but also for the sheer For the years 1999 to 2002, Fannie Mae that the market share estimates can be volume of refinancing ($2.5 trillion), which estimated a low-mod market share of 52–53 examined in terms of less favorable market led to record single-family mortgage percent. (This is their estimate assuming no conditions than have existed during the 1993 originations ($3.8 trillion) that year. missing data; see their Table I.9, page I–34.) For 1998, the baseline model assumed a to 2003 period. As also explained below, This compares with HUD’s estimate of 53.7 HUD is publishing in the Federal Register an multifamily mix of 14.0 percent (see Section percent to 54.5 percent. As discussed in C) and a mortgage investor share of 6.8 Advance Notice of Proposed Rulemaking that Section C.6, Fannie Mae assumes a rather advises the public of HUD’s intention to percent (see Section D). Under these low MF mix (approximately 10 percent) in assumptions, the 1998 market estimate is consider by separate rulemaking a provision the model that generates its historical that recognizes and takes into consideration 51.9 percent. If the MF mix for 1998 had been estimates. 13.0 (12.0) percent, instead of the baseline of the impact of high volumes of refinance Given that HUD did not receive 2003 transactions on the GSEs’ ability to achieve 14.0 percent, then the estimated low-mod HMDA data until August 2004, it was not market share for 1998 would be 51.3 (50.8) the housing goals in certain years, and possible to develop a complete projection solicits proposals on how such a provision percent. For 2001, the baseline model model for 2003. Still, HUD developed some assumed a multifamily mix of 13.5 percent should be structured and implemented. rough projections for 2003. Given the huge Volatility of the Market. Changing and a mortgage investor share of 7.8 percent. volume of single-family originations ($3.8 Under these assumptions, the 2001 market economic conditions can affect the validity of trillion), the 1998 MF mix was likely rather HUD’s market estimates as well as the estimate is 53.4 percent. If the MF mix for low. In fact, Fannie Mae estimates the MF 2001 had been 12.5 (12.0) percent, instead of feasibility of the GSEs’ accomplishing the mix dropped to five percent in 2003. Thus, housing goals. The volatile nature of the the baseline of 13.5 percent, then the the estimates of the low-mod market share for estimated low-mod market share for 2001 mortgage market in the past few years 2003 are presented for different assumptions suggests a degree of uncertainty around would be 52.9 (52.7) percent. For 2002, the about the MF mix, recognizing that firm data baseline model assumed a multifamily mix of projections of the origination market. Large on the 2003 multifamily market are not swings in refinancing, consumers switching slightly over 11.0 percent and a mortgage available. Combining an investor mortgage investor share of 7.8 percent. Under these between adjustable-rate mortgages and fixed- share of 8.2 from HMDA (from HMDA) with rate mortgages, and increased first-time assumptions, the 2002 low-mod market is different MF mixes produces the following 47 homebuyer activity due to record low interest estimated to be 53.2 percent. A lower MF estimates: 51.9 percent (MF mix of 8 mix—for example, 10.5 (9.5) percent instead rates, have all characterized the mortgage percent); 51.4 percent (MF mix of 7 percent); market during the nineties. These conditions and 51.0 percent (MF mix of 6.0 percent). are beyond the control of the GSEs but they 47 The baseline estimates for 1998 (51.9 percent), As shown by both the simulation results would affect their performance on the 2001 (53.4 percent) and 2002 (53.2 percent) are and by the actual experience during 1998 and housing goals. A mortgage market dominated lower than those (53.8 percent, 54.9 percent and 2001–2003, the low-mod share declines 54.1 percent, respectively) reported in Table D.9 of by heavy refinancing on the part of middle- Appendix D of the proposed rule. As explained when refinances dominate the mortgage income homeowners would reduce the GSEs’ earlier, the differences between the results in the market. The above estimates place the low- ability to reach a specific target on the Low- proposed rule and this Final Rule are mainly due mod average during these four years of heavy and Moderate-Income Goal, for example. A to the treatment of single-family rental mortgages. refinancing at 52 percent, with practically all jump in interest rates would reduce the (In addition, the SF owner percentages for 2002 of the estimates of annual low-mod shares availability of very-low-income mortgages for were also lowered by approximately 0.5 percentage varying between 51 and 53 percent. As noted the GSEs to purchase. But on the other hand, point in the Final Rule.) Notice that in 1998, the above, the estimates for 2003 (around 51 the next few years may be favorable to investor mortgage share dropped to 6.8 percent, or percent) are somewhat speculative. 3.2 percentage points lower than that assumed in achieving the goals because of the high the proposed rule; this differential accounts for the The various market estimates presented in refinancing activity in 2001, 2002, and 2003. reduction of 1.9 percentage points (53.8 percent to Table D.10 for a home purchase environment A period of low-to-moderate interest rates 51.9 percent) in the low-mod market estimate for and reported above for a refinance would sustain affordability levels without 1998. environment are not all equally likely. Most causing the rush to refinance seen earlier in
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00287 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63866 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
1998 and 2001–2003. A high percentage of and economic conditions and the financial As discussed in Appendix A, the GSEs are potential refinancers have already done so, condition of the enterprise) the achievement involved in this market both through specific and are less likely to do so again. However, of the housing goal was or is feasible.’’ This program offerings and through purchases of these same predictions were made after the provision of FHEFSSA clearly allows for a securities backed by subprime loans 1998 refinance wave, which indicates the finding by HUD that a goal was not feasible (including B&C loans as well as A-minus uncertainty of making predictions about the due to market conditions, and no subsequent loans). The B&C loans experience much mortgage market. actions would be taken. As HUD noted in higher delinquency rates than A-minus Recent years have been characterized by both the 1995 and 2000 GSE Rules, it does loans.51 record affordability conditions due to low not set the housing goals so that they can be The market estimates reported in Section interest rates and economic expansion. Thus, met even under the worst of circumstances. F.3.a–b exclude the B&C portion of the as Section F.3.a indicates, HUD also Rather, as explained above, HUD has subprime market; or conversely, they include examined potential changes in the market conducted numerous sensitivity analyses for the A-minus portion of the subprime market. shares under very different macroeconomic economic and market affordability This section explains how these ‘‘adjusted’’ environments, including periods of environments much more adverse than has market shares are calculated from recession, high interest rates, and heavy existed in recent years. If macroeconomic ‘‘unadjusted’’ market shares that include B&C refinancing (accompanied by low interest conditions change even more dramatically, loans. rates). A recessionary environment would the levels of the goals can be revised to There are two possible approaches for likely be characterized by a reduction in reflect the changed conditions. FHEFSSA adjusting for the effects of B&C owner loans single-family activity (or an increase in the and HUD recognize that conditions could in the projection model. First, readers could multifamily share of the market) and a change in ways that require revised choose a single-family low-mod percentage reduction in the low-mod shares of the expectations. (that is, one of the percentages in the first single-family-owner market. The home HUD received a number of public column in Table D.10) that they believe is purchase percentage can fall as low as 36 comments seeking a regulatory solution to adjusted for B&C loans and then obtain a percent—about four-fifths of the 1994–2003 the issue of the ability of the GSEs to meet rough estimate of the overall market estimate average—and the low- and moderate-income the housing goals during a period when from the second to fourth columns market share would still be 49 percent. If the refinances of home mortgages constitute an corresponding to different multifamily mixes. low-mod share of the owner market were unusually large share of the mortgage market. For instance, if one believes the appropriate reduced more modestly to 39 percent, the As explained in the Preamble, HUD is not single-family-owner percentage adjusted for low-mod share for the overall market would addressing the refinance issue in this final B&C loans (or adjusted for any other market fall to 51.5 percent, assuming a multifamily rule. Elsewhere in this Federal Register, sectors that the reader thinks appropriate) is mix of 15.0 percent. (See Table D.10.) HUD is publishing an Advance Notice of 44 percent, then the low-mod market As discussed in Appendix A, record low Proposed Rulemaking that advises the public estimate is 55.4 percent assuming a interest rates, a more diverse socioeconomic of HUD’s intention to consider by separate multifamily mix of 15 percent. While group of households seeking rulemaking a provision that recognizes and intuitively appealing, such an approach homeownership, and affordability initiatives takes into consideration the impact of high would provide inaccurate results, as of the private sector have encouraged first- volumes of refinance transactions on the explained next. time buyers and low-income borrowers to GSEs’ ability to achieve the housing goals in Second, readers could choose a single- enter the market since the mid-1990s. Over certain years, and solicits proposals on how family-owner percentage directly from the past eight years, the affordable lending such a provision should be structured and HMDA data that is unadjusted for B&C loans market has demonstrated an underlying implemented. HUD believes that it would and then rely on HUD’s methodology strength that suggests it will continue, benefit from further consideration and (described below) for excluding the effects of particularly given demographic projections of additional public input on this issue. HUD B&C loans. This is the approach taken in increased minorities and immigrants in the also notes (see above) that FHEFSSA Table D.10. The advantage of the second mortgage market. However, a significant provides a mechanism by which HUD can approach is that HUD’s methodology makes increase in interest rates over recent levels take into consideration market and economic the appropriate adjustments to the various would reduce the presence of low-income conditions that may make the achievement of property shares (i.e., the owner versus rental families in the mortgage market and the housing goals infeasible in a given year. (See percentages) that result from excluding availability of low-income mortgages for 12 U.S.C. 1336(b)(e).) single-family B&C loans from the analysis. According to HUD’s methodology, dropping purchase by the GSEs. As noted above, the c. Treatment of B&C Loans and Other B&C loans would reduce the various low- 51–56 percent range for the low-mod market Technical Market Issues share covers economic and market mod market estimates by less than half of a affordability conditions much less favorable B&C Mortgages. As discussed in Appendix than recent conditions of low interest rates A, the market for subprime mortgages has (overestimates) the share of A-minus (B&C) loans. and economic expansion. The low-mod share experienced rapid growth over the past 6–7 According to data obtained by the Mortgage of the single-family home purchase market years, rising from an estimated $65 billion in Information Corporation (see next footnote), 57 could fall to 38 percent, which is five 1995 to $174 billion in 2001, $213 billion in percent of all subprime loans were labeled A-minus 2002 and $332 billion in 2003.49 In terms of (as of September 30, 2000). According to Inside B&C percentage points lower than its 1995–2003 Lending, which is published by Inside Mortgage average level of 43 percent, and the low-mod credit risk, subprime loans include a wide range of mortgage types. ‘‘A-minus’’ loans, Finance, the A-minus share of the subprime market market share would only be slightly below 51 was 61.6 percent in 2000, 70.7 percent in 2001 (see percent. The above analysis of 1998 and the which represent at least half of the subprime 50 March 11, 2002 issue), 75 percent in 2002 (see the 2001–2003 period suggests that 51 percent is market, make up the least risky category. September 15, 2003 issue), and 82 percent during a reasonable minimum low-mod share for the first nine months of 2003 (see the December 8, years of heavy refinancing. 49 Estimates of the subprime market for all years 2003 issue). A more recent analysis by Inside Feasibility Determination. As stated in the since 1995 are as follows (dollar and market share): Mortgage Finance found that 81.4 percent of 1995 ($65 billion, 10 percent); 1996 ($96.5 billion, subprime loans originated during the first quarter 2000 Rule, HUD is well aware of the 12.3 percent); 1997 ($125 billion, 15 percent); 1998 of 2002 were A-minus or better (see Inside B&C volatility of mortgage markets and the ($150 billion, 10 percent; 1999 ($160 billion, 12.5 Lending, Vol. 9, Issue 12, June 14, 2004). possible impacts on the GSEs’ ability to meet percent); 2000 ($138 billion, 12.1 percent); 2001 51 The Mortgage Information Corporation (MIC) the housing goals. FHEFSSA allows for ($174 billion, 8.5 percent); 2002 ($213 billion, 8.6 reports the following serious delinquency rates changing market conditions.48 If HUD has set percent), and 2003 ($332 billion, 8.7 percent). The (either 90 days past due or in foreclosure) by type a goal for a given year and market conditions uncertainty about what these various estimates of subprime loan: 3.36 percent for A-minus; 6.67 change dramatically during or prior to the include should be emphasized; for example, they percent for B; 9.22 percent for C; and 21.03 percent year, making it infeasible for the GSE to may include second mortgages and home equity for D. The D category accounted for only 2 percent loans as well as first mortgages, which are the focus attain the goal, HUD must determine of subprime loans and of course, is included in the of this analysis. The source for these estimates is ‘‘B&C’’ category referred to in this appendix. By ‘‘whether (taking into consideration market Inside Mortgage Finance (various years). comparison, MIC reports a seriously delinquent rate 50 The one-half assumption for A-minus loans is of 3.63 percent for FHA loans. See MIC, The Market 48 Section 1336(b)(3)(A). conservative because it probably underestimates Pulse, Winter 2001, page 6.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00288 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63867
percentage point. This minor effect is due to Applying the goals-qualifying percentages to owner loans (as well as the smaller number (a) the fact that the low-mod share of B&C the 590,489 owner B&C loans gives the of B&C rental units) leads to a reduction in loans is similar to that of the overall market; following estimates of B&C owner loans that the underserved areas market share of 1.0 and (b) the offsetting effects of the increase qualified for each of the housing goals: Low- percentage points, from 38.0 percent to 37.0 in the rental market share when single-family mod (346,027), special affordable (165,337), percent. (If this analysis were conducted in B&C loans are dropped from the market and underserved areas 614,109. The process terms of 1990-Census data, the one- totals. for the smaller number (48,998) of investor percentage point reduction would be from As noted above, if one assumes the single- B&C loans is similar. It is assumed that 90 about 33.0 percent to 32.0 percent.) family-owner percentages in the first column percent (44,098) of these B&C rental units Dropping B&C loans from HUD’s projection of Table D.10 are unadjusted for B&C loans, qualify for the low-mod goal, 58 percent model changes the mix between rental and then the overall low-mod market estimates (28,419) qualify for the special affordable owner units in the final market estimate; must be adjusted to exclude these loans. The goal, and 74 percent (36,259) qualify for the rental units accounted for 26.7 percent of effects of deducting the B&C loans from the underserved areas goal (based on 2000 total units after dropping B&C loans projection model can be illustrated using an Census data). compared with 25.6 percent before dropping example of a low-mod percentage of 44 Adjusting HUD’s model to exclude B&C B&C loans. Since practically all rental units percent for single-family-owner loans. Again, owner loans and B&C financed rental units qualify for the low-mod goal, their increased as explained earlier, this 44 percent figure involves subtracting the above eight figures— importance in the market partially offsets the could reflect a mortgage market environment two for the overall owner and rental B&C negative effects on the goals-qualifying shares where home purchase and refinance loans market and six for B&C owner units and of any reductions in B&C owner loans. Thus, had similar low-mod percentages (i.e., 44 rental units that qualify for each of the three another way of explaining why the goals- percent) or a mortgage market environment housing goals—from the corresponding qualifying market shares are not affected so where home purchase and refinance loans figures estimated by HUD for the total single- much by dropping B&C owner loans is that had different low-mod market percentages family and multifamily market inclusive of the rental share of the overall market that together resulted in a 44 percent average B&C owner loans and B&C dwelling units. increases as the B&C owner units are for the single-family-owner market. HUD’s model projects that 10,478,681 single- dropped from the market. Since rental units As Table D.10 shows, a 44 percent low- family and multifamily units will be have very high goals-qualifying percentages, mod share for owner mortgages translates financed; of these, 5,842,313 (55.8 percent) their increased importance in the market into an overall low-mod market share of 55.4 qualified for the low-mod goal, 2,801,179 partially offsets the negative effects on the goals-qualifying shares of any reductions in percent. It is assumed that the subprime (26.7 percent) for the special affordable goal, B&C owner loans. In fact, this rental mix market accounts for 12 percent of all and 3,983,005 (38.0 percent) for the effect would come into play with any mortgages originated, which would be $204 underserved areas goal. Deducting the B&C reduction in owner units from HUD’s model. billion based on $1,700 billion for the owner and rental market estimates produces A similar analysis can be used to mortgage market. This $204 billion estimate the following adjusted market estimates: A demonstrate the effects of deducting the for the subprime market is reduced by 20 total market of 9,839,193, of which 5,452,188 remaining, A-minus portion of the subprime percent to arrive at $163.2 billion for (55.4 percent) qualified for the low-mod goal, market from the market estimates. Of course, subprime loans that will be less than the 2,607,423 (26.5 percent) for the special deducting A-minus loans as well as B&C conforming loan limit. Dividing this figure by affordable goal, and 3,639,692 (37.0 percent) loans is equivalent to deducting all subprime the average loan amount for subprime loans for the 2000-based underserved areas goal. gives 1,256,361 subprime loans in the loans from the market. In the example given The low-mod market share estimate above (44 percent low-mod percentage for conventional market. HMDA data indicate exclusive of B&C loans (55.4 percent) is only that six percent of these are SF investor loans owners), deducting all subprime loans would slightly lower than the original market further reduce the overall low-mod market (75,382) and the remaining ones are SF estimate (55.8 percent from above), as is also owner loans (1,180,979). Since this analysis estimate to 55.0 percent. Thus, the the special affordable market estimate (26.7 unadjusted low-mod market estimate is 55.8 retains half of subprime loans (i.e., the A- percent versus 26.5 percent). This occurs minus portion of that market), these figures percent, the estimate adjusted for B&C loans because the B&C owner loans that were is 55.4 percent (reported in Table D.10), and are reduced by one-half to arrive at 590,489 dropped from the analysis have similar low- owner B&C loans and 37,691 investor B&C the estimate adjusted for all subprime loans mod and special affordable percentages as is 55.0 percent. loans. The investor loans are placed on a unit the overall (both single-family and basis by multiplying by 1.3 (units per As discussed in the 2000 Rule, there are multifamily) market. For example, the low- caveats that should be mentioned concerning mortgage), yielding 48,998 financed dwelling mod share of B&C loans was projected to be units in the investor B&C market. the above adjustments for the B&C market. 58.6 percent and HUD’s market model The adjustment for B&C loans depends on HMDA data was used to provide an (unadjusted for B&C loans) projected the estimate of the portion of the 590,489 owner several estimates relating to the single-family overall low-mod share to be practically the mortgage market, derived from various B&C loans that would qualify for each of the same, 55.8 percent. Thus, dropping B&C housing goals. HMDA data does not identify sources. Different estimates of the size of the owner loans from the market totals does not B&C market or the goals-qualifying shares of subprime loans, much less divide them into significantly reduce the overall low-mod their A-minus and B&C components. As the B&C market could lead to different share of the market. Because they qualify at estimates of the goals-qualifying shares for explained in Appendix A, Randall such a high rate (e.g., 90 percent on low- Scheessele in HUD’s Office of Policy the overall market. The goals-qualifying mod), dropping B&C rental loans tends to shares of the B&C market were based on Development and Research has identified reduce the market share estimates. However, almost 200 HMDA reporters that primarily HMDA data for selected lenders that they are relatively small in number—B&C primarily originate subprime loans; since originate subprime loans. Based on 1999– owner loans dominate the results because 2002 HMDA data, the goals-qualifying these lenders are likely originating both A- they account for 92.3 percent (590,489 minus and B&C loans, the goals-qualifying percentages of loans originated by these divided by 639,487) of the total B&C owner percentages used here may not be accurately subprime lenders were as follows: 58.6 and rental units dropped from the market measuring the goals-qualifying percentages percent qualified for the low-mod goal, 28.0 totals. for only B&C loans. The above technique of percent for the special affordable goal, and The situation is different for the dropping B&C loans also assumes that the 52.0 percent for the underserved areas goal.52 underserved areas goal. Underserved areas coverage of B&C and non-B&C loans in account for 52.0 percent of the B&C owner HMDA’s metropolitan area data is the same; 52 The goals-qualifying percentages for subprime loans, which is a higher percentage than the however, it is likely that HMDA coverage of lenders are much higher than the percentages for underserved area share of the overall market non-B&C loans is higher than its coverage of the overall single-family conventional conforming 53 market; for example, the 1999–2003 average low- (38.0 percent). Thus, dropping the B&C B&C loans. Despite these caveats, it also mod percentage for all single-family owner loans was 44 percent. For further analysis of subprime HF–009. Office of Policy Development and 53 Dropping B&C loans in the manner described lenders, see Randall M. Scheessele, 1998 HMDA Research, U.S. Department of Housing and Urban in the text results in the goals-qualifying Highlights, Housing Finance Working Paper No. Development, October 1999. Continued
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00289 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63868 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
appears that reasonably different estimates of The effects can be considered separately. increase or decrease the size of the low- and the various market parameters would not Dropping only manufactured housing loans moderate-income market during that period. likely change, in any significant way, the would reduce the market estimates by G. Size of the Conventional Conforming above estimates of the effects of excluding approximately three-quarters of a percentage Market Serving Central Cities, Rural Areas, B&C loans in calculating the goals-qualifying point. ICF argued that loans with less than and Other Underserved Areas shares of the market. As discussed in other $15,000 should be excluded. The impact of sections, HUD provides a range of estimates doing this on the market estimates would be The following discussion presents for the goals-qualifying market shares to less than half a percentage point. ICF also estimates of the size of the conventional account for uncertainty related to the various considered scenarios where one-half of conforming market for the Central City, Rural parameters included in its projection model manufactured loans would be dropped, as Areas, and other Underserved Areas Goal; for the mortgage market. well as small loans less than $15,000. The this housing goal will also be referred to as Manufactured Housing Loans and Small impact of doing this on the market estimates the Underserved Areas Goal. The first three Loans. HUD includes the effects of would be less than three-quarters of a sections, which analyze historical data going manufactured housing loans (at least those percentage point. back to the early 1990’s, necessarily used financing properties in metropolitan areas) in The estimated reductions in goals- 1990 Census geography to define its market estimates. However, sensitivity qualifying shares due to excluding underserved census tracts and underserved analyses are conducted to determine the counties. The first two sections focus on effects of excluding these loans. Excluding manufactured housing would be even lower during the heavy refinance years such as underserved census tracts in metropolitan manufactured housing loans as well as small areas, as Section 1 presents underserved area loans (loans less than $15,000) reduces the 1998 and 2001–2003. It should also be percentages for different property types while overall market estimates reported in Table mentioned that manufactured housing in Section 2 presents market estimates for D.10 by about one percentage point. This is non-metropolitan areas is not included in metropolitan areas. Section 3 discusses B&C estimated as follows. First, excluding these HUD’s analysis due to lack of data; including loans and rural areas. But as explained in loans reduces the low-mod percentage for that segment of the market would increase single-family-owner mortgages in the goals-qualifying shares of the overall Appendix B, HUD will be defining metropolitan areas by about 1.9 percentage market. Thus, the analyses of manufactured underserved areas based on 2000 Census points, based on analysis of recent home housing reported above and throughout the geography beginning in 2005, the first year purchase environments (1995–97 and 1999 this final rule pertain only to manufactured covered by this final rule. Therefore, Section and 2000). Multiplying this 1.9 percentage housing loans in metropolitan areas, as 4 repeats much of the analyses in Sections 1– point differential by the property share measured by loans originated by the 21 3 but in terms of 2000 Census geography, (0.745) of single-family-owner units yields manufactured housing lenders identified by rather than 1990 Census geography 1.4 percentage points, which serves as a Randy Scheessele at HUD. 1. Underserved Areas Goal Shares by proxy for the reduction in the overall low- The above analyses of the effects of less Property Type mod market share due to dropping affordable market conditions, different For purposes of the Underserved Areas manufactured home loans from the market assumptions about the size of the rental Goal, underserved areas in metropolitan analysis. The actual reduction will be market, and dropping different categories of areas are defined as census tracts with: somewhat less because dropping loans from the market definition suggest that (a) Tract median income at or below 90 manufactured home loans will increase the 51–56 percent is a reasonable range of percent of the MSA median income; or share of rental units, which increases the estimates for the low- and moderate-income (b) A minority composition equal to 30 overall low-mod market share, thus partially market. This range covers markets without offsetting the 1.4 percent reduction. The net percent or more and a tract median income B&C and allows for market environments that no more than 120 percent of MSA median effect is probably a reduction of about one would be much less affordable than recent percentage point. income. market conditions. The next section presents Owner Mortgages. The first set of numbers additional analyses related to market in Table D.12 are the percentages of single- percentages for the non-B&C market being volatility and affordability conditions. underestimated since HMDA coverage of B&C loans family-owner mortgages that financed is less than that of non-B&C loans and since B&C d. Conclusions About the Size of Low- and properties located in underserved census loans have higher goals-qualifying shares than non- Moderate-Income Market tracts of metropolitan areas between 1992 B&C loans. For instance, the low-mod shares of the Based on the above findings as well as and 2003. There are several interesting market reported in the text underestimate (to an numerous sensitivity analyses, HUD patterns in these data. During 1999 and 2000, unknown extent) the low-mod shares of the market 28–30 percent of mortgages (both home inclusive of B&C loans; so reducing the low-mod concludes that 51–56 percent is a reasonable range of estimates of the mortgage market’s purchase and refinance loans) financed owner shares by dropping B&C loans in the manner properties located in these areas; this described in the text would provide an low- and moderate-income share for the year underestimate of the low-mod share of the non-B&C 2005 and beyond. The range covers much percentage fell to 25.7 percent in 2001, 25.0 owner market. A study of 1997 HMDA data in more adverse economic and market percent in 2002, and 25.3 percent in 2003, Durham County, North Carolina by the Coalition for affordability conditions than have existed figures that were slightly below the average Responsible Lending (CRL) found that loans by recently, allows for different assumptions (26.8 percent) between 1994 and 1998. In mortgage and finance companies are often not 1992 and 1993, the underserved areas share reported to HMDA. For a summary of this study, about the single-family and multifamily rental markets, and excludes the effects of of single-family-owner mortgages was only see ‘‘Renewed Attack on Predatory Subprime 20 percent. Lenders’’ in Fair Lending/CRA Compass, June 9, B&C loans. HUD recognizes that shifts in 1999. economic conditions and refinancing could BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00290 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63869
In most years, refinance loans are more of home purchase loans. This 1.8 percentage properties in underserved areas, compared to likely than home purchase loans to finance point refinance-home-purchase differential is 25.1 (24.6) percent of home purchase loans. properties located in underserved census mostly due to the influence of subprime Thus, excluding B&C (subprime) loans tracts. Between 1994 and 2003, 27.3 percent loans. Excluding B&C (all subprime) loans reduces the differential from 1.8 percentage of refinance loans were for properties in and considering the same time period, 26.1 points to 1.0 (0.3) percentage point. In the underserved areas, compared to 25.5 percent (24.9) percent of refinance loans were for year (2000) with the largest differential,
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00291 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.095 63870 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
excluding B&C (all subprime) loans reduced market, it appears that the underserved percent). HUD’s model excludes B&C the refinance-home-purchase differential market has certainly maintained itself at a investor loans in the same manner it from 8.1 percent to 6.9 (5.7) percent; in this high level over the past four years. excludes B&C owner loans (see earlier case, a significant differential remained after Renter Mortgages. The second and third explanation). excluding B&C (subprime) loans. In the sets of numbers in Table D.12 are the heavy refinance years of 1998, 2001, 2002, underserved area percentages for single- 2. Market Estimates for Underserved Areas in and 2003 underserved areas accounted for family rental mortgages and multifamily Metropolitan Areas about 25 percent of total (both home mortgages, respectively. Based on HMDA Table D.13 reports HUD’s estimates of the purchase and refinance) owner loans. data for single-family, non-owner-occupied market share for underserved areas based on The underserved areas share for home (i.e., investor) loans, the underserved area the projection model discussed earlier. The purchase loans has been in the 25–26 percent share of newly-mortgaged single-family estimates in Table D.13 exclude the effects of range since 1995, except for 2000 and 2002 rental mortgages has averaged about 44 B&C owner loans and B&C investor loans. when it increased to over 27 percent, and in percent (over nine or ten years). HMDA data The percentage of single-family-owner 2003 when it increased to 28.5 percent. also show that about half of newly-mortgaged mortgages financing properties in Considering all (both home purchase and multifamily rental units are located in underserved areas is the most important refinance) loans during recent ‘‘home underserved areas. HUD’s baseline assumes determinant of the overall market share for purchase’’ environments, the underserved that 42.5 percent of single-family investor this goal. Therefore, Table D.13 reports areas share was a high 28–30 percent during loans and 48 percent of multifamily loans are market shares for different single-family- 1999–2000, compared with a 27 percent located in underserved areas. The GSEs and owner percentages ranging from 30 percent average between 1995 and 1997; excluding ICF argued that HUD had overstated these B&C and other (i.e. A-minus) subprime loans underserved area percentages; Section G.4 (2000 level) to 20 percent (1993 level) to 19 places 1999 on par with the earlier years, below, which focuses on the 2000-based percent. Considering a 15.0-percent MF mix with only the year 2000 showing a higher underserved area percentage, will discuss and a 8.5-percent investor mortgage share, level of underserved area lending than and respond to their concerns. Fannie Mae the market share estimate is 31–32 percent if occurred during 1995–97. These data also said that subprime (or B&C) loans should the overall (both home purchase and indicate that the single-family-owner market be taken out of the SF investor loans. As refinance) single-family-owner percentage for in underserved areas has remained strong shown in Table D.12, deducting B&C loans underserved areas is at its 1994–2003 HMDA since the 2000 Rule was written. While it is reduces the underserved area percentage for average of 26.6 percent. The overall market recognized that economic and housing SF investor mortgages by almost one share for underserved areas peaks at 35 affordability conditions could change and percentage point (the 1993–2003 unweighted percent when the single-family-owner reduce the size of the underserved areas average falls from 44.0 percent to 43.1 percentage is at its 2000 level of 30 percent.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00292 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63871
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00293 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.096 63872 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
The analysis can also be conducted in 15.0 (14.0), then the estimate of the 1999 Additional sensitivity analyses were terms of the home purchase percentages underserved areas market share would be conducted to reflect the volatility of the reported in Table D.13. Again, considering a only slightly lower at 32.9 (32.6) percent.54 economy and mortgage market. Recession 15.0-percent MF mix and an 8.5-percent For 2000, the baseline model assumed a and high interest rate scenarios assumed a investor mortgage share, the underserved multifamily mix of 17.2 percent and a significant drop in the underserved area area market estimates reported in Table D.13 mortgage investor share of 9.1 percent. Under percentage for single-family-owner are: 33.3 percent if the owner percentage is these assumptions, the 2000 underserved mortgages. The single-family-owner 28.5 percent (home purchase share for 2003); areas market is estimated to be 34.9 percent. percentage can go as low as 24 percent— 32.1 if the owner percentage is 27 percent A lower MF mix of 16.0 (15.0) percent would which is 3 percentage points lower than the (home purchase share in 2000 and 2002 reduce the estimated 2000 underserved areas 1995–2003 average of 27 percent—and the slightly above the 1999–2003 average home market share slightly to 34.6 (34.4) percent.55 estimated market share for underserved areas purchase share of 26.8 percent); 31.3 percent The heavy refinance scenarios discussed remains at almost 30 percent. In a more if the owner percentage is 26 percent (home for the low-mod market were also projected severe case, the overall underserved market purchase share for 1999 and 2001); and 30.5 for the underserved areas market. Since the share would be 27.5 percent if the single- percent if the owner percentage is 25 percent impact of a heavy refinancing period on the family-owner share fell to 21 percent (its (home purchase average from 1994–98). This underserved areas market share will be 1992 level), which is 7–9 percentage points analysis assumes that the underserved areas covered in Section G.4, which incorporates lower than its 1999–2000 levels. share of refinance loans is the same as those 2000 Census data, there is no need for a 3. Adjustments: B&C Loans, the Rural listed above for home purchase loans. But, as detailed discussion in this section’s analysis Underserved Areas Market, and Table D.12 shows, the underserved areas based on 1990 Census data. Still, it is useful Manufactured Housing Loans share of refinance loans tends to be higher to provide a quick review of the 1990-based than that for home purchase loans. And in underserved area estimates for three heavy B&C Loans. The procedure for dropping the year 2000, the overall underserved areas refinancing environments (1998, 2001, 2002, B&C loans from the projections is the same share for owner loans reached 30 percent; as and 2003). For 1998, the baseline model as described in Section F.3.b for the Low- noted in the previous paragraph, the overall assumed a multifamily mix of 14.0 percent and Moderate-Income Goal. The underserved market estimate is 34.6 percent in this case. and a mortgage investor share of 6.8 percent. area percentage for B&C loans is 44.5 percent, However, the next highest overall owner Under these assumptions, the 1998 market which is much higher than the projected share is the 28.2 percent share in 1999, estimate is 29.9 percent. If the MF mix for percentage for the overall market (which peaks at 35 percent as indicated in Table which yields a market estimate of 1998 had been 12.0 percent, instead of the D.13). Thus, dropping B&C loans will reduce approximately 33 percent. baseline of 14.0 percent, then the estimated the overall market estimates. Consider the Sensitivity Analyses. Unlike the Low- and underserved area market share for 1998 case of a single-family-owner percentage of Moderate-Income and Special Affordable would be 29.4 percent. For 2001, the baseline 27 percent, which yields an overall market Goals, the market estimates differ only model assumed a multifamily mix of 13.5 estimate for underserved areas of 33.1 slightly as one moves from a 13.5 percent MF percent and a mortgage investor share of 7.8 percent, including B&C loans. When B&C mix to 16.0 percent MF mix. For example, percent. Under these assumptions, the 2001 loans are excluded from the projection reducing the assumed multifamily mix from market estimate is 32.1 percent, dropping to model, the underserved areas market share 16.0 percent to 13.5 percent reduces the 31.7 percent if the MF mix was 12.0 percent. For 2002, the baseline model assumed a falls by 0.9 percentage points to 32.2 percent, overall market projection for underserved which is the figure reported in Table D.13. multifamily mix of slightly over 11.0 percent areas by only 0.5–0.6 percentage points. This Non-metropolitan Areas. Underserved and a mortgage investor share of 7.8 percent. is because the underserved area differentials rural areas are non-metropolitan counties Under these assumptions, the 2002 between owner and rental properties are not with: underserved areas market is estimated to be as large as the low- and moderate-income (a) County median income at or below 95 31.6 percent, dropping to 31.1 percent if the differentials reported earlier. percent of the greater of statewide non- MF mix is 9.5 percent. This analysis suggests Similarly, the market estimates differ only metropolitan median income or nationwide that the underserved areas market based on slightly with changes in the investor non-metropolitan income; or 1990 Census data will be about 29–32 mortgage share. Reducing the investor mix (b) A minority composition equal to 30 from 9.5 percent to 8.0 percent reduces the percent range during periods of heavy 56 percent or more and a county median income overall market projection for underserved refinancing. no more that 120 percent of statewide non- areas by only 0.2–0.3 percentage points. metropolitan median income. Case 2 (see Table D.9) considered slightly 54 Table D.15 of the 2000 GSE Rule also reported HMDA’s limited coverage of mortgage data smaller underserved area percentages for underserved area shares of 33.9 percent for 1995 in non-metropolitan counties makes it rental properties (40 percent for SF rentals and 1997 and 33.4 percent for 1996. These estimates, after adjustments for a lower HMDA- impossible to estimate the size of the and 46 percent for MF rentals), as compared mortgage market in rural areas. However, all with the baseline Case 1, which assumed based mortgage investor share and a lower-than- baseline MF mix, would still remain in the 32–33 indicators suggest that underserved counties 42.5 percent and 46.0 percent, respectively. percent range. To provide some confirmation for in non-metropolitan areas comprise a larger Incorporating these Case 2 assumptions this, HUD went back and re-estimated the model for share of the non-metropolitan mortgage reduces the underserved areas market 1997. As shown in Table D.15 of the 2000 GSE market than the underserved census tracts in estimate by only 0.6 percentage points. For Rule, HUD had earlier estimated an underserved metropolitan areas comprise of the example, if the SFO home purchase share is areas share of 33.9 percent for 1997 (which was the metropolitan mortgage market. For instance, 28 percent, then the overall underserved area same as the 33.9-percent underserved areas estimate underserved counties within rural areas for 1995 and similar to the 33.4-percent estimate for estimate is 32.3 percent under Case 2, as include 54 percent of non-metropolitan compared with 32.9 percent under Case 1 1996). With a lower investor share (8.4 percent instead of 10.0 percent) and other changes homeowners; on the other hand, underserved (see Table D.13). mentioned in the text, the new estimate for the 1997 census tracts in metropolitan areas account Examples of Home Purchase and underserved areas market was 32.7 assuming a for only 34 percent of metropolitan Refinance Environments. The above multifamily mix of 19.3 percent. If the multifamily homeowners. projection results for a home purchase mix is reduced to 17.3 (16.3) percent, the During 1999–2003, 38.3 percent of the environment can be compared with actual underserved areas share of the 1997 market is 32.3 GSEs’ single-family-owner (SFO) purchases results for two home purchase years, 1999 (32.0) percent. Thus, this 32.0–32.7 percent range in non-metropolitan areas were in and 2000 (see earlier description of these two for 1997 is consistent with a 32–33 percent range underserved counties while 23.1 percent of for 1995–1997. years in the low-mod section, F.3.a). For 55 their SFO purchases in metropolitan areas 1999, the baseline model assumed a The baseline 34.9 percent estimate for 2000 is 0.4 percentage points lower than the 35.3 percent were in underserved census tracts. These multifamily mix of 16.0 percent and a share reported in Table D.9 of the proposed rule. figures suggest the market share for mortgage investor share of 8.2 percent. Under The difference is mostly explained by the different these assumptions, the projected 1999 market treatment of single-family rental mortgages. percent. (See their Table I.9, page I–34.) This estimate (based on 1990-Census data) is 33.1 56 For the years 1999 to 2002, Fannie Mae compares with HUD’s estimate of 32.5 percent to percent; if the 1999 MF mix was lower at estimated an underserved areas share of 32–33 32.9 percent for the same period.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00294 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63873
underserved counties in rural areas is higher adjustment of 1.25 percentage points was range for the Geographically Targeted Goal than the market share for underserved census made in Table D.13 for the 2005–2008 based on past origination activity in tracts in metropolitan areas. Thus, using a projection model. The non-metropolitan area underserved areas and on scenarios that metropolitan estimate to proxy the overall issue will be discussed further in Section G.4 cover a variety of economic and mortgage market for this goal, including rural areas, is below, which incorporates the effects of the market conditions. That analysis, which conservative.57 new 2000 Census data. included historical data going back to the The limited HMDA data available for non- Small Loans and Manufactured Housing early 1990’s, necessarily used 1990 Census metropolitan counties also suggest that the Loans. Excluding manufactured housing geography to define underserved census loans and small loans (less than $15,000) underserved areas market estimate would be tracts. As explained in Appendix B, HUD reduces the overall underserved area market higher if complete data for non-metropolitan will be defining underserved areas based on counties were available. According to estimates reported in Table D.13 by less than 2000 Census geography beginning in 2005, HMDA, underserved counties accounted for one percentage point. This is estimated as the first year covered by this final rule. 41.6 percent of SFO mortgages originated in follows. First, excluding these loans reduces Appendix B also explains that the number of non-metropolitan areas between 1999 and the unadjusted underserved areas percentage census tracts in metropolitan areas covered 2003. By contrast, underserved census tracts for single-family-owner mortgages in accounted for approximately 24.9 percent of metropolitan areas by about 1.2 percentage by HUD’s underserved area definition will SFO mortgages originated in metropolitan points, based on analysis of recent home increase from 21,587 tracts (based on 1990 areas between 1999 and 2003.58 Since non- purchase environments (1995–97 and 1999 Census) to 26,959 tracts (based on 2000 metropolitan areas account for 13 percent of and 2000). Multiplying this 1.2 percentage Census and OMB’s respecification of all single-family-owner mortgages 59 and point differential by the property share of metropolitan areas). This increase in the estimating that the single-family-owner single-family-owner units (74.5 percent) number of tracts defined as underserved market accounts for 74.5 percent of newly- yields 0.9 percentage points, which serves as means that the market estimate for the mortgaged dwelling units, then the non- a proxy for the reduction in the overall Geographically Targeted Goal will be about metropolitan underserved area differential of underserved area market share due to five percentage points higher than the 1990- 16.7 percent would raise the overall market dropping manufactured home loans from the based market estimate. Thus, this section estimate by 1.6 percentage point—16.7 market analysis. The actual reduction will be provides a new range of market estimates for percentage points times 0.13 (non- somewhat less because dropping underserved areas defined in terms of 2000 manufactured home loans will increase the metropolitan area mortgage market share) Census data. share of rental units, which increases the times 0.745 (single-family owner mortgage For the years 1999 to 2003, Table D.14a. overall underserved areas market share, thus market share). Based on this calculation, if reports the underserved areas share of the partially offsetting the 0.8 percent reduction. the 16.7 point differential reflected actual mortgage market for single-family-owner, The net effect is probably a reduction of investor (non-owner), and multifamily market conditions, then the underserved about three-quarters of a percentage point. properties, with comparisons between 1990- areas market share estimated using The small loan and manufactured housing metropolitan area data should be increased effects can be considered separately. based and 2000-based measures of by 1.6 percentage points to account for the Dropping only manufactured housing loans underserved areas. HMDA data, which is the effects of underserved counties in non- would reduce the market estimates by source of the mortgage data, were reported in 60 metropolitan areas. A more conservative approximately three-fourths of a percentage terms of 1990 census tracts. For the years point. ICF argued that loans with less than 1999 to 2002, HUD used various 57 Between 1999 and 2001, the non-metropolitan $15,000 should be excluded. The impact of apportionment techniques to re-allocate portion of the Underserved Areas Goal contributed doing this on the market estimates would be 1990-based HMDA mortgage data into census 1.1 to 1.4 (0.7 to 1.3) percentage points to Freddie about one-third of a percentage point. ICF tracts as defined by the 2000 Census; 2003 Mac’s (Fannie Mae’s) overall performance (i.e., also considered scenarios where one-half of HMDA data were defined in terms of 2000 including both metro and non-metro loans), manufactured loans would be dropped, as Census tracts, so no reallocation was compared with a goals-counting system that only included metropolitan areas. well as small loans less than $15,000. The required. The 1990-based underserved area 58 These data do not include loans originated by impact of doing this on the market estimates market shares reported in Table D.14.a. are lenders that specialize in manufactured housing would be three-fifths of a percentage point. the same data reported earlier in Table D.12, loans, as well as estimated B&C loans. The next section discusses changes as a while the 2000-based underserved area 59 Federal Housing Finance Board data. result of switching from 1990 to 2000 Census market shares result from re-allocating 1999– 60 Mortgage Interest Rate Survey (MIRS) data geography. 2002 HMDA data into 2000 Census reported by the Federal Housing Finance Board 4. 2000-Based Underserved Area Market geography. In addition, the data are defined separate conventional home purchase loans by their Shares in terms of the new OMB metropolitan area metropolitan and non-metropolitan location. The definitions. average non-metropolitan share between 1999 and The above analysis has concluded that 29– 2002 was about 13 percent. 34 percent would be a reasonable market BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00295 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63874 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00296 Fmt 4701 Sfmt 4725 C:\02NOR2.SGM 02NOR2 ER02NO04.097
Single-Family-Owner Loans. First, consider respect to the underserved areas shares of with a mortgage, (c) all SF rental properties the market shares for single-family-owner single-family rental properties: 52.0% for with a conventional conforming mortgage, (d) properties in the top portion of Table D.14a. Case 1 (baseline), 50.0% for Case 2, and all SF Rental properties with a new first In 2002, the underserved area percentage for 54.0% for Case 3. With respect to multifamily mortgage, and (e) all SF rental properties home purchase loans increases from 27.1 properties, the following assumptions were with a new conventional conforming first percent (1990-based) to 32.8 percent (2000- made with respect to underserved areas mortgage. The underserved areas share for based), an increase of 5.7 percentage points; shares: 58.0% for Case 1 (baseline), 56.0% for each of the groups of SF rental properties was the corresponding percentages for refinance Case 2, and 59.0% for Case 3. ICF criticized approximately 50 percent. Adding a five loans were 24.2 percent (1990-based) and HUD’s baseline assumptions (52 percent for percent adjustment to reflect 2000-based 29.4 percent (2000-based), or an increase of SF investors and 58 percent for MF rentals) geography (see Table D.14a) would increase 5.2 percentage points. Considering total as being too high.61 ICF’s best estimate was owner loans (i.e., both home purchase and 50 percent for SF investors and 55 percent for these estimates to 55 percent. While this refinance owner loans), the weighted average MF rentals.62 Since SF rentals account for information is dated, it is consistent with of the ‘‘Differences’’ reported in Table D.14a. 10.6 percent of financed units, reducing the HUD’s 52.0 percent baseline and its 54.0 is 5.4 percentage points in 2002 for the underserved area share by two percentage percent assumption in Case 3. Abt Associates conforming market. Between 1999 and 2003, points from HUD’s 52 percent to ICF’s 50 also reported similar data for MF rental 30.3 percent of mortgage originations were percent would reduce the overall categories (a)–(c). In this case the originated in underserved areas based on underserved areas goal by 0.21 percentage underserved areas share ranged from 51–54 2000 geography, compared with 25.2 percent point. Since MF rentals account for 15.0 percent; adding 7–8 percent adjustment to based on 1990 geography—yielding the percent of financed units (in HUD’s baseline reflect 2000-based geography would increase overall differential of 5.1 percentage points. model), reducing the underserved area share these estimates to 55–62 percent, again (The unweighted 1999–2003 differential is by three percentage points from HUD’s 58 providing support for HUD’s baseline (58 4.9 percent.) percent to ICF’s 55 percent would reduce the percent) and Case 3 (59 percent) The first column of Table D.14a. reports overall underserved areas goal by an assumptions.64 the 2000-based underserved areas share for additional 0.45 percentage point. Thus, the HUD had Census Bureau staff use the home purchase loans for the years, 1999 to combined effect of ICF’s assumptions would geocoded 2003 AHS file to calculate the 2003. The share was about 31 percent in 1999 be a 0.66 percentage point reduction in the distribution of the rental housing stock across and 2001 and in the 32.6–33.7 percent range underserved areas goal. Fannie Mae did not served and underserved areas. This analysis, during 2000, 2002, and 2003. Notice that the comment directly on this parameter other which was conducted in terms of 1990- peak share (33.7 percent) for home purchase than to emphasize that HUD’s Case 2 is the Census geography, showed that 55.8 percent loans occurred in the most recent year, 2003. ‘‘most likely set of assumptions’’ for of the SF rental housing stock was located in It should be recalled that there was no need estimating the underserved areas share to re-apportion the 2003 data from 1990- (Fannie Mae Appendix, p. I–38). HUD’s Case underserved areas, as was 51.4 percent of the based tracts to 2000-based tracts, as these 2 (see above) would drop the baseline MF rental housing stock. Adding a five (7– 2003 data were already defined in terms of underserved area share for both SF and MF 8) percent adjustment to reflect 2000-based 2000 census geography. Whether this fact by two percentage points; therefore, Fannie geography would increase these SF (MF) affects the various differentials between 2003 Mae’s assumptions are similar to ICF’s. rental estimates to 60.8 (58.4–59.4) percent. and earlier years is not clear. The years 1999 In this analysis supporting the Final Rule, HUD also had Census Bureau staff use the and 2000 exhibited higher underserved area HUD is retaining the same underserved areas geocoded, 2001 Residential Finance Survey shares for refinance loans than for home shares for SF and MF rental properties that (RFS) to calculate the distribution of rental purchase loans; as discussed earlier, this it used in the 2004 proposed GSE rule. HUD mortgages and financed units across served pattern was largely, but not entirely, due to conducted several additional analyses that and underserved areas. (See Table D.14b.) subprime refinance loans. support its SF rental baseline of 52 percent Unlike the AHS analysis mentioned above, Single-Family Rental and Multifamily and its MF rental baseline of 58 percent. this analysis was conducted in terms of 2000 Loans. Next, consider the underserved area These analyses are summarized below. Census geography. In 2001, 54.1 percent of market shares reported for single-family A report by Abt Associates 63 calculated newly-mortgaged SF rental units were rental (or non-owner) and multifamily 1990-based underserved areas shares using located in underserved areas, as were 61.5 of properties in the middle and bottom portions the 1995 AHS and POMS data, for (a) all SF newly mortgaged MF rental units. Similar of Table D.14a. In 2002, the underserved area rental properties, (b) all SF rental properties underserved area percentages were obtained percentage for home purchase investor loans for SF investor and MF loans that were increases from 42.0 percent (1990-based) to 61 ICF incorrectly said HUD’s baseline originated in 1999 and 2000 and still 47.7 percent (2000-based), an increase of 5.7 underserved areas share for MF rentals was 60 percentage points; the corresponding surviving at the time of the RFS survey in percent, rather than 58 percent (ICF Appendix, p. 2001. 65 percentages for refinance loans were 45.6 47). percent (1990-based) and 50.8 percent (2000- 62 Freddie Mac says ‘‘ICF estimates the 64 based), or an increase of 5.3 percentage multifamily underserved share to be just 56 percent As shown in Table D.12, excluding B&C points. The multifamily differentials are and the single-family renter underserved area share investor loans reduces the market’s underserved area share for SF investor loans. An adjustment for somewhat higher at approximately 7–8 to be just 50 percent’’ (at Appendix IV–24). However, ICF uses a 50 percent share in its B&C investor loans is made within HUD’s model, percentage points. Between 1999 and 2003, along the same lines as that B&C owner loans are 60 percent (unweighted average) of projection model (ICF Appendix, p. 133); therefore, 55 percent is used here as the ICF number. Also, excluded from the analysis. See Section F.3.c for multifamily originations were originated in ICF’s lower (upper bound) projection was 47 (53) further explanation. underserved areas based on 2000 geography, percent for SF rental properties and 56 (58) percent 65 It is encouraging that the RFS underserved area compared with 52.6 percent based on 1990 for multifamily properties. percentage (31.7 percent) for SF-owner mortgages geography. 63 ‘‘Affordability and Geographic Distribution of originated in metropolitan areas during 2001 was In the 2004 proposed GSE Rule, HUD made the Housing Stock and the Use of Mortgage similar to the corresponding percentage (31.0 the following 2000-based assumptions with Finance,’’ Abt Associates, October 22, 2001. percent) reported by HMDA.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00297 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63876 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
Finally, HUD examined the GSEs’ own SF and MF rental properties. Specifically, the a 15.0-percent MF mix and a 8.5-percent data. Between 1999 and 2003, 58 percent of baseline underserved area share for SF rental investor mortgage share, the market share the SF rental units financed by GSE units is 52 percent and that for MF rental estimate is 36.9 percent if the overall (both purchases were located in underserved areas. units is 58 percent. home purchase and refinance) single-family- Between 1999 and 2002, 57 percent of the 2000-Based Underserved Area Market owner percentage for underserved areas is 31 multifamily units financed by GSE purchases Estimates. Table D.15 reports the results of were located in underserved areas. the projection model assuming 2000 percent, which is the estimated 1994–2003 Based on the above analyses, HUD retained geography. Since Table D.15 has the same HMDA average as well as the recent 1999– the assumptions from the 2004 GSE proposed interpretation as Table D.13, there is no need 2003 HMDA average.66 rule concerning underserved areas location of for a detailed explanation of it. Considering
66 In this case, the 2000-based underserved area underserved area percentages reported in Table owner loans reported in Table D.14. This procedure percentages for years prior to 1999 (i.e. 1994 to 1998 D.12. The 4.9 percent is the unweighted difference will be used throughout this section. in this example) are estimated by adding 4.9 of the 2000-based and 1990-based underserved area percent to the corresponding 1990-based shares for total (home purchase and refinance) SFO
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00298 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.098 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63877
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00299 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.099 63878 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
The above results are based on averages 16.0 percent MF mix. For example, reducing underserved area estimate is 37.9 percent across both home purchase and heavy the assumed multifamily mix from 16.0 under Case 2, as compared with 38.4 percent refinance environments. The analysis can percent to 13.5 percent reduces the overall under Case 1 (see Table D.15). As discussed also be conducted in terms of home purchase market projection for underserved areas by earlier, the baseline Case 1 assumptions offer environments, focusing on the underserved only 0.6–0.7 percentage points. This smaller a reasonable approach for estimating the area percentages for home purchase loans MF mix effect occurs because the underserved area market shares. reported in the first column of Table D.15. underserved area differentials between owner Examples of Home Purchase Years. The Again, considering a 15.0-percent MF mix and rental properties are not as large as the above projection results for a home purchase and a 8.5-percent investor mortgage share, low- and moderate-income and special environment can be compared with actual the underserved area market estimates affordable differentials reported earlier. For results for two home purchase years, 1999 reported in Table D.15 are: 37.8 percent if the example, the 1999–2003 average SF-owner and 2000 (see earlier description of these two SFO owner underserved area percentage is underserved areas share (30.3 percent in years in the low-mod section, F.3.a). For 32.3 percent (1999–2003 average home Table D.14a) is only 22 percentage points less 1999, the baseline model assumed a purchase share); 67 37.6 if the SF owner than the baseline SF-Rental underserved multifamily mix of 16.0 percent (see Section percentage is 31.8 percent (estimated average areas share (52.0); on the other hand, the C) and a mortgage investor share of 8.2 home purchase share from 1994–2003); 36.9 1999–2003 average SF-owner special percent (see Section D). Under these percent if the owner percentage is 31 percent affordable share (15.7 percent) is about 42 assumptions, the projected 1999 market (approximate home purchase share in 1999 percentage points less than the baseline SF- estimate (based on 2000-Census data) is 37.6 and 2001); 38.0 percent if the owner Rental special affordable share (58.0 percent). percent; lowering the MF mix to 15.0 (14.0) percentage is 32.5 percent (approximate As shown in Table D.15, ICF’s MF mix of percent instead of 16.0 percent reduces the home purchase percentage in 2000 and 14.25 percent produces results intermediate estimate only slightly to 37.3 (36.9) percent. 2002); and 39.0 percent if the owner between HUD’s 13.5 percent and 15.0 For 2000, the baseline model assumed a percentage is 33.7 percent (home purchase percent. Estimates of the underserved areas multifamily mix of 17.2 percent and a percentage in 2003). This analysis assumes based on a MF mix of 14.2 percent are only mortgage investor share of 9.1 percent. Under that the underserved areas share of refinance 0.2 percentage points less than those based these assumptions, the 2000 underserved loans is the same as those listed above for on a MF mix of 15.0 percent. areas market is estimated to be 39.7 percent. home purchase loans. But, as Table D.14a Investor Mortgage Share. Similarly, the A lower MF mix—for example, 16.0 (15.0) shows, in recent home purchase market estimates differ only slightly with percent instead of 17.2 percent—would environments, the underserved areas share of changes the investor mortgage share. reduce the estimated 2000 underserved areas refinance loans has been higher than that for Reducing the investor mix from 9.5 percent market share slightly to 39.4 (39.2) percent.68 home purchase loans, largely but not totally to 8.0 percent reduces the overall market For 1999, the 2000-based underserved area due to subprime refinance loans (see earlier projection for underserved areas by only 0.2– estimate (37.6 percent) is 4.8 percentage discussion). In the year 2000, for example, 0.4 percentage points. If the 10.0 percent points greater than the earlier-reported 1990 the overall underserved areas share for SFO baseline from the 2004 proposed GSE rule based estimate (32.8 percent); for the year owner loans reached 34.2 percent; in this were used in this analysis, the market 2000, the differential is 5.0 percentage points case, the market estimate is 39.4 percent in estimates would be approximately 0.3 (0.2) (39.7 versus 34.7). This approximately five this case. However, the next highest overall percentage points higher relative to the percentage point differential can be used to (both home purchase and refinance loans) results reported in Table D.15 for a baseline obtain estimates of 2000-based underserved owner share is the 31.9 percent share in of 8.5 (9.0) percent. Fannie Mae’s model area shares for the earlier home purchase 1999, which yields at an overall market combined a MF mix of 12.3 percent with an years, 1995 to 1997. Table D.9 of the estimate of approximately 37.5 percent. investor mortgage share of 8.0 percent. If the proposed GSE rule reported 1990-based Fannie Mae reports its estimates of the underserved area share of home purchase underserved area shares of 33.9 percent for 2000-Census-based underserved areas market loans is 32.3 percent (the average for 1999– 1995 and 1997 and 33.4 percent for 1996. in Table I.13 on page I–40. For SFO 2003), then the estimate for the overall These estimates, after adjustments for a lower percentages of 30 percent and 32 percent underserved areas market is 37.0 percent HMDA-based mortgage investor share and a (obtained by adding five percentage points to based on Fannie Mae’s assumptions. In lower-than-baseline MF mix, would remain Fannie Mae’s 1990-Census-based SFO contrast, HUD’s estimates (with a MF mix of in the 32–33 percent range. Adding five percentages of 25 percent and 27 percent, 15.0 percent and 8.5 percent investor share) percentage points would place these respectively), Fannie Mae projects are 37.8 percent—almost one percentage estimates in the 37–38 percent range in terms underserved area market shares of 35.1 point higher. If the underserved areas share of 2000 Census geography.69 ICF’s best percent and 36.8 percent, respectively. (It is of home purchase loans is at its 2003 level estimates were approximately 37 percent for interesting that these are the exact same (33.7 percent), then Fannie Mae’s 1994–1997 and 39 percent for 1999 (ICF market shares projected by HUD in Table assumptions result in a market estimate of Appendix, p. 77); its lower bound estimates D.15 for the ‘‘Fannie Mae assumptions’’ of 38.3 percent while HUD’s assumptions (see were approximately 34 percent during 1994– 12.2-percent MF mix and an 8.0-percent 1997 and 1999, and 37 percent in 2000 (ICF investor mortgage share—suggesting that previous sentence) result in a market estimate of 39.0 percent. In its projection Appendix, p.82). As noted earlier, ICF fills its Fannie Mae’s model produces the same report with numerous minimums that often results as HUD’s model when the input model, ICF assumed an underserved areas assumptions are the same.) Fannie Mae share of 31.5 percent for SF owner loans and 68 concluded that the higher 36.8 percent produced an estimate of almost 37 percent The baseline 39.7 percent estimate for 2000 is 0.7 percentage points lower than the 40.4 percent market share was not appropriate because the for the overall underserved areas market during 2005–2008 (ICF Appendix, p.133). share reported in Section G.4 of Appendix D of the SFO percentage of 32 percent was too high. proposed rule, mainly for the reasons discussed in However, as shown in Table D.14a, the 2000- Different Underserved Area Shares for Rental Properties. Case 2 (see Table D.9) the previous footnote. The difference is mostly based underserved area percentage for SFO explained (a) by the different treatment of single- home loans was greater than 32 percent in considered slightly smaller underserved area family rental mortgages and (b) by a 0.4 percentage 2000, 2002, and 2003. percentages for rental properties (50 percent point decline in HUD’s projections (in terms of the Multifamily Mix. As discussed earlier, for SF rentals and 56 percent for MF rentals), 2000 Census data) of the 2000 underserved areas compared with the low-mod and special as compared with the baseline Case 1, which percentage for SF owners. affordable market estimates, the underserved assumed 52 percent and 58 percent, 69 As explained earlier in Section G.2, HUD re- area market estimates exhibit less variation as respectively. Case 2 includes ICF’s estimated the underserved areas share for 1997 under the new assumptions (e.g., a lower, HMDA- one moves from a 13.5 percent MF mix to assumption (50 percent) for SF Rentals and is close to ICF’s assumption (55 percent) for based mortgage share for investor loans), obtaining MF Rentals. Incorporating these Case 2 a range of 32.0 percent (with a 16.3 MF mix) to 32.7 67 The market share estimates are interpolated percent (with a 19.3 percent MF mix). These from Table D.15. For example, the overall market assumptions reduces the underserved areas estimates assume 1990 Census geography. Adding estimate for a SFO percentage of 32.3 percent is market estimate by only 0.5 percentage five percentage points to reflect 2000 Census obtained by adding [.3*(38.4 minus 37.6)] to 37.6, points. For example, if the SFO home geography yields estimates of 37.0 percent to 37.7 to obtain the 37.6 figure reported in the text. purchase share is 33 percent, then the overall percent for the 1997 underserved areas market.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00300 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63879
appear unbelievable, such as the 32.8 percent share of market activity. The right-hand suggest that the underserved areas share will projection for the overall underserved market column of Table D.11 shows that the rental not likely fall below 35 percent, although, as in 2000 (ICF Appendix, p. 83), a time when share falls to the 17–22 percent range, or 4– noted above, the estimates for 2003 (around the SF owner underserved areas percentage 9 percentage points less than the almost 26- 35 percent) are somewhat speculative. was 35.7 percent itself (see Table 14a)—in percent rental share in HUD’s baseline Similar to 1999 and 2000, the 2001 and this case, the rental portion of the market was model. This contributes to the underserved 2002 differences between the 1990-based and below the underserved share for owners, areas share of the market typically falling to 2000-based underserved area market rather than the typical case where the rental 34–36 percent during a heavy refinancing estimates are about five percentage points. portion is more ‘‘goals rich’’ than the owner period. For 2001, the 2000-based baseline estimate portion. Model estimates were also made for the (36.9 percent) is 5.0 percentage points greater Market Volatility. Additional sensitivity recent refinancing years of 2001, 2002, and than the earlier-reported 1990 based estimate analyses were conducted to reflect the 2003. For 2001, the baseline model assumed of 31.9 percent); for the year 2002, the volatility of the economy and mortgage a multifamily mix of 13.5 percent and a differential is 4.9 percentage points (36.2 market. Recession and high interest rate mortgage investor share of 7.8 percent. Under versus 31.3).73 scenarios assumed a significant drop in the these assumptions, the 2001 market estimate The analysis in this section suggests that a underserved area percentage for single- is 36.9 percent.70 If the MF mix for 2001 had reasonable range for the overall market share family-owner mortgages. The single-family- been 12.5 (12.0) percent, then the estimated for underserved areas based on 2000 owner home purchase percentage can go as underserved areas market share for 2001 geography might be 35–39 percent, which is low as 29 percent—which is almost 2.8 would be 36.6 (36.4) percent. For 2002, the consistent with the 30–34 percent range percentage points lower than the 1994–2003 baseline model assumed a multifamily mix of estimated earlier based on 1990-based average of 31.8 percent, 3.3 percentage points slightly over 11.0 percent and a mortgage geography. lower than the 1999–2003 average of 32.3 investor share of 7.8 percent. Under these Feasibility of Underserved Areas Goal in a percent, and 4.7 percentage points lower than assumptions, the 2002 underserved areas Period of Heavy Refinancing. HUD received the underserved areas share of home market is estimated to be 36.2 percent.71 A a number of public comments seeking a purchase loans in 2003—and the estimated lower MF mix—for example, 10.5 (9.5) regulatory solution to the issue of the ability market share for underserved areas remains percent instead of 11 percent—would reduce of the GSEs to meet the housing goals during about 35 percent. In a more severe case, the the estimated 2002 underserved areas market a period when refinances of home mortgages overall underserved market share would be share to 36.0 (35.7) percent. ICF’s best constitute an unusually large share of the 33–34 percent if the single-family-owner estimates for 1998, 2001, and 2002 were in mortgage market. As explained in the home purchase share fell to 27 percent (its the 34–35 percent range while its lower- Preamble, HUD is not addressing the 1992 level), which is 5.3 percentage points bound estimates were in the 32–33 percent refinance issue in this final rule. Elsewhere 72 lower than its 1999–2002 average. range. in the Federal Register, HUD is publishing Table D.11 shows the impact on the As noted in Section F.3.b, HUD did not an Advance Notice of Proposed Rulemaking underserved areas market share under receive 2003 HMDA data until early August that advises the public of HUD’s intention to different assumptions about a refinancing 2004 and therefore HUD has not been able to consider by separate rulemaking a provision environment. See the earlier discussion of develop a complete projection model for that recognizes and takes into consideration 2003. Still, some rough projections for 2003 the low-mod goal in Section F.2b for an the impact of high volumes of refinance are provided here for different assumptions explanation of the various model transactions on the GSEs’ ability to achieve about the MF mix, recognizing that firm data assumptions necessary to simulate a heavy the housing goals in certain years, and on the 2003 multifamily market are not refinance environment. The discussion solicits proposals on how such a provision available. Combining an investor mortgage focuses on the 65-percent refinance rate since should be structured and implemented. HUD share of 8.2 from HMDA with different MF that has characterized recent refinance believes that it would benefit from further mixes produces the following estimates of waves. With respect to the underserved area consideration and additional public input on the underserved areas market for 2003: 35.1 characteristics of SF owner loans, two this issue. HUD also notes that FHEFSSA percent (MF mix of 8 percent); 34.7 percent scenarios were considered: (A) Scenario A provides a mechanism by which HUD can (MF mix of 7 percent); and 34.4 percent (MF represents the average underserved area take into consideration market and economic mix of 6.0 percent). conditions that may make the achievement of percentages for the last four refinance years As shown by both the simulation results in (1998, 2001, 2002, and 2003)—32 percent for housing goals infeasible in a given year. (See Table D.10 and by the actual experience 12 U.S.C. 1336(b)(e).) home purchase loans and 30 percent for during 2001–2003, the underserved area refinance loans; and (B) Scenario B B&C Loans. The procedure for dropping share declines when refinances dominate the B&C loans from the projections is the same represents the average underserved mortgage market. The above estimates percentages for the two most recent refinance as described in Section F.3.c for the Low- and Moderate-Income Goal. The underserved years (2002, and 2003)—33 percent for home 70 purchase loans and 29 percent for refinance The baseline 36.9 percent estimate for 2001 is areas percentage for B&C loans is 52.0 0.8 percentage point lower than the 37.7 percent loans. Thus, there is a 2–4 percentage point percent, which is larger than the projected share reported in Section G.4 of Appendix D of the percentages for the overall market given in differential between home purchase loans proposed rule. The difference is mostly explained and refinance loans in a heavy refinancing (a) by the different treatment in this Final Rule of Table D.15. Thus, dropping B&C loans (as environment. single-family rental mortgages and (b) by a 0.2 well as all subprime loans) will appreciably Under Scenario A, the underserved areas percentage point decline in HUD’s projections (in reduce the overall market estimates. Consider market shares varied by almost two terms of the 2000 Census data) of the 2001 the case of a single-family-owner percentage percentage points (i.e., 1.6 percent), from underserved areas percentage for SF owners. of 32 percent, which yields an overall market 71 36.0 percent with a 12 percent MF mix to The baseline 36.2 percent estimate for 2002 is estimate for the underserved areas of 38.6 34.4 percent with a 6 percent MF mix. These one percentage point lower than the 37.2 percent percent if B&C loans are included in the share reported in Section G.4 of Appendix D of the underserved area market shares are 3–5 analysis. Dropping B&C loans from the proposed rule. The difference is mostly explained projection model reduces the market share by percentage points lower than the (a) by the different treatment in this Final Rule of underserved areas shares reported in Table single-family rental mortgages and (b) by a 0.4 one percentage point to 37.6 percent, as D.15 for HUD’s baseline home purchase percentage point decline in HUD’s projections (in reported in Table D.15. Dropping all environment. (The results were similar for terms of the 2000 Census data) of the 2002 Scenario B.) Notice that under Scenario A, underserved areas percentage for SF owners. 73 The differentials reported in Table D.14 for the the underserved areas share remains in the 72 For the years 1999 to 2002, Fannie Mae three individual property types tend to be greater 34–35 percent range even if the MF mix falls estimated a 2000-Census-based underserved areas than five percentage points, which raises the to 6–8 percent. In addition to higher-income share of 37–38 percent, obtained by adding five question of why the overall differential is only five percentage points to Fannie Mae’s 32–33 percent percentage points. As explained later, the upward borrowers dominating the single-family estimate for the underserved areas market based on adjustment to account for underserved areas in non- market, the share of the ‘‘goals rich’’ rental 1990 Census data. (See their Table I.9, page I–34.) metropolitan areas is about 0.65 percentage point market declines in a refinancing wave, which This compares with HUD’s estimate of 37.1 percent less using the 2000-based Census data than it was tends to further reduce the underserved areas to 37.6 percent for the same period. using the 1990-based Census data.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00301 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63880 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
subprime loans (A-minus as well as B&C) manufactured housing effects can be market is 23–27 percent of the conventional would reduce the underserved areas market considered separately. Dropping only conforming market. projection to 37.4 percent. manufactured housing loans would reduce HUD is proposing to establish each GSE’s Non-metropolitan Areas. As explained in the market estimates by approximately three- special affordable multifamily subgoal as 1.0 Section G.3, in order to account for the much fourths of a percentage point. ICF argued that percent of its average annual dollar volume larger coverage of underserved areas in non- loans with less than $15,000 should be of total (single-family and multifamily) metropolitan areas, 1.25 percent was added excluded. The impact of doing this on the mortgage purchases over the 2000–2002 to the market share based on metropolitan market estimates would be about one-third of period. In dollar terms, the Department’s area data, in order to arrive at a nationwide a percentage point. ICF also considered proposal is $5.49 billion per year in special estimate of the market share for underserved scenarios where one-half of manufactured affordable multifamily purchases for Fannie areas. According to HMDA, underserved loans would be dropped, as well as small Mae, and $3.92 billion for Freddie Mac. The counties accounted for 42.7 percent of single- loans less than $15,000. The impact of doing multifamily special affordable goal, as well as family-owner mortgages originated in non- this on the market estimates would be three- the special affordable home purchase metropolitan areas during the 1999-to-2002 fifths of a percentage point. subgoal, are discussed further in Appendix C. period, based on 1990 geography. With 2000 The above analyses of the effects of less Section F described HUD’s methodology geography and the new tract-based definition affordable market conditions, different for estimating the size of the low- and of underserved areas in non-metropolitan assumptions about the size of the rental moderate-income market. Essentially the areas, the market share falls by 2.3 percentage same methodology is employed here except points to 39.6 percent. This 2000-based market, and dropping different categories of loans from the market definition suggest that that the focus is on the very-low-income underserved areas percentage of 39.6 percent market (0–60 percent of Area Median for non-metropolitan areas is about eight the 35–39 percent range described earlier is a reasonable range for the market estimate for Income) and that portion of the low-income percentage points less than the comparable market (60–80 percent of Area Median percentage for metropolitan areas.74 This underserved areas based on the projection model described earlier. This range Income) that is located in low-income census eight-point differential is lower than the 16- tracts. Data are not available to estimate the incorporates market affordability conditions point differential used in the earlier 1990- number of renters with incomes between 60 that are more adverse than have existed based Census analysis. Assuming that non- and 80 percent of Area Median Income who recently and it excludes B&C loans from the metropolitan areas account for 13 percent of live in projects that meet the tax credit market estimates. all single-family-owner mortgages and thresholds. Thus, this part of the Special estimating that the single-family-owner 5. Conclusions Affordable Housing Goal is not included in market accounts for 74.5 percent of newly- the market estimate. mortgaged dwelling units, then the non- Based on the above findings as well as metropolitan underserved area differential of numerous sensitivity analyses, HUD 1. Special Affordable Shares by Property 8 percent would raise the overall market concludes that 35–39 percent is a reasonable Type estimate by 0.78 percentage point—8 estimate of mortgage market originations that The basic approach involves estimating for percentage points times 0.13 (non- would qualify toward achievement of the each property type the share of dwelling metropolitan area mortgage market share) Geographically Targeted Goal if purchased by units financed by mortgages that are times 0.745 (single-family owner mortgage a GSE. The 35–39 percent range is higher occupied by very-low-income families or by market share). Based on this calculation, if than the market range in the 2000 Rule low-income families living in low-income the 8 point differential reflected actual mainly because it is based on 2000 Census areas. HUD combined mortgage information market conditions, then the underserved geography which includes more underserved from HMDA, the American Housing Survey, areas market share estimated using census tracts than 1990 Census geography. the Property Owners and Managers Survey metropolitan area data should be increased HUD recognizes that shifts in economic and and the recently released 2001 Residential by 0.78 percentage point to account for the housing market conditions could affect the Finance Survey in order to estimate these effects of underserved counties in non- size of this market; however, the market special affordable shares. estimate allows for the possibility that metropolitan areas, based on 2000 geography. a. Special Affordable Owner Percentages A more conservative adjustment of 0.65 adverse economic conditions can make percentage points was made in Table D.15, housing less affordable than it has been in HMDA data for the percentage of single- which reports the results of the projection the last few years. In addition, the market family-owners that qualify for the Special model. estimate incorporates a range of assumptions Affordable Goal are reported in Table D.16. Section G.3 reported that excluding about the size of the multifamily market and That table also reports data for the two manufactured housing loans (as well as small excludes B&C loans. components of the Special Affordable Goal— very-low-income borrowers and low-income loans less than $15,000) reduced the overall H. Size of the Conventional Conforming underserved area market estimates based on borrowers living in low-income census tracts. Market for the Special Affordable Housing Focusing first on home purchase loans, 1990 geography by less than one percentage Goal point (roughly three-quarters of a percentage HMDA data show that the special affordable point). Excluding manufactured housing This section presents estimates of the share of the market has followed a pattern loans leads to a similar reduction for the conventional conforming mortgage market for similar to that discussed earlier for the low- market estimates based on 2000 geography. the Special Affordable Housing Goal. The and moderate-income loans. The percentage As reported earlier, the small loan and special affordable market consists of owner of special affordable borrowers increased and rental dwelling units which are occupied significantly between 1992 and 1994, from 10.4 percent of the conforming market in 74 Between 1999 and 2002, 2000-based by, or affordable to: (a) Very-low-income underserved census tracts accounted for 31.4 families; or (b) low-income families in low- 1992 to 12.6 percent in 1993, and then to percent (unweighted annual average) of all income census tracts; or (c) low-income 14.1 percent in 1994. Between 1995 and mortgages in metropolitan areas. This 1999–02 families in multifamily projects that meet 1998, the special affordable market was in average percentage for single-family owners in minimum income thresholds patterned on the 14–16 percent range, averaging 15.1 metropolitan area is lower than the underserved the low-income housing tax credit (LIHTC).75 percent. Over the past five years (1999–2003), area percentage reported in previous paragraphs. To HUD estimates that the special affordable the special affordable share of the home be comparable with the non-metropolitan data, purchase loans has averaged 16.4 percent. It these metropolitan area data do not include loans was about 17 percent during 1999 and 2000 originated by lenders that specialize in 75 There are two LIHTC thresholds: at least 20 manufactured housing loans and B&C loans, percent of the units are affordable at 50 percent of and 16 percent during the most recent three excluding these loans lowers the underserved areas AMI or at least 40 percent of the units are affordable years, 2001 to 2003. share. at 60 percent of AMI. BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00302 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63881
Considering all (home purchase and affordable market has changed in higher special affordable percentages for refinance) loans during recent ‘‘home fundamental ways from the mortgage market refinance loans are reduced or even purchase’’ environments, the special of the early 1990s. eliminated if subprime loans are excluded affordable share averaged 18.7 percent during Except for the four years of heavy from the analysis. As shown in Table D.16, 1999–2000, over three percentage points refinancing (1998, 2001, 2002, and 2003), the excluding B&C loans from the data more than the 15.4 percent average between special affordable share of the refinance practically eliminates the refinance-home- 1995 and 1997. Excluding B&C (all subprime) market has recently been higher than the purchase differential for 1999 and reduces loans from the analysis reduces this special affordable share of the home purchase the differential for 2000 to 4.2 percentage differential only slightly to 2.8 (2.4) market—a pattern discussed in Section F for points (from 5.7 percentage points). Going percentage points. As mentioned earlier, low-mod and very-low-income loans. During further and excluding A-minus loans from lending patterns could change with sharp 1999 (2000), for example, the special the year 2000 data would reduce the changes in the economy, but the fact that affordable share of the refinance market was differential to 2.7 percentage points. HUD’s there have been several years of strong 19.2 (22.6) percent, compared with 17.3 projection model excludes B&C loans and affordable lending suggests that the special (16.9) percent for the home loan market. The sensitivity analyses will show the effects on
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00303 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.100 63882 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
the overall special affordable market of very-low-income families. According to the The baseline analysis in HUD’s proposed excluding all single-family subprime loans. AHS, 59 percent of single-family units and 53 GSE rule assumed that 8 percent of the New 2000-Based Census Geography and percent of multifamily units were affordable single-family rental units and 11.0 percent of New OMB Metropolitan Area Definitions. to very-low-income families in 1997. The multifamily units are affordable at 60–80 Going forward, HUD will be re-benchmarking corresponding average values for the AHS’s percent of AMI and located in low-income its median incomes for metropolitan areas six surveys between 1985 and 1997 were 58 areas. and non-metropolitan counties based on percent and 47 percent, respectively. As Combining the assumed very-low-income 2000 Census incomes, will be defining low- discussed earlier in Section F, an important percentage of 50 percent (47 percent) for income census tracts (which are included in issue concerns whether rent data based on single-family rental (multifamily) units with the definition of special affordable) in terms the existing rental stock from the AHS can be the assumed low-income-in-low-income-area of the 2000 Census geography, and will be used to proxy rents of newly mortgaged percentage of 8 percent (11 percent) for incorporating the effects of the new OMB rental units. HUD’s analysis of POMS data single-family rental (multifamily) units yields metropolitan area definitions. As discussed during the 2000 rule-making process the special affordable percentage of 58 earlier in Section F, HUD projected the suggested that it could—estimates from percent (58 percent) for single-family rental effects of these three changes on the special POMS of the rent affordability of newly- (multifamily) units. This was the baseline affordable shares of the market for the years mortgaged rental properties are quite case in the 2004 proposed GSE rule. 1999–2003; the results for special affordable consistent with the AHS data on the d. Comments on the Special Affordable loans are reported in the top portion of Table affordability of the rental stock. Fifty-six (56) Rental Share and Additional Analysis D.8b. Under the historical MSA-based data, percent of single-family rental properties Both ICF and Fannie Mae commented that the (unweighted) average special affordable with new mortgages between 1993 and 1995 HUD overstated the special affordable share share of the conventional conforming market were affordable to very-low-income families, of the single-family rental and multifamily was 16.4 (16.3) percent for home purchase as were 51 percent of newly-mortgaged rental markets. They argued that updated (total) loans (see Table D.16); the multifamily properties. These percentages for 2001 AHS data showed that the affordability corresponding average with the CBSA-based newly-mortgaged properties from the POMS of the rental housing stock had declined projected data was 16.4 (16.4) percent, or are similar to those reported above from the since HUD had conducted its POMS and practically the same. Given these small AHS for the rental stock. Based on this POMS AHS analyses in 1995 and 1997, respectively. differences there is no need to adjust the analysis, HUD’s baseline model in the 2004 For both single-family (SF) and multifamily overall market estimates reported below to proposed GSE rule assumed that 50 percent (MF) rentals, ICF used a special affordable account for the new data. However, it should of newly-mortgaged, single-family rental range of 47–53 percent, with a baseline of 50 be noted that the most recent year of 2003 units, and 47 percent of multifamily units, percent. ICF’s special affordable range is does show a rather larger difference—the were affordable to very-low-income families. much less than both HUD’s 53–61 percent special affordable share of home purchase (See further discussion of this issue in range (58 percent baseline) for single-family loans under the projected CBSA approach is Section H.1.d) rentals and HUD’s 54–62 percent range for 16.9 percent, which is a full percentage point multifamily rentals (also a 58 percent higher than the special affordable share of c. Low-Income Renters in Low-Income Areas baseline). Since SF and MF rentals account 15.9 percent under historical data.76 HMDA does not provide data on low- for about 25 percent of financed units in For the other two property types (single- income renters living in low-income census HUD’s model, reducing the SF and MF family rental and multifamily), comparisons tracts. As a substitute, HUD used the POMS baselines from 58 percent (HUD’s baseline) to between projected and historical special and AHS data. As explained in the 2000 GSE 50 percent (ICF’s baseline) would reduce the affordable percentages were made using the Rule, the share of single-family and overall special affordable market estimate by GSEs’ data. For single-family rental multifamily rental units affordable to low- two percentage points. Thus, this is an mortgages, the weighted average of Fannie income renters at 60–80 percent of area important issue. Mae’s (Freddie Mac’s) special affordable median income (AMI) and located in low- Based on its analysis of the AHS (see percentage for the years 1999 to 2003 was income tracts was calculated using the Fannie Mae Appendix, I–31–I–32), Fannie 48.2 (48.7) percent using the historical data, internal Census Bureau AHS and POMS data Mae concluded that the very-low-income compared with 49.6 (49.5) percent using the 77 files. The POMS data showed that 8.3 share for single-family rental properties had projected data. For multifamily mortgages, percent of the 1995 single-family rental stock, fallen from 58.3 percent in 1997 to 53.0 the weighted average of Fannie Mae’s and 9.3 percent of single-family rental units percent in 2001; similarly, the very-low- (Freddie Mac’s) special affordable percentage receiving financing between 1993 and 1995, income (VLI) share of multifamily rental for the years 1999 to 2003 was 50.9 (48.7) were affordable at the 60–80 percent level properties had fallen from 52.0 percent to percent using historical data, compared with and were located in low-income census 44.9 percent over this same period. (By 51.6 (51.5) percent using the projected data. tracts. The POMS data also showed that 12.4 comparison, ICF estimated that 47 percent of These comparisons suggest little difference percent of the 1995 multifamily stock, and the SF rental stock and 42 percent of the MF between the historical and projected special 13.5 percent of the multifamily units rental stock were affordable to VLI families.) affordable shares for rental properties. HUD receiving financing between 1993 and 1995, In its analysis, Fannie Mae provides a weight also projected the overall special affordable were affordable at the 60–80 percent level of 0.07 to the VLI share (25.7 percent) of percentage for each GSE. For the overall 78 and located in low-income census tracts. recently-constructed single-family rental special affordable goal (considering all three units in the AHS, and the residual 0.93 property types), the unweighted average of 77 Affordability was calculated as discussed weight to the VLI share (53.6 percent) of the Fannie Mae’s (Freddie Mac’s) special earlier in Section F, using AHS monthly housing remaining existing units in the AHS. While affordable percentage for the years 1999 to cost, monthly rent, number of bedrooms, and MSA Fannie Mae uses a VLI share of 46 percent 2002 was 20.0 (18.9) percent using the location fields. Low-income tracts were identified for single-family rentals in its market sizing projected data, compared with 20.0 (18.9) using the income characteristics of census tracts models, applying these weights to the 2001 percent using the historical data. There is from the 1990 Census of Population, and the census AHS data (reported by Fannie Mae in Table little difference in the GSEs’ average special tract field on the AHS file was used to assign units I.7 on p. I–32) yields approximately 52 affordable performance between the in the AHS survey to low-income tracts and other percent for the VLI share of single-family projected and historical data. tracts. POMS data on year of mortgage origination were utilized to restrict the sample to properties rental properties. Similarly, for multifamily b. Very-Low-Income Rental Percentages mortgaged during 1993–1995. properties, Fannie Mae provides a weight of Table D.14 in Appendix D of the 2000 Rule 78 During the 1995 rule-making process, HUD 0.11 to the VLI share (22.2 percent) of reported the percentages of the single-family examined the rental housing stock located in low- recently-constructed multifamily rental units income zones of 41 metropolitan areas surveyed as rental and multifamily stock affordable to in the AHS, and the residual 0.89 weight to part of the AHS between 1989 and 1993. While the the VLI share (45.7 percent) of the remaining low-income zones did not exactly coincide with 76 As noted earlier, this discrepancy could be due low-income tracts, they were the only proxy readily existing units in the AHS. In this case, to mis-measurement from the technique for available to HUD at that time. Slightly over 13 apportioning 2003 data, which is defined in 2000- percent of single-family rental units were both located in low-income zones; almost 16 percent of census geography, to a 1990-based geography. affordable at the 60–80 percent of AMI level and multifamily units fell into this category.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00304 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63883
applying the above weights to the 2001 AHS analysis of POMS data and its analysis of 2. Size of the Special Affordable Market data yields 43 percent for the VLI share of 2003 AHS geocoded data suggest that low- The size of the special affordable market multifamily rental properties—a figure income SF renters in low-income areas depends in large part on the size of the similar to the 41-percent VLI share that account for 22 percent of all SF low-income single-family rental and multifamily markets Fannie Mae uses in its market sizing models. renters; GSE data for 2001 and 2002 suggest and on the special affordable percentages of After computing a VLI share of 46 percent for 79 a slightly higher percentage. With respect both owners and renters. Therefore, this SF rentals, Fannie Mae adds 8 percent to to MF properties, HUD’s analysis of POMS section conducts several sensitivity analyses account for low-income renters living in low- data and its analysis of 2003 AHS geocoded around these market parameters. As in the income census tracts (the second component data suggest that low-income MF renters in previous sections, this section initially of the special affordable category); this yields low-income areas accounted for 24–25 assumes a refinance rate of 35 percent, which 54 percent for the special affordable share of percent of all MF low-income renters; GSE means that it initially focuses on a home SF rentals. After computing a VLI share of 41 data for 2001 and 2002 suggest a slightly percent for MF rentals, Fannie Mae adds 11 purchase or low-refinancing environments. lower percentage (21 percent). These shares After presenting these results, market percent to account for low-income renters can be applied to the 2003 AHS results for living in low-income census tracts; this estimates reflecting a heavy refinance low-income renters. For SF rentals, the 22 yields 52 percent for the special affordable environment will be presented. In the 2000 percent share for low-income renters living share of MF rentals. Thus, Fannie Mae’s GSE Rule, HUD assumed that the special in low-income census tracts can be estimates are intermediate between ICF’s (50 affordable share of refinance loans was 1.4 percent) and HUD’s (58 percent). Since SF multiplied by the 20 percent figure that the percentage points lower than the special rentals account for 10.6 percent of financed 2003 AHS produces for low-income SF affordable share of borrowers purchasing a units in HUD’s model, reducing the SF renters, yielding estimate of 4.4 percent. This home. However, as discussed earlier, the baseline from 58 percent (HUD’s baseline) to 4.4 percent is added to the VLI percentage of special affordable share of refinance loans 54 percent (Fannie Mae’s baseline) would 67 percent for SF rentals to arrive at a special equaled or was greater than the special reduce the overall special affordable market affordable estimate of 71 percent, based on affordable share of home purchase loans estimate by 0.42 percentage points. Since MF the 2003 AHS. For MF rentals, the 25 percent during home purchase environments such as rentals account for 15.0 percent of financed share for low-income renters living in low- 1995–97 or 1999–2000; thus, the assumption units in HUD’s model, reducing the MF income census tracts can be multiplied by of a lower special affordable share for baseline from 58 percent (HUD’s baseline) to the 27 percent figure that the 2003 AHS refinance loans is initially dropped from the 52 percent (Fannie Mae’s baseline) would produces for low-income MF renters, analysis but will be reintroduced during the reduce the overall special affordable market yielding an estimate of 6.7 percent.80 This 6.7 sensitivity analysis and the discussion of estimate by 0.90 percentage points. percent is added to the VLI percentage of 57 heavy refinancing environments. If the Combining these two reductions yields a 1.32 percent for MF rentals to arrive at a special special affordable share of refinance loans percentage point reduction in the overall affordable estimate of 63 percent, based on were assumed to be one percentage point less special affordable market. the 2003 AHS. These 2003 AHS special than that of home purchase loans, then the HUD is retaining its baseline of 58 percent affordable shares—67 percent for SF rental market shares in Table D.17 would be for the special affordable share of both SF units and 63 percent for MF rental units— approximately one-quarter percentage point and MF rentals. Several sets of analyses led support HUD’s use of a 58-percent baseline lower.81 to this decision. as the special affordable share of both SF and Considering a 15.0-percent MF mix and a HUD updated its analysis with 2001 and MF rental properties. 8.5-percent investor mortgage share, the 2003 AHS data. Using ICF’s assumptions for It is interesting to compare HUD’s 58- special affordable market estimates reported an AHS analysis (see ICF Appendix, p. 45), percent baseline with the actual performance in Table D.17 are: 27.3 percent if the owner the 2003 AHS data showed that 57 percent of Fannie Mae and Freddie Mac. For single- percentage is 17 percent (home purchase (67 percent) of single-family (multifamily) family rental mortgages, the weighted average share for 1999 and 2000); 26.8 if the owner rental units would qualify as being affordable of both Fannie Mae’s and Freddie Mac’s percentage is 16.4 percent (average home to VLI families. This analysis of the 2003 special affordable percentage for the years purchase share from 1999–2003); 26.5 percent if the owner percentage is 16 percent AHS used a new geocoded file that identified 1999 to 2003 was about 50 percent using (home purchase share for 1998, 2001, 2002, the specific metropolitan area or county projected CBSA data. For multifamily and 2003); and 25.7 percent if the owner location for each observation in the AHS. mortgages, the weighted average of Fannie percentage is 15 percent (home purchase This allowed HUD to link accurate area Mae’s special affordable percentage for the average from 1995–97). Considering a range median incomes (used to determine same years was 49 percent, while Freddie of 13.5–16.0 for the MF mix and a range of affordability) to each AHS observation, Mac’s percentage was 52 percent. As ICF 8.5–9.0 for the investor mortgage share, the which represents a substantial improvement notes, the GSEs’ below market performance special affordable market estimates reported over previous AHS analyses that did not have may be due to their limited participation in in Table D.17 are: 26.7–27.9 percent if the the specific household location and thus had the small multifamily market (ICF Appendix, owner percentage is 17 percent; 26.2–27.4 to rely on estimates of area median income p. 47). in order to compute affordability ratios. This percent if the owner percentage is 16.4 more accurate approach appears to produce percent; 25.9–27.1 percent if the owner 79 higher affordability estimates than earlier Fannie Mae’s data exhibited some variation, percentage is 16 percent; and 25.1–26.3 analyses based on the non-geocoded AHS. standing at 33 percent in 2001 and 19 percent in percent if the owner percentage is 15 percent. 2001. Freddie Mac’s percentage was 29 percent in BILLING CODE 4210–27–P To derive an overall special affordable both years. percentage, one must add the second 80 These adjustments for low-income renters component of the special affordable living in low-income areas may be conservative. For 81 This is obtained by multiplying (a) 1.0 category—low-income renters living in low- SF (MF) rentals, the 2001 and 2002 figures for the percentage point by (b) the refinance rate of 0.35 by income areas—to the VLI share. HUD’s GSEs were in the nine (eight) percent range. (c) the 0.745 property share for SF owner loans.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00305 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63884 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
BILLING CODE 4210–27–C
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00306 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 ER02NO04.101 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63885
If the special affordable percentage for market based on a MF mix of 14.2 percent estimates reported in Table D.9 of Appendix home purchase loans fell to 13 percent—or are only 0.3 percentage points less than those D of the proposed rule for the three home by three percentage points below its 1995– based on a MF mix of 15.0 percent. Fannie purchase environments prior to 1999—28.9 2003 average level of approximately 16 Mae’s model combined an even lower MF percent for 1995, 28.7 for 1996, and 28.8 percent—then the overall market estimate mix of 12.3 percent with an investor percent for 1997.82 Given (a)–(c) and the fact would be about 24 percent under the baseline mortgage share of 8.0 percent. If the special that the HMDA-reported mortgage investor assumptions. Thus, 24 percent is consistent affordable share of home purchase loans is share was approximately eight percent with a rather significant decline in the 16.4 percent (the 1999–2003 average), then during these three years (instead of the special affordable share of the single-family the estimate for the overall special affordable assumed 10 percent in the earlier 1995–97 home purchase market. A 24 percent market market is 25.2 percent based on Fannie Mae’s analysis), these estimates should probably be estimate allows for the possibility that assumptions. In contrast, HUD’s estimates reduced by the above-mentioned 1.3 adverse economic and housing affordability (with a MF mix of 15.0 percent and 8.5–9.0 percentage points, which would place them conditions could keep special affordable percent investor share) are 26.8–27.0 percent at 27–28 percent assuming no adjustment in families out of the housing market. On the ‘‘about one and a half percentage points the baseline MF mix, and at 26–27 percent other hand, if the special affordable home higher. If the special affordable share of home assuming a MF mix three percentage points purchase percentage stays at its recent levels purchase loans is 16 percent (its recent 2001– lower than the baseline MF mix.83 (15–17 percent), the market estimate is in the 2003 level), then Fannie Mae’s assumptions For 2000, the baseline model assumed a 26–27 percent range. result in a market estimate of 25.2 percent multifamily mix of 17.2 percent and a Different Special Affordable Shares for while HUD’s assumptions (see previous mortgage investor share of 9.1 percent. Under Rental Properties. Case 2 (see Table D.9) sentence) result in market estimates of 26.5– these assumptions, the 2000 special considered smaller special affordable 26.7 percent. affordable market is estimated to be 29.1 percentages for rental properties (53 percent Investor Mortgage Share. As shown in percent. A lower MF mix—for example, 15.0 for SF rentals and 54 percent for MF rentals), Table D.17, increasing the investor mortgage percent instead of 17.2 percent—would as compared with the baseline Case 1, which share by one percentage point from 8.0 reduce the estimated 2000 low-mod market 84 assumed 58 percent for both property types. percent to 9.0 percent increases the special share to 28.2 percent. Case 2 assumptions are close to Fannie Mae’s affordable market estimate by approximately ICF’s best estimates for the special assumptions—54 percent for SF Rentals and 0.4–0.5 percentage point. If the 10.0 percent affordable market were 25–26 percent in 52 for MF Rentals. Incorporating the Case 2 baseline from the 2004 proposed GSE rule 1995, 1997, 1999, and 2000, and a assumptions reduces the special affordable were used in this analysis, the market particularly low 23 percent for 1996 (ICF market estimate by 1.2 percentage points. For estimates would be approximately 0.6 (0.4) Appendix, p. 94). Its lower bound estimates were 22–23 percent for 1997 and 1999, 24 example, if the SFO home purchase share is percentage points higher relative to the percent for 1995 and 2000, and 21 percent for 17 percent, then the overall special affordable results reported in Table D.15 for a baseline 1996 (ICF Appendix, p. 99). As discussed estimate is 26.1 percent under Case 2, as of 8.5 (9.0) percent. earlier, two percentage points of the HUD– compared with 27.3 percent under Case 1 Examples of Home Purchase Years. The ICF differential involves ICF’s lower (see Table D.17). above projection results for a home purchase assumptions about the special affordable ICF’s assumptions were even lower, 50 environment can be compared with actual characteristics of rental loans. Given that the percent for both SF and MF rentals, a figure results for two home purchase years, 1999 SFO percentage was 18–19 percent during that is eight percentage points lower than and 2000, which were characterized by 1999 and 2000 (see Table D.16), ICF’s 23–24 HUD’s baseline Case 1 assumption of 58 refinance rates of 34 percent and 29 percent, estimates for 1999 and 2000 are in need of percent for each of these two property types. respectively. For 1999, the baseline model further explanation. Given that these two property types account assumed a multifamily mix of 16.0 percent Heavy Refinancing Environments. The for 25 percent of all financed dwelling units, and a mortgage investor share of 8.2 percent. special affordable share of the overall market using ICF’s 50-percent assumption (instead of Under these assumptions, the 1999 market declines when refinances dominate the HUD’s 58-percent assumption) would reduce estimate is 27.9 percent; if the 1999 MF mix market. Section F.3c, which presents the low- the overall special affordable market shares was lower—for example, 15.0 (14.0) percent mod market estimates, explained the in Table D.17 by two percentage points. As instead of 16.0 percent—then the estimate of assumptions for incorporating a refinance discussed above, HUD’s baseline Case 1 the 1999 special affordable market share environment into the basic projection model assumptions offer a reasonable approach for would be 27.5 (27.2) percent. for 2005–08. Briefly, they are: the refinance estimating the special affordable market The 2004 proposed rule (Table D.9 in share of single-family mortgages was shares. Appendix D) reported a higher baseline increased to 65 percent (from 35 percent); the Multifamily Mix. The volume of market estimate for 1999 of 29.2 percent, as multifamily mix was allowed to vary from 6 multifamily activity is also an important compared with the 27.9 percent reported in to 12 percent; the market share for subprime determinant of the size of the special market. the previous paragraph—a differential of 1.3 While Section C explained the rationale for percentage points. The difference is largely 82 These three estimates were initially reported in HUD’s 15.0 percent range, it is useful, given due to the treatment of single-family rental HUD’s 2000 Final Rule, and repeated in Table D.9 the uncertainty surrounding the size of the mortgages. For example, using the proposed of Appendix D of the 2004 proposed GSE rule. multifamily market, to consider the effects of rule’s 10-percent assumption for the 83 To provide some confirmation for these 1995– lower multifamily mix assumptions, even in mortgage investor share (instead of the lower 1997 estimates, HUD went back and re-estimated a home purchase environment. Assuming a 8.2 percent HMDA-based mortgage investor the model for 1997. As shown in Table D.9 of the 13.5 percent MF mix reduces the overall shares reported in the text) would increase 2004 GSE Proposed Rule (as well as in Table D.15 special affordable market estimates by 0.4 the 1999 estimate by 0.8 percentage points to of the 2000 GSE Rule), HUD had earlier estimated percentage points compared with a 15 28.7 percent, only 0.5 percentage points a special affordable share of 28.8 percent for 1997 percent MF mix, and by 1.0 percentage point lower than the 29.2 percent reported in the (which was practically the same as the 28.9-percent share estimated for 1995 and the 28.7-percent share compared with a 16.0 percent mix. For proposed rule. Other more minor changes estimated for 1996). With a lower investor share example, when the special affordable share of that lower market estimate included: (a) (8.4 percent instead of 10.0 percent) and other the home purchase market is at 16.4 percent Further reducing the SF mortgage investor changes mentioned in the text, the new estimate for (its 1999–2003 average), the special share by excluding B&C investor loans from the 1997 special affordable market was 28.0 affordable share of the overall market is 26.2 the HMDA data (see Section C); (b) using 1.6 assuming a multifamily mix of 19.3 percent. If the percent assuming a 13.5 percent multifamily percent (instead of 2.0 percent) for the multifamily mix is reduced to 17.3 (16.3) percent, mix, compared with 26.8 (27.4) percent mortgage share of single-family 2–4 property the special affordable share of the 1997 market is assuming a 15 (16.0) percent multifamily owners; and (c) using slightly lower 27.1 (26.7) percent. The 26.7–28.0 percent range for 1997 is consistent with the 1995–1997 ranges mix. dwelling-units-per-mortgage assumptions for reported in the text. As shown in Table D.17, the ICF’s MF mix SF 2–4 properties (2.20 instead of 2.25) and 84 Using the projected CBSA data (instead of the of 14.2 percent produces results intermediate for SF investor mortgages (1.30 instead of historical 1990-based MSA data) did not change the between HUD’s 13.5 percent and 15.0 1.35). These changes, leading to this 1.3 special affordable market estimate in either 1999 or percent. Estimates of the special affordable percentage point differential, also affect the 2000.
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00307 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 63886 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations
loans was reduced to 8.5 percent (from 12 lower MF mix—for example, 10.5 (9.5) implemented. HUD believes that it would percent); and the mortgage investor share was percent instead of 11 percent—would reduce benefit from further consideration and set at 8.0 percent (its average during recent the estimated 2002 special affordable market additional public input on this issue. HUD refinancing waves). With respect to MF share to 24.2 (23.7) percent. 86 87 also notes (see above) that FHEFSSA mixes, it is likely that an 11–12 percent MF As explained in Section F.3b, HUD has not provides a mechanism by which HUD can mix characterized 2001, 9–11 percent yet completed its analysis of 2003 data. take into consideration market and economic characterized 2002, and less than 7 percent However, HUD developed some rough conditions that may make the achievement of characterized 2003, although there is some projections for different assumptions about housing goals infeasible in a given year. (See uncertainty with these estimates. In a the MF mix. Combining an investor mortgage 12 U.S.C. 1336(b)(e).) refinancing wave, the special affordable share of 8.2 from HMDA with different MF B&C Loans. The procedure for dropping percent is typically lower for refinance loans mixes (ranging from 6 percent to 8 percent) B&C loans from the projections is the same than home purchase loans, as middle- and produced estimates of 22.6 percent (MF mix as described in Section F.3.c for the Low- and high-income borrowers dominate the market. of 6 percent) to 23.5 percent (MF of 8 Moderate-Income Goal. The special With respect to the special affordable percent). affordable percentage for B&C loans is 28.0 characteristics of SF owner loans, the As shown by both the simulation results in percent, which is similar to the projected refinancing analysis assumed 16 percent for Table D.17 and the actual experience during percentages for the overall market given in home purchase loans and 14 percent for 2001–2003, the special affordable share of the Table D.17. Thus, dropping B&C loans (as refinance loans, which were the average overall market declines when refinances well as all subprime loans) does not special affordable percentage for the last four dominate the market. The special affordable appreciably reduce the overall market refinance years (1998, 2001, 2002, and 2003). share was approximately 24 percent during estimates. Consider the case of a single- There has been a two percentage point 2001 and 2002 and 23 percent in 2003 family-owner percentage of 16 percent, differential between home purchase loans (although there is some uncertainty with the which yields an overall market estimate for and refinance loans during a heavy 2003 estimate). Special Affordable Goal of 26.7 percent if refinancing environment. The various market estimates presented in B&C loans are included in the analysis. As shown in Table D.11, the special Table D.17 for a home purchase environment Dropping B&C loans from the projection affordable shares varied by over two and reported above for a refinance model reduces the special affordable market percentage points, from 24.1 percent with a environment are not all equally likely. Most share by 0.2 percentage points to 26.5, as 12 percent MF mix to 21.7 percent with a 6 of them equal or exceed 23 percent. In the reported in Table D.17. Dropping all percent MF mix. These special affordable home purchase environment, estimates subprime loans (A-minus as well as B&C) market shares are 3–5 percentage points below 23 percent would require the special would reduce the special affordable market lower than the special affordable shares affordable share for home purchase loans to projection to 26.2 percent. reported in Table D.17 for HUD’s baseline drop to 12 percent which would be 4 Manufactured Housing Loans and Small home purchase environment. Notice that the percentage points lower than the 1995–2003 Loans. Excluding manufactured housing special affordable share remains in the 22–23 average for the special affordable share of the loans and small loans (loans less than percent range even if the MF mix falls to 6– home purchase market. As shown in Table $15,000) reduces the overall market estimates 8 percent. In addition to higher-income D.11, dropping below 23 percent would be reported in Table D.17 by less than one borrowers dominating the single-family more likely in a heavy refinance percentage point. This is estimated as market, the share of the ‘‘goals rich’’ rental environment, particularly those characterized follows. First, excluding these loans reduces market declines in a refinancing wave, which by extremely low MF mixes of 7 percent or the special affordable percentage for single- tends to further reduce the special affordable less. family-owner mortgages in metropolitan of market activity. The right-hand column of As stated in Sections F and G above, HUD areas by about 1.5 percentage points, based Table D.11 shows that the rental share falls received a number of public comments on analysis of recent home purchase to the 17–22 percent range, or 4–9 percentage seeking a regulatory solution to the issue of environments (1995–97 and 1999 and 2000). points less that the almost 26-percent rental the ability of the GSEs to meet the housing Multiplying this 1.5 percentage point share in HUD’s baseline model. goals during a period when refinances of differential by the property share (0.745) of Model estimates were also made for the home mortgages constitute an unusually single-family-owner units yields 1.1 recent refinancing years of 1998, 2001, 2002, large share of the mortgage market. As percentage points, which serves as a proxy and 2003. For 1998, the baseline model explained in the Preamble, HUD is not for the reduction in the overall special assumed a multifamily mix of 14.0 percent addressing the refinance issue in this final affordable market share due to dropping and a mortgage investor share of 6.8 percent. rule. Elsewhere in the Federal Register, HUD manufactured home loans from the market Under these assumptions, the 1998 market is publishing an Advance Notice of Proposed analysis. The actual reduction will be estimate is 24.0 percent. If the MF mix for Rulemaking that advises the public of HUD’s somewhat less because dropping 1998 had been 13.0 (12.0) percent then the intention to consider by separate rulemaking manufactured home loans will increase the a provision that recognizes and takes into estimated special affordable market share for share of rental units, which increases the consideration the impact of high volumes of 1998 would be 23.5 (23.1) percent. For 2001, overall special affordable share, thus partially refinance transactions on the GSEs’ ability to the baseline model assumed a multifamily offsetting the 1.1 percent reduction. The net achieve the housing goals in certain years, mix of 13.5 percent and a mortgage investor effect is probably a reduction of three- and solicits proposals on how such a share of 7.8 percent. Under these quarters to one percentage point. provision should be structured and assumptions, the 2001 market estimate for The effects can be considered separately. special affordable loans is 25.0 percent. If the Dropping only manufactured housing loans MF mix for 2001 had been 12.0 percent, mortgages. In addition, the SF0 percentage for home would reduce the market estimates by instead of the baseline of 13.5 percent, then purchase loans originated during 2002 was lowered approximately one-half of a percentage point. the estimated special affordable market share by approximately 0.2 percentage point in the Final ICF argued that loans with less than $15,000 for 2001 would be 24.4 percent. For 2002, the Rule. should be excluded. The impact of doing this baseline model assumed a multifamily mix of 86 Using the projected CBSA data (instead of the on the market estimates would be about one- slightly over 11.0 percent and a mortgage historical 1990-based MSA data) resulted in only third to four-fifths of a percentage point. ICF investor share of 7.8 percent. Under these small changes in the special affordable market estimates for 2001 (a 0.1 percentage point decline) also considered scenarios where one-half of assumptions, the 2002 special affordable and 2002 (a 0.5 percentage point decline). manufactured loans would be dropped, as market is estimated to be 24.3 percent.85 A 87 For the years 1999 to 2002, Fannie Mae well as small loans less than $15,000. The estimated a special affordable market share of 23– impact of doing this on the market estimates 85 The baseline estimates for 2001 (25.0 percent) 25 percent. (This is their estimate assuming no would be three-fifths to three-quarters of a and 2002 (24.3 percent) are lower than those (26.5 missing data; see their Table I.9, page I–34.) This percentage point. percent and 25.8 percent, respectively) reported in compares with HUD’s estimate of 25.9 percent to Table D.9 of Appendix D of the proposed rule. As 26.6 percent. As discussed in Section C.6, Fannie The above analyses of the effects of less explained earlier, the differences between the Mae assumes a rather low MF mix (approximately affordable market conditions, different results in the proposed rule and this Final Rule are 10 percent) in the model that generates its historical assumptions about the size of the rental mainly due to the treatment of single-family rental estimates. market, and dropping different categories of
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00308 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2 Federal Register / Vol. 69, No. 211 / Tuesday, November 2, 2004 / Rules and Regulations 63887
loans from the market definition suggest that Affordable Housing Goal is rather small. For type, and for various assumptions in the 23–27 percent is a reasonable range of instance, adding the tax credit condition market projection model. These analyses estimates for the low- and moderate-income increased Fannie Mae’s performance as suggest that 23–27 percent is a reasonable market. This range covers markets without follows: 0.42 percentage point in 1999 (from estimate of the size of the conventional B&C and allows for market environments that 17.20 to 17.62 percent); 0.59 percentage point conforming market for the Special Affordable would be much less affordable than recent in 2000 (from 18.64 to 19.23 percent); and Housing Goal. This estimate excludes B&C 0.43 percent point in 2001 (from 19.29 to market conditions. loans and allows for the possibility that 19.72 percent). The increases for Freddie Mac Tax Credit Definition. Data are not homeownership will not remain as affordable available to measure the increase in market have been lower (ranging from 0.24 to 0.38 percentage point during the same period). as it has over the past six years. In addition, share associated with including low-income the estimate covers a range of projections 3. Conclusions units located in multifamily buildings that about the size of the multifamily market. meet threshold standards for the low-income Sensitivity analyses were conducted for the housing tax credit. Currently, the effect on market shares of each property type, for the [FR Doc. 04–24101 Filed 11–1–04; 8:45 am] GSE performance under the Special very-low-income shares of each property BILLING CODE 4210–27–P
VerDate jul<14>2003 18:20 Nov 01, 2004 Jkt 205001 PO 00000 Frm 00309 Fmt 4701 Sfmt 4700 C:\02NOR2.SGM 02NOR2