Monday, May 3, 2004

Part II

Department of Housing and Urban Development 24 CFR Part 81 HUD’s Proposed 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; Proposed Rule

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00001 Fmt 4717 Sfmt 4717 E:\FR\FM\03MYP2.SGM 03MYP2 24228 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

DEPARTMENT OF HOUSING AND HUD; codifying a process for handling General Counsel for Government URBAN DEVELOPMENT errors, omissions or discrepancies in a Sponsored Enterprises/RESPA, Office of GSE’s current year-end data the General Counsel, Room 9262, 24 CFR Part 81 submissions; clarifying that HUD may telephone 202–708–3137. The address [Docket No. FR–4790–P–01] exercise its goal counting authority by for all of these persons is Department of adjusting a GSE’s housing goals Housing and Urban Development, 451 RIN 2501–AC92 performance for a current year by Seventh Street, SW., Washington, DC, deducting miscredits from a previous 20410. Persons with hearing and speech HUD’s Proposed Housing Goals for the year caused by errors, omissions or impairments may access the phone Federal National Mortgage Association discrepancies in a GSE’s prior year data numbers via TTY by calling the Federal (Fannie Mae) and the Federal Home submissions (including the AHAR); and Information Relay Service at (800) 877– Loan Mortgage Corporation (Freddie clarifying that HUD may take 8399. Mac) for the Years 2005–2008 and enforcement action against the GSEs, as SUPPLEMENTARY INFORMATION: Amendments to HUD’s Regulation of authorized by FHEFSSA and as Fannie Mae and Freddie Mac implemented by HUD’s regulations at 24 I. General AGENCY: Office of the Assistant CFR part 81, subpart G, for the A. Statutory and Regulatory Background Secretary for Housing—Federal Housing submission of non-current, inaccurate or In 1968, at the time Fannie Mae was Commissioner, HUD. incomplete report(s), data or chartered in its current form as a information. ACTION: government sponsored enterprise (GSE), Proposed rule. In addition, HUD is proposing in this Congress assigned the Department of rulemaking to amend the definitions of SUMMARY: Through this proposed rule, Housing and Urban Development ‘‘Underserved area’’, ‘‘Metropolitan the Department of Housing and Urban (‘‘HUD’’ or ‘‘the Department’’) area’’ and ‘‘Minority’’, and to add a new Development is proposing new housing regulatory authority over Fannie Mae definition of the term ‘‘Home Purchase goal levels for the Federal National pursuant to section 802(ee) of the Mortgage’. Mortgage Association (Fannie Mae) and Housing and Urban Development Act of the Federal Home Loan Mortgage The rulemaking also invites comments on whether HUD should have 1968 (Pub. L. 90–448, approved August Corporation (Freddie Mac) (collectively, 1, 1968, 82 Stat. 476, 541) (HUD Act of the Government Sponsored Enterprises, a standard econometrically based method for imputing the distribution of 1968). In 1989, Congress granted the or GSEs) for calendar years 2005 Department essentially identical through 2008. The new housing goal GSE-purchased mortgages that lack income data, and whether HUD should authority over another GSE, Freddie levels are proposed in accordance with Mac, pursuant to section 731 of the the Federal Housing Enterprises revise its definitions or other rules (including the counting rules) to ensure Financial Institutions Reform, Recovery, Financial Safety and Soundness Act of and Enforcement Act of 1989 (FIRREA) 1992 (FHEFSSA) and govern the that only those large scale GSE transactions that are consistent with the (Pub. L. 101–73, approved August 9, purchase by Fannie Mae and Freddie 1989), which amended the Federal Mac of mortgages financing low- and statute and its purposes qualify under the goals. Home Loan Mortgage Corporation moderate-income housing, special Charter Act, Pub. L. 91–351, approved DATES: Comments must be submitted on affordable housing, and housing in July 24, 1970 (the ‘‘Freddie Mac Charter or before: July 2, 2004. central cities, rural areas and other Act’’). underserved areas. ADDRESSES: Interested persons are Under section 802(ee) of the HUD Act To increase homeownership invited to submit written comments of 1968, HUD was authorized to require opportunities for families targeted by regarding this proposed rule to the that a ‘‘reasonable portion’’ of Fannie the three housing goals, this rule also Regulations Division, Office of General Mae’s mortgage purchases be related to would establish new subgoals for the Counsel, Room 10276, Department of the national goal of providing adequate GSEs’ acquisitions of home purchase Housing and Urban Development, 451 housing for low- and moderate-income loans that qualify for each of the Seventh Street, SW., Washington, DC families. Accordingly, in 1978, the housing goals. Under the proposed rule, 20410. All communications should refer Department established by regulation performance under these subgoals to the above docket number and title. two housing goals for Fannie Mae: a would be calculated as percentages of Facsimile (FAX) comments and e-mail goal for mortgages on low- and the GSEs’ total acquisitions of home comments are not acceptable. A copy of moderate-income housing and a goal for purchase mortgages for single-family, each communication submitted will be mortgages on housing located in central owner-occupied properties located in available for public inspection and cities (see 24 CFR 81.16(d) and 81.17 of metropolitan areas meeting each of the copying between 8 a.m. and 5 p.m. HUD’s former rules at 43 FR 39203, three housing goals. weekdays at the above address. published August 15, 1978). HUD The Department also proposes to FOR FURTHER INFORMATION CONTACT: established each goal at the level of 30 revise the existing rule to provide Sandra Fostek, Director, Office of percent of Fannie Mae’s conventional enhanced requirements to ensure GSE Government Sponsored Enterprises, mortgage purchases. data integrity by: codifying the existing Office of Housing, Room 3150, Similar housing goals for Freddie Mac authority that authorizes HUD to telephone 202–708–2224. For questions were proposed by the Department in independently verify the accuracy and on data or methodology, contact John L. 1991 (at 56 FR 41022, published August completeness of data, information and Gardner, Director, Financial Institutions 16, 1991) but were not finalized prior to reports provided by the GSEs; Regulation Division, Office of Policy October 1992, when Congress enacted establishing certification requirements Development and Research, Room 8212, FHEFSSA and revised the Department’s for the submission of the GSEs’ Annual telephone (202) 708–1464. For legal GSE regulatory authorities, including Housing Activities Report (AHAR) and questions, contact Kenneth A. Markison, establishing new requirements for the for such other report(s), data Assistant General Counsel for housing goals. submission(s) or information for which Government Sponsored Enterprises/ Specifically, FHEFSSA established certification is requested in writing by RESPA or Paul S. Ceja, Deputy Assistant the Office of Federal Housing Enterprise

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00002 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24229

Oversight (OFHEO) as the GSEs’ safety GSEs through 1995 while the thereafter, the annual Housing Goals for and soundness regulator and affirmed, Department completed its development each of those years were to be clarified and expanded the Secretary of of post-transition Housing Goals (see 59 established at 50 percent, 31 percent, Housing and Urban Development’s GSE FR 61504). and 20 percent, respectively; regulatory authority. FHEFSSA also In 1995, the Department issued a (2) Made temporary bonus points provided that, except for certain proposed rule (60 FR 9154, published available for the GSEs’ purchases of exclusive authorities of the Director of February 16, 1995) and, several months mortgages for small multifamily OFHEO, and all other matters relating to later, a final rule (60 FR 61846, properties with 5 to 50 units, and, above the GSEs’ safety and soundness, the published December 1, 1995) (the a threshold, for single-family 2- to 4-unit Secretary had general regulatory power ‘‘Housing Goals 1995 final rule’’) owner-occupied rental properties, for over the GSEs. (See section 1321 of establishing the Housing Goals for the calendar years 2001 through 2003 (but FHEFSSA, 12 U.S.C. 4541.) years 1996 through 1999, along with not for subsequent years, unless Further, FHEFSSA detailed and regulations implementing FHEFSSA. determined by HUD); expanded the Department’s The Housing Goals 1995 final rule (3) Established a temporary responsibilities to establish, monitor, provided that the Housing Goals for adjustment factor (‘‘TAF’’) for Freddie and enforce housing goals for the GSEs’ 1999 would continue beyond 1999 if the Mac’s purchases of mortgages on large purchases of mortgages that finance Department elected not to change the multifamily properties (over 50 units) housing for low- and moderate-income Housing Goals, and that HUD could for calendar years 2001 through 2003; families (the ‘‘Low- and Moderate- change the level of the Housing Goals (4) Prohibited high-cost mortgage Income Housing Goal’’), housing located for the years 2000 and beyond based loans with predatory features from in central cities, rural areas, and other upon HUD’s experience and in receiving Housing Goals credit; underserved areas (the ‘‘Underserved accordance with HUD’s statutory (5) Established and clarified counting Areas Housing Goal’’), and special authority and responsibility. rules under the Housing Goals for the affordable housing, affordable to very The Housing Goals 1995 final rule treatment of missing affordability data, low-income families and low-income established counting requirements to purchases of seasoned mortgage loans, families in low-income areas (the calculate performance under the purchases of federally insured mortgage ‘‘Special Affordable Housing Goal’’) Housing Goals. The Housing Goals 1995 loans and purchases of mortgage loans (collectively, the ‘‘Housing Goals’’ or, final rule also: (1) Prohibited the GSEs on properties with expiring assistance individually, the ‘‘Housing Goal’’). (See, from discriminating in any manner, on contracts; generally, sections 1331–1334 of any prohibited basis, in their mortgage (6) Established procedures for HUD’s FHEFSSA, 12 U.S.C. 4561–4564.) There purchases; (2) implemented procedures review of transactions to determine is also a subgoal under the Special for the exercise of HUD’s new program appropriate Housing Goal treatment; Affordable Housing Goal for multifamily review authority; (3) established and housing. reporting requirements and a public use (7) Made certain definitional and Under FHEFSSA, the Department is data base of the GSEs’ mortgage technical corrections to the Housing required to establish each Housing Goal purchase activities; (4) provided Goals 1995 final rule. after consideration of certain factors that protections for GSE confidential and The Housing Goals 2000 final rule are relevant to the particular Housing proprietary information; and (5) provided for the award of temporary Goal, including: (a) National housing established enforcement procedures. bonus points (double credit) toward the needs; (b) economic, housing and On March 9, 2000, HUD published a Housing Goals for both GSEs’ mortgage demographic conditions; (c) the proposed rule to establish new Housing purchases that financed single-family, performance and efforts of the GSEs Goal levels for Fannie Mae and Freddie owner-occupied 2–4 unit properties and toward achieving the Housing Goal in Mac for calendar years 2000 through 5–50 unit multifamily properties. Under previous years; (d) the size of the market 2003 (see 65 FR 12632–12816). On the TAF, the rule also awarded Freddie for mortgages targeted by the Housing October 31, 2000, after analyzing over Mac 1.2 units credit for each Goal relative to the overall conventional 250 comments, HUD issued a final rule multifamily unit in property over 50 mortgage market; (e) the ability of the establishing the new Housing Goals (the units.1 The Housing Goals 2000 final GSEs to lead the industry in making ‘‘Housing Goals 2000 Final Rule,’’ 65 FR rule made clear, however, that both of credit available for mortgages targeted 65044–65229). these measures were temporary, by the Housing Goal; and (f) the need to The Housing Goals 2000 final rule intended to encourage the GSEs to ramp maintain the sound financial condition increased the level of the Housing Goals up their efforts to meet financing needs of the GSEs. (See sections 1332(b), for Fannie Mae and Freddie Mac. that had not been well served. During 1333(a)(2), 1334(b) of FHEFSSA; 12 Specifically, this rule: the three years for which the temporary U.S.C. 4562(b); 12 U.S.C. 4563(a)(2); and (1) Increased the level of the Housing bonus points and TAF were established, 12 U.S.C. 4564.) (There are slight Goals for calendar years 2001 through HUD expected the GSEs to develop new, differences among the three Housing 2003 as follows: sustainable business relationships and Goals in the statutory specification of • The Low- and Moderate-Income purchasing strategies for the targeted the factors. In particular, for the Special Housing Goal increased to 50 percent; needs. Affordable Housing Goal factors (b) and • The Underserved Areas Housing At the end of the three years (2001– (d) are absent, and there is a factor for Goal increased to 31 percent; 2003), the Department determined not • data submitted in previous years to the The Special Affordable Housing to extend the bonus points or the TAF, Secretary in connection with the Goal increased to 20 percent; • after careful review of the facts and Housing Goal.) The Special Affordable Multifamily circumstances of performance under the For the transition period of 1993– Subgoal increased to the respective Housing Goals. Data indicate that both 1994, FHEFSSA required HUD to average of one percent of each GSE’s GSEs increased their financing of units establish interim Housing Goals, which total mortgage purchases during the

HUD did in 1993 (at 53 FR 53048). In period of 1997 Through 1999; and 1 • Congress increased the level of the TAF to 1.35 November 1994, HUD extended the Pending establishment of annual per unit, section 1002 of Pub. L. 106–554 (December 1994 interim Housing Goals for both Housing Goals for the year 2004 and 21, 2000).

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00003 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24230 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

targeted by the bonus points and the 303(e) of the Freddie Mac Charter Act Specifically, as indicated, the GSEs’ TAF. and section 309(c)(2) of the Fannie Mae Charter Acts require them to continually Charter Act). assist in the efficient functioning of the B. Background: Fannie Mae and Freddie secondary market for residential Mac Fannie Mae and Freddie Mac engage in two principal businesses: purchasing mortgages, including mortgages for low- Fannie Mae and Freddie Mac were and otherwise investing in residential and moderate-income families that may chartered by the Congress as mortgages and guaranteeing securities involve a reasonable economic return government sponsored enterprises. backed by residential mortgages. that is less than the economic return on Pursuant to section 301 of the Federal While the securities that the GSEs other mortgages. The GSEs also are National Mortgage Association Charter guarantee, and the debt instruments required to promote access to mortgage Act (the ‘‘Fannie Mae Charter Act’’, 12 they issue, are explicitly not backed by credit throughout the nation, including U.S.C. 1716, et seq.) and section 301(b) the full faith and credit of the United central cities, rural areas, and other of the Federal Home Loan Mortgage States, and nothing in this proposed underserved areas. These statutory Corporation Act (the ‘‘Freddie Mac rule should be construed otherwise, mandates obligate the GSEs to work to Charter Act’’, 12 U.S.C. 1451, et seq.), such securities and instruments trade at ensure that everyone in the nation has the GSEs were chartered expressly to: yields only a few basis points over those a reasonable opportunity to enjoy access (1) Provide stability in the secondary of U.S. Treasury securities with to the mortgage financing benefits market for residential mortgages; comparable terms. Moreover, these resulting from the activities of these (2) Respond appropriately to the securities also offer yields lower than enterprises. private capital market; The GSEs have achieved an important (3) Provide ongoing assistance to the those for securities issued by fully private firms that are more highly part of their mission: providing stability secondary market for residential and liquidity to large segments of the mortgages (including activities relating capitalized but otherwise comparable. These factors, in addition to the fact housing finance markets. They have also to mortgages on housing for low- and increased their purchases of loans moderate-income families involving a that the market does not require that individual GSE securities be rated by a affordable to low-income families over reasonable economic return that may be the past decade since the affordable less than the return earned on other national rating agency, evidence that investors perceive that GSE-guaranteed housing goals were put in place under activities) by increasing the liquidity of FHEFSSA. Through partnership efforts, securities have inherent advantages over mortgage investments and improving new product offerings, and flexible other types of guaranteed securities in the distribution of investment capital underwriting and purchase standards, light of the GSEs’ relationship to the available for residential mortgage both enterprises have reached out to Federal Government, including their financing; and underserved borrowers, as discussed public purposes, their Congressional (4) Promote access to mortgage credit below in this preamble and in the charters, and the explicit benefits throughout the nation (including central appendices. cities, rural areas, and other provided in their charters as described The major premise of this proposed underserved areas) by increasing the above. rule is that the GSEs must further utilize liquidity of mortgage investments and Consequently, the GSEs are able to their entrepreneurial talents and power improving the distribution of fund their operations at lower cost than in the marketplace to genuinely ‘‘lead investment capital available for other private firms with similar the mortgage finance industry’’ and to residential mortgage financing. financial characteristics. In a recent ‘‘ensure that citizens throughout the As a result of their status as GSEs, report, the Congressional Budget Office country enjoy access to the public Fannie Mae and Freddie Mac receive (CBO) estimated this funding advantage benefits provided by these federally significant explicit benefits that are not for the year 2003 to be a $19.6 billion related entities.’’ (See, S. Rep. No. 282, enjoyed by fully private shareholder- annual combined subsidy for both GSEs. 102d Cong., 2d Sess. 34 (1992).) owned corporations in the mortgage Of this amount, CBO estimated that the For example, despite the record market. These benefits include: GSEs retained about $6.2 billion, or national homeownership rate of 67.9 • Conditional access to a $2.25 billion approximately one-third of the subsidy, percent in 2002, certain segments of the line of credit from the U.S. Treasury (see for their officers and shareholders, population clearly have not benefited to section 306(c)(2) of the Freddie Mac while the remainder accrued to the same degree that others have from Charter Act and section 304(c) of the borrowers.3 the advantages and efficiencies Fannie Mae Charter Act); C. Secretary’s Approach To Regulating provided by Fannie Mae and Freddie • Exemption from the securities the GSEs Mac. Problems continue to persist for registration requirements of the low-income families and certain Securities and Exchange Commission In return for the public benefits they minorities: and the States (see section 306(g) of the receive, Congress has mandated in the • Lower homeownership rates prevail Freddie Mac Charter Act and section GSEs’ Charter Acts that the GSEs carry for certain minorities, especially for 304(d) of the Fannie Mae Charter Act); 2 out public purposes not required of African-American households (47.9 and other private sector entities in the percent) and Hispanics (48.2 percent). • Exemption from all State and local housing finance industry. These gaps are only partly explained by taxes except property taxes (see section differences in income, age, and other 3 ‘‘Updated Estimates of the Subsidies to the socioeconomic factors. Disparities in 2 Fannie Mae and Freddie Mac have both Housing GSEs’’, attachment to a letter from Douglas mortgage lending are reflected in loan announced their intention voluntarily to register Holtz-Eakin, Director, Congressional Budget Office, their common stock with the Securities and to the Honorable Richard C. Shelby, Chairman, denial rates of minority groups when Exchange Commission (SEC) under section 12(g) of Committee on Banking, Housing, and Urban Affairs, compared to white applicants. Denial the Securities Exchange Act of 1934. Fannie Mae’s United States Senate, April 8, 2004. A related recent rates for conventional home purchase registration became effective March 31, 2003. study is Wayne Passmore, ‘‘The GSE Implicit mortgage loans (excluding Freddie Mac has stated that it will complete the Subsidy and Value of Government Ambiguity,’’ process of voluntarily registering its common stock Board of Governors of the Federal Reserve System, manufactured housing loans) in 2002 once it resumes timely reporting of its financial Finance and Economics Discussion Series, FEDS were 19.9 percent for African results. Working Paper 2003–64, December 2003. Americans, 14.0 percent for Native

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00004 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24231

American applicants, 15.1 percent for received the benefits of mainstream achieve a Housing Goal. Regulating two Hispanic applicants, 8.9 percent for lenders supported by an active exceedingly large financial enterprises Asian applicants, and 7.9 percent for secondary market; in a dynamic market requires that HUD White applicants. • Expanding their penetration in the provide the GSEs with sufficient • While Fannie Mae and Freddie Mac following market segments: (1) latitude to use their innovative cannot be expected to solve all these Borrowers with credit blemishes, or capacities to determine how best to problems, they have both the resources with little traditional credit history; (2) develop products to carry out their and the expertise to improve credit first-time homebuyers; (3) Community respective missions. HUD’s regulations access for low- and moderate-income Reinvestment Act (‘‘CRA’’)-related are intended to allow the GSEs the families, minority families, and families loans, which are loans to low- and flexibility to respond quickly to market in underserved areas. The GSEs also moderate-income populations and opportunities. At the same time, the have the ability to increase the financing neighborhoods in a financial Department must ensure that the GSEs’ of affordable multifamily rental housing. institution’s assessment area as strategies address national credit needs, Yet, studies by HUD and others show established under the CRA; (4) the especially as they relate to housing for that the GSEs generally have been less rental property market; and (5) the low- and moderate-income families and active in historically underserved market for rehabilitation loans; and housing located in underserved markets where there is a need for • Increasing their outreach to, and geographical areas. The addition of additional sources of financing to achieving greater efficiency in, the Home Purchase Subgoals to the address persistent housing and credit above identified markets, as well as in regulatory structure provides an needs, and fully private companies, other markets that serve low-income additional means of encouraging the operating without the benefits of GSE and moderate-income families and GSEs’ affordable housing activities to status, perform better in these markets. families living in underserved areas. address identified, persistent credit • Between 1999 and 2002, special Under the present rulemaking, the needs while leaving to the GSEs the affordable housing borrowers accounted Department is proposing new, higher specific approaches to meeting these for 14.4 percent of Fannie Mae’s levels for the Housing Goals, needs. acquisitions of home purchase mortgage accompanied by subgoals under each of (3) Discrimination in lending—albeit loans and 14.5 percent of Freddie Mac’s the Housing Goals for purchases of sometimes subtle and unintentional— acquisitions, at the same time that such home purchase mortgages on owner- has denied racial and ethnic minorities mortgages accounted for 16.4 percent of occupied properties in metropolitan the same access to credit to purchase a home purchase loans originated in the areas. (The subgoals are hereafter home that has been available to overall conventional, conforming referred to in this rule as ‘‘Home similarly situated non-minorities. As market (excluding B&C loans) in Purchase Subgoal’’ or ‘‘Subgoal’’.) The noted above, troublesome gaps in metropolitan areas. Department’s purpose in proposing homeownership remain for minorities • During the same period, mortgage higher Housing Goals and in even after record growth in affordable purchases on properties located in establishing new Home Purchase lending and homeownership during the underserved areas accounted for 24.0 Subgoals in this rulemaking is to nineties. Studies indicate that, over the percent and 22.9 percent of Fannie encourage the GSEs to facilitate greater next few years, minorities will account Mae’s and Freddie Mac’s acquisitions of financing and homeownership for a growing share of the families home purchase loans, respectively, and opportunities for families and seeking to buy their first home. HUD’s 25.8 percent of home purchase neighborhoods targeted by the Housing analyses indicate, however, that Fannie mortgages originated in the primary Goals. In developing these regulations, Mae and Freddie Mac account for a market. the Department was guided by, and re- relatively small share of the minority • Both Fannie Mae and Freddie Mac affirms, the following principles first-time homebuyer market. The GSEs have lagged the market in funding first- established in the Housing Goals 1995 have a responsibility to promote access time homebuyers. Between 1999 and final rule: to capital for minorities and others who 2002, first-time homebuyers accounted (1) The GSEs should fulfill are seeking their first homes, and to for 27 percent of each GSE’s purchases FHEFSSA’s intent that they lead the demonstrate the benefits of such lending of home purchase loans, compared with industry in ensuring that access to to industry and borrowers alike. The 38 percent for home purchase loans mortgage credit is made available for GSEs also have an integral role in originated in the conventional very low-, low- and moderate-income eliminating predatory mortgage lending conforming market. families and residents of underserved practices. Fannie Mae and Freddie Mac have areas. HUD recognizes that, to lead the (4) In addition to the GSEs’ purchases increased their role in providing mortgage industry over time, the GSEs of single-family home mortgages, the financing for the low-income end of the will have to stretch to reach certain GSEs also must continue to assist in the mortgage market, but the GSEs need to Housing Goals and to close gaps creation of an active secondary market increase their efforts further and between the secondary mortgage market for mortgages on multifamily rental demonstrate their capacity to be and the primary mortgage market for housing. Affordable rental housing is industry leaders. There are ample various categories of loans. This essential for those families who cannot market opportunities for them to do so, approach is consistent with the afford to become, or who choose not to including: Congress’ directive that ‘‘the enterprises become, homeowners. For this reason, • Continuing to introduce new will need to stretch their efforts to the GSEs must assist in making capital products, and providing greater achieve’’ the goals (see S. Rep. No. 282, available to assure the continued flexibility in their purchase and 102d Cong., 2d Sess., 35 (1992)). development of single-family and underwriting guidelines, to better (2) The Department’s role as a multifamily rental housing. address the unique circumstances of regulator is to set broad performance With these principles in mind, the low-income families; standards for the GSEs through the Department is proposing levels of the • Continuing to look for sound Housing Goals, but not to dictate the Housing Goals that will bring the GSEs investment opportunities in those specific products or delivery to a position of market leadership in a lower-income sectors that have not yet mechanisms the GSEs will use to range of foreseeable economic

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00005 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24232 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

circumstances related to the future address the areas formerly targeted by 2005–2008, at 1.0 percent of their course of interest rates and consequent these measures. The business respective average dollar volumes of fluctuations in origination rates on relationships that the GSEs established mortgage purchases in calendar years home purchase and refinance when these provisions were in place 2000, 2001, and 2002. This would mortgages—both multifamily and will be necessary to meet the higher increase the dollar value to $5.49 billion single-family. For each Goal, HUD has Housing Goals. annually for Fannie Mae and $3.92 projected Goal-qualifying percentages of The Department’s proposals to billion annually for Freddie Mac. mortgage originations in terms of ranges increase the levels of the Housing Goals, The Housing Goal percentages that are that cover a variety of economic and to establish new Home Purchase proposed in this rule reflect the scenarios. The objective of HUD’s Subgoals, are predicated upon its application of area median incomes and proposed Housing Goals is to bring the recognition that the GSEs not only have minority percentages based on 2000 GSEs’ performance to the upper end of the ability to achieve these Housing Census data, the Census Bureau’s HUD’s market range estimate for each Goals but, also, that they are fully specification of census tract boundaries Goal, consistent with the statutory consistent with the statutory factors for the 2000 Census, and the Office of criterion that HUD should consider the established under FHEFSSA. In Management and Budget’s specification GSEs’ ability to lead the market for each addition, these proposals are supported of metropolitan area boundaries based Goal. To enable the GSEs to achieve this by the Department’s comprehensive on the 2000 Census. leadership, the Department is proposing analyses of the size of the mortgage 2. HUD’s Consideration of Statutory modest increases in Housing Goal levels market, the opportunities available to Factors in Setting the Housing Goals for 2005 which will increase further, the GSEs, America’s unmet housing year-by-year through 2008, to achieve needs, and identified credit gaps. As discussed above, HUD considered the ultimate objective for the GSEs to The Department anticipates that, as six statutory factors before it decided lead the market under a range of the GSEs’ businesses grow, the upon the levels of the Housing Goals foreseeable economic circumstances by increased level of the Housing Goals, being proposed in this rulemaking, as 2008. Such a program of staged and the new Home Purchase Subgoals, described in Section III(B) of this increases is consistent with the statutory will enable the GSEs to continue to preamble and proposed rule amendment requirement that HUD consider the past address new markets and persistent, numbers 3–5 of this proposed rule. A performance of the GSEs in setting the unmet housing finance needs. summary of HUD’s findings relative to Goals. Staged annual increases in the each factor follows. More detailed Goals will provide the enterprises with II. Implementation discussion of these points is included in opportunity to adjust their business A. Affordable Housing Goals Appendices A, B, and C. models and prudently try out business 1. Proposed Changes to Housing Goal a. Demographic, Economic, and Housing strategies, so as to meet the required Conditions 2008 levels without compromising other Levels business objectives and requirements. The current Housing Goal levels are (i) Demographic Trends. Changing The Department believes that the 50 percent for the Low- and Moderate- population demographics will result in Home Purchase Subgoals that it Income Housing Goal, 31 percent for the a need for the primary and secondary proposes to establish under this Underserved Areas Housing Goal, and mortgage markets to meet nontraditional rulemaking are necessary and 20 percent for the Special Affordable credit needs, respond to diverse housing warranted. Increasing homeownership Housing Goal. The Special Affordable preferences and overcome information is a national priority. As detailed below, Housing Goal includes a Subgoal for and other barriers that many immigrants the GSEs must apply greater efforts to mortgage purchases financing dwelling and minorities face. increasing homeownership for low- and units in multifamily housing which is The U.S. Census Bureau has projected moderate-income families, families 1.0 percent of the average annual dollar that the U.S. population will grow by an living in underserved areas, and very- volume of mortgages (both single-family average of 2.5 million persons per year low income families and low-income and multifamily) purchased by the between 2000 and 2025, resulting in families living in low-income areas. The respective GSE in 1997, 1998, and about 1.2 million new households per addition of Home Purchase Subgoals to 1999—$2.85 billion annually for Fannie year. The aging of the baby-boom the regulatory structure will serve to Mae and $2.11 billion annually for generation and the entry of the baby- better focus the GSEs’ efforts in a clear Freddie Mac. bust generation into prime home-buying and transparent manner and better The Department is proposing in this age will have a dampening effect on allow the government and public alike rulemaking to increase the Housing Goal housing demand. Growing housing to monitor the GSEs’ efforts in meeting levels as follows: demand from minorities, immigrants the nation’s homeownership needs. • The proposed level of the Low- and and non-traditional homebuyers will Moreover, the Department reaffirms Moderate-Income Housing Goal is 52 help offset declines in the demand for its view that neither the award of bonus percent in 2005, 53 percent in 2006, 55 housing caused by the aging of the points for particular mortgage purchases percent in 2007, and 57 percent in 2008; population. nor the temporary adjustment factor for • The proposed level of the The continued influx of immigrants Freddie Mac’s multifamily purchases Underserved Areas Housing Goal is 38 will increase the demand for rental are necessary. At this point, their percent in 2005, 39 percent in 2006, 39 housing, while those who immigrated continued use would only result in percent in 2007, and 40 percent in 2008; during the 1980s and 1990s will be in misleading information about the extent and the market for homeownership. to which the GSEs are, in fact, meeting • The proposed level of the Special Immigrants and minorities—who the Housing Goals. The decision to Affordable Housing Goal is 22 percent accounted for nearly 40 percent of the increase the levels of the Housing Goals in 2005, 24 percent in 2006, 26 percent growth in the nation’s homeownership substantially in a staged manner under in 2007, and 28 percent in 2008. rate over the past five years—will be this proposal and, at the same time, not • In addition, HUD is proposing to responsible for almost two-thirds of the to renew the bonus points or TAF, will retain the Special Affordable growth in the number of new ensure that the GSEs continue to Multifamily Subgoal for calendar years households over the next ten years.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00006 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24233

Non-traditional households have million units for 2004 and 5.8 million Hispanics and African Americans being become more important, as overall for 2005, compared to their record 6 told that housing units were unavailable household formation rates have slowed. million level in 2003. when non-Hispanic whites found them With later marriages, divorce, and non- (iii) Mortgage Market Conditions. Low to be available. The study also found traditional living arrangements, the interest rates and record levels of other worrisome trends of fastest growing household groups have refinancing caused mortgage discrimination in metropolitan housing been single-parent and single-person originations to soar from $2.2 trillion in markets that persisted in 2000, for households. By 2025, non-family 2001 to $2.9 trillion in 2002 and around example, geographical steering households will make up a third of all $3.8 trillion in 2003. Fannie Mae experienced by African-American households. The role of traditional 25- projects that mortgage originations will homebuyers, and real estate agents who to-34 year-old married, first-time drop to $2.4 trillion in 2004 and $1.7 provided less assistance in obtaining homebuyers in the housing market will trillion in 2005, as refinancing returns to financing for Hispanic homebuyers than be smaller in the current decade due to more normal levels. The volume of for non-Hispanic whites.4 Racial the aging of the population. Between home purchase mortgages was $910 disparities in mortgage lending are also 2000 and 2025, the Census Bureau billion to $1.1 trillion between 1999 and well documented. HUD-sponsored projects that the largest growth in 2001 before jumping to $1.2 trillion in studies of the pre-qualification process households will occur among 2002 and $1.3 trillion in 2003. As with conclude that African Americans and householders 65 and over. housing starts, the home purchase Hispanics face a significant risk of As these demographic factors play origination market is expected to exhibit unequal treatment when they visit out, the overall effect on housing sustained growth. mainstream mortgage lenders. Studies demand will likely be continued growth b. National Housing Needs have shown that mortgage denial rates and an increasingly diverse household are substantially higher for African population from which to draw new (i) Affordability Problems. Data from Americans and Hispanics, even after renters and homeowners. A greater the 2000 Census and the American controlling for applicant income and a diversity in the housing market will, in Housing Surveys demonstrate that there host of underwriting characteristics, turn, require greater adaptation by the are substantial housing needs among such as the credit record of the primary and secondary mortgage low- and moderate-income families. applicant.5 markets. Many of these households are burdened The existence of substantial (ii) Economic and Housing by high homeownership costs or rent neighborhood disparities in Conditions. While most other sectors of payments and, consequently, are facing homeownership and mortgage credit is the economy were weak or declining serious housing affordability problems. also well documented for metropolitan during 2001 and 2002, the housing There is evidence of persistent areas. HUD’s analysis of HMDA data sector showed remarkable strength. The housing problems for Americans with shows that mortgage credit flows in housing market continued at a record the lowest incomes. HUD’s analysis of metropolitan areas are substantially pace during 2003. American Housing Survey data reveals lower in high-minority and low-income In 2002, the U.S. economy moved into that, in 2001, 5.1 million households neighborhoods and mortgage denial recovery, with real Gross Domestic had ‘‘worst case’’ housing needs, rates are much higher for residents of Product (GDP) growing 2.2 percent, defined as housing costs greater than 50 these neighborhoods. Studies have also although measures of unemployment percent of household income or severely documented that mainstream lenders continued to rise. In October 2002, the inadequate housing among unassisted often do not operate in inner-city average 30-year home mortgage interest very-low-income renter households. minority neighborhoods, leaving their rate slipped below 6 percent for the first Among these households, 90 percent residents with only high-cost lenders as time since the mid-1960s. Favorable had a severe rent burden, 6 percent options. Too often, residents of these financing conditions and solid increases lived in severely inadequate housing, same neighborhoods have been in house prices were the key supports and 4 percent suffered from both subjected to the abusive practices of to record housing markets during both problems. Among the 34 million renters predatory lenders. 2002 and 2003. By the end of 2003, the in all income categories, 6.3 million (19 These troublesome disparities mostly industry had set new records in single- percent) had a severe rent burden and affect those families (minorities and family permits, new home sales, over one million renters (3 percent) immigrants) who are projected to existing home sales, interest rates, and lived in housing that was severely account for almost two-thirds of the homeownership. Other indicators—total inadequate. growth in the number of new permits, starts, completions, and (ii) Disparities in Housing and households over the next ten years. affordability—reached levels that were Mortgage Markets. Despite the strong (iii) Single-Family Market: Trends in among the highest in the past two growth in affordable lending over the Affordable Lending and decades. past ten years, there are families who Homeownership. Many younger, Over the near term, the are not being adequately served by the minority and lower-income families did Administration’s forecast for real GDP nation’s housing and mortgage markets. not become homeowners during the growth is 4.0 percent for 2004, while the Serious racial and income disparities 1980s due to the slow growth of Congressional Budget Office (CBO) remain. The homeownership rate for earnings, high real interest rates, and projects that real GDP will grow at an minorities is 25 percentage points below continued house price increases. Over average rate of 3.2 percent from 2005 that for whites. A major HUD-funded the past ten years, economic expansion, through 2008. The ten-year Treasury study of discrimination in the sales and accompanied by low interest rates and rate is projected to average 5.5 percent rental markets found that while between 2005 and 2008 compared to its discrimination against minorities was 4 Margery Austin Turner, All Other Things Being average of 4.6 percent in 2002 and 4.0 generally down since 1989, it remained Equal: A Paired Testing Study of Mortgage Lending percent in 2003. Standard & Poor’s at unacceptable levels in 2000. The most Institutions, The Urban Institute Press, April 2002. Appendix A includes further discussion of this expects housing starts to average 1.8 prevalent form of discrimination against study. million units in 2004–05. Fannie Mae Hispanic and African-American home 5 These studies are discussed in section B.1 of projects existing home sales at 6.1 seekers observed in the study was Appendix B.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00007 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24234 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

increased outreach on the part of the history; lack of access to mainstream investment they can make, far ahead of mortgage industry, has improved lenders; little understanding of the 401(k)s, other retirement accounts, and affordability conditions for these home buying process; a limited supply stocks. Forty-two percent of African- families. of modestly priced homes; and American families reported that they As this preamble and the appendices continued discrimination in housing were ‘‘very or fairly likely’’ to buy a note, there has been a ‘‘revolution in markets and mortgage lending. These home in the next three years, up from affordable lending’’ that has extended barriers are discussed in Appendix A. 38 percent in 1998 and 25 percent in homeownership opportunities to (iv) Single-Family Market: Potential 1997. Among Hispanics and Hispanic historically underserved households. Homeowners. As already noted, the immigrants, the numbers reached 37 The mortgage industry, including the potential homeowner population over percent and 34 percent, respectively. GSEs, has offered more customized the next decade will be highly diverse, The survey also reported that more than mortgage products, more flexible as growing housing demand from half of Hispanic renters cite underwriting, and expanded outreach to immigrants (both those who are already homeownership as being ‘‘one of their low-income and minority borrowers. in this country and those who are top priorities.’’ HMDA data suggest that the industry projected to arrive), minorities, and non- In spite of these trends, potential and GSE initiatives are increasing the traditional homebuyers will help to minority and immigrant homebuyers see flow of credit to underserved borrowers. offset declines in the demand for more obstacles to buying a home than Between 1993 and 2002, conventional housing caused by the aging of the does the general public. Typically, the loans to low-income and minority population. primary barriers to homeownership are families increased at much faster rates Fannie Mae reports that, between credit issues and a lack of funds for a than loans to upper-income and non- 1980 and 1995, the number of new downpayment and closing costs. minority families. Conventional home immigrant owners increased by 1.4 However, other barriers also exist, such purchase originations to African- million and, between 1995 and 2010, as a lack of affordable housing, little Americans more than doubled between that figure is expected to rise by more understanding of the home buying 1993 and 2002 and those to Hispanic than 50 percent to 2.2 million. These process, and language barriers. Thus, borrowers more than tripled during this trends do not depend on the future the new group of potential homeowners period. Home loans to low-income inflow of new immigrants, as will have unique needs. borrowers and to low-income and high- immigrants do not, on average, enter the The GSEs can play an important role minority census tracts also more than home purchase market until they have in tapping this potential homeowner doubled during this period. been in this country for eleven years. population. Along with others in the Thus, the 1990s and the early part of Fannie Mae staff note that there are industry, they can address these needs the current decade have seen the enough immigrants already in this on several fronts, such as expanding development of a strong affordable country to keep housing demand strong education and outreach efforts, lending market. Homeownership for several years. introducing new products, and statistics show similar trends. After Thus, the need for the GSEs and other adjusting current underwriting declining during the 1980s, the industry participants to meet standards to better reflect the special homeownership rate has increased nontraditional credit needs, respond to circumstances of these new households. every year since 1994, reaching a record diverse housing preferences, and to These efforts will be necessary if the mark of 67.9 percent in 2002. The overcome the information barriers that Administration’s goal of expanding number of households owning their many immigrants face will take on minority homeownership by 5.5 million own home in 2002 was 10.6 million added importance. A new or recent families by the end of the decade is to greater than in 1994. Gains in immigrant may have no credit history be achieved. (In this regard, the Joint homeownership rates have been or, at least, may not have a credit history Center for Housing Studies has stated widespread over the last eight years, that can be documented by traditional that, if favorable economic and housing with the homeownership rate for methods. In order to address these market trends continue, and if African American households needs, the GSEs and the mortgage additional efforts to target mortgage increasing from 42.5 percent to 47.9 industry have been developing lending to low-income and minority percent, for Hispanic households from innovative products and seeking to households are made, the 41.2 percent to 48.2 percent, for non- extend their outreach efforts to attract homeownership rate could reach 70 Hispanic white households from 50.8 these homebuyers, as discussed in percent by 2010.) percent to 55.1 percent, and for central Appendix A. The single-family mortgage market city residents from 48.5 percent to 51.8 In addition, the current low has been very dynamic over the past few percent from 1994 to 2002. homeownership rates in inner cities (as years, experiencing volatile swings in Despite the record gains in compared with the suburbs) also suggest originations (with the 1998 and 2001– homeownership since 1994, a that urban areas may be a potential 2003 refinancing waves), witnessing the substantial gap in the homeownership growth market for lenders. As explained rapid growth in new types of lending rate of approximately 25 percentage in Appendix A, lenders are beginning to (such as subprime lending), points prevails for African-American recognize that urban borrowers and incorporating new technologies (such as and Hispanic households as compared properties have different needs than automated underwriting systems), and to white non-Hispanic households. suburban borrowers and properties. facing serious challenges (such as Studies show that these lower CRA-type lending will continue to be abusive predatory lending). Fannie Mae homeownership rates are only partly important in our inner cities. and Freddie Mac have played a major accounted for by differences in income, Surveys indicate that these role in the ongoing changes in the age, and other socioeconomic factors. demographic trends will be reinforced single-family market and in helping the In addition to low income, barriers to by the fact that most Americans desire, industry address the problems and homeownership that disproportionately and plan, to become homeowners. challenges that have arisen. affect minorities and immigrants According to Fannie Mae’s 2002 The appendices to this proposed rule include: lack of capital for down National Housing Survey, Americans discuss the various roles that Fannie payment and closing costs; poor credit rate homeownership as the best Mae and Freddie Mac have played in

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00008 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24235

the single-family market. A wide range has major implications for the Low- and rates, and market rents of multifamily of topics is examined, including the Moderate-Income Housing and Special properties appear to be highly correlated GSEs’ automated underwriting Affordable Housing Goals, since high with local job market conditions, technology used throughout the percentages of multifamily units have creating greater sensitivity in loan industry, their many affordable lending affordable-level rents and can count performance to economic conditions partnerships and underwriting toward one or both of these Housing than may be experienced for single- initiatives aimed at extending credit to Goals. However, the potential of the family mortgages. underserved borrowers, their GSEs to lead the multifamily mortgage There is a need for an ongoing GSE development of new targeted low- industry has not been fully developed. presence in the multifamily secondary downpayment products, their entry into The GSEs’ purchases between 1999 and market, both to increase liquidity and to new markets such as subprime lending, 2002 accounted for only 30 percent of further affordable housing efforts. The and their attempts to reduce predatory the multifamily units that received potential for an increased GSE presence lending. As that discussion emphasizes, financing during this period. Certainly is enhanced by the fact that an the GSEs have the ability to bring there are ample opportunities and room increasing proportion of multifamily increased efficiencies to a market and to for expansion of the GSEs’ share of the mortgages are now originated in attract mainstream lenders into markets. multifamily mortgage market. accordance with secondary market (Readers are referred to Appendices A– The GSEs’ size and market position standards. Small multifamily properties, C for further discussion of the GSEs’ between loan originators and mortgage and multifamily properties with role in different segments of the single- investors make them the logical significant rehabilitation needs, have family mortgage market.) institutions to identify and promote historically experienced difficulty (v) Multifamily Mortgage Market. The needed innovations and to establish gaining access to mortgage financing, market for financing of multifamily standards that will improve market and the flow of capital into multifamily apartments has reached record volume. efficiency. As their role in the housing for seniors has been historically The favorable long-term prospects for multifamily market continues to grow, characterized by volatility. The GSEs apartments, combined with record low the GSEs will have the knowledge and can play a role in promoting liquidity interest rates, have kept investor market presence to push simultaneously for multifamily mortgages and demand for apartments strong and have for standardization and for increasing the availability of long-term, also supported property prices. programmatic flexibility to meet special fixed rate financing for these properties. Fannie Mae and Freddie Mac have needs and circumstances, with the been among those boosting their ultimate goal of increasing the c. GSEs’ Past Performance and Effort volumes of multifamily financing and availability and reducing the cost of Toward Achieving the Housing Goals both have introduced new programs to financing for affordable and other Both Fannie Mae and Freddie Mac serve the multifamily market. Fannie multifamily rental properties. have improved their affordable housing Mae and, especially (considering its The long-term outlook for the loan performance over the past ten early withdrawal from the market), multifamily rental market is sustained, years, since the enactment of FHEFSSA Freddie Mac have rapidly expanded moderate growth, based on favorable their presence in the multifamily demographics. The minority population, and HUD’s establishment in 1993 of the mortgage market under the Housing especially Hispanics, provides a Housing Goals. However, the GSEs’ Goals. growing source of demand for affordable mortgage purchases have generally Freddie Mac has successfully rebuilt rental housing. ‘‘Lifestyle renters’’ lagged, and not led, the overall primary its multifamily acquisition program, as (older, middle-income households) are market in providing financing for shown by the increase in its purchases also a fast-growing segment of the rental affordable housing to low- and of multifamily mortgages: from $27 population. moderate-income families and million in 1992 to $3 billion in 1997 At the same time, the provision of underserved borrowers and their and then to approximately $7 billion affordable housing units will continue neighborhoods, indicating that there is annually during the next three years to challenge suppliers of multifamily more that the GSEs can do to improve (1998 to 2000), before rising further to rental housing as well as policy makers their performance. $11.9 billion in 2001 and $13.3 billion at all levels of government. Low (i) Performance on the Housing Goals. in 2002. Multifamily units accounted for incomes, combined with high housing The year 2001 was the first year under 8.4 percent of all dwelling units (both costs, define the difficult situation of the higher levels of the Housing Goals owner and rental) financed by Freddie millions of renter households. Housing established in the Housing Goals 2000 Mac between 1999 and 2002. cost reductions are constrained by high final rule. Both GSEs met all three Concerns regarding multifamily land prices and construction costs in Housing Goals in 2001 and 2002. Their capabilities no longer constrain Freddie many markets. Regulatory barriers at the performance is discussed further in a Mac’s performance with regard to the state and local level have an enormous later section of this preamble. Housing Goals. Although Fannie Mae impact on the development of affordable (ii) The GSEs’ Efforts in the Home never withdrew from the multifamily rental housing. Government action— Purchase Mortgage Market. The market, it has stepped up its activities through land use regulation, building Appendices include a comprehensive in this area substantially, with codes, and occupancy standards—is a analysis of each GSE’s performance in multifamily purchases rising from $3.0 major contributor to high housing costs. funding home purchase mortgages for billion in 1992 to $9.4 billion in 1999, Since the early 1990s, the multifamily borrowers and neighborhoods targeted and $18.7 billion in 2001, and then mortgage market has become more by the three Housing Goals—special declining slightly to $18.3 billion in closely interconnected with global affordable and low- and moderate- 2002. Multifamily units accounted for capital markets, although not to the income borrowers and underserved 9.2 percent of all dwelling units (both same degree as the single-family areas. The GSEs’ role in the first-time owner and rental) financed by Fannie mortgage market. Loans on multifamily homebuyer market is also analyzed. Mae between 1999 and 2002. properties are still viewed as riskier by Because homeownership opportunities The increased role of Fannie Mae and some than mortgages on single-family are integrally tied to the ready Freddie Mac in the multifamily market properties. Property values, vacancy availability of affordable home purchase

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00009 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24236 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

loans, the main findings from that performance during 2001 and 2002, the of all loans originated for African- analysis are provided below: first two years under HUD’s higher American and Hispanic first-time • Both Fannie Mae and Freddie Mac Housing Goal targets. Evaluating their homebuyers, or one-third of their share have increased their purchases of activity relative to the market depends, (42 percent) of all home purchase loans affordable loans since the Housing Goals to some extent, on the way in which originated during that period. were put into effect, as indicated by the GSE activity is measured. Under the Considering conventional conforming increasing share of their business going purchase-year approach for measuring originations during the same time to the three Goals-qualifying categories. GSE activity (in which characteristics of period, it is estimated that the GSEs Between 1992 and 2002, the special mortgages purchased by a GSE in a purchased only 31 percent of loans for affordable share of Fannie Mae’s particular year, including mortgages African-American and Hispanic first- purchases of home purchase loans in originated in prior years, are compared time homebuyers, or about one-half of metropolitan areas more than doubled, with characteristics of mortgages their share (57 percent) of all home rising from 6.3 percent to 16.3 percent, originated just within the year), Fannie purchase loans in that market. A large while the underserved areas share Mae’s average performance during 2001 percentage of the lower-income loans increased more modestly, from 18.3 and 2002 matched the market in the purchased by the GSEs had relatively percent to 26.7 percent. The figures for low- and moderate-income category and low loan-to-value ratios and Freddie Mac are similar. The special approached the market in the special consequently high down payments, affordable share of Freddie Mac’s affordable and underserved areas which may explain the GSEs’ limited business rose from 6.5 percent to 15.8 categories. For example, during 2001 role in the first-time homebuyer market. percent, while the underserved areas and 2002, loans for special affordable share increased more modestly, from borrowers accounted for 15.6 percent of d. Size of the Mortgage Market That 18.6 percent to 25.8 percent. Fannie Mae’s purchases, compared with Qualifies for the Housing Goals • While both GSEs improved their 16.0 percent of market originations. As The Department estimates the size of performance, they have lagged the explained in Appendix A, conclusions the conventional, conforming market for primary market in providing affordable about Fannie Mae’s recent performance loans that would qualify under each loans to low-income borrowers and relative to the market depend Housing Goal category. The market underserved neighborhoods. Freddie significantly on whether GSE activity is estimates (which reflect 2000 Census Mac’s average performance, in measured on a ‘‘purchase year’’ basis or data and geography) are as follows: particular, fell far short of market on an ‘‘origination year’’ basis (in which • 51–57 percent for the Low- and performance during the 1990s. Fannie characteristics of mortgages originated Moderate-Income Housing Goal Mae’s performance was better than in a particular year are compared with • 24–28 percent for the Special Freddie Mac’s during 1993–2002, as characteristics of mortgages that were Affordable Housing Goal • well as during 1996–2002, which covers originated in that year and purchased by 35–40 percent for the Underserved the period under HUD’s currently- a GSE in that year or a subsequent year). Areas Housing Goal (based on 2000 defined Housing Goals. For the 1996– Fannie Mae matched the market in the Census geography). 2002 period, 21.7 percent of Freddie low- and moderate-income category in These market estimates exclude the Mac’s purchases financed properties in 2002, using the more consistent B&C (subprime loans that are not A underserved neighborhoods, compared ‘‘origination year’’ approach. (See minus grade) portion of the subprime with 23.5 percent of Fannie Mae’s Appendix A for further discussion.) market. The estimates, expressed as purchases, 24.9 percent of loans • While Freddie Mac has consistently ranges, allow for economic and market originated by depository institutions improved its performance relative to the affordability conditions that are more (i.e., banks and savings associations), market, it continued to lag the market in adverse than recent conditions. The and 25.4 percent of loans originated in all three Housing Goal categories during market estimates are based on several the conventional conforming market 2001 and 2002. For example, during mortgage market databases such as (i.e., loans below the conforming loan 2001 and 2002, loans financing Home Mortgage Disclosure Act (HMDA) limit that are not government insured or properties in underserved areas and American Housing Survey data. The guaranteed). accounted for 24.1 percent of Freddie Department’s estimates of the size of the • During the more recent 1999-to- Mac’s purchases, compared with 25.9 conventional mortgage market for each 2002 period, both Fannie Mae and percent of market originations. Housing Goal are discussed in detail in Freddie Mac fell significantly below the • Appendix A to this rule compares Appendix D. market in funding special affordable the GSEs’ funding of first-time The GSEs have substantial room for loans. During that period, special homebuyers with that of primary growth in serving the affordable housing affordable loans accounted for 14.4 lenders in the conventional conforming mortgage market. The Department percent of Fannie Mae’s purchases, 14.5 market. Both Fannie Mae and Freddie estimates that the two GSEs’ mortgage percent of Freddie Mac’s purchases, and Mac lag the market in funding first-time purchases accounted for 49 percent of 16.4 percent of loans originated in the homebuyers, and by a rather wide the total (single-family and multifamily) market. Thus, the ‘‘Fannie Mae-to- margin. Between 1999 and 2002, first- conventional, conforming mortgage market’’ ratio was 0.88 (14.4/16.4), as time homebuyers accounted for 27 market between 1999 and 2002. In was the ‘‘Freddie Mac-to-market’’ ratio. percent of each GSE’s purchases of contrast, GSE purchases comprised 42 Between 1999 and 2002, underserved home loans, compared with 38 percent percent of the low- and moderate- area loans accounted for 24.0 percent of for home loans originated in the income market, 41 percent of the Fannie Mae’s purchases, 22.9 percent of conventional conforming market. underserved areas market, and a still Freddie Mac’s purchases, and 25.8 • The GSEs account for a small share smaller 35 percent of the special percent of loans originated in the of the market for important groups such affordable market. Thus, 58–65 percent market, resulting in a ‘‘Fannie Mae-to- as minority first-time homebuyers. of the Goals-qualifying markets have not market’’ ratio of 0.93 and a ‘‘Freddie Considering all mortgage originations yet been touched by the GSEs. Mac-to-market’’ ratio of 0.89. (both government and conventional) The GSEs’ presence in mortgage • Both GSEs, but particularly Fannie between 1999 and 2001, it is estimated markets for rental properties, where Mae, markedly improved their that the GSEs purchased only 14 percent much of the nation’s affordable housing

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00010 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24237

is concentrated, is below that in the period. Obviously, there is room for the relative to HUD’s market estimates for single-family-owner market. The GSEs’ GSEs to increase their presence in the 1999–2002, market projections for share of the rental market (including single-family rental and multifamily 2005–2008, and the proposed Housing both single-family and multifamily) was rental markets. Goal levels for 2005–2008. only 30 percent during the 1999-to-2002 Table 1 summarizes the Department’s BILLING CODE 4210–27–P findings regarding GSE performance

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00011 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24238 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00012 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.000 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24239

The analysis reflected in Table 1 is the high end of the range for HUD’s Goals-rich’’ owner segment of the based on 2000 Census data on area 2005–2008 market projections is the market. As shown in Figure 1, GSE median incomes and minority same as or within one percentage point mortgage purchases represented only 27 concentrations, with the metropolitan of the 1999–2002 average of the market percent of single-family rental units area boundaries specified in June 2003 levels for the Housing Goals. financed between 1999 and 2002, and by the Office of Management and Second, it is evident from this table only 30 percent of multifamily units Budget. This affects the market that the proposed initial new level for financed during that time period—both percentages for all three Housing Goals, the Special Affordable Housing Goal (22 figures are much lower than their 57 as well as the figures on area median percent) is below the low end of HUD’s percent market share for single-family incomes and minority percentage projected market range for 2005–2008 owner-occupied properties. (Figure 2 figures that will be used to measure GSE (24 percent). The proposed initial level provides unit-level detail comparing the performance on the Housing Goals of the Low- and Moderate-Income GSEs’ purchases with originations in the beginning in 2005. For example, Housing Goal (52 percent) is at the low- conventional conforming market.) expressing the Underserved Areas end of HUD’s market estimate range. Typically, about 90 percent of rental Housing Goal in terms of 2000 Census Third, the proposed initial units in single-family rental and data adds approximately 5 percentage Underserved Areas Housing goal level is multifamily properties qualify for the points to the Housing Goal and market more consistent than the current Goal Low- and Moderate-Income Housing levels, compared with analysis using level with the market range now Goal, compared with about 44 percent 1990 Census data with Metropolitan projected by HUD for the Housing Goals of owner units. Corresponding figures Statistical Areas as defined prior to using 2000 Census data. for the Special Affordable Housing Goal 2000. Fourth, the GSEs’ performance on all are approximately 60 percent of rental The GSEs’ baseline performance of the Housing Goals was significantly units and 16.4 percent of owner units. figures in Table 1 exclude the effects of below the market average for 1999– Thus, one reason that the GSEs’ the bonus points for small multifamily 2002. The higher Housing Goals are performance under the Low- and and single-family 2–4 unit owner- intended to move the GSEs closer to or Moderate-Income Housing and Special occupied properties and the Temporary within the market range for 2005 and to Affordable Housing Goals has fallen Adjustment Factor for Freddie Mac the upper end of the market range short of HUD’s market estimates is that which were applied in official scoring projection by 2008. the GSEs have had a relatively small toward the Housing Goals in 2001–2003. An analysis of the GSEs’ mortgage presence in the two rental market The Department did not extend these purchases by property type shows that segments, notwithstanding that these adjustments beyond 2003. they have had much less presence in the market segments are important sources Table 1 reveals several features of ‘‘Goals-rich’’ rental segments of the of affordable housing and important HUD’s proposed Housing Goals. First, market, as compared with the ‘‘less- components in HUD’s market estimates.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00013 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24240 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00014 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.001 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24241

In the overall conventional community groups, leadership must Similarly, HUD estimates that Fannie conforming mortgage market, rental always involve increasing the Mae and Freddie Mac accounted for units in single-family properties and in availability of financing for only 27 percent of single-family rental multifamily properties are expected to homeownership and affordable rental units financed between 1999 and 2002. represent approximately 30 percent of housing. Thus, the GSEs’ obligation to In this case, the GSEs’ presence in the the overall mortgage market, 45 percent ‘‘lead the industry’’ entails leadership in single-family rental mortgage market of the units that collateralize mortgages facilitating access to affordable credit in was less than one-half their presence in qualifying for the Low- and Moderate- the primary market for borrowers at the market for mortgages on single- Income Housing Goal, and 60 percent of different income levels, and with family owner-occupied properties. the units that collateralize mortgages different housing needs, as well as in Clearly there is room for the GSEs to qualifying for the Special Affordable underserved urban and rural areas. increase their presence in the single- Housing Goal. Yet between 1999 and Because the GSEs’ market presence family rental and multifamily rental 2002, units in such properties varies significantly by property type, the markets. As explained above, these accounted for only 17 percent of the Department examined whether the GSEs markets are an important source of low- GSEs’ overall purchases, 31 percent of have led the industry in three different and moderate-income housing since the GSEs’ purchases meeting the Low- market sectors served by the GSEs: these units qualify for the Housing and Moderate-Income Housing Goal, single-family-owner, single-family Goals in a greater proportion than do and 44 percent of the GSEs’ purchases rental (those with at least one rental unit single-family owner-occupied meeting the Special Affordable Housing and no more than four units in total), properties. Thus, Fannie Mae and Goal.6 The continuing weakness in GSE and multifamily rental. Freddie Mac can improve their purchases of mortgages on single-family The GSEs’ purchases between 1999 performance on each of the three rental and multifamily properties is a and 2002 financed almost 60 percent of Housing Goals if they increase their significant factor explaining the the approximately 35 million owner- purchases of mortgages on rental shortfall between GSE performance and occupied units financed in the properties. that of the primary mortgage market. conventional conforming market during As discussed in Section B below with e. Ability of the GSEs To Lead the that period. The GSEs’ state-of-the-art respect to the Home Purchase Subgoals, Industry technology, staff resources, share of the the GSEs should be able to lead the total conventional conforming market, market for single-family owner- An important factor in determining and financial strength strongly suggest occupied properties. The GSEs are the overall Housing Goal level is the that they have the ability to lead the already dominant players in this market ability of the GSEs to lead the industry industry in making home purchase which, unlike the rental markets, is in making mortgage credit available for credit available for low-income families their main business activity. However, Housing Goals-qualifying populations and underserved neighborhoods. From as already discussed, research studies and areas. the analysis in Appendices A–D, it is conducted by HUD and academic The legislative history of FHEFSSA clear that the GSEs are able to improve researchers conclude that the GSEs have reflects Congress’s strong concern that their performance and lead the primary not been leading this market, but have the GSEs need to do more to benefit market in financing Housing Goals- historically lagged behind the primary low- and moderate-income families and qualifying home purchase mortgages. market in financing owner-occupied residents of underserved areas that lack As discussed in Appendix A, there housing for low-income families, first- access to credit. (See, e.g., S. Rep. 102– are a wide variety of quantitative and time homebuyers, and housing in 282 at 34.) The Senate Report on qualitative indicators that demonstrate underserved areas. FHEFSSA emphasized that the GSEs that the GSEs have ample, indeed f. Need To Maintain the Sound should ‘‘lead the mortgage finance robust, financial strength to improve Financial Condition of the GSEs industry in making mortgage credit their affordable lending performance. available for low- and moderate-income For example, the combined net income Based on HUD’s economic analysis families.’’ (See S. Rep. 102–282 at 34.) of the GSEs has risen steadily over the and review by the Office of Federal Thus, FHEFSSA specifically requires last 15 years, from $677 million in 1987 Housing Enterprise Oversight, the that HUD consider the ability of the to $10.4 billion in 2002. This financial Department has concluded that the GSEs to lead the industry in establishing strength provides the GSEs with the proposed levels of the Housing Goals the level of the Housing Goals. resources to lead the industry in making will not adversely affect the sound FHEFSSA also clarified the GSEs’ mortgage financing available for families financial condition of the GSEs. Further responsibility to complement the and neighborhoods targeted by the discussion of this issue is found in the requirements of the CRA (see section Housing Goals. economic analysis that accompanies 1335(a)(3)(B) of FHEFSSA, 12 U.S.C. The GSEs have been much less active this rule. 4565(a)(3)(B)), and fair lending laws (see in providing financing for the section 1325 of FHEFSSA, 12 U.S.C. 3. Other Factors Considered by HUD in multifamily rental housing market. 4545) in order to expand access to Proposing the New Housing Goals Between 1999 and 2002, the GSEs capital to those historically underserved financed 2.2 million multifamily HUD considered a number of by the housing finance market. dwelling units, which represented additional factors in connection with its While leadership may be exhibited approximately 30 percent of the 7.0 proposal to establish the new Housing through the GSEs’ introduction of million multifamily dwelling units that Goals described in this rule. These innovative products, technology, and were financed in the conventional additional factors also were relevant to processes, and through their market during this period. Thus, the HUD’s proposal to establish the new establishment of partnerships and GSEs’ share of the multifamily mortgage Home Purchase Subgoals. The alliances with local communities and market was just slightly over one-half of Department describes these additional 6 These percentage shares are computed from their share of the market for mortgages factors in Section B of this preamble Table A.30 in Appendix A. Note that B&C loans are on single-family owner-occupied (see, ‘‘Home Purchase Subgoals’’ excluded from these data. properties. immediately below).

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00015 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24242 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

B. Home Purchase Subgoals the Housing Goals. Although Fannie property. Specifically, for each GSE the Given the need for, and the Mae’s average performance during 2001 following proposed Subgoals would Administration’s emphasis on, and 2002 matched the market in the apply. (A ‘‘home purchase mortgage’’ is increasing homeownership low- and moderate-income category, defined as a residential mortgage for the opportunities, including those for low- and approached the market in the purchase of an owner-occupied single- and moderate-income and minority special affordable and underserved family property.) borrowers, HUD is proposing also to set areas categories, the Department’s • 45 percent of home purchase Subgoals for GSE mortgage purchase analysis shows that there is ample room mortgages purchased by the GSE in activities to increase financing for both Fannie Mae and Freddie Mac to metropolitan areas must qualify under improve their performance in opportunities for low- and moderate- the Low- and Moderate-Income Housing purchasing home loans that qualify for income, underserved, and special Goal in 2005, with this share rising to these Housing Goals, particularly in affordable borrowers who are 46 percent in 2006 and 47 percent in important market segments such as the purchasing single-family homes. both 2007 and 2008; minority, first-time homebuyer market. Specifically, the Department is • As detailed in Appendix A, evidence 33 percent of home purchase proposing Subgoals for home purchase suggests that there is a significant mortgages purchased by the GSE in loans that qualify for the Housing Goals. population of potential homebuyers metropolitan areas must qualify under The purpose of the Home Purchase who are likely to respond well to the Underserved Areas Housing Goal in Subgoals is to assure that the GSEs focus increased homeownership opportunities 2005, with this share rising to 34 on financing home purchases for the produced by increased GSE purchases percent in 2006 and 35 percent in both homeowners targeted by the Housing in this area. Immigrants and minorities, 2007 and 2008; and Goals. The Department believes that the in particular, are expected to be a major • 17 percent of home purchase establishment of Home Purchase source of future homebuyers. mortgages purchased by the GSE in Subgoals will place the GSEs in an Furthermore, studies indicate the metropolitan areas must qualify under important leadership position in the existence of a large untapped pool of the Special Affordable Housing Goal in Housing Goals categories, while also potential homeowners among the rental 2005, with this share rising to 18 facilitating homeownership. The GSEs population. Indeed, the GSEs’ recent percent in 2006 and 19 percent in both have years of experience in providing experience with new outreach and 2007 and 2008. secondary market financing for single- affordable housing initiatives confirms family properties and are fully capable Counting toward the Subgoals will be in the existence of this potential. terms of numbers of mortgages, not of exerting such leadership. Thus, the Department is proposing to numbers of units. This is consistent The focus of these Subgoals on home establish Subgoals for home purchase with the basis of reporting in HMDA purchase loans meeting the Housing loans that qualify for the three Housing data, which were HUD’s point of Goals will also help address the racial Goals to encourage the GSEs to take a reference in establishing the Subgoal and income disparities in leadership position in creating levels. HMDA data are reported in terms homeownership that exist today. homeownership financing opportunities of numbers of mortgages. Although minority homeownership has within the categories that Congress grown, the homeownership rate for expressly targeted with the Housing These proposed Subgoals are shown African Americans and Hispanic Goals. in Table 2, along with information on families is still approximately 25 what the GSEs’ performance on the percentage points below that for non- 1. Proposed Home Purchase Subgoals Subgoals would have been if they had Hispanic white families. The focus of Under this proposed rule, been in effect for 1999–2002 (under the the Subgoals on home purchase will performance on the Home Purchase proposed scoring rules for 2005–08). also increase the GSEs’ support of first- Subgoals would be calculated as Table 2 also presents HUD’s estimates of time homebuyers, a market segment Housing Goal-qualifying percentages of the average shares of mortgages on where they have lagged primary lenders. the GSEs’ total purchases of mortgages owner-occupied single-family properties The Department’s analysis suggests that finance purchases of single-family, in metropolitan areas that were that the GSEs have not been leading the owner-occupied properties located in originated in 1999–2002 that would market in purchasing single-family, metropolitan areas, based on the have qualified for these Subgoals. owner-occupied loans that qualify for owner’s income and the location of the BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00016 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24243

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00017 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.002 24244 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

2. HUD’s Determinations Regarding the GSEs have shown that they have the for a downpayment, credit problems, Home Purchase Subgoal Levels capacity to operate in underserved and discrimination. Immigrants and minorities are Current law does not require that neighborhoods and to reach out to projected to account for almost two- HUD consider the statutory factors set lower-income families seeking to buy a thirds of the growth in the number of forth in FHEFSSA prior to establishing home. Both Fannie Mae and Freddie new households over the next ten years. or setting the level of Subgoals. Mac have the staff expertise and As emphasized throughout this FHEFSSA authorizes HUD to establish financial resources to make the extra preamble and the Appendices, changing Subgoals within the Low- and effort to lead the primary market in population demographics will result in Moderate-Income Housing Goal and the funding single-family-owner mortgages for low- and moderate-income, special a need for the primary and secondary Underserved Areas Housing Goal. affordable, and underserved area mortgage markets to meet nontraditional However, under current law, Subgoals mortgages. credit needs, respond to diverse housing under these two Goals are not (b) The GSEs Have Lagged the Market. preferences and overcome information enforceable. Also, FHEFFSA authorizes Even though the GSEs have the ability and other barriers that many immigrants HUD to establish Subgoals within the to lead the market, they have not done and minorities face. The GSEs must Special Affordable Housing Goal and so under the Housing Goals. As noted increase their efforts towards providing these Subgoals are enforceable. The earlier, the Department and financing for these families. Administration has proposed, as part of independent researchers have published (d) There Are Ample Opportunities GSE regulatory reform, that Congress numerous studies examining whether or for the GSEs to Improve Their authorize HUD to establish a separate not the GSEs have been leading the Performance in the Home Purchase Home Purchase Goal that would include single-family market in terms of funding Market. Home purchase loans that enforceable components. Pending the loans that qualify for the three Housing qualify for the Housing Goals are enactment of any such legislation, HUD Goals. While the GSEs have available for the GSEs to purchase, is proposing the Subgoals described in significantly improved their which means they can improve their this proposed rule under its current performance, they have lagged the performance and lead the primary statutory authority. primary market in funding Housing market in purchasing loans for lower- The following sections provide an Goals-qualifying loans since FHEFSSA income borrowers and properties in overview of HUD’s reasons for was enacted in 1992. underserved areas. Three indicators of establishing the Subgoals, which are As also noted above, the type of this have already been discussed. detailed in the Appendices. improvement needed to meet the new First, the affordable lending market (a) The GSEs Have the Ability to Lead Subgoals was demonstrated by Fannie has shown an underlying strength over the Market. The GSEs have the ability to Mae during 2001 and 2002, when its the past few years that is unlikely to lead the primary market for mortgages average performance matched the vanish (without a significant increase in on single-family owner-occupied primary market in funding low- and interest rates or a decline in the properties, which are the ‘‘bread-and- moderate-income families and economy). Since 1999, the shares of the butter’’ of their business. Both GSEs approached the market in funding home purchase market accounted for by have long experience in the home special affordable families and the three Housing Goal categories are as purchase mortgage market, and properties in underserved areas. follows: 16.4 percent for special therefore there is no issue of the degree (c) Disparities in Homeownership and affordable, 32.3 for underserved areas, to which they have penetrated the Credit Access Remain. There remain and 44.2 percent for low- and moderate- market, as there is with the single- troublesome disparities in our housing income. family rental and multifamily mortgage and mortgage markets, even after the Second, market share data reported in markets. In addition, because the ‘‘revolution in affordable lending’’ and Section G of Appendix A show that over Subgoals focus on homeownership the growth in homeownership that has half of newly-originated loans that opportunities and, thus, do not include taken place since the mid-1990s. The qualify for the Housing Goals are not refinance loans, there is no issue homeownership rate for African- purchased by the GSEs. As noted above, regarding potentially large year-to-year American and Hispanic households the situation is even more extreme for changes in refinance mortgage volumes, remains 25 percentage points below that special sub-markets, such as the which affect the magnitude of the of white households. In 2002, the minority first-time homebuyer market denominator in calculating performance mortgage denial rate for African- where the GSEs have only a minimal percentages under the Housing Goals, as American borrowers was over twice that presence. In terms of the overall experienced in the heavy refinance for white borrowers, even after mortgage market (both conventional and years of 1998 and 2001–2003. controlling for the income of the government), the GSEs funded only 24 Both GSEs have not only been borrower. percent of all first-time homebuyers and operating in the single-family owner There is growing evidence that inner 17 percent of minority first-time mortgage market for years, they have city neighborhoods are not always being homebuyers between 1999 and 2001. been the dominant players in that adequately served by mainstream Similarly, during the same period, the market, funding 57 percent of mortgages lenders. Some have concluded that a GSEs funded only 40 percent of first- on single-family owner-occupied dual mortgage market has developed in time homebuyers in the conventional residences financed between 1999 and our nation, with conventional conforming market, and only 33 percent 2002. As discussed in Section G of mainstream lenders serving mainly of minority first-time homebuyers in Appendix A, their underwriting white families living in the suburbs and that market. guidelines are industry standards and FHA and subprime lenders serving Finally, the GSEs’ purchases that can their automated mortgage systems are minority families concentrated in inner count toward the Subgoal are not widely used in the mortgage industry. city neighborhoods. In addition to the limited to new mortgages that are Through their new low-downpayment unavailability of mainstream lenders, originated in the current calendar year. products and various underwriting families living in high-minority The GSEs can purchase loans from the initiatives, and through their various neighborhoods generally face many substantial, existing stock of affordable partnership and outreach efforts, the additional hurdles, such as lack of cash loans held in lenders’ portfolios, after

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00018 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24245

these loans have seasoned and the GSEs improvement in the affordable lending whether a rural area qualifies as an have had the opportunity to observe performance of the GSEs, and underserved area, there is no longer any their payment performance. In fact, particularly Fannie Mae, further need to distinguish underserved areas based on Fannie Mae’s recent demonstrates the GSEs’ capacity to lead located in New England from experience, the purchase of seasoned the home purchase market. underserved areas in other areas of the loans appears to be one useful strategy country. For this reason, the Department 3. Counting of Mortgages for the Home for purchasing Housing Goals-qualifying is proposing to eliminate from the Purchase Subgoals loans. definition of ‘‘Underserved area’’ the The current low homeownership rate The Department is proposing to current distinct regulatory treatment for of minorities and others living in inner amend § 81.15 to add a new paragraph New England. cities suggests that there will be (i) that would clarify that the procedures considerable growth in the origination in § 81.15 generally govern the counting D. Adequacy of Borrower Income Data of CRA loans in urban areas. For banks of home purchase mortgages toward the Accurate measurement of the GSEs’ and thrifts, selling their CRA Home Purchase Subgoals in §§ 81.12, performance under the three Housing originations will free up capital to make 81.13 and 81.14. The new paragraph Goals depends on the completeness of new CRA loans. As a result, the CRA provides, however, that the numerator data on borrower income (or, in the case market segment provides an opportunity and denominator for purposes of of non-owner-occupied units, the rent) for the GSEs to expand their affordable counting performance under the and property location. As between these lending programs. As explained in Subgoals are comprised of numbers of two, property location is reported by the Appendix A, Fannie Mae and Freddie home purchase mortgages in GSEs on most of the mortgages they Mac have already started developing metropolitan areas, rather than numbers purchase—a less than one percent programs to purchase CRA-type loans of dwelling units. Paragraph (i) also incidence of missing or incomplete on a flow basis as well as after they have provides that, for purposes of geographical data between 2000 and seasoned. addressing missing data or information 2002 for each GSE. The incidence of While the GSEs can choose any for each Subgoal, the procedures in missing borrower income data has been strategy for leading the market, this § 81.15(d) shall be implemented using greater—on the order of several leadership role can likely be numbers of home purchase mortgages in percentage points each year. accomplished by building on the many metropolitan areas and not single-family One reason for the increase in missing initiatives and programs that the owner-occupied dwelling units. Finally, income data is the recent increased use enterprises have already started, the new paragraph provides that where of mortgages for which the borrower is including: (1) Their outreach to a single home purchase mortgage not required to provide income underserved markets and their finances the purchase of two or more information. For some of these partnership efforts that encourage owner-occupied units, the mortgage mortgages the borrower presents mainstream lenders to move into these shall count once toward each Subgoal information on assets but not income markets; (2) their incorporation of that applies to the GSE’s mortgage because of circumstances that make greater flexibility into their purchase purchase. assets easier to document. Other and underwriting guidelines, (3) their C. Definition of Underserved Area for mortgages are originated entirely on the development of new products for Rural Areas basis of a credit report, property borrowers with little cash for a appraisal, and cash for the downpayment and for borrowers with The rule proposes to change the downpayment. These mortgages credit blemishes or non-traditional definition of ‘‘Underserved Area’’ for typically require relatively large credit histories; (4) their targeting of purposes of determining whether a downpayments and often require a important markets where they have had ‘‘Rural Area’’ is an ‘‘Underserved Area.’’ higher interest rate than fully only a limited presence in the past, such The definition of a ‘‘Rural Area’’ that is documented mortgages. as the markets for minority first-time an ‘‘Underserved Area’’ would be a The Housing Goals 2000 Final Rule homebuyers; (5) their purchases of both census tract, Federal or State American provided that the GSEs may exclude newly-originated and seasoned CRA Indian Reservation or tribal or from the denominator owner-occupied loans; and (6) their use of automated individual trust land, or the balance of units lacking mortgagor income data underwriting technology to qualify a census tract excluding the area within which are located in low-or moderate- creditworthy borrowers that would have any Federal or State American Indian income census tracts, i.e., tracts whose been deemed not creditworthy under reservation or tribal or individual trust median income is no greater than the traditional underwriting rules. land, having: (i) A median income at or median income of the metropolitan area The experience of Fannie Mae and below 120 percent of the greater of the or, for properties located outside of Freddie Mac in the subprime market State non-metropolitan median income metropolitan areas, the larger of the indicates that they have the expertise or nationwide non-metropolitan median median incomes of the county or the and experience to develop technologies income and a minority population of 30 statewide non-metropolitan area (see 24 and new products that allow them to percent or greater, or (ii) a median CFR 81.15(d)).7 enter new markets in a prudent manner. income at or below 95 percent of the In view of the increasing use of loans Given the innovativeness of Fannie Mae greater of the State non-metropolitan made without obtaining income and Freddie Mac, other strategies will median income or nationwide non- information from the borrower, there is be available as well. In fact, a wide metropolitan income. a question whether HUD’s existing variety of quantitative and qualitative This is essentially the same definition counting rules for missing-data indicators suggest that the GSEs have that was established in HUD’s Housing the expertise, resources and financial Goals 2000 final rule, except that census 7 For rental units, the 2000 Housing Goals Final strength to improve their affordable tracts, rather than counties, are the basic Rule also established counting rules which allow lending performance enough to lead the spatial unit for determining whether an the GSEs to estimate rents or exclude units from the denominator when rent data are missing. See 24 home purchase market for special area is underserved. Because HUD’s CFR 81.15(e)(6)(i) on the rules applicable to affordable, low- and moderate-income, proposed amendment would establish multifamily units and 24 CFR 81.15(e)(6)(ii) on the and underserved areas loans. The recent uniform standards for determining rules for single-familly rental units.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00019 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24246 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

situations are adequately reliable and the GSEs’ performance under the by the GSEs, but as a practical matter create no more than a negligible Housing Goals pursuant to HUD’s most verification of data, information statistical bias in the GSEs’ Housing counting rules under 24 CFR 81.15 and and reports occurs well after their Goals performance figures relative to the 81.16. Both the counting rules and submission to the Department, which values that they would have if complete definitions are designed to ensure renders this current verification income data could be obtained, and consistency with the statute and its provision a useful but not immediately whether a more precise method for purposes of increasing the availability of effective regulatory control. Indeed, in imputing incomes could be employed. financing for homeowners targeted by the case of data and information needed In order to inform HUD’s consideration the Goals. to calculate Housing Goals performance, of this issue, HUD requests comments In light of HUD’s interest in ensuring verification occurs only after such from the public on the following that transactions are appropriately Housing Goals performance has been question: Would it be desirable for HUD counted under the law and in calculated. Likewise, the information to have a standard, econometrically- accordance with its purposes, HUD asks provided in reports ordinarily would based method for imputing the income whether the definition of ‘‘mortgage not be verified until well after the report distribution of mortgages purchased by purchase’’ in § 81.2 should be revised in is submitted. each GSE that lack income data, based the final rule. Should HUD, for example, For these reasons, the Department has on known characteristics of the loan and further define ‘‘transactions in which a concluded that, to ensure the integrity the tract? Income distribution GSE bought or otherwise acquired with of the report(s), data submission(s) and information would be needed that cash or other thing of value, a mortgage other information provided to the shows proportions of units that are in for its portfolio or for securitization’’ for Department, additional measures are the very-low-income range (below 60 purposes of ensuring appropriate necessary. Accordingly, as described percent of area median), low- but not counting of large transactions and, if so, more fully below, the Department is very-low income (60–80 percent) and how? HUD also asks what changes, if proposing to revise § 81.102 to: (1) Re- moderate income (80–100 percent), to any, to HUD’s regulations (including, codify in paragraph (a) the existing support estimating proportions of but not limited to, changes to the authority under § 81.102 which missing-data loans for both the Low- counting rules at §§ 81.15 and 81.16) are authorizes HUD to independently verify and Moderate-Income Housing Goal and warranted to ensure that the GSEs’ large the accuracy and completeness of data, the Special Affordable Housing Goal. scale transactions further the information and reports provided by the For example, the mortgage amount as a requirements and purposes of the GSEs; (2) establish in paragraph (b) percentage of average loan amounts in Housing Goals. Do commenters believe certification requirements for the the tract, or home prices in the local HUD’s current rules are sufficiently submission of the GSEs’ Annual market, might be used in the estimation specific to determine which seasoned Housing Activities Report (AHAR) and process. Depending on the type of mortgage transactions, including large- for such other report(s), data methodology that is developed, such a scale transactions, are substantially submission(s) or information for which procedure might be applied on a equivalent to mortgage purchases? If certification is requested in writing by geographical level from census tracts up commenters believe the rules are not HUD; (3) codify in paragraph (c) HUD’s to the United States as a whole. In the sufficiently specific, how should the process for handling errors, omissions latter case one national estimate would rules be changed? or discrepancies in the GSEs’ current be created for the proportion of owner- year-end data submissions (including occupied units lacking income data that F. Verification and Enforcement of GSE the AHAR); (4) clarify in paragraph (d) qualify for each Goal, for each GSE. Data Integrity—Revised § 81.102 that HUD may exercise its Housing Goal 1. Summary counting authority by adjusting Goals E. Possible Changes to GSE Counting performance for a current year by Rules The Department’s ability to monitor deducting miscredits from a previous FHEFSSA establishes housing goals effectively the GSEs’ performance under year caused by errors, omissions or for the GSEs’ purchases of mortgages for the Housing Goals, and otherwise to discrepancies in a GSE’s prior year data low- and moderate-income families, carry out its regulatory functions, submissions (including the AHAR); and special affordable housing (very-low depends in large measure upon the (5) clarify in paragraph (e) that HUD income families and low-income submission of accurate, complete and may take enforcement action against the families in low-income areas) and current data, information and reports by GSEs under section 1341 of FHEFSSA families with properties in underserved Fannie Mae and Freddie Mac. The (12 U.S.C. 4581) and section 1345 of areas (see sections 1332–1334) in order GSEs’ Charter Acts require Fannie Mae FHEFSSA (12 U.S.C. 4585), as to ensure that the GSEs increase the and Freddie Mac to submit data, implemented by subpart G (‘‘Procedures availability to these borrowers of the information and reports on Housing for Actions and Review of Actions’’) of lower cost financing available through Goals performance under subsections HUD’s regulations at 24 CFR part 81 for the GSEs. With increasing frequency, 307(e) and (f) of the Freddie Mac the submission of non-current, the GSEs have entered into large-scale Charter Act and subsections 309(m) and inaccurate or incomplete information or transactions with lenders involving (n) of the Fannie Mae Charter Act. data. seasoned mortgages to achieve the FHEFSSA also requires the GSEs to housing goals. It is possible that some of submit reports (see section 1327 of 2. Background these transactions may include broad FHEFSSA, 12 U.S.C. 4547), and other Under section 1336 of FHEFSSA (12 buyback arrangements with the seller authorities necessitate that the GSEs U.S.C. 4566), HUD is required to for the transaction. submit information for HUD’s review monitor and enforce compliance with HUD’s rules at 24 CFR 81.2 define a (see, for example, section 1325 of the Housing Goals. The GSEs each ‘‘mortgage purchase’’ to mean a FHEFSSA, 12 U.S.C. 4545). submit quarterly information and semi- transaction in which a GSE bought or HUD’s current GSE regulations at 24 annual loan-level data on their mortgage otherwise acquired with cash or other CFR 81.102 make clear that HUD may purchases pursuant to their Charters thing of value a mortgage for its verify the accuracy and completeness of and the requirements of 24 CFR part 81. portfolio or securitization. HUD counts data, information and reports submitted To fulfill its monitoring responsibility,

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00020 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24247

HUD conducts two types of verification As the Department noted in the and other information submitted to the procedures for this data and preamble to its Housing Goals 1995 final SEC. Moreover, the recently enacted information. rule, the authority to verify information Sarbanes-Oxley Act of 2002 (P.L. 107– The first procedure is a recalculation is derived in part from section 1321 of 204, approved July 30, 2002) requires process whereby HUD, using the loan- FHEFSSA (12 U.S.C. 4541), which certification as a means of ensuring level data provided by the GSEs, accords the Secretary ‘‘general corporate accuracy in, and reconstructs each GSE’s Housing Goals regulatory power over each enterprise.’’ accountability for, the financial performance for the reporting period by The Secretary’s general regulatory information provided by a corporation applying current counting rules and power is in addition to the enumerated to its regulators and to the public (see Housing Goal eligibility criteria to the powers conferred on the Secretary by 15 U.S.C. 7241). data provided. These recalculations are FHEFSSA and the GSEs’ Charter Acts. The Department’s proposal requiring conducted immediately upon receipt of The Department also regards the GSEs to submit a certification in the GSEs’ loan-level data. If adjustments verification authority as necessary and connection with their AHARs and such in performance data are necessary incidental to its authority under section other report(s), data submission(s) or because a GSE has improperly applied 1336 of FHEFSSA to monitor and information for which certification is counting rules, or HUD discovers some enforce compliance with the Housing requested in writing by the Department, other error during the recalculation Goals. is reasonably related to the process, the Department makes these Accordingly, the rule would retain in Department’s performance of its adjustments at the time recalculation paragraph (a) of § 81.102 its existing statutory duties under FHEFSSA and is work is done and calculates the GSE’s regulatory authority to independently well supported by both statutory and official Housing Goals performance verify the accuracy and completeness of regulatory authority. Specifically, as stated, section 1321 of based on the adjustment. HUD data, information and reports submitted FHEFSSA grants the Secretary ‘‘general publishes the GSEs’ official Housing by a GSE. regulatory power’’ over the GSEs and Goal performance figures for the year on 4. Certification—§ 81.102(b) directs the Secretary to ‘‘make such its Web site, usually within six months The Department is proposing in this rules and regulations as shall be of the end of the reporting year, and rule to require the GSEs to provide a necessary and proper’’ to carry out the includes these figures in other certification in connection with their purposes of FHEFSSA and the GSEs’ published HUD management and AHARs submitted under sections 309 Charter Acts. The Supreme Court has performance reports. (m) and (n) of the Fannie Mae Charter repeatedly held that a grant to an agency The second type of verification Act or section 307(e) and (f) of the of ‘‘general regulatory authority’’ procedure consists of performance Freddie Mac Charter Act, as applicable, extends to the agency those reviews, including audit procedures, that, among other things, the AHAR is unenumerated powers that are which occur after the reporting year is current, complete and does not contain ‘‘reasonably related to the purposes of closed and Housing Goal results have any untrue statement of a material fact the enabling legislation.’’ (See Mourning been announced. Performance reviews as detailed below. The rule would also v. Family Publications Service, Inc., 411 evaluate the GSEs’ internal controls and make clear that the Department could U.S. 356, 369 (1973) (quoting Thorpe v. related business practices relative to the require such certification for such other Housing Authority of City of Durham, accuracy, completeness, and report(s), data submission(s) or 393 U.S. 268, 280–281 (1969).) This appropriateness of the information and information for which certification is standard has been accepted by every data that were provided to HUD and requested in writing by HUD. Federal Court of Appeals. (See, e.g., upon which Housing Goals performance Because of the post facto nature of Action on Smoking and Health v. CAB, was based. These reviews also include performance reviews, such reviews 699 F.2d 1209, 1212 (D.C. Cir. 1983).) sampling tests of source documents and cannot be the sole means of preventing Moreover, under section 1336 of data testing to determine the accuracy of the submission of incorrect data. HUD FHEFSSA, the Secretary is expressly reported data and to review the believes that certification requirements mandated by Congress to ‘‘monitor and transactions a GSE relied upon to better serve the end of assuring the enforce [the GSEs’] compliance with the develop the data. Due to the timing of integrity of data, information and housing goals established under * * * these reviews, which can begin no report(s) (including the AHAR) [FHEFSSA]’’ and the GSEs’ Charter Acts earlier than the close of a reporting year, submitted at the outset and such require the GSEs to submit a report to and the extensive sampling work requirements are consistent with current designated Congressional committees involved, it may take up to 24 months practice. and to the Secretary ‘‘on [their] from the date of the report under review Pursuant to its regulatory authority, activities under subpart B of * * * for HUD to develop its findings on a HUD has in the past, with regard to [FHEFSSA].’’ (See section 309(n) of the reporting year. certain specific matters, required that Fannie Mae Charter Act, 12 U.S.C. Fannie Mae and Freddie Mac certify the 3. Independent Verification Authority— 1723a(n); section 307(f) of the Freddie accuracy, currency and completeness of § 81.102(a) Mac Charter Act, 12 U.S.C.1456(f).) information and data submitted to the Also, section 309(n)(2)(L) of the Fannie As indicated, the Department is first Department. Other financial regulators, Mae Charter Act and section 307(f)(2)(L) proposing to recodify existing § 81.102 such as the Office of Federal Housing of the Freddie Mac Charter Act as paragraph (a) in the revised § 81.102. Enterprise Oversight (OFHEO), the expressly grant the Secretary the Paragraph (a) would retain HUD’s Securities and Exchange Commission discretion to require the GSEs to submit current regulatory authority to (SEC), and the Federal Deposit in their AHARs ‘‘any other information independently verify the accuracy and Insurance Corporation (FDIC) require that the Secretary considers completeness of data, information and similar certifications to ensure the appropriate’’ with respect to their reports submitted by a GSE, thereby accuracy of information submitted to activities under subpart B of FHEFSSA. retaining the Department’s authority to them. Similarly as the GSEs register (Emphasis added.) conduct on-site verifications, and to their stock with the SEC, they will be The Secretary also is accorded by carry out performance reviews. required to certify financial statements statute a number of fact finding

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00021 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24248 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

functions. These include the authority report(s), data submission(s) or GSE’s compliance with the Housing to require reports (see section 1327 of information may differ from HUD’s Goals. In addition, this provision is FHEFSSA), to gather data from the GSEs written articulations of its counting predicated upon the Department’s on their mortgage purchases (see rules including, but not limited to, the existing regulatory authority under 24 sections 309(m) and (n) of the Fannie regulations under 24 CFR part 81, and CFR 81.102 to independently verify the Mae Charter Act and sections 307(e) and any other areas of ambiguity. accuracy and completeness of data, (f) of the Freddie Mac Charter Act), to information and reports submitted by a 5. Adjustment To Correct Current Year- monitor and enforce compliance with GSE. End Errors, Omissions or the housing goals (see section 1336 of Discrepancies—§ 81.102(c) 6. Adjustment To Correct Prior Year FHEFSSA), and to issue subpoenas (see Reporting Errors—§ 81.102(d) section 1348 of FHEFSSA). These The Department is proposing to add a functions in turn permit the Secretary to new paragraph (c) to § 81.102 that The Department is proposing to add a make factual determinations, such as: would largely codify its administrative new paragraph (d) to § 81.102 that (1) Whether a GSE is complying with practice regarding errors, omissions or would provide for effective regulatory the Housing Goals; (2) whether a GSE discrepancies it discovers relative to oversight and enforcement when it has made a good-faith effort to comply HUD’s regulations and/or other determines that a GSE has, in a prior with a housing plan; and (3) whether a guidance concerning how current year year, improperly calculated its GSE has submitted the mortgage data are reported by a GSE and provide performance under one or more Housing information and reports required under the GSEs with a mechanism upon which Goals and/or Subgoals as a result of sections 309(m) and (n) of the Fannie to comment. errors, omissions or discrepancies in its Mae Charter Act, sections 307(e) and (f) Under this paragraph, the Department data submissions (including its AHAR). of the Freddie Mac Charter Act and is proposing to notify the GSE initially As background for this proposal, section 1327 of FHEFSSA. The by telephone or e-mail transmission of notably unlike financial reporting where Secretary also is charged with the errors, omissions or discrepancies in results are cumulative from year to year authority to initiate enforcement actions current year-end data reporting relative and the results of adjustments in prior upon determining that the law has been to HUD’s regulations and other years carry forward to the current year, violated. guidance. The GSE has five business the GSEs’ Housing Goal performance Since all of these functions days to respond to such notification. If reports (the Annual Housing Activity necessitate the submission of current, each error, omission or discrepancy is Reports) impact only the current complete and accurate information, data not resolved to the Department’s reporting year. This means that, unlike and reports, a certification requirement satisfaction, HUD will then notify the financial reporting, if corrections are not is necessary to carrying out these GSE in writing and seek clarification or made prior to release of HUD’s official functions. additional information to correct the performance data for the reporting year, For these reasons, the Department is error, omission or discrepancy. The GSE any subsequent corrections to that data proposing to amend § 81.102 by adding will have 10 business days from the date for that year are likely to go unnoticed a new paragraph (b) that requires the of HUD’s written notice to respond in by the public and policy makers. GSE senior officer responsible for writing to the request (or such longer In addition, if a correction is such that submitting to HUD the AHAR and such time as HUD may establish, not to it would have caused failure under a other report(s), data submission(s) or exceed 30 business days). If the GSE Housing Goal that was previously information for which a certification is fails to submit a written response to reported as having been achieved, requested in writing by HUD (referred to HUD within the 10-day (or longer) time HUD’s enforcement remedies under in the rule as the ‘‘GSE Certifying period, or if HUD determines that the section 1336 of FHEFSSA would have Official’’) to submit a certification in GSE’s written response fails to explain little relevance as they only require a connection with such documents. or correct the error, omission or GSE to submit a housing plan to ensure The rule would require that the GSE discrepancy in its current year-end compliance with the Housing Goals in certification provide: (1) The GSE reported data submissions (including the current or subsequent calendar year. Certifying Official has reviewed the the AHAR) to HUD’s satisfaction, the For these reasons, it is not practical to particular AHAR, other report(s), data Department will determine the correct overstatements in performance submission(s) or information; (2) to the appropriate adjustments to the data that were reported in previous best of the GSE Certifying Official’s numerator and the denominator to years by adjusting performance for a knowledge and belief, the particular calculate performance under the prior year. On the other hand, AHAR, other report(s), data applicable Housing Goal(s) and/or adjustments to current year performance submission(s) or information are Subgoal(s). The Department’s are an effective means of assuring current, complete and do not contain determination may involve excluding accuracy in counting under the Housing any untrue statement of a material fact; the unit(s) or mortgage(s) from the Goals in a manner that makes the public (3) to the best of the GSE Certifying numerator and including them in the aware of the adjustment. Accordingly, Official’s knowledge and belief, the denominator of the applicable Housing the Department is proposing to add a AHAR or other report(s), data Goal(s) and/or Subgoal(s). The new paragraph (d) to § 81.102 that submission(s) and information fairly Department may also pursue additional would enable it to reduce a GSE’s present in all material respects the enforcement actions against the GSE current year credit toward its Housing GSE’s performance, as required to be under § 81.102(e), if it determines that Goals performance based on errors, reported by section 309(m) or (n) of the such action is warranted. omissions or discrepancies that the Fannie Mae Act, section 307(e) or (f) of The Department’s legal authority to Department discovers in a GSE’s prior the Freddie Mac Charter Act, or other implement this provision also is based year’s data submissions (including its applicable legal authority; and (4) to the upon its general regulatory power over AHAR). best of the GSE Certifying Official’s each enterprise pursuant to section 1321 This procedure, to be known as an knowledge and belief, the GSE has of FHEFSSA and its explicit statutory ‘‘adjustment to correct prior year identified in writing any areas in which authority under section 1336 of reporting errors, omissions or the GSE’s particular AHAR, other FHEFSSA to monitor and enforce the discrepancies,’’ would provide the

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00022 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24249

Department with a mechanism for The Department proposes that this and/or to levy civil money penalties in ensuring the continued accuracy, provision will become effective upon connection with a GSE’s failure to completeness and currency of each publication of the final rule for comply with its statutory obligations GSE’s performance results. The reporting periods occurring on or after under its Charter Act and FHEFSSA. Department anticipates that the the rule’s effective date. It will not be III. Discussion of Proposed Regulatory procedure would be used infrequently. retroactive to reporting periods that Changes Even so, given the increasing preceded publication of the final rule. complexity of each GSE’s business as Should any adjustment cause a failure A. Subpart A—General well as the complexity of many of the under a Housing Goal in the current Section 81.2—Definitions transactions that the GSEs use to meet year, then current year Housing Goals their Housing Goals, the Department performance would be subject to The proposed regulation would believes that the proposed procedure is enforcement under sections 1336, 1341, change several current definitions in both reasonable and necessary. Should and 1345 of FHEFSSA, and subpart G of § 81.2, and add a new definition to this its use become necessary, the proposed part 81. section. First, to conform HUD’s procedure will provide a means for As noted, section 1321 of FHEFSSA regulations to changes in data collection HUD to effect corrections in a manner grants the Secretary ‘‘general regulatory practices made by the Office of that is appropriate and obvious to those power over each enterprise’’ which Management and Budget (OMB), HUD’s who track the GSEs’ performance includes the authority to ‘‘make such proposed regulation would change the annually, and it will help to ensure that rules and regulations as shall be current definitions of ‘‘Metropolitan the GSEs continue to exercise necessary and proper to ensure that area’’ and ‘‘Minority.’’ Second, the appropriate diligence in their Housing [Part 2, Subtitle A, of FHEFSSA] and the proposed regulation would modify the Goals reporting. purposes of [the GSEs’ Charter Acts] are current definition of ‘‘Underserved The Department’s proposed procedure accomplished.’’ The Secretary’s general area.’’ Finally, the proposed regulation would provide that the Department may regulatory power under section 1321 is would add a new definition for ‘‘Home adjust a GSE’s current year Housing in addition to the specific enumerated Purchase Mortgage’’ consistent with this Goal performance to correct for any powers conferred on the Secretary by proposal. overstatement in Housing Goals FHEFSSA and the GSEs’ Charter Acts. ‘‘Metropolitan area’’—The proposed reporting discovered in the course of Moreover, also as noted, section 1336 regulation would change the current performance reviews or otherwise of of FHEFSSA—under which the definition of ‘‘metropolitan area’’ to any previous year’s Annual Housing Secretary is mandated by Congress to remove the term ‘‘primary metropolitan Activity Report that were the result of ‘‘monitor and enforce compliance with statistical area (‘‘PMSA’’)’’ since this is errors, omissions or discrepancies. the housing goals established under a term that is no longer used by the Should the Department determine that sections 1332, 1333, and 1334, as Office of Management and Budget an adjustment to current year data for a provided in this section * * *’’— (OMB) in defining ‘‘metropolitan area.’’ prior year error, omission or expressly authorizes HUD to establish See Office of Management and Budget, discrepancy in Housing Goal reporting guidelines to measure the extent of Standards for Defining Metropolitan is warranted, the Department would compliance with the Housing Goals. and Micropolitan Statistical Areas, 65 communicate its initial findings and Section 1336 further authorizes HUD to FR 82228–82238 (December 27, 2000). determinations in writing to the GSE ‘‘assign full credit, partial credit, or no ‘‘Minority’’—The proposed regulation within 24 months of the end of the credit toward achievement of the would also change the definition of the relevant reporting year. The GSE would Housing Goals to different categories of term ‘‘minority’’ in light of significant have 30 days from the date of HUD’s mortgage purchase activities of the changes in reporting conventions for initial letter to respond in writing, with enterprises, based on such criteria as race and ethnicity, in accordance with supporting documentation, to contest the Secretary deems appropriate.’’ OMB guidance. the determination. Within 60 days of the (Emphasis added.) Currently, ‘‘minority’’ is defined in date of the GSE’s written response, the The Department’s proposal to grant HUD regulations as ‘‘any individual Department would issue a final only partial credit to a GSE in its current who is included within any one’’ of the determination letter to the GSE (unless year performance report to correct for a following list of racial and ethnic HUD determines that good cause exists prior year’s error constitutes an categories (emphasis added). The to extend this period for an additional appropriate counting criterion to assure proposed regulation would change the 30 days.) the accuracy of data used to assess GSE definition of minority to ‘‘any If the GSE fails to submit a written performance under the Housing Goals. individual who is included within any response to HUD within the 30-day one or more’’ of the following list of 7. Additional Enforcement Provisions— period, or if the Department otherwise racial and ethnic categories (emphasis determines that an adjustment is § 81.102(e) added). This change is consistent with warranted, the GSE would be required Finally, the rule would make clear a decision made by OMB in 1997, to reflect an adjustment in its Annual that a GSE’s submission of data, revising federal data classification Housing Activity Report for the current information, or reports required by standards on race and ethnicity, to year, as directed by HUD. The section 307(e) or (f) of the Freddie Mac allow individuals, in federal data adjustment would be reflected in the Charter Act, section 309(m) or (n) of the collection, to identify themselves in GSE’s year-end performance under the Fannie Mae Charter Act or subpart E of more than one category. See Office of applicable Housing Goal(s) or Subgoal(s) part 81 that are incomplete, not current, Management and Budget, Revisions to for the current reporting year by or contain an untrue statement of the Standards for the Classification of deducting the number of units or material fact shall be regarded by the Federal Data on Race and Ethnicity, 62 mortgages that HUD has determined Department as equivalent to failing to FR 58781–58790 (October 30, 1997). were erroneously counted in a previous submit such data, information or Also, consistent with OMB year from the numerator (but not the reports. For such a non-submission, the determinations, the proposed regulation denominator) for the relevant Housing Department may bring under subpart G would change the current definition of Goal or Subgoal. of part 81 an order to cease and desist ‘‘minority’’ so that: (1) ‘‘American

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00023 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24250 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Indian’’ would be defined to include • A Low- and Moderate-Income 26–30 percent for 1997–2002, and 26–27 persons with origins in any of the Housing Goal, which focuses on percent for 2001–2002. The underserved original peoples of South and Central mortgages on housing for families with areas market defined in terms of 1990 America; (2) ‘‘Asian or Pacific Islander’’ incomes no greater than area median Census data and pre-2003 metropolitan would be divided into separate income (as defined by HUD),8 and area boundaries accounted for 31–35 categories—’’Asian,’’ which would which is set at 50 percent of total units percent for 1997–2002 and 32–33 include examples of countries of origin, financed by each of the GSEs’ mortgage percent for 2001–2002. With 2000 and ‘‘Pacific Islander’’ which would be purchases; Census data and the metropolitan area • An Underserved Areas Housing included in a new definition with boundaries established in June, 2003, Goal, which focuses on mortgages on ‘‘Native Hawaiian’’ (which would these figures become 37–40 percent for include ‘‘peoples having origins in any properties located in ‘‘underserved areas,’’ defined as low-income and/or 1999–2002 and 37–39 percent for 2001– of the original peoples of Hawaii, Guam, 2002. Samoa, or other Pacific Islands;’’ (3) high-minority census tracts and rural ‘‘African-American’’ would be changed counties (excluding high-income, high- In accordance with FHEFSSA, HUD to ‘‘Black or African American;’’ and (4) minority tracts), and which is set at 31 has re-estimated the market shares of ‘‘Hispanic’’ would be changed to percent of total units financed by each the mortgages in the primary ‘‘Hispanic or Latino.’’ of the GSEs’ mortgage purchases in conventional market that would qualify ‘‘Underserved area’’—As discussed 2001–2004; for each of the GSEs’ Housing Goals for • more fully above (see section II.C), the A Special Affordable Housing Goal, the years 2005 through 2008.9 HUD proposed regulation would change the which focuses on mortgages on housing estimates that for the years 2005 through definition of ‘‘Underserved area’’ for for very low-income families and low- 2008 the low- and moderate-income purposes of determining whether a income families living in low-income share of the conventional market will be areas, and which is set at 20 percent of ‘‘Rural area’’ is an underserved area. 51–57 percent, the underserved areas total units financed by each of the GSEs’ ‘‘Home Purchase Mortgage’’— share of the market will be 35–40 Consistent with the proposed mortgage purchases in 2001–2004; and • A Special Affordable Multifamily percent, and the special affordable share establishment of Home Purchase will be 24–28 percent. Appendix D, Subgoals, the proposed regulation Subgoal, which focuses on mortgages on housing for very low-income families ‘‘Estimating the Size of the would add a definition for ‘‘Home Conventional Conforming Market for Purchase Mortgage,’’ which would be and low-income families living in low- Each Housing Goal,’’ provides an defined to mean a residential mortgage income areas, in multifamily properties extensive analysis of the Department’s for the purchase of an owner-occupied (defined as properties with five or more market share estimates. single-family property. units), and which is set at a fixed amount of 1.0 percent of the average The gaps between the current Goal B. Subpart B—Housing Goals total dollar volume of mortgages levels and HUD’s latest market estimates 1. Background purchased by each GSE in the years indicate that the Goals should be higher 1997, 1998, and 1999. This formula and that there are ample opportunities The Department is required to results in a Subgoal of special affordable available for the GSEs to meet the new establish, by regulation, annual Housing multifamily mortgage purchases totaling Goals for each GSE. The Goals include $2.85 billion per year for Fannie Mae initial Goals in 2005 as they institute a Low- and Moderate-Income Housing and $2.11 billion per year for Freddie measures to ensure that they will attain Goal, a Special Affordable Housing Mac for each calendar year from 2001 the increased goal levels in 2006–2008. Goal, and a Central Cities, Rural Areas, through 2004. Moreover, HUD’s new market estimates and Other Underserved Areas Housing These Housing Goals, excluding the allow for more adverse economic and Goal (the Underserved Areas Housing Special Affordable Multifamily Subgoal, affordability conditions than recently Goal). Section 1331(a) of FHEFSSA share common characteristics: (1) The experienced. For example, the lower requires HUD to establish these Goals in Goal levels are the same for both GSEs; end—51 percent—of the range for the a manner consistent with sections (2) they are percentage based Goals low- and moderate-income market 301(3) of the Fannie Mae Charter Act defined in terms of percentages of estimate is consistent with low- and and 301(b)(3) of the Freddie Mac housing units financed; and (3) one unit moderate-income borrowers accounting Charter Act, which require the GSEs ‘‘to may qualify for one or more Goals. In for 38 percent of home purchase loans provide ongoing assistance to the addition, under the current regulation, in the single-family owner-occupied secondary market for residential Goals were established based on market. (The remainder of the low- and mortgages (including * * * mortgages consideration of the statutory factors moderate-income market share estimate on housing for low- and moderate- and set for a three-year period from includes multifamily and single-family income families involving a reasonable 2001 through 2003 to allow the GSEs rental properties.) Since the 1995–2002 economic return that may be less than time to develop long-range strategies. average for the low- and moderate- the return earned on other activities).’’ A key factor in determining the level income share of the home purchase Under section 1331(c) of FHEFSSA, of the Goals was and is the estimated market was 43.5 percent, and the more HUD may, by regulation, adjust any size of the conventional market for each recent 1999–2002 average was 44.6 Housing Goal from year to year. Goal. This determination is discussed In October 2000, HUD established above and in Appendix D. HUD percent, the initial Goals for 2005 allow Housing Goals for the GSEs for 2001– estimates that the low- and moderate- leeway for more adverse income and 2003, revising and restructuring the income market accounted for 54–59 interest rate conditions. Goals that had been in effect for 1996– percent of all mortgages originated 2000. The current Housing Goal levels, during the 1997 to 2002 period, and for 9 The Goal-qualifying market shares are estimated which were in place for 2001–2003 and 54–55 percent in 2001 and 2002. The for the years 2005–2008 under several projections about the relative sizes of the single-family and extended through 2004 without the special affordable market accounted for multifamily markets. Numerous sensitivity analyses bonus points and Temporary that consider alternative market and economic Adjustment Factor, are: 8 24 CFR 81.2. conditions are examined in Appendix D.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00024 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24251

2. Low- and Moderate-Income Housing for housing affordable to low- and range, rather than a specific point Goal, § 81.12 moderate-income families for 2005, estimate, to account for the projected rising to 46 percent in 2006, 47 percent effects of different economic and This section discusses the in 2007, and 47 percent in 2008. affordability conditions that can Department’s consideration of the A short discussion of the statutory reasonably be anticipated. HUD statutory factors in arriving at the new factors reviewed to establish the Goal estimates that low- and moderate- Housing Goal level for the Low- and follows. More detailed information income share of the market averaged 57 Moderate-Income Housing Goal, which analyzing each of the statutory factors is percent between 1999 and 2002. targets mortgages on housing for provided in Appendix A, ‘‘Departmental families with incomes at or below the Considerations to Establish the Low- b. Past Performance of the GSEs under area median income. After analyzing the and Moderate-Income Housing Goal,’’ the Low- and Moderate-Income Housing statutory factors, this proposed rule and Appendix D, ‘‘Estimating the Size Goal would establish (a) a Goal of 52 percent of the Conventional Conforming Market for the percentage of the total number of for each Housing Goal.’’ As discussed above, a number of dwelling units financed by each GSE’s changes in Goal-counting procedures mortgage purchases for housing a. Market Estimate for the Low- and were adopted as part of HUD’s Housing affordable to low- and moderate-income Moderate-Income Housing Goal Goals 2000 final rule. Thus, it is families for 2005, rising to 53 percent in The Department estimates that necessary to provide information using 2006, 55 percent in 2007, and 57 dwelling units serving low- and several different measures in order to percent in 2008, and (b) a Subgoal of 45 moderate-income families will account track performance on the Low- and percent of the total number of owner- for 51–57 percent of total units financed Moderate-Income Housing Goal over the occupied dwelling units financed by in the overall conventional conforming 1996–2002 period. Table 3 shows each GSE’s purchases of home purchase mortgage market during the period 2005 performance under these measures. mortgages in metropolitan areas that are through 2008. HUD has developed this BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00025 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24252 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00026 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.003 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24253

Specifically, the following changes end of 2003) are not applied. These eligibility for Goal credit of certain were made in counting procedures for figures are termed the ‘‘2001–03 qualifying government-backed loans. measuring performance on the Low- and baseline assumptions.’’ For 1996–2000 That is, the Department does not plan to Moderate-Income Housing Goal for these figures differ from the official change the 2001–03 baseline 2001–03. HUD: performance figures because they assumptions for scoring loans under the (a) Established ‘‘Bonus points’’ incorporate the revised counting Low- and Moderate-Income Housing (awarding double credit) for purchases procedures described under point (c), Goal. of low- and moderate-income mortgages above, which were not reflected in the Beneath the 2001–03 baseline figures on small (5–50 unit) multifamily official performance figures at that time. in Table 3 is another row of figures properties and, above a threshold level, For 2001 and 2002 both sets of figures designated ‘‘With 2005 Assumptions.’’ mortgages on 2–4 unit owner-occupied incorporate the revised counting These figures show the effects of properties; procedures, but the baseline does not applying 2000 Census data and the new (b) Established a ‘‘temporary incorporate the bonus points and the specification of Metropolitan Statistical adjustment factor’’ (1.35 units credit, as Freddie Mac Temporary Adjustment Areas released by the Office of revised by Congress for 2001–03 from Factor. Management and Budget in 2003 to the HUD’s 1.2 unit credits in the 2000 rule) In terms of the 2001–2003 baseline measurement of Low- and Moderate- that applied to Freddie Mac’s purchases measure, both Fannie Mae and Freddie Income purchase percentages with the (but not Fannie Mae’s purchases) of Mac’s low- and moderate-income same counting rules that were used for low- and moderate-income mortgages on performance reached its maximum in the 2001–03 baseline. The effect is to large (more than 50-unit) multifamily 2000 (Fannie Mae at 51.3 percent and reduce the Goal-qualifying percentage properties; and Freddie Mac at 50.6 percent) before by an average of 0.5 percentage points (c) Revised procedures that HUD had declining somewhat in 2001 and 2002. for Fannie Mae and 0.8 percentage instituted regarding the treatment of Both GSEs’ baseline performance in points for Freddie Mac, over the four- missing data on unit affordability, the 2001 exceeded the level attained in year period. use of imputed or proxy rents for 1999. However, Freddie Mac’s baseline c. Proposed Low- and Moderate-Income determining Goal credit for multifamily performance fell further in 2002, to Home Purchase Subgoal for 2005–2008 mortgages, and the eligibility for Goals approximately the same level as in credit for certain qualifying government- 1999. Fannie Mae’s baseline The Department proposes to establish backed loans. performance was essentially unchanged a Subgoal of 45 percent of each GSE’s Based on the counting rules in effect in 2002. purchases of home purchase mortgages at that time for 1996–2000, as shown Overall, both GSEs’ performance on single-family owner-occupied under ‘‘official performance’’ for 1996– exceeded HUD’s Low- and Moderate- properties in metropolitan areas which 2000 in Table 3, Low- and Moderate- Income Housing Goals by significant are for low- and moderate-income Income Housing Goal performance for margins in 1996–99, and by wide families in 2005, with this Subgoal Fannie Mae was consistently in the 44– margins in 2000. New, higher Goals rising to 46 percent in 2006 and 47 46 percent range over the 1996–1999 were established for 2001–03, and percent in both 2007 and 2008. The period, before jumping to a peak of 49.5 despite somewhat lower performance purpose of this Subgoal is to encourage percent in 2000. Freddie Mac’s than the level attained in 2000, both the GSEs to increase their acquisitions performance started at a lower level, but GSEs’ official performance exceeded the of home purchase loans for low- and then increased in several steps, from new goal levels in 2001 and 2002, with moderate-income families, many of 41–43 percent in 1996–98 to 46.1 the inclusion of the bonus points and whom are expected to enter the percent in 1999, and a record level of the TAF. homeownership market over the next 49.9 percent in 2000. That was the only The decline in baseline performance few years. If the GSEs meet this Subgoal, year prior to 2001 in which Freddie in 2001 and 2002 can be attributed in in 2005 they will be leading the primary Mac’s performance has exceeded Fannie large measure to the mortgage refinance market by approximately one percentage Mae’s performance on this Goal. wave that occurred in those years. point, based on the income Based on the then current counting Fannie Mae’s overall volume of characteristics of home purchase loans rules, including the bonus points and mortgage purchases (in terms of reported in HMDA. Between 1999 and TAF, as shown under ‘‘official numbers of housing units) rose from 2.2 2002, HMDA data show that low- and performance’’ in Table 3, Low- and million in 2000 to 4.7 million in 2001, moderate-income families accounted for Moderate-Income Housing Goal and then to 6.0 million in 2002. an average of 44.3 percent of single- performance in 2001 was 51.5 percent Similarly, Freddie Mac’s volume rose family-owner loans originated in the for Fannie Mae and 53.2 percent for from 1.6 million in 2000 to 3.3 million conventional conforming market of Freddie Mac. Low- and Moderate- in 2001, and then to 4.3 million in 2002. metropolitan areas. Loans in the B&C Income Housing Goal performance in For each GSE the increase in volume portion of the subprime market are not 2002 was 51.8 percent for Fannie Mae each year can be largely attributed to included in these averages. To reach the and 51.4 percent for Freddie Mac. increases in purchase volumes for 45-percent Subgoal for 2005, both GSEs Immediately beneath the official Low- refinance mortgages relative to home must improve their average and Moderate-Income Housing Goal purchase mortgages. For each GSE, the performance, as shown in Table 2— performance percentages in Table 3 are fraction of mortgages that qualified as Fannie Mae by about one percentage figures showing the GSEs’ low- and Low- and Moderate-Income was less for point over its average performance of moderate-income purchase percentages refinance mortgages than for home 44.2 percent during 2001 and 2002, and on a consistent basis for the entire purchase mortgages. Freddie Mac by 2.4 percentage points 1996–2002 period. The assumptions For 2005–2008 HUD does not propose over its average performance of 42.6 used were the scoring rules established to change the current procedures percent; these required improvements in HUD’s Housing Goals 2000 Final regarding the treatment of missing data will increase further by one percentage Rule except that bonus points and the on unit affordability, the use of imputed point in 2006 and an additional one Freddie Mac Temporary Adjustment or proxy rents for determining Goal percentage point in 2007–08 under Factor (which were terminated at the credit for multifamily mortgages, or the HUD’s proposal.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00027 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24254 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

As explained above, HUD will be re- mortgages comprise the ‘‘bread-and- underserved by the mortgage finance benchmarking its median incomes for butter’’ of their business, the GSEs have system. The 1995 rule provided that metropolitan areas and non- historically lagged behind the primary mortgage purchases count toward the metropolitan counties based on 2000 market in financing mortgages for low- Underserved Areas Housing Goal if such Census median incomes, and will be and moderate-income families. Because purchases finance properties that are incorporating the effects of the new home purchase loans account for a located in underserved census tracts. At OMB metropolitan area definitions. major share of the GSEs’ purchases, the 24 CFR 81.2 of HUD’s current rules, HUD projected the effects of these two establishment of this Subgoal will aid HUD defines ‘‘underserved areas’’ for changes on the low- and moderate- their performance under the overall metropolitan areas (in central cities and income shares of the single-family- Low- and Moderate-Income Housing other underserved areas) as census owner market for the years 1999–2002. Goal. tracts where either: (1) the tract median These estimates will be referred to as For the foregoing reasons, the income is at or below 90 percent of the ‘‘projected data’’ while the 1990-based Department believes that the GSEs can area median income (AMI); or (2) the data reported above will be referred to do more to raise the share of their home minority population is at least 30 as ‘‘historical data.’’ The average low- loan purchases serving low- and percent and the tract median income is mod share of the home purchase market moderate-income families. This can be at or below 120 percent of AMI. The (without B&C loans) was 43.1 percent accomplished by building on efforts that AMI ratio is calculated by dividing the based on projected data, as compared the enterprises have already started, tract median income by the MSA with 44.3 percent based on historical including their new affordable lending median income. The minority percent of data. Thus, based on projected data, the products, their many partnership efforts, a tract’s population is calculated by proposed 45-percent Home Purchase their outreach to inner city dividing the tract’s minority population Subgoal for 2005 is approximately two neighborhoods, their incorporation of by its total population. percentage points above the 1999–2002 greater flexibility into their For properties in non-metropolitan market average. Fannie Mae’s average underwriting guidelines, and their (rural) areas, mortgage purchases count low-mod performance between 1999 purchases of seasoned CRA loans. A toward the Underserved Areas Housing and 2002 based on the projected data wide variety of quantitative and Goal where such purchases finance was 41.4 percent, compared with 42.5 qualitative indicators indicate that the properties that are located in percent based on historical data. To GSEs’ have the resources and financial underserved counties. These are defined reach the 45-percent Subgoal based on strength to improve their affordable as counties where either: (1) the median projected data, Fannie Mae would have lending performance enough to lead the income in the county does not exceed to improve its performance in 2005 by market serving low- and moderate- 95 percent of the greater of the median 2.3 percentage points over its projected income families. incomes for the non-metropolitan portions of the state or of the nation as average performance of 42.7 percent in d. Proposed Goal Levels for 2005–2008 2001 and 2002, or by 1.4 percentage a whole; or (2) minorities comprise at The Department is proposing to points over its projected 2002 low-mod least 30 percent of the residents and the increase the Low- and Moderate-Income performance of 43.6 percent. Freddie median income in the county does not Housing Goal to 52 percent for 2005, 53 Mac’s average low-mod performance exceed 120 percent of the greater of the percent in 2006, 55 percent in 2007, and between 1999 and 2002 based on the median incomes for the non- 57 percent in 2008. The reasons for metropolitan portions of the state or of projected data was 40.9 percent, increasing the Low- and Moderate- the nation as a whole. compared with 42.3 percent based on Income Housing Goal are discussed in This proposed rule bases its proposed historical data. To reach the 45-percent Section a, above. While the GSEs have level for the Underserved Areas Housing Subgoal based on projected data, lagged the primary market in funding Goal on 2000 Census data on area Freddie Mac would have to improve its low- and moderate-income loans, they median incomes and minority performance in 2005 by 4.0 percentage appear to have ample room to improve percentages for census tracts, counties, points over its projected average their performance in that market. The MSAs, and the non-metropolitan performance of 41.0 percent in 2001 and GSEs’ mortgage purchases between 1999 portions of states and of the entire 2002, or by 2.9 percentage points over and 2002 accounted for 49 percent of nation. HUD’s analysis, which is its projected 2002 low-mod performance the total (single-family and multifamily) sketched below and described in greater of 42.1 percent. conforming mortgage market, but they detail in Appendix B, has revealed that Section II.B.2 of this preamble and accounted for only 42 percent of the the effect of using 2000 Census data Section I of Appendix A discuss the low- and moderate-income market. A rather than 1990 data to determine reasons why the Department is wide variety of quantitative and whether areas are underserved increase establishing the Subgoal for low- and qualitative indicators demonstrate that the percentages of the GSEs’ mortgage moderate-income loans, as follows: (1) the GSEs’ have the expertise, resources purchases in underserved areas by an The GSEs’ have the resources and the and financial strength to improve their estimated average of 5 percentage points ability to lead the market in providing low- and moderate-income lending for Fannie Mae and 4 percentage points mortgage funding for low- and performance and close their gap with for Freddie Mac, based on the moderate-income families; (2) the GSEs the market. geographic locations of the GSEs’ have generally not led the market, even mortgage purchases in 1999 through though they have the ability to do so; (3) 3. Central Cities, Rural Areas, and Other 2002. This change reflects geographical troublesome disparities in our housing Underserved Areas Goal, § 81.13 shifts in population concentrations by and mortgage markets indicate a This section discusses the income and minority status from 1990 continuing need for increased GSE Department’s consideration of the to 2000. It is for this reason that HUD’s activity; and (4) there are ample statutory factors in arriving at the proposed level of the Underserved opportunities for the GSEs to improve proposed new housing goal level for the Areas Housing Goal is greater than the their low- and moderate-income Underserved Areas Housing Goal. existing level by several percentage performance in the home purchase The Underserved Areas Housing Goal points more than the increase in the market. Although single-family-owner focuses on areas of the nation currently other two Goals.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00028 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24255

After analyzing the statutory factors, A short discussion of the statutory economic and affordability conditions this proposed rule would: (a) Establish factors reviewed in establishing the Goal that can reasonably be anticipated. HUD a Goal of 38 percent for the percentage follows. Additional information estimates that the underserved areas of the total number of dwelling units analyzing each of the statutory factors is market averaged 39 percent between financed by each GSE’s mortgage provided in Appendix B, ‘‘Departmental 1999 and 2002. purchases for properties located in Considerations to Establish the Central b. Past Performance of the GSEs under underserved areas for 2005, 39 percent Cities, Rural Areas, and Other the Underserved Areas Housing Goal for 2006 and 2007, and 40 percent for Underserved Areas Goal,’’ and 2008; (b) establish census tracts as the Appendix D, ‘‘Estimating the Size of the As discussed above, a number of spatial basis for establishing whether Conventional Conforming Market for changes in goal-counting procedures properties in non-metropolitan (rural) Each Housing Goal.’’ were adopted as part of HUD’s Housing areas count toward the Underserved Goals 2000 final rule. Thus it is a. Market Estimate for the Underserved Areas Housing Goal, in place of counties necessary to provide information using Areas Housing Goal as in the definition stated above, for the several different measures in order to reasons described below; and (c) also The Department estimates that track changes in the GSEs’ performance establish a Subgoal of 33 percent of the dwelling units in underserved areas will on the Underserved Areas Housing Goal total number of dwelling units financed account for 35–40 percent of total units over the 1996–2002 period. These are by each GSE’s purchases of home financed in the overall conventional shown in Table 4. The same changes in purchase mortgages in metropolitan conforming mortgage market during the counting rules described for the Low- areas for properties located in period 2005 through 2008. HUD has and Moderate-Income Housing Goal are underserved areas of metropolitan areas developed this range, rather than a applicable to the Underserved Areas for 2005, rising to 34 percent for 2006, specific point estimate, to accommodate Housing Goal. and 35 percent for 2007 and 2008; the projected effects of different BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00029 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24256 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00030 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.004 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24257

Based on the counting rules in effect Goals were established for 2001–03, and specification of Metropolitan Statistical at that time, as shown under ‘‘official despite somewhat lower performance Areas released by the Office of performance’’ for 1996–2000 in Table 4, than the level attained in 2000 (largely Management and Budget in 2003 to the Underserved Areas Housing Goal due to the 2001–02 refinance wave), identification of underserved areas for performance for Fannie Mae generally both GSEs’ performance exceeded the purposes of measuring historical GSE fluctuated in the range between 27 and new Goal levels in 2001 and 2002. goal performance. The second of the two 29 percent over the 1996–99 period, Appendix B includes a lines also incorporates the effects of the before rising to a peak of 31.0 percent comprehensive analysis of the GSEs’ Department’s proposed change from in 2000. Freddie Mac’s performance performance in funding mortgages for counties to census tracts as the basis for started at a lower level, but then single-family-owner properties in identifying underserved areas outside of increased in several steps, from 25–26 underserved areas. (The data reported metropolitan areas beginning in 2005. percent in 1996–98 to 27.5 percent in there are based on 2000 Census HUD’s determination of underserved 1999, and a record level of 29.2 percent geography, which produces underserved areas for purposes of computing the in 2000. Freddie Mac’s performance in area figures slightly over five percentage GSEs’ performance on the Underserved 1999 was the only year prior to 2001 in points higher than 1990-based Areas Housing Goal has through 2002 which it exceeded Fannie Mae’s geography.) Between 1999 and 2002, been based on area median incomes and performance on this Goal. 28.3 percent of Freddie Mac’s purchases area minority percentages from the 1990 Based on current counting rules, and 29.5 percent of Fannie Mae’s Census. HUD applied the existing including the bonus points and the purchases financed properties in numerical thresholds for minority TAF, as shown under ‘‘official underserved neighborhoods, compared percentages and median incomes to performance’’ for 2001 in Table 4, with 31.5 percent home purchase loans 2000 Census data and ascertained that Underserved Areas Housing Goal originated in the conventional the proportion of underserved census performance in 2001 was 32.6 percent conforming market (excluding B&C tracts and the proportion of housing for Fannie Mae and 31.7 percent for loans). Thus, Freddie Mac performed at units in underserved census tracts in Freddie Mac. Underserved Areas 90 percent of the market level, while metropolitan areas increases Housing Goal performance in 2002 was Fannie Mae performed at 94 percent of significantly from 1990 levels: from 47.5 32.8 percent for Fannie Mae and 31.9 the market level—both results similar to percent to 54.9 percent of census tracts percent for Freddie Mac. those reported in Appendix B for underserved and from 44.3 percent to Immediately beneath the official underserved areas based on 1990 52.5 percent of population in Underserved Areas Housing Goal Census geography. The 2000-based underserved census tracts (including the performance percentages in Table 4 are results also show that Fannie Mae has effects of the 2003 re-specification of figures showing the GSEs’ purchase improved its performance and matched Metropolitan Statistical Areas). percentages under this Goal on a the primary market in funding Comparable shifts at the county level in consistent basis for the entire 1996– underserved areas during 2002. The non-metropolitan areas were found to be 2002 period. The assumptions used share of Fannie Mae’s purchases going of much smaller magnitude. Further, were the scoring rules established in to underserved areas increased from HUD estimated the spatial distribution HUD’s Housing Goals 2000 Final Rule, 25.7 in 1999 to 32.3 percent in 2002, of GSE mortgage purchases across except that bonus points and the which placed it at the market level of metropolitan census tracts and non- Freddie Mac Temporary Adjustment 32.3 percent. However, the 2000-based metropolitan counties for recent years. Factor (which terminated at the end of results show that, like Freddie Mac, The findings were that for 2000, 2001, 2003) are not applied. These figures are Fannie Mae’s longer-term performance and 2002, Fannie Mae’s performance termed the ‘‘2001–03 baseline’’ (since 1996) as well as its recent average figures are an estimated 7.2 percent, 6.0 assumptions. For 1996–2000 these performance (1999 to 2001) has percent, and 5.5 percent higher in terms figures differ from the official consistently been below market levels. of 2000 Census geography than with performance figures because they But, it is encouraging that Fannie Mae 1990 Census geography. The incorporate the revised counting significantly improved its performance corresponding figures for Freddie Mac procedures, which were not reflected in relative to the market during the first are 5.6 percent, 5.1 percent, and 5.1 the official performance figures at that two years of HUD’s higher Housing Goal percent larger, respectively. With a time. For 2001 and 2002 both sets of levels. further shift to tract-based definitions figures incorporate the revised counting In evaluating the GSEs’ past the figures for Fannie Mae are reduced procedures, but the baseline does not performance, it should be noted that by 0.7 percentage points in each of the incorporate the bonus points and while borrowers in underserved three years, and for Freddie Mac 0.7, Freddie Mac Temporary Adjustment metropolitan areas tend to have much 0.8, and 0.7 percentage points, Factor. lower incomes than borrowers in other respectively. HUD has taken account of In terms of the 2001–2003 baseline areas, this does not mean that GSE these shifts in establishing the level of measure, both Fannie Mae and Freddie mortgage purchases in underserved the Underserved Areas Housing Goal for Mac’s Underserved Areas Housing Goal areas must necessarily be mortgages on 2005 and beyond. performance reached its maximum in housing for lower income families. HUD originally adopted its current 2000 (Fannie Mae at 31.0 percent and Between 1999 and 2001, housing for county-based definition for targeting Freddie Mac at 29.2 percent) before above median-income households GSE purchases to underserved non- declining somewhat in 2001 and 2002. accounted for nearly 60 percent of the metropolitan areas primarily based on Both GSEs’ baseline performance in single-family owner-occupied mortgages information that rural lenders did not 2001 and 2002 exceeded the level the GSEs purchased in underserved perceive their market areas in terms of attained in 1999. areas. census tracts, but rather, in terms of Overall, both GSEs’ official Beneath the 2001–03 baseline figures counties. A further concern was an performance exceeded their in Table 4 are two additional rows of apparent lack of reliability of geocoding Underserved Areas Housing Goal by figures designated ‘‘2005 Assumptions.’’ software applied to non-metropolitan significant margins in 1996–99, and by These figures show the effects of areas. Recent research summarized in wide margins in 2000. New, higher applying 2000 Census data and the new Appendix B indicates that a tract-based

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00031 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24258 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

system would improve the extent to Section II.B.2 of this preamble and 4. Special Affordable Housing Goal, which the underserved area definition Section I of Appendix B discuss the § 81.14 distinguishes areas by key reasons why the Department is This section discusses the socioeconomic and demographic establishing a Subgoal for home Department’s consideration of the characteristics such as median family purchase mortgages in underserved statutory factors in arriving at the income, poverty, unemployment, school areas namely: (1) The GSEs’ have the proposed Housing Goal level for the dropout rates, and minority resources and the ability to lead the Special Affordable Housing Goal, which populations. Under a tract-based market in providing funding in counts mortgages on housing for very definition underserved areas stand out underserved neighborhoods; (2) the low-income families and low-income more as areas of lower income and low GSEs have not led the market, even families living in low-income areas. economic activity and as having though they have the ability to do so; (3) After analyzing the statutory factors, somewhat larger minority population troublesome disparities in our housing this proposed rule would establish: (a) proportions. A tract-based definition and mortgage markets indicate a A Goal of 22 percent for the percentage would also improve the targeting of the continuing need for increased GSE of the total number of dwelling units goal to areas with relatively greater activity; and (4) there are ample financed by each GSE’s mortgage housing needs. Based on these findings, opportunities for the GSEs to improve purchases that are for special affordable which are detailed in Appendix B, HUD their underserved area performance in housing, affordable to very low-income is proposing to re-specify the definition the home purchase market. Although families and families living in low- of underserved areas within non- single-family-owner mortgages comprise income areas for 2005, rising to 24 metropolitan (rural) areas to be based on the ‘‘bread and butter’’ of the GSEs’ percent in 2006, 26 percent in 2007, and census tracts rather than counties. business, the GSEs have lagged behind 28 percent in 2008; (b) a Subgoal of 1 the primary market in financing c. Proposed Underserved Areas Home percent of each GSE’s combined annual properties in underserved areas. For the Purchase Subgoal for 2005–2008 average mortgage purchases in 2000, foregoing reasons, the Secretary believes 2001, and 2002, for each GSE’s special The Department believes the GSEs that the GSEs can do more to raise the affordable mortgage purchases that are can play a leadership role in share of their home loan purchases in for multifamily housing in 2005–2008; underserved markets. To facilitate this underserved areas. This can be and (c) a Subgoal of 17 percent of the leadership, the Department is proposing accomplished by building on efforts that total number of each GSE’s purchases of a Subgoal of 33 percent for each GSE’s the enterprises have already started, home purchase mortgages in acquisitions of home purchase including their new affordable lending metropolitan areas that are for housing mortgages on properties located in the products, their many partnership efforts, affordable to very low income families underserved census tracts of their outreach to inner city and low-income families in low-income metropolitan areas for 2005, rising to 34 neighborhoods, their incorporation of areas for 2005, rising to 18 percent in percent in 2006 and 35 percent in 2007 greater flexibility into their 2006, 19 percent in 2007, and 19 and 2008. The purpose of this Subgoal underwriting guidelines, and their percent in 2008. is to encourage the GSEs to improve purchases of seasoned CRA loans. A A short discussion of the statutory their purchases of mortgages for wide variety of quantitative and factors for establishing the Goal follows. homeownership in underserved areas, qualitative indicators demonstrate that Additional information analyzing each thus providing additional credit and the GSEs have the resources and of the statutory factors is provided in capital for neighborhoods that financial strength to improve their Appendix C, ‘‘Departmental historically have not been adequately affordable lending performance enough Considerations to Establish the Special served by the mortgage industry. If the to lead the market in underserved areas. Affordable Housing Goal,’’ and GSEs meet this Subgoal, they will be d. Proposed Goal Levels for 2005–2008 Appendix D, ‘‘Estimating the Size of the leading the primary market, based on Conventional Conforming Market for the census tract characteristics of home The Department is proposing to Each Housing Goal.’’ purchase loans reported in HMDA. increase the Underserved Areas Housing Between 1999 and 2002, HMDA data Goal to 38 percent for 2005, 39 percent a. Market Estimate for the Special show that underserved areas accounted for 2006 and 2007, and 40 percent for Affordable Housing Goal for 32.3 percent of single-family-owner 2008. The reasons for increasing the The Department estimates that loans originated in the conventional Underserved Areas Housing Goal are dwelling units serving very low-income conforming market of metropolitan discussed in Sections I.C and II.A of this families and low-income families living areas. To reach the 33 percent Subgoal preamble. While the GSEs have lagged in low-income areas will account for for 2005, both GSEs would have to the primary market in funding loans in 24–28 percent of total units financed in improve their performance, as shown in underserved areas, they appear to have the overall conventional conforming Table 2—Fannie Mae by 1.9 percentage ample room to improve their mortgage market during the period 2005 points over its average performance of performance in that market. The GSEs’ through 2008. HUD has developed this 31.1 percent, and Freddie Mac by 3.5 mortgage purchases between 1999 and range, rather than a point estimate, to percentage points over its average 2002 accounted for 49 percent of the account for the projected effects of performance of 29.5 percent during total (single-family and multifamily) different economic conditions that can 2001 and 2002. These required conforming mortgage market, but they reasonably be anticipated. HUD also improvements would increase further accounted for only 41 percent of the estimates that the special affordable by one percentage point in 2006 and by underserved areas market. A wide market averaged 28 percent between an additional one percentage point in variety of quantitative and qualitative 1999 and 2002. 2007–08 under HUD’s proposal. The indicators demonstrate that the GSEs Subgoal applies only to the GSEs’ have the expertise, resources and b. Past Performance of the GSEs Under purchases in metropolitan areas because financial strength to improve their the Special Affordable Housing Goal the HMDA-based market benchmark is performance in underserved areas and As discussed above, a number of only available for metropolitan areas. to close their gap with the market. changes in Goal-counting procedures

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00032 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24259

were adopted as part of HUD’s Housing several different measures in order to the 1996–2002 period. These are shown Goals 2000 final rule. Thus, it is track changes in performance on the in Table 5. necessary to provide information using Special Affordable Housing Goal over BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00033 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24260 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00034 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.005 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24261

Based on the counting rules in effect level data for home purchase mortgages Fannie Mae and Freddie Mac because of at that time, as shown under ‘‘official on single-family owner-occupied the relatively high percentage of performance’’ for 1996–2000 in Table 5, properties in metropolitan areas to multifamily units meeting the Special Special Affordable Housing Goal compare the GSEs’ performance in Affordable Housing Goal. For example, performance for Fannie Mae generally special affordable lending to the between 1999 and 2002, 53 percent of fluctuated in the range between 14 and performance of depositories and other units financed by Fannie Mae’s 17 percent over the 1996–99 period, lenders in the conventional conforming multifamily mortgage purchases met the before rising to a peak of 19.2 percent market. There are two main findings Special Affordable Housing Goal, in 2000. Freddie Mac’s performance with respect to the special affordable representing 27 percent of units counted started at a lower level, but then category. First, Fannie Mae and Freddie toward the Special Affordable Housing increased in several steps, from 14–16 Mac have historically lagged Goal, during a period when multifamily percent in 1996–98 to 17.2 percent in depositories and the overall market in units represented only 10 percent of its 1999, and to a record level of 20.7 providing mortgage funds for special total purchase volume. For Freddie Mac, percent in 2000. That was the only year affordable housing. Between 1993 and 49 percent of units financed by prior to 2001 in which Freddie Mac’s 2002, 11.8 percent of Freddie Mac’s multifamily mortgage purchases met the performance exceeded Fannie Mae’s mortgage purchases, 12.7 percent of Special Affordable Housing Goal, performance on this Goal. Fannie Mae’s purchases, 15.4 percent of representing 23 percent of units counted Based on current counting rules, as loans originated by depositories, and toward the Special Affordable Housing shown under ‘‘official performance’’ for 15.4 percent of loans originated in the Goal, during a period when multifamily 2001 in Table 5, Special Affordable conventional conforming market units represented only 9 percent of its Housing Goal performance in 2001 was (without estimated B&C loans) were for total purchase volume. 21.6 percent for Fannie Mae and 22.6 special affordable housing. percent for Freddie Mac. Special Second, while both GSEs have c. Proposed Special Affordable Home Affordable Housing Goal performance in improved their performance over the Purchase Subgoal for 2005–2008 2002 was 21.4 percent for Fannie Mae past few years, Fannie Mae has made The Secretary believes the GSEs can and 21.4 percent for Freddie Mac. more progress than Freddie Mac in play a leadership role in the special Immediately beneath the official closing its gap with the market. The affordable market generally and the Special Affordable Housing Goal share of Fannie Mae’s purchases going home purchase special affordable performance percentages in Table 5 are to special affordable loans increased market in particular. Thus, the figures showing the GSEs’ special from 12.5 percent in 1999 to 16.3 Department is proposing a Subgoal of 17 affordable purchase percentages on a percent in 2002, the latter figure being percent for each GSE’s purchases of consistent basis for the entire 1996– at the 2002 market level of 16.3 percent. home purchase mortgages for special 2002 period. The assumptions used The share of Freddie Mac’s purchases affordable housing located in were the scoring rules established in going to special affordable loans metropolitan areas for 2005, rising to 18 HUD’s Housing Goals 2000 Final Rule increased from 12.8 percent in 1999 to percent in 2006, and 19 percent in 2007 except that bonus points and the 15.8 percent in 2002, the latter figure and 2008. The purpose of this Subgoal Freddie Mac Temporary Adjustment being below the 2002 market level of is to encourage the GSEs to improve Factor (which were terminated at the 16.3 percent. their purchases of home purchase end of 2003) are not applied. These are Section G in Appendix A discusses mortgages on special affordable housing, termed the ‘‘2001–03 baseline’’ the role of the GSEs both in the overall thus expanding homeownership assumptions. In terms of this measure, special affordable market and in the opportunities for very-low-income both Fannie Mae and Freddie Mac’s different segments (single-family owner, borrowers and low-income borrowers in special affordable performance reached single-family rental, and multifamily low-income areas, including minority its maximum in 2000 (Fannie Mae at rental) of the special affordable market. first-time homebuyers who are expected 21.4 percent and Freddie Mac at 21.0 The GSEs’ special affordable purchases to enter the housing market over the percent) before declining somewhat in accounted for 35 percent of all special next few years. If the GSEs meet this 2001 and then declining further in 2002. affordable owner and rental units that Subgoal, they will be leading the Both GSEs’ baseline performance in were financed in the conventional primary market, based on the income 2002 exceeded the level attained in conforming market between 1999 and characteristics of home purchase loans 1999. 2002. The GSEs’ 35-percent share of the reported in HMDA. Between 1999 and Overall, both GSEs’ performance special affordable market was below 2002, HMDA data show that special exceeded HUD’s Special Affordable their 49-percent share of the overall affordable housing accounted for an Housing Goals by significant margins in market. Even in the owner market, average of 16.4 percent of single-family- 1996–99, and by wide margins in 2000. where the GSEs account for 57 percent owner home purchase loans originated New, higher Goals were established for of the market, their share of the special in the conventional conforming market 2001–03, and despite somewhat lower affordable market was only 49 percent. in metropolitan areas. Loans in the B&C performance than the level attained in While the GSEs improved their market portion of the subprime market are not 2000 (largely due to the 2001–02 shares during 2002, the analysis included in these averages. To reach the refinance wave, as discussed under the suggests that the GSEs are not leading 17 percent Subgoal, both GSEs would Low- and Moderate-Income Housing the single-family market in purchasing have to improve their performance in Goal), both GSEs’ performance exceeded loans that qualify for the Special 2005, as shown in Table 2—Fannie Mae the new Goal levels in 2001–02. Affordable Housing Goal. There is room by 1.4 percentage points over its average The Special Affordable Housing Goal and ample opportunity for the GSEs to performance of 15.6 percent during is designed, in part, to ensure that the improve their performance in 2001 and 2002, and Freddie Mac by 1.9 GSEs maintain a consistent focus on purchasing affordable loans at the percentage points over its performance serving the low- and very low-income lower-income end of the market. of 15.1 percent during the same period. portion of the housing market where The multifamily market is especially These required improvements would housing needs are greatest. Appendices important in the establishment of the increase further by one percentage point A and B use HMDA data and GSE loan- Special Affordable Housing Goal for in 2006 and by an additional one

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00035 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24262 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

percentage point in 2007–08 under 1998, 313 percent in 1999, 294 percent increasing the Special Affordable HUD’s proposal. As discussed in 2000, and, under the new Subgoal Housing Goal are discussed above in previously, the Subgoal applies only to level, 258 percent in 2001, and 266 this preamble. Since the GSEs have the GSEs’ purchases in metropolitan percent in 2002. historically lagged the primary market areas because the HMDA-based market Freddie Mac’s total eligible in funding special affordable loans, they benchmark is only available for multifamily mortgage purchase volume have ample room to improve their metropolitan areas. increased even more sharply, from $0.2 performance in that market. The GSEs’ Section II.B.2 of this preamble and billion in 1993 to $6.6 billion in 1998, mortgage purchases between 1999 and Section D of Appendix C discuss and then jumped further in 2001 to 2002 accounted for 49 percent of the reasons why the Department set the $11.8 billion and $18.3 billion in 2002. total (single-family and multifamily) Subgoal for special affordable loans. Its special affordable multifamily conforming mortgage market, but they mortgage purchases followed a similar d. Special Affordable Housing Goal: accounted for only 35 percent of the path, rising from $0.1 billion in 1993 to Multifamily Subgoals special affordable market. A wide $2.7 billion in 1998, and also jumping variety of quantitative and qualitative Based on the GSEs’ past performance sharply to $4.6 billion in 2001 and $5.2 indicators demonstrate that the GSEs on the Special Affordable Multifamily billion in 2002. As a result of its strong have the expertise, resources and Subgoals, and on the outlook for the performance, Freddie Mac’s purchases financial strength to improve their multifamily mortgage market, HUD is have also been at least twice its special affordable lending performance proposing that these Subgoals be minimum Subgoal in every year since and close their gap with the market. retained for the 2005–2008 period. 1998—272 percent of the Subgoal in Unlike the overall Goals, which are that year, 229 percent in 1999, 243 C. Subpart I—Other Provisions expressed in terms of minimum Goal- percent in 2000, and, under the new Section 81.102—Independent qualifying percentages of total units Subgoal level, 220 percent in 2001, and verification authority. financed, these Subgoals for 2001–03 247 percent in 2002. See Section II of this preamble for a and in prior years have been expressed The Special Affordable Multifamily complete discussion of the Department’s in terms of minimum dollar volumes of Subgoals set forth in this proposed rule proposal to amend § 81.102 to provide Goal-qualifying multifamily mortgage are reasonable and appropriate based on additional means of verifying and purchases. Specifically, each GSE’s the Department’s analysis of this enforcing GSE data submissions. special affordable multifamily Subgoal market. The Department’s decision to IV. Findings and Certifications is currently equal to 1.0 percent of its retain these Subgoals is based on HUD’s average total (single-family plus analysis which indicates that Executive Order 12866 multifamily) mortgage volume over the multifamily housing still serves the The Office of Management and Budget 1997–99 period. Under this formulation, housing needs of lower-income families (OMB) reviewed this proposed rule in October 2000 the Subgoals were set and families in low-income areas to a under Executive Order 12866, at $2.85 billion per year for Fannie Mae greater extent than single-family Regulatory Planning and Review, which and $2.11 billion per year for Freddie housing. By retaining the Special the President issued on September 30, Mac, in each of calendar years 2001 Affordable Multifamily Subgoal, the 1993. This rule was determined to be through 2003. These Subgoals are also Department ensures that the GSEs economically significant under E.O. in effect for 2004. These represented continue their activity in this market, 12866. Any changes made to this increases from the Goals for 1996–2000, and that they achieve at least a proposed rule subsequent to its which were $1.29 billion annually for minimum level of special affordable submission to OMB are identified in the Fannie Mae and $0.99 billion annually multifamily mortgage purchases that are docket file, which is available for public for Freddie Mac. affordable to lower-income families. The inspection between 8 a.m. and 5 p.m. HUD’s Determination. The Department proposes to retain each weekdays in the Office of the Rules multifamily mortgage market and both GSE’s Special Affordable Multifamily Docket Clerk, Office of General Counsel, GSEs’ multifamily transactions volume Subgoal at 1.0 percent of its average Room 10276, Department of Housing grew significantly over the 1993–2002 annual dollar volume of total (single- and Urban Development, 451 Seventh period, indicating that both enterprises family and multifamily) mortgage Street, SW., Washington, DC. The have provided increasing support for purchases over the 2000–2002 period. In Economic Analysis prepared for this the multifamily market, and that they dollar terms, the Department’s proposal rule is also available for public have the ability to continue to provide is $5.49 billion per year in special inspection in the Office of the Rules further support for the market. affordable multifamily mortgage Docket Clerk and on HUD’s Web site at Specifically, Fannie Mae’s total purchases for Fannie Mae, and $3.92 http://www.hud.gov. eligible multifamily mortgage purchase billion per year in special affordable volume increased from $4.6 billion in multifamily mortgage purchases for Congressional Review of Major Proposed 1993 to $12.5 billion in 1998, and then Freddie Mac. These Subgoals would be Rules jumped sharply to $18.7 billion in 2001 less than actual special affordable This rule is a ‘‘major rule’’ as defined and $18.3 billion in 2002. Its special multifamily mortgage purchase volume in Chapter 8 of 5 U.S.C. At the final rule affordable multifamily mortgage in 2001 and 2002 for both GSEs. Thus, stage, the rule will be submitted for purchases followed a similar path, the Department believes that they would Congressional review in accordance rising from $1.7 billion in 1993 to $3.5 be feasible for the 2005–2008 period. with this chapter. billion in 1998 and $4.0 billion in 1999, and also jumping sharply to $7.4 billion e. Proposed Special Affordable Housing Paperwork Reduction Act in 2001 and $7.6 billion in 2002. As a Goal Levels for 2005–2008 HUD’s collection of information on result of its strong performance, Fannie The Department is proposing to the GSEs’ activities has been reviewed Mae’s purchases have been at least increase the Special Affordable Housing and authorized by the Office of twice its minimum subgoal in every Goal to 22 percent for 2005, 24 percent Management and Budget (OMB) under year since 1997—247 percent of the for 2006, 26 percent for 2007, and 28 the Paperwork Reduction Act of 1995 Subgoal in that year, 274 percent in percent for 2008. The reasons for (44 U.S.C. 3501–3520), as implemented

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00036 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24263

by OMB in regulations at 5 CFR part or on the private sector, within the of the original peoples of Hawaii, Guam, 1320. The OMB control number is meaning of UMRA. Samoa, or other Pacific Islands. 2502–0514. List of Subjects in 24 CFR Part 81 * * * * * Environmental Impact Underserved area means * * * Accounting, Federal Reserve System, (2) For purposes of the definition of This proposed rule would not direct, Mortgages, Reporting and recordkeeping ‘‘Rural area,’’ a whole census tract, a provide for assistance or loan and requirements, Securities. Federal or State American Indian mortgage insurance for, or otherwise For the reasons discussed in the reservation or tribal or individual trust govern or regulate real property preamble, HUD proposes to amend 24 land, or the balance of a census tract acquisition, disposition, lease, CFR part 81 as follows: excluding the area within any Federal or rehabilitation, alteration, demolition, or State American Indian reservation or new construction; nor would it PART 81—THE SECRETARY OF HUD’S tribal or individual trust land, having: establish, revise, or provide for REGULATION OF THE FEDERAL (i) A median income at or below 120 standards for construction or NATIONAL MORTGAGE ASSOCIATION percent of the greater of the State non- construction materials, manufactured (FANNIE MAE) AND THE FEDERAL metropolitan median income or the housing, or occupancy. Accordingly, HOME LOAN MORTGAGE nationwide non-metropolitan median under 24 CFR 50.19(c)(1) of HUD’s CORPORATION (FREDDIE MAC) income and a minority population of 30 regulations, this proposed rule is percent or greater; or categorically excluded from 1. The authority citation for 24 CFR (ii) A median income at or below 95 environmental review under the part 81 continues to read as follows: percent of the greater of the State non- National Environmental Policy Act of Authority: 12 U.S.C. 1451 et seq., 1716– metropolitan median income or 1969 (42 U.S.C. 4321). 1723h, and 4501–4641; 42 U.S.C. 3535(d) and nationwide non-metropolitan median 3601–3619. income. Regulatory Flexibility Act 2. In § 81.2, revise the definitions of * * * * * The Secretary, in accordance with the ‘‘Metropolitan area,’’ ‘‘Minority,’’ and 3. In § 81.12, revise the last sentence Regulatory Flexibility Act (5 U.S.C. paragraph (2) of the definition of of paragraph (b) and revise paragraph 605(b)), has reviewed this rule before ‘‘Underserved area,’’ and add a new (c), to read as follows: publication and by approving it certifies definition of the term ‘‘Home Purchase that this rule would not have a § 81.12 Low- and Moderate-Income Mortgage,’’ in alphabetical order, to read significant economic impact on a Housing Goal. as follows: substantial number of small entities. * * * * * This rule is applicable only to the GSEs, § 81.2 Definitions. (b) Factors. * * * A statement which are not small entities for * * * * * documenting HUD’s considerations and findings with respect to these factors, purposes of the Regulatory Flexibility Home Purchase Mortgage means a entitled ‘‘Departmental Considerations Act. Therefore, the rule does not have a residential mortgage for the purchase of to Establish the Low- and Moderate- significant economic impact on a an owner-occupied single-family Income Housing Goal,’’ was published substantial number of small entities property. within the meaning of the Regulatory in the Federal Register on date of * * * * * Flexibility Act. publication of final rule in the Federal Metropolitan area means a Register]. Executive Order 13132, Federalism metropolitan statistical area (‘‘MSA’’), or (c) Goals. The annual goals for each Executive Order 13132 (‘‘Federalism’’) a portion of such an area for which GSE’s purchases of mortgages on prohibits, to the extent practicable and median family income estimates are housing for low- and moderate-income permitted by law, an agency from published annually by HUD. families are: promulgating a regulation that has Minority means any individual who is (1) For the year 2005, 52 percent of federalism implications and either included within any one or more of the the total number of dwelling units imposes substantial direct compliance following racial and ethnic categories: financed by that GSE’s mortgage costs on state and local governments (1) American Indian or Alaskan purchases unless otherwise adjusted by and is not required by statute, or Native—a person having origins in any HUD in accordance with FHEFSSA. In preempts state law, unless the relevant of the original peoples of North and addition, as a Low- and Moderate- requirements of section 6 of the South America (including Central Income Housing Home Purchase executive order are met. This proposed America), and who maintains tribal Subgoal, 45 percent of the total number rule does not have federalism affiliation or community attachment; of home purchase mortgages in implications and does not impose (2) Asian—a person having origins in metropolitan areas financed by that substantial direct compliance costs on any of the original peoples of the Far GSE’s mortgage purchases shall be home state and local governments or preempt East, Southeast Asia, or the Indian purchase mortgages in metropolitan state law within the meaning of the subcontinent, including, for example, areas which count toward the Low- and executive order. Cambodia, China, India, Japan, Korea, Moderate-Income Housing Goal in the Malaysia, Pakistan, the Philippine year 2005 unless otherwise adjusted by Unfunded Mandates Reform Act Islands, Thailand, and Vietnam; HUD in accordance with FHEFSSA; Title II of the Unfunded Mandates (3) Black or African American—a (2) For the year 2006, 53 percent of Reform Act of 1995 (12 U.S.C. 1531— person having origins in any of the the total number of dwelling units 1538) (UMRA) establishes requirements black racial groups of Africa; financed by that GSE’s mortgage for federal agencies to assess the effects (4) Hispanic or Latino—a person of purchases unless otherwise adjusted by of their regulatory actions on state, Cuban, Mexican, Puerto Rican, South or HUD in accordance with FHEFSSA. In local, and tribal governments, and the Central American, or other Spanish addition, as a Low- and Moderate- private sector. This proposed rule culture or origin, regardless of race; and Income Housing Home Purchase would not impose any federal mandates (5) Native Hawaiian or Other Pacific Subgoal, 46 percent of the total number on any state, local, or tribal government, Islander—a person having origins in any of home purchase mortgages in

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00037 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24264 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

metropolitan areas financed by that to Establish the Central Cities, Rural Areas, and Other Underserved Areas GSE’s mortgage purchases shall be home Areas, and Other Underserved Areas Home Purchase Subgoal, 35 percent of purchase mortgages in metropolitan Housing Goal,’’ was published in the the total number of home purchase areas which count toward the Low- and Federal Register on [date of publication mortgages in metropolitan areas Moderate-Income Housing Goal in the of final rule in the Federal Register]. financed by that GSE’s mortgage year 2006 unless otherwise adjusted by (c) Goals. The annual goals for each purchases shall be home purchase HUD in accordance with FHEFSSA; GSE’s purchases of mortgages on mortgages in metropolitan areas which (3) For the year 2007, 55 percent of housing located in central cities, rural count toward the Central Cities, Rural the total number of dwelling units areas, and other underserved areas are: Areas, and Other Underserved Areas financed by that GSE’s mortgage (1) For the year 2005, 38 percent of Housing Goal in the year 2008 unless purchases unless otherwise adjusted by the total number of dwelling units otherwise adjusted by HUD in HUD in accordance with FHEFSSA. In financed by that GSE’s mortgage accordance with FHEFSSA; and addition, as a Low- and Moderate- purchases unless otherwise adjusted by (5) For the year 2009 and thereafter Income Housing Home Purchase HUD in accordance with FHEFSSA. In HUD shall establish annual goals. Subgoal, 47 percent of the total number addition, as a Central Cities, Rural Pending establishment of goals for the of home purchase mortgages in Areas, and Other Underserved Areas year 2009 and thereafter, the annual metropolitan areas financed by that Home Purchase Subgoal, 33 percent of goal for each of those years shall be 40 GSE’s mortgage purchases shall be home the total number of home purchase percent of the total number of dwelling purchase mortgages in metropolitan mortgages in metropolitan areas units financed by that GSE’s mortgage areas which count toward the Low- and financed by that GSE’s mortgage purchases in each of those years. In Moderate-Income Housing Goal in the purchases shall be home purchase addition, as a Central Cities, Rural year 2007 unless otherwise adjusted by mortgages in metropolitan areas which Areas, and Other Underserved Areas HUD in accordance with FHEFSSA; count toward the Central Cities, Rural Home Purchase Subgoal, 35 percent of (4) For the year 2008, 57 percent of Areas, and Other Underserved Areas the total number of home purchase the total number of dwelling units Housing Goal in the year 2005 unless mortgages in metropolitan areas financed by that GSE’s mortgage otherwise adjusted by HUD in financed by that GSE’s mortgage purchases unless otherwise adjusted by accordance with FHEFSSA; purchases shall be home purchase HUD in accordance with FHEFSSA. In (2) For the year 2006, 39 percent of mortgages in metropolitan areas which addition, as a Low- and Moderate- the total number of dwelling units count toward the Central Cities, Rural Income Housing Home Purchase financed by that GSE’s mortgage Areas, and Other Underserved Areas Subgoal, 47 percent of the total number purchases unless otherwise adjusted by Housing Goal in each of those years of home purchase mortgages in HUD in accordance with FHEFSSA. In unless otherwise adjusted by HUD in metropolitan areas financed by that addition, as a Central Cities, Rural accordance with FHEFSSA. GSE’s mortgage purchases shall be home Areas, and Other Underserved Areas * * * * * purchase mortgages in metropolitan Home Purchase Subgoal, 34 percent of 5. In § 81.14, revise the last sentence areas which count toward the Low- and the total number of home purchase of paragraph (b) and revise paragraph Moderate-Income Housing Goal in the mortgages in metropolitan areas (c), to read as follows: year 2008 unless otherwise adjusted by financed by that GSE’s mortgage § 81.14 Special Affordable Housing Goal. HUD in accordance with FHEFSSA; and purchases shall be home purchase (5) For the year 2009 and thereafter mortgages in metropolitan areas which * * * * * HUD shall establish annual goals. count toward the Central Cities, Rural (b) * * * A statement documenting Pending establishment of goals for the Areas, and Other Underserved Areas HUD’s considerations and findings with year 2009 and thereafter, the annual Housing Goal in the year 2006 unless respect to these factors, entitled goal for each of those years shall be 57 otherwise adjusted by HUD in ‘‘Departmental Considerations to percent of the total number of dwelling accordance with FHEFSSA; Establish the Special Affordable units financed by that GSE’s mortgage (3) For the year 2007, 39 percent of Housing Goal,’’ was published in the purchases in each of those years. In the total number of dwelling units Federal Register on [date of publication addition, as a Low and Moderate financed by that GSE’s mortgage of final rule in the Federal Register]. Income Housing Home Purchase purchases unless otherwise adjusted by (c) Goals. The annual goals for each Subgoal, 47 percent of the total number HUD in accordance with FHEFSSA. In GSE’s purchases of mortgages on rental of home purchase mortgages in addition, as a Central Cities, Rural and owner-occupied housing meeting metropolitan areas financed by that Areas, and Other Underserved Areas the then-existing, unaddressed needs of GSE’s mortgage purchases shall be home Home Purchase Subgoal, 35 percent of and affordable to low-income families in purchase mortgages in metropolitan the total number of home purchase low-income areas and very low-income areas which count toward the Low- and mortgages in metropolitan areas families are: (1) For the year 2005, 22 percent of Moderate-Income Housing Goal in each financed by that GSE’s mortgage the total number of dwelling units of those years unless otherwise adjusted purchases shall be home purchase financed by each GSE’s mortgage by HUD in accordance with FHEFSSA. mortgages in metropolitan areas which 4. In § 81.13, revise the last sentence count toward the Central Cities, Rural purchases unless otherwise adjusted by of paragraph (b) and revise paragraph Areas, and Other Underserved Areas HUD in accordance with FHEFSSA. The (c), to read as follows: Housing Goal in the year 2007 unless goal for the year 2005 shall include otherwise adjusted by HUD in mortgage purchases financing dwelling § 81.13 Central Cities, Rural Areas, and accordance with FHEFSSA; units in multifamily housing totaling Other Underserved Areas Housing Goal. (4) For the year 2008, 40 percent of not less than 1.0 percent of the average * * * * * the total number of dwelling units annual dollar volume of combined (b) Factors. * * * A statement financed by that GSE’s mortgage (single family and multifamily) documenting HUD’s considerations and purchases unless otherwise adjusted by mortgages purchased by the respective findings with respect to these factors, HUD in accordance with FHEFSSA. In GSE in 2000, 2001, and 2002, unless entitled ‘‘Departmental Considerations addition, as a Central Cities, Rural otherwise adjusted by HUD in

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00038 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24265

accordance with FHEFSSA. In addition, goal for the year 2008 shall include housing goal. The denominator of each as a Special Affordable Housing Home mortgage purchases financing dwelling fraction shall be the total number of Purchase Subgoal, 17 percent of the units in multifamily housing totaling home purchase mortgages in total number of home purchase not less than 1.0 percent of the average metropolitan areas financed by each mortgages in metropolitan areas annual dollar volume of combined GSE’s mortgage purchases in a financed by each GSE’s mortgage (single-family and multifamily) particular year. For purposes of each purchases shall be home purchase mortgages purchased by the respective Subgoal, the procedure for addressing mortgages in metropolitan areas which GSE in 2000, 2001, and 2002, unless missing data or information, as set forth count toward the Special Affordable otherwise adjusted by HUD in in paragraph (d) of this section, shall be Housing Goal in the year 2005 unless accordance with FHEFSSA. In addition, implemented using numbers of home otherwise adjusted by HUD in as a Special Affordable Housing Home purchase mortgages in metropolitan accordance with FHEFSSA; Purchase Subgoal, 19 percent of the areas and not single-family owner- (2) For the year 2006, 24 percent of total number of home purchase occupied dwelling units. the total number of dwelling units mortgages in metropolitan areas (2) Special counting rule for financed by each GSE’s mortgage financed by each GSE’s mortgage mortgages with more than one owner- purchases unless otherwise adjusted by purchases shall be home purchase occupied unit. For purposes of counting HUD in accordance with FHEFSSA. The mortgages in metropolitan areas which mortgages toward the Home Purchase goal for the year 2006 shall include count toward the Special Affordable Subgoals, where a single home purchase mortgage purchases financing dwelling Housing Goal in the year 2008 unless mortgage finances the purchase of two units in multifamily housing totaling otherwise adjusted by HUD in or more owner-occupied units in a not less than 1.0 percent of the average accordance with FHEFSSA; and metropolitan area, the mortgage shall annual dollar volume of combined (5) For the year 2009 and thereafter count once toward each Subgoal that (single-family and multifamily) HUD shall establish annual goals. applies to the GSE’s mortgage purchase. mortgages purchased by the respective Pending establishment of goals for the 7. Remove and reserve § 81.16(c)(1) GSE in 2000, 2001, and 2002, unless year 2009 and thereafter, the annual and (c)(11). otherwise adjusted by HUD in goal for each of those years shall be 28 8. Revise § 81.102 to read as follows: accordance with FHEFSSA. In addition, percent of the total number of dwelling as a Special Affordable Housing Home units financed by each GSE’s mortgage § 81.102 Verification and enforcement to Purchase Subgoal, 18 percent of the purchases in each of those years. The ensure GSE data integrity. total number of home purchase goal for each such year shall include (a) Independent verification authority. mortgages in metropolitan areas mortgage purchases financing dwelling The Secretary may independently verify financed by each GSE’s mortgage units in multifamily housing totaling the accuracy and completeness of the purchases shall be home purchase not less than 1.0 percent of the annual data, information, and reports provided mortgages in metropolitan areas which average dollar volume of combined by each GSE, including conducting on- count toward the Special Affordable (single-family and multifamily) site verification, when such steps are Housing Goal in the year 2006 unless mortgages purchased by the respective reasonably related to determining otherwise adjusted by HUD in GSE in the years 2000, 2001, and 2002. whether a GSE is complying with 12 accordance with FHEFSSA; In addition, as a Special Affordable U.S.C. 4541’4589 and the GSE’s Charter (3) For the year 2007, 26 percent of Housing Home Purchase Subgoal, 19 Act. the total number of dwelling units percent of the total number of home (b) Certification. The senior officer of financed by each GSE’s mortgage purchase mortgages in metropolitan each GSE who is responsible for purchases unless otherwise adjusted by areas financed by each GSE’s mortgage submitting to HUD the AHAR under HUD in accordance with FHEFSSA. The purchases shall be home purchase section 309(m) and (n) of the Fannie goal for the year 2007 shall include mortgages in metropolitan areas which Mae Act or section 307(e) and (f) of the mortgage purchases financing dwelling count toward the Special Affordable Freddie Mac Charter Act, as applicable, units in multifamily housing totaling Housing Goal in each of those years or for submitting to HUD such other not less than 1.0 percent of the average unless otherwise adjusted by HUD in report(s), data submission(s), or annual dollar volume of combined accordance with FHEFSSA. information for which certification is (single-family and multifamily) * * * * * requested in writing by HUD (‘‘GSE mortgages purchased by the respective 6. Add § 81.15(i), to read as follows: Certifying Official’’) shall certify in GSE in 2000, 2001, and 2002, unless connection with each such report(s), otherwise adjusted by HUD in § 81.15 General requirements. data submission(s) or information that: accordance with FHEFSSA. In addition, * * * * * (1) The GSE Certifying Official has as a Special Affordable Housing Home (i) Counting mortgages toward the reviewed the particular AHAR, other Purchase Subgoal, 19 percent of the Home Purchase Subgoals. (1) General. report(s), data submission(s) or total number of home purchase The requirements of this section, except information; mortgages in metropolitan areas for paragraphs (b) and (e) of this section, (2) To the best of the GSE Certifying financed by each GSE’s mortgage shall apply to counting mortgages Official’s knowledge and belief, the purchases shall be home purchase toward the Home Purchase Subgoals at particular AHAR, other report(s), data mortgages in metropolitan areas which §§ 81.12–81.14. However, performance submission(s) or information are count toward the Special Affordable under the Subgoals shall be counted current, complete and do not contain Housing Goal in the year 2007 unless using a fraction that is converted into a any untrue statement of a material fact; otherwise adjusted by HUD in percentage for each Subgoal and the (3) To the best of the GSE Certifying accordance with FHEFSSA; numerator of the fraction for each Official’s knowledge and belief, the (4) For the year 2008, 28 percent of Subgoal shall be the number of home particular AHAR, other report(s), data the total number of dwelling units purchase mortgages in metropolitan submission(s) or information fairly financed by each GSE’s mortgage areas financed by each GSE’s mortgage present in all material respects the purchases unless otherwise adjusted by purchases in a particular year that count GSE’s performance, as required to be HUD in accordance with FHEFSSA. The towards achievement of the applicable reported by section 309(m) or (n) of the

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00039 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24266 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Fannie Mae Act or section 307(e) or (f) of the Freddie Mac Charter Act, as its AHAR for the current year, as of the Freddie Mac Charter Act, or other applicable) that result in an directed by HUD. applicable legal authority; and overstatement of GSE housing goal (6) Effect of failure to meet a housing (4) To the best of the GSE Certifying performance. goal, or substantial probability of such Official’s knowledge and belief, the GSE (2) Applicability. This paragraph failure. has identified in writing any areas in applies to errors, omissions or (i) Procedural requirements. In the which the GSE’s particular AHAR, other discrepancies in a GSE’s data event HUD determines that a GSE has report(s), data submission(s) or submissions, including its AHAR, as failed, or that there is a substantial information may differ from HUD’s provided in this section. It does not probability that the GSE will fail, to written articulations of its counting apply to the process applicable to meet any housing goal(s) in the current rules including, but not limited to, the HUD’s review of current year reporting year as a result of an regulations under this part, and any performance, as described in paragraph adjustment under paragraph (d) (5) of other areas of ambiguity. (c) of this section. this section for previously overstated (c) Adjustment to correct current year- (3) Limitations. This paragraph housing goals performance, HUD shall end errors, omissions or discrepancies. applies only to GSE reporting periods provide written notice to the GSE and If HUD finds errors, omissions or occurring on or after [effective date of otherwise comply with the procedural discrepancies in a GSE’s current year- final rule]. requirements set forth in 12 U.S.C. end data submissions (including data 4566(b). (4) Procedural requirements. In the reported in the GSE’s AHAR under (ii) Remedies. If HUD determines event HUD determines that an section 309(m) and (n) of the Fannie pursuant to 12 U.S.C. 4566(b) that a GSE adjustment to correct an error, omission Mae Act or section 307(e) and (f) of the has failed, or that there is a substantial or discrepancy in a GSE’s prior year’s Freddie Mac Charter Act, as applicable) probability that the GSE will fail, any data submissions (including data relative to HUD’s regulations or other housing goal(s) in the current reporting reported in the AHAR), as provided in guidance, HUD will first notify the GSE year as a result of an adjustment under paragraph (d)(1) of this section is by telephone or e-mail transmission of paragraph (d) (5) of this section to warranted, it will provide the GSE with each such error, omission or correct for an overstatement of a prior an initial letter containing its written discrepancy. The GSE must respond year’s goals performance, and that the within five business days of such findings and determinations within 24 achievement of the housing goal was or notification. If each error, omission or months of the end of the relevant GSE is feasible, it may pursue one or both of discrepancy is not resolved to HUD’s reporting year. The GSE shall have an the following remedies: satisfaction, HUD will then notify the opportunity, not to exceed 30 days from (A) Housing plan. HUD may require GSE in writing and seek clarification or the date of HUD’s initial letter, to the GSE to submit a housing plan for additional information to correct the respond in writing, with supporting approval by the Secretary pursuant to 12 error, omission or discrepancy. The GSE documentation, to contest the initial U.S.C. 4566(c) and § 81.22; and shall have 10 business days (or such determination that there were errors in (B) Additional enforcement options. longer period as HUD may establish, not a prior year’s data submissions HUD may, after complying with the to exceed 30 business days) from the (including the AHAR). HUD shall then procedural requirements set forth in date of this written notice to respond in issue a final determination letter within subpart G of this part, seek a cease-and- writing to the request. If the GSE fails 60 days of the date of the GSE’s written desist order or civil money penalties to submit a written response to HUD response. HUD may, upon a against the GSE as described in within this period, or if HUD determination of good cause, extend the paragraph (e) of this section. determines that the GSE’s written period for issuing a final determination (e) Additional enforcement options. response fails to explain or correct each letter by an additional 30 days. (1) General. In the event the Secretary error, omission or discrepancy in its (5) Adjustments. If the GSE failed to determines, either as a result of its current year-end reported data to HUD’s submit a written response to HUD’s independent verification authority satisfaction, HUD will determine the initial determination letter within the described in paragraph (a) of this appropriate adjustments to the 30-day time period, or if, after reviewing section or by other means, that the data numerator and the denominator of the a GSE’s written response to the initial submissions, information or report(s) applicable housing goal(s) and determination letter, HUD determines submitted by a GSE to HUD pursuant to Subgoal(s). Should the Department that a GSE’s prior year’s data subpart E of this part, section 309(m) or determine that additional enforcement submissions (including data reported in (n) of the Fannie Mae Charter Act, or action against the GSE is warranted, it the AHAR as provided in paragraph section 307(e) and (f) of the Freddie Mac may pursue additional remedies under (d)(1) of this section) resulted in an Charter Act, as applicable, are not paragraph (e) of this section. overstatement of its performance under current, are incomplete or otherwise (d) Adjustment to correct prior year one or more housing goals or Subgoals contain an untrue statement of material reporting errors, omissions or for a previous reporting period, HUD fact, the Secretary may regard this as discrepancies. will direct the GSE to correct the equivalent to the GSE’s failing to submit (1) General. HUD may, in accordance overstatement by adjusting its level of such data and, accordingly, may take with its authority in 12 U.S.C. 4566(a) performance under the applicable the enforcement action authorized to measure the extent of compliance housing goal(s) and/or Subgoal(s) in the under paragraph (e)(2) of this section. with the housing goals, adjust a GSE’s current year AHAR prior to submitting (2) Remedies. After HUD makes a current year-end performance under a such report to HUD. The adjustment final determination pursuant to housing goal to deduct credit under the will be made by excluding the number paragraph (e) of this section that a GSE current goals and/or Subgoals to the of units or mortgages that HUD has has submitted report(s), data extent caused by errors, omissions or determined were erroneously counted submission(s) or information that are discrepancies in a GSE’s prior year’s in a previous year from the numerator not current, are incomplete, or that data submissions (including the AHAR (but not the denominator) of each contain untrue statement(s) of material under section 309(m) and (n) of the applicable housing goal and/or Subgoal. fact, it may pursue any or all of the Fannie Mae Act or section 307(e) and (f) The GSE shall reflect the adjustment in following remedies:

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00040 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24267

(i) HUD may obtain a cease-and-desist families relative to the size of the overall Low- and Moderate-Income Housing Goal. order against the GSE for failing to conventional mortgage market; Section I also gives the rationale for a low- submit the report(s), data submission(s) (5) The ability of the enterprises to lead the and moderate-income subgoal for home or information, as applicable, required industry in making mortgage credit available purchase loans. for low- and moderate-income families; and The consideration of the factors in this by subsection (m) or (n) of section 309 (6) The need to maintain the sound Appendix has led the Secretary to the of the Fannie Mae Charter Act or financial condition of the enterprises. following conclusions: subsection (e) or (f) of the Freddie Mac The Secretary also considered these factors • Changing population demographics will Charter Act, and as authorized by 12 in establishing a low- and moderate-income result in a need for primary and secondary U.S.C. 4581(a)(3), § 81.82, and subpart E subgoal for home purchase loans on single- mortgage markets to meet nontraditional of this part; family-owner properties in metropolitan credit needs, respond to diverse housing (ii) HUD may seek civil money areas. preferences, and overcome information and other barriers that many immigrants and penalties against the GSE for failing to 2. Underlying Data submit the report(s), data submissions, minorities face. Growing housing demand In considering the statutory factors in or information, as applicable, required from immigrants (both those who are already establishing these goals, HUD relied on data here and those projected to come) and non- by subsection (m) or (n) of section 309 from the 2001 American Housing Survey, the traditional homebuyers will help to offset of the Fannie Mae Charter Act or 2000 Censuses of Population and Housing, declines in the demand for housing caused subsection (e) or (f) of the Freddie Mac the 1991 Residential Finance Survey (RFS), by the aging of the population. Immigrants Charter Act, and as authorized by 12 the 1995 Property Owners and Managers and other minorities—who accounted for U.S.C. 4585(a)(3), 24 CFR 81.83 and Survey (POMS), other government reports, nearly 40 percent of the growth in the Subpart E of this part. reports submitted in accordance with the nation’s homeownership rate over the past (iii) HUD may seek any other Home Mortgage Disclosure Act (HMDA), and five years—will be responsible for almost remedies or penalties against the GSE the GSEs. In order to measure performance two-thirds of the growth in the number of that may be available to the Secretary by toward achieving the Low- and Moderate- new households over the next ten years. As Income Housing Goal in previous years, HUD virtue of the GSE’s failure to provide these demographic factors play out, the analyzed the loan-level data on all mortgages overall effect on housing demand will likely data submissions, information and/or purchased by the GSEs for 1993–2002 in be sustained growth and an increasingly report(s) in accordance with the accordance with the goal counting provisions diverse household population from which to requirements of this section. established by the Department in the draw new renters and homeowners. (3) Procedures. HUD shall comply December 1995 and October 2000 rules (24 • Despite the record national with the procedures set forth in Subpart CFR part 81). homeownership rate of 67.9 percent in 2002, G of this part in connection with any 3. Conclusions Based on Consideration of the much lower rates prevailed for minorities, enforcement action that it initiates Factors especially for African-American households against a GSE under this paragraph. (47.9 percent) and Hispanics (48.2 percent), The discussion of the first two factors and these lower rates are only partly Dated: April 2, 2004. covers a range of topics on housing needs accounted for by differences in income, age, John C. Weicher, and economic and demographic trends that and other socioeconomic factors. Assistant Secretary for Housing—Federal are important for understanding mortgage • In addition to low incomes, barriers to Housing Commissioner. markets. Information is provided which homeownership that disproportionately describes the market environment in which affect minorities and immigrants include lack Note: The following appendices will not the GSEs must operate (for example, trends of capital for down payments and closing appear in the Code of Federal Regulations. in refinancing activity). In addition, the costs, poor credit history, lack of access to severe housing problems faced by lower- mainstream lenders, little understanding of Appendix A—Departmental income families are discussed, as are the the home buying process, and continued Considerations To Establish the Low- barriers that minorities face when attempting discrimination in housing markets and and Moderate-Income Housing Goal to become homeowners. This discussion mortgage lending. serves to provide useful background • A HUD-published study of A. Introduction information for the discussion of the discrimination in the rental and owner Sections 1 and 2 provide a basic Underserved Areas and Special Affordable markets found that while differential description of the rule process. Section 3 Housing Goals in Appendixes B and C, as treatment between minority and white home discusses conclusions based on consideration well as for the Low- and Moderate-Income seekers had declined over the past ten years, of the factors. Housing Goal in this Appendix. it continued at an unacceptable level in the The third factor (past performance) and the year 2000. In addition, disparities in 1. Establishment of Low- and Moderate- fifth factor (ability of the GSEs to lead the mortgage lending continued across the nation Income Goal industry) are also discussed in some detail in in 2002, when the loan denial rate was 7.8 In establishing the Low- and Moderate- this Appendix. With respect to home percent for white mortgage applicants, but Income Housing Goals for the Federal purchase mortgages, the past performance of 20.1 percent for African Americans and 15.5 National Mortgage Association (Fannie Mae) the GSEs and their ability to lead the percent for Hispanics.1 and the Federal Home Loan Mortgage industry are examined for all three housing • Americans with the lowest incomes face Corporation (Freddie Mac), collectively goals; that analysis provides the basis for persistent housing problems. Recent HUD referred to as the Government-Sponsored establishing the three subgoals for the GSEs’ analysis reveals that in 2001, 5.1 million Enterprises (GSEs), Section 1332 of the acquisitions of home loans on single-family- households had ‘‘worst case’’ housing needs, Federal Housing Enterprises Financial Safety owner properties. defined as housing costs greater than 50 and Soundness Act of 1992 (12 U.S.C. 4562) The fourth factor (size of the market) and percent of household income or severely (FHEFSSA) requires the Secretary to the sixth factor (need to maintain the GSEs’ inadequate housing among unassisted very- consider: sound financial condition) are mentioned low-income renter households. Among these (1) National housing needs; only briefly in this Appendix. Detailed households, 90 percent had a severe rent (2) Economic, housing, and demographic analyses of the fourth factor and the sixth burden, 6 percent lived in severely conditions; factor are contained in Appendix D and in inadequate housing, and 4 percent suffered (3) The performance and effort of the the economic analysis of this rule, from both problems. enterprises toward achieving the Low- and respectively. Moderate-Income Housing Goal in previous The factors are discussed in sections B 1 Mortgage denial rates are based on 2002 HMDA years; through H of this appendix. Section I data for home purchase loans; manufactured (4) The size of the conventional mortgage summarizes the findings and presents the housing lenders are excluded from these market serving low- and moderate-income Department’s conclusions concerning the comparisons.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00041 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24268 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

• Over the past ten years, there has been 2001 and 2002, the first two years of HUD’s • The long run outlook for the multifamily a ‘‘revolution in affordable lending’’ that has higher housing goal levels. Fannie Mae’s rental market is sustained, moderate growth, extended homeownership opportunities to average performance during 2001 and 2002 based on favorable demographics. The historically underserved households. Fannie approached the market on the special minority population, especially Hispanics, Mae and Freddie Mac have been a substantial affordable and underserved areas categories provides a growing source of demand for part of this ‘‘revolution in affordable and matched the market on the low-mod affordable rental housing. ‘‘Lifestyle renters’’ lending.’’ During the mid-to-late 1990s, they category. Under one measure of GSE and (older, middle-income households) are also a added flexibility to their underwriting market activity, Fannie Mae matched the fast-growing segment of the rental guidelines, introduced new low-down- market during 2002 on the special affordable population. Provision of affordable housing, payment products, and worked to expand the category and slightly outperformed the however, will continue to challenge use of automated underwriting in evaluating market on the low-mod and underserved suppliers of multifamily rental housing and the creditworthiness of loan applicants. areas categories. In this case, which is policy makers at all levels of government. HMDA data suggest that the industry and referred to in the text as the ‘‘purchase year’’ Low incomes combined with high housing GSE initiatives are increasing the flow of approach, Fannie Mae’s performance is based costs define a difficult situation for millions credit to underserved borrowers. Between on comparing its purchases of all loans (both of renter households. Housing cost 1993 and 2002, conventional loans to low- seasoned loans and newly-originated reductions are constrained by high land income and minority families increased at mortgages) during a particular year with prices and construction costs in many much faster rates than loans to upper-income loans originated in the market in that year. markets. Government action—through land and non-minority families. When Fannie Mae’s performance is measured use regulation, building codes, and • The Low- and Moderate-Income Goal on an ‘‘origination year’’ basis (that is, occupancy standards—are major contributors was set at 50 percent beginning in 2001. allocating Fannie Mae’s purchases in a to those high costs. Effective on January 1, 2001, several changes particular year to the year that the purchased- • The market for financing multifamily in counting requirements came into effect, loan was originated), Fannie Mae matched apartments has grown to record volumes. including (1) ‘‘bonus points’’ (double credit) the market in the low- and moderate-income Fannie Mae and Freddie Mac have been for purchases of mortgages on small (5–50 category during 2002, and lagged the market among those boosting volumes and unit) multifamily properties and, above a slightly on the other two categories. introducing new programs to serve the threshold level, mortgages on 2–4 unit • Both Fannie Mae and Freddie lag the multifamily market. Fannie Mae’s owner-occupied properties; and (2) a conventional conforming market in funding multifamily purchases jumped from about ‘‘temporary adjustment factor’’ (1.35 unit first-time homebuyers, and by a rather wide $10 billion in 1999 and 2000 to $18.7 billion credit) for Freddie Mac’s purchases of margin. Between 1999 and 2001, first-time during the heavy refinancing year of 2001, mortgages on large (more than 50 units) homebuyers accounted for 27 percent of each and $18.3 billion in 2002. multifamily properties. With these two GSE’s purchases of home loans, compared • Freddie Mac has re-entered the counting rules, Fannie Mae’s performance with 38 percent for home loans originated in multifamily market, after withdrawing for a was 51.5 percent in 2001 and 51.8 percent in the conventional conforming market. time in the early 1990s. Concerns regarding 2002, and Freddie Mac’s performance was • The GSEs have accounted for a Freddie Mac’s multifamily capabilities no 53.2 percent in 2001 and 51.4 percent in significant share of the total (government as longer constrain its performance with regard 2002; thus, both GSEs surpassed this higher well as conventional) market for home to the housing goals. Freddie Mac’s goal in both years. purchase loans, but their market share for multifamily purchases increased from a • The bonuses and temporary adjustment each of the affordable lending categories (e.g., relatively low $3 billion in 1997 to factor expired at the end of 2003. Without low-income borrowers and census tracts, approximately $7 billion during the next these rules, Fannie Mae’s performance would high-minority census tracts) has been less three years (1998 to 2000), before rising have been 51.3 percent in 2000, 49.2 percent than their share of the overall market. further to $11.9 billion in 2001 and $13.3 in 2001, and 49.0 percent in 2002. Freddie • The GSEs also account for a very small billion in 2002. Mac’s performance would have been 50.6 share of the market for important groups such • The overall presence of both GSEs in the percent in 2000, 47.7 percent in 2001, and as minority first-time homebuyers. rental mortgage market falls short of their 46.5 percent in 2002. Thus, both Fannie Mae Considering the total mortgage market (both involvement in the single-family owner and Freddie Mac would have surpassed the government and conventional loans), it is market. Between 1999 and 2002, the GSEs’ 50 percent goal in 2000 and fallen short in estimated that the GSEs purchased only 14 purchases totaled for 57 percent of the owner 2001 and 2002. percent of loans originated between 1999 and market, but only 27 percent of the single- • This Appendix includes a 2001 for African-American and Hispanic family rental market and 30 percent of the comprehensive analysis of each GSE’s first-time homebuyers, or one-third of their multifamily market. Certainly there is room performance in funding home purchase share (42 percent) of all home purchase loans for expansion of the GSEs in supporting the mortgages for borrowers and neighborhoods originated during that period. Considering nation’s rental markets, and that expansion is covered by the three housing goals—special the conventional conforming market and the needed if the GSEs are to make significant affordable and low- and moderate-income same time period, it is estimated that the progress in closing the gaps between the borrowers and underserved areas. In GSEs purchased only 31 percent of loans affordability of their mortgage purchases and addition, the role of the GSEs in the first-time originated for African-American and that of the overall conventional conforming homebuyer market is examined. While Hispanic first-time homebuyers, or market. Freddie Mac has improved its affordable approximately one-half of their share (57 • Considering both owner and rental lending performance in recent years, it has percent) of all home purchase loans in that properties, the GSEs’ presence in the goals- consistently lagged the conventional market. The GSEs’ small share of the first- qualifying market has been significantly less conforming market in funding affordable time homebuyer market could be due to the than their presence in the overall home purchase loans for borrowers and preponderance of high (over 20 percent) conventional conforming mortgage market. neighborhoods targeted by the housing goals. downpayment loans in their mortgage Specifically, HUD estimates that the GSEs However, Freddie Mac’s recent performance purchases. accounted for 49 percent of all owner and (1999–2002) has been much closer to the • This Appendix discusses the dynamic rental units financed in the primary market market than its earlier performance. nature of the single-family mortgage market between 1999 and 2002, but only 32 percent • In general, Fannie Mae’s affordable and the numerous changes that that this of units qualifying for the low-mod goal, 41 lending performance has been better than market has undergone over the past few percent of units qualifying for the Freddie Mac’s. But like Freddie Mac, Fannie years. Some important trends that will likely underserved areas goal, and 35 percent of Mae’s average performance during past factor into the GSEs’ performance in meeting units qualifying for special affordable goal. periods (e.g., 1993–2002, 1996–2002, 1999– the needs of underserved borrowers include 2002) has been below market levels. the growth of the subprime market, the B. Factor 1: National Housing Needs However, it is encouraging that Fannie Mae increasing use of automated underwriting This section reviews the general housing markedly improved its affordable lending systems, and the introduction of risk-based needs of lower-income families that exist performance relative to the market during pricing into the market. today and are expected to continue in the

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00042 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24269

near future. Affordability problems that homeowners under 65 was twelve times that affordability were seen in 2001 as lower lower-income families face in both the rental of a similar renter.8 Thus a homeownership interest rates and modest income growth and owner markets are examined. The gap continues to translate directly into a reduced the average monthly mortgage section also describes racial disparities in wealth gap. payment from its year-ago level.14 However, homeownership and the causes of these High rates of homeownership support many households still lack the earning power disparities. It also notes some special economic stability within housing and to take advantage of today’s home buying problems, such as the need to rehabilitate our related industries, sectors that contributed opportunities. Several trends have older urban housing stock, that are discussed nearly one-half of the total gain in real GDP contributed to the reduction in the real throughout this appendix. in 2001.9 In addition to economic benefits earnings of young adults without college such as jobs and residential investment, education over the last 15 years, including 1. Homeownership Gaps studies show that the better living technological changes that favor white-collar Despite recent record homeownership environment associated with owning a home employment, losses of unionized rates, many Americans, including has positive impacts on children, in terms of manufacturing jobs, and wage pressures disproportionate numbers of racial and lower rates of teenage pregnancy and higher exerted by globalization. Over 42 percent of ethnic minorities, are shut out of reading other test scores. The current the nation’s population between the ages of homeownership opportunities. Although the literature substantiates that the benefits of 25 and 34 had no advanced education in national homeownership rate for all homeownership extend beyond individual 200015 and were therefore at risk of being Americans stood at 68.3 percent at the end homeowners and their families to society at unable to afford homeownership. African of 2003, the rate for minority households was large. Homeownership promotes social and Americans and Hispanics, who have lower lower—for example, just 48.5 percent of community stability by increasing the average levels of educational attainment than African-American households and 48.3 number of stakeholders and reducing whites, are especially disadvantaged by the percent of Hispanic households owned a disparities in the distributions of wealth and erosion in wages among less educated home. Differences in income and age income. The empirical literature is generally workers. between minorities and whites do not fully supportive of a relationship between Immigrants and other minorities, who explain these gaps. The Joint Center for homeownership and greater investment in accounted for nearly 40 percent of the growth Housing Studies estimated that if minorities property.10 Homeownership is also in the homeownership rate over the past five owned homes at the same rates as whites of associated with neighborhood stability (lower years, will be responsible for two-thirds of similar age and income, a homeownership mobility), greater participation in voluntary the growth in new households over the next 2 11 gap of 10 percentage points would still exist. and political activities, and links to ten years. These groups have unique housing 12 a. Importance of Homeownership entrepreneurship. needs and face numerous hurdles in Homeownership is one of the most b. Barriers to Homeownership 13 becoming homeowners. In addition to low common forms of property ownership as well Insufficient income, high debt burdens, income, barriers to homeownership that as savings.3 Historically, home equity has and limited savings are obstacles to disproportionately affect minorities and been the largest source of wealth for most homeownership for younger families. As immigrants include: • Americans, and wealth gains in housing have home prices skyrocketed during the late Lack of capital for down payment and been more widely distributed among the 1970s and early 1980s, real incomes also closing costs; • population than gains in the stock market.4 stagnated, with earnings growth particularly Poor credit history; • With stocks appreciating faster than home slow for blue collar and less educated Lack of access to mainstream lenders; • prices over the past decade, home equity as workers. Through most of the 1980s, the Complexity and fear of the home buying a share of family assets fell from 38 percent combination of slow income growth and process; and, • in 1989 to 33 percent in 1998.5 Many of the increasing rents made saving for home Continued discrimination in housing gains in the stock market were erased after purchase more difficult, and relatively high markets and mortgage lending. 1999 however, and housing returned to its interest rates required large fractions of (i) Lack of Cash for Down Payment. In the place as the most significant asset in the family income for home mortgage payments. 2002 Fannie Mae National Housing Survey, household balance sheet in 2001.6 Even with Thus, during that period, fewer households 40 percent of Hispanics reported not having a bull market through most of the 1990s, 59 had the financial resources to meet down enough money for a down payment as an percent of all homeowners in 1998 held more payment requirements, closing costs, and obstacle to buying a home versus 32 percent than half of their net wealth in the form of monthly mortgage payments. of all Americans.16 A study by Gyourko, home equity.7 Among low-income Economic expansion and lower mortgage Linneman, and Wachter found significant homeowners (household income less than rates substantially improved homeownership racial differences in homeownership rates in $20,000), home equity accounted for about 72 affordability during the 1990s. Many young, ‘‘wealth-constrained’’ households while percent of household wealth, and low-income, and minority families who were finding no racial differences in approximately 55 percent for homeowners closed out of the housing market during the homeownership rates among households with incomes between $20,000 and $50,000. 1980s re-entered the housing market during with wealth sufficient to meet down payment Median net wealth for low-income the last decade. Even with an economic and closing costs.17 Minorities and slowdown in 2000–2001, improvements in immigrants are much less likely to receive gifts and inheritances from their parents to 2 Joint Center for Housing Studies of Harvard University, State of the Nation’s Housing 2003, 8 U.S Department of Housing and Urban assist them in becoming a homeowner. 2003, p. 16. Development, ‘‘Economic Benefits of Increasing (ii) Poor Credit History. Poor credit history 3 According to the National Association of Minority Homeownership,’’ p. 7. also differentially affects minority Realtors, Housing Market Will Change in New 9 Mark Zandi, ‘‘Housing’s Rising Contribution,’’ Millennium as Population Shifts, November 7, June 2002, p. 3. 14 Joint Center for Housing Studies of Harvard 1998. Forty-five percent of U.S. household wealth 10 Robert Dietz and Donald Haurin, ‘‘The Social University, State of the Nation’s Housing 2002, p. was in the form of home equity in 1998. Since 1968, and Private Consequences of Homeownership,’’ 14. home prices have increased each year, on average, May 2001, p. 51. 15 U.S. Census Bureau, Current Population at the rate of inflation plus two percentage points 11 William M. Rohe, George McCarthy, and Survey, March 2000. 4 Todd Buchholz, ‘‘Safe At Home: The New Role Shannon Van Zandt, ‘‘The Social Benefits and Costs 16 Fannie Mae, Fannie Mae National Housing of Housing in the U.S. Economy,’’ a paper of Homeownership,’’ May 2000, p. 31. Survey, 2002, p. 11. commissioned by the Homeownership Alliance, 12 U.S. Department of Housing and Urban 17 Joseph Gyourko, Peter Linneman, and Susan 2002. Development, ‘‘Economic Benefits of Increasing Wachter. ‘‘Analyzing the Relationships among Race, 5 Federal Reserve Board, ‘‘Recent Changes in U.S. Minority Homeownership,’’ p. 8–9. Wealth, and Home Ownership in America,’’ Journal Family Finances: Results from the 1998 Survey of 13 For a discussion of the causes of existing of Housing Economics 8 (2), p. 63–89, as discussed Consumer Finances,’’ January 2000, p. 15. disparities in homeownership, see the various in Thomas P. Boehm and Alan M. Schlottmann. 6 Mark Zandi, ‘‘Housing’s Rising Contribution,’’ articles in Nicolas P. Retsinas and Eric S. Belsky ‘‘Housing and Wealth Accumulation: June 2002, p. 5. (Eds), Low-Income Homeownership: Examining the Intergenerational Impacts,’’ in Low-Income 7 Joint Center for Housing Studies of Harvard Unexamined Goal, Washington, D.C.: Brookings Homeownership: Examining the Unexamined Goal, University, State of the Nation’s Housing 1998. Institution Press, 2002. Brookings Institution Press (2002), p. 408.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00043 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24270 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

households. In the same Fannie Mae survey, homeownership is fear and a lack of at unacceptable levels. The greatest share of nearly a third of African-American understanding about the buying process and discrimination for Hispanic and African- respondents said their credit rating would be the risks of ownership. Many Americans American home seekers can still be attributed an obstacle to buying a home versus 23 could become homeowners if provided with to being told units are unavailable when they percent of all Americans.18 Because African- information to correct myths, are available to non-Hispanic whites, and American and Hispanic borrowers are more misinformation, and concerns about the being shown and told about fewer units than likely than others to have little traditional mortgage process. Some potential comparable non-minority home seekers. credit history or a poorer credit history, they homeowners, particularly minorities, are Although discrimination is down on most face increased difficulties in being accepted unaware that they may already qualify for a areas for African-American and Hispanic for mortgage credit. This is because credit mortgage they can afford. The 2002 Fannie homebuyers, there remain worrisome upward history scores (such as a FICO score) are a Mae survey revealed that 30 percent of trends of discrimination in the areas of major component of the new automated Americans believe erroneously that they geographic steering for African Americans mortgage scoring systems. These systems are need to pay 20 percent of the cost of a home and, relative to non-Hispanic whites, the more likely to refer minority borrowers for up-front. In addition, Fannie Mae reported amount of help agents provide to Hispanics more intensive manual underwriting, rather that half of Americans are only ‘‘somewhat’’ with obtaining financing. On the rental side, than to automatically accept them for the less or ‘‘not at all’’ comfortable with mortgage Hispanics are more likely in 2000 than in costly, expedited processing. In these terms.20 Freddie Mac reports that six of 10 1989 to be quoted a higher rent than their situations, there is the additional concern Hispanics are uncomfortable with home white counterpart for the same unit. that ‘‘referred’’ borrowers may not always buying terminology, and think they need Another HUD-sponsored study asked receive a manual underwriting for the loan ‘‘perfect credit’’ to buy; and less than four in respondents to a nationwide survey if they that they initially applied for, but rather be 10 are aware that lenders are not required by ‘‘thought’’ they had ever been discriminated directed to a high-cost subprime loan law to give them the lowest interest rate against when trying to buy or rent a house product. 21 possible. A study using focus groups with or an apartment.24 While the responses were (iii) Lack of Access to Mainstream Lenders. renters found that even among those whose subjective, they are consistent with the Minorities face heightened barriers in financial status would make them capable of accessing credit because of their often limited findings of the HDS. African Americans and homeownership, many felt that the buying Hispanics were considerably more likely access to mainstream lenders. Access to process was insurmountable because they lenders becomes difficult when mainstream than whites to say they have suffered feared rejection by the lender or being taken discrimination—24 percent of African financial institutions are not located in advantage of.22 neighborhoods where minorities live. The Americans and 22 percent of Hispanics (v) Discrimination in the Housing and perceived discrimination, compared to only growth in subprime lending over the last Mortgage Markets. Finally, differential several years has benefited credit-impaired 13 percent of whites. treatment of minorities in the sales and rental Mortgage Lending Market. Research based borrowers—those who may have blemishes markets and in the mortgage lending market in their credit record, insufficient credit on Home Mortgage Disclosure Act (HMDA) has been well documented. The continued history, or non-traditional credit sources. data suggests pervasive and widespread discrimination in these markets is discussed Subprime lenders have allowed these disparities in mortgage lending across the in the next section. borrowers to access credit that they could not Nation. For 2001, the mortgage denial rate for otherwise obtain in the prime credit market. 2. Disparities in Housing and Mortgage white mortgage applicants was 23 percent, However, studies by HUD, The Woodstock Markets while 36 percent of African-American and 35 Institute and others have shown that percent of Hispanic applicants were denied. Sales and Rental Markets, In 2002, HUD subprime lending is disproportionately Two recent HUD-sponsored studies of released its third Housing Discrimination concentrated in low-income and minority paired-testing at the mortgage pre-application Study (HDS) in the sale and rental of neighborhoods.19 While these studies stage also points to discrimination by recognize that differences in credit behavior housing. The study, entitled Discrimination mortgage lenders. Based on its review of pair explain some of the disparities in subprime in Metropolitan Housing Markets: National tests conducted by the National Fair Housing lending across neighborhoods, they argue Results from Phase I of The Housing Alliance, the Urban Institute concluded that that the absence of mainstream lenders has Discrimination Study was conducted by the differential treatment discrimination at the 23 also contributed to the concentration of Urban Institute. The results of this HDS pre-application level occurred at significant subprime lending in low-income and were based on 4,600 paired tests of minority levels in at least some cities.25 Minorities minority neighborhoods. More competition and non-minority home seekers conducted were less likely to receive information about by prime lenders in inner city neighborhoods during 2000 in 23 metropolitan areas loan products, received less time and could lower the borrowing costs of families nationwide. The report showed large information from loan officers, and were who currently have only the option of a high- decreases between 1989 and 2000 in the level quoted higher interest rates in most of the cost subprime loan. This issue of the lack of of discrimination experienced by Hispanics cities where tests were conducted. A second mainstream lenders in inner city and African Americans seeking to buy a HUD-sponsored study by the Urban Institute neighborhoods is discussed further in home. There has also been a modest decrease used the paired testing methodology in Los subsection 2, below, in connection with in discrimination toward African Americans Angeles and Chicago and found similar disparities between neighborhoods. seeking to rent a unit. This downward trend, results. African Americans and Hispanics (iv) Complexity and Fear of Home Buying however, has not been seen for Hispanic faced a significant risk of unequal treatment Process. An additional barrier to renters, who now are more likely to when they visited mainstream mortgage experience discrimination in their housing lending institutions to make pre-application 18 Fannie Mae, Fannie Mae National Housing search than do African-American renters. But inquiries.26 Survey, 2002, p. 11. while generally down since 1989, the report 19 See Dan Immergluck, Stark Differences: The found that housing discrimination still exists 24 Martin D. Abravanel and Mary K. Cunningham, Explosion of the Subprime Industry and Racial How Much Do We Know? Public Awareness of the Hypersegmentation in Home Equity Lending. 20 Fannie Mae, Fannie Mae National Housing Nation’s Fair Housing Laws. A report prepared for Woodstock Institute, October 2000; and Daniel Survey, 2002, p. 9. HUD by the Urban Institute, Washington, DC, April Immergluck and Marti Wiles, Two Steps Back: The 21 Dual Mortgage Market, Predatory Lending, and the See ‘‘Immigration Changes Won’t Hurt 2002. Undoing of Community Development, Woodstock Housing,’’ in National Mortgage News, January 27, 25 Margery Austin Turner, John Yinger, Stephen Institute, Chicago, IL, November 1999. For a 2003, page 8. Ross, Kenneth Temkin, Diane Levy, David Levine, national analyses, see the HUD report Unequal 22 Donald S. Bradley and Peter Zorn, ‘‘Fear of Robin Ross Smith, and Michelle deLair, What We Burden: Income and Racial Disparities in Subprime Homebuying: Why Financially Able Households Know About Mortgage Lending Discrimination. The Lending in America, April 2000; and Randall M. May Avoid Ownership,’’ Secondary Mortgage Urban Institute, contract report for the Department Scheessele, Black and White Disparities in Markets, 1996. of Housing and Urban Development, December Subprime Mortgage Refinance Lending, Housing 23 Margery Austin Turner, Stephen L. Ross, 1998. Finance Working Paper No. HF–114, Office of George Galster, and John Yinger, ‘‘Discrimination in 26 Margery Austin Turner, All Other Things Being Policy Development and Research, U.S. Department Metropolitan Housing Markets,’’ The Urban Equal: A Paired Testing Study of Mortgage Lending of Housing and Urban Development, April 2002. Institute Press, November 2002. Institutions, The Urban Institute Press, April 2002.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00044 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24271

Several possible explanations for these These geographical disparities can be the a. Problems Faced by Owners lending disparities have been suggested. A result of cost factors, such as the difficulty of Of the 68.8 million owner households in study by the Boston Federal Reserve Bank appraising houses in these areas because of 1999, 5.8 million (8 percent) confronted a found that racial disparities cannot be the paucity of previous sales of comparable severe cost burden and another 8.7 million explained by reported differences in homes. Sales of comparable homes may also (12.7 percent) faced a moderate cost burden. creditworthiness.27 In other words, be difficult to find due to the diversity of There were 870,000 households with severe minorities are more likely to be denied than central city neighborhoods. The small loans physical problems, 2 million with moderate whites with similar credit characteristics, prevalent in low-income areas are less physical problems and 905,000 that were which suggests lender discrimination. In profitable to lenders because up-front fees to overcrowded. The report found that 25 addition, loan officers, who may believe that loan originators are frequently based on a percent of American homeowners faced at race is correlated with credit risk, may use percentage of the loan amount, although the least one severe or moderate problem. race as a screening device to save time, rather costs incurred are relatively fixed. As noted Not surprisingly, problems were most than devote effort to distinguishing the above, racial disparities in mortgage access common among very low-income owners.31 creditworthiness of the individual may be due to the fact that mainstream Almost a third of these households (31 applicant.28 This violates the Fair Housing lenders are not doing business in certain percent) faced a severe cost burden, and an Act. inner city neighborhoods. There is evidence additional 22 percent faced a moderate cost Underwriting rigidities may fail to that mainstream lenders active in white and burden. And 8 percent of these families lived accommodate creditworthy low-income or upper-income neighborhoods are much less in severely or moderately inadequate minority applicants. For example, under housing, while 2 percent faced overcrowding. active in low-income and minority traditional underwriting procedures, Only 42 percent of very-low-income owners neighborhoods—often leaving these applicants who have conscientiously paid reported no problems. rent and utility bills on time but have never neighborhoods to unregulated subprime Over time the percentage of owners faced used consumer credit would be penalized for lenders. Geographical disparities in mortgage with severe or moderate physical problems having no credit record. Applicants who lending are discussed further in Section C.8 has decreased, as has the portion living in have remained steadily employed, but have below (which examines subprime lending) overcrowded conditions. However, changed jobs frequently, would also be and in Appendix B (which examines the affordability problems have become more penalized. As discussed in Section C below, Underserved Areas Goal). common—the shares facing severe lenders, private mortgage insurers, and the 3. Affordability Problems and Worst Case (moderate) cost burdens were only 3 percent GSEs have been adjusting their underwriting Housing Needs (5 percent) in 1978, but rose to 5 percent (11 guidelines to take into account these special percent) in 1989 and 8 percent (13 percent) circumstances of lower-income families. The severe affordability problems faced by in 1999. The increase in affordability Many of the changes recently undertaken by low-income homeowners and renters are problems apparently reflects a rise in the industry focused on finding alternative documented in HUD’s ‘‘Worst Case Housing mortgage debt in the late 1980s and early underwriting guidelines to establish Needs’’ reports. These reports, which are 1990s, from 21 percent of homeowners’ creditworthiness that do not disadvantage prepared biennially for Congress, are based equity in 1983 to 36 percent in 1995.32 The creditworthy minority or low-income on the American Housing Survey (AHS), Joint Center for Housing Studies also applicants. However, because of the conducted every two years by the Census attributes this to the growing gap between enhanced roles of credit scoring and Bureau for HUD. The latest detailed report housing costs and the incomes of the nation’s automated underwriting in the mortgage analyzes data from the 1999 AHS. Although poorest households.33 As a result of the origination process, it is unclear to what it focuses on the housing problems faced by increased incidence of severe and moderate degree the reduced rigidity in industry very-low-income renters, it also presents cost burdens, the share of owners reporting standards will benefit borrowers who have basic data on families and households in no problems fell from 84 percent in 1978 to been adversely impacted by the traditional owner-occupied housing.30 78 percent in 1989 and 75 percent in 1999. guidelines as discussed in section C.7, some The ‘‘Worst Case’’ report measures three b. Problems Faced by Renters types of problems faced by homeowners and industry observers have expressed a concern Problems of all three types listed above are that the greater flexibility in the industry’s renters: more common among renters than among written underwriting guidelines may not be 1. Cost or rent burdens where housing homeowners. In 1999 there were 6.3 million reflected in the numerical credit and costs or rent exceed 50 percent of income (a renter households (19 percent of all renters) mortgage scores which play a major role in ‘‘severe burden’’) or range from 31 percent to who paid more than 50 percent of their the automated underwriting systems that the 50 percent of income (a ‘‘moderate burden’’); income for rent.34 Another 7.1 million faced GSEs and others have developed. 2. The presence of physical problems a moderate rent burden. Thus in total 40 Disparities Between Neighborhoods. involving plumbing, heating, maintenance, percent of renters paid more than 30 percent Mortgage credit also appears to be less hallway, or the electrical system, which may of their income for rent. accessible in low-income and high-minority lead to a classification of a residence as Among very-low-income renters, 71 neighborhoods. As discussed in Appendix B, ‘‘severely inadequate’’ or ‘‘moderately percent faced an affordability problem, 2001 HMDA data show that mortgage denial inadequate;’’ and, including 40 percent who paid more than rates are nearly twice as high in census tracts 3. Crowded housing, where there is more half of their income in rent. Almost one-third with low-income and/or high-minority than one person per room in a residence. (31 percent) of renters with incomes between composition, as in other tracts (16.8 percent The study reveals that in 1999, 4.9 million 51 percent and 80 percent of area median versus 8.7 percent). Numerous studies have households had ‘‘worst case’’ housing needs, found that mortgage denial rates are higher defined as housing costs greater than 50 31 Very-low-income households are defined as in low-income census tracts, even accounting percent of household income or severely those whose income, adjusted for household size, 29 for other loan and borrower characteristics. inadequate housing among unassisted very- does not exceed 50 percent of HUD-adjusted area low-income renter households. Among the 34 median income. This differs from the definition 27 Alicia H. Munnell, Geoffrey M.B. Tootell, Lynn million renters in all income categories, 6.3 adopted by Congress in the GSE Act of 1992, which E. Browne, and James McEneaney, ‘‘Mortgage million (19 percent) had a severe rent burden uses a cutoff of 60 percent and which does not Lending in Boston: Interpreting HMDA Data,’’ and over one million renters (3 percent) lived adjust income for family size for owner-occupied dwelling units. American Economic Review, 86, March 1996. in housing that was severely inadequate. 28 See Charles W. Calomeris, Charles M. Kahn and 32 Edward N. Wolff, ‘‘Recent Trends in the Size Stanley D. Longhofer, ‘‘Housing Finance Distribution of Household Wealth,’’ The Journal of Intervention and Private Incentives; Helping Evidence from HMDA, Working Paper Series 94–21, Economic Perspectives, 12(3), (Summer 1998), p. Minorities and the Poor,’’ Journal of Money, Credit Federal Reserve Bank of Cleveland, December 1994. 137. and Banking, 26, August 1994, pp. 634–74, for more 30 HUD has published an update on ‘‘worst case 33 Joint Center for Housing Studies, The State of discussion of this phenomenon, which is called housing needs,’’ which found that the number of the Nation’s Housing: 2000, June 2000, p. 24. ‘‘statistical discrimination’’ such households rose from 4.86 million in 1999 to 34 Rent is measured in this report as gross rent, 29 Robert B. Avery, Patricia E. Beeson and Mark 5.07 million in 2001. However, detailed tables for defined as contract rent plus the cost of any utilities E. Sniderman, Understanding Mortgage Markets: 2001 have not been published. that are not included in contract rent.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00045 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24272 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

family income also paid more than 30 multifamily markets, as well as in record high in 2001 and again in 2002. New percent of their income for rent. Appendices B–D. The discussion will cover home sales, which increased an average 6.3 Affordability problems have increased over a wide range of topics, such as subprime percent per year between 1992 and 2002, time among renters. The shares of renters lending, predatory lending, automated reached a record high of 976,000 units in with severe or moderate rent burdens rose underwriting systems, manufactured 2002, an increase of 7.5 percent over 2001 from 32 percent in 1978 to 36 percent in 1989 housing, the special needs of the single- sales. The market for new homes has been and 40 percent in 1999. family rental market, and challenges strong throughout the nation. The share of households living in associated with producing affordable The National Association of Realtors inadequate housing in 1999 was higher for multifamily housing—just to name a few. reported that nearly 5.6 million existing renters (11 percent) than for owners (4 homes were sold in 2002, overturning the old percent), as was the share living in C. Factor 2: Economic, Housing, and record set in 2001 by 5 percent, and setting overcrowded housing (5 percent for renters, Demographic Conditions: Single-Family an all-time high in the 34-year history of the but only 1 percent for owners). Crowding and Mortgage Market series. Sales of existing homes reached record inadequate housing were more common This section discusses economic, housing, levels in three of the four major regions of the among lower-income renters, but among even and demographic conditions that affect the nation and came within 96 percent of the the lowest income group, affordability was single-family mortgage market. After a review record in the Northeast in 2001. Combined the dominant problem. The prevalence of of housing trends and underlying new and existing home sales also set a inadequate and crowded rental housing demographic conditions that influence national record of 6.2 million last year. diminished over time until 1995, while homeownership, the discussion focuses on One of the strongest sectors of the housing affordability problems grew. specific issues related to the single-family market in past years had been manufactured Other problems faced by renters discussed owner mortgage market. This subsection homes, but that sector has declined recently. in the most recent detailed ‘‘Worst Case’’ includes descriptions of recent market Between 1991 and 1996, manufactured home report include a sharp decline (of 2.3 million, interest rate trends, refinance and home shipments more than doubled, peaking in or 14 percent) between 1991 and 1999 in the purchase activity, homebuyer characteristics, 1998 at 373,000. However, shipments fell number of rental units affordable to very-low- and the state of affordable lending. Other more than 20 percent in both 2000 and 2001. income families, and a worsening of the special topics examined include the growth In 2002, the industry shipped 169,000 new manufactured homes, down 12.4 percent national shortage of units affordable and in subprime lending, the increased use of from 2001. This was the lowest number of available to extremely-low-income families automated underwriting, and the remaining manufactured home shipments since 1963. (those with incomes below 30 percent of area homeownership potential among existing Homeownership Rate. In 1980, 65.6 median income). Shortages of units renters. Section D follows with a discussion percent of Americans owned their own affordable and available to extremely-low- of the economic, housing, and demographic home, but due to the unsettled economic income households were most pressing in the conditions affecting the mortgage market for conditions of the 1980s, this share fell to 63.8 West and Northeast, especially in multifamily rental properties. percent by 1989. But since 1994, gains in the metropolitan areas in those regions. 1. Recent Trends in the Housing Market homeownership rate have occurred in each year, with the rate reaching another record 4. Rehabilitation and Other National Housing While most other sectors of the economy Needs mark of 67.9 percent in 2002. The number of were weak or declining during 2001 and households owning their own home in 2002 In addition to the broad housing needs 2002, the housing sector showed remarkable was 10.6 million greater than in 1994. discussed above, there are additional needs strength. Despite the recession in 2001, Gains in homeownership have been confronting specific sectors of the housing factors such as record-low interest rates and widespread over the last eight years.35 As a and mortgage markets. One example of these continued price stability contributed to a result, the homeownership rate rose from: specific needs concerns the rehabilitation of record year in the housing market. In 2002, • 42.0 percent in 1993 to 47.9 percent in the nation’s older housing stock. A major the U.S. economy moved into recovery with 2002 for African-American households, problem facing lower-income households is real GDP growing 2.4 percent. In October • 39.4 percent in 1993 to 48.2 percent in that low-cost housing units continue to 2002, the 30-year home mortgage rate slipped 2002 for Hispanic households, disappear from the existing housing stock. below 6 percent for the first time since the • 73.7 percent in 1993 to 78.9 percent in Older properties are in need of upgrading mid-1960s. Favorable financing conditions 2002 for married couples with children, and rehabilitation. These aging properties are and solid increases in house prices were the • 65.1 percent in 1993 to 68.6 percent in concentrated in central cities and older inner key supports to another record housing 2002 for household heads aged 35–44, and suburbs, and they include not only detached market during 2002. In fact, the year 2002 • 48.9 percent in 1993 to 51.8 percent in single-family homes, but also small was among the strongest years experienced 2002 for central city residents. multifamily properties that have begun to by the housing industry. By the end of 2002 However, as these figures demonstrate, deteriorate. But obtaining the funds to fix up the industry set many new records in single- sizable gaps in homeownership remain. older properties can be difficult. The owners family permits, new home sales, existing Economy/Housing Market Prospects. The of small rental properties in need of home sales, interest rates, and economy grew at a rate of 2.2 percent in 2002 rehabilitation may be unsophisticated in homeownership. Other indicators—total and was less robust than in past U.S. obtaining financing. The properties are often permits, starts, completions, and recoveries.36 In response, the Federal Reserve occupied, and this can complicate the affordability—reached levels that were has lowered interest rates to record lows, rehabilitation process. Lenders may be among the highest in the past two decades. supporting housing affordability. reluctant to extend credit because of a Single-Family Permits, Starts, and The Blue Chip consensus forecast for real sometimes-inaccurate perception of high Completions. Builders took out 1,319,100 GDP growth is 4.2 percent for 2004.37 The credit risk involved in such loans. The GSEs single-family permits in 2002, up 6.8 percent Congressional Budget Office (CBO) 38 projects and other market participants have recently from 2001. The 2002 level was the highest begun to pay more attention to these needs number of single-family permits ever 35 Homeownership rates prior to 1993 are not for financing of affordable rental housing reported in the 43-year history of this series. strictly comparable with those beginning in 1993 rehabilitation. However, extra effort is Single-family starts totaled 1,359,700 housing because of a change in weights from the 1980 required, due to the complexities of units, up 6.8 percent from 2001, and the Census to the 1990 Census. rehabilitation financing, as there is still a highest number of single-family starts since 36 National Association of Realtors, ‘‘Near Record need to do more. 1978. Construction was completed on Home Sales Projected for 2003,’’ December 3, 2002. 37 The rehabilitation of our aging housing 1,328,400 single-family housing units, up 5.8 Blue Chip Economic Indicators, Vol. 28, No. 11, November 10, 2003. stock is but one example of the housing and percent from 2001. This is the highest 38 mortgage issues that need to be addressed. number of single-family completions in 24 Real GDP, unemployment, inflation, and treasury note interest rate projections are obtained Several other examples will be provided years. for fiscal years 2003–2013 from The Budget and throughout the following sections on the Sales of New and Existing Homes. After Economic Outlook: An Update, Washington, DC economic, housing, and demographic leveling out in 2000, housing sales have Congressional Budget Office. (August 2003). conditions in the single-family and boomed in the past two years, reaching a http://www.cbo.gov/showdoc.cfm.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00046 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24273

that real GDP will grow at an average rate of As explained below, the role of traditional more affordable. While this cohort has 3.3 percent from 2005 through 2008, down first-time homebuyers, 25-to-34 year-old achieved a homeownership rate equal to the from their forecasted rate of 3.8 percent in married couples, in the housing market will middle baby-boomers, they live in larger, 2004. Inflation, as measured by the Consumer be smaller in the current decade due to the more expensive homes. As the baby-boom Price Index (CPI), is projected to remain aging of the population. For the first time in generation ages, demand for housing from modest during the same period, averaging 2.5 history, the population will have roughly this group is expected to wind down.49 percent. The unemployment rate is expected equal numbers of people in every age group. The baby-boom generation was followed by to ease from 2003–2004 levels, averaging 5.4 Between 2000 and 2025, the Census Bureau the baby-bust generation, from 1965 through percent over the forecast period. The projects that the largest growth in households 1977. Since this population cohort is smaller remainder of this subsection focuses on will occur among householders 65 and than that of the baby boom generation, it future prospects for the housing market. over.46 Thus, an increasing percentage of the reduced housing demand in the preceding Fannie Mae expects existing home sales to population will be past their homebuying decade and is expected to do the same in the reach a record level of 6 million in 2003 and peak in the next two decades. However, current decade, though, as discussed below, decline only slightly to 5.7 million in 2004 because homeownership rates do not peak other factors kept the housing market very and 2005.39 Projected at 1.84 million in 2003, until population groups reach 65 to 74 years strong in the 1990s. However, the echo baby- the National Association of Home Builders of age, this age cohort will continue to boom generation (the children of the baby- expects housing starts to decline to 1.77 provide housing demand. According to boomers, who were born after 1977), while million in 2004 and 1.71 million in 2005.40 Riche, the increasing presence of older smaller than the baby-boom generation, will The Mortgage Bankers Association forecasts households should increase the proportion of reach peak home buying age later in the first that 2004 housing starts will total 1.73 the population that owns, rather than rents decade of the millennium. million units and the 30-year fixed mortgage housing.47 Immigrant Homebuyers. Past, present, and rate will average 6.1 percent.41 After more Growing housing demand from immigrants future immigration will also contribute to than doubling from a relative trough in 2000 and non-traditional homebuyers will help to gains in the homeownership rate. During the to an estimated $2.6 trillion in 2002, Fannie offset declines in the demand for housing 1990s, 9.8 million legal immigrants entered Mae forecasts that mortgage originations will caused by the aging of the population. the United States, as compared to 6.3 million rise to a record high $3.7 trillion in 2003 Riche’s study estimates that minorities will entering in the 1980s and 4.2 million during before dropping to $1.8 trillion in 2004 and account for two-thirds of the growth in U.S. the 1970s. Overall, the increase in the 42 $1.5 trillion in 2005. households over the next 25 years, and by immigrant population directly accounted for 2. Underlying Demographic Conditions 2025, non-family households will make up a 35 percent of the nation’s rise in population third of all households. The ‘‘echo baby- in the 1990s.50 As a result, the foreign-born Between 2000 and 2025, the U.S. boom’’ (that is, children of the baby-boomers) population of the United States more than population is expected to grow by an average will also add to housing demand in the tripled from 9.6 million in 1970 to 31.1 of 2.5 million per year.43 This will likely current and next decades. Finally, the million in 2000. Immigrants who become result in 1.1 million new households per growing income inequality between people citizens buy homes at rates nearly as high as year, increasing the number of households 26 with and without a post-secondary education their same-aged native-born counterparts. percent in the period, and creating a will continue to affect the housing market. Moreover, U.S.-born children of immigrants continuing need for additional housing.44 The Baby-Boom Effect. The demand for often have higher homeownership rates than This section discusses important housing during the 1980s and 1990s was the same-age children of native-born demographic trends behind these overall driven, in large part, by the coming of parents.51 However, there are concerns about household numbers that will likely affect homebuying age of the baby-boom the assimilation into homeownership of housing demand in the future. These generation, those born between 1945 and recent Hispanic immigrants who are less demographic forces include the baby-boom, educated than earlier cohorts of immigrants. baby-bust and echo baby-boom cycles; 1964. Homeownership rates for the oldest of the baby-boom generation, those born in the Many immigrants also locate in high-priced immigration trends; non-traditional and housing markets, which makes it more single households; ‘‘trade-up buyers;’’ and 1940s, rival those of the generation born in the 1930s. Due to significant house price difficult for them to achieve homeownership. the growing income inequality between Although net foreign immigration is people with different levels of education. appreciation in the late-1970s and 1980s, older baby-boomers have seen significant projected to decline in the current decade HUD’s Office of Policy Development and after 2002, high levels of immigration in the Research funded a study, Issue Papers on gains in their home equity and subsequently have been able to afford larger, more late 1980s and throughout the 1990s will Demographic Trends Important to Housing, have lasting positive effects on housing which analyzes effects of demographic expensive homes. Circumstances were not so favorable for the middle baby-boomers. demand. New immigration in the current and conditions on the housing market. The next decades is projected to create 6.9 findings are presented throughout the Housing was not very affordable during the 1980s, their peak homebuying age period. As million net new households, but the majority sections that follow.45 a result, the homeownership rate, as well as of household growth in the period (16.9 million) will come from people already 39 wealth accumulation, for the group of people Fannie Mae, ‘‘Berson’s Economic and Mortgage born in the 1950s lags that of the generations resident in the U.S. including the foreign- Market Development Outlook,’’ December 2003. 48 born population.52 While immigrants tend to http://www.fanniemae.com/media/pdd/berson/ before them. monthly2003/121203.pdf. As the youngest of the baby-boomers (those rent their first homes upon arriving in the 40 http://www.nahb.org. born in the 1960s) reached their peak home United States, homeownership rates are 41 Mortgage Bankers Association of America, buying years in the 1990s, housing became substantial for those that have lived here for Mortgage Finance Forecast, December 17, 2003. at least 6 years. In 1996, the homeownership http://www.mbaa.org/marketdata/forecasts/ Papers on Demographic Trends Important to mffore1103.pdf. Housing. Urban Institute Final Report to the Office 49 Ibid. p. 15. 42 Fannie Mae, ‘‘Berson’s Economic and Mortgage of Policy Development and Research, U.S. 50 Federation for American Immigration Reform, Market Development Outlook,’’ December 2003. Department of Housing and Urban Development, , 43 U.S. Census Bureau, Population Projections September 2002. site visited December 13, 2002. Table NP–T1. 46 Martha Farnsworth Riche, ‘‘How Changes in 51 Joint Center for Housing Studies of Harvard 44 Martha Farnsworth Riche, ‘‘How Changes in the Nation’s Age and Household Structure Will University, State of the Nation’s Housing 2002, pp. the Nation’s Age and Household Structure Will Reshape Housing Demand in the 21st Century,’’ in 16–17. Reshape Housing Demand in the 21st Century,’’ in Issue Papers on Demographic Trends Important to 52 George S. Masnick and Zhu Xiao Di, Issue Papers on Demographic Trends Important to Housing. Urban Institute Final Report to the U.S. ‘‘Projections of U.S. Households By Race/Hispanic Housing, Urban Institute Final Report to the Office Department of Housing and Urban Development, Origin, Age, Family, Type, and Tenure to 2020: A of Policy Development and Research, U.S. September 2002, p. 4. Sensitivity Analysis,’’ in Issue Papers on Department of Housing and Urban Development, 47 Ibid. p. 6. Demographic Trends Important to Housing. Urban September 2002, p. 5. 48 Joint Center for Housing Studies of Harvard Institute Final Report to the U.S. Department of 45 Barry Chiswick, Paul Miller, George Masnick, University, State of the Nation’s Housing 1998, p. Housing and Urban Development, September 2002, Zhu Xiao Di, and Martha Farnsworth Riche, Issue 14. p. 5.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00047 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24274 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

rate for recent immigrants was 14.7 percent potential to ‘‘trade up’’ to more expensive likely be sustained growth and an while it was 66.9 percent for foreign-born housing.56 increasingly diverse household population naturalized citizens after six years.53 Higher- Growing Income Inequality. The Census from which to draw new homeowners. There than-average foreign-born fertility rates and Bureau recently reported that the top 5 are continuing concerns about the increasing high rates of homeownership for immigrants percent of American households received income inequality of our population and living in the country for several years and 22.4 percent of aggregate household income those recent immigrants and other persons among the children of immigrants suggest in 2001, up from 21.4 percent in 1998 and who have limited education. that past immigration will continue to create up sharply from 16.1 percent in 1977. The share accruing to the lowest 80 percent of 3. Basic Trends in the Single-Family housing demand. Mortgage Market Past and future immigration will lead to households fell from 56.5 percent in 1977 to increasing racial and ethnic diversity, 50.8 percent in 1998 and again to 49.8 Mortgage lending in the nation is growing especially among the young adult percent in 2001. The share of aggregate at unprecedented levels. Residential population. As immigrant minorities account income accruing to households between the mortgage originations soared to $2.5 trillion for a growing share of first-time homebuyers 80th and 95th percentiles of the income in 2002, a 22 percent increase over the in many markets, HUD and others will have distribution was virtually unchanged from previous record of $2.06 trillion set in 57 2001.60 This boom in lending can be to intensify their focus on removing 1977 to 2001. attributed to low mortgage interest rates and discrimination from the housing and The increase in income inequality over a record number of refinances. mortgage finance systems. The need to meet past decades has been especially significant Approximately 40 percent of mortgage debt nontraditional credit needs, respond to between those with and those without post- secondary education. The Census Bureau outstanding, or $2.5 trillion, was refinanced diverse housing preferences, and overcome reports that by 1999, the annual earnings of during the 2001–02 refinance boom. The last the information barriers that many workers with a bachelor’s degree were 1.8 refinancing record was set in 1998 when immigrants face will take on added times the annual earnings of workers with a roughly 20 percent of mortgage debt importance. In order to address these needs, high school education.58 The inflation- outstanding was refinanced.61 This section the mortgage industry must offer innovative adjusted median earnings of high school focuses on recent interest rate trends, the products and improve outreach efforts to graduates were at the same level in 2001 as refinance market, the home purchase market, attract minority homebuyers. in 1991 while the earnings of bachelor and first-time homebuyers. The section Nontraditional and Single Homebuyers. degree-holders rose nearly 9 percent over the concludes by examining the GSEs’ While overall growth in new households has same period.59 acquisitions as a share of the primary single- slowed down, nontraditional households So, while homeownership is highly family mortgage market, and provides have become more important in the affordable, those without post-secondary mortgage market prospects. homebuyer market. As the population ages education often lack the financial resources a. Mortgage Characteristics both relatively and absolutely, the nation’s to take advantage of the opportunity. As households will become smaller and more discussed earlier, the days of the well-paying Interest Rate Trends and Volatility. diverse. Riche notes that in 2000, traditional unionized factory job have passed. They have Historically low mortgage interest rates in the family households represented fewer than given way to technological change that favors late 1990s and 2001–2003 helped maintain one in four households and were surpassed white-collar jobs requiring college degrees, consumer confidence in the housing sector as by both single-person households and and wages in the manufacturing jobs that the economy emerged from its first recession married couples without children. With later remain are experiencing downward pressures in almost a decade. After high and marriages and more divorces, single-parent from economic globalization. The effect of fluctuating mortgage rates in the 1980s and and single-person households have increased this is that workers without the benefit of a early 1990s, recent years have seen a period rapidly. In fact, single-parent households post-secondary education find their demand of lower and more stable rates. The 1980s grew from 4 percent of family households in for housing constrained. This is especially began with interest rates on mortgages for 1950 to 12 percent in 2000. Single-person problematic for recent immigrants who are new homes above 12 percent but quickly rose 62 households are now the nation’s second most more likely to have limited educational to more than 15 percent. By 1987–88, rates numerous household type, accounting for attainment and English language proficiency. dipped into single digits but were rising over 25 percent of all households. In the Summary. Over the next two-and-a-half again by 1989–90. Rates declined in the early future, longer life expectancies and the decades, the number of U.S. households is 1990s, reaching a low of 6.8 percent in late continuing preference for one or two children projected to increase by nearly 27 million. Of 1993. An upturn in rates in 1994 and 1995 will make households without children even these new households, non-Hispanic white peaked at 8.3 percent in early 1995. By 1998, more numerous. Projected to compose 80 and traditional households will contribute 30-year fixed conventional mortgages percent of all households by 2025, only one-third and one-tenth of the growth, averaged 6.95 percent, the lowest level since nontraditional family households will play respectively. As the baby-boomers aged out 1968 but saw a rise in 1999 to 7.44 percent. Mortgage rates then continued to rise in an increasingly important role in the housing of their peak home buying stage and the 2000, averaging 8.05 percent for the year, market.54 baby-bust generation aged into their peak before falling to a low of 6.62 percent in Trade-up Buyers. Due to weak house price home buying stage in the late 1980s, demand October 2001 and averaging 6.97 percent for appreciation, traditional ‘‘trade-up buyers’’ for housing was dampened by demographic 2001 as a whole.63 Rates averaged 6.54 stayed out of the market during the early factors during the 1990s. (Of course, other percent during 2002, reaching a low of 6.05 1990s. Their absence may explain, in part, factors such as low interest rates propelled the large representation of nontraditional the housing market to record levels during 60 homebuyers during that period. However, this period.) As the echo baby-boomers begin ‘‘Mortgage Originations Hit Record-Busting to enter their peak home buying age, housing $2.5 Trillion in 2002, IMF Numbers Reveal,’’ Inside since 1995 home prices have increased more Mortgage Finance, January 24, 2003, p. 3. 55 demand should pick up again through the than 30 percent. The greater equity 61 Economy.com, ‘‘The Economic Contribution of resulting from recent increases in home remainder of the current decade and into the next. As these demographic factors play out, the Mortgage Refinancing Boom,’’ December 2002, prices should lead to a larger role for ‘‘trade- the overall effect on housing demand will p. 2. up buyers’’ in the housing market during the 62 Interest rates in this section are effective rates next 10 to 15 years. In addition, the growing paid on conventional home purchase mortgages on number of higher-income, mid-life 56 Riche, 2002, p.17. new homes, based on the Monthly Interested Rate 57 households will increase households’ All data in this paragraph are from the U.S. Survey (MIRS) conducted by the Federal Housing Census Bureau’s Historical Income Table H2. Finance Board and published by the Council of 58 Jennifer Cheeseman Day and Eric C. Economic Advisers annually in the Economic 53 Fred Flick and Kate Anderson, ‘‘Future of Newburger, The Big Payoff: Educational Attainment Report of the President and monthly in Economic Housing Demand: Special Markets,’’ Real Estate and Synthetic Estimates of Work-Life Earnings, U.S. Indicators. These are average rates for all loan types, Outlook, 1998, p. 6. Bureau of the Census, Current Population Reports encompassing 30-year and 15-year fixed-rate 54 Riche, 2002, p. 1. P23–210, July 2002, p.3. mortgages and adjustable rate mortgages. 55 Average new-home price: U.S. Census Bureau, 59 U.S. Census Bureau, Historical Income Table 63 U.S. Housing Market Conditions, 2nd Quarter H13. 2002, August 2002, Table 14.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00048 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24275

percent in December of that year. Falling b. Refinance Mortgages these conditions, the after-tax cost saving on further to 5.23 in June of 2003, mortgage Refinancing has fueled the growth in total a new, lower-rate loan is much greater than interest rates remained low throughout last mortgage originations, which were $638 the transaction costs of refinancing. In 64 year, averaging 5.79 through September. billion in 1995 (a period of low refinance addition, the appreciation of housing prices Other Loan Terms. When mortgage rates activity), but topped $2.5 trillion in 2002 (a has also contributed to the increase in are low, most homebuyers prefer to lock in period of heavy refinance activity). The refinancing. Over the past five years, the a fixed-rate mortgage (FRM). Adjustable-rate refinance share of total mortgage originations value of housing rose by approximately $5 mortgages (ARMs) are more attractive when rose to 50 percent in 1998, then decreased to rates are high, because they carry lower rates trillion, and the rise in value has enabled 19 percent in 2000 before jumping to 57 lenders to service refinancing homeowners than FRMs and because buyers may hope to percent in 2001.67 Over the past ten years, refinance to a FRM when mortgage rates because of greater confidence in the refinance booms occurred three times, during 71 decline. The Federal Housing Finance Board 1992–93, 1998, and 2001–02. During the creditworthiness of borrowers. (FHFB) reports that the ARM share of the 2001–02 refinance boom, approximately 40 Over the past few years, homeowners have market fell from 20 percent in 1993 to a percent of the $2.5 trillion in mortgage debt become more willing to draw on the rising record low of 12 percent in 1998, before outstanding was refinanced. The last equity in their homes. According to Fannie rising back to 21 percent in 1999. The ARM refinancing record was set in 1998 when Mae’s 2002 National Housing Survey, share continued to rise to 24 percent in 2000, roughly 20 percent of mortgage debt homeowners that refinanced during 2001 but then fell dramatically to a low of 12 outstanding was refinanced.68 withdrew about $110 billion in accumulated percent in 2001 as mortgage rates decreased. In 1989–90 interest rates exceeded 10 home equity wealth.72 Freddie Mac estimates In 2001, the term-to-maturity was 30 years percent, and refinancings accounted for less that more than one-half of all refinance for 83 percent of conventional home than 25 percent of total mortgage mortgages in the past two years involved purchase mortgages, after steadily climbing originations.69 The subsequent sharp decline cash-out refinancing.73 to a high of 90 percent in 2000. The other in mortgage rates drove the refinance share maturities in 2001 included 15 years (13 The refinancing boom contributed to an over 50 percent in 1992 and 1993 and estimated one-fifth of the national economy’s percent), 20 years (3 percent), and 25 years propelled total single-family originations to real GDP growth since late 2000.74 During (1 percent). more than $1 trillion in 1993—twice the level 2001 and 2002, roughly $270 billion was Low- and no-point mortgages continue to attained just three years earlier. be a popular option for mortgage purchases. The refinance wave subsided after 1993, raised in cash-out refinancing. FHFB reports that average initial fees and because most homeowners who found it Approximately one-half of cash from cash- charges (‘‘points’’) have decreased from 2.5 beneficial to refinance had already done so out refinancing has enabled consumers to percent of loan balance in the mid-1980s to and because mortgage rates rose once again.70 finance more spending for expenses such as 2 percent in the late-1980s, 1.5 percent in the Total single-family mortgage originations home improvements, medical payments, early 1990s, and less than 1 percent in 1995– bottomed out at $638 billion in 1995, when education, and vehicles during a weakened 97. The downward trend continued the refinance share was only 21 percent. economy. Roughly one-third of the cash from throughout the late 1990s with the average Total originations, driven by the volume of cash-out refinancing has allowed consumers initial fees and charges reaching a low of one- refinancings, amounted to $1.507 trillion in to repay other debt.75 The remaining cash half percent in 2001. Coupled with declining 1998, nearly 50 percent higher than the from cash-out refinancing has enabled interest rates, these lower transactions costs previous record level of $1.02 trillion consumers to invest in other assets. have increased the propensity of attained in 1993. homeowners to refinance their mortgages.65 Refinancing households save approximately The refinance wave from late 1997 through $10 billion in their annual interest payments Another major change in the conventional early 1999 reflected other factors besides on their mortgage and consumer installment home mortgage market has been the interest rates, including greater borrower liabilities. proliferation of high loan-to-value ratio (LTV) awareness of the benefits of refinancing, a mortgages. According to data from the highly competitive mortgage market, and the Although the refinancing boom may Federal Housing Finance Board, loans with enhanced ability of the mortgage industry, quickly fade if mortgage rates rise in 2004, LTVs greater than 90 percent (that is, down utilizing automated underwriting and the boom will have lingering effects. payments of less than 10 percent) made up mortgage origination systems to handle an Mortgage borrowers that were able to secure less than 10 percent of the market in 1989– unprecedented volume of originations. The low long-term interest rates through fixed 91, but 25 percent of the market in 1994–97, refinance share decreased to 19 percent in rate mortgages will have more of their gradually decreasing to an average of 21 2000 before jumping to a record 57 percent budgets to spend on other items. Meanwhile, percent of the market in 2001. Loans with in 2001. cash-out borrowers, who are just receiving LTVs less than or equal to 80 percent fell Historically low interest rates and their money, will spend this year. It must be from three-quarters of the market in 1989–91 declining mortgage transaction costs have noted there is some concern regarding the to an average of 56 percent of the market in driven the latest refinancing boom. Given 1994–97, but then rose to an average of 63 potential for increased credit risk stemming percent of mortgages originated in 1998– from mortgage debt from cash out borrowers. separate FHA-insured loans from conventional 2001. As a result, the average LTV rose from According to a 2002 Regional Finance mortgages. In addition, the statistics cited above Review article, the mortgage liabilities of 75 percent in 1989–91 to nearly 80 percent pertain only to home purchase mortgages. in 1994–97, and then declined to 76.2 Refinance mortgages generally have shorter terms households have been growing at a rate more percent in 2001.66 and lower loan-to-value ratios than home purchase than double the growth in household mortgages. incomes. However, this potential credit risk 64 Mortgage Bankers Association website. MBA 67 The source for the refinance share and total is moderated by the strong growth in housing Weekly Survey of Mortgage Applications, Monthly mortgage originations was the Mortgage Bankers values. The ratio of mortgage debt to housing Average Interest Rates on 30-Year Fixed-Rate Association. Mortgages. http://www.mortgagebankers.org/ 68 Economy.com, ‘‘The Economic Contribution of marketdata/index.html. the Mortgage Refinancing Boom,’’ December 2002, 71 Economy.com, ‘‘The Economic Contribution of 65 This is discussed in more detail in Paul p. 2. the Mortgage Refinancing Boom,’’ December 2002, Bennett, Richard Peach, and Stavros Peristani, 69 Refinancing data is taken from Freddie Mac’s p. 4. Structural Change in the Mortgage Market and the monthly Primary Mortgage Market Survey. 72 Fannie Mae, 2002 Fannie Mae National Propensity to Refinance, Staff Report Number 45, 70 There is some evidence that lower-income Housing Survey. , September 1998. boom as much as higher-income borrowers—see 4, 2002, p. 2. 66 Other sources of data on loan-to-value ratios Paul B. Manchester, Characteristics of Mortgages 73 Economy.com, ‘‘The Economic Contribution of such as the American Housing Survey and the Purchased by Fannie Mae and Freddie Mac: 1996– the Mortgage Refinancing Boom,’’ December 2002, Chicago Title and Trust Company indicate that 97 Update, Housing Finance Working Paper No. p. 4. high-LTV mortgages are somewhat more common in HF–006, Office of Policy Development and 74 Mark M. Zandi, ‘‘Refinancing Boom,’’ Regional the primary market than the Finance Board’s Research, Department of Housing and Urban Finance Review, December 2002, p. 11. survey. However, the Chicago Title survey does not Development, August 1998, pp. 30–32. 75 Ibid. p. 14.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00049 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24276 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

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.76 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 $1.34 trillion in financing available to them.’’ 80 were below the age of 35, compared with less 2004 as the home purchase share rises to 54 These homeownership barriers pose a than one-quarter of repeat buyers in the percent of all originations.77 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.81 city and more likely to be a female-headed calculated as the ratio of median household 85 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 current 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 82 rates being offset by gains in median family in 1980 to 25.4 percent in 1992. Even the average repeat buyer made a income.78 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 sales. According to Chicago Title data for percent of first-time homebuyers were single 143 in 1998, reflecting the most affordable 86 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 83 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 2003.79 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 84 households living in lower-income percent in 2001. and Asian immigration will also contribute 87 communities increased over the past decade, In the 1990s, lenders developed special much to this growth. 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 low-income families with special Purchases by the GSEs of single-family continues to rise, reaching $158,100 in 2002 mortgages amounted to $519 billion during and $170,000 in 2003. The production of circumstances. The disproportionate growth in the number of first-time homebuyers and the heavy refinancing year of 1993, stood at affordable housing and low interest rates $215 billion in 1995, and were at $618 billion could offset the negative impact of rising minority homebuyers largely drove the rising trend in total home purchases. Analysis of during the heavy refinancing year of 1998. house prices, which undermine housing Purchases then fell to $395 billion in 2000 affordability for many Americans, the American Housing Survey (AHS) indicates there were 1.3 million new first- before reaching record levels during the particularly in several high-cost markets on heavy refinancing years of 2001 ($961 the east and west coasts. time homebuyers during 1991, in comparison with over two million in each year between billion) and 2002 ($1,090 billion). Purchases by Fannie Mae decreased from $316 billion 76 Economy.com, ‘‘The Economic Contribution of in 1999 to $227 billion in 2000, before rising 80 Fannie Mae, September 4, 2002, p. 2. the Mortgage Refinancing Boom,’’ December 2002, to $568 billion in 2001 and $848 billion in p. 9. 81 Ibid. 2002. Freddie Mac’s single-family mortgage 77 Mortgage Bankers Association, ‘‘Mortgage 82 U.S. Department of Commerce, Bureau of the Finance Forecast’’, March 15, 2004. http:// Census, Money Income of Households, Families, purchases followed a similar trend, falling www.mortgagebankers.org/marketdata/forecasts/ and Persons in the United States: 1992, Special mffore1203.pdf. Studies Series P–60, No. 184, Table B–25, October 85 U.S. Housing Market Conditions, 3rd Quarter 78 Housing affordability varies markedly between 1993. 2001, November 2001, Table 4. regions, ranging in January 2004 from 194 in the 83 Chicago Title and Trust Family of Insurers, 86 National Association of Realtors. ‘‘New NAR Midwest to 107 in the West, with the South and Who’s Buying Homes in America, 1998. Survey of Home Buyers and Sellers Shows Growing Northeast falling in between. 84 National Association of Realtors. ‘‘New NAR Web Use in a Dynamic Housing Market.’’ http:// 79 National Association of REALTORS. Housing Survey of Home Buyers and Sellers Shows Growing www.realtor.org. Affordability Index, http://www.realtor.org/ Web Use in a Dynamic Housing Market.’’ http:// 87 Joint Center for Housing Studies at Harvard Research.nsf/Pages/HousingInx, 2003. www.realtor.org. University, State of the Nation’s Housing 2002, p.2.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00050 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24277

from $233 billion in 1999 to $168 billion in Fannie Mae and Freddie Mac have been a In 2000, Fannie Mae launched the 2000, and then rising to $393 billion in 2001 part of this ‘‘revolution in affordable ‘‘MyCommunityMortgage’’ suite of products, and $475 billion in 2002.88 lending.’’ During the mid-to-late 1990s, they which provides high loan-to-value product The Office of Federal Housing Enterprise added flexibility to their purchase guidelines, options for low- and moderate-income Oversight (OFHEO) estimates that the GSEs’ they introduced new low-down-payment borrowers. In 2002, Fannie Mae purchased or share of total originations in the conventional products, and they worked to expand the use securitized more than $882.5 million of single-family mortgage market, measured in of credit scores and automated underwriting MyCommunityMortgage products, which dollars, declined from 37 percent in 1996 to in evaluating the creditworthiness of loan helped provide affordable housing solutions 32 percent in 1997—well below the peak of applicants. These major trends reflect for 7,866 households. In addition, Fannie 51 percent attained in 1993. OFHEO changes in the GSEs’ underwriting that have Mae created new tailored solutions to attributes the 1997 downturn in the GSEs’ impacted affordable lending. Through these MyCommunityMortgage including a rural role to increased holdings of mortgages in trends, Fannie Mae and Freddie Mac have housing program, a ‘‘Community Solutions’’ portfolio by depository institutions and to attempted to increase their capacity to serve program offering flexible income increased competition with Fannie Mae and low- and moderate-income homebuyers. requirements consistent with targeted Freddie Mac by private label issuers. This section summarizes recent initiatives professions and an ‘‘Energy Efficient However, OFHEO estimates that the GSEs’ undertaken by the GSEs and others in the Mortgage’’ program.92 share of the conventional market rebounded industry to expand affordable housing. The Fannie Mae also expanded its ‘‘Flexible’’ sharply in 1998–99, to 43–42 percent. The end of this section will present evidence that product line with the ‘‘Flexible 100’’ product, GSEs’ share then decreased to approximately these new industry initiatives are working, as which eliminates the requirement for a down 30 percent of the single-family conventional increased mortgage credit has been flowing to payment by providing 100 percent loan-to- mortgages originated in 2000, and then low-income and minority families. The value financing. The borrower is required to increased sharply to 40 percent in 2001. following section will continue the affordable make at least a three percent contribution to Total GSE purchases, including loans lending theme by examining the performance closing costs; the funds for the contribution originated in prior years, amounted to 46 of different market sectors (e.g., depositories, may come from a variety on sources such as percent of conventional originations in GSEs, etc.) in funding loans for low-income gifts, grants, or unsecured loans from 2001.89 and minority families. That section will also relatives, employers, public agencies, or e. Mortgage Market Prospects discuss the important role that FHA plays in nonprofits. Lenders delivered 17,206 The Mortgage Bankers Association (MBA) making affordable housing available to ‘‘Flexible 100’’ loans to Fannie Mae totaling 93 reports that mortgage originations in 2001 historically underserved groups as well as $2.2 billion in 2001. were $2.0 trillion, which is almost twice the the continuing concern that participants in In 2001, Fannie Mae launched the eZ TM volume of originations in 2000. Mortgage the conventional market could be doing even Access product pilot. This product is originations then increased to record levels of more to help underserved families. targeted to 11 underserved markets and $2.5 trillion in 2002 and $3.8 trillion in 2003, allows lenders to qualify borrowers who may a. Lowering Down Payments and Up-Front have less than perfect credit and limited with refinancings representing 66 percent of Costs originations and the purchase volume available funds for down payment. Through amounting to $1.3 trillion. Estimates indicate Numerous studies have concluded that December 2002, eZ Access helped 400 that ARMs accounted for 19 percent of total saving enough cash for a down payment and underserved families through Fannie Mae’s mortgage originations in 2003.90 In its March for up-front closing costs is the greatest purchase of $57.1 million in loans.94 15, 2004 forecast, MBA predicts that single- barrier that low-income and minority In 2000, Freddie Mac introduced its family mortgage originations will amount to families face when considering ‘‘Freddie Mac 100’’ product, which is $2.5 trillion in 2004 and $1.9 trillion in 2005, homeownership.91 To assist in overcoming designed to assist borrowers who have good with refinancings representing 46 percent this barrier, the industry (including lenders, credit but lack the ability to provide a large and 25 percent of originations respectively. private mortgage insurers and the GSEs) down payment. ‘‘Freddie Mac 100’’ allows a began offering in 1994 mortgage products 100 percent loan-to-value ratio with the 4. Affordable Lending in the Mortgage that required down payments of only 3 condition that the borrower has the funds for Market: New Products and Outreach percent, plus points and closing costs. Other closing costs. Another Freddie Mac product, Extending homeownership opportunities industry efforts to reduce borrowers’ up-front ‘‘Affordable Gold 100’’ provides 100 percent to historically underserved households has costs included zero-point-interest-rate financing to low- and moderate-income been a growing concern for conventional mortgages and monthly insurance premiums borrowers for the purchase price of a home lenders, private mortgage insurers and the with no up front component. These new in California. ‘‘Affordable Gold 100’’ GSEs. The industry has responded in what plans eliminated large up-front points and combines mortgage insurance benefits some have called a ‘‘revolution in affordable premiums normally required at closing. provided by a state insurance fund, the lending.’’ The industry has offered more During 1998, Fannie Mae introduced its secondary mortgage market, and a team of the customized mortgage products, more flexible ‘‘Flexible 97’’ and Freddie Mac introduced its nation’s leading mortgage lenders.95 underwriting, and expanded outreach so that ‘‘Alt 97’’ low down payment lending b. Partnerships—Fannie Mae the benefits of the mortgage market can be programs. Under these programs, borrowers extended to those who have not been In addition to developing new affordable were required to put down only 3 percent of products, lenders and the GSEs have been adequately served through traditional the purchase price. The down payment, as products, underwriting, and marketing. entering into partnerships with local well as closing costs, could be obtained from governments and nonprofit organizations to a variety of sources, including gifts, grants or increase mortgage access to underserved 88 The source of the GSE data for 2001 and earlier loans from a family member, the government, years is the Office of Federal Housing Enterprise borrowers. Fannie Mae’s partnership offices a non-profit agency and loans secured by life in 54 central cities, which coordinate Fannie Oversight (OFHEO), Report to Congress, 2002 (see insurance policies, retirement accounts or Tables 1 and 11). The 2002 data are taken from Mae’s programs with local lenders and ‘‘Fannie and Freddie Roll to Nearly $1.5 Trillion in other assets. Fannie Mae continues to offer affordable housing groups, are an example of New Business, Portfolios Continue Growing’’ in the ‘‘Flexible’’ line of products, and Freddie this initiative. Inside Mortgage Finance, January 31, 2003, pages 6– Mac continues to list ‘‘Alt 97.’’ Fannie Mae continues to reach out to 7. It should be noted that the Inside Mortgage national groups and work with local affiliates Finance data for 2001 was 13 percent higher than 91 See Charles, K. K. and E. Hurst (2002). ‘‘The the OFHEO data for 2001; therefore, the 2002 data Transition to Home Ownership and the Black-White 92 may be overstated. Wealth Gap.’’ The Review of Economics and Fannie Mae, 2002 Annual Housing Activities 89 Office of Federal Housing Enterprise Oversight. Statistics, 84(2): 281–297; Mayer, C. and G. Report, 2003, pp. 8–9. ‘‘Mortgage Markets and The Enterprises in 2001,’’ Engelhardt (1996). ‘‘Gift Down Payments and 93 Fannie Mae, 2001 Annual Housing Activities August 2002, p. 13 Housing Affordability.’’ Journal of Housing Report, 2002, pp. 5–7. 90 Mortgage market projections from the MBA’s Research, 7(1): 59–77; and Quercia, R. G., G. W. 94 Fannie Mae, 2002 Annual Housing Activities MBA Mortgage Finance Forecast, December 17, McCarthy, et al. (2003). ‘‘The Impacts of Affordable Report, 2003, p. 8. 2003. 2000 and 2001 numbers from the MBA’s MBA Lending Efforts on Homeownership Rates.’’ Journal 95 Freddie Mac, 2002 Annual Housing Activities Mortgage Finance Forecast, January 10, 2002. of Housing Economics, 12(1): 29–59. Report, 2003, p.57.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00051 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24278 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

to expand homeownership. In 2002, Fannie partnerships, bringing the total number of products to meet the needs of Hispanic Mae enhanced 5 partnerships with national lenders committed since 2000 to 16, with an borrowers. Mortgage products include low organizations and maintained 13 national estimated $180 billion of American Dream down payments, flexible credit underwriting partnership agreements. For example, Fannie Commitment business pledged to be and debt-to-income ratios, and streamlined Mae maintains a partnership with the delivered. Examples of lender partnerships processing for resident alien borrowers.105 National Urban League (NUL) and the Chase under this initiative include J.P. Morgan In 2002, Freddie Mac joined with the City Manhattan Mortgage Corporation to increase Chase & Co. with a $35 billion national of Boston and the U.S. Conference of Mayors NUL’s homeownership counseling capacity investment initiative designed to increase to make available the ‘‘Don’t Borrow by providing the necessary technology and homeownership opportunities for Trouble’’ predatory lending educational underserved communities and improve tools to support the effort, and to purchase campaign to approximately 1,100 cities. In affordable homeownership options for $50 million in mortgage products over five addition, Freddie Mac joined with Rainbow/ years that are specifically targeted to African immigrants and minorities, and Bank One PUSH and the National Urban League to Americans and other minorities in with a $12.5 billion community lending promote the ‘‘CreditSmartSM’’ financial underserved areas. In 2002, NUL originated alliance to help low- and moderate-income $20 million in loans. Another example is families purchase homes with a total educational curriculum that helps consumers Fannie Mae’s partnership with the AFL-CIO designated commitment of at least 25% understand, obtain and maintain good credit, Housing Investment Trust (HIT) and toward increasing homeownership among thereby preparing them for homeownership Countrywide Mortgage, which launched minorities.99 and other personal financial goals. In 2002, ‘‘HIT HOME’’ in 2001. HIT HOME is an Through these partnerships, a strategic Freddie Mac also joined with the American affordable home mortgage initiative that effort was made to eliminate language, credit, Community Bankers and the Credit Union targets 13 million union members in 16 cities and other barriers to minority National Association in strategic alliances throughout the nation to provide union homeownership and to reach underserved that will better enable member banks and members with a variety of affordable communities. In 2002, Fannie Mae helped credit unions access to the secondary mortgage choices that enable them to qualify serve 984,276 minority families by providing market.106 for competitively priced loans with new re- $136.2 billion in mortgage financing.100 In June 2002, President George W. Bush payment terms. As of December 2002, over According to Fannie Mae, its lending challenged the nation’s housing industry to $244 million in loans have been originated partners realize that multicultural markets invest more than $1 trillion to make may differ from traditional markets, and thus through this initiative, serving 2,076 homeownership a reality for 5.5 million more they offer various flexible mortgage products households.96 minority households for the decade. Freddie to reach out to minority and immigrant In order to meet the needs of underserved Mac responded to the challenge with ‘‘Catch and low- and moderate-income populations, homebuyers. Some of these mortgage products require only a $500 contribution the Dream,’’ which is a comprehensive set of Fannie Mae has targeted specific populations 25 major initiatives aimed at accelerating the for initiatives. These include minority and from the borrower for closing costs. Others have flexible qualifying guidelines that use growth in minority homeownership. The women-owned lenders (MWOL), Native initiatives range from homebuyer education Americans, working Americans, and alternative sources of income like rent and 101 and outreach to new technologies with borrowers served by community part-time employment. innovative mortgage products. Catch the development financial institutions and c. Partnerships—Freddie Mac Dream represents a collaborative effort with public housing agencies. In 2002, through the Freddie Mac does not have a partnership lenders, nonprofit housing and community- MWOL Initiative, Fannie Mae purchased $9 office structure similar to Fannie Mae’s, but billion in mortgages originated by MWOLs; based organizations, and other industry it has undertaken a number of initiatives in participants to expand homeownership 97% of this amount reached minority 102 specific metropolitan areas. In 2001, opportunities for America’s minorities.107 households. The Employer Assisted Housing Freddie Mac joined the Congressional Black Freddie Mac has committed to providing Initiative reached 116 employers in 2002 in Caucus to launch a new initiative, ‘‘With $400 billion in mortgage financing for industries ranging from health care to Ownership Wealth,’’ designed to increase minority families by the end of the decade.108 education. The Community Development African-American homeownership with one Financial Institutions Initiative committed to million new families by 2005; Freddie Mac In 2002, Freddie Mac purchased mortgages invest $17.1 million in 2002, which was has pledged to purchase qualified mortgages for 576,000 minority families, a total of expected to generate more than 980 originated under this initiative.103 In 2002, 17.3% of their single-family, owner-occupied 109 additional units of affordable housing. The Freddie Mac launched more than 30 new mortgage purchases for the year. In Section 8 Homeownership Initiative helped alliances and initiatives and continued addition, in 2002, minority- or women- 35 families make the transition from Section working with existing alliances.104 Freddie owned lenders comprised 2.7% of Freddie 8 rental housing to homeownership in 2002. Mac has partnered with the National Council Mac’s network of lenders. $5.5 billion in The Native American Initiative has served of La Raza (NCLR), 20 community based loans were purchased from these lenders, more than 3,376 Native American families NCLR affiliated housing counseling financing housing for 45,000 families.110 living on reservations and trust lands since organizations, the National Association of The programs mentioned above are its inception, while providing $290 million Hispanic Real Estate Professionals examples of the partnership efforts in mortgage financing.97 (NAHREP), EMT Applications and undertaken by the GSEs. There are more Fannie Mae’s American Dream participating Freddie Mac Seller/Servicers partnership programs than can be adequately Commitment’s Opportunity for All Strategy including Bank of America, U.S. Bank and described here. Fuller descriptions of these and National Minority Homeownership Wells Fargo Home Mortgage on the ‘‘En Su programs are provided in their Annual Initiative has pledged to contribute at least Casa’’ initiative. This $200 million Housing Activity Reports. $700 billion in private capital to serve 4.6 homeownership initiative combines million families towards President George W. technology tools with flexible mortgage 105 Bush’s goal of expanding homeownership to Freddie Mac, 2002 Annual Housing Activities 5.5 million new minority Americans by the Report, 2003, p. 61. 99 Fannie Mae, 2002 Annual Housing Activities 106 end of the decade.98 This marks a 66% Freddie Mac, 2002 Annual Housing Activities Report, 2003, pp. 15–16. Report, 2003, pp. 35–38. increase in Fannie Mae’s earlier commitment 100 Fannie Mae, 2002 Annual Housing Activities 107 Freddie Mac. Corporate Information. ‘‘Our of $420 billion. Towards this goal, in 2002, Report, 2003, p 5. Homeownership Commitment.’’ http:// Fannie Mae announced 10 new lender 101 Fannie Mae, ‘‘Minority Homeownership,’’ www.freddiemac.com/corporate/about/dream/ 2002. expanding_minority_homeownership.htm. 96 Fannie Mae, 2002 Annual Housing Activities 102 Freddie Mac, News Release, January 15, 1999. 108 Freddie Mac, 2002 Annual Housing Activities Report, 2003, pp. 12–15. 103 Freddie Mac, 2002, pp. 41–42, and Freddie Report, 2003, p. 28. 97 Fannie Mae, 2002 Annual Housing Activities Mac, 2002 Annual Housing Activities Report, 2003, 109 Freddie Mac, 2002 Annual Housing Activities Report, 2003, pp. 16–18. p. 62. Report, 2003, p. 32. 98 Fannie Mae, 2002 Annual Housing Activities 104 Freddie Mac, 2002 Annual Housing Activities 110 Freddie Mac, 2002 Annual Housing Activities Report, 2003, p. 15. Report, 2003, p. 60. Report, 2003, p. 15.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00052 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24279

d. Underwriting and GSE Purchase underwriting guidelines that allow lower scoring section, there are concerns that the Guidelines down payments, higher debt-to-income ratios codification of certain underwriting Lenders, mortgage insurers, and the GSEs and poorer credit histories than allowed by guidelines could result in unintentional have also been modifying their mortgage the GSEs’ guidelines. discrimination or disparate treatment across underwriting standards to address the needs From this and other evidence, the Urban groups. In response to the potential disparate of families who have historically found it Institute concluded that the GSEs were impact of automated underwriting, Freddie difficult to qualify under traditional lagging the market in servicing low- and Mac have launched initiatives to make the guidelines. In addition to the changes in moderate-income and minority borrowers. mortgage process more transparent by underwriting standards, the use of automated Furthermore, the Urban Institute found ‘‘that disclosing both credit and non-credit factors underwriting has dramatically transformed the GSEs’’ efforts to increase underwriting that Loan Prospector consider when the mortgage application process. This flexibility and outreach has been noticed and evaluating a loan application. In 2000, section focuses on changes to traditional is applauded by lenders and community Freddie Mac launched an initiative that advocates. Despite the GSEs’ efforts in recent underwriting standards and recent GSE published a list of all of the factors that Loan years to review and revise their underwriting initiatives for credit-impaired borrowers. Prospector uses to analyze loans, and put the criteria, however, they could do more to Subsequent sections will provide more list on the Freddie Mac Web site.115 serve low- and moderate-income borrowers details on the impact of automated In 2002, Fannie Mae released two versions underwriting. and to minimize disproportionate effects on minorities.’’112 Since the Urban Institute of its automated underwriting service, The GSEs modified their underwriting ‘‘Desktop Underwriter’’ (DU), to expand its standards to address the needs of families study, Freddie Mac and Fannie Mae have been playing a larger role in financing low- mortgage product offerings and to update who find qualifying under traditional income and minority borrowers. (See Section underwriting guidelines. These guidelines difficult. The goal of these E.2.) enhancements—labeled DU 5.2 and DU underwriting changes is not to loosen In addition to offering low-down-payment 5.2.1—increased homeownership underwriting standards, but rather to identify programs, the GSEs’ recent efforts have also opportunities for low- and moderate-income creditworthiness by alternative means that centered around their automated borrowers and borrowers with small more appropriately measures the unique underwriting systems and their treatment of downpayments by enhancing DU’s risk circumstances of low-income, immigrant, borrowers with blemished credit, the latter assessment capabilities for certain high loan- and minority households. Examples of being perhaps the most controversial to-value loans. For example, DU 5.2.1 changes that the GSEs and others in the underwriting issue over the past few years. enhanced its Expanded ApprovalTM policies industry have made to their underwriting Freddie Mac recently launched a variety of to allow 100 percent loan-to-value limited standards include the following: • new products aimed at providing borrowers cash-out refinances and the origination of 5/ Using a stable income standard rather with impaired credit more mortgage product 1 ARMs.116 The Expanded Approval feature than a stable job standard (or minimum choices. The new products include: and Timely Payment Rewards option in DU period of employment). This particularly ‘‘CreditWorks,’’ which helps borrowers with were created by Fannie Mae in 1999 to benefits low-skilled applicants who have excessive debt and impaired credit to qualify enable lenders to more comprehensively successfully remained employed, even with for a prime market rate mortgage more review a borrower’s creditworthiness. The frequent job changes. quickly than before, and ‘‘LeasePurchase Plus Timely Payment Rewards option reduces the • Using an applicant’s history of rent and Initiative,’’ which provides closing cost and interest rate of qualified borrowers of up to utility payments as a measure of down payment assistance in addition to one percent after making timely mortgage creditworthiness. This measure benefits extensive counseling for borrowers who have payments for a given time period.117 With lower-income applicants who have not had bad credit or who have never established these options, lenders can offer mortgage established a credit history. a credit history.113 During 2002, Freddie Mac loans to many borrowers previously unable • Allowing pooling of funds for entered into several new markets under the to receive financing from a mainstream qualification purposes. This change benefits ‘‘LeasePurchase Plus Initiative’’ and lender. A borrower who is recommended for applicants with extended family members. 114 purchased more than $16 million in loans. approval for either of these features would be Freddie Mac, for example, allows income According to Freddie Mac, its automated eligible for an initial mortgage rate that is from relatives who live together to pool their underwriting system, ‘‘Loan Prospector’’ has lower than that available through the funds to cover downpayment and closing reduced costs, made approving mortgages 118 costs and to combine their incomes for use subprime market. Automated mortgage easier and faster, and increased the scoring and the potential for disparate in calculating the borrower’s stable monthly consistency of the application of objective income. impacts on borrowers will be further underwriting criteria. In addition, Freddie discussed in a later section. These underwriting changes have been Mac states that ‘‘Loan Prospector’’ extends accompanied by homeownership counseling the benefits of the mortgage finance system 5. Affordable Single-Family Lending: Data to ensure homeowners are ready for the to borrowers with less traditional credit Trends responsibilities of homeownership. In profiles and limited savings by more a. 1993–2002 Lending Trends addition, the industry has engaged in accurately measuring risk. Freddie Mac intensive loss mitigation to control risks. reports that its automated underwriting HMDA data suggest that the industry and In 1999, HUD commissioned a study by the system, Loan Prospector, has resulted in GSE initiatives are increasing the flow of Urban Institute to examine the underwriting higher approval rates for minority borrowers credit to underserved borrowers. Between criteria that the GSEs use when purchasing than under traditional manual underwriting 1993 and 2002, conventional loans to low- mortgages from primary lenders.111 because of improved predictive powers. As income and minority families increased at According to the study, while the GSEs had mentioned in Section C.7, the 2000 version much faster rates than loans to higher income improved their ability to serve low- and of LP approved 87.1 percent of loans and non-minority families. As shown below, moderate-income borrowers, it did not generated through affordable housing conventional home purchase originations to appear at that time that they had gone as far programs, compared to 51.6 percent African Americans more than doubled as some primary lenders to serve these approved by manual underwriting. The between 1993 and 2002 and those to borrowers. From the Urban Institute’s Freddie Mac study found automated Hispanic borrowers more than tripled. Home discussion with lenders, it was found that mortgage scoring less discriminatory and loans to low-income borrowers and to low- primary lenders were originating mortgages more accurate in predicting risk. However, as income and high-minority census tracts also to lower-income borrowers using noted below in the automated mortgage more than doubled during this period.

111 Kenneth Temkin, Roberto Quercia, George 113 Freddie Mac, 2001 Annual Housing Activities 116 Fannie Mae, 2002 Annual Housing Activities Galster, and Sheila O’ Leary, A Study of the GSEs’ Report, 2002, p. 28. Report, 2003, p. 10. Single Family Underwriting Guidelines: Final 114 Freddie Mac, 2002 Annual Housing Activities 117 Ibid. p. 6. Report. Washington DC: U.S. Department of Report, 2003, p. 35. 118 Housing and Urban Development, April 1999. Fannie Mae, 2002 Annual Housing Activities 115 112 Temkin, et al. 1999, p. 28. Ibid. p. 57. Report, 2003, p. 32.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00053 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24280 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

1993–2002 1993–2002 Growth rate: con- Growth rate: all ventional home home loans loans P (per cent) P (per cent)

African-American Borrowers ...... 80 133 Hispanic Borrowers ...... 186 245 White Borrowers ...... 30 43 Low-Income Borrower (Less than 80% of AMI) ...... 91 119 Upper-Income Borrower (More than 120% of AMI) ...... 66 81 Low-Income Census Tract ...... 99 143 Upper-Income Census Tract ...... 64 78 High-Minority Tract (50% or more minority) ...... 113 167 Predominantly-White Tract (Less than 10% minority) ...... 53 64

GSE purchases showed similar trends, as initiatives such as the GSE housing goals and lending market is highlighted and questions indicated by the following 1993–to–2002 the Community Reinvestment Act have also are raised about whether the conventional percentage point increases for metropolitan played a role in the growth of affordable conforming market could be doing a better areas: African-American borrowers (193 lending over the past 10 years. job helping low-income and minority percent), Hispanic borrowers (208 percent), b. Affordable Lending Shares by Major borrowers obtain access to mortgage credit. and low-income borrowers (193 percent). Market Sector Table A.1 reports borrower characteristics While their annual purchases of all home and Table A.2 reports neighborhood loans increased by 57 percent between 1993 Section E below compares the GSEs’ performance with the performance of characteristics for home purchase mortgages and 2001, their purchases of mortgages that insured by FHA, purchased by the GSEs, qualify for the three housing goals increased primary lenders in the conventional originated by depository institutions (mainly as follows: Special affordable by 264 percent; conforming market. To provide a useful banks and thrift), and originated in the low- and moderate-income by 142 percent; context for that analysis, this section and underserved areas by 112 percent. examines the role of the conventional conventional conforming market and in the While low interest rates and economic conforming market in funding low-income total market for owner-occupied properties in 119 expansion certainly played an important role and minority families and their metropolitan areas. In this case, the ‘‘total’’ in the substantial increase in conventional neighborhoods. Information on the mortgage market consists of both the conventional affordable lending in recent years, most market’s funding of homes purchased by conforming market and the government observers believe that the efforts of lenders, first-time homebuyers is also provided. In (mainly FHA and VA loans) market; ‘‘jumbo’’ private mortgage insurers, and the GSEs were addition, this section compares the GSEs loans above the conventional conforming also important contributors. In addition, with other sectors of the mortgage market. loan limit are excluded from this analysis.120 many observers believe that government The important role of FHA in the affordable BILLING CODE 4210–22–P

119 Table A.3 also provides the same average loans. Thus, it provides a complete picture of below the conforming loan limit of $240,000 in (1999 to 2002) information as Tables A.1 and A.2 overall mortgage activity. 1999, $252,700 in 2000, $275,000 in 2001, and but for total (both home purchase and refinance) 120 The ‘‘Total Market’’ is defined as all loans $300,700 in 2002. (including both government and conventional)

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00054 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24281

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00055 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.006 24282 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00056 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.007 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24283

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00057 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.008 24284 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

HMDA is the source of the FHA, The main insights from the ‘‘distribution of percent of FHA-insured loans and 18.8 depository, and market data, while the GSEs business’’ percentages in Tables A.1 and A.2 percent of all loans originated in the total provide their own data. Low-income, pertain to four topics. (government and conventional conforming) African-American, Hispanic, and minority (i) FHA-Insured Loans. FHA has market. Not surprisingly, the minority borrowers are covered in Table A.1. Table traditionally been the mechanism used by lending performance of conventional lenders A.2 provides information on four types of borrowers who face difficulty obtaining has been subject to much criticism. Recent neighborhoods—low-income census tracts, mortgage financing in the private studies contend that primary lenders in the tracts where minorities (or African conventional market. FHA has long been conventional market are not doing their fair Americans) account for more than 30 percent recognized as the major source of funding for share of minority lending which forces of the census tract population, and first-time, low-income and minority minorities, particularly African-American underserved areas as defined by HUD. The homebuyers who are not often able to raise and Hispanic borrowers, to rely on more average data reported in Tables A.1 and A.2 cash for large downpayments.122 Tables A.1 costly FHA and subprime loans.124 Thus, it for the years 1999 to 2002 offer a good and A.2 show that FHA places much more appears that conventional lenders could be summary of recent lending to low-income emphasis on affordable lending than the doing a better job helping minority borrowers and minority borrowers and their other market sectors. Between 1999 and obtain access to mortgage credit. communities.121 Individual year data are also 2002, low-income borrowers accounted for • The GSEs’ funding of minority loans can provided. 50.7 percent of FHA-insured loans, compared be compared with mortgages originated for The focus of different market sectors on with 27.1 percent of the home loans minority borrowers in the conventional affordable lending is summarized by the purchased by the GSEs, 29.2 percent of home conforming market, although the latter may percentages reported in Tables A.1 and A.2. loans originated by depositories, and 29.5 be a poor benchmark, as discussed above. These percentages show each sector’s percent of all originations in the Between 1999 and 2002, home purchase ‘‘distribution of business,’’ defined as the conventional conforming market (see Table loans to African-American and Hispanic share of loans originated (or, for the GSEs, A.1 ). Likewise, 40.9 percent of FHA-insured borrowers accounted for 10.3 percent of purchased) that had a particular borrower or loans were originated in underserved census Freddie Mac’s purchases, 13.0 percent of neighborhood characteristic. The tracts, while only 23.5 percent of the GSE- Fannie Mae’s purchases, and 14.3 percent of interpretation of the ‘‘distribution of purchased loans, 25.7 percent of home loans loans originated in the conventional business’’ percentages can be illustrated originated by depositories, and 26.5 percent conforming market (or 13.7 percent if B&C using the FHA percentage for low-income of conventional conforming loans were loans are excluded from the market 123 borrowers: Between 1999 and 2002, 50.7 of originated in these tracts (see Table A.2). definition). Thus, since 1999, the African- all FHA-insured home purchase loans in As discussed in Section E, FHA’s share of the American and Hispanic share of the GSEs’ metropolitan areas were originated for minority lending market is particularly high. purchases has been lower than the borrowers with an income less than 80 While FHA insured only 18 percent of all corresponding share for the conventional percent of the local area median income. home purchase mortgages originated below conforming market.125 These percentages are to be contrasted with the conforming loan limit in metropolitan • As the above comparisons show, Fannie ‘‘market share’’ percentages, which are areas between 1999 and 2002, it is estimated Mae has had a much better record than presented below in Section E. A ‘‘market that FHA insured 33 percent of all home Freddie Mac in funding loans for minority share’’ percentage is the share of loans with loans originated for African-American and families. And Fannie Mae significantly a particular borrower or neighborhood Hispanic borrowers. increased its purchases of loans for African- (ii) Conventional and GSE Minority characteristic that was funded by a particular American and Hispanic borrowers during Lending. The affordable lending shares for market sector (e.g., FHA-insured, GSEs, 2001, raising the share of its purchases to the conventional conforming sector are low depositories). As will discussed below, market levels—13.7 percent for both Fannie for minority borrowers, particularly African- FHA’s ‘‘market share’’ for low-income Mae and the conforming market (without American and Hispanic borrowers. These borrowers during the 1999-to-2002 period B&C loans). In 2002, Fannie Mae surpassed borrowers accounted for only 14.3 percent of was estimated to be 26 percent which is the conventional conforming market in all conventional conforming loans originated funding African-American and Hispanic interpreted as follows: Of all home purchase between 1999 and 2002, compared with 34.7 loans originated for low-income borrowers in borrowers—a 15.8 percent share for Fannie metropolitan areas between 1999 and 2002, Mae and a 15.2 share for the market. When 122 26 percent were FHA-insured loans. Thus, in Almost two-thirds of the borrowers with an all minority borrowers are considered, FHA-insured home purchase loan make a this example, the ‘‘distribution of business’’ downpayment less than five percent, and over 80 Fannie Mae has purchased mortgages for percentage measures the importance (or 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) 124 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 concepts are important for evaluating Development and Research, U.S. Department of of the Racial Patterns in Home Purchase Lending Housing and Urban Development, 1995; and (b) in the Baltimore Market. Report for The Public performance—for an industry sector such as Office of Policy Development and Research, ‘‘FHA’s Justice Center, May 2000; and The Patterns of GSE FHA or the GSEs to have a significant impact 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 targeted group and that sector must be of Urban Development, December 2000. For data on Lending, GSE Study No. 11, U.S. Department of some size. The discussion below will focus the credit characteristics of FHA borrowers, see Housing and Urban Development, September 2000. on the degree to which different mortgage Harold L. Bunce, William J. Reeder and Randall For analysis suggesting some minorities receiving sectors concentrate on targeted groups, while Scheessele, ‘‘Understanding Consumer Credit and FHA loans could qualify for conventional loans, see Mortgage Scoring: A Work in Progress at HUD’’, Anthony Pennington-Cross, Anthony Yezer, and Section E will also provide estimates of U.S. Department of Housing and Urban Joseph Nichols, Credit Risk and Mortgage Lending: market shares. Development, Unpublished Paper, 1999. Who Uses Subprime and Why? Working Paper No. 123 FHA, which focuses on low downpayment 00–03. Research Institute for Housing America, 121 The affordable market shares reported in Table loans and also accepts borrowers with credit 2000. Also see the series of recent studies A.1 for the ‘‘Conventional Conforming Market W/ blemishes, experiences higher mortgage defaults concerning the lack of mainstream lenders in O B&C’’ were derived by excluding the estimated than conventional lenders and the GSEs. Still, the minority neighborhoods. number of B&C loans from the market data reported FHA system is actuarially sound because it charges 125 For a comprehensive analysis of the GSEs’ by HMDA. Because B&C lenders operate mainly in an insurance premium that covers the higher purchases of minority loans through 1999, see the refinance sector, excluding these loans from the default costs. For the results of FHA’s actuarial Harold L. Bunce, An Analysis of GSE Purchases of conforming market has litte impact on the home analysis, see Deloitte & Touche, Actuarial Review of Mortgages for African-American Borrowers and purchase percentages reported in Table A.1. The MMI Fund as of FY 2000, report for the U.S. their Neighborhoods, Housing Finance Working method for excluding B&C loans is explained in Department of Housing and Urban Development, Paper No. 11, Office of Policy Development and Section E below and Appendix D. January 2001. Research, HUD, December 2000.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00058 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24285

minority borrowers at a higher rate (years income families and their neighborhoods.126 originated in the conventional conforming 2001 and 2002) than these loans were Between 1999 and 2002, underserved areas market. The AHS defines a first-time originated by primary lenders in the accounted for 26.8 percent of loans held in homebuyer as someone who has never conventional conforming market (without depository portfolios, which compares owned a home. Using a more liberal B&C loans). Freddie Mac, on the other hand, favorably with the underserved areas definition of a first-time homebuyer lagged behind both the market and Fannie percentage (26.5 percent) for the overall (someone who has not owned a home in the 127 Mae in funding loans for minority borrowers conventional conforming market. past three years), FHA reports that first-time Depository lenders have extensive knowledge during 2001 and 2002, as well as during the homebuyers accounted for 80.5 percent of all of their communities and direct interactions entire 1999-to-2002 period. The share of home loans that it insured between 1999 and Freddie Mac’s purchases for African- with their borrowers, which may enable them to introduce flexibility into their 2001 and the GSEs report that first-time American and Hispanic borrowers declined homebuyers accounted for 26.5 percent of the from 10.9 percent in both 2000 and 2001 to underwriting standards without unduly increasing their credit risk. The Community home loans purchased by each GSE during 10.1 percent in 2002. that same period. Given FHA’s low • Considering the minority census tract Reinvestment Act provides an incentive for banks and thrifts to initiate affordable downpayment requirements, it is not data reported in Table A.2, Fannie Mae lending programs with underwriting surprising that FHA focuses on first-time lagged behind the conforming market flexibility and to reach out to lower income homebuyers. The GSEs, on the other hand, (without B&C loans) in high-minority families and their communities.128 Many of fall at the other end of the continuum, with neighborhoods and in high-African-American the CRA loans are held in portfolio by their first-time homebuyer share (26.5 neighborhoods during the 1999-to-2002 lenders, rather than sold to Fannie Mae or percent) falling far short of the first-time period. However, Fannie Mae improved its Freddie Mac.129 mortgage purchases in African-American homebuyer share (37.6 percent) of the (v) First-time Homebuyers. As explained in conventional conforming market. Section E neighborhoods during 2001 and 2002 to Section E, market information on first-time exceed market levels by 0.1 percentage point will include a more detailed comparison of homebuyers is not as readily available as the the GSEs and the conventional conforming (e.g., 4.7 percent of Fannie Mae’s purchases HMDA data reported in Tables A.1 and A.2 market in serving first-time homebuyers. In and 4.6 percent of market originations were on the income and racial characteristics of addition, Section E will conduct a market in high African-American tracts in 2002). borrowers and census tracts served by the And during 2001 and 2002, Fannie Mae also mortgage market. However, the limited share analysis that examines the funding of purchased loans in high-minority census market data that are available from the minority first-time homebuyers. Consistent tracts at a higher rate than loans were American Housing Survey, combined with with the earlier discussion, that analysis originated by conventional lenders in these the first-time homebuyer data reported by suggests that conventional lenders and the tracts. While Freddie Mac has generally FHA and the GSEs, indicate a rather large GSEs have played a relatively small role in lagged the primary market in funding variation in the funding of first-time the market for minority first-time minority neighborhoods, note in Table A.2 homebuyers across the different sectors of the homebuyers. One analysis reported in that high African-American tracts increased mortgage market. Based on the American Section E estimates that mortgage purchases from 3.9 percent of Freddie Mac’s purchases Housing Survey (AHS), it is estimated that by the GSEs between 1999 and 2001 totaled in 2001 to 5.3 percent in 2002, placing first-time homebuyers accounted for 42.3 41.5 percent of all home loans originated, but Freddie Mac above the conventional percent of all home purchase loans originated they accounted for only 14.3 percent of home conforming market level (4.6 percent) in throughout the market between 1999 and loans originated for first-time African- 2002. 2001,130 and for 37.6 percent of home loans American and Hispanic homebuyers. (iii) Low-Income Lending by the GSEs. c. Community Reinvestment Act Information is also provided on the GSEs’ 126 Tables A.1, A.2, and A.3 include data for all purchases of home loans for low-income home loans originated by depositories as well as for The Community Reinvestment Act (CRA) borrowers (A.1) and for families living in the subset of loans originated but not sold, the latter requires depository institutions to help meet being a proxy for loans held in depository the credit needs of their communities.131 low-income neighborhoods (A.2). portfolios. (See the notes to Table A.1 for Historically, the GSEs have lagged behind the definitions of the depository data.) CRA loans are typically made to low-income conventional conforming market in funding 127 However, as shown in Table A.1 , depository borrowers earning less than 80 percent of affordable loans for these groups. During the institutions resemble other conventional lenders in area median income, and in moderate- 1999-to-2002 period, low-income borrowers their relatively low level of originating loans for income neighborhoods. CRA provides an (census tracts) accounted for 27.2 (9.6) African-American, Hispanic and minority incentive for lenders to initiate affordable borrowers. Within the conventional conforming lending programs with underwriting percent of Freddie Mac’s purchases, 27.1 market, Fannie Mae has done a better job than (9.8) percent of Fannie Mae’s purchases, 29.2 depositories in funding minority borrowers, flexibility. CRA loans are usually smaller (11.1) percent of loans originated by particularly Hispanic borrowers and minority than typical conventional mortgages and also depositories, and 29.3 (11.1) percent of home borrowers as a group. During the last two years, are more likely to have a higher LTV, higher loans originated by conventional conforming Fannie Mae has also funded African-American debt-to-income ratios and no payment lenders (without B&C loans). By the end of borrowers at a higher rate than have depository reserves, and may not be carrying private this period, Fannie Mae had significantly institutions. 128 mortgage insurance (PMI). Generally, at the improved its performance relative to the CRA loans are typically made to low-income borrowers earning less than 80 percent of area time CRA loans are originated, many do not market. In 2002, low-income borrowers median income, and in moderate-income meet the underwriting guidelines required in (census tracts) accounted for 29.7 (11.0) of neighborhoods. For a comprehensive analysis of order for them to be purchased by one of the Fannie Mae’s purchases, compared with 29.6 CRA and its impact on affordable lending, see GSEs. Therefore, many of the CRA loans are (11.1) percent for the conforming market. It Robert E. Litan, Nicolas P. Retsinas, Eric S. Belsky held in portfolio by lenders, rather than sold is also interesting that even though Freddie and Susan White Haag, The Community to Fannie Mae or Freddie Mac. Evidence is Reinvestment Act After Financial Modernization: A Mac lagged the market in funding home loans growing that CRA-type lending to low- for low-income borrowers during 2002 (28.6 Baseline Report, U.S. Department of Treasury, 2000. 129 income families can be profitable, percent versus 29.6 percent), it surpassed the Evidence is growing that CRA-type lending to low-income families can be profitable, particularly particularly when combined with intensive market in financing properties in low-income when combined with intensive loss mitigation loss mitigation efforts to control credit risk. census tracts (11.3 percent versus 11.1 efforts to control credit risk. In a survey conducted In a recent survey conducted by the Federal percent). A more complete analysis of the by the Federal Reserve, lenders reported that most Reserve, lenders reported that most CRA GSEs’ recent improvements in purchasing CRA loans are profitable although not as profitable home loans that qualify for the housing goals as the lenders’ standard products. See Board of is provided below in Section E. Governors of the Federal Reserve System. The 131 For a comprehensive analysis of CRA and its (iv) Depositories. Within the conventional Performance and Profitability of CRA-Related impact on affordable lending, see Robert E. Litan, Lending. Washington, DC, 2000. conforming market, depository institutions Nicolas P. Retsinas, Eric S. Belsky and Susan White 130 In this case, the market includes all Haag, The Community Reinvestment Act After (mainly banks and thrifts) are important government and conventional loans, including Financial Modernization: A Baseline Report, U.S. providers of affordable lending for lower- jumbo loans. Department of Treasury, 2000.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00059 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24286 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

loans are profitable although not as profitable for 7,866 households.136 In addition, Freddie and perhaps even 10 more years.’’ 139 As as the lenders’ standard products.132 Mac is also purchasing seasoned affordable these demographic factors play out, the Some anticipate that the big growth market mortgage portfolios originated by overall effect on housing demand will likely over the next decade for CRA-type lending depositories to help meet their CRA be sustained growth and an increasingly will be urban areas. There has been some objectives. In 2002, Freddie Mac developed diverse household population from which to movement of population back to cities, credit enhancements that enable depositories draw new homeowners. consisting of aging Baby Boomers (so-called to profitably sell their loans to Freddie Mac— Surveys indicate that these demographic ‘‘empty nesters’’), the children of Baby these transactions facilitate targeted trends will be reinforced by the fact that most Boomers (the Echo Boomers aged 18–25), and affordable lending activity by providing Americans desire, and plan, to become immigrants, particularly Hispanics but also immediate liquidity. Freddie Mac also homeowners. According to the 2002 Fannie Asians.133 The current low homeownership increased its ability to purchase smaller Mae Foundation annual National Housing in inner cities (compared with the suburbs) portfolios opening this option to many Survey, Americans rate homeownership as also suggests that urban areas may be a community banks that otherwise would not the best investment they can make, far ahead potential growth market for lenders. Lenders have an outlet for their portfolios.137 The of 401Ks, retirement accounts, and stocks. are beginning to recognize that urban billions of dollars worth of CRA loans that The percentage of Americans who said it was borrowers are different from suburban will be originated, as well as the CRA loans a good time to buy a home was at its highest borrowers. A new or recent immigrant may being held in bank and thrift portfolios, offer level since 1994 at 75 percent, a jump of 21 have no credit history or, more likely, a loan- both GSEs an opportunity to improve their percentage points since May 2001.140 In worthy credit history that can’t be performance in the single-family area. addition, the survey found that 27 percent of substantiated by the usual methods.134 6. Potential Homebuyers Americans report they are likely to buy in the Products for duplexes and four-plexes are not next three years, and 23 percent of those have the same as a mortgage for a subdivision While the growth in affordable lending and started to save or have saved enough money house in the suburbs. Programs are being homeownership has been strong in recent for a down payment.141 years, attaining this Nation’s homeownership implemented to meet the unique needs of Further increases in the homeownership goals will not be possible without tapping urban borrowers. One program emphasizing rate depend on whether or not recent gains into the vast pool of potential homebuyers. urban areas was initiated by the American in the home owning share(s) of specific Due to record low interest rates, expanded Community Bankers (ACB). Under the ACB groups are maintained. Minorities accounted homeownership outreach, and new flexible program, which made $16.2 billion in loans for 17 percent of owner households in 2001, mortgage products, the homeownership rate in 2002, lenders originated a variety of but the Joint Center for Housing Studies reached an annual record of 67.9 percent in affordable products for first-time homebuyers reports that minorities were responsible for 2002, reaching 68.3 percent in the fourth and non-traditional borrowers that are then more than 40 percent (a total of 5.2 million) sold to Fannie Mae, Freddie Mac, quarter of 2002. This section discusses the of the net growth in homeowners between Countrywide, or other investors that are potential for further increases beyond those 1993 and 2002.142 As reported by the Fannie partnering with the ACB. It is reported that resulting from current demographic trends. Mae survey, 42 percent of African-American some lenders are making these non- The potential homeowner population over families reported that they were ‘‘very or traditional loans for the first time. the next decade will be highly diverse, as fairly likely’’ to buy a home in the next three For banks and thrifts, selling their CRA growing housing demand from immigrants years, up from 38 percent in 1998 and 25 loans will free up capital to make new CRA (both those who are already here and those percent in 1997. Among Hispanics and loans. As a result, the CRA market segment projected to come) and non-traditional Hispanic immigrants, the numbers reached provides an opportunity for Fannie Mae and homebuyers will help to offset declines in 37 percent and 34 percent respectively. The Freddie Mac to expand their affordable the demand for housing caused by the aging lending programs. Section E.3c below of the population. As noted in the above 2002 survey also reports that more than half presents data showing that purchasing discussion of CRA, many of these potential of Hispanic renters cite homeownership as targeted seasoned loans has been one strategy homeowners will be located in urban areas. being ‘‘one of their top priorities.’’ In that Fannie Mae has chosen to improve its Immigrants and other minorities—who addition, nearly a third (31 percent) of baby goals performance. Fannie Mae has been accounted for nearly 40 percent of the growth boomers said they are ‘‘very or fairly likely’’ offering CRA programs since mid-1997, when in the nation’s homeownership rate over the to buy a home in the next three years. it launched a pilot program, ‘‘Community past five years—will be responsible for In spite of these trends, potential minority Reinvestment Act Portfolio Initiative,’’ for almost two-thirds of the growth in the homebuyers see more obstacles to buying a purchasing seasoned CRA loans in bulk number of new households over the next ten home, compared with the general public. transactions, taking into account track record years (between 2000 and 2010), as well as Typically, the primary barriers to ownership as opposed to relying just on underwriting over the next 25 years (between 2000 and are credit issues and a lack of funds for a guidelines. Fannie Mae also started another 2025).138 By 2025, non-family households downpayment and closing costs. But Freddie pilot program in 1998, involving purchases of will make up a third of all households. Non- Mac staff emphasize that ‘‘immigrants and CRA loans on a flow basis, as they are Hispanic white and traditional households minorities face additional hurdles, including originated. By 2001, Fannie Mae was will contribute only one-third and one-tenth a lack of affordable housing, little investing $10.3 billion in initiatives targeted of the growth in new households, understanding of the home buying process, to aid financial institutions in meeting their respectively. Fannie Mae staff report that and continuing financial obligations in their CRA obligations. One CRA-eligible product between 1980 and 1995, the number of new home countries.’’ 143 In the Fannie Mae in 2002 included the MyCommunityMortgage immigrant owners increased by 1.4 million; survey, minority groups reported suite, which provides flexible product and between 1995 and 2010, that figure is misconceptions about the difficulty of options for low- to moderate-income expected to rise to by more than 50 percent becoming a homeowner such as beliefs about borrowers purchasing one- to four-unit to 2.2 million. These trends do not depend the amount of down payment required and homes.135 In 2002, Fannie Mae purchased or on the future inflow of new immigrants, as mortgage lending practices, a lack of securitized more than $882.5 million of immigrants don’t enter the housing market confidence about the homebuying process, MyCommunityMortgage products, which until they have been in this country for poor credit ratings, and language barriers. In helped provide affordable housing solutions eleven years. As noted by Fannie Mae staff, addition, there are continuing concerns about ‘‘there are enough immigrants already in this the limited education and low-income levels 132 Board of Governors of the Federal Reserve country to keep housing strong for at least six System. The Performance and Profitability of CRA- 139 Ibid. Related Lending. Washington, DC, 2000. 136 Fannie Mae, 2002 Annual Housing Activities 140 Fannie Mae, Fannie Mae National Housing 133 This discussion of urban lending draws from Report, p. 9. Survey, 2002, p. 6. Jeff Siegel, ‘‘Urban Lending Helps Increase Volume 137 Fannie Mae, 2002 Annual Housing Activities 141 Ibid. p. 8. and Meet CRA Requirements,’’ Secondary Report, p. 59. 142 Joint Center for Housing Studies of Harvard Marketing Executive, February 2003, pp. 21–23. 138 This section draws from ‘‘Immigration University, State of the Nation’s Housing 2003, p. 134 Ibid. Changes Won’t Hurt Housing,’’ Nation Mortgage 15. 135 Fannie Mae, (2002), p. 5. News, January 27, 2003, p. 8. 143 ‘‘Immigration Changes. * * *’’ Op. cit.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00060 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24287

of recent immigrants and other minorities. applicants. Applicants who would otherwise constrained applicants. The GSEs’ new Thus, the new group of potential be rejected by manual underwriting are being ‘‘timely reward’’ products for subprime homeowners will have unique needs. To tap qualified for mortgages with automated borrowers (discussed later) are integrated this potential homeowner population, the mortgage scoring in part because the with their mortgage scoring systems. mortgage industry will have to address these scorecard allows an applicant’s weaker areas Automated mortgage scoring presents the needs on several fronts, such as expanding to be offset by stronger characteristics. opportunity to remove discrimination from education and outreach efforts, introducing Typically, applicants whose projected mortgage underwriting, to accept all new products, and adjusting current monthly debt payment (mortgage payment applicants, and to bring fair, objective, underwriting standards to better reflect the plus credit card payment plus automobile statistically based competitive pricing, special circumstances of these new loan payment and so on) comprise a high greatly reducing costs for all risk groups. households. percentage of their monthly income would be Some institutions have sought to better The Bush administration has outlined a turned down by a traditional underwriting model and automate marginal and higher-risk plan to expand minority homeownership by system that relied on fixed debt-to-income loans, which have tended to be more costly 5.5 million families by the end of the decade. ratios (such as 36 percent). In a mortgage to underwrite and more difficult to The Joint Center for Housing Studies has scoring system, these same applicants might automate.147 stated that if favorable economic and housing be automatically accepted for a loan due to Along with the promise of benefits, market trends continue, and if additional their stellar credit record or to their ability however, automated mortgage scoring has efforts to target mortgage lending to low- to raise more cash for a down payment. The raised concerns. These concerns are related income and minority households are made, entity funding or insuring the mortgage (i.e., to the possibility of disparate impact and the the overall homeownership rate could reach a lender, private mortgage insurer, or a GSE) proprietary nature of the mortgage score 70 percent by 2010.144 allows these positive characteristics to offset inputs. The first concern is that low-income 7. Automated Underwriting Systems and the negative characteristics because its and minority homebuyers will not score well Mortgage Scorecards confidence in the ability of the empirically- enough to be accepted by the automated based mortgage scorecard to accurately underwriting system, resulting in their This, and the following two sections, identify those applicants who are more likely getting fewer loans. African-American and discuss special topics that have impacted the or less likely to eventually default on their Hispanic borrowers, for example, tend to primary and secondary mortgage markets in loan. have a poorer credit history record than other recent years. They are automated mortgage Automated mortgage scoring was borrowers, which means they are more likely scoring, subprime loans, and risk-based developed as a high-tech tool with the to be referred (rather than automatically pricing. The GSEs’ use of automated purpose of identifying credit risks in a more accepted) by automated mortgage scoring underwriting and mortgage scoring systems efficient manner. Automated mortgage systems that rely heavily on credit history was briefly discussed in the earlier section on scoring has grown as competition and measures such as a FICO score. There is also underwriting standards. This section decreased profit margins have created a significant statistical relationship between expands on issues related to automated demands to reduce loan origination costs. As credit history scores and the minority underwriting, a process that has spread a result, automated mortgage scoring has composition of an area, after controlling for throughout the mortgage landscape over the become the predominant (around 60 to 70 other locational characteristics.148 past five years, due mainly to the efforts of percent) mortgage underwriting method.146 The second concern relates to the ‘‘black Fannie Mae and Freddie Mac. As time and cost are reduced by the box’’ nature of the scoring algorithm. The According to Freddie Mac economists, automated system, the hope was that more scoring algorithm is proprietary and therefore automated mortgage scoring has enabled time would be devoted by underwriters to it is difficult for applicants to know the lenders to expand homeownership qualifying marginal loan applicants that are reasons for their scores. However, it should opportunities, particularly for underserved referred by the automated system for a more be noted that the GSEs have taken steps to 145 populations. There is growing evidence intensive, manual underwriting review. make their automated underwriting systems that automated mortgage scoring is more Fannie Mae and Freddie Mac are in the more transparent. Both Fannie Mae and accurate than manual underwriting in forefront of new developments in automated Freddie Mac have published the factors used predicting borrower risks. Mortgage mortgage scoring technology. Both to make loan purchase decisions in Desktop scorecards express the probability that an enterprises released automated underwriting Underwriter and Loan Prospector, applicant will default as a function of several systems in 1995—Freddie Mac’s Loan respectively. In response to criticisms aimed underwriting variables such as the level of Prospector and Fannie Mae’s Desktop at using FICO scores in mortgage down payment, monthly-payment-to-income Underwriter. Each system uses numerical underwriting, Fannie Mae’s new version of ratios, cash reserves, and various indicators credit scores, such as those developed by Desktop Underwriter (DU) 5.0 replaces credit of an applicant’s creditworthiness or credit Fair, Isaac, and Company, and additional scores with specific credit characteristics and history. Mortgage scorecards are statistically data submitted by the borrower, such as loan- provides expanded approval product estimated regression-type equations, based to-value ratios and available assets, to offerings for borrowers who have blemished on historical relationships between mortgage calculate a mortgage score that evaluates the credit. The specific credit characteristics foreclosures (or defaults) and the likelihood of a borrower defaulting on the include variables such as past delinquencies; underwriting variables. The level of down loan. The mortgage score is in essence a credit records, foreclosures, and accounts in payment and credit history indicators, such recommendation to the lender to accept the collection; credit card line and use; age of as a FICO score, are typically the most application, or to refer it for further review accounts; and number of credit inquiries.149 important predictors of default in mortgage through manual underwriting. Accepted With automated mortgage scoring replacing scoring systems. loans benefit from reduced document traditional manual underwriting comes the This increased accuracy in risk assessment requirements and expedited processing. fear that the loss of individual attention of mortgage scorecards has allowed risk As explained above, automated mortgage poses a problem for people who have managers to set more lenient risk standards, scoring allows tradeoffs between risk factors inaccuracies on their credit report or for and thus originate more loans to marginal to be quantified more precisely, providing members of cultural groups or recent the industry more confidence in ‘‘pushing immigrants who do not use traditional credit 144 Joint Center for Housing Studies of Harvard the envelope’’ of acceptable expected default and do not have a credit score. Some University, State of the Nation’s Housing 1998, p. rates. The GSEs’ willingness to offer low- subprime lenders and underwriters have 20. 145 down-payment programs was based on their claimed that their manual underwriting of Peter M. Zorn, Susan Gates, and Vanessa belief that their scoring models could Perry, ‘‘Automated Underwriting and Lending Outcomes: The Effect of Improved Mortgage Risk identify the more creditworthy of the cash- 147 Ibid. pp. 208–217. Assessment on Under-Served Populations. Program 148 Robert B. Avery, Raphael W. Bostic, Paul S. on Housing and Urban Policy,’’ Conference Paper 146 John W. Straka, ‘‘A Shift in the Mortgage Calem, and Glenn B. Canner, Credit Scoring: Issues Series, Fisher Center for Real Estate and Urban Landscape: The 1990s Move to Automated Credit and Evidence from Credit Bureau Files, mimeo, Economics. University of California Berkeley, 2001, Evaluations,’’ Journal of Housing Research, 2000, 1998, p. 24. p. 5. (11)2: p. 207. 149 Fannie Mae, September 4, 2002, p. 33.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00061 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24288 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

high-risk borrowers cannot be automated mortgage scoring. The lenders reported they Despite the gains in automated mortgage with mortgage scoring. Although automated formulated their policies and procedures to scoring and other innovations, minorities are mortgage scoring has greatly reduced the cost make certain that borrowers receive the best still less likely to be approved for a loan. The of many lower-risk loans that are easier to mortgage, according to product eligibility. difference in minority and non-minority rate, the cost of manually underwriting gray- This study will be further referenced in a accept rates may reflect greater social area and higher-risk applicants still remains following section regarding subprime inequities in financial capacity and credit, high.150 There is also the fear that applicants markets. which are integral variables in both manual who are referred by the automated system 2001 Freddie Mac Study. According to a and automated underwriting. In the future, will not be given the full manual Freddie Mac study published by the Fisher the accuracy of automated mortgage scoring underwriting for the product that they Center for Real Estate and Urban Economics will hinge on updating the models and initially applied for—rather they might be at University of California at Berkeley, making them more predictive while reducing pushed off to higher priced products such as underserved populations have benefited from the disparate impact on low-income and a subprime or FHA loan. In this case, the automated mortgage scoring because of the minority borrowers.156 The fairness of applicant may have had special increased ability to distinguish between a automated scoring systems will also depend circumstances that would have been clarified range of credit risks. In this paper, Freddie importantly on whether referred applicants by the traditional manual underwriting, thus Mac economists compared the manual and receive a traditional manual underwriting for enabling the applicant to receive a prime automated mortgage scoring approval rates of the loan that they initially applied for, rather loan consistent with his or her a sample of minority loans originated in than being immediately offered a higher creditworthiness. 1993–94 and purchased by Freddie Mac. priced loan that does not recognize their true Banking regulators and legal analysts While manual underwriters rated 51 percent creditworthiness. acknowledge the value of automated of the minority loans in the sample as accept, In addition to using automated mortgage scoring, although some skeptics automated mortgage scoring would have underwriting systems as a tool to help have noted concerns regarding fair lending, rated 79 percent of the loans as accept.154 determine whether a mortgage application potential fraud, privacy issues, and the In comparison to manual underwriting, should be approved, the GSEs’ automated ability of models to withstand changing this study found automated mortgage scoring underwriting systems are being further economic conditions.151 With the rise of not only less discriminatory but also more adapted to facilitate risk-based pricing. With automated mortgage scoring, the great accurate in predicting risk. Two versions of risk-based pricing, mortgage lenders can offer difference in Internet usage known as the Freddie Mac’s automated underwriting each borrower an individual rate based on ‘‘digital divide’’ could result in informational system, Loan Prospector (LP), were used to his or her risk. The division between the disadvantages for less educated and lower- review three groups of mortgage loans subprime and the prime mortgage market 155 income consumers. In addition to the digital purchased by Freddie Mac. The study will begin to fade with the rise of risk-based divide, the lack of financial literacy in the found that LP was a highly accurate predictor pricing, which is discussed in the next United States may also result in a disparate of mortgage default. The resulting improved section on the subprime market. impact on low-income and minority accuracy translates into benefits for borrowers.152 borrowers, who would otherwise be rejected 8. Subprime Lending 2002 Urban Institute Study. The Urban by manual underwriting to qualify for The subprime mortgage market provides Institute submitted a report to HUD in 2002 mortgages. mortgage financing to credit-impaired on subprime markets, the role of GSEs, and Analysis of the first group of loans showed borrowers—those who may have blemishes risk-based pricing.153 The study took a that loans rated as ‘‘caution’’ were four times in their credit record, insufficient credit preliminary look at the use of automated more likely to default than the average for all history, or non-traditional credit sources. underwriting systems for a small sample of loans. Minority borrowers whose loans were This section examines several topics related lenders. After conducting interviews with rated as ‘‘caution’’ were five times more to subprime lending including (a) the growth both subprime and prime lenders, the report likely to default, and low-income borrowers and characteristics of subprime loans, (b) the noted that all of the lenders in the study had whose loans were rated as ‘‘caution’’ were neighborhood concentration of subprime implemented some type of automated four times more likely to default than the lending, (c) predatory lending, and (d) underwriting system. These lenders stated average for all loans. The 2000 version of LP purchases of subprime mortgages by the that automated underwriting raised their approved 87.1 percent of loans generated GSEs. Section C.9 follows with a discussion business volume and streamlined their through affordable housing programs, of risk-based pricing. approval process. In addition, the lenders compared to a 51.6 percent approval rate reported they were able to direct more when the same loans were assessed using a. The Growth and Characteristics of underwriting resources to borderline manual underwriting procedures. Further, Subprime Loans applications despite an increase in business the study found LP more accurate than The subprime market has grown rapidly volume. manual underwriting at predicting default over the past several years, increasing from Even with the use of automated mortgage risk even with a higher approval rate. The an estimated $35 billion in 1994 to $160 scoring, the lenders in the study continued study also demonstrated that Freddie Mac’s billion in 1999 and $173.3 billion in 2001, to conduct at least a cursory review to year 2000 version of LP was more accurate before rising to $213 billion in 2002. The validate the application material. The in predicting risk than its 1995 version. subprime share of total market originations majority of the lenders still used manual Concluding Observations. Automated rose from 4.6 percent in 1994 to a high of 15 underwriting to originate loans not underwriting has enabled lenders to reach percent in 1999, and then fell to 8.5 percent recommended for approval with automated new markets and expand homeownership in both 2001 and 2002.157 Various factors opportunities, as illustrated by the 2001 have led to the rapid growth in the subprime Freddie Mac study. Increased accuracy with 150 Kenneth Temkin, Jennifer E.H. Johnson, and market: federal legislation preempting state Diane Levy, Subprime Markets, The Role of GSEs, automated mortgage scoring has led to the restrictions on allowable rates and loan and Risk-Based Pricing, Washington: The Urban development of new mortgage products that features, the tax reform act of 1986 which Institute. Report Prepared for the U.S. Department would have been previously considered too encouraged tax-exempt home equity of Housing and Urban Development, 2002. risky. For example, Freddie Mac uses Loan financing of consumer debt, increased 151 Allen J. Fishbein, ‘‘Is Credit Scoring a Winner Prospector to approve Alt A loans, which demand for and availability of consumer for Everyone?’’ Stone Soup, 2000, 14(3): pp. 14–15. tend to have nontraditional documentation; debt, a substantial increase in homeowner See also Fitch IBCA, Inc., Residential Mortgage A-minus loans, which pose a higher risk of Credit Scoring, New York, 1995 and Jim Kunkel, equity due to house price appreciation, and default; and other higher-risk mortgages, like a ready supply of available funds through ‘‘The Risk of Mortgage Automation,’’ in Mortgage 100 percent LTV loans. Both GSEs have and Banking, 1995, 57(8): pp. 69–76. continue to add new products to develop 152 156 Zorn et al., 2001, pp. 19–20. their automated underwriting systems to Ibid. pp. 18–19. 153 157 Kenneth Temkin, Jennifer E.H. Johnson, and reach more marginal borrowers. Subprime origination data are from Inside Diane Levy, Subprime Markets, The Role of GSEs, Mortgage Finance. For the 2002 estimates, see and Risk-Based Pricing, Washington: The Urban ‘‘Subprime Origination Market Shows Strong Institute. Report Prepared for the U.S. Department 154 Zorn, et al., 2001, pp. 14–15. Growth in 2002,’’ Inside B&C Lending, published by of Housing and Urban Development, 2002. 155 Ibid. p. 5. Inside Mortgage Finance, February 3, 2003, page 1.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00062 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24289

Wall Street securitization.158 It is important loan-to-value (LTV) ratios indicate, one loss analysis for the year 2000 yields the to note that subprime lending grew in the mitigation technique used by subprime following trends: 166 1990s mostly without the assistance of lenders is a high down payment requirement. • In 2000, 36 percent of refinance Fannie Mae and Freddie Mac. Some housing advocates have expressed mortgages in low-income neighborhoods Generally, there are three different types of concern that the perceptions about the risk of were subprime, compared with only 16 products available for subprime borrowers. subprime loans may not always be accurate, percent in upper-income neighborhoods. • These include: home purchase and refinance for example, creditworthy borrowers in inner Subprime lending accounted for 50 mortgages designed for borrowers with poor city neighborhoods may be forced to use percent of refinance loans in majority credit histories; ‘‘Alt A’’ mortgages that are subprime lenders because mainstream African-American neighborhoods—compared with only 21 percent in predominantly white usually originated for borrowers who are lenders are not doing business in their areas (less than 30 percent of population is unable to document all of the underwriting neighborhoods (see below). information but who may have solid credit African-American). Subprime borrowers are much more likely records; and high loan-to-value mortgages • The most dramatic view of the disparity to be low income and be a minority than originated to borrowers with fairly good in subprime lending comes from comparing credit. Fannie Mae and Freddie Mac are more other borrowers. Between 1999 and 2001, homeowners in upper-income African- likely to serve the first two types of subprime 43.1 percent of subprime loans in the American and white neighborhoods. Among borrowers.159 conventional conforming market went to homeowners living in the upper-income Borrowers use subprime loans for various low-income borrowers, compared with 29.5 white neighborhoods, only 16 percent turned purposes, which include debt consolidation, percent of conventional conforming loans. to subprime lenders in 2000. But 42 percent home improvements, and an alternative During that same period, 19.9 percent of of homeowners living in upper-income source of consumer credit. Between 1999 and subprime loans were for African-American African-American neighborhoods relied upon 2001, about two-thirds of subprime loans borrowers, compared with 6.5 percent of all subprime refinancing which is substantially were refinance loans. It has been estimated conventional conforming loans. However, more than the rate (30 percent) for that 59 percent of refinance loans were ‘‘cash what distinguishes subprime loans from homeowners living in low-income white out’’ loans.160 According to a joint HUD- other loans is their concentration in African- neighborhoods. Treasury report, first liens accounted for American neighborhoods. • Similar results are obtained when the analysis is conducted for borrowers instead more than three out of four loans in the b. The Neighborhood Concentration of of neighborhoods. Upper-income African- subprime market. Subprime Lending The subprime market is divided into American borrowers are twice as likely as different risk categories, ranging from least The growth in subprime lending over the low-income white borrowers to have risky to most risky: A-minus, B, C, and D. last several years has benefited credit- subprime loans. Over one-half (54 percent) of While there are no clear industry standards impaired borrowers as well as those low-income African-American borrowers for defining the subprime risk categories, borrowers who choose to provide little turn to subprime lenders, as does over one- Inside Mortgage Finance defines them in documentation for underwriting. However, third (35 percent) of upper-income African- terms of FICO scores—580–620 for A-minus, studies showing that subprime lending is American borrowers. By comparison, only 24 560–580 for B, 540–560 for C, and less than disproportionately concentrated in low- percent of low-income white borrowers and 540 for D. The A-minus share of the income and minority neighborhoods have 12 percent of upper-income white borrowers, subprime market rose from 61.6 percent in raised concerns about whether mainstream rely upon subprime lenders for their 167 2000 to 70.7 percent in 2001.161 For the first lenders are adequately serving these refinance loans. nine months of 2002, the A-minus share neighborhoods. A study of subprime lending It does not seem likely that these high accounted for 74 percent of the market, while in Chicago by The Woodstock Institute market shares by subprime lenders in low- the B share accounted for 11 percent, the C concluded that a dual, hyper-segmented income and African-American share accounted for 7.2 percent, and the D mortgage market existed in Chicago, as neighborhoods can be justified by a heavier share accounted for 7.9 percent of the mainstream lenders active in white and concentration of households with poor credit in these neighborhoods. Rather it appears market.162 upper-income neighborhoods were much less that subprime lenders may have attained Delinquency rates by type of subprime loan active in low-income and minority such high market shares by serving areas are as follows: 3.36 percent for A-minus neighborhoods—effectively leaving these loans, 6.67 percent for B, 9.22 percent for C, where prime lenders do not have a neighborhoods to unregulated subprime significant presence. The above finding that and 21.03 percent for D, according to the 165 lenders. As part of the HUD-Treasury Task upper-income black borrowers rely more Mortgage Information Corporation.163 Force on Predatory Lending, HUD’s Office of heavily on the subprime market than low- Because of their higher risk of default, Policy Development and Research released a income white borrowers suggests that a subprime loans typically carry much higher national level study—titled Unequal Burden: portion of subprime lending is occurring mortgage rates than prime mortgages. Recent Income and Racial Disparities in Subprime with borrowers whose credit would qualify quotes for a 30-year Fixed Rate Mortgage Lending in America—that showed families them for lower cost conventional prime were 8.85 percent for A-minus (with an 85 living in low-income and African-American loans. A lack of competition from prime percent LTV), 9.10 percent for B credit (with neighborhoods in 1998 relied lenders in low-income and minority an 80 percent LTV), and 10.35 percent for C disproportionately on subprime refinance neighborhoods has increased the chances credit (with a 75 percent LTV).164 As the low lending, even after controlling for that borrowers in these communities are neighborhood income. An update of that paying a high cost for credit. As explained 158 Temkin et. al, 2002, p.1. 159 Kenneth Temkin, Jennifer E.H. Johnson, Diane 165 166 See Randall M. Scheessele, Black and White Levy, Subprime Markets, The Role of GSEs, and Daniel Immergluck, The Predatory Lending Disparities in Subprime Mortgage Refinance Risk Based Pricing, Washington: The Urban Crisis in Chicago: The Dual Mortgage Market and Lending, Housing Finance Working Paper HF–014, Institute. Report Prepared for the Department of Local Policy, testimony before the Chicago City Office of Policy Development and Research, U.S. Housing and Urban Development, 2002, p. 4. Council, April 5, 2000. Immergluck found that subprime lenders received 74 percent of refinance Department of Housing and Urban Development, 160 U.S. Department of Housing and Urban applications in predominantly black tracts April 2002. Development/U.S. Department of the Treasury, compared to 21 percent in predominantly white 167 For an update to 2001, see The Association of Curbing Predatory Lending Report, 2000, p. 31. tracts in 1998. According to Immergluck, these Community Organizers for Reform Now (ACORN), 161 ‘‘Wholesale Dominates Subprime Market racial disparities provide evidence that the Separate and Unequal Predatory Lending in Through 3rd Quarter ’02,’’ Inside B&C Lending, residential finance market in Chicago is America, 2002. In 2001, subprime lenders published by Inside Mortgage Finance, December hypersegmented, resulting in the increased originated 27.8 percent of all conventional 16, 2002, pp. 1–2. likelihood that minorities receive mortgage credit refinance loans for African-Americans, 13.6 percent 162 Inside B&C Lending, November 16, 2002, p. 2. from a subprime, rather than a prime, lender in for Hispanic homeowners, and just 6.3 percent for 163 Mortgage Information Corporation, The Chicago. Also see Daniel Immergluck, Stark white homeowners. Overall, African-Americans Market Pulse, Winter 2001, pp. 4–6. Differences: The Explosion of the Subprime were 4.4 times more likely to use a subprime lender 164 Inside B&C Lending, published by Inside Industry and Racial Hypersegmentation in Home than whites, and Hispanics were 2.2 times more Mortgage Finance, February 17, 2003, page 13. Equity Lending, Woodstock Institute, October 2000. likely to do so.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00063 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24290 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

next, there is also evidence that the higher underwriting system.170 Fannie Mae has HUD and others in recent research studies.172 interest rates charged by subprime lenders stated that half of all mortgage borrowers These studies have found that foreclosures by cannot be fully explained solely as a function steered to the high-cost subprime market are subprime lenders grew rapidly during the of the additional risks they bear. Thus, a in the A-minus category, and therefore are 1990s and now exceed the subprime lenders’ greater presence by mainstream lenders prime candidates for Fannie Mae.171 share of originations. In addition, the studies could possibly reduce the high up-front fees c. Predatory Lending indicate that foreclosures of subprime loans and interest rates being paid by residents of occur much more quickly than foreclosures low-income and minority neighborhoods. Predatory lending has been a disturbing on prime loans, and that they are The Freddie Mac study presented evidence part of the growth in the subprime market. concentrated in low-income and African- that subprime loans bear interest rates that Although questions remain about its American neighborhoods. Of course, given are higher than necessary to offset the higher magnitude, predatory lending has turned the riskier nature of these loans, a higher credit risks of these loans.168 The study homeownership into a nightmare for far too foreclosure rate would be expected. With the compared (a) the interest rate on subprime many households. The growing incidence of information available it is not possible to loans rated A-minus by the lenders abusive practices has been stripping evaluate whether the disparities in originating these loans with (b) the interest borrowers of their home equity, threatening foreclosure rates are within the range of what rates on prime loans purchased by Freddie families with foreclosure, and destabilizing would be expected for loans prudently Mac and rated A-minus by a Freddie Mac neighborhoods. Also, in some cities, there are originated within this risk class. But findings underwriting model. Despite the fact that indications that unscrupulous realtors, from these studies about the high rate of both loan groups were rated A-minus, on mortgage brokers, appraisers, and lenders are mortgage foreclosure associated with average the subprime loans bore interest rates duping some FHA borrowers into purchasing subprime lending reinforce the concern that that were 215 basis points higher. Even homes at an inflated price or with significant predatory lending can potentially have devastating effects for individual families assuming that the credit risk of the subprime undisclosed repairs. The problems associated and their neighborhoods. loans was in fact higher than the prime loans, with home equity fraud and other mortgage At this time, there are open questions the study could not account for such a large abuses are not new ones, but the extent of about the effectiveness of the different discrepancy in interest rates. Assuming that this activity seems to be increasing. The expansion of predatory lending practices approaches being proposed for eradicating default rates might be three to four times predatory lending and the appropriate roles higher for the subprime loans would account along with subprime lending is especially troubling since subprime lending is of different governmental agencies—more for a 90 basis point interest rate differential. legislation versus increased enforcement of Assuming that servicing the subprime loans disproportionately concentrated in low- and very-low income neighborhoods, and in existing laws, long-run financial education would be more costly would justify an versus mortgage counseling, Federal versus additional 25 basis point differential. But African-American neighborhoods. The term ‘‘predatory lending’’ is a short state and local actions. In its recent issuance even after allowing for these possible of predatory lending standards for national differences, the Freddie Mac researchers hand term that is used to encompass a wide range of abuses. While there is broad public banks, the Office of the Comptroller of the concluded that the subprime loans had an Currency (OCC) cited the efforts of Fannie unexplained interest rate premium of 100 agreement that predatory lending should have no place in the mortgage market, there Mae and Freddie Mac’ in reducing predatory basis points on average.169 lending.173 The OCC advised banks against are differing views about the magnitude of Banking regulators have recognized the abusive practices, such as rolling single- the problem, or even how to define practices link between the growth in subprime lending premium life insurance into a loan. The that make a loan predatory. The joint HUD- and the absence of mainstream lenders and agency cited guidelines developed by Fannie Treasury report, Curbing Predatory Home have urged banks and thrifts that lending in Mae and Freddie Mac as a ‘‘useful reference’’ these neighborhoods not only demonstrates Mortgage Lending, concluded that a loan can or starting point for national banks. responsible corporate citizenship but also be predatory when lenders or brokers: charge Following publication of HUD’s proposed profitable lending. Ellen Seidman, former borrowers excessive, often hidden fees 2000 Rule inviting comments on disallowing Director of the Office of Thrift Supervision, (called ‘‘packing fees’’); successively goals credit for high cost mortgage loans, stated that, ‘‘Many of those served by the refinance loans at no benefit to the borrower Fannie Mae and Freddie Mac told lenders subprime market are creditworthy borrowers (called ‘‘loan flipping’’); make loans without they would no longer purchase loans with who are simply stuck with subprime loans or regard to a borrower’s ability to repay; and, certain abusive practices, such as excessive subprime lenders because they live in engage in high-pressure sales tactics or fees and failing to consider a borrower’s neighborhoods that have too few credit or outright fraud and deception. These practices ability to repay the debt. banking opportunities.’’ are often combined with loan terms that, It is important to re-emphasize that With respect to the question of whether alone or in combination, are abusive or make predatory lending generally occurs in borrowers in the subprime market are the borrower more vulnerable to abusive neighborhoods where borrowers have limited sufficiently creditworthy to qualify for more practices. Vulnerable populations, including access to mainstream lenders. While traditional loans, Freddie Mac has said that the elderly and low-income individuals, and predatory lending can occur in the prime one of the promises of automated low-income or minority neighborhoods, market, it is ordinarily deterred in that underwriting is that it might be better able to appeared to be especially targeted by market by competition among lenders, identify borrowers who are unnecessarily unscrupulous lenders. greater homogeneity in loan terms and assigned to the high-cost subprime market. One consequence of predatory lending is greater financial information among Freddie Mac has estimated that 10–30 that borrowers are stripped of the equity in their homes, which places them at an percent of borrowers who obtain mortgages 172 For an overview of these studies, see Harold in the subprime market could qualify for a increased risk of foreclosure. In fact, high L. Bunce, Debbie Gruenstein, Christopher E. conventional prime loan through Loan foreclosure rates for subprime loans provide Herbert, Randall M. Scheessele, Subprime Prospector, Freddie Mac’s automated the most concrete evidence that many Foreclosures: The Smoking Gun of Predatory subprime borrowers are entering into Lending, 2000. Also see Abt Associates Inc., mortgage loans that they simply cannot Analyzing Trends in Subprime Originations and 168 Howard Lax, Michael Manti, Paul Raca, and afford. The high rate of foreclosures in the Foreclosures: A Case Study of the Atlanta Metro Peter Zorn, ‘‘Subprime Lending: An Investigation of subprime market has been documented by Area, February 2000 and Analyzing Trends in Economic Efficiency,’’ February 25, 2000. Subprime Originations and Foreclosures: A Case 169 It should also be noted that higher interest Study of the Boston Metro Area, September 2000; rates are only one component of the higher cost of 170 Freddie Mac, We Open Doors for America’s National Training and Information Center, Preying subprime loans since borrowers also often face Families, Freddie Mac’s Annual Housing Activities on Neighborhoods: Subprime Mortgage Lenders and higher origination points. The Freddie Mac study Report for 1997, March 16, 1998, p. 23. Chicagoland Foreclosures, 2000; and the HUD did not find a large differential between prime and 171 Rommy Fernandez, ‘‘Fannie Mae Eyes Half of study, Unequal Burden in Baltimore: Income and subprime loans in points paid, but the study notes the Subprime Market,’’ in The American Banker, Racial Disparities in Subprime Lending, May 2000. that subprime loans often have points rolled into March 1, 2002. Also see ‘‘Fannie Mae Vows More 173 ‘‘OCC Cites Fannie, Freddie Predatory Lending the loan principal, which cannot be identified with Minority Lending,’’ Washington Post, March 16, Rules As Model,’’ Dow Jones Business News, their data. 2000, p. EO1. February 25, 2003.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00064 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24291

borrowers. Thus, one solution to address this reduced by 100 basis points if the borrower Mae, although a high proportion of borrowers problem would be to encourage more makes 24 consecutive monthly payments in the subprime market could qualify for less mainstream lenders to do business in our without a delinquency. Fannie Mae has costly prime mortgages, it remains unclear inner city neighborhoods. revamped its automated underwriting system why these borrowers end up in the subprime 184 d. Purchases of Subprime Mortgages by the (Desktop Underwriter) so loans that were market. Fannie Mae and Freddie Mac GSEs traditionally referred for manual believe they can bring more efficiency to the underwriting are now given four risk subprime market by creating standardized Fannie Mae and Freddie Mac have shown classifications, three of which identify underwriting and pricing guidelines in the increasing interest in the subprime market potential subprime (A-minus) loans.177 subprime market. An expanded GSE since the latter half of the 1990s. The GSEs Fannie purchased about $600 million of presence in the subprime market could be of entered this market by purchasing securities subprime loans on a flow basis in 2000.178 significant benefit to lower-income and backed by non-conforming loans. Freddie Fannie Mae securitized around $0.6 billion minority families if it attracted more Mac, in particular, increased its subprime in subprime mortgages in 2000, before mainstream lenders and competition to those business through structured transactions, increasing to $5.0 billion in 2001 and 7.3 inner-city neighborhoods that are currently with Freddie Mac guaranteeing the senior billion in 2002.179 served mainly by subprime lenders. classes of senior/subordinated securities. The In terms of total subprime activity (both Many subprime lenders do not think it is two GSEs also purchase subprime loans on flow and securitization activities), Fannie appropriate for Fannie Mae and Freddie Mac a flow basis. Fannie Mae began purchasing Mae purchased $9.2 billion in 2001 and over to increase their role in the subprime market subprime loans through its Timely Payment $15 billion in 2002, the latter figure because they do not view the subprime Reward Mortgage program in June 1999, and representing about 10 percent of the market, market as inefficient. Some officials in Freddie Mac rolled out a similar product, according to Fannie Mae staff.180 subprime mortgage markets claim the higher Affordable Merit Rate, in May 2000 A greater GSE role in the subprime lending prices paid by borrowers in the subprime (described below). In addition to purchasing market will most likely have a significant market appropriately reflect the risks that subprime loans for borrowers with blemished impact on the subprime market. Currently, come from extending credit to riskier credit, the GSEs also purchase another non- the majority of subprime loans are not borrowers. These officials believe it is unfair conforming loan called an Alternative-A or purchased by GSEs, and the numbers of for GSEs to enter an efficient, private market ‘‘Alt-A’’ mortgage. These mortgages are made lenders originating subprime loans typically that provides a necessary service to credit- to prime borrowers who do not want to do not issue a large amount of prime loans. impaired borrowers. Opponents of a larger provide full documentation for loans. The Partly in response to higher affordable GSE role in the subprime market argue GSEs GSEs’ interest in the subprime market has housing goals set by HUD in its new rule set have an unfair competitive advantage coincided with a maturation of their in 2000, the GSEs are increasing their because they can secure capital at cheaper traditional market (the conforming business in the subprime market. In the 2000 rates.185 Because the GSEs have a funding conventional mortgage market), and their GSE Rule, HUD identified subprime advantage over other market participants, development of mortgage scoring systems, borrowers as a market that can assist Fannie they have the ability to under price their which they believe allows them to accurately Mae and Freddie Mac in reaching their competitors and increase their market model credit risk. Although the GSEs account higher affordable housing goals while also share.186 This advantage, as has been the case for only a modest share of the subprime helping establish more standardization in the in the prime market, could allow the GSEs market today, some market analysts estimate subprime market. According to an Urban to eventually play a significant role in the that they could purchase as much as half of Institute study in 2002, many subprime subprime market. Many subprime lenders the overall subprime market in the next few lenders believe that successful companies fear they will be unfairly driven out of 174 years. serving high-risk borrowers need to have business as the GSEs increase their role in Precise information on the GSEs’ purchases specialized expertise in outreach, servicing, the subprime market. of subprime loans is not readily available. and underwriting, which is lacking among Data can be pieced together from various most prime lenders.181 These lenders do not 9. Risk-Based Pricing sources, but this can be a confusing exercise believe the more standardized approaches of The expanded use of automated because of the different types of non- prime lenders and the GSEs will work with underwriting and the initial uses of risk- conforming loans (Alt-A and subprime) and subprime borrowers, who require the more based pricing are changing the mortgage the different channels through which the customized and intensive origination and lending environment, often blurring the GSEs purchase these loans (through loan servicing processes currently offered by distinctions between the prime and subprime securitizations and through their ‘‘flow- experienced subprime lenders. market. Prime lenders are now using based’’ product offerings). Freddie Mac, As noted above, both Fannie Mae and automated underwriting systems that are which has been the more aggressive GSE in Freddie Mac make the claim that the being adapted to facilitate risk-based pricing. the subprime market, purchased subprime market is inefficient, pointing to For some time, the majority of prime approximately $12 billion in subprime loans evidence indicating that subprime borrowers mortgage borrowers have received loan rates during 1999—$7 billion of A-minus and pay interest rates, points, and fees in excess based on average cost pricing. Generally, alternative-A loans through its standard flow of the increased costs associated with serving borrowers receive roughly the same Annual programs and $5 billion through structured riskier borrowers in the subprime market. 182 Percentage Rate 187 (APR), regardless of the 175 transactions. In 2000, Freddie Mac A recent Freddie Mac study found automated risk of loss to the lender. The risk of all purchased $18.6 billion of subprime loans on mortgage scoring less discriminatory and borrowers is averaged together, and the price a flow basis in addition to another $7.7 more accurate in predicting risk than manual is determined by the average risk. billion of subprime loans through structured systems such as those currently used by In contrast, risk-based pricing enables 176 transactions. Freddie Mac securitized $9 subprime lenders.183 According to Fannie mortgage lenders to offer each borrower an billion in subprime and Alt-A product in individualized interest rate based on his or 2001 and $11.1 billion in 2002. 177 her risk. Or, more broadly, to offer interest Fannie Mae initiated its Timely Payments See Lederman, et al., Op cit. 178 rates based on whether or not the borrower product in September 1999, under which Kenneth Temkin, Jennifer E. H. Johnson, and Diane K. Levy, ‘‘Subprime Markets, the Role of borrowers with slightly damaged credit can GSEs, and Risk-Based Pricing,’’ Urban Institute, 184 qualify for a mortgage with a higher interest Fannie Mae, Remarks Prepared for Delivery by August 2001, p. 1. Franklin Raines, Chairman and CEO of Fannie Mae rate than prime borrowers. Under this 179 Inside Mortgage Finance’s, ‘‘Inside MBS & to the National Community Reinvestment Coalition. product, a borrower’s interest rate will be ABS,’’ December 15, 2000 and March 8, 2002. Washington, DC, March 20, 2000. 180 Statement by Mercy Jimenez of Fannie Mae in 185 Temkin et al., 2002, p. 1. 174 Temkin et al., 2002, p. 1. ‘‘Fannie Mae: Forges Ahead in Subprime,’’ 186 For an explanation of the GSEs funding 175 David A. Andrukonis, ‘‘Entering the Subprime Secondary Marketing Executive, February 2003, advantage see Government Sponsorship of FNMA Arena,’’ Mortgage Banking, May 2000, pp. 57–60. p.15. and FHLMC, United States Department of the 176 Subprime Lenders Mixed on Issue of GSE 181 Temkin et al., 2002, p. 1 Treasury, July 11, 1996. Mission Creep,’’ Inside B and C Lending, March 19, 182 See Lax et al., 2000. 187 Annual Percentage Rate takes into account 2001. 183 Zorn, et al., 2001, p. 5. points, fees, and the periodic interest rate.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00065 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24292 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

falls into a certain category of risk, such as Risk-based pricing also poses challenges on multifamily construction about unchanged, specific loan-to-value and FICO score the mortgage market because some of the vacancies have risen and rents have softened. combination or specified mortgage score more risky borrowers (who are currently Provision of decent housing affordable to range. Lenders could also set the interest rate cross-subsidized by less risky borrowers) may households of moderate or low incomes is a based on various factors including the not be able to afford their higher, risk-based challenge even in strong economic times, and probability of prepayment and characteristics interest rate. Also, the adoption of an with the unemployment rate up nearly two of the underlying collateral, as well as the automated risk-based pricing system may percentage points since late 2000, default risk of the borrower. Borrowers that have an uncertain effect on minority groups, affordability problems have increased for pose a lower risk of loss to the lender would who tend to have lower credit scores, as many, despite the softness in rents. then be charged a comparatively lower rate discussed earlier. On the other hand, if Despite the recent weakness in the than those borrowers with greater risk. Rather minorities are eligible for prime financing, apartment property market, the market for than lower risk borrowers cross-subsidizing the cost of financing minorities may fall as financing of apartments has grown to record higher risk borrowers like in average cost will the potential for subprime lenders to volumes. The favorable long-term prospects pricing, lower risk borrowers pay a relatively draw minorities to their higher-priced for apartment investments, combined with lower rate. products. record low interest rates, has kept investor In response to the expanded use of As the GSEs become more comfortable demand for apartments strong and supported automated underwriting and pressures from with subprime lending, the line between property prices. Refinancings too have the GSEs, other purchasers of loans, mortgage what today is considered a subprime loan grown, and credit quality has remained very insurers, and rating firms, lenders are versus a prime loan will likely deteriorate, high. Fannie Mae and Freddie Mac have been increasing their use of risk-based pricing.188 making expansion by the GSEs look more among those boosting volumes and In today’s markets, some form of differential like an increase in the prime market. This introducing new programs to serve the multifamily market. pricing exists for the various subprime melding of markets could occur even if many This section will review these market categories, for new products targeted at of the underlying characteristics of subprime developments, interpret the performance of credit-impaired borrowers (such as Fannie borrowers and the market’s evaluation of the Fannie and Freddie within that market Mae’s Timely Payments product), and for risks posed by these borrowers remain unchanged. Increased involvement by the context, and discuss future prospects for the private mortgage insurance across all credit multifamily rental market, its financing, and ranges. For example, private mortgage GSEs in the subprime market will result in more standardized underwriting guidelines the GSE role. The intention here is only to insurers use FICO scores and ‘‘Accept’’ update the discussion from 2000. For general determinations from the GSEs’’ automated and the increased participation of traditional lenders. In fact, there are indications that background information on the multifamily underwriting systems to make adjustments to mortgage market and the GSEs, see the 2000 insurance premiums.189 Rating agencies vary mainstream players are already increasing their activity in this market. According to Rule and the HUD-sponsored research report, subordination requirements based on the Study of Multifamily Underwriting and the credit qualify of the underlying collateral. staff from Moody’s Investors Service, the growing role of large mortgage aggregators in GSEs’ Role in the Multifamily Market (Abt Many believe there is cross-subsidization Associates, 2001). within the crude risk categories used in the subprime market has been a key factor in today’s market. For example, some of the the improving credit qualify on deals issued 2. The Multifamily Rental Housing Market: 191 better quality subprime borrowers in the A- in 2002. According to a representative 2000–2003 minus category may be inappropriately from Washington Mutual, subprime credit qualify has also improved as lenders carve The definition of ‘‘good’’ market conditions assigned to the subprime market. The GSEs in multifamily rental housing depends on and others are attempting to learn more about out new loan categories that fall somewhere between the large Alt A market and one’s perspective. Investors and lenders like the subprime market, and their initial efforts traditional subprime business.192 As the low vacancies, steady rent increases, and suggest that there will be an increase in the subprime market becomes more rising property values. Developers like strong use of risk-based pricing within this market, standardized, market efficiencies will reduce demand for new construction and favorable although it is recognized that the expansion borrowing costs. Lending to credit-impaired terms on construction financing. Consumers, of risk-based pricing depends importantly on borrowers will, in turn, increasingly make in contrast, prefer low rents and a wide these parties gaining a better understanding good business sense for the mortgage market. selection of available apartments. of the subprime borrower and the ability of The mid- to late-1990s were among the their mortgage scoring systems to predict risk C. Factor 2: Economic, Housing, and most successful of recent history, in that within this market. It must be noted that the Demographic Conditions: Multifamily apartment market conditions were generally power of the underlying algorithm in Mortgage Market good for all of these interest groups. automated underwriting systems determines Investment returns were favorable, the ability of these systems to accurately 1. Introduction construction volumes were steady at predict risk and set prices. At the time of the previous GSE sustainable levels, and many consumers had If prime lenders adopted risk-based rulemaking in 2000, the multifamily rental income gains in excess of their rent increases. pricing, many would be willing to lend to housing market was coming off several years Market conditions for multifamily rental riskier subprime borrowers because their risk of generally positive performance. Vacancies housing began to weaken toward the end of would now be offset with an increase in were low in most markets and rent increases 2000. Early warnings came from the publicly price. In theory, the mortgage market should were matching or exceeding economy-wide traded apartment companies, some of which expand because all mortgages will be inflation. A key to this strong performance reported easing in demand growth in the first approved at a price commensurate with risk, was the volume of new multifamily months of 2001, coinciding with a slowdown rather than setting a risk floor and approving construction, which was at a level consistent in job growth to its lowest level since 1992. no one beneath the floor. Risk-based pricing with demand growth. Job growth and income By the second quarter of 2001, most could also expand the prime lenders’ market gains helped many renters pay the higher apartment market indicators were reflecting by enabling them to reach a new group of rents without undue burden. As always, the slowdown. Vacancies were up, underserved customers.190 Taking advantage conditions varied from region to region, and approaching 10 percent for all multifamily of GSEs’ lower cost of capital, GSEs may be across market segments, but the overall tone (5+ units in structure) rental housing, able to offer borrowers who could not afford of the apartment market was quite healthy. according to the Census Bureau, and about a rate in the subprime market a rate they can Much has changed in the subsequent three half that rate among the larger apartment afford resulting from risk-based pricing. years. The economic slowdown has reduced properties monitored by private market apartment demand, and with new research firms. The FDIC’s Survey of Real 188 Temkin et al., 2002, p. 29. Estate Trends detected the first signs of 189 For example, see Radian’s product offerings at 191 ‘‘Improving Credit Quality, Maturing Business weakening in the first half of 2001, followed http://www.radiangroupinc.com. Stoke Confidence in Subprime MBS Market,’’ Inside by a big falloff in second half of the year and 190 Vanessa Bush, ‘‘Risk-Based Pricing Trend MBS & ABS, published by Inside Mortgage Finance, a continuing slide in the first half of 2002. Could Make Mortgage Lending More Efficient,’’ February 21, 2003. Apartments—especially those serving the America’s Community Banker, October 1, 1998. 192 Ibid. top end of the rental market—appear to have

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00066 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24293

performed worse than other rental housing in demand.196 The maturing of the ‘‘Baby Boom 27 metropolitan markets for which estimates the past four years, after several years of rent Echo’’ generation will increase the number of are available from the CPI ranged from a low growth and occupancies surpassing the rental persons under age 25 who will seek rental of ¥0.3 percent to a high of 6.7 percent in market averages. The multifamily vacancy housing, immigration is expected to continue the first half of 2003 relative to a year earlier. rate has increased more than the overall to fuel demand for rental housing, and And across the 75 metropolitan areas for rental market vacancy rate in each of the minority populations, while increasing their which rental vacancy rates (apartments plus years 2000, 2002, and 2003. In 2001, the homeownership rates, are growing and will other rentals combined) are available, rates vacancy rates increased at an equivalent rate. contribute to higher absolute demand for for the year 2002 ranged from 2.4 percent to For example, the Census Bureau’s estimate of rental housing. Thus demographic trends 15.4 percent, according to the Census Bureau. a 1.2 percentage point increase in vacancies support an improvement in the long-run In a historical context, this variation is for apartments in the year ending in the third demand for rental demand, which is likely to moderate, although up somewhat since the quarter of 2003 exceeds the overall rental include higher multifamily rental demand. late 1990s. vacancy rate of .9%. Similarly, while rent Supply growth has been maintained, even Conditions in the ‘‘affordable’’ segment of growth has decelerated slightly for all rental though the current reduced multifamily the apartment market are harder to track than in the high-end segment because of lesser housing according to the CPI, industry demand warrants less new construction. investor interest and analyst coverage. Data surveys of apartment rents show year-over- Total multifamily starts (2+ units) have been for the late 1990s analyzed by the National year declines in rents in many local running approximately 325-to-350 thousand 193 annually for the past six years, according to Housing Conference saw affordability markets. In 2003, asking rents remained problems continuing, although a study of flat nationally, as multifamily completions Census Bureau statistics, adding about 1 percent annually to the total multifamily apartment renters by the National Multi declined 5 percent.194 stock. Most of these new units are built for Housing Council saw some improvement in a. Apartment Demand and Supply rental use, with only about 20 percent in affordability during the strong economic 197 The primary reason for the softening in the recent years reported as being built as for-sale growth of 1997–1999. Other work noted multifamily rental market has been a condominium units. that rent to income ratios for the lowest income quintile of renters rose during the reduction in the growth of consumer demand The reduced short-term demand has shown late 1990s even as these ratios were stable or for apartment housing. The general through in absorption speeds for new declining for other renters.198 Harvard’s State slowdown in economic activity meant fewer apartments. The percentage of newly of the Nation’s Housing report for 2002 apartment customers, with less money, than completed unfurnished apartments rented within three months of completion fell from highlighted the variability of the affordability if the economy were vigorously expanding. problem from place to place.199 Persistent low interest rates have also enticed 71 percent during the first quarter of 2000 to 64 percent during the first quarter of 2001 Little research is available on affordability renters into the home purchase market as trends since 1999. However, tabulations from evidenced by the U.S. homeownership rate, and to 58 percent during the first quarter of 2002, according to the Census Bureau. This the 2001 American Housing Survey indicate which grew to 68.4 percent in 2003, further that income growth between 1999 and 2001 contributing to a weakness in rental demand. percentage rose slightly to 59 percent in the first quarter of 2003. in the lowest quintile of renter households The reduced demand is most evident in the continued to lag that of higher income national statistics on employment. Job b. Performance by Market Segments renters, and fell short of the average rent growth began decelerating in late 2000 and Some segments of the multifamily rental increases during this period. Together, these throughout 2001, turning negative late that market have been more affected than others statistics suggest that affordability has year. The largest year-over-year job loss of the by the recent softening. As mentioned earlier, deteriorated early this decade among at least economic downturn occurred in February the top end of the apartment market seems this group of very low-income renters. Other 2002, and year-over-year losses have especially hard hit, as measured by rising work using the AHS found that the number continued through October 2003. Apartment vacancies and reduced rent growth. This of low-to moderate-income working families demand seems particularly sensitive to labor segment is particularly dependent on job with severe rental cost burdens increased 24 market conditions, given the importance of growth and transfers for new customers, and percent between 1999 and 2001.200 rental housing to mobile individuals and is particularly vulnerable to losses of The low-income housing tax credit families accepting new jobs or transfers. Reis, residents and prospective customers to home (LIHTC) continues to finance much of the Inc., a real estate market research firm, purchase. According to reports by apartment newly built multifamily rental housing that estimates that the total number of occupied REITs and other investors, these top-end is affordable to households with moderate apartments (in properties with 40+ units) properties have not been getting the job- income. Restricted to households with actually declined in both 2001 and 2002 in related in-movers, but have still been losing incomes no greater than 60 percent of the the large markets nationwide that are a lot of customers to home purchase. These local median, this program financed monitored by the company.195 properties generally have annual resident approximately 75,000 units in 2001, Households, not jobs, fill apartments, and turnover rates of above 50 percent, and thus according to the National Council of State for this reason household formations are a are particularly quickly influenced by Housing Agencies, after running in the mid- preferable indicator of demand for changes in demand. Furthermore, this is the to high-60 thousand range the previous three apartments as well as other types of housing. segment of the apartment market into which years. About 70 percent of these units are The Census Bureau estimates that the total most of the new construction is built. newly built, and the rest are renovations of number of renter households nationwide has Performance has varied geographically as existing units. been essentially unchanged at approximately well. Some of the coastal markets, especially Expenditures for improvements to existing 34.8 million since 1996. Yet during the late in Northern California, saw the double-digit rental apartments have grown in recent years. 1990s apartment demand was expanding, rent increases of the late 1990s replaced by and apartments were apparently picking up double-digit declines, before stabilizing more 197 Center for Housing Policy/National Housing market share from other rental housing. The recently. ‘‘Supply constrained markets’’ had Conference, ‘‘Housing America’s Working Families: past two or three years may have seen a been preferred by apartment investors during A Further Exploration,’’ New Century Housing, Vol. reversal of that trend in share. the 1990s, but recent market performance has 3, No. 1, March 2002; Mark Obrinsky and Jill Long-term demographic trends are reminded investors and analysts that all Meron, ‘‘Housing Affordability: The Apartment expected to be favorable for rental housing markets have their day. For example, Universe,’’ National Multi Housing Council, 2002. Houston posted the biggest year-over-year 198 ‘‘Housing Affordability in the United States: rent increase of any major apartment market Trends, Interpretations, and Outlook,’’ a report 193 See, for example, Marcus & Millichap prepared for the Millennial Housing Commission by Research Services, National Apartment Report, in 2001, despite a long-run history of J. Goodman, November, 2001. January 2003. moderate rent growth and few barriers to new 199 Joint Center for Housing Studies of Harvard 194 Marcus & Millichap Research Services, apartment construction. Rent changes in the University, The State of the Nation’s Housing, 2002. National Apartment Report, January 2004. 200 Center for Housing Policy/National Housing 195 ‘‘Apartment Landlords Gather to Dreary 196 Mortgage Bankers Association of America, Conference, ‘‘America’s Working Families and the Outlook for Sector,’’ Wall Street Journal, January 15, ‘‘MBA News Link: Rental Market Demographics Housing Landscape 1997–2001,’’ New Century 2003, Section B. ‘‘Favorable,’’ Report Says,’’ January 2003. Housing, Vol. 3, No. 2, November 2002.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00067 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24294 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

In 2001 the total of $11.3 billion was nearly has been the lack of attractive returns on Realtors’ monthly estimates. But surveys by twice the figure of three years earlier, many financial assets and other alternative the National Multi Housing Council and according to the Census Bureau, and more investments. Despite the current weak other apartment industry reports indicate than a third as large as construction spending performance of apartments, investors that transactions volume dipped during 2001 for newly built multifamily structures, apparently are looking through to the long- and has since stabilized but not yet returned including owner-occupied condos. Many of run outlook for these assets, which is to the levels of the late 1990s. Even if the these improvements are to older properties in generally thought to be favorable, as number of transactions is off, the dollar high-demand neighborhoods. Improvements indicated most recently by investor surveys volume may well have risen, depending on to the physical structures have external fielded by the Urban Land Institute and by the mix and prices of properties sold. benefits. But often the renovations are in LendLease and PriceWaterhouseCoopers.201 Mortgage lending volumes have recently connection with re-positionings that move The net change in mortgage debt been boosted by shifts in property the apartments into a higher rent range and outstanding is defined as loan originations ownership. Publicly traded real estate bring changes in the demographic less repayments and charge offs. As investment trusts had been the big gainers composition of the resident base. discussed in Appendix D, net change is a during most of the 1990s, and by 1999 owned In 2002, expenditures on total lower bound on originations. By all accounts, nearly 6 percent of all apartments nationwide improvements to existing apartments originations—for which no single source of and a considerably larger share of all big declined to $9.8 billion, while new estimates is available—are much higher than (100+ unit) properties. But beginning in 1999 construction spending increased $2 billion. net change in most years. High levels of capital market developments made private This shift further suggests a re-positioning to refinancings of existing multifamily buyers more competitive. Since then the apartments with a higher rent range. mortgages in recent years has been a factor number of apartments owned by large REITs Excluding units financed with tax credits or in originations exceeding the net change in has declined about 5 percent, with diverse other subsidies, most of the multifamily debt outstanding. private interests apparently picking up rental construction in recent years has been Most mortgage lending is in the market share. targeted on the upper end of the market, ‘‘conventional’’ market. Multifamily loan Private investors are able to use more often the only segment for which programs of the Federal Housing leverage—greater debt—in financing their unsubsidized new construction is Administration accounted about $7 billion in transactions than the market permits the economically feasible. The median asking new insured mortgages in fiscal year 2003— public REITs. As a result, the very low rent on new unfurnished apartments up from $6 billion in fiscal year 2002 and $5 mortgage rates recently have given them an completed in 2001 was $877, up 11 percent billion in fiscal 2001. Despite the recent advantage in bidding on properties. In over the previous two years. In 2002 median increase in FHA originations, and the likely addition, equity funding costs of REITs rose asking rent for these properties was $905. Of continued strong performance for FHA as their stock prices flattened or moved down those units brought to market in 2002, 45 multifamily programs in the foreseeable as part of the broader equity market percent were at rents at or above $950. future,202 FHA remains but a small portion correction. of the total multifamily mortgage market. Refinancings have, by all accounts, also 3. Multifamily Financing Trends Outstanding FHA-insured multifamily been strong. Despite the lockout provisions In contrast to the softening observed in the mortgage debt was $55 billion at the end of and yield maintenance agreements that demand/supply balance for multifamily, the first quarter of 2003—only about 11 constrain early refinancings of many mortgage financing of these properties has percent of all multifamily mortgage debt multifamily loans, lenders reported very been at a record pace in the past three years. outstanding. strong refinancing activity in 2001 and a. Lending Volume Multifamily lending has been spurred by continuing into 2002. Although refinancing new apartment construction, property sales, volume data for the entire market are not Total multifamily mortgage debt and refinancings. New multifamily available, the trends in refinance volume for outstanding increased 9.5 percent in 2000 construction was valued at $32.6 billion in FHA and the GSEs show very strong (Q4/Q4), 11.4 percent in 2001, and 8.6 2002, according to the Census Bureau, up 14 increases in refinance activity during 2002 percent in 2002, according to the Federal percent from 2000. The number of new and 2003. For example, FHA’s Section Reserve’s Flow of funds accounts. This trend multifamily units completed over this period 223(a)(7) program, which is limited to continued through the third quarter of 2003, actually declined 6 percent, and the refinancing of FHA multifamily mortgages, which saw a 12.4 percent annualized increased expenditures reflect higher costs experienced an increase in origination increase. The dollar volumes were above per unit. The increase in asking rents volume of 133 percent in Fiscal Year 2003 those of any previous year, and far exceeded described earlier suggests higher property and 181 percent in Fiscal Year 2002. ($1.73 the lending volumes of all years other than values and greater debt carrying capacity. billion in FY 2003, $0.74 billion in FY 2002, 1998 and the frenzied period 1985–86. The b. Property Sales and Refinancings and $0.26 billion in FY 2001). Similarly, the pace has picked up slightly in 2003, with GSEs increased their combined volume of figures through the first two quarters Sales of existing apartment properties tend refinances by 83 percent from 1999–2000 to indicating annualized growth of about 9 to be pro-cyclical. Increasing asset values 2001–2002, from $17.6 billion to $32.1 percent. Furthermore, a survey by the bring buyers to the market and tempt sellers billion. Refinancings, especially when Mortgage Bankers Association of America to realize their capital gains. In soft markets, motivated by a desire to lower interest shows that of 48 member firms surveyed, in contrast, the bid-ask spread generally expense rather than to extract equity, do not representing all large mortgage banking firms widens and the volume of sales declines, as add as much to debt outstanding as do and a cross section of smaller mortgage sellers perceive current offers as beneath the purchase loans, which often are much larger companies, multifamily origination volume property’s long run value and buyers are than the seller’s existing mortgage that is increased by 16 percent in 2002—from $35 reluctant to pay for past performance or the repaid at the time of sale. Nonetheless, billion in 2001 to $41 billion in 2002. hope of future gains. Sales tend to increase refinancings represent a significant part of all The apparent inconsistency between mortgage debt, because the loan originated to multifamily mortgage lending. current market fundamentals and financing finance the purchaser’s acquisition is can be explained by low interest rates. The typically considerably larger than the c. Sources of Financing and Credit Quality same financial forces that lowered the mortgage retired by the seller. The sources of funding of multifamily mortgage rates for home purchasers to record No source of apartment property sales mortgages shifted somewhat in the past few lows by 2002 also reduced the financing statistics matches the comprehensive years, judging from the Flow of Funds costs of multifamily properties. The ten year national coverage of the single-family market accounts. As shown in Table A.4, four Treasury yield, a common benchmark for provided by the National Association of categories of lenders have dominated multifamily loan pricing, fell to a 45-year low multifamily mortgage lending since the mid- of 3.3 percent in June 2003 from 6.3 percent 201 Urban Land Institute, The ULI Forecast, 2002; 1990s. Of those four, commercial banks have as recently as the end of 1999. Lendlease and Prive WaterhouseCoopers, Emerging played a lesser, although still substantial, Another feature boosting investor demand Trends in Real Estate, 2003. role in recent years, providing 20 percent of for apartment properties and the resulting 202 Merrill Lynch, A New Look at FHA the $86 billion in net additional funding of demand for debt to finance those purchases Prepayments and Defaults, September 2002. multifamily mortgages during 2000 and 2001.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00068 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24295

The portfolio holdings of the GSEs, by growth in 2000–01 than in 1995–99, but net credit extended in 2000 and 2001, contrast, have been much more important between them still accounted for nearly half compared to all of it in the previous five-year than during the last half of the 1990s. of all the net credit extensions. Some slight period. Mortgage backed securities, both from the broadening of the base of multifamily lending BILLING CODE 4210–27–P GSEs and especially from other issuers, in the past two years, as these four lender accounted for proportionally less of the groups accounted for only 85 percent of the

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00069 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24296 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00070 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.009 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24297

The market values of apartment properties demand during the last half of the 1990s, current business activity show sharp gains of have generally held up well, although the construction never rose above its long-run over 70 percent in 2001, following small most recent indicators suggest some sustainable level, unlike the rampant decreases in activity in 2000. In 2002, the flattening. Properties in the portfolios of overbuilding that plagued the industry in the GSEs combined posted small declines for pension funds continued to appreciate into mid- and late-1980s. both measures. Measures of multifamily gross the second half of 2002, according to the mortgage purchases and new security National Council of Real Estate Investment 4. Recent GSE Involvement in Multifamily Finance issuance diverged for the two GSEs in 2002. Fiduciaries, although at a reduced annual Fannie Mae experienced declines in these rate of less than 2 percent. And the sales As the multifamily mortgage market has balance sheet and new business indicators in price per square foot of ‘‘Class A’’ properties expanded since 1999, Fannie Mae and 2002 while Freddie Mac experienced gains, monitored by Global Real Analytics rose Freddie Mac have increased their lending, until turning down in early 2002, posting a picked up market share, introduced new particularly in new security issuance. As 1.6 percent year over year decline in the programs, and enhanced others. discussed earlier, the credit quality of GSE second quarter. Beginning with their whole loans, the GSEs multifamily loans has remained very high The continuing value of collateral has added 34 percent to their combined holdings even with the large gains in loan volume. helped keep loan quality high on multifamily of multifamily loans in 2001, and another 26 Despite the substantial pickup in GSE mortgages. Delinquency rates from all major percent in 2002 (see Table A.6 below). The multifamily activity, the position of these reporters are at or near record lows, and well growth in multifamily MBS volume was companies in the multifamily mortgage below the rates reported for single-family nearly as dramatic, increasing 26 percent in market remains well below their dominance mortgages and commercial properties. At 2001 and another 14 percent in 2002. The in single-family mortgage finance. At the end commercial banks, the FDIC reports a non- gains resulted in the GSEs increasing their of 2002, the GSEs’ market share of single current loan percentage of 0.38 in the second share (whole loans and securities combined) family debt outstanding was 44 percent, quarter of 2002. In life insurance company of all multifamily debt outstanding to 22.8 twice the share of multifamily debt held or portfolios only .05 percent of residential percent by the third quarter of 2003, up from securitized by these two companies, mortgages were overdue at the end of 2002, 19 percent at year-end 2001, 15 percent at according to Federal Reserve statistics. and as of the third quarter of 2002 the GSEs year-end 1999 and 11 percent at the end of Furthermore, the multifamily share of all were both reporting similarly miniscule 1995. By this combined measure of portfolio housing units financed by the GSEs delinquency rates of below 0.1 percent; all of holdings and MBS outstanding, at year-end combined has declined from its 1997 level these rates are below those of a year earlier. 2002 Fannie Mae had nearly twice ($65 (Table A.5), although the annual statistics are Multifamily lenders have remained billion versus $37 billion) the multifamily heavily influenced by the volume of cautious in their underwriting and, together business of Freddie Mac, although Freddie refinancings in the single-family market, with their regulators, have avoided repeating was growing its multifamily business more the mistakes of the 1980s. Many of the senior rapidly (67 percent increase between 2000 which spiked in 1998 and again in 2001 and loan officers surveyed quarterly by the and 2002, compared to 46 percent increase 2002 in response to the big decline in Federal Reserve have reported tightening for Fannie Mae). mortgage rates in those years. Because of their terms on commercial mortgages, and Measures that focus on new multifamily lock-out agreements and other loan that shift likely has occurred in their activity, specifically gross mortgage purchase covenants, multifamily loans are not as prone multifamily lending as well. Perhaps the best volumes and new security issuance, vary to rate-induced refinancings as are single- indicator of discipline in multifamily lending across recent years and between the GSEs. family mortgages. is the fact that, despite the strong apartment For the GSEs combined, these measures of BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00071 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24298 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00072 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.010 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24299

a. Contrasting Business Models growth. As shown in Table A.6, most of had almost four dollars of outstanding MBSs While both Fannie Mae and Freddie Mac Fannie Mae’s multifamily growth has come for every dollar of portfolio holdings. Freddie have significantly increased their multifamily in MBS products, whereas Freddie Mac has Mac, on the other hand, more than three activities in recent years, they have pursued relied more on loans purchased and held in times as much volume in portfolio as it had distinct business models in achieving that its portfolio. At the end of 2002, Fannie Mae in MBS outstanding.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00073 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24300 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00074 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.011 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24301

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.203 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’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 multifamily rentals, as described above. Yet b. Affordable Multifamily Lending Servicers who are authorized to originate 15 percent of the units qualifying as low- and loans meeting Fannie Mae’s requirements for Because most of the GSEs’ multifamily moderate-income purchases were sale to the GSE without prior approval of lending is on properties affordable to multifamily, according to Table 1 of the individual transactions. These ‘‘DUS’’ households with low- or moderate incomes, GSEs’ activity reports for 2002. lenders retain part of the credit risk on the financing of affordable multifamily housing The GSEs increased the volume of their loans sold to Fannie. by the GSEs has increased almost as much as affordable multifamily lending dramatically Freddie Mac has taken a different approach their total multifamily lending. in 2001, the first year of the new, higher to credit underwriting. In the wake of large Approximately 86 percent of Fannie Mae’s affordable housing goals set for the GSEs. As credit losses on its multifamily business in multifamily lending volume in 2002 measured by number of units financed, the the late 1980s and 1990, Freddie Mac qualified as affordable to low- or moderate total affordable lending (shown in the ‘‘low- essentially withdrew from the market. When income households, according to Fannie mod total’’ rows of Table A.7) more than it re-entered in late 1993, the company Mae’s annual Housing Activity Report, as did doubled from a year earlier, especially after elected to retain all underwriting in-house 93 percent of Freddie Mac’s multifamily application of the upward adjustment factor and not delegate this function to the loan units financed. For the entire multifamily authorized for Freddie Mac in the 2000 Rule. originators participating in Freddie Mac’s rental market, HUD estimates that 90 percent In 2002, the GSEs maintained a high volume Program Plus network. Because Freddie of all housing units qualify as affordable to of affordable multifamily lending with families at 100 percent of the area median, Fannie Mae showing a slight decrease and 203 ‘‘No Mistaking GSEs for Twins in the standard upon which the low- and Freddie Mac a slight increase. Multifamily,’’ American Banker, October 2, 2002. moderate-income housing goal is defined. BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00075 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24302 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00076 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.012 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24303

The figures in Table A.7 are exclusive of to these striking numerical results. Delegated for 2001 and 2002 indicate attempts by the the ‘‘Temporary Adjustment Factor (TAF)’’ Underwriting and Servicing (DUS) is Fannie GSEs to promote market standards that will granted to Freddie Mac as part of the 2000 Mae’s principle product line for purchasing reduce the transactions costs of multifamily Rule. The TAF was a response to Freddie individual multifamily loans. This product lending while also providing programs that Mac’s limited opportunities for refinancing line is offered through 26 lenders with have the flexibility needed to deal with business because of its minimal involvement expertise in financing multifamily properties. unique circumstances. in the multifamily market in the early and In 2002, 92% of the DUS loan activity served 5. Future Prospects mid-1990s. 204 The TAF, which expired at affordable housing needs, 41% of DUS loans the end of 2003, provided a 20 percent in underserved markets, and 51% addressed The outlook for the multifamily rental upward adjustment to multifamily units in ‘‘special affordable’’ needs.207 Fannie Mae housing market is marked by near-term risks properties with 50 or more units, for markets its specialized 3MaxExpress product and longer-run optimism, according to most purposes of the affordable housing goals. line for loans worth less than or equal to $3 observers. The prospects for the next few Multifamily financing made major million. This program helped secure $4.1 quarters are dominated by the contributions not only to the GSEs billion in financing since 2001, which has macroeconomy. In particular, job growth, attainment of the overall goal for affordable assisted 130,000 families living in small with its implications for formations of lending in 2002, but also to the ‘‘underserved multifamily properties.208 Fannie Mae households, will be a key for the resumption areas’’ goal and ‘‘special affordable’’ goal. As additionally has federal Low-Income Housing of growth in apartment demand. Many shown in Table A.7, the 2001 increases in Tax Credit (LIHTC) programs and special forecasters would ascribe to the Federal lending in each of these categories were financing projects for special use properties Reserve’s forecast of a slight increase in GDP substantial at both Fannie Mae and Freddie such as Seniors Housing.209 growth to 4.3 percent in 2004,213 while also Mac, again leveling off for both in 2002. The During 2002, Freddie Mac used innovative agreeing with the Fed’s warning that ‘‘An GSEs also met the special multifamily financing structures combined with prudent, unusual degree of uncertainty attends the affordable subgoal set in the 2000 Rule in flexible multifamily lending practices, which economic outlook at present, in large both 2001 and 2002. were targeted at affordable initiatives through measure, but not exclusively, because of 214 c. Multifamily Initiatives of the GSEs its Program Plus network of lenders resulting potential geopolitical developments.’’ in record levels of multifamily mortgage When consumer demand does pick up, Fannie Mae and Freddie Mac have taken a recovery should be reasonably fast. While the number of steps since 2000 to expand their purchases. The GSEs face strong competition in this market from small banks and other recent production levels have outpaced multifamily lending and to respond demand, they have been near the middle of specifically to the goals established in the depository institutions that prefer to hold 210 the long run historical range and very close 2000 Rule. These initiatives are summarized these loans in their own portfolios. The 2000 Rule discussed other ways in to the average of the last half of the 1990s. in the annual activity reports filed by the Judging from the firm tone to rents and GSEs.205 which the GSEs might help promote financing of affordable multifamily housing. vacancies during that period, total One focus of the 2000 Rule was on lending multifamily completions production of to small (5-to-50 units) multifamily Two of those were lending for property rehabilitation and leadership in establishing 275,000 to 350,000 units is a sustainable properties, which the Rule identified as an level of annual production—that is, the level underserved market. HUD-sponsored standards for affordable multifamily lending. Many affordable properties are old and in consistent with long run demographic trends research has found that the supply of and replacement of units lost from the stock. mortgage credit to small properties was need of capital improvements if they are to remain in the housing stock. Rehabilitation Because new construction has remained impeded by the substantial fixed costs of moderate, there is no massive overhang of multifamily loan originations, by owners’ lending is a specialized field, and one in which the GSEs for a variety of reasons have product that will need to be absorbed. With insufficient documentation of property increased demand, vacancies should fall and not been major players. Less than 1 percent income and expense, and by the limited rents firm reasonably promptly. A key of all GSE multifamily lending in 2002 was opportunities for fees for underwriting and assumption behind this forecast for vacancies 206 for property rehabilitation. In 2002, Fannie servicing small loans. As a result, many and rents is that new apartment construction Mae hosted its first ever Preservation multifamily lenders focus on larger not rise appreciably from its current level. Advisory Meeting with leaders in the properties, which were found to have more Recovery in the apartment market may housing and real estate finance industry to loan products available to them and to pay also, perversely, be promoted by the recent identify best practices and formulate real lower interest rates than did small properties. unprecedented strength of the single-family world solutions to this critical policy In an attempt to promote the supply of market. Typically, economic recoveries bring issue.211 credit to small properties, the 2000 Rule strong growth in single-family housing provided incentives for the GSEs to step up Setting standards for affordable demand, some of that coming from apartment their involvement in this segment of the multifamily lending was identified in the renters seeking more space. With single- multifamily mortgage market. The incentives 2000 Rule as another area where the GSEs family activity already near record highs, likely contributed to the huge increases in could provide greater leadership. It was also boosted by historically low mortgage interest small property lending posted by both Fannie noted, based on HUD-sponsored research rates and despite the recently soft economy, 212 Mae and Freddie Mac in 2001 and continuing underway at that time, that market it is uncertain how much higher single- into 2002 (Table A.7). The combined total of participants believe the GSEs to be family demand—and the accompanying these units financed in 2001 and 2002 was conservative in their approaches to affordable losses of apartment customers to almost 8 times those financed in the previous property lending and underwriting. Actions homeownership—can go. two years. This lifted the percentage of all described in the GSEs’ annual activity reports Whenever the recovery comes, it will put GSE multifamily lending that was on small the multifamily rental market back onto a properties to their highest levels ever. 207 Fannie Mae, 2002 Annual Housing Activities long-run path that appears to promise Programs introduced or enhanced by the Report, 2003, p. 25. sustained, moderate growth. As discussed in GSEs in the past two years have contributed 208 Fannie Mae, 2002 Annual Housing Activities the 2000 Rule, the demographic outlook is Report, 2003, p. 25. favorable for apartment demand. Even if the 204 For background information on the Freddie 209 Fannie Mae, 2002 Annual Housing Activities homeownership rate increases further and Mac TAF, see pages 65054 and 65067–65068 of the Report, 2003, p. 26–27. the total number of renter households grows 210 2000 Rule. ‘‘Fannie Courting Multifamily Sellers; Small only slowly, as described in the discussion 205 Banks Balking,’’ American Banker, January 13, Fannie Mae’s 2002 Annual Housing Activities of the single-family housing market earlier in Report, pages 24–27; Freddie Mac’s Annual 2003. Housing Activities Report for 2002, pages 41–47. 211 Fannie Mae, 2002 Annual Housing Activities this Rule, apartment demand can be expected 206 Abt Associates Inc., An Assessment of the Report, 2003, p. 27. Availability and Cost of Financing for Small 212 Abt Associates, ‘‘Study of Multifamily 213 Federal Reserve, Survey of Professional Mulifamily Properties, a report prepared for the U.S. Underwriting and the GSEs’ Role in the Multifamily Forecasters, November 2003. Department of Housing and Urban Development, Market,’’ Final Report to the U.S. Department of 214 Board of Governors of the Federal Reserve Office of Policy Development and Research, August Housing and Urban Development, Office of Policy System, Monetary Policy Report to the Congress, 2001. Development and Research, August 2001. February 11, 2003, page 4.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00077 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24304 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

to increase more rapidly than that for other outlook of most market participants is for The main finding of this section rental housing, owing to the likely changes ample supply of mortgage financing at concerning the overall housing goals is that in age composition and reductions in average historically low interest rates.217 Yet both Fannie Mae and Freddie Mac surpassed household size. One estimate projects the complacency would be a mistake. the Department’s Low- and Moderate-Income annual growth in apartment households to be Responding to both market incentives and Housing Goals for each of the seven years one percent.215 their public charters, Fannie Mae and during this period. Specifically: • a. The Outlook for Multifamily Housing Freddie Mac can be expected to build on The goal was set at 40 percent for 1996; Supply their recent records of increased multifamily Fannie Mae’s performance was 45.6 percent lending and continue to be leaders in and Freddie Mac’s performance was 41.1 Regarding supply, one of the secrets of the financing volumes, in program innovations, percent. success of the multifamily sector during the and in standards setting. Certainly there is • The goal was set at 42 percent for 1997– 1990s was that production never rose above 2000. Fannie Mae’s performance was 45.7 its long-run sustainable level. The discipline room for expansion of the GSEs’ share of the multifamily mortgage market, which, as percent in 1997, 44.1 percent in 1998, 45.9 of developers, investors, and their lenders percent in 1999, and 49.5 percent in 2000; that brought that result needs to be continued mentioned earlier, is by the measure of dollar volume outstanding currently only about half and Freddie Mac’s performance was 42.6 if the apartment market is to maintain percent in 1997, 42.9 percent in 1998, 46.1 stability. the market share enjoyed by the GSEs in single-family lending. And from the percent in 1999, and 49.9 percent in 2000. Multifamily housing may benefit in the • In the October 2000 rule, the low- and future from more favorable public attitudes perspective of units financed, the statistics from Table A.5 combined with data from the moderate-income goal was set at 50 percent and local land use regulation. Higher density for 2001–03. As of January 1, 2001, several housing is a potentially powerful tool for 2001 American Housing Survey indicate that, while the GSEs financed 7.2 percent of all the changes in counting provisions took effect for preserving open space, reducing sprawl, and the low- and moderate-income goal, as promoting transportation alternatives to the nation’s year-round housing units that year, the percentage of multifamily rental units follows: ‘‘bonus points’’ (double credit) for automobile. The recently heightened purchases of goal-qualifying mortgages on attention to these issues may increase the (that is renter-occupied units and vacant rental units in structures with at least five small (5–50 unit) multifamily properties and, acceptance of multifamily rental construction above a threshold level, mortgages on 2–4 to both potential customers and their units) was only 5.7 percent. The sharp gains since 2000 in small unit owner-occupied properties; a prospective neighbors. ‘‘temporary adjustment factor’’ (1.20 units Provision of affordable housing will property lending by Fannie Mae and Freddie Mac demonstrate that it is feasible for this credit, subsequently increased by Congress to continue to challenge suppliers of 1.35 units credit) for Freddie Mac’s multifamily rental housing and policy important segment of the affordable housing market to be served by the GSEs. Building on purchases of goal-qualifying mortgages on makers at all levels of governments. Low large (more than 50 units) multifamily incomes combined with high housing costs the expertise and market contacts gained in the past three years, the GSEs should be able properties; changes in the treatment of define a difficult situation for millions of missing data; a procedure for the use of to make even greater in-roads in small renter households. Housing cost reductions imputed or proxy rents for determining goal property lending, although the challenges are constrained by high land prices and credit for multifamily mortgages; and noted earlier will continue. construction costs in many markets. eligibility of purchases of certain qualifying The GSEs’ size and market position Government action—through land use government-backed loans to receive goal between loan originators and mortgage regulation, building codes, and occupancy credit. These changes are explained below. investors makes them the logical institutions standards—are major contributors to those Fannie Mae’s low-mod goal performance was to identify and promote needed innovations high costs, as is widely recognized by market 51.5 percent in 2001 and 51.8 percent in and to establish standards that will improve participants, including the leaders of the 2002, and Freddie Mac’s performance was 216 market efficiency. As their presence in the GSEs. Reflecting the preferences of the 53.2 percent in 2001 and 51.4 percent in electorate, these regulated constraints are multifamily market continues to grow, the 2002, thus both GSEs surpassed this higher unlikely to change until voter attitudes GSEs will have both the knowledge and the goal in both years. This section discusses the change. ‘‘clout’’ to push simultaneously for market October 2000 counting rule changes in detail b. The Future Role of the GSEs standardization and for programmatic below, and provides data on what goal flexibility to meet special needs and Regarding the mortgage financing of performance would have been in 2001–02 circumstances, with the ultimate goal of without these changes.219 multifamily rental apartments, it is hard to increasing the availability and reducing the anticipate events that might disrupt the flow After the discussion of the overall housing cost of financing for affordable and other goals in Sections E.1 to E.5, Sections E.6 to or alter the sources of mortgage credit to multifamily rental properties. apartments. In the past, certain events have E.12 examine the role of the GSEs in funding triggered such changes—notably the savings E. Factor 3: Performance and Effort of the home purchase loans for lower-income and loan debacle of the 1980s and Freddie GSEs Toward Achieving the Low- and borrowers and for first-time homebuyers. A Mac’s withdrawal from the market following Moderate-Income Housing Goal in Previous summary of the main findings from that large losses in the early 1990s—but these are, Years analysis is given in Section E.6. Section E.13 by definition, surprises. The current structure then summarizes some recent studies on the This section first discusses each GSE’s GSEs’ market role and section E.14 discusses and performance of the multifamily mortgage performance under the Low- and Moderate- market provide some comfort that the risks the GSEs’ role in the financing of single- Income Housing Goal over the 1996–2002 family rental properties. are slight. The lender base is not overly period.218 The data presented are ‘‘official dependent on any one institution or lender results’’—i.e., they are based on HUD’s 1. Performance on the Low- and Moderate- type for either loan originations or funding. analysis of the loan-level data submitted to Income Housing Goal in 1996–2002 Lending discipline appears to have been the Department by the GSEs and the counting maintained, given the low mortgage HUD’s December 1995 rule specified that provisions contained in HUD’s regulations in in 1996 at least 40 percent of the number of delinquency rates even during the weak 24 CFR part 81, subpart B. As explained economy of the past two years. The near term units financed by each of the GSEs that were below, in some cases these ‘‘official results’’ eligible to count toward the Low- and differ from goal performance reported by the Moderate-Income Goal should qualify as low- 215 Jack Goodman, ‘‘The Changing Demography of GSEs in the Annual Housing Activities or moderate-income, and at least 42 percent Multifamily Rental Housing,’’ Housing Policy Reports (AHARs) that they submit to the of such units should qualify in 1997–2000. Debate, Winter 1999. Department. 216 Remarks by Franklin D. Raines, Chairman and HUD’s October 2000 rule made various CEO, Fannie Mae, to the Executive Committee of the National Association of Home Builders, January 217 ‘‘Capital Markets Outlook 2003,’’ Apartment 219 To separate out the effects of changes in 18, 2003. See also Edward Glaeser and Joseph Finance Today, Vol. 7, No. 1 (January/February counting rules that took effect in 2001, this section Gyourko, ‘‘The Impact of Zoning on Housing 2003). also compares performance in 2001 to estimated Affordability,’’ Working Paper 8835, National 218 Performance for the 1993–95 period was performance in 2000 if the 2001 counting rules had Bureau of Economic Research, March 2002. discussed in the October 2000 rule. been in effect in that year.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00078 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24305

changes in the goal counting rules, as percentage points and 3.7 percentage points point. Freddie Mac showed a gain in discussed below, and increased the Low- and in 1996 and 1997, respectively, while performance to 46.1 percent in 1999, Moderate-Income Goal to 50 percent for Freddie Mac surpassed the goals by narrower exceeding its previous high by 3.2 percentage 2001–03. margins, 1.1 and 0.6 percentage points. points. Fannie Mae’s performance in 1999 Table A.8 shows low-mod goal During the heavy refinance year of 1998, was 45.9 percent, which, for the first time, performance over the 1996–2002 period, Fannie Mae’s performance fell by 1.6 slightly lagged Freddie Mac’s performance in based on HUD’s analysis. The table shows percentage points, while Freddie Mac’s that year. that Fannie Mae surpassed the goals by 5.6 performance rose slightly, by 0.3 percentage BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00079 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24306 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00080 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.013 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24307

Both GSEs exhibited sharp gains in goal In 2002, Freddie Mac’s performance on the determine whether or not a property is performance in 2000—Fannie Mae’s low mod-goal (51.4 percent) fell short of located in an underserved area, all units in performance increased by 3.6 percentage Fannie Mae’s performance (51.8 percent), such a property are included in the points, to a record level of 49.5 percent, even though Freddie Mac had the advantage denominator, but not in the numerator, in while Freddie Mac’s performance increased of the Temporary Adjustment Factor. The gap calculating performance on this goal. even more, by 3.8 percentage points, which would have been wider without this factor, • Missing data and proxy rents for also led to a record level of 49.9 percent. and in fact Freddie Mac’s performance would multifamily properties. In the past, if a GSE Fannie Mae’s performance was 51.5 percent have been short of the goal, at 49.2 percent. lacked data on rent for rental units in in 2001 and 51.8 percent in 2002; Freddie multifamily properties whose mortgages it Mac’s performance was 53.2 percent in 2001 2. Changes in the Goal Counting Rules for 2001–03 purchased, such units were included in the and 51.4 percent in 2002. However, as denominator, but not in the numerator, in discussed below, using consistent accounting A number of changes in the counting rules calculating goal performance. Since some of rules for 2000–02, each GSE’s performance in underlying the calculation of low- and these units likely would have qualified for 2001–02 was below its performance in 2000. moderate-income goal performance took one or more of the housing goals, this rule The official figures for low-mod goal effect beginning in 2001, as follows: performance presented above differ from the • Bonus points for multifamily and single- lowered goal performance. Under the new corresponding figures presented by Fannie family rental properties. During the 2001–03 counting rules that took effect in 2001, if rent Mae and Freddie Mac in their Annual period the Department awarded ‘‘bonus is missing for multifamily units, a GSE may Housing Activity Reports to HUD by 0.2–0.3 points’’ (double credit in the numerator) for estimate ‘‘proxy rents,’’ and, up to a ceiling percentage point in both 1996 and 1997, goal-qualifying units in small (5–50 unit) of 5 percent of total multifamily units reflecting minor differences in the multifamily properties and, above a financed, may apply these proxy rents in application of counting rules. These threshold, 2–4 unit owner-occupied determining whether such units qualify for differences also persisted for Freddie Mac for properties whose loans were purchased by the low- and moderate income goal and 1998–2000, but the goal percentages shown the GSEs. By letters dated December 24, special affordable goal. If such proxy rents above for Fannie Mae for these three years 2003, the Department notified the GSEs that cannot be estimated, these multifamily units are the same as the results reported by Fannie these bonus points would not be in effect are excluded from the denominator in Mae to the Department. Fannie Mae reported after December 31, 2003. calculating performance under these goals. its performance in 2001 as 51.6 percent and • Freddie Mac’s Temporary Adjustment No multifamily loans can be excluded from Freddie Mac reported its performance as 53.6 Factor. As part of the Consolidated the denominator in calculating performance percent—both were slightly above the Appropriations Act of 2000, Congress on the underserved areas goal—that is, if a corresponding official figures of 51.5 percent required the Department to award 1.35 units GSE does not have sufficient information to and 53.4 percent, respectively. For 2002, of credit for each unit financed in ‘‘large’’ determine whether or not a property is Fannie Mae’s reported performance was the multifamily properties (i.e., those with 51 or located in an underserved area, all units in same as reported by HUD (51.8 percent), more units) in the numerator in calculating such a property are included in the while Freddie Mac’s reported performance performance on the housing goals for Freddie denominator, but not in the numerator, in was 51.3 percent, slightly below HUD’s Mac for 2001–03.220 This ‘‘temporary calculating performance on this goal. official figure of 51.4 percent. adjustment factor’’ (TAF) did not apply to • Fannie Mae’s performance on the Low- and Purchases of certain government-backed goal performance for Fannie Mae during this Moderate-Income Goal was in the range loans. Prior to 2001, purchases of period. By letters dated December 24, 2003, between 44 percent and 46 percent between government-backed loans were not taken into the Department notified Freddie Mac that 1996 and 1999, but jumped sharply in just account in determining performance on the this factor would not be in effect after one year, from 45.9 percent in 1999 to 49.5 GSEs’ low- and moderate-income and December 31, 2003. underserved area housing goals. That is, all percent in 2000. Freddie Mac’s performance • was in the range between 41 percent and 43 Missing data for single-family properties. such loans were excluded from both the percent between 1996 and 1998, and then In the past, if a GSE lacked data on rent for numerator and the denominator in rose to 46.1 percent in 1999 and 49.9 percent rental units or on borrower income for calculating goal performance on these two in 2000. As discussed above, official owner-occupied units in single-family goals, and in accordance with Section performance rose for both GSEs in 2001–02, properties whose mortgages it purchased, 1333(b)(1)(A) of the Federal Housing but this was due to one-time changes in the such units were included in the Enterprises Financial Safety and Soundness counting rules—abstracting from counting denominator, but not in the numerator, in Act of 1992, purchases of only certain rule changes, performance fell for both GSEs. calculating goal performance. Since some of government-backed loans were included in Fannie Mae’s performance on the Low- and these units likely would have qualified for determining performance on the GSEs’ Moderate-Income Goal surpassed Freddie one or more of the housing goals, this rule special affordable goals. In October 2000 the Mac’s in every year through 1998. This lowered goal performance. Under the new Department took steps to encourage the pattern was reversed in 1999, as Freddie Mac counting rules for the low- and moderate- enterprises to play more of a role in the income goal and the special affordable goal surpassed Fannie Mae in goal performance secondary market for several types of that took effect in 2001, the GSEs are allowed for the first time, though by only 0.2 government-backed loans where it appeared to exclude loans with missing borrower percentage point. This improved relative that greater GSE involvement could increase income from the denominator if the property performance of Freddie Mac was due to its the liquidity of such mortgages. Home equity is located in a below-median income census increased purchases of multifamily loans, as conversion mortgages (HECMs) were it re-entered that market, and to increases in tract. This exclusion is subject to a ceiling of 1 percent of total owner-occupied units developed in the late-1980s by the Federal the goal-qualifying shares of its single-family Housing Administration (FHA); these mortgage purchases. Freddie Mac’s financed. The enterprises are also allowed to exclude single-family rental units with mortgages allow senior citizens to draw on performance also slightly exceeded Fannie the equity in their homes to obtain monthly Mae’s performance in 2000, 49.9 percent to missing rental information from the payments to supplement their incomes. Thus 49.5 percent. Freddie Mac’s official denominator in calculating performance for purchases of FHA-insured HECMs now count performance also exceeded Fannie Mae’s these two goals; there is no ceiling or toward the low- and moderate-income official performance in 2001, but this restriction to properties located in below- reflected a difference in the counting rules median income census tracts for this housing goals if the mortgagor’s income is applicable to the two GSEs that was enacted exclusion of single-family rental units. No less than median income for the area. by Congress; if the same counting rules were single-family loans can be excluded from the Similarly, purchases of mortgages on applied to both GSEs (that is, Freddie Mac denominator in calculating performance on properties on tribal lands insured under did not receive the 1.35 Temporary the underserved areas goal—that is, if a GSE FHA’s Section 248 program or HUD’s Section Adjustment Factor), Fannie Mae’s does not have sufficient information to 184 program may qualify for the GSEs’ performance would have exceeded Freddie housing goals. And purchases of mortgages Mac’s performance, by 51.5 percent to 50.5 220 See Congressional Record, December 15, 2000, under the Rural Housing Service’s Single percent. pp. H12295–96. Family Housing Guaranteed Loan Program

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00081 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24308 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

may also count toward all of the housing 3. Effects of Changes in the Counting Rules rules; this may be compared with official goals.221 on Goal Performance in 2001–02 performance in 2001–02—an ‘‘apples-to- Because of the changes in the low- and apples comparison.’’ HUD has also calculated moderate-income goal counting rules that what performance would have been in 2001– took effect in 2001, direct comparisons 02 under the 1996–2000 rules; this may be between official goal performance in 2000 compared with official performance in and 2001–02 are somewhat of an ‘‘apples-to- 2000—an ‘‘oranges-to-oranges comparison.’’ 221 Prior to the October 2000 rule, purchases of oranges comparison.’’ For this reason, the These comparisons are presented in Table these government-backed mortgages were only Department has calculated what performance A.9. eligible for credit under the special affordable goal. would have been in 2000 under the 2001–03 BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00082 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24309

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00083 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.014 24310 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Specifically, Table A.9 shows performance • Freddie Mac. The largest impact of the the housing goals was counted as two units under the low- and moderate-income goal in counting rule changes on Freddie Mac’s goal in the numerator (and one unit in the three ways. Baseline A represents performance was due to the application of denominator) in calculating goal performance performance under the counting rules in the temporary adjustment factor for for that goal. For example, if a GSE financed effect in 1996–2000. Baseline B incorporates purchases of mortgages on large multifamily a mortgage on a 40-unit property in which 10 the technical changes in counting rules— properties, as enacted by Congress; this of the units qualified for the low- and changes in the treatment of missing data added 2.7 percentage points to goal moderate-income goal, 20 units would be (including use of proxy rents), and eligibility performance in 2001, as shown in Table A.9. entered in the numerator and 40 units in the for the goals of certain government-backed Bonus points for purchases of mortgages on denominator for this property in calculating loans. Baseline C incorporates in addition to small multifamily properties added 1.5 goal performance. the technical changes the bonus points and, percentage points to performance, and bonus Small multifamily bonus points thus for Freddie Mac, the temporary adjustment points for purchase of mortgages on owner- encouraged the GSEs to play a larger role in factor. Baseline B corresponds to the occupied 2–4 unit rental properties added 1.4 this market, and also to purchase mortgages counting approach proposed in this rule to percentage points to performance. The on such properties in which large shares of take effect in 2005. Boldface figures under remaining impact (0.5 percentage point) was the units qualified for the housing goals. Baseline A for 1999–2000 and under Baseline due to technical changes in counting rules— Some evidence may be gleaned from the data C for 2001–02 indicate official goal primarily, the exclusion of single-family provided to HUD by the GSEs for 2001–02. performance, based on the counting rules in units with missing information from the Fannie Mae financed 37,403 units in small effect in those years—e.g., for Fannie Mae, denominator in calculating goal performance. multifamily properties in 2001 that were 45.9 percent in 1999, 49.5 percent in 2000, Credit for purchases of qualifying eligible for the low- and moderate-income 51.5 percent in 2001, and 51.8 percent in government-backed loans played a minor role goal, and 58,277 such units in 2002, a two- 2002. in determining Freddie Mac’s goal year increase of more than 700 percent from • Performance on the Low- and Moderate- performance. These same patterns also the 7,196 such units financed in 2000. Small Income Goal under 1996–2000 Counting appeared in 2002. multifamily properties also accounted for a Rules Plus Technical Changes. If the • Fannie Mae. The temporary adjustment greater share of Fannie Mae’s multifamily ‘‘Baseline B’’ counting approach had been in factor applies to Freddie Mac’s goal business in 2001–02—7.4 percent of total effect in 2000–02 and the GSEs had performance, but not to Fannie Mae’s multifamily units financed in 2001 and 13.2 purchased the same mortgages that they performance, thus counting rule changes had percent in 2002, up from 2.5 percent in 2000. actually did purchase in those years, both less impact on its performance than on However, HUD’s 2000 rule reported Fannie Mae and Freddie Mac would have Freddie Mac’s performance in 2001. The information from the 1991 Residential surpassed the low- and moderate-income largest impact of the counting rule changes Finance Survey that small multifamily goal in 2000 and fallen short in 2001 and on Fannie Mae’s goal performance was due properties accounted for 37 percent of all 2002. Specifically, Fannie Mae’s performance to the application of bonus points for multifamily units, thus Fannie Mae was still would have been 51.3 percent in 2000, 49.2 purchases of mortgages on owner-occupied less active in this market than in the market percent in 2001, and 49.0 percent in 2002. 2–4 unit rental properties, which added 1.6 for large multifamily properties.223 Freddie Mac’s performance would have been percentage points to performance, and for Within the small multifamily market, there 50.6 percent in 2000, 47.7 percent in 2001, purchases of mortgages on small multifamily was no evidence that Fannie Mae targeted and 46.5 percent in 2002. properties, which added 0.7 percentage point affordable properties to a greater extent in • Performance on the Low- and Moderate- to performance. The remaining impact (1.3 2001–02 than in 2000. That is, 87 percent of Income Goal under 2001–2003 Counting percentage points) was due to technical Fannie Mae’s small multifamily units Rules. If the 2001–03 counting rules had been changes—primarily, the exclusion of single- qualified for the low- and moderate-income in effect in 2000–02 and the GSEs had family units with missing information from goal in 2000; this fell to 75 percent in 2001, purchased the same mortgages that they the denominator in calculating goal but rose to 89 percent in 2002. actually did purchase in those years (i.e., performance.222 Credit for purchases of Freddie Mac financed 50,299 units in small abstracting from any behavioral effects of qualifying government-backed loans and the multifamily properties in 2001 that were ‘‘bonus points,’’ for example), both GSEs use of proxy rent for multifamily properties eligible for the low- and moderate-income would have substantially surpassed the low- played a minor role in determining Fannie goal and 42,772 such units in 2002, a two- and moderate-income goal in all three years, Mae’s goal performance. These same patterns year increase of more than 1300 percent from but both GSEs’ performance figures would also appeared in 2002 for Fannie Mae. the 2,996 units financed in 2000. Small have deteriorated somewhat from 2000 to multifamily properties also accounted for a 2001, and, for Freddie Mac, from 2001 to 4. Bonus Points for the Low- and Moderate- significantly greater share of Freddie Mac’s 2002. Specifically, Fannie Mae’s ‘‘Baseline Income Goal multifamily business in 2001—16.1 percent C’’ performance would have been 52.5 As discussed above, the Department of total multifamily units financed in 2001 percent in 2000, 51.5 percent in 2001, and established ‘‘bonus points’’ to encourage the and 13.4 percent in 2002, up from 1.8 percent 51.8 percent in 2002. Freddie Mac’s GSEs to step up their activity in 2001–03 in in 2000. performance would have been 55.1 percent two segments of the mortgage market—the Within the small multifamily market, there in 2000, surpassing its official performance small (5–50 unit) multifamily mortgage was some evidence that Freddie Mac targeted level of 53.2 percent in 2001 and 51.4 percent market, and the market for mortgages on 2– affordable properties to a greater extent in in 2002. Measured on this consistent basis, 4 unit properties where 1 unit is owner- 2001–02 than in 2000. That is, 87 percent of then, Fannie Mae’s performance fell by 1.0 occupied and 1–3 units are occupied by Freddie Mac’s small multifamily units percentage point in 2001, and Freddie Mac’s renters. Bonus points did not apply to qualified for the low- and moderate-income by 1.9 percentage points in 2001 and an purchases of mortgages for owner-occupied goal in 2000; this rose to 96 percent in 2001 additional 1.8 percentage points in 2002. 1-unit properties, for investor-owned 1–4 and 94 percent in 2002. These reductions were primarily due to unit properties, and for large (more than 50 In summary, then, there is evidence that 2001–02 being years of heavy refinance units) multifamily properties, although as bonus points for small multifamily properties activity. also discussed above, a ‘‘temporary had an impact on Fannie Mae’s role in this Details of Effects of Changes in Counting adjustment factor’’ applied to Freddie Mac’s market in 2001–02 and an even larger impact Rules on Goal Performance in 2001–02. As purchases of qualifying mortgages on large on Freddie Mac’s role in this market. In discussed above, counting rule changes that multifamily properties. addition, Fannie Mae has announced a took effect in 2001 had significant positive Bonus points for small multifamily program to increase its role in this market impacts on the performance of both GSEs on properties. Each unit financed in a small further in future years.224 the low- and moderate-income goal in that multifamily property that qualified for any of year—3.8 percentage points for Fannie Mae, 223 Federal Register, October 31, 2000, Footnote and 6.0 percentage points for Freddie Mac. 222 Exclusion of loans with missing information 145, p. 65141. This section breaks down the effects of these had a greater impact on Fannie Mae’s goal 224 ‘‘Fannie Courting Multifamily Sellers; Small changes on goal performance for both GSEs; performance than on Freddie Mac’s goal Banks Balking,’’ American Banker, January 13, results are shown in Table A.9. performance. 2003, p. 1.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00084 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24311

Bonus points for single-family rental • The mortgagors’ income at the time of 2005 Procedure. Relative to the above properties. Above a threshold, each unit mortgage origination. procedure, scoring of loans purchased by the financed in a 2–4 unit property with at least • The median income of an area specified GSEs in and after 2005 will be affected by one owner-occupied unit (referred to as as follows: (i) For properties located in two factors. First, the Economic and Market ‘‘OO24s’’ below) that qualified for any of the Metropolitan Statistical Areas (MSAs), the Analysis Division has begun to incorporate housing goals was counted as two units in area is the MSA; and (ii) for properties data from the 2000 census into its procedure the numerator (and one unit in the located outside of MSAs, the area is the for estimating annual area median incomes denominator) in calculating goal performance county or the non-metropolitan portion of the and American Community Survey data are for that goal in 2001–03. The threshold was State in which the property is located, becoming available at increasingly finer equal to 60 percent of the average number of whichever has the larger median income, as levels of geographical detail for use in annual such qualifying units over the previous five of the year of mortgage origination (which updating. Beginning in 2005 Bureau of Labor years. For example, Fannie Mae financed an may be for the current year or a prior year). Statistics data on rates of inflation in average average of 50,030 low- and moderate-income For rental units in single-family properties wages will not be used. For 2005, the units in these types of properties between with rent data are available (assuming no procedure for estimating area median 1996 and 2000, and 101,423 such units in income data available for actual or incomes will be to adjust 2000 census data 2001. Thus Fannie Mae received 71,405 prospective tenants): on 1999 area median incomes to 2003 using bonus points in this area in 2001—that is, • The unit rent (or average rent for units data from the Census Bureau’s American 101,423 minus 60 percent of 50,030. So of the same type) at the time of mortgage Community Survey (ACS) on rates of change 172,828 units were entered in the numerator origination. in average incomes for States between 1999 for these properties in calculating low- and • The area median income as specified for and 2003, with a further adjustment to 2005 moderate-income goal performance. single-family owner-occupied units. based on an appropriate annual inflation Single-family rental bonus points thus For rental units in multifamily properties factor.228 Increasingly more detailed ACS encouraged the GSEs to play a larger role in where rent data are available. data will be available and will be used in this market, and also to purchase mortgages • The unit rent (or the average rent for subsequent years, as ACS estimates for on such properties in which large shares of units of the same type) at the time of metropolitan and micropolitan areas and the units qualified for the housing goals. As mortgage acquisition by the GSE. counties become available. for small multifamily bonus points, again • The area median income as specified for The second factor is the Office of some evidence may be gleaned from the data single-family owner-occupied units, but as of Management and Budget’s June, 2003, re- provided to HUD by the GSEs for 2001–02. the year the GSE acquired the mortgage. specification of MSA boundaries based on Fannie Mae financed 175,103 units in For rental units in multifamily properties analysis of 2000 census data.229 OO24s in 2001 that were eligible for the low- where rent data are not available, the GSE Analysis. For purposes of specifying the and moderate-income goal and 229,632 such may apply HUD-estimated rents which are level of the Low- and Moderate-Income units in 2002, a two-year increase of nearly based on the following area data; Housing Goal, HUD developed a • 200 percent from the 77,930 units financed The median rent in the census tract methodology for scoring loans purchased by in 2000. However, Fannie Mae’s total single- where the property is located, as of the most the GSEs in past years through 2002 as family business increased at approximately recent decennial census. though the re-benchmarking of area median • the same rate as its OO24 business in 2001 The area median income as specified for income estimates to the 2000 census and the and 2002, thus the share of its business single-family owner-occupied units, but as of 2003 re-designation of MSAs had been in accounted for by OO24s was the same in the most recent decennial census. effect and HUD had been using an ACS-based 2001–02 as in 2000—4 percent. Thus, scoring loans under the Low- and estimation procedure at the time the Within the OO24 market, there was no Moderate-Income Goal requires a data series estimates for these years were prepared. For evidence that Fannie Mae targeted affordable showing annual median incomes for MSAs, this purpose, HUD created a series of annual properties to a greater extent in 2001–02 than non-metropolitan counties, and the non- estimates of median incomes for MSAs, non- in 2000. That is, approximately 55–60 metropolitan portions of states; and metropolitan counties, and the non- percent of Fannie Mae’s OO24 units qualified decennial census data on median incomes for metropolitan portions of states. For 2000, the for the low- and moderate-income goal in census tracts.225 estimates were 1999 census medians trended each of these three years. For scoring loans purchased by the GSEs by three-fourths of the 4.0 percent annual Freddie Mac financed 96,050 units in year-by-year from 1993 through 2002, area OO24s in 2001 that were eligible for the low- median income estimates produced by HUD’s FY [year] Median Family Incomes’’ for years 1993 and moderate-income goal and 146,222 such Economic and Market Analysis Division were through 2002, issued by the Economic and Market units in 2002, also a two-year increase of used. An example will illustrate the Analysis Division, Office of Economic Affairs, nearly 200 percent from the 49,993 units estimation procedure. To generate the area PD&R, U.S. Department of Housing and Urban Development. financed in 2000. However, Freddie Mac’s median income estimates that were used to 227 The procedure applicable to the decennial total single-family business increased at score GSE loans in 2002, data from the 1990 census on 1989 area median incomes were census data used to generate estimated rents is approximately the same rate as its OO24 explained in connection with data used to define business in 2001–02, thus the share of its adjusted to 2002 using Bureau of Labor Underserved Areas in Appendix B. business accounted for by OO24s was the Statistics survey data on rates of change in 228 Transition from the 2002 methodology to the same in 2002 as in 2000—4 percent. average incomes for MSAs and counties 2005 methodology is occurring in stages in 2003 As for Fannie Mae, within the OO24 between 1989 and 1999, data from the and 2004. To generate the area median income market there was no evidence that Freddie Census Bureau’s Current Population Survey estimates used to score GSE loans in 2003, data Mac targeted affordable properties to a on rates of change in median family incomes from the 2000 census on 1999 area median incomes greater extent in 2001–02 than in 2000. That for the nine Census Divisions between 1989 were adjusted to 2001 using Bureau of Labor and 2000, and an assumed 4.0 percent per Statistics survey data on rates of change in average is, 68–69 percent of Fannie Mae’s OO24 units incomes for MSAs and counties between 1999 and qualified for the low- and moderate-income year inflation factor between 2000 and 226, 227 2000, data on rates of change in median incomes goal in each year from 2000 through 2002. 2002. for the United States and individual States between 1999 and 2001 from Census Bureau’s Current 5. Effects of 2000 Census on Scoring of Loans 225 In New England, MSAs were defined through Population Survey and American Communities Toward the Low- and Moderate-Income mid-2003 in terms of Towns rather than Counties, Survey, and an assumed 3.5 percent per year Housing Goal and the portion of a New England county outside inflation factor between 2001 and 2003. (See ‘‘HUD Background. Scoring of housing units of any MSA was regarded as equivalent to a county Methodology for Estimating FY 2003 Median under the Low- and Moderate-Income in establishing the metropolitan or non- Family Incomes,’’ issued by the Economic and metropolitan location of a property. The MSA Market Analysis Division, op cit.) A similar Housing Goal is based on data for mortgagors’ definitions established by the Office of Management procedure has been used to generate area median incomes for owner-occupied units, rents for and Budget (OMB) in June, 2003 defined MSAs in income estimates for scoring GSE loans in 2004. rental units, and area median incomes, as New England in terms of counties. 229 HUD has deferred application of the 2003 follows: 226 The procedure is explained in detail in annual MSA specification to 2005, pending completion of For single-family owner-occupied units: releases entitled ‘‘HUD Methodology for Estimating the present rulemaking process.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00085 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24312 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

trending factor (to adjust the figures from between 1989 and 1999 census data. The GSE analysis are provided in Table A.10. The mid-1999 to April 1, 2000). For 2001, the 2003 OMB MSA designations were applied. results of the GSE-HMDA comparative estimates were based on one-and-three- The resulting estimates of area median analysis are presented in the next section. fourths years of trending, since no data incomes for MSAs, non-metropolitan Table A.10 shows three sets of estimates would have been available to use for counties, and the non-metropolitan parts of for each GSE, based respectively on the updating. The 2002 estimates would have States, were used to re-score loans purchased counting rules in place in 2001–2002 (but by the GSEs between 1999 and 2002, and disregarding the bonus points and Temporary used one year of data and 1.75 years of were used further in estimating the share of Adjustment Factor), on the addition of 2000 trending. The 2003 estimates would have loans originated in metropolitan areas that census re-benchmarking, and finally on the used two years of data plus 1.75 years of would be eligible to score toward the Low- addition of both 2000 census re- trending. Area median incomes from 1989 to and Moderate-Income Housing Goal, from benchmarking and 2003 MSA specification. 1999 were estimated based on trend-lines HMDA data. The results of the retrospective BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00086 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24313

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00087 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.015 24314 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

6. GSEs Compared With the Primary Freddie Mac’s. But like Freddie Mac, Fannie through 12; they are grouped under the Conventional Conforming Mortgage Market Mae’s average performance during past following five topic-headings: This section and the next five sections periods (e.g., 1993–2002, 1996–2002, 1999– (b.1) Longer-term Performance of the GSEs; (Sections E.7 to E.12) provide a detailed 2002) has been below market levels. (b.2) Performance of the GSEs During analysis of the extent to which the GSEs’ loan However, it is encouraging that Fannie Mae Recent Years; purchases mirror or depart from the patterns markedly improved its affordable lending (b.3) The GSEs’ Funding First-time found in the primary mortgage market. As in performance relative to the market during Homebuyer Loans; Section C.5, the GSEs’ affordable lending 2001 and 2002, the first two years of HUD’s (b.4) Performance of the GSEs Based on performance is also compared with the higher housing goal levels. Fannie Mae’s Total (Home Purchase and Refinance) Loans; performance of depository lenders such as average performance during 2001 and 2002 (b.5) GSE Market Shares; and, (b.6) Additional Findings. commercial banks and thrift institutions. approached the market on the special Dimensions of lending considered include affordable and underserved areas categories (b.1) Longer-Term Performance of the GSEs the three ‘‘goals-qualifying’’ categories— and matched the market on the low-mod category. Under one measure of GSE and The longer-run performance of the GSEs is special affordable borrowers, less-than- examined between 1993 and 2002 (which median income borrowers, and underserved market activity, Fannie Mae matched the market during 2002 on the special affordable covers the period since the housing goals areas. The special affordable category were put into effect) and between 1996 and consists mainly of very-low-income category and slightly outperformed the market on the low-mod and underserved 2002 (which covers the period under the borrowers, or borrowers who have an annual current definitions of the housing goals). Of income less than 60 percent of area median areas categories. In this case, which is referred to in the text as the ‘‘purchase year’’ the two borrower-income goals, the analysis income. Because this category is more below will typically focus on the special targeted than the broadly-defined less-than- approach, Fannie Mae’s performance is based on comparing its purchases of all loans (both affordable category, which is a more targeted median-income (or low-mod) category, the category than the rather broadly defined low- discussion below will often focus on the seasoned loans and newly-originated mortgages) during a particular year with and moderate-income category. special affordable category as well as the (1) Since the early nineties, the mortgage underserved areas category which adds a loans originated in the market in that year. When Fannie Mae’s performance is measured industry has introduced new affordable neighborhood dimension (low-income and lending programs and has allowed greater high-minority census tracts) to the analysis. on an ‘‘origination year’’ basis (that is, allocating Fannie Mae’s purchases in a flexibility in underwriting lower-income This section will also compare the loans. There is evidence that these programs performance of Fannie Mae and Freddie Mac particular year to the year that the purchased- loan was originated), Fannie Mae matched are paying off in terms of more mortgages for in funding first-time homebuyers with that of low-income and minority borrowers. As primary lenders in the conventional the market in the low- and moderate-income category during 2002, and lagged the market noted earlier, Fannie Mae and Freddie Mac conforming market. have played an active role in this upsurge of The remainder of this introductory section slightly on the other two categories. 3. Both Fannie Mae and Freddie lag the affordable lending, as indicated by the high E.6 provides a list of the major and specific growth rates of their goals-qualifying findings which are presented in detail in the conventional conforming market in funding first-time homebuyers, and by a rather wide business. following Sections E.7 through 12. Sections • Between 1993 and 2002, the GSEs’ 7 and 8 define the primary mortgage market margin. Between 1999 and 2001, first-time homebuyers accounted for 27 percent of each purchases of home loans in metropolitan and discuss some technical issues related to areas increased by 57 percent.231 Their the use of the GSE and HMDA data. Sections GSE’s purchases of home loans, compared with 38 percent for home loans originated in purchases of home loans for the three 8 and 9 compare the GSEs’ performance with housing goals increased at much higher market performance for home purchase and the conventional conforming market. 4. The GSEs have accounted for a rates—264 percent for special affordable first-time homebuyer loans, while Section 10 loans, 142 percent for low- and moderate- does the same for total single family loans significant share of the total (government as well as conventional) market for home income loans, and 112 percent for loans in (that is, refinance loans and home purchase underserved census tracts. loans). Section 11 examines GSE purchases purchase loans, but their market share for each of the affordable lending categories (e.g., (2) Both Fannie Mae and Freddie Mac have in individual metropolitan areas. Following improved their purchases of affordable loans these analyses, Section 12 examines the low-income borrowers and census tracts, high-minority census tracts) has been less since the housing goals were put in place, as overall market share of the GSEs in important indicated by the increasing share of their submarkets such as first-time homebuyers. than their share of the overall market. 5. The GSEs also account for a very small business going to the three goals-qualifying a. Main Findings on GSEs’ Performance in categories. (See Table A.15 in Section E.9.) share of the market for important groups such • the Single-Family Market as minority first-time homebuyers. Between 1992 and 2002, the special affordable share of Fannie Mae’s business There are six main findings from this Considering the total mortgage market (both more than doubled, rising from 6.3 percent analysis concerning the GSEs’ purchases of government and conventional loans), it is to 16.3 percent, while the underserved areas single-family-owner mortgages: estimated that the GSEs purchased only 14 share increased more modestly, from 18.3 1. While Freddie Mac has improved its percent of loans originated between 1999 and percent to 26.7 percent. The figures for affordable lending performance in recent 2001 for African-American and Hispanic Freddie Mac are similar. The special years, it has consistently lagged the first-time homebuyers, or one-third of their affordable share of Freddie Mac’s business conventional conforming market in funding share (42 percent) of all home purchase loans rose from 6.5 percent to 15.8 percent, while affordable home purchase loans for special originated during that period. Considering the underserved areas share also increased affordable and low-moderate-income the conventional conforming market and the but more modestly, from 18.6 percent to 25.8 borrowers and underserved neighborhoods same time period, it is estimated that the percent. targeted by the housing goals.230 However, GSEs purchased only 31 percent of loans originated for African-American and (3) While both GSEs improved their Freddie Mac’s recent performance (2001 and performance, they have lagged the primary 2002) has been much closer to the market Hispanic first-time homebuyers, or about than its earlier performance. one-half of their share (57 percent) of all 231 2. In general, Fannie Mae’s affordable home purchase loans in that market. Throughout this analysis, the terms ‘‘home loan’’ and ‘‘home mortgage’’ will refer to a ‘‘home lending performance has been better than 6. The GSEs’ small share of the first-time homebuyer market could be due to the purchase loan,’’ as opposed to a ‘‘refinance loan.’’ As noted earlier, the mortgage data reported in this 230 The ‘‘affordable lending performance’’ of preponderance of high (over 20 percent) paper are for metropolitan areas, unless stated Fannie Mae and Freddie Mac refers to the downpayment loans in their mortgage otherwise. Restricting the GSE data to metropolitan performance of the GSEs in funding loans for low- purchases. areas is necessary to make it comparable with the income and underserved borrowers through their b. Specific Findings on GSE Performance in HMDA-reported conventional primary market data, purchase (or guarantee) of loans originated by the Single-Family Market which is more reliable for metropolitan areas. The primary lenders. It does not, of course, imply that analysis of first-time homebuyers in Sections E.9 the GSEs themselves are lenders originating loans This section presents 17 specific findings and E.12 cover both metropolitan and non- in the primary market. from the analyses reported in Sections E.7 metropolitan areas.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00088 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24315

market in providing affordable loans to low- (5) After experiencing declines from 1997 Mae-to-Freddie-Mac’’ ratios for special income borrowers and underserved to 1999, Fannie Mae’s affordable lending affordable and low-mod loans to above one neighborhoods. Freddie Mac’s average performance improved between 2000 and (1.03 for both), indicating better performance performance, in particular, fell far short of 2002. for Fannie Mae. The ‘‘Fannie-Mae-to-Freddie- market performance during the 1990s. Fannie • After declining from 23.0 percent in Mac’’ ratio (1.03) for the underserved area Mae’s average performance was better than 1997 to 20.4 percent in 1999, the share of category remained above one in 2002. Freddie Mac’s during the 1993–2002 period Fannie Mae’s purchases financing properties (8) While Freddie Mac has consistently as well as during the 1996–2002 period, in underserved areas jumped by three improved its performance relative to the which covers the period under HUD’s percentage points to 23.4 percent in 2000, market, it continued to lag the market in currently-defined housing goals. and then increased further to 26.7 percent by funding affordable home loans in 2001 and • Between 1993 and 2002, 11.8 percent of 2002. 2002. Freddie Mac’s mortgage purchases were for • After declining from 13.2 percent in • Unlike Fannie Mae, Freddie Mac had not special affordable borrowers, compared with 1998 to 12.5 percent in 1999, the share of made any progress through 1997 in closing 12.7 percent of Fannie Mae’s purchases, 15.4 Fannie Mae’s purchases going to special its gap with the market. The ‘‘Freddie Mac- percent of loans originated by depositories, affordable loans rebounded to 13.3 percent in to-market’’ ratio for the special affordable and 15.4 percent of loans originated in the 2000, 14.9 percent in 2001, and 16.3 percent category actually declined from 0.63 in 1992 conventional conforming market (without in 2002. to 0.59 in 1997. But Freddie Mac’s sharp estimated B&C loans).232 (6) Freddie Mac’s performance on the two improvement in special affordable purchases • Considering the underserved areas borrower-income categories improved resulted in the ‘‘Freddie-Mac-to-market’’ ratio category for the 1996–2002 period, 21.7 between 2000 and 2002, but not as much as rising to 0.88 by 2000. After declining from percent of Freddie Mac’s purchases financed Fannie Mae’s performance. Freddie Mac’s 0.84 in 1992 to 0.80 in 1997, the ‘‘Freddie- properties in underserved neighborhoods, performance on the underserved areas Mac-to-market’’ ratio for underserved areas compared with 23.5 percent of Fannie Mae’s category increased substantially between had risen only modestly to 0.84 by the year purchases, 24.9 percent of loans originated 2001 and 2002. 2000. Thus, Freddie Mac’s improvements • by depositories, and 25.4 percent of loans The share of Freddie Mac’s single- prior to 2001 allowed it to close its gap with originated in the conventional conforming family-owner business going to special the market, mainly for the special affordable market. affordable home loans increased from 9.2 in category where its gap had been the widest. 1997 to 14.7 percent in 2000 before falling to • (b.2) Performance of the GSEs During Recent During 2001 and 2002, Freddie Mac 14.4 percent in 2001 and rising to 15.8 continued to close its gap with the market. Years percent in 2001. • By 2002, all three ‘‘Freddie-Mac-to-market’’ The recent performance of the GSEs is Freddie Mac’s purchases of underserved ratios were higher than in 2000, although area loans increased at a modest rate from examined for the four-year period between they continued to fall below one: special 1999 and 2002 and then for 2001 and 2002, 19.8 percent in 1997 to 22.3 percent in 2001, affordable (0.97), low-mod (0.97), and which were the first two years that the GSEs before sharply jumping to 25.8 percent in underserved areas (0.98). Thus, during 2002, operated under the higher goal targets 2002. Freddie Mac lagged the market on all three established by HUD in the 2000 Rule. As (7) The long-standing pattern of Fannie goals-qualifying categories. explained below, the most interesting recent Mae outperforming Freddie Mac was (9) Through 1998, Fannie Mae had trend concerned Fannie Mae, which reversed during 1999 and 2000. But that significantly improved its performance improved its performance during 2001 and pattern returned in 2001 and 2002 when relative to the market. But as a result of shifts 2002, at a time when the conventional Fannie Mae outperformed Freddie Mac on all in its purchases of affordable loans, Fannie conforming market was showing little change three goals-qualifying categories. Mae lagged the market even further in 2000 in affordable lending. • Fannie Mae and Freddie Mac had (4) During the recent 1999-to-2002 period, practically the same performance in 1992 on than it had in some earlier years. During 2001 both Fannie Mae and Freddie Mac fell the three housing goal categories—special and 2002, Fannie Mae again improved its performance relative to the market. significantly below the market in funding affordable loans accounted for 6.3 percent of • affordable loans. Fannie Mae’s purchases and 6.5 percent of The above analysis and the data reported • Between 1999 and 2002, special Freddie Mac’s purchases, for a ‘‘Fannie-Mae- under this specific finding (9) are based on affordable loans accounted for 14.4 percent of to-Freddie-Mac’’ ratio of 0.97. The 1992 ratio the ‘‘purchase year’’ approach for measuring Fannie Mae’s purchases, 14.5 percent of for underserved areas was also 0.98 and that GSE activity. The purchase year approach Freddie Mac’s purchases, and 16.4 percent of for low-mod, 1.02. Reflecting Fannie Mae’s assigns GSE purchases of both prior-year loans originated in the market; thus, the much better performance, the special (seasoned) and newly-originated mortgages to ‘‘Fannie-Mae-to-market’’ ratio was 0.88 and affordable ‘‘Fannie-Mae-to-Freddie-Mac’’ the calendar year in which they were the ‘‘Freddie-Mac-to-market’’ ratio was also ratio had risen to 1.27 by 1997, the purchased by the GSE; this results in an 0.88. underserved area ratio to 1.17, and the low- inconsistency with the HMDA-reported • During the same period, underserved mod ratio to 1.10. market data, which covers only newly- area loans accounted for 24.0 percent of • However, in 1999, the ‘‘Fannie-Mae-to- originated mortgages. Sections E.9 and E.10 Fannie Mae’s purchases, 22.9 percent of Freddie-Mac’’ ratio for each of the three also report the results of an alternative Freddie Mac’s purchases, and 25.8 percent of goals-qualifying categories fell to slightly ‘‘origination year’’ approach that assigns GSE loans originated in the market; the ‘‘Fannie- below one. 1999 was the first year since 1992 purchases to their year of origination, placing Mae-to-market’’ ratio was 0.93 and the that Freddie Mac had outperformed Fannie them on a more consistent basis with the ‘‘Freddie-Mac-to-market’’ ratio was only Mae in purchasing affordable home loans HMDA-reported market data. The findings 0.89.233 (although only by a very slight margin). from the origination-year approach are • In 2000, Freddie Mac’s sharper increases discussed under specific finding (10). • Fannie Mae’s decline in performance 232 Unless otherwise noted, the conventional in special affordable and low-mod purchases conforming market data reported in this section further reduced the ‘‘Fannie-Mae-to-Freddie- during 1999 resulted in the ‘‘Fannie-Mae-to- exclude an estimate of B&C loans; the less-risky A- Mac’’ ratios for these two categories to 0.90 market’’ ratio falling sharply to 0.74 for minus portion of the subprime market is included and 0.96, respectively. Fannie Mae’s sharper special affordable and to 0.81 for in the market definition. See Section E.7 and increase in underserved areas funding underserved areas. In 2000, Fannie Mae Appendix D for a discussion of primary market resulted in the ‘‘Fannie-Mae-to-Freddie-Mac’’ improved and reversed its declining trend, as definitions and the uncertainty surrounding ratio rising from slightly below one (0.98) in the ‘‘Fannie-Mae-to-market’’ ratios increased estimates of the number of B&C loans in HMDA 1999 to 1.06 in 2000. to 0.79 for special affordable purchases and data. As noted there, B&C loans are much more • Fannie Mae’s stronger performance to 0.89 for underserved area purchases. likely to be refinance loans rather than home • purchase loans. during 2001 and 2002 returned the ‘‘Fannie- During 2001, Fannie Mae increased its 233 Fannie Mae had a particularly poor year special affordable percentage by 1.6 during 1999. Therefore, the text also reports performance is closer to the market, it continues to percentage points to 14.9 percent, which was averages for 2000–2002, dropping the year 1999 (see fall below market levels during the 2000–2002 only 0.7 percentage point below the market’s Table A.13 in Section E.9). While Fannie Mae’s period. performance of 15.6 percent. Fannie Mae

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00089 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24316 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

increased its low-mod percentage from 40.8 have improved their performance they have combined market share, downpayment, and percent to 42.9 percent at the same time that consistently lagged the market in funding default data, concluded that the GSEs play a the low-mod share of the primary market was loans (home purchase and refinance) that very minimal role in providing credit support falling from 44.4 percent to 42.9 percent, qualify for the housing goals. (See Table A.20 and assuming credit risk for low-income and placing Fannie Mae at the market’s of Section E.10, which is based on the minority borrowers; for example, the study performance. Similarly, Fannie Mae purchase-year approach for measuring GSE concluded that in 1995 the GSEs provided increased its underserved area percentage activity.) only four percent of the credit support going from 23.4 percent in 2000 to 24.4 percent in • 1999–2002. During the recent 1999-to- to African-Americans and Hispanic 2001 while the underserved area share of the 2002 period, both Fannie Mae and Freddie borrowers. primary market was falling from 26.4 percent Mac fell significantly below the market in • Section V of this study begins to to 25.2 percent, placing Fannie Mae at 0.8 funding affordable loans. Between 1999 and reconcile these different results by examining percentage point from the market’s 2002, special affordable loans accounted for the role of the GSEs in the first-time performance. 13.8 percent of Fannie Mae’s purchases, 13.8 homebuyer market and the downpayment • During 2002, Fannie Mae continued to percent of Freddie Mac’s purchases, and 15.7 characteristics of mortgages purchased by the improve its performance on all three goals percent of loans originated in the market; GSEs. categories. Using the purchase-year approach thus, the ‘‘Fannie-Mae-to-market’’ ratio and (14) The market role of the GSEs appears to measure GSE performance, Fannie Mae the ‘‘Freddie-Mac-to-market’’ ratio were each to be particularly low in important market matched the market on the special affordable 0.88 during this period. segments such as minority first-time category (16.3 percent for both), led the • During the same period, underserved homebuyers. market on the low-mod category (45.3 area loans accounted for 23.8 percent of • Recent analysis has estimated that the percent for Fannie Mae compared with 45.2 Fannie Mae’s purchases, 23.1 percent of GSEs’ share of the market for first-time percent for the market), and led the market Freddie Mac’s purchases, and 25.7 percent of African-American and Hispanic homebuyers on the underserved area category (26.7 loans originated in the market; thus, the was only 14.3 percent between 1999 and percent for Fannie Mae versus 26.4 percent ‘‘Fannie-Mae-to-market’’ ratio was 0.93 and 2001, or about one-third of their share (41.5 for the market). As explained in the next the ‘‘Freddie-Mac-to-market’’ ratio was percent) of all home purchases during that specific finding, measuring Fannie Mae’s 0.90.234 period. This analysis includes the total performance on the more consistent • 2002. During this year of heavy market, including government and origination-year basis gives somewhat refinancing, Fannie Mae’s performance conventional loans. different results. approached but fell below market • A similar market share analysis was (10) This analysis addresses several performance. The ‘‘Fannie-Mae-to-market’’ conducted for the conventional conforming technical issues involved in measuring GSE ratios were 0.98 for special affordable loans, market. Between 1999 and 2001, the GSEs’ performance. The above analysis was based 0.99 for low-mod loans, and 0.99 for purchases accounted for 56.6 percent of all on the ‘‘purchase year’’ approach, as defined underserved area loans. The ‘‘Freddie-Mac- home loans originated in the conventional in (9) above. An alternative ‘‘origination to-market’’ ratios were 0.04–0.05 lower: 0.93 conforming market of both metropolitan year’’ approach has also been utilized, which for special affordable loans, 0.94 for low-mod areas and non-metropolitan areas. Their assigns GSE purchases to their year of loans, and 0.94 for underserved area loans. purchases of first-time homebuyer loans, on origination, placing them on a more the other hand, accounted for only 39.8 (b.5) GSE Market Shares consistent basis with the HMDA-reported percent of all first-time homebuyer loans market data. While the average results (e.g., This analysis includes an expanded originated in that market. 1999–2002 GSE performance) are similar ‘‘market share’’ analysis that documents the • The GSEs have funded an even lower under the two reporting approaches, GSE GSEs’ contribution to important segments of share of the minority first-time homebuyer performance in any particular year can be the home purchase and first-time homebuyer market in the conventional conforming affected, depending on the extent to which markets. market. Between 1999 and 2001, the GSEs the GSE has purchased goals-qualifying (13) The GSEs account for a significant purchases of African-American and Hispanic seasoned loans in that particular year. share of the total (government as well as first-time homebuyer loans represented 30.9 • The choice of which approach to follow conventional conforming) market for home percent of the conventional conforming particularly affected conclusions about purchase loans. However, the GSEs’ market market for these loans. Thus, while the GSEs Fannie Mae’s performance relative to the share for each of the affordable lending have accounted for 56.6 percent of all home market. Under the origination-year approach, categories is much less than their share of the loans in the conventional conforming market, Fannie Mae lagged the market on all three overall market. they have accounted for only 30.9 percent of housing goal categories during 2001 and on • The GSEs’ purchases were estimated to loans originated in that market for African- the special affordable and underserved area be 46 percent of all home loans originated in American and Hispanic first-time categories during 2002. In 2002, Fannie Mae metropolitan areas between 1999 and 2002 homebuyers. essentially matched the market on the low- but only 29 percent of loans originated for (15) A noticeable pattern among the lower- mod category (45.4 percent for Fannie Mae African-American and Hispanic borrowers, income-borrower loans purchased by the compared with 45.2 percent of the market). 37 percent of loans originated for low-income GSEs is the predominance of loans with high (b.3) The GSEs’ Funding of First-Time borrowers, and 36 percent for properties in downpayments. This pattern of purchasing Homebuyer Loans underserved areas. The GSEs’ market share mainly high downpayment loans is one for the various affordable lending categories (11) The GSEs’ funding of first-time factor explaining why the Fed study found increased during 2001 and 2002, but the such a small market role for the GSEs. It may homebuyers has been compared to that of above-mentioned pattern remained. primary lenders in the conventional be the explanation for the small role of • A study by staff from the Federal Reserve Fannie Mae and Freddie Mac in the first-time conforming market. Both Fannie Mae and Board suggests that the GSEs have a much Freddie lag the market in funding first-Time homebuyer market. Further study of this more limited role in the affordable lending issue is needed. homebuyers, and by a rather wide margin. market than is suggested by the data • • First-time homebuyers account for 27 During 2001 and 2002, approximately 50 presented above.235 The Fed study, which percent of each GSE’s purchases of home percent of Fannie Mae’s special affordable, low-mod, and underserved areas loans had loans, compared with 38 percent for home 234 loans originated in the conventional As explained in Section E.9, deducting B&C downpayments of at least 20 percent, a loans from the market totals has more impact on the conforming market. percentage only slightly smaller than the market percentages for total (both home purchase corresponding percentage (53 percent) for all (b.4) Performance of the GSEs Based on and refinance) loans than for only home purchase Fannie Mae’s home loan purchases. Similar Total (Home Purchase and Refinance) Loans loans. The effects of excluding B&C loans from the total market can be seen by comparing the third and patterns of high downpayments on the goals- (12) The GSEs’ acquisitions of total loans sixth columns of data in Table A.19 in Section E.10. qualifying loans were evident in Freddie (including refinance loans as well as home 235 See Glenn B. Canner, Wayne Passmore, and purchase loans) were also examined. The Brian J. Surette, ‘‘Distribution of Credit Risk Among Minority Homebuyers’’ in Federal Reserve Bulletin, main results indicate that while the GSEs Providers of Mortgages to Lower-Income and 82(12): 1077–1102, December, 1996.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00090 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24317

Mac’s 2001 and 2002 purchases, as well as 2002—these are called ‘‘conventional for B&C loans, this analysis follows HUD’s in prior years for both GSEs. conforming loans.’’ The GSEs’’ purchases of 2000 Rule which assumed that the B&C (b.6) Additional Findings FHA-insured, VA-guaranteed, and Rural portion of the subprime market accounted for Housing Service loans are excluded from this one-half of the loans originated by the This analysis examines two additional analysis. The conventional conforming subprime lenders included in Scheessele’s topics related to minority first-time market is used as the benchmark against list.240 As shown below, the effects of homebuyers and the use of HMDA data for which to evaluate the GSEs because that is adjusting the various market percentages for measuring the characteristics of loans the market definition Congress requires that B&C loans are minor mostly because the originated in the conventional conforming HUD consider when setting the affordable analysis in this section focuses on home market. housing goals. However, as discussed in purchase loans, which historically have (16) The share of the GSEs’ purchases for Section II, some have questioned whether accounted for less than one quarter of the minority first-time homebuyers was much lenders in the conventional market are doing mortgages originated by subprime lenders— less than the share of newly-originated an adequate job meeting the credit needs of the subprime market is mainly a refinance mortgages in the conventional conforming minority borrowers, which suggests that this market.241 market for those homebuyers. 237 market provides a low benchmark. Lender-Purchased Loans in HMDA. When • Between 1999 and 2001, minority first- Manufactured Housing Loans. In their analyzing HMDA data, Fannie Mae includes time homebuyers accounted for 6.6 percent comments on the proposed 2000 Rule, both in its market totals those HMDA loans of Fannie Mae’s purchases of home loans, 5.8 GSEs raised questions about whether loans identified as having been purchased by the percent of Freddie Mac’s purchases, and 10.6 on manufactured housing should be reporting lender, above and beyond loans percent of home loans originated in the excluded when comparing the primary that were originated by the reporting conventional conforming market. For this market with the GSEs. The GSEs purchase lender.242 Fannie Mae contends that there are subgroup, Fannie Mae’s performance is 62 these loans, but they have not played a a subset of loans originated by brokers and percent of market performance, while significant role in the manufactured housing subsequently purchased by wholesale Freddie Mac’s performance is 55 percent of loan market. As emphasized by HUD in its lenders that are neither reported by the market performance. 2000 GSE Rule, manufactured housing is an (17) Some studies have concluded that important source of home financing for low- brokers nor the wholesale lenders as HMDA data overstate the share of market income families and for that reason, should originations but are reported by the loans going to low-income borrowers and be included in any analysis of affordable wholesale lenders as purchased loans. underserved areas. This analysis does not lending. However, for comparison purposes, According to Fannie Mae, these HMDA- support that conclusion. data are also presented for the primary reported purchased loans should be added to • This analysis compares the low-income market defined without manufactured HMDA-reported originated loans to arrive at and underserved areas characteristics of the housing loans. Because this analysis focuses an estimate of total mortgage originations. GSEs’ purchases of newly-originated on metropolitan areas, it does not include the This rule’s market definition includes only (‘‘current-year’’) loans as reported both by the substantial number of manufactured housing HMDA-reported originations; purchased GSEs’’ own data and by HMDA data.236 For loans originated in non-metropolitan areas. loans are excluded from the market recent years, HMDA data on loans sold to the Subprime Loans. Both GSEs also raised definition. While some purchased loans may GSEs do not always have higher percentages questions about whether subprime loans not be reported as originations in HMDA (the of low-income and underserved areas loans should be excluded when comparing the Fannie Mae argument), there are several than the GSEs’ own data on their purchases primary market with their performance. In its reasons for assuming that most HMDA- of newly-originated mortgages. For example, final 2000 GSE Rule, HUD argued that reported purchased loans are also reported in from 1996–2002, both HMDA and Fannie borrowers in the A-minus portion of the HMDA as market originations. First, Fed staff Mae reported that special affordable loans subprime market could benefit from the have told HUD that including purchased accounted for about 13 percent of Fannie standardization and lower interest rates that loans would result in double counting Mae’s purchases of newly-originated loans. typically accompany an active secondary mortgage originations.243 Second, HMDA reported a 21.9 underserved areas market effort by the GSEs. A-minus loans are percentage for Fannie Mae, which was rather not nearly as risky as B&C loans and the are available at http://www.huduser.org/ similar to the underserved areas percentage GSEs have already started purchasing A- publications/hsgfin.html. (22.4 percent) reported by Fannie Mae itself. minus loans (and likely the lower ‘‘B’’ grade 240 The one-half estimate is conservative as some Given that similar patterns were observed for subprime loans as well). The GSEs observers estimate that B&C loans account for only Freddie Mac’s mortgage purchases, it appears themselves have mentioned that a large 30–40 percent of the subprime market. However, varying the B&C share from 50 percent to 30 percent that there is no upward bias in the HMDA- portion of borrowers in the subprime market could qualify as ‘‘A credit.’’ This analysis does not significantly change the following analysis based market benchmarks used in this study. of home purchase loans because subprime loans are includes the A-minus portion of the 7. Definition of Primary Market mainly for refinance purposes. Overstating the subprime market, or conversely, excludes the share of B&C loans in this manner also allows for Conventional Conforming Market. The B&C portion of that market. any differences in HMDA reporting of different market analysis section is based mainly on Unfortunately, HMDA does not identify types of loans—for example, if B&C loans account HMDA data for mortgages originated in the subprime loans, much less separate them into for 35 percent of all subprime loans, then assuming conventional conforming market of their A-minus and B&C components.238 that they account for 50 percent is equivalent to metropolitan areas during the years 1992 to Randall M. Scheessele at HUD has identified assuming that B&C loans are reported in HMDA at approximately 200 HMDA reporters that 70 percent of the rate of other loans. 2002. Only conventional loans with a 241 primarily originate subprime loans and The reductions in the market shares are more principal balance less than or equal to the significant for total loans, which include refinance conforming loan limit are included; the account for about 60–70 percent of the 239 as well as home purchase loans; for data on total conforming loan limit was $300,700 in subprime market. To adjust HMDA data loans, see Table A.19 in Section 10. Subprime lenders have been focusing more on home purchase 236 In this comparison, a higher special affordable 237 The market definition in this section is loans recently. The home purchase share of loans percentage for HMDA-reported mortgage narrower than the ‘‘Total Market’’ data presented originated by the subprime lenders in Scheessele’s originations that lenders report as also being sold earlier in Tables A.1 and A.2, which included all list increased from 26 percent in 1999 to 36 percent to the GSEs—as compared with the special home loans below the conforming loan limit, that in 2000 before dropping to about 30 percent during affordable percentage for newly-originated is, government loans as well as conventional the heavy refinancing years of 2001 and 2002. mortgages that the GSEs report as being actually conforming loans. The market share analysis 242 In 2001 (2002), lenders reported in HMDA that purchased by them—would suggest that HMDA reported in Section E.12 also examine the GSEs’ they purchased 851,735 (906,684) conventional market data are biased; that is, in this situation, the role in the overall market. conforming, home purchase loans in metropolitan special affordable percentage for all mortgage 238 And there is some evidence that many areas; this compares with 2,763,230 (2,929,197) originations reported in HMDA would likely be subprime loans are not even reported to HMDA, loans that these same lenders reported that they larger than the special affordable percentage for all although there is nothing conclusive on this issue. originated in metropolitan areas. new mortgage originations, including those not See Fair Lending/CRA Compass, June 1999, p. 3. 243 See Randall M. Scheeselle, HMDA Coverage of reported in HMDA as well as those reported in 239 The list of subprime lenders as well as the Mortgage Market, Housing Finance Working HMDA. Scheessele’s list of manufactured housing lenders Continued

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00091 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24318 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

comparisons of HMDA-reported FHA data 8. Technical Issues: Using HMDA Data To data do not overstate the share of goals- with data reported by FHA supports the Measure the Characteristics of GSE qualifying loans in the market. The Fed’s conclusion. For instance, FHA’s own Purchases and Mortgage Market discussion below of the GSEs’ purchases of data indicate that during 2001 FHA insured Originations 247 prior-year and current-year loans also 752,319 home purchase loans in This section discusses important technical highlights the strategy of purchasing metropolitan areas; the sum of HMDA- issues concerning the use of HMDA data for seasoned loans that qualify for the housing reported purchased home loans and HMDA- measuring the GSEs’ performance relative to goals. The implications of this strategy for reported originated home loans in the characteristics of mortgages originated in understanding recent shifts in the relative metropolitan areas alone yields a much the primary market. The first issue concerns performance of Fannie Mae and Freddie Mac higher figure of 845,176 FHA-insured loans the reliability of HMDA data for measuring are discussed below in Section E.9. during 2001.244 While these calculations are the borrower income and census tract a. GSEs’ Purchases of ‘‘Prior-Year’’ and for the FHA market (rather than the characteristics of loans sold to the GSEs. ‘‘Current-Year’’ Mortgages conventional market), they suggest that Fannie Mae, in particular, contends that There are two sources of loan-level including HMDA-reported purchased loans HMDA data understates the percentages of its information about the characteristics of in the market definition would overstate business that qualify for the three housing mortgages purchased by the GSEs—the GSEs mortgage origination totals. Third, Abt goals. In its comments on the proposed 2000 themselves and HMDA data. The GSEs Associates surveyed nine wholesale lenders Rule, Fannie Mae questioned HUD’s reliance provide detailed data on their mortgage and questioned them concerning their on HMDA data for measuring its purchases to HUD on an annual basis. As guidelines for reporting in HMDA loans performance. As discussed below, HMDA part of their annual HMDA reporting purchased from brokers. Most of these data on loans sold to the GSEs do not include responsibilities, lenders are required to lenders said brokered loans were reported as prior-year (seasoned) loans that are sold to indicate whether their new mortgage originations if they [the wholesale lender] the GSEs. Since about one-fourth of GSE originations or the loans that they purchase make the credit decision; this policy is purchases in any particular year involve (from affiliates and other institutions) are consistent with the Fed’s guidelines for loans originated in prior years, HMDA data sold to Fannie Mae, Freddie Mac or some HMDA reporting. Abt Associates concluded will not provide an accurate measure of the other entity. There have been numerous that ‘‘brokered loans do seem more likely to goals-qualifying characteristics of the GSEs’ studies by HUD staff and other researchers be reported as originations * * *.’’ 245 total purchases when the characteristics of that use HMDA data to compare the borrower Finally, it should be noted that including prior-year loans differ from those of newly- and neighborhood characteristics of loans purchased loans in the market definition originated, current-year loans. sold to the GSEs with the characteristics of does not significantly change the goals- A related issue concerns the appropriate all loans originated in the market. One qualifying shares of the market, mostly definition of the GSE data when making question is whether HMDA data, which is because borrower income data are missing for annual comparisons of GSE performance widely available to the public, provides an the majority of purchased loans. In addition, with the market. On the one hand, the GSE accurate measure of GSE performance, as the low-income and underserved area shares annual data can be expressed on a purchase- compared with the GSEs’ own data.249 for purchased and originated loans are rather year basis, which means that all GSE Fannie Mae has argued that HMDA data similar. In 2001, the following shares for the purchases in a particular year would be understate its past performance, where conventional conforming home purchase assigned to that particular year. performance is defined as the percentage of market were obtained for purchased and Alternatively, the GSE annual data can be Fannie Mae’s mortgage purchases accounted originated loans: Low-income (25.8 percent expressed on an origination-year basis, which for by one of the goal-qualifying categories. for purchased loans, 28.3 percent for market means that GSE purchases in a particular As explained below, over the past six years, originations), low-mod income (41.3 percent, year would be assigned to the calendar year HMDA has provided rather reliable national- 43.2 percent), and underserved areas (24.2 that the GSE-purchased mortgage was level information on the goals-qualifying percent, 25.8 percent). In 2002, the originated; for example, a GSE’s purchase percentages for the GSEs’ purchases of comparisons were as follows: low-income during 2001 of a loan originated in 1999 ‘‘current-year’’ (i.e., newly-originated) loans, (26.6 percent for purchased loans, 29.7 would be assigned to 1999, the year the loan but not for their purchases of ‘‘prior-year’’ percent for market originations), low-mod was originated. These two approaches are loans.250 income (42.3 percent, 45.3 percent), and discussed further below. In any given calendar year, the GSEs can underserved areas (28.8 percent, 27.2 A final technical issue concerns the purchase mortgages originated in that percent).246 reliability of HMDA for measuring the calendar year or mortgages originated in a percentage of goals-qualifying loans in the prior calendar year. In 2001 and 2002, for Paper No. HF–007. Office of Policy Development primary market. Both GSEs refer to findings example, purchases of prior-year mortgages and Research, U.S. Department of Housing and from a study by Jim Berkovec and Peter Zorn accounted for approximately 20 percent of Urban Development, July, 1998. concerning potential bias in HMDA data.248 244 In this example, HMDA-reported purchased Based on a comparison of the borrower and 249 For another discussion of this issue, see loans insured by FHA have been reduced from census tract characteristics between Freddie- Randall M. Scheessele, HMDA Coverage of the 411,930 to 100,251 by a procedure that accounts for Mac-purchased loans (from Freddie Mac’s Mortgage Market, Housing Finance Working Paper missing data and overlapping purchased and own data) and loans identified in 1993 HF–007, Office of Policy Development and originated loans. See Harold L. Bunce, The GSEs’ HMDA data as sold to Freddie Mac, Berkovec Research, Department of Housing and Urban Funding of Affordable Loans: A 2000 Update, and Zorn conclude that HMDA data overstate Development, July 1998. Scheessele reports that Working Paper HF–013, Office of Policy and HMDA data covered 81.6 percent of the loans Development and Research, HUD, April 2002, for the percentage of conventional conforming acquired by Fannie Mae and Freddie Mac in 1996. an alternative analysis showing that a market loans originated for lower-income borrowers The main reason for the under-reporting of GSE estimate based on adding HMDA-reported and for properties located in underserved acquisitions is a few large lenders failed to report purchased loans to HMDA-reported originations census tracts. If HMDA data overstate the the sale of a significant portion of their loan would substantially overstate the volume of FHA percentage of goals-qualifying loans, then originations to the GSEs. Also see the analysis of mortgage originations in metropolitan areas. HUD’s market benchmarks (which are based HMDA coverage by Jim Berkovec and Peter Zorn. 245 See Chapter III, ‘‘Reporting of Brokered and on HMDA data) will also be overstated. The ‘‘Measuring the Market: Easier Said than Done,’’ Correspondent Loans under HMDA’’, in Exploratory analysis below does not support the Berkovec Secondary Mortgage Markets. McLean VA: Freddie Study of the Accuracy of HMDA Data, by Abt Mac, Winter 1996, pp. 18–21; as well as the Associates Inc. under contract for the Office of and Zorn findings—it appears that HMDA Berkovec and Zorn study cited in the above Policy Development and Research, HUD, February footnote. 12, 1999, page 18. 247 Readers not interested in these technical 250 Between 1993 and 1996, the GSEs’ purchases 246 The percentage shares for purchased loans are issues may want to proceed to Section E.9, which of prior-year loans were not as targeted as they were obtained after eliminating purchased loans without compares GSE performance to the primary market. after 1996; thus, during this period, HMDA data and purchased loans that overlap with 248 See Jim Berkovec and Peter Zorn, ‘‘How provided reasonable estimates of the goals- originated loans. The calculations included 138,536 Complete is HMDA? HMDA Coverage of Freddie qualifying percentages of the GSEs’ purchases of all purchased loans for 2001 and 182,290 purchased Mac Purchases,’’ The Journal of Real Estate (both current-year and prior-year) loans, with a few loans for 2002. Research, Vol. II, No. 1, Nov. 1, 1996. exceptions (see Table A.11).

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00092 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24319

the home loans purchased by each GSE.251 HMDA data provide information mainly on purchases, as measured by HMDA data and newly-originated mortgages that are sold to by the GSEs’ own data. Table A.11 divides 251 During the 1990s, the GSEs increased their the GSEs’that is, HMDA data on loans sold each of the GSEs’ goals-qualifying purchases of seasoned loans; see Paul B. to the GSEs will not include many of their percentages for a particular acquisition year Manchester, Goal Performance and Characteristics purchases of prior-year loans. The into two components, the percentage for implications of this for measuring GSE of Mortgages Purchased by Fannie Mae and Freddie ‘‘prior-year’’ loans and the percentage for performance can be seen in Table A.11, Mac, 1998–2000, Housing Finance Working Paper ‘‘current-year’’ loans. No. HF–015, Office of Policy Development and which provides annual data on the borrower Research, HUD, May 2001. and census tract characteristics of GSE BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00093 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24320 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00094 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.016 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24321

Consider Fannie Mae’s special affordable prior-year mortgages explains why Fannie percent of current-year loans purchased by purchases in 2002. According to Fannie Mae’s own data show an overall (both prior- Fannie Mae financed properties in Mae’s own data, 16.3 percent of its purchases year and current-year) percentage of special underserved areas. However, Freddie Mac during 2002 were special affordable loans. affordable loans that is higher than that reported that 21.0 percent of the current-year reported for Fannie Mae in HMDA data. According to HMDA data, only 15.5 percent loans it purchased between 1996 and 2002 of loans sold to Fannie Mae fell into the b. Reliability of HMDA Data financed properties in underserved areas, a special affordable category. In this case, With the above explanation of the basic figure somewhat higher than the 19.5 percent HMDA data underestimate the special differences between GSE-reported and that HMDA reported as underserved area affordable share of Fannie Mae’s purchases HMDA-reported loan information, issues during 2002. What explains these different loans sold to Freddie Mac during that related to the reliability of HMDA data can period.252 patterns in the GSE and HMDA data? The now be discussed. Table A.12 presents the reason that HMDA data underestimate the same information as Table A.11, except that special affordable percentage of Fannie Mae’s the data are aggregated for the years 1993–5, 2002 purchases can be seen by disaggregating 1996–2002, and 1999–2002. Comparing Fannie Mae’s purchases during 2002 into HMDA-reported data on GSE purchases with their prior-year and current-year GSE-reported current-year data suggests that, components. Table A.11 shows that the on average, HMDA data have provided overall figure of 16.3 percent for special reasonable estimates of the goals-qualifying affordable purchases is a weighted average of percentages for the GSEs’ current-year 18.8 percent for Fannie Mae’s purchases purchases (with the exception of Freddie during 2002 of prior-year mortgages and 15.8 Mac’s underserved area loans, as discussed percent for its purchases of current-year below). For example, Fannie Mae reported 252 Freddie Mac’s underserved area figure for purchases. The HMDA-reported figure of 15.5 that 13.0 percent of the current-year loans it 2002 showed a particularly large discrepancy—as percent is based mainly on newly-mortgaged purchased between 1996 and 2002 were for shown in Table A.11, Freddie Mac reported that (current-year) loans that lenders reported as special affordable borrowers. In their HMDA 25.0 percent of the current-year loans it purchased being sold to Fannie Mae during 2002. The submissions, lenders reported a nearly during 2002 financed properties in underserved HMDA figure is similar in concept to the identical figure of 12.7 percent for the special areas, a figure much higher than the 21.4 percent current-year percentage from the GSEs’ own affordable share of loans that they sold to that HMDA reported as underserved area loans sold data. And the HMDA figure and the GSE to Freddie Mac during 2002. This is the largest Fannie Mae. The corresponding numbers for discrepancy in Table A.11, and it is not clear what current-year figure are practically the same in Freddie Mac were 12.4 percent reported by explains it. This downward bias for HMDA data, is this case (15.5 versus 15.8 percent). Thus, the them and 11.9 percent reported by HMDA. the opposite of that suggested by Berkovec and relatively large share of special affordable During the same period, both Fannie Mae Zorn, who argued that affordability percentages mortgages in Fannie Mae’s purchases of and HMDA reported that approximately 22 from HMDA data are biased upward.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00095 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24322 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00096 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.017 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24323

The facts that the GSE (both Fannie Mae that the purchase actually takes place; this 9. Affordable Lending by the GSEs: Home and Freddie Mac) and HMDA figures for approach is also consistent with the statutory Purchase Loans special affordable and low-mod loans are requirement for measuring GSE performance This section compares the GSEs’ affordable similar, and that the Fannie Mae and HMDA under the housing goals. However, this lending performance with the primary figures for underserved areas are similar, approach results in an obvious ‘‘apples to market for the years 1993–2002. The analysis suggest that the Berkovec and Zorn oranges’’ problem with respect to the HMDA- in this section begins by presenting the GSE conclusions about HMDA being upward based market data, which include only data on a purchase-year basis. As discussed biased are wrong.253 For the 1996-to-2002 newly-originated mortgages (i.e., current-year above, the GSE data that are reported to HUD period, the discrepancies reported in Table mortgages). To place the GSE and market include their purchases of mortgages A.11 as well as Table A.12 are mostly data on an ‘‘apples to apples’’ basis, HUD has originated in prior years as well as their consistent with HMDA being biased in a also used an alternative approach that purchases of mortgages originated during the downward direction, not an upward expresses the GSE annual data on an 254 current year. The market data reported by direction as Berkovec and Zorn contend. origination-year basis. In this case, all In particular, the Freddie-Mac-reported purchases by a GSE in any particular year HMDA include only mortgages originated in underserved area percentage being larger would be fully reported but they would be the current year. This means that the GSE- than the HMDA-reported underserved area allocated to the year that they were versus-market comparisons are defined percentage suggests a downward bias in originated, rather than to the year they were somewhat inconsistently for any particular HMDA. The more recent and complete purchased. Under this approach, a GSE’s data calendar year. Each year, the GSEs have (Fannie Mae data as well as Freddie Mac for the year 2000 would not only include that newly-originated loans available for data) analysis does not support the Berkovec GSE’s purchases during 2000 of newly- purchase, but they can also purchase loans and Zorn finding that HMDA overstates the originated mortgages but also any year-2000- from a large stock of seasoned (prior-year) goals-qualifying percentages of the market.255 originations purchased in later years (i.e., loans currently being held in the portfolios c. Purchase-Year Versus Origination-Year during 2001 and 2002 in this analysis). This of depository lenders. One method for Reporting of GSE Data approach places the GSE and the market data making the purchase-year data more In comparing the GSEs’ performance to the on a consistent, current-year basis. In the consistent is to aggregate the data over primary market, HUD has typically expressed above example, the market data would several years, instead of focusing on annual the GSEs’ annual performance on a purchase- present the income and underserved area data. This provides a clearer picture of the year basis. That is, all mortgages (including characteristics of mortgages originated in types of loans that have been originated and both current-year mortgages and prior-year 2000, and the GSE data would present the are available for purchase by the GSEs. This mortgages) purchased by a GSE in a same characteristics of all year-2000- approach is taken in Table A.13, which is particular year are assigned to the year of mortgages that the GSE has purchased to date discussed below. Another method for making GSE purchase. The approach of including a (i.e., through year 2002).256 the GSE and market data consistent is to GSE’s purchases of both ‘‘current-year’’ and Below, results will be presented for both express the GSE data on an origination-year ‘‘prior-year’’ mortgages gives the GSE full the purchase-year and origination-year basis; that approach is taken in Table A.16, credit for their purchase activity in the year approaches. Following past HUD studies that which is discussed after presenting the have compared GSE performance with the annual results on a purchase-year basis. 253 The data in Table A.12 that support Berkovec primary market, most of the analysis in this a. Longer-Term Performance, 1993–2002 and and Zorn are the 1993–95 special affordable and section reports the GSE data on a purchase- 1996–2002 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 how well the GSEs have been doing relative the ten-year period between 1993 and 2002. reporting targeted loans. In fact, the recent years’ Data are also presented for two important data suggest a downward bias in HMDA’s reporting to the market. of targeted loans. sub-periods: 1993–95 (for showing how much the GSEs have improved their 254 Of course, on an individual year basis, the 256 Under the origination-year approach, GSE 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 1996–2002 (for analyzing their other annual data reported in Table A.11 show a year (say year 2002) 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 home purchase mortgages, and the next percentage is smaller than the GSE current year affordable share for the market was 16.8 percent in percentage (e.g., Freddie Mac’s underserved areas 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 255 Table A.12 also includes aggregates for the Mae continues to purchase year-2000 originations lending performance of Freddie Mac and more recent period, 1999–2002. The ratios of during 2003 and the following years. Of course, Fannie Mae over the two longer-term periods, HMDA-reported-to-GSE-reported averages for this whether Fannie Mae’s future purchases result in it 1993–2002 and 1996–2002. sub-period are similar to those reported for 1996– ever leading the 2000-year market is not known at 2002. this time. BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00097 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24324 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00098 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.018 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24325

Freddie Mac lagged both Fannie Mae and moderate-income borrowers, compared with special affordable, 0.91 for low-mod, and the primary market in funding affordable 41.2 percent of Fannie Mae’s purchases, 43.1 0.85 for underserved areas. home loans in metropolitan areas between percent of loans originated by depositories, The above analysis has defined the market 1993 and 2002. During that period, 11.8 and 43.6 percent of loans originated in the to exclude B&C loans, which HUD believes percent of Freddie Mac’s mortgage purchases conventional conforming market. Over the is the appropriate market definition. were for special affordable (mainly very-low- same period, 21.7 percent of Freddie Mac’s However, to gauge the sensitivity of the income) borrowers, compared with 12.7 purchases financed properties in results to how the market is defined, Table percent of Fannie Mae’s purchases, 15.4 underserved neighborhoods, compared with A.14 shows the effects on the market percent of loans originated by 23.5 percent of Fannie Mae’s purchases, 24.9 percentages for different definitions of the 257 depositories, and 15.4 percent of loans percent of depository originations, and 25.4 conventional conforming market, such as originated in the conforming market without percent of loans originated in the primary excluding manufactured housing loans, small B&C loans.258 market. loans, and all subprime loans (i.e., the A- Although Freddie Mac consistently Fannie Mae’s affordable lending minus portion of the subprime market as well improved its performance during the 1990s, performance was better than Freddie Mac’s as the B&C portion). For example, the average a similar pattern characterized the 1996–2002 over the 1993 to 2002 period as well as special affordable (underserved area) market period. During that period, 39.8 percent of during the 1996 to 2002 period. However, percentage for 1996–2002 would fall by about Freddie Mac’s purchases were for low- and Fannie Mae lagged behind depositories and 0.2 (0.6) percentage point if the remaining the overall market in funding affordable subprime loans (i.e., the A-minus loans) were 257 As shown in Table A.13, the depository loans during both of these periods (see above percentage is higher (16.9 percent) if the analysis paragraph). Between 1996 and 2002, the also excluded from the market totals. is restricted to those newly-originated loans that ‘‘Fannie-Mae-to-market’’ ratio was only 0.84 Excluding manufactured housing loans in depositories do not sell (the latter being a proxy for on the special affordable category, obtained metropolitan areas would reduce the above loans held in depositories’ portfolios). Note that by dividing Fannie Mae’s performance of market percentage for special affordable during the recent, 1999-to-2002 period (also 13.5 percent by the market’s performance of (underserved area) loans by 1.5 (1.1) reported in Table A.13), there is less difference percentage points. The above findings with between the two depository figures. 16.0 percent. Fannie Mae’s market ratio was respect to the GSEs’ longer-term performance 258 Unless stated otherwise, the market in this 0.94 on the low-mod category and 0.93 on the section is defined as the conventional conforming underserved area category. The ‘‘Freddie- are not much affected by the choice of market market without estimated B&C loans. Mac-to-market’’ ratios were lower’0.80 for definition.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00099 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24326 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00100 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.019 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24327

b. Recent Performance, 1999–2002 percent. In addition, Freddie Mac purchased the 1999–2002 market percentage for special This and the next subsection focus on the low-mod loans at about the same rate as affordable loans from 16.4 percent to 15.2 average data for 1999–2002 in Table A.13 and Fannie Mae during this period—42.3 percent percent, which would raise the GSEs’ market the annual data reported in Table A.14. As for the Freddie Mac, 42.5 percent for Fannie ratios from approximately 0.88 to 0.95. explained below, the annual data are useful Mae, and 44.3 percent for the market. Freddie Similarly, excluding manufactured housing for showing shifts in the relative positions of Mac (22.9 percent) purchased underserved loans would reduce the 1999–2002 market Fannie Mae and Freddie Mac that began in area loans at a lower rate than Fannie Mae percentage for underserved areas from 25.8 1999, and for highlighting the improvements (24.0 percent) and the primary market (25.8 percent to 25.0 percent, which would raise made by Fannie Mae during 2001 and 2002 percent). As these figures indicate, both Fannie Mae’s market ratio from 0.93 to 0.96 (which were the first two years under HUD’s Fannie Mae and Freddie Mac continued to and Freddie Mac’s, from 0.89 to 0.92. As higher goal levels) and by Freddie Mac lag the market during this recent four-year shown in Table A.14, Fannie Mae is even during 2002. Between 1993 and 1998, period. Both GSEs’ market ratios were 0.88 closer to the market averages if the year 1999 Freddie Mac’s performance fell below Fannie for special affordable loans and is dropped—over the 2000–2002 period, Mae’s, but a sharp improvement in Freddie approximately 0.95 for low-mod loans. Fannie Mae’s performance on the Mac’s performance during 1999 pushed it Although less than one (where one indicates underserved area category is practically at pass Fannie Mae on all three goals-qualifying equal performance with the market), the market levels under the alternative categories. In 2000, Fannie Mae improved its ‘‘Fannie-Mae-to-market’’ ratio (0.93) for the definitions of the market, and its underserved areas performance enough to underserved area category was higher than performance on the special affordable and surpass Freddie Mac on that category, while the ‘‘Freddie-Mac-to-market’’ ratio (0.89). low-mod categories to close to market levels. Fannie Mae had an uncharacteristically Freddie Mac continued to out-perform c. GSEs’ Performance—Annual Data Fannie Mae on the borrower-income poor year in 1999. Thus, averages for 2000– categories (special affordable and low-mod). 2002 are also presented in Table A.13, Freddie Mac’s Annual Performance. As During 2001 and 2002, Fannie Mae improved dropping 1999. These data show an increase shown by the annual data reported in Table its performance enough to surpass Freddie in Fannie Mae’s performance relative to the A.15, Freddie Mac significantly improved its Mac on all three goals-qualifying categories market, particularly on the special affordable purchases of goals-qualifying loans during and to essentially match the market during and underserved areas categories. Between the 1990s. Its purchases of loans for special these two years. 2000 and 2002, special affordable affordable borrowers increased from 6.5 Consider first the average data for 1999– (underserved area) loans accounted for 15.0 percent of its business in 1992 to 9.2 percent 2002 reported in Table A.13. During this percent (24.9 percent) of Fannie Mae’s in 1997, and then jumped to 14.7 percent in recent period, Freddie Mac’s average purchases, compared with 16.2 percent (26.0 2000 before falling slightly to 14.4 percent in performance was similar to Fannie Mae’s percent) for the market. 2001 and rising again to 15.8 percent in 2002. performance for the borrower income Table A.14 shows the effects on the market The underserved areas share of Freddie categories. Between 1999 and 2002, 14.5 percentages for 1999–2002 (as well as 2000– Mac’s purchases increased at a more modest percent of Freddie Mac’s purchases and 14.4 2002) of different definitions of the rate, rising from 18.6 percent in 1992 to 22.3 percent Fannie Mae’s mortgage purchases conventional conforming market. Excluding percent by 2001; it then jumped to 25.8 consisted of special affordable loans, manufactured housing loans (as well as B&C percent in 2002. compared with a market average of 16.4 loans) in metropolitan areas would reduce BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00101 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24328 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00102 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.020 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24329

With its improved performance, Freddie Mae’s purchases and 24.2 percent of market 2000 to 14.9 percent in 2001, and then Mac closed its gap with the market in originations, for a higher ‘‘Fannie Mae-to- increased it further to 16.3 percent in 2002, funding goals-qualifying loans. In 2002, market’’ ratio of 0.94.261 the latter being the same as the market’s special affordable loans accounted for 15.8 The year 1999 saw a shift in the above performance of 16.3 percent. The ‘‘Fannie- percent of Freddie Mac’s purchases and 16.3 patterns, with Fannie Mae declining in Mae-to-market’’ ratio for special affordable percent of loans originated in the overall performance while the share of goals- loans jumped from 0.79 in 2000 to 1.00 in conventional conforming market, which qualifying loans in the market increased. 2002. Between 2000 and 2001, Fannie Mae produces a ‘‘Freddie-Mac-to-market’’ ratio of Between 1998 and 1999, the special increased its low-mod percentage from 40.8 0.97 (15.8 divided by 16.3). Table A.15 shows affordable share of Fannie Mae’s business percent to 42.9 percent at the same time that the trend in the ‘‘Freddie-Mac-to-market’’ declined from 13.2 percent to 12.5 percent the low-mod share of the primary market was ratio from 1992 to 2002 for each of the goals- while this type of lending in the market falling from 44.4 percent to 42.9 percent, qualifying categories. For the special increased from 15.4 percent to 17.0 percent. placing Fannie Mae at the market’s affordable and low-mod categories, Freddie For this reason, the ‘‘Fannie-Mae-to-market’’ performance in 2001. During 2002, the low- Mac’s performance relative to the market ratio for special affordable loans declined mod share of Fannie Mae’s purchases of remained flat (at approximately 0.60 and sharply from 0.86 in 1998 to 0.74 in 1999. home loans increased further to 45.3 percent, 0.80, respectively) through 1997; by 2002, the The share of Fannie Mae’s purchases in placing Fannie Mae 0.1 percentage point ‘‘Freddie-Mac-to-market’’ ratios had risen to underserved areas also declined, from 22.7 above the market performance of 45.2 0.97 for both the special affordable and low- percent in 1998 to 20.4 percent in 1999, percent. Fannie Mae increased its mod categories. which lowered the ‘‘Fannie-Mae-to-market’’ underserved area percentage from 23.4 Surprisingly, Freddie Mac did not make ratio from 0.94 to 0.81. percent in 2000 to 24.4 in 2001 percent while much progress during the 1990s closing its After declining in 1999, Fannie Mae’s the underserved area share of the primary gap with the market on the underserved areas performance rebounded in 2000, particularly market was falling from 26.4 percent to 25.2 category. The ‘‘Freddie-Mac-to-market’’ ratio on the underserved areas category. Fannie percent, placing Fannie Mae at less than one for underserved areas was approximately the Mae’s underserved areas percentage jumped percentage point from the market’s same in 2000 (0.83) as it was in 1992 (0.84). by three percentage points from 20.4 percent performance. The ‘‘Fannie-Mae-to-market’’ While it rose to 0.88 in 2001, that was due in 1999 to 23.4 percent in 2000. The 2000 ratio for underserved area loans was 0.97 in more to a decline in the market level than to figure was similar to its level in 1997 but 2001. During 2002, the underserved area an improvement in Freddie Mac’s below Fannie Mae’s peak performances of share of Fannie Mae’s purchases of home performance. However, due to a substantial 24–25 percent during 1994 and 1995. loans increased further to 26.7 percent, increase in Freddie Mac’s underserved area Between 1999 and 2000, the ‘‘Fannie-Mae-to- placing Fannie Mae slightly ahead of market percentage from 22.3 percent in 2001 to 25.8 market’’ ratio for underserved areas increased performance (26.4 percent). percent in 2002, Freddie Mac’s performance from 0.82 to 0.89. Fannie Mae improved its Table A.14 reports Fannie Mae’s 2001 and approached market performance (26.4 performance on the special affordable goal at 2002 performance under alternative percent) during 2002. 259 In the ten years a more modest rate. Fannie Mae’s special definitions of the primary market. As shown under the housing goals, the year 2002 affordable percentage increased by 0.8 there, the above results of Fannie Mae’s represented the first time that Freddie Mac’s percentage points from 12.5 percent in 1999 improvement relative to the market during performance in purchasing home loans in to 13.3 percent in 2000. The 2000 figure was 2001 and 2002 are further reinforced when underserved areas had ever been within two similar to its previous peak level (13.2 lower market percentages are used. percentage points of the market’s percent) in 1998). The ‘‘Fannie-Mae-to- Changes in the ‘‘Fannie-Mae-to-Freddie- performance.260 market’’ ratio for special affordable loans Mac’’ Performance Ratio. The above Fannie Mae’s Annual Performance. With increased from 0.74 in 1999 to 0.79 in 2000, discussion documents shifts in the relative respect to purchasing affordable loans, with the latter figure remaining below Fannie performance of Fannie Mae and Freddie Mac Fannie Mae followed a different path than Mae’s peak market ratio (0.86) in 1998. over the past few years. To highlight these Freddie Mac. Fannie Mae improved its Fannie Mae continued its improvement in changing patterns, Table A.15 reports the performance between 1992 and 1998 and purchasing targeted home loans during 2001, ratio of Fannie Mae’s performance to Freddie made much more progress than Freddie Mac at a time when the conventional conforming Mac’s performance for each goals category for in closing its gap with the market. In fact, by market was experiencing a decline in the years 1992 to 2002. As shown there, the 1998, Fannie Mae’s performance was close to affordable lending, and again in 2002, at a ‘‘Fannie-Mae-to-Freddie-Mac’’ ratio for the that of the primary market for some time when the conventional conforming special affordable category increased from important components of affordable lending. market was increasing enough to return approximately one in 1992 (indicating equal In 1992, special affordable loans accounted approximately to its year-2000 level. Thus, performance) to over 1.3 during the 1994–97 for 6.3 percent of Fannie Mae’s purchases during the 2000-to-2002 period, Fannie Mae period, indicating that Fannie Mae clearly and 10.4 percent of all loans originated in the significantly improved its targeted out-performed Freddie Mac during this conforming market, giving a ‘‘Fannie Mae-to- purchasing performance while the primary period. Between 1997 and 2000, Freddie Mac market’’ ratio of 0.61. By 1998, this ratio had market originated targeted home loans at substantially increased its special affordable risen to 0.86, as special affordable loans had about the same rate in 2002 as it did in 2000. share (from 9.2 percent to 14.7 percent), increased to 13.2 percent of Fannie Mae’s As a result, Fannie Mae’s performance during causing the ‘‘Fannie-Mae-to-Freddie-Mac’’ purchases and to 15.4 percent of market 2001 approached the market on the special ratio to fall from 1.27 in 1997 to 0.90 in 2000 originations. A similar trend in market ratios affordable and underserved area categories (indicating Freddie Mac surpassed Fannie can be observed for Fannie Mae on the and matched the market on the low-mod Mae). But Fannie Mae’s stronger performance underserved areas category. In 1992, category. In 2002, Fannie Mae matched the during 2001 and 2002 returned the ratio to underserved areas accounted for 18.3 percent market on the special affordable category, above one (1.03 in both years), indicating of Fannie Mae’s purchases and 22.2 percent and slightly outperformed the market on the slightly better performance for Fannie Mae of market originations, for a ‘‘Fannie Mae-to- low-mod and underserved areas categories. (e.g., 16.3 percent in 2002 versus 15.8 percent market’’ ratio of 0.82. By 1998, underserved As shown in Table A.15, Fannie Mae for Freddie Mac). The ‘‘Fannie-Mae-to- areas accounted for 22.8 percent of Fannie increased its special affordable percentage by Freddie-Mac’’ performance ratio for low-mod 1.6 percentage points, from 13.3 percent in loans followed a similar pattern, standing at 259 Table A.14 reports annual market percentages 1.03 in 2002 (45.3 percent for Fannie Mae that exclude the effects of manufactured housing, 261 Freddie Mac, on the other hand, fell further versus 44.0 percent for Freddie Mac). small loans, and subprime loans. Freddie Mac’s behind the market during this period. In 1992, Prior to 2000, the ‘‘Fannie-Mae-to-Freddie- performance is closer to the market average under Freddie Mac had a slightly higher underserved Mac’’ ratio for underserved areas had also the alternative market definitions, particularly areas percentage (18.6 percent) than Fannie Mae followed a pattern similar to that outlined during 2001 and 2002. (18.3 percent). However, Freddie Mac’s 260 Prior to 2002, Freddie Mac’s performance on underserved areas percentage had only increased to above for special affordable loans, but at a the underserved areas category had not approached 19.8 percent by 1998 (versus 22.7 percent for lower overall level—rising from about one in the market even under the alternative market Fannie Mae). Thus, the ‘‘Freddie Mac-to-market’’ 1992 (indicating equal performance) to definitions reported in Table A.14. ratio fell from 0.84 in 1992 to 0.82 in 1998. approximately 1.2 during the 1994–97

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00103 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24330 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

period, before dropping to slightly below one declined somewhat. The result was that affordable share of Fannie Mae’s purchases of (0.98) in 1999. However, Fannie Mae during 2001 Fannie Mae outperformed newly-originated mortgages was only 14.2 increased its underserved areas percentage Freddie Mac on all three goals-qualifying percent, or 1.4 percentage points below the from 20.4 percent in 1999 to 24.4 percent in categories, and even matched the market on market average of 15.6 percent. Again, Fannie 2001 while Freddie Mac only increased its the low-mod category. During 2002, both Mae improved its overall performance by percentage from 20.9 percent to 22.3 percent. Fannie Mae and Freddie Mac again improved purchasing seasoned loans with a high This resulted in the ‘‘Fannie-Mae-to-Freddie- their performance; Fannie Mae continued to percentage (18.1) of special affordable loans, Mac’’ ratio rising from 0.98 in 1999 to 1.09 outperform Freddie Mac and even matched enabling Fannie Mae to reduce its gap with in 2001. But during 2002, Freddie Mac’s the market on the special affordable category underserved area percentage jumped by 3.5 and slightly outperformed the market on the the market to 0.7 percentage points—14.9 percentage points to 25.8 percent, while low-mod and underserved area categories. percent versus 15.6 percent. Fannie Mae’s increased at a more modest rate While Freddie Mac lagged the market on all As shown in Table A.11, Freddie Mac also (by 2.3 percentage points) to 26.7 percent, three goals-qualifying categories during 2002, followed a strategy of purchasing seasoned with the result being that the ‘‘Fannie-Mae- it had significantly closed its gap with the special affordable loans mainly during 2000 to-Freddie-Mac’’ ratio for underserved area market by the end of 2002, particularly on and 2001. Prior to 2000, Freddie Mac had not loans fell from 1.09 in 2001 to 1.03 in 2002. the underserved area category. pursued such a strategy, or at least not to the To conclude, while Freddie Mac ended the GSE Purchases of Seasoned Loans. When same degree as Fannie Mae. During the 1997– 1990s on a more encouraging note than the GSE data are expressed on a purchase- 99 period, Freddie Mac’s purchases of prior- Fannie Mae, the past three years (2000, 2001, year basis (as in the above analysis), one year mortgages and newly-originated and 2002) have seen a substantial factor which affects each GSE’s performance mortgages had similar percentages of special improvement in Fannie Mae’s performance concerns their purchases of seasoned (prior- affordable (and low-mod) borrowers. Over on all three goals-qualifying categories. year) loans. As shown in Table A.11, Fannie time, there have been small differentials Fannie Mae ended the 1990s with a decline Mae followed a strategy of purchasing between Freddie Mac’s prior-year and newly- in affordable lending performance at the targeted seasoned loans between 1996 and originated mortgages for the underserved same time that Freddie Mac was improving 1998, and again during the past three years— and the share of goals-qualifying loans was all years when Fannie Mae improved its areas category but they have been smaller increasing in the market. Both GSEs’ overall affordable lending performance. For than the differentials for Fannie Mae (see performance during 2000 was encouraging— example, consider Fannie Mae’s underserved Table A.11). Freddie Mac continued to improve, area performance of 24.4 percent during d. GSEs’ Annual Purchases of Home Loans— particularly with respect to the borrower- 2001, which was helped by its purchases of Origination-Year Basis income categories, while Fannie Mae seasoned mortgages on properties located in Table A.16 reports GSE purchase data for reversed its declining performance, underserved neighborhoods. The 1996 to 2002 on an origination-year basis. particularly with respect to underserved underserved area percentage for Fannie Recall that in this case, mortgages purchased areas. During 2000, Freddie Mac Mae’s purchases of newly-originated outperformed Fannie Mae on the special (current-year) mortgages was only 23.3 by a GSE in any particular calendar year are affordable and low-mod categories, while percent in 2001, or 1.9 percentage points allocated to the year that the mortgage was Fannie Mae purchased a higher percentage of below the market average of 25.2 percent. originated, rather than to the year that the loans in underserved areas. During 2001, Fannie Mae obtained its higher overall mortgage was purchased (as in subsections Fannie Mae continued to improve its percentage (24.4 percent) by purchasing C.1–C.3 above). This approach places the performance while Freddie Mac’s seasoned loans with a particularly high GSE and the market data on a consistent, performance remained about the same and concentration (28.3 percent) in underserved current-year basis, as explained earlier. the market’s originations of affordable loans areas. Similarly, during 2001, the special BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00104 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24331

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00105 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.021 24332 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

In general, the comparisons of Freddie first-time homebuyers is one approach for Table A.17 compares the first-time Mac’s and the market’s performance are helping young families enter the homebuyer share of GSE purchases with the similar to those discussed in Sections homeownership market. Therefore, this corresponding share of home loans originated E.9.a–c above, except for some differences on section briefly compares the GSEs’ funding of in the conventional conforming market. the special affordable category. The ‘‘Freddie first-time homebuyer loans with that of Readers are referred to recent work by Bunce Mac to market’’ ratios in Table A.16 show primary lenders in the conventional and Gardner 264 for the derivation of the that Freddie Mac has improved its conforming market. estimates of first-time homebuyer market performance but has also consistently lagged During the past few years, the GSEs have shares reported in Table A.17. This analysis the primary market in funding mortgages increased their purchases of first-time does not include year 2002 data because covered by the housing goals. homebuyer loans. Fannie Mae’s annual market data from the American Housing The ‘‘Fannie Mae to market’’ ratios in purchases of first-time homebuyer loans Survey are not yet available for that year. Table A.16 show that Fannie Mae has increased from approximately 287,000 in Between 1999 and 2001, first-time improved its performance, and has generally 1999 to 373,000 in 2002, while Freddie Mac’s homebuyers accounted for 26.5 percent of outperformed Freddie Mac, but has lagged annual purchases increased from 199,000 to Fannie Mae’s purchases of home loans, 26.5 the primary market in funding mortgages 259,000 during the same period.262 However, percent of Freddie Mac’s, and 37.6 percent of covered by the housing goals. Under the since 1999, the first-time homebuyer share of home loans originated in the conventional origination-year approach, Fannie Mae the GSEs’ purchases of home loans has conforming market. Thus, both Fannie Mae lagged the market on all three housing goal remained relatively flat, varying within the and Freddie Mac fell substantially short of categories during 2001 and on the special 25–28 percent range.263 the primary market in financing first-time affordable and underserved area categories homebuyers during this time period. The during 2002. In 2002, low- and moderate- 262 These figures include estimates of first-time income loans accounted for 45.4 percent of homebuyer loans for those home purchase loans GSEs’ performance was only 70.5 percent of Fannie Mae’s purchases and 45.2 percent of with a missing first-time homebuyer indicator; the market performance (26.5 percent divided by the market originations, placing Fannie Mae estimates were obtained by multiplying the GSE’s 37.6 percent). 0.2 percentage points above the market. first-time homebuyer share (based only on data with BILLING CODE 4210–27–P a first-time homebuyer indicator) by the number of e. GSEs’ Purchases of First-Time Homebuyer loans with a missing first-time homebuyer Mortgages—1999 to 2001 indicator. and 1997 before dropping to about 25 percent in 263 The first-time homebuyer share for Fannie 1998 and 1999. While not a specific housing goal category, Mae was almost 35 percent between 1996 and 1998; 264 See Harold L. Bunce and John L. Gardner, mortgages for first-time homebuyers are an it then dropped to 30 percent in 1998 and to 26 ‘‘First-time Homebuyers in the Conventional important component of the overall home percent in 1999. The first-time homebuyer share for Conforming Market: The Role of the GSEs’’ loan market. Making financing available for Freddie Mac was approximately 29 percent in 1996 (unpublished paper), January 2004.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00106 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24333

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00107 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.022 24334 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Table A.17 also reports first-time conforming market of metropolitan areas based on projected data, Fannie Mae would homebuyer shares for African Americans and between 1999 and 2002; the figure is 43.6 have to improve its performance by 2.3 (4.3) Hispanics and for all minorities. Between percent if the average is computed for the percentage points over its estimated average 1999 and 2001, African-American and years between 1996 and 2002. Loans in the performance of 42.7 percent in 2001 and Hispanic first-time homebuyers accounted B&C portion of the subprime market are 2002, or by 1.4 (3.4) percentage points over for 4.0 percent of Fannie Mae’s purchases of excluded from these market averages. To its estimated 2002 low-mod performance of home loans, 3.4 percent of Freddie Mac’s reach the proposed 45-percent subgoal for 43.6 percent. purchases, and 6.9 percent of home loans 2005, both GSEs would have to improve their The estimated low-mod percentages originated in the conventional conforming historical performance—Fannie Mae by 0.8 between 1999 and 2002 for Freddie Mac were market. For this subgroup, Fannie Mae’s percentage points over its average as follows (with the historical percentages performance is 58 percent of market performance of 44.2 percent in 2001 and from Table A.15 in parentheses): 40.0 (40.8) performance, while Freddie Mac’s 2002, and Freddie by 2.4 percentage points percent for 1999; 41.7 (42.7) percent for 2000; performance is 49 percent of market over its average performance of 42.6 percent 39.8 (41.3) percent for 2001; and 42.1 (44.0) performance. The group of all minority first- during the same period. To reach the 47 percent for 2002; Freddie Mac’s average low- time homebuyers accounted for 6.6 percent percent subgoal in 2007–08, each GSE’s mod performance between 1999 and 2002 of Fannie Mae’s purchases of home loans, 5.8 performance would have to increase by an based on the projected data was 40.9 percent, percent of Freddie Mac’s purchases, and 10.6 additional two percentage points. compared with 42.3 percent based on percent of home loans originated in the As explained earlier, HUD will be re- historical data. To reach the 45-percent conventional conforming market. In this case, benchmarking its median incomes for subgoal based on projected data, Freddie Mac Fannie Mae’s performance is 62 percent of metropolitan areas and non-metropolitan would have to improve its performance by counties based on 2000 Census median market performance, while Freddie Mac’s 4.0 percentage points over its projected incomes, and will be incorporating the effects performance is 55 percent of market average performance of 41.0 percent in 2001 of the new OMB metropolitan area performance. and 2002, or by 2.9 percentage points over its definitions. As explained in Appendix D, Section E.12 below will continue this projected 2002 low-mod performance of 42.1 HUD projected the effects of these two examination of first-time homebuyers by percent. presenting market share analysis that changes on the low- and moderate-income shares of the single-family-owner market for The subgoal applies only to the GSEs’ estimates the GSEs’ overall importance in the purchases in metropolitan areas because the funding of first-time homebuyers. the years 1999–2002. These estimates will be referred to as ‘‘projected data’’ while the HMDA-based market benchmark is only f. Low- and Moderate-Income Subgoal for 1990-based data reported in the various available for metropolitan areas. HMDA data Home Purchase Loans tables will be referred to as ‘‘historical data.’’ for non-metropolitan areas are not reliable The Department is proposing to With the historical data, the average low-mod enough to serve as a market benchmark. The establishing a subgoal of 45 percent for each share of the conventional conforming market Department is also setting home purchase GSE’s purchases of home purchase loans for (without B&C loans) was 44.3 percent for subgoals for the other two goals-qualifying low- and moderate-income families in the home purchase loans (weighted average of categories, as explained in Appendices B and single-family-owner market of metropolitan 1999–2002 percentages in Table A.13); the C. areas for 2005, with the proposed subgoal corresponding average with the projected 10. GSEs Purchases of Total (Home Purchase rising to 46 percent for 2006 and 47 percent data was 43.1 percent, a differential of 1.2 and Refinance) Loans for 2007 and 2008. If the GSEs meet this percentage points. The projected low-mod subgoal, they will be leading the primary percentages for each year between 1999 and Section E.9 examined the GSEs’ market by approximately one percentage 2002 were as follows (with the historical acquisitions of home purchase loans, which point in 2005 and by three percentage points percentages from Table A.15 in parentheses): is appropriate given the importance of the in 2007–08, based on historical data (see 44.0 (44.8) percent for 1999; 43.7 (43.7) GSEs for expanding homeownership below). This home purchase subgoal will percent for 2000; 41.6 (42.9) percent for 2001; opportunities. To provide a complete picture encourage the GSEs to expand and 43.1 (45.2) percent for 2002. The of the GSEs’ mortgage purchases in homeownership opportunities for lower- differentials between the projected and metropolitan areas, Tables A.18, A.19, A.20, income homebuyers who are expected to historical data are larger in 2001 (1.3 and A.21 report the GSEs’ purchases of all enter the housing market over the next few percentage points) and 2002 (2.1 percentage single-family-owner mortgages, including years. As detailed in Section I, there are four points) than in 1999 (0.8 percentage point) both home purchase loans and refinance specific reasons for establishing this subgoal: and 2000 (0.7 percentage point). It appears loans.265 (1) The GSEs have the expertise, resources, that the low-mod share for single-family- Table A.18 provides a long-run perspective and ability to lead the single-family-owner owners in the conventional conforming on the GSEs’ overall performance. Between market, which is their ‘‘bread and butter’’ market will be at least one percentage point 1993 and 2002, as well as during the 1996– business; (2) the GSEs have historically less due to the re-benchmarking of area 2002 period, each GSE’s performance was lagged the primary market for low- and median incomes and the new OMB 80–86 percent of market performance for the moderate-income loans, not lead it; (3) the definitions of metropolitan areas. Thus, special affordable category, 91–95 percent of GSEs can improve their funding of first-time based on projected data, the 45-percent (47 market performance for the low-mod homebuyers and help reduce troublesome percent) subgoal for 2005 (2007) is category, and 88–92 percent of market disparities in homeownership and access to approximately two (four) percentage points performance for the underserved areas mortgage credit; and (4) there are ample above the 1999–2002 market average. category. For example, between 1996 and opportunities for the GSEs to expand their The estimated low-mod percentages 2002, underserved areas accounted for 23.2 purchases in important and growing market between 1999 and 2002 for Fannie Mae were percent of Fannie Mae’s purchases and 22.4 segments such as the market for minority as follows (with the historical percentages percent of Freddie Mac’s purchases, first-time homebuyers. Sections E.9 and G of from Table A.15 in parentheses): 39.2 (40.0) compared with 25.5 percent for the this appendix provide additional information percent for 1999; 40.1 (40.8) percent for 2000; conventional conforming market (without on opportunities for an enhanced GSE role in 41.7 (42.9) percent for 2001; and 43.6 (45.3) B&C loans). Similarly, for special affordable the home purchase market and on the ability percent for 2002; Fannie Mae’s average low- loans, both GSEs lagged the market during of the GSEs to lead that market. mod performance between 1999 and 2002 the 1996–2002 period—Fannie Mae and As shown in Tables A.13 and A.15, low- based on the projected data was 41.4 percent, Freddie Mac averaged approximately 13.0 and moderate-income families accounted for compared with 42.5 percent based on percent while the market was over two an average of 44.3 percent of home purchase historical data. To reach the 45-percent percentage points higher at 15.2 percent. loans originated in the conventional subgoal (47 percent) subgoal for 2005 (2007) BILLING CODE 4210–27–P

265 purchase-year basis; Table A.21 presents similar The GSE total (home purchase and refinance) data on an origination-year basis. data in Tables A.18–A.20 are presented on a

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00108 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24335

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00109 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.023 24336 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Similar to the patterns discussed for home special affordable loans, approximately 0.95 definitions of the conventional conforming purchase loans, Fannie Mae has tended to for low-mod loans, and slightly over 0.90 for market. For example, the average 1999–2002 outperform Freddie Mac. This can be seen by underserved area loans. As with home market share for special affordable examining the various ‘‘Fannie-Mae-to- purchase loans, dropping the year 1999 and (underserved areas) loans would fall to 15.1 Freddie-Mac’’ ratios in Table A.18, which are characterizing recent performance by the (25.3) percent if manufactured housing loans all equal to or greater than one. Over the 2000–2002 period improves the performance recent 1999–2002 period, Fannie Mae and of both GSEs relative to the market, but in metropolitan areas were excluded from the Freddie Mac continued to lag the overall particularly Fannie Mae. Over the 2000–2002 market definition along with B&C loans. In market on all three goals-qualifying period, the ‘‘Fannie-Mae-to-market’’ ratio was this case, the market ratio for Fannie Mae categories. Special affordable (underserved 0.93 for Special Affordable loans, 0.98 for (Freddie Mac) would be was 0.91 (0.91) for area) loans averaged 13.8 (23.8) percent of low-mod loans, and 0.96 for underserved special affordable loans, 0.97 (0.96) for low- Fannie Mae’s purchases, 13.8 (23.1) percent area loans. mod loans, and 0.94 (0.91) for underserved of Freddie Mac’s purchases, and 15.7 (25.7) The above analysis has defined the market area loans. percent of market originations. Considering to exclude B&C loans. Table A.19 shows the BILLING CODE 4210–27–P both GSEs, the market ratio was 0.88 for effects on the market percentages of different

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00110 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24337

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00111 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.024 24338 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Shifts in performance occurred during affordable lending during the refinancing percentage point declines for the 2001 and 2002, the first two years under wave than did either of the GSEs. Fannie Mae underserved areas category were 1.0 for HUD’s higher housing goal targets. Table stood out in 2001 because of its particularly Fannie Mae, 1.9 for Freddie Mac, and 4.0 for A.20 shows that both GSEs improved their small decline in affordable lending. Between the market. By the end of 2001, Fannie Mae overall performance between 1999 and 2000, 2000 and 2001, Fannie Mae’s special led Freddie Mac in all three goals-qualifying but they each fell back a little during the affordable lending fell by only 0.6 percentage categories, and had erased its gap with the heavy refinancing year of 2001. But the points while Freddie Mac’s fell by 2.8 low-mod market, but continued to lag the primary market (without B&C loans) percentage points and the market’s fell by 3.8 market on the special affordable and experienced a much larger decline in percentage points. The corresponding underserved areas categories.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00112 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24339

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00113 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.025 24340 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

During the refinancing wave of 2002, and underserved area categories, in A.19, accounting for B&C loans in this Fannie Mae improved slightly on the special particular. manner reduces the year 2001 HMDA- affordable and low-mod categories and Freddie Mac significantly lagged the reported goal-qualifying shares of the total declined slightly on the underserved area single-family (home purchase and refinance (home purchase and refinance) conforming category. Freddie Mac showed slight loans combined) market during 2001 and market as follows: special affordable, from improvement on the special affordable and 2002. In 2002, the ‘‘Freddie-Mac-to-market’’ 15.0 to 14.5 percent; low-mod, from 42.3 to underserved area categories and remained ratios were 0.93 for special affordable loans, 41.6 percent; and underserved areas, from about the same on the low-mod category. The 0.94 for low-mod loans, and 0.94 for 25.7 to 24.9 percent. Obviously, the GSEs’ market showed the same pattern as Fannie underserved area loans. performance relative to the market will Mae. The end result of these changes can be Subprime Loans. Table A.14 in Section E.9 depend on which market definition is used seen by considering the market ratios in showed that the goals-qualifying shares of the (much as it did with the earlier examples of Table A.20. In 2002, special affordable loans home purchase market did not change much excluding manufactured housing loans in accounted for 14.3 percent of Fannie Mae’s when originations by subprime lenders are metropolitan areas from the market purchases and 14.6 percent of loans excluded from the analysis; the reason is that definition). For example, defining the subprime lenders operate primarily in the originated in the non-B&C portion of the conventional conforming market to exclude refinance market. Therefore, in this section’s conventional conforming market, yielding a subprime loans, rather than only B&C loans, analysis of the total market (including ‘‘Fannie-Mae-to-market’’ ratio of 0.98. Since would increase Fannie Mae’s 2002 special refinance loans), one would expect the Fannie Mae’s market ratio for the special affordable (underserved area) market ratio treatment of subprime lenders to significantly affordable category stood at 0.79 in 2000, affect the market estimates and, indeed, this from 0.98 to 1.01 (0.99 to 1.03). Similarly, it Fannie Mae substantially closed its gap with is the case. For the year 2001, excluding would increase Freddie Mac’s special the market during 2001 and 2002. During this subprime loans reduced the goal-qualifying affordable (underserved area) market ratio period, Fannie Mae also mostly eliminated shares of the total market as follows: special from 0.93 to 0.96 (0.94 to 0.98). For the its market gap for the other two goals- affordable, from 15.0 to 13.9 percent; low- broader-defined low-mod category, qualifying categories. In 2002, underserved mod, from 42.3 to 40.9 percent; and redefining the market to exclude subprime area loans accounted for 24.0 percent of underserved areas, from 25.7 to 23.9 percent. loans, rather than only B&C loans, would Fannie Mae’s purchases and 24.3 percent of (See Table A.19.) Similar declines take place increase Fannie Mae’s (Freddie Mac’s) loans originated in the non-B&C portion of in 2002. market ratio from 0.99 to 1.01 (0.94 to 0.96). the conventional conforming market, As explained earlier, the comparisons in Table A.21 reports GSE purchase data for yielding a ‘‘Fannie-Mae-to-market’’ ratio of this appendix have defined the market to total (home purchase and refinance) loans on 0.99, or approximately one. During 2002, exclude the B&C portion of the subprime an origination-year basis. The ‘‘Freddie Mac- low-mod loans accounted for 42.2 percent of market. Industry observers estimate that A- to-market’’ ratios in Table A.21 show that Fannie Mae’s purchases and 42.6 percent of minus loans account for about two-thirds of Freddie Mac has lagged the primary market loans originated in the market, yielding a all subprime loans while the more risky B&C in funding mortgages covered by the housing ‘‘Fannie-Mae-to-market’’ ratio of 0.99, or loans account for the remaining one-third. As goals. The ‘‘Fannie Mae-to-market’’ ratios in approximately one (also note that Fannie explained earlier, this analysis reduces the Table A.21 show that except for the low-mod Mae slightly outperformed the low-mod goal-qualifying percentages from the HMDA category in 2002 Fannie Mae has lagged the market during 2001). Thus, while Fannie data by half the differentials between (a) the primary market in funding home purchase Mae continued to lag the market in 2002 on market (unadjusted) and (b) the market and refinance mortgages covered by the each of the three goals-qualifying categories, without the specialized subprime lenders housing goals. it was close to the market on the low-mod identified by Scheessele. As shown in Table BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00114 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24341

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00115 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.026 24342 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

11. GSE Mortgage Purchases in Individual of percentages are the same as those used in purchase loans. A GSE’s performance is Metropolitan Areas the aggregate analysis discussed in the above determined to be lagging the market if the While the above analyses, as well as earlier sections. Third, the ‘‘GSE-to-market’’ ratio is ratio of the GSE housing goal loan purchases studies, concentrate on national-level data, it then calculated by dividing each GSE to their overall purchases is less than 99 is also instructive to compare the GSEs’ percentage by the corresponding market percent of that same ratio for the market. purchases of mortgages in individual percentage. For example, if it is calculated (The analysis was conducted where the ‘‘lag’’ metropolitan areas (MSAs). In this section, that one of the GSEs’ purchases of low- and determination is made at 98 percent instead the GSEs’ purchases of single-family owner- moderate-income loans in a particular MSA of 99 percent and the results showed little occupied home purchase loans are compared is 40 percent of their overall purchases in change.) In the example given in the above that MSA, while 44 percent of all home loans to the market in individual MSAs. There are paragraph, that GSE would be considered (excluding B&C loans) in that MSA qualify as three steps. First, goals-qualifying lagging the market. Tables A.22 (2000), A.13 percentages for conventional conforming low-mod, then the GSE-to-market ratio is 40/ (2001) and A.24 report the number of MSAs mortgage originations (without B&C loans) 44 (or 0.91). The goals-qualifying ratios for are computed for each year and for each Fannie Mae and Freddie Mac can be in which each GSE under-performs the MSA, based on HMDA data. Second, compared for each MSA in a similar manner. market with respect to each of the three corresponding goals-qualifying percentages Tables A.22, A.23, and A.24 summarize the housing goal categories. The following points are computed for each GSE’s purchases for performance of the GSEs within MSAs for can be made from this data: each year and for each MSA. These two sets 2000, 2001 and 2002 originations of home BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00116 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24343

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00117 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.027 24344 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00118 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.028 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24345

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00119 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.029 24346 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Fannie Mae’s improvement between 2000 determine if this reflected particularly poor groups such as minority first-time and 2002 shows up clearly in these tables. In performance in a few large MSAs or if it homebuyers. But as this section documents, 2000, Fannie Mae lagged the market in 296 reflected shortfalls in many MSAs. In this the GSEs have been increasing their share of (89 percent) of the 331 MSAs in the purchase case, an analysis of individual MSA data the low-income and minority market, which of underserved area loans; this number increases public understanding of that GSE’s provides an optimistic note on which to go decreased to 267 (81 percent) MSAs in 2001 performance. forward. and to 248 (75 percent) MSAs in 2002. Section E.12.a uses HMDA and GSE data Fannie Mae’s improvement was even greater 12. GSE Market Shares: Home Purchase and to estimate the GSEs’ share of home loans for special affordable and low-mod loans; in First-Time Homebuyer Loans originated for low-income and minority the latter case, Fannie Mae lagged the market This section examines the role that the borrowers and their neighborhoods. Sections in 133 (40 percent) MSAs in 2002, compared GSEs have played in the overall affordable E.12.b and E.12.c summarize recent research with 269 (81 percent) MSAs in 2000. lending market for home loans. There are two on the role of the GSEs in the first-time Freddie Mac’s improvement between 2000 differences from the above analyses in homebuyer market. Section E.12.d examines and 2002 was greater for underserved area Sections E.9 and E.10. The first difference is the downpayment characteristics of home loans. In 2000, Freddie Mac lagged the that this section focuses on ‘‘market share’’ loans purchased by the GSEs, a potentially market in 292 (88 percent) of the 331 MSAs percentages rather than ‘‘distribution of important determinant of the GSEs’ ability to in the purchase of underserved area loans; business’’ percentages. A ‘‘market share’’ reach first-time homebuyers. this number decreased to 260 (79 percent) percentage measures the share of loans with a. GSEs’ Share of Home Purchase Lending MSAs in 2001 and to 193 (58 percent) MSAs a particular borrower or neighborhood in 2002. Freddie Macs made less characteristic that is funded by a particular Table A.25 reports market share estimates improvement on the special affordable and market sector (such as FHA or the GSEs). In derived by combining HMDA market data low-mod categories; in the former case, other words, a ‘‘market share’’ percentage with GSE and FHA loan-level data. To Freddie Mac lagged the market in 234 (71 measures a sector’s share of all home loans understand these estimates, consider the GSE percent) MSAs in 2002, compared with 282 originated for a particular targeted group. The market share percentage of 46 percent for (85 percent) MSAs in 2000. ‘‘market share’’ of a sector depends not only ‘‘All Home Purchase Loans’’ at the bottom of Freddie Mac outperformed Fannie Mae on the degree to which that sector the first column in the table. That market during 2002 in 65 percent of the MSAs, even concentrates its business on a targeted group share percentage is interpreted as follows: though Freddie Mac’s average national (i.e., its ‘‘distribution of business’’ It is estimated that home loans acquired by performance was below Fannie Mae’s in that percentage) but also on the size, or overall Fannie Mae and Freddie Mac during the year (see Table A.16 in Section E.9.d); this mortgage volume, of the sector. If an industry years 1999 to 2002, totaled 46 percent of all suggests that Freddie Mac performs better in sector has a large ‘‘market share’’ for a home loans originated in metropolitan areas small MSAs, as compared with Fannie Mae. targeted group, then that sector is making an during that period. This is also consistent with the fact that important contribution to meeting the credit It should be noted that ‘‘all home loans’’ Fannie Mae lagged the market in 75 percent needs of the group. Both ‘‘distribution of refers to all government (FHA and VA) loans of the MSAs during 2002, even though its business’’ and ‘‘market share’’ data are plus all conventional loans less than the average national performance was only important for evaluating the GSEs‘‘ conforming loan limit; in other words, only slightly below market performance (see Table performance. In fact, given the large size of ‘‘jumbo loans’’ are excluded from this A.16); this suggests Fannie Mae does better the GSEs’, one would expect that a ‘‘market analysis.266 The analysis is restricted to in large MSAs, as compared with small share’’ analysis would highlight their metropolitan areas because HMDA data (the MSAs. importance to the affordable lending market. source of the market estimates) are reliable In its comments on the 2000 Proposed The second difference is that this section only for metropolitan areas. B&C originations Rule, Fannie Mae raised several concerns also examines the role of the GSEs in the are included in the market data, since the about HUD’s comparisons between Fannie total market for home loans, as well as in the purpose here is to gauge the GSEs’ role in the Mae and the primary market at the conventional conforming market. Such an overall mortgage market. As discussed in metropolitan statistical area level. approach provides a useful context for Section E.9, excluding B&C loans, or even all Essentially, Fannie Mae questioned the commenting on the contribution of Fannie subprime loans, would not materially affect relevance of any analysis at the local level, Mae and Freddie Mac to overall affordable this analysis of the home loan market since given that the housing goals are national- lending, particularly given evidence that subprime loans are mainly for refinance level goals. HUD believes that its conventional lenders have done a relatively purposes. The analysis below frequently metropolitan-area analyses support and poor job providing credit access to combines purchases by Fannie Mae and clarify the national analyses on GSE disadvantaged families, which renders the Freddie Mac since previous sections have performance. While official goal performance conventional market a poor benchmark for compared their performance relative to each is measured only at the national level, HUD evaluating GSE performance. The analysis of other. believes that analyses of, for example, the first-time homebuyers conducts the market BILLING CODE 4210–27–P numbers of MSAs where Fannie Mae and share analysis in terms of both the total

Freddie Mac lead or lag the local market market Section E.12.b) and the conventional 266 increases public understanding of the GSEs’ conforming market (Section E.12.c). Following the purchase-year approach used in Sections E.9 and E.10, the GSE purchase data performance. For example, if the national While the GSEs have accounted for a large include their acquisitions of ‘‘prior-year’’ as well as aggregate data showed that one GSE lagged share of the overall market for home ‘‘current-year’’ mortgages, while the market data the market in funding loans in underserved purchase loans, they have accounted for a include only newly-originated (or ‘‘current year’’) areas, it would be of interest to the public to very small share of the market for important mortgages.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00120 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24347

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00121 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.030 24348 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

The GSE market share percentage for a significant margin. For instance, homebuyers.269 Because HMDA does ‘‘Low-Income Borrowers’’ at the top of the between 1999 and 2001, FHA insured not require lenders to report information first column of Table A.25 has a similar 26 percent of all mortgages originated in on first-time homebuyers, Bunce used interpretation: low-income census tracts, which was data from the American Housing Survey It is estimated that home loans for low- only eight percentage points less than to estimate the number of first-time income borrowers acquired by Fannie Mae homebuyers in the market. Using and Freddie Mac between 1999 and 2002 the GSEs’ market share of 34 percent in totaled 37 percent of all home loans low-income census tracts. FHA’s share American Housing Survey data on home originated for low-income borrowers in of the market was particularly high for purchases from 1997 to 1999, Bunce metropolitan areas. African-American and Hispanic estimated that the GSEs’ share of the According to the data in Table A.25, borrowers, as FHA insured 33 percent of market for first-time African-American the GSEs account for a major portion of all home loans originated for these and Hispanic homebuyers was only 10– the market for targeted groups. For borrowers between 1999 and 2002—a 11 percent, or less than one-third of example, purchases by Fannie Mae and figure four percentage points higher their share (36 percent) of all home Freddie Mac represented 37 percent of than the GSEs’ share of 29 percent.268 purchases during that period. FHA’s the low-income-borrower market and Thus, during the 1999–2002 period, share of this market was 36 percent, or 34–37 percent of the markets in low- FHA’s overall market share was only twice its share (18 percent) of all home 270 income, high-minority, and underserved two-fifths (39 percent) of the GSEs’ purchases. These data highlight the census tracts. Thus, access to credit in combined market share, but its share of small role that the GSEs have played in these historically underserved markets the market for loans to African the important market for minority first- depends importantly on the purchase Americans and Hispanics was 14 time homebuyers. Bunce, Neal and Vandenbroucke activities of Fannie Mae and Freddie percent larger than the GSEs’ share of (BNV) recently updated through 2001 Mac. However, the data in Table A.25 that market. the study by Bunce. In addition, BNV show that the GSEs’ role in low-income The data for the two recent years developed an improved methodology and minority markets is significantly (2001 and 2002) indicate a larger market that combined industry, HMDA and less than their role in the overall home role for Fannie Mae and Freddie Mac AHS data to estimate the number of loan market. Fannie Mae and Freddie relative to FHA. While the GSEs first-time homebuyers (by race and Mac accounted for 46 percent of all continued to have a much larger share ethnicity) in the mortgage market during of the overall market than FHA (48–50 home loans but only 36 percent of the the years 1996 to 2001.271 BNV’s percent for the GSEs versus 14–17 loans financing properties in analysis includes the total mortgage underserved neighborhoods. Their percent for FHA), their share of home loans for African Americans and market share was even lower for loans 269 See Harold L. Bunce, The GSEs’ Funding of to African-American and Hispanic Hispanics jumped to 32–34 percent Affordable Loans: A 2000 Update, Housing Finance borrowers—29 percent, or 17 percentage during 2001 and 2002, which was Working Paper No. HF–013, Office of Policy points less than the GSEs’ overall higher than the percentage share for Development and Research, HUD, April 2002. market share of 46 percent. FHA (27–32 percent). The differentials 270 Bunce explains numerous assumptions and caveats related to combining American Housing An encouraging finding is that the in market share between FHA and the Survey data on homebuyers with FHA and GSE GSEs have increased their presence in GSEs on the other affordable lending data on mortgages. For example, the American the affordable lending market during categories listed in Table A.25 were Housing Survey (AHS) data used by Bunce 2001 and 2002, when they accounted for lower in 2001 and 2002 than in earlier included both financed home purchases and homes purchased with cash. If only financed home 38–45 percent of the loans financing years. purchases were used, the market shares of both properties in low-income, high- FHA and the GSEs would have been slightly higher b. The GSEs’ Share of the Total First- minority, and underserved (although the various patterns would have Time Homebuyer Market neighborhoods and for 32–34 percent of remained the same). The AHS defines first-time homebuyers as buyers who have never owned a loans for African-American and This section summarizes two recent home, while FHA and the GSEs define a first-time Hispanic borrowers. These market share analyses of mortgage lending to first- homebuyer more expansively as buyers who have figures for the GSEs are much higher time homebuyers; these two studies not owned a home in the past three years. If it were possible to re-define the FHA and GSE data to be than their performance during the two examine the total mortgage market, consistent with the AHS data, the FHA and GSE earlier years, 1999 and 2000. including both government and first-time homebuyer shares would be lower (to an To provide additional perspective, conventional loans originated unknown degree). For additional caveats with the Table A.25 also reports market share throughout the U.S. (i.e., in both AHS data, also see David A. Vandenbroucke, Sue 267 G. Neal, and Harold L. Bunce, ‘‘First-Time estimates for FHA. During the 1999– metropolitan areas and non- Homebuyers: Trends from the American Housing 2002 period, FHA’s overall market share metropolitan areas). Section E.12.c will Survey,’’ November 2001, U.S. Housing Market was less than half of the GSEs’ market summarize a third study of first-time Condition, a quarterly publication of the Office of share, as FHA insured only 18 percent homebuyers that focuses on the Policy Development and Research at HUD. In some years, home purchases as measured by the AHS of all home mortgages originated in conventional conforming market. All declined while home purchases as measured by metropolitan areas. However, FHA’s three studies are market share studies other data sources (e.g., HMDA) increased. In share of the underserved segments of that examine the GSEs’ role in the first- addition, the AHS home purchase data for separate the market are not far below the GSEs’ minority groups (e.g., African-Americans, time homebuyer market. Hispanics) sometimes exhibited shifts inconsistent share, and in one case actually higher by First, a study by Bunce concluded with other sources. that the GSEs have played a particularly 271 BNV’s methodology for estimating first-time 267 As explained in Section E.7, the GSEs’ small role in funding minority first-time borrowers consists of three steps: (1) Estimate the affordable lending performance is evaluated relative total number of home purchase loans originated to the conventional conforming market, as required during a particular year using a mortgage market by Congress in the 1992 GSE Act that established 268 As explained in the notes to Table A.25, model that they develop; (2) disaggregate the home the housing goals. However, it is insightful to HMDA data are the source of the market figures. It purchase loans in step (1) into racial and ethnic examine their overall role in the mortgage market is assumed that HMDA data cover 85 percent of all groups using HMDA data for metropolitan areas; and to contrast them with other major sectors of the mortgage originations in metropolitan areas. If and (3) for each racial and ethnic group in step (2), market such as FHA. There is no intention here to HMDA data covered higher (lower) percentages of estimate the number of first-time homebuyers using imply that the GSEs should purchase the same market loans, then the market shares for both the mortgage and first-time homebuyer information types of loans that FHA insures. GSEs and FHA would be lower (higher). from the American Housing Survey.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00122 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24349

market, that is, the government, A.26 is interpreted as follows: purchases time homebuyer market was 24.5 conventional conforming, and jumbo of home loans by Fannie Mae and percent during the 1996–to–2001—a sectors of the mortgage market. Freddie Mac totaled 40.7 percent of all market share significantly lower than Table A.26 presents the key market home loans financed between 1996 and their overall market share of 40.7 shares estimated by BNV for the GSEs 2001. Going down the first column percent. and FHA. The first figure (40.7) in Table shows that the GSEs’ share of the first- BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00123 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24350 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00124 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.031 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24351

FHA’s greater focus on first-time market share was two-fifths of the GSEs’ GSEs’ share of the market for first-time homebuyers is also reflected in the share of the overall home purchase African-American and Hispanic market share data reported in Table market (16.4 percent versus 41.5 homebuyers jumped from about 11–12 A.26. While FHA insured only 16.6 percent), FHA’s market share was over percent during 1999 and 2000 to 19.7 percent of all home loans originated three times the GSEs’ share of the percent in 2001. Fannie Mae’s share of between 1996 and 2001, it insured 30.9 market for first-time African-American this market almost doubled during this percent of all first-time-homebuyer and Hispanic homebuyers (46.5 percent period, rising from 7.0 percent in 1999 loans during that period. The GSEs, on versus 14.3 percent). This finding that to 12.6 percent in 2001. Thus, while the the other hand, accounted for a larger the GSEs have played a relatively minor GSEs continue to play a relatively small share (40.7 percent) of the overall home role in the first-time minority market is role in the minority first-time purchase market but a smaller share similar to the conclusion reached by the homebuyer market, during 2001 they (24.5 percent) of the first-time Fed researchers (see below) and Bunce improved their performance in this homebuyer market. (2002) that the GSEs have provided little area.273 Table A.26 also reports home credit support to this underserved purchase and first-time homebuyer borrower group. c. The GSEs’ Share of the Conventional information for minorities. During the The results reported in Table A.26 for Conforming, First-Time Homebuyer more recent 1999-to-2001 period, the the year 2001 suggest some optimism Market GSEs’ loan purchases represented 41.5 concerning the GSEs’ role in the first- Bunce and Gardner (2004) recently percent of all home mortgages but only time homebuyer market. As explained conducted an analysis of first-time 24.3 percent of home loans for African- in earlier sections, both GSEs, but homebuyers for the conventional American and Hispanic families, and particularly Fannie Mae, improved their conforming market. The Bunce and just 14.3 percent of home loans for affordable lending performance during Gardner analysis used a similar African-American and Hispanic first- 2001, at a time when the overall methodology to the study by Bunce, time homebuyers. During this period, market’s performance was slightly Neal, and Vandenbroucke of first-time the GSEs’ role in the market for first- declining. This improvement is homebuyers in the total mortgage time African-American and Hispanic reflected in the higher first-time market market. Bunce and Gardner restricted homebuyers was only one-third of their shares for the GSEs during the year their analysis to the funding of first-time role in the overall home loan market 2001, compared with the two previous homebuyers in the conventional (14.3 percent versus 41.5 percent). years, 1999 and 2000 (not reported). The conforming market, which is the market FHA, on the other hand, accounted where Fannie Mae and Freddie Mac for a much larger share of the minority conventional conforming market in metropolitan operate. Their market share results are areas calculated using the same methodology as first-time homebuyer market than it did Table A.25 with (b) the 1999–2001 market share summarized in Table A.27. of the overall homebuyer market. estimates reported in Table A.25 for the entire BILLING CODE 4210–27–P Between 1999 and 2001, FHA insured mortgage market (including jumbo loans and covering non-metropolitan areas as well as 46.5 percent of all loans for African- 273 metropolitan areas). The results are strikingly For other analyses of the GSEs’ market role, American and Hispanic first-time consistent. For the 1999–to–2001 period, the FHA see the following study by economists at the homebuyers—a market share that was share of the overall (African American and Federal Reserve Board: Glenn B. Canner, Wayne almost three times its overall market Hispanic) home loan market is estimated to be 19.0 Passmore, and Brian J. Surette, ‘‘Distribution of Credit Risk among Providers of Mortgages to Lower- share of 16.4 percent.272 While FHA’s percent (35.8 percent) under (a) versus 16.4 percent (31.2 percent) under (b). Lower percentage shares Income and Minority Homebuyers’’ in Federal are expected for (b) because (b) includes jumbo Reserve Bulletin, 82(12): 1077–1102, December, 272 See Bunce, Neal, and Vandenbroucke, op. cit., loans. For the same period, the GSE share of the 1996. This study considered several characteristics for comparisons of various estimates of the market overall (African American and Hispanic) home loan of the GSEs’ loan purchases (such as amount of shares for FHA and the GSEs using different data market is estimated to be 46.0 percent (25–28 downpayment) and concluded that the GSEs have bases and estimation methods. One can compare (a) percent) under (a) versus 41.5 percent (24.3 percent) played a minimal role in providing credit support the 1999–2001 market shares for FHA and the under (b). for underserved borrowers.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00125 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24352 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00126 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.032 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24353

Between 1999 and 2001, the GSEs’ these loans. Thus, while the GSEs have time homebuyers accounted for 6.9 purchases accounted for 56.6 percent of accounted for 56.6 percent of all home percent of Fannie Mae’s purchases of all home loans originated in the loans in the conventional conforming home loans between 1999 and 2001, a conventional conforming market of both market, they have accounted for only figure higher than Freddie Mac metropolitan areas and non- 30.9 percent of loans originated in that percentage of 5.3 percent. Loans for metropolitan areas. In other words, market for African-American and African-American and Hispanic first- Fannie Mae and Freddie Mac funded Hispanic first-time homebuyers. time homebuyers accounted for 10.2 almost three out of every five The market share data in Table A.27 percent of all home loans originated in homebuyers entering the conventional show some slight differences between the conventional conforming market. conforming market between 1999 and the Freddie Mac and Fannie Mae in d. Downpayments on Loans Purchased 2001. Their purchases of first-time serving minority first-time homebuyers. by the GSEs homebuyer loans, on the other hand, During the 1999-to-2001 period, Freddie accounted for only 39.8 percent of all Mac’s share (11.9 percent) of the The level of downpayment can be an first-time homebuyer loans originated in African-American and Hispanic first- important obstacle to young families that market. Thus, while the GSEs time homebuyer market was only one- seeking their first homes. Examining the funded approximately two out of every half of its share (24.0 percent) of the downpayment characteristics of the five first-time homebuyers entering the home loan market. On the other hand, mortgages purchased by the GSEs might conventional conforming market, their Fannie Mae’s share (19.0 percent) of the help explain why they have played a market share (39.8 percent) for first-time African-American and Hispanic first- rather limited role in the first-time homebuyers was only 70 percent of time homebuyer market was almost 60 homebuyer market. their market share (56.6 percent) for all percent of its share (32.5 percent) of the Table A.28 reports the loan-to-value home buyers. home loan market. Thus, while Fannie (LTV) distribution of home purchase As shown in Table A.27, the GSEs Mae performance in serving minority mortgages acquired by the GSEs have funded an even lower share of the first-time homebuyers has been poor, it between 1997 and 2002. In Table A.29, minority first-time homebuyer market. has been better than Freddie Mac’s. This LTV data are provided for the GSEs’ Between 1999 and 2001, the GSEs difference in performance between purchases of home loans that qualify for purchases of African-American and Fannie Mae and Freddie Mac was also the three housing goals’special Hispanic first-time homebuyer loans seen in the portfolio percentages affordable, low-mod, and underserved represented 30.9 percent of the reported earlier in Table A.17. Loans for areas. Three points stand out. conventional conforming market for African-American and Hispanic first- BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00127 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24354 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00128 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.033 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24355

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00129 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.034 24356 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

First, the GSEs (and particularly remained in the 12–15 percent range. analysis to test whether the Fannie Mae) have recently increased Under the more expansive, over-90- Government-Sponsored Enterprises their purchases of home loans with low percent-LTV definition, almost one- (GSEs) Fannie Mae and Freddie Mac downpayments. After remaining about 4 third of Fannie Mae’s goals-qualifying purchases of home mortgages in percent of Fannie Mae’s purchases purchases during 2001 would be neighborhoods traditionally between 1997 and 2000, over-95- considered low downpayment, as would underserved by financial institutions percent-LTV loans (or less-than-five- a slightly smaller percentage of Freddie stimulate housing market activity in percent downpayment loans) jumped to Mac’s purchases. However, during 2002, those neighborhoods. Specifically, this 7.1 percent during 2001 and 7.7 percent Freddie Mac’s over-90-percent-LTV study analyzes data of single-family in 2002. It is interesting that this jump shares for the goals-qualifying loans fell home sales volumes and prices of in less-than-five-percent downpayment to 23–24 percent. mortgages originated from 1993–1999 in loans occurred in the same years that Third, a noticeable pattern among Cleveland, OH. Fannie Mae improved its purchases of goals-qualifying loans purchased by the The study concludes that aggressive loans for low-income homebuyers, as GSEs is the predominance of loans with secondary market purchasing behavior discussed in earlier sections. As a share high downpayments. For example, 55.9 by non-GSE entities stimulated sales of Freddie Mac’s purchases, over-95- percent of special affordable home loans volumes and prices of homes in low- percent-LTV loans increased from 1.1 purchased by Freddie Mac during 2002 income and predominantly minority- percent in 1997 to 5.9 percent in 2000, had a downpayment of at least 20 occupied neighborhoods of Cleveland. before falling to 4.3 percent in 2001 and percent, a percentage not much lower The study results also showed a positive 4.8 percent in 2002. If the low- than the high-downpayment share (59.1 relationship between home transaction downpayment definition is expanded to percent) of all Freddie Mac’s home loan activity and the actions of the secondary ten percent (i.e., over-90-percent-LTV purchases. Similarly, 46.8 percent of the mortgage market, and concludes that the loans), Freddie Mac had about the same home loans purchased by Fannie Mae in secondary mortgage market (and the percentage (25 percent) of low- underserved areas during 2002 had a 20 non-GSE sector in particular) purchases downpayment loans during 2001 as percent or higher downpayment, of mortgages had a positive effect on the Fannie Mae. In fact, under the more compared with 53.0 percent of all home number of sales transactions one year expansive definition, Freddie Mac had loans purchased by Fannie Mae. later. However, the study also concludes the same share of over-90-percent-LTV Thus, the data in Tables A.28 and that although non-GSE purchases of loans in 2001 as it did in 1997 (about A.29 show a preponderance of high non-home purchase mortgages appeared 25 percent), while Fannie Mae exhibited downpayment loans, even among lower- to boost prices one and two years later, only a modest increase in the share of income borrowers who qualify for the no consistent impacts of purchasing its purchases with low downpayments housing goals. The past focus of the rates on sales prices could be observed. (from 23.2 percent in 1997 to 25.4 GSEs on high-downpayment loans In addition, there was no robust percent in 2001). The share of over-90- provides some insight into a study by evidence that GSE purchasing rates percent-LTV loans in Freddie Mac’s staff at the Federal Reserve Board who were positively associated with single- purchases fell sharply from 25.0 percent found that the GSEs have offered little family home transactions volumes or in 2001 to 21.9 percent in 2002, while credit support to the lower end of the sales prices during any periods. Urban Institute Rural Markets the share in Fannie Mae’s purchases fell mortgage market.274 The fact that Study.276 A study by Jeanette Bradley, more modestly from 25.4 percent in approximately half of the goals- Noah Sawyer, and Kenneth Temkin uses 2001 to 24.2 percent in 2002. qualifying loans purchased by the GSEs both quantitative and qualitative data to have a downpayment of over 20 percent Second, loans that qualify for the explore the issue of GSE service to rural is also consistent with findings reported housing goals have lower areas. The study first summarizes the earlier concerning the GSEs’ minimal downpayments than non-qualifying existing research on rural lending and service to first-time homebuyers, who loans. In 2001 and 2002, over-95- GSE service to rural areas. It then experience the most problems raising percent-LTV loans accounted for about reviews the current underwriting cash for a downpayment. On the other 15 percent of Fannie Mae’s purchases of guidelines of Fannie Mae, Freddie Mac, hand, the recent experience of Fannie special affordable loans, 13 percent of the USDA Rural Housing Service, and Mae suggests that purchasing low- low-mod loans, and 12 percent of Farmer Mac, focusing on issues relevant downpayment loans may be one underserved area loans, compared with to rural underwriting. The GSE public- about 7.5 percent of Fannie Mae’s technique for reaching out and funding use database is used to analyze GSE purchases of all home loans. (See Table low-income and minority families who non-metropolitan loan purchasing A.29.) These low-downpayment shares are seeking to buy their first home. patterns from 1993–2000. Finally, the for 2001 and 2002 were almost double 13. Other Studies of the GSEs study presents the results of a series of those for 2000 when over-95-percent- Performance Relative to the Market discussions conducted with key LTV loans accounted for 8.4 percent of This section summarizes briefly the national industry and lender experts Fannie Mae’s purchases of special and local experts in three rural sites in affordable loans and about 7 percent of main findings from other studies of the GSEs’ affordable housing performance. south-central Indiana, southwestern its purchases of low-mod and New Mexico and southern New underserved area loans. Fannie Mae’s These include studies by the HUD and the GSEs as well as studies by Hampshire chosen for the diversity of low-downpayment shares during 2001 their region, population, economic were higher than Freddie Mac’s shares academics and research organizations. Freeman and Galster Study.275 A structures, and housing markets. of 12.3 percent for special affordable The authors of the study conclude recent study by Lance Freeman and loans and about 8 percent for low-mod that while Fannie Mae and Freddie Mac George Galster uses econometric and underserved area loans. Between have increased their lending to rural 2001 and 2002, Freddie Mac’s over-95- areas since 1993, their non-metropolitan 274 Canner, et al., op. cit. percent-LTV shares fell sharply to 4–5 275 The Impact of Secondary Mortgage Market loan purchases still lag behind their role percent for the three housing goal and GSE Purchases on Underserved Neighborhood categories, while Fannie Mae’s shares Housing Markets: Final Report to HUD. July 2002. 276 GSE Service to Rural Areas, 2002.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00130 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24357

in metropolitan loan purchases, conventional conforming market has loans from subprime and manufactured particularly in regard to the percentage increased, especially for lower income housing lenders. of affordable loans. From the borrowers and neighborhoods. The The study concludes that the GSEs are discussions with experts, the authors of study also concludes that the affordable not leading: They do not purchase the study make the following policy housing goals have an impact on the relatively more underserved market recommendations: underserved purchase decisions of Fannie Mae and loans than the primary market makes populations and rural areas should be Freddie Mac. The study also finds that nor do they purchase as many of these specifically targeted at the census-tract interest rates are lower in markets in loans as their secondary market level; HUD should set manufactured which Fannie Mae and Freddie Mac competitors. Additionally, the study housing goals; HUD should consider purchase a higher proportion of concludes that the disparities between implementing a survey of small rural conventional loans. Finally, the study’s the GSEs and the primary market are lenders or setting up an advisory group analysis shows that overall lending even greater once the growing role of of small rural lenders in order to volume in a metropolitan area increases subprime and manufactured housing is determine their suggestions for creating when the GSEs purchase seasoned considered. The authors admit that stronger relationships between the GSEs loans. there have been signs of progress, and rural lenders with the goal of Specifically, that homeownership particularly in 1999 and 2000 when increasing GSE non-metropolitan rates increased at a faster rate for low- primary market lending to underserved purchase rates. income families when compared to all markets increased and GSE purchases of Urban Institute GSE Impacts Study.277 families, and that in a subset of MSAs, underserved market loans increased A report by Thomas Thibodeau, Brent minority homeownership rates also even faster. Regardless, the study Ambrose, and Kenneth Temkin analyzes grew faster when compared to overall concludes that there continues to be the extent to which the GSEs’ responses homeownership changes in those MSAs. significant racial, economic, and to The Federal Housing Enterprises Finally, the affordable housing goal geographic disparities in the way that Financial Safety and Soundness Act’s effects were examined for 80 MSAs in the benefits of GSE activities are (FHEFSSA) affordable housing goals relation to the homeownership rate distributed and that the benefits of GSE have had their intended effect of making changes between 1991 and 1997. The activities still go disproportionately to low- and moderate-income families study found that the GSEs, by members of served rather than better off. Specifically the report purchasing loans originated to low- underserved markets. examines several methodologies income families, helped to reduce the 14. The GSEs’ Support of the Mortgage determining that the conceptual model disparity between homeownership rates Market for Single-Family Rental created by Van Order in 1996 278 for lower and higher income families, Properties provided the most complete description suggesting that the liquidity created The 1996 Property Owners and of how the primary and secondary when the GSEs purchase loans markets interact. This model was then Managers Survey reported that 49 originated to low-income families is percent of rental units are found in the applied in a narrow scope to capital recycled into more lending targeted to market outcomes which included GSE ‘‘mom and pop shops’’ of the rental lower income homebuyers. market’’single-family’’ rental properties, market shares and effective borrowing The authors of the study qualify their costs, and housing market outcomes that containing 1–4 units. These small results by stating that they are based on properties are largely individually- include low- and moderate-income available data that does not provide the homeownership rates. Finally, owned and managed, and in many cases level of detail necessary to conduct a the owner-managers live in one of the metropolitan American Housing Survey fully controlled national assessment. (AHS) data for eight cities were used to units in the property. They include Williams and Bond Study.279 Richard conduct empirical analyses of the two many properties in older cities, in need Williams and Carolyn Bond examine categories of outcomes. These cities of financing for rehabilitation. Single- GSE leadership of the mortgage finance included areas surveyed in 1992, the family rental units play an especially industry in making credit available for year before HUD adopted the affordable important role in lower-income housing, low- and moderate-income families. housing goals, to provide the baseline over half of such units are affordable to Specifically, it asks if the GSEs are for the analysis. Four metropolitan areas very low-income families. doing relatively more of their business were surveyed in 1992 and again in There is not, however, a strong with underserved markets than other 1996: Cleveland, Indianapolis, Memphis secondary market for single-family financial institutions, and whether the and Oklahoma City. Four cities were rental mortgages. While single-family GSEs’ leadership helps to narrow the surveyed in 1992 and again in 1998: rental properties comprise a large gap in home mortgage lending that Birmingham, Norfolk, Providence and segment of the rental stock for lower- exists between served and underserved Salt Lake City. income families, they make up a small The study’s empirical analysis markets. The study uses HMDA data for portion of the GSEs’ business. In 2001, suggests that the GSE affordable goals metropolitan areas and the Public Use the GSEs purchased $84 billion in have helped to make homeownership Data Base at HUD for compilations of mortgages for such properties, but this more attainable for target families. The GSE data sets for the entire nation (GSE represented 6 percent of the total dollar assessment of the effects of the PUDB File B) to conduct descriptive and volume of the enterprises’ 2002 business affordable goals on capital markets multivariate analyses of nationwide and 10 percent of total single-family showed that the GSE share of the lending between 1993 and 2000. units financed by each GSE. It follows Additionally, separate analyses are that since single-family rentals make up 277 An Analysis of the Effects of the GSE conducted that include and exclude such a small part of the GSEs business, Affordable Goals on Low- and Moderate-Income they have not penetrated the single- Families, 2001. 279 Are the GSEs Leading, and if So Do They Have family rental market to the same degree 278 Van Order, Robert. 1996. ‘‘Discrimination and Any Followers? An Analysis of the GSEs’ Impact on that they have penetrated the owner- the Secondary Mortgage Market.’’ In John Goering Home Purchase Lending to Underserved Markets and Ronald Wienk, eds. Mortgage Discrimination, During the 1990s. University of Notre Dame occupant market. Table A.30 in Section Race, and Federal Policy. The Urban Institute Press, Working Paper and Technical Series Number 2003– G below shows that between 1999 and Washington, DC: 335–363. 2. 2002 2002, the GSEs financed 57 percent of

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00131 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24358 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

owner-occupied dwelling units in the mortgage market during 2005–2008, the conventional conforming mortgage conventional conforming market, but period for which the Low- and market and the share of the goal- only 27 percent of single-family rental Moderate-Income Housing Goal is qualifying markets (low-mod, special units. proposed. The market estimates exclude affordable, and underserved areas) There are a number of factors that B&C loans and allow for much more accounted for by the GSEs’ purchases. have limited the development of the adverse economic and market This analysis, which is conducted by secondary market for single-family affordability conditions than have product type (single-family owner, rental property mortgages thus existed recently. Between 1999 and single-family rental, and multifamily), explaining the lack of penetration by the 2002 the low-mod market averaged shows the relative importance of the GSEs. Little is collectively known about about 57 percent. The detailed analyses GSEs in each of the goal-qualifying these properties as a result of the wide underlying these estimates are markets. spatial dispersion of properties and presented in Appendix D. owners, as well as a wide diversity of 1. GSEs’ Role in Major Sectors of the characteristics across properties and G. Factor 5: GSEs’ Ability To Lead the Mortgage Market individuality of owners. This makes it Industry difficult for lenders to properly evaluate FHEFSSA requires the Secretary, in Tables A.30 and A.31 compare GSE the probability of default and severity of determining the Low- and Moderate- mortgage purchases with HUD’s loss for these properties. Income Housing Goal, to consider the estimates of the numbers of units Single-family rental properties could GSEs’ ability to ‘‘lead the industry in financed in the conventional be important for the GSEs housing goals, making mortgage credit available for conforming market. Table A.30 presents especially for meeting the needs of low- and moderate-income families.’’ aggregate data for 1999–2002 while lower-income families. In 2002 around Congress indicated that this goal should Table A.31 presents more summary 70 percent of single-family rental units ‘‘steer the enterprises toward the market share data for individual years qualified for the Low- and Moderate- development of an increased capacity 2000 and 2002.282 HUD estimates that Income Goals, compared with 40 and commitment to serve this segment there were 48,270,415 owner and rental percent of one-family owner-occupied of the housing market’’ and that it ‘‘fully units financed by new conventional properties. This heavy focus on lower- expect[ed] [that] the enterprises will conforming mortgages between 1999 income families meant that single- need to stretch their efforts to achieve and 2002. Fannie Mae’s and Freddie family rental properties accounted for [these goals].’’281 Mac’s mortgage purchases financed 15 percent of the units qualifying for the The Department and independent 23,580,594 of these dwelling units, or 49 Low- and Moderate-Income Goal, even researchers have published numerous percent of all dwelling units financed. though they accounted for10 percent of studies examining whether or not the As shown in Table A.30, the GSEs have the total units (single-family and GSEs have been leading the single- played a smaller role in the goals- multifamily) financed by the GSEs. family market in terms of their qualifying markets than they have affordable lending performance. This Given the large size of this market, the played in the overall market. Between high percentage of these units which research, which is summarized in 1999 and 2002, new mortgages were qualify for the GSEs’ housing goals, and Section E, concludes that the GSEs have originated for 27,158,020 dwelling units the weakness of the secondary market generally lagged behind primary lenders that qualified for the Low- and for mortgages on these properties, an in funding first-time homebuyers, Moderate-Income Goal; the GSEs low- enhanced presence by Fannie Mae and lower-income borrowers and Freddie Mac in the single-family rental underserved communities. As required mod purchases financed 11,408,692 mortgage market would seem by FHEFSSA, the Department has dwelling units, or 42 percent of the low- warranted.280 Single-family rental produced estimates of the portion of the mod market. Similarly, the GSEs’ housing is an important part of the total (single-family and multifamily) purchases accounted for 41 percent of housing stock because it is an important mortgage market that qualifies for each the underserved areas market, but only source of housing for lower-income of the three housing goals (see 35 percent of the special affordable households. Appendix D). Congress intended that market. Obviously, the GSEs have not the Department use these market been leading the industry in financing F. Factor 4: Size of the Conventional estimates as one factor in setting the units that qualify for the three housing Conforming Mortgage Market Serving percentage target for each of the housing goals. They need to improve their Low- and Moderate-Income Families goals. The Department’s estimate for the performance and it appears that there is Relative to the Overall Conventional size of the Low- and Moderate-Income ample room in the non-GSE portions of Conforming Market market is 51–57 percent, which is the goals-qualifying markets for them to The Department estimates that higher than the GSEs’ performance on do so. For instance, the GSEs were not dwelling units serving low- and that goal. involved in almost two-thirds of the moderate-income families will account This section provides another special affordable market during the for 51–57 percent of total units financed perspective on the GSEs’ performance 1999-to-2002 period. in the overall conventional conforming by examining the share of the total BILLING CODE 4210–27–P

280 A detailed discussion of the GSEs’ activities in Specifically, it considers GSE purchases of: (a) 1999 1999 of both new 1999 originations and originations this area is contained in Theresa R. Diventi, The mortgage originations during 1999 and 2000; (b) during previous years (the latter called ‘‘prior-year’’ GSEs’ Purchases of Single-Family Rental Property 2000 originations during 2000 and 2002; and (c) or seasoned loans). Either approach is a valid Mortgages, Housing Finance Working Paper No. 2002 originations during 2002 (and 2002 will be method for examining GSE purchases; in fact, when added when those data become available in March HF–004, Office of Policy Development and analyzing aggregated data such as the combined 2003). In other words, this analysis looks at the Research, Department of Housing and Urban GSEs’ purchases of a particular origination year 1999–2002 data in Table A.30, the two approaches Development, (March 1998). cohort over a two-year period. This approach yield somewhat similar results. HUD’s methodology 281 Senate Report 102–282, May 15, 1992, p. 35. contrasts with the approach that examines GSE for deriving the market estimates is explained in 282 Tables A.30 and A.31 examine GSE purchases purchases on a ‘‘backward looking basis by Appendix D. B&C loans have been excluded from on a ‘‘going forward basis by origination year.’’ purchase year’’, for example, GSE purchases during the market estimates in Tables A.30 and A.31.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00132 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24359

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00133 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.035 24360 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00134 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.036 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24361

While the GSEs are free to meet the Single-family Rental Market. Single- single-family-owner loans accounted for Department’s goals in any manner that family rental housing is a major source 83 percent of all dwelling units financed they deem appropriate, it is useful to of low-income housing. As discussed in by the GSEs during this period, versus consider their performance relative to Appendix D, data on the size of the 73 percent of all units financed in the the industry by property type. The GSEs primary market for mortgages on these conventional conforming market. accounted for 57 percent of the single- properties is limited, but available While it is recognized that the GSEs family owner market but only 30 information indicate that the GSEs are have been increasing their multifamily percent of the multifamily market and much less active in this market than in purchases, a further enlargement of their 27 percent of the single-family rental the single-family owner market. HUD role in the multifamily market seems market (or a combined 29 percent share estimates that GSE purchases between feasible and appropriate, particularly in of the rental market). 1999 and 2002 totaled only 27 percent the affordable (lower rent) end of the Single-family Owner Market. As of all newly-mortgaged single-family market. As noted in Section D.3, market stated in the 2000 Rule, the single- rental units that were affordable to low- participants believe that the GSEs have family-owner market is the bread-and- and moderate-income families. been conservative in their approaches to butter of the GSEs’ business, and based As explained in the 2000 Rule, many affordable multifamily lending and 284 on the financial and other factors of these properties are ‘‘mom-and-pop’’ underwriting. Certainly the GSEs face discussed below, the GSEs clearly have operations, which may not follow a number of challenges in better meeting the ability to lead the primary market in financing procedures consistent with the needs of the affordable multifamily providing credit for low- and moderate- the GSEs’ guidelines. Much of the market. For example, thrifts and other income owners of single-family financing needed in this area is for depository institutions may sometimes properties. However, the GSEs have rehabilitation loans on 2–4 unit retain their best loans in portfolio, and historically lagged behind the market in properties in older areas, a market in the resulting information asymmetries funding single-family-owner loans that which the GSEs’ have not played a may act as an impediment to expanded major role. However, this sector could secondary market transaction qualify for the housing goals and, as 285 discussed in Section E, they have certainly benefit from an enhanced role volume. However, the GSEs have played a rather small role in funding by the GSEs, and the data in Table A.30 demonstrated that they have the depth minority first-time homebuyers. The indicate that there is room for such an of expertise and the financial resources market share data reported in Table enhanced role, as approximately three- to devise innovative solutions to fourths of this market remains for the A.30 for the single-family-owner market problems in the multifamily market. GSEs to enter. The GSEs can build on their recent tell the same story. The GSEs’ purchases Multifamily Market. Fannie Mae is the records of increased multifamily of single-family-owner loans largest single source of multifamily lending and innovative products to represented 57 percent of all single- finance in the United States, and make further in-roads into the affordable family-owner loans originated between Freddie Mac has made a solid reentry market. As explained in Section D.3, the 1999 and 2002, compared with 53 into this market over the last nine years. GSEs have the expertise and market percent of the low-mod loans that were However, there are a number of presence to push simultaneously for originated, 52 percent of underserved measures by which the GSEs lag the market standardization and for area loans, and 49 percent of the special multifamily market. For example, the programmatic flexibility to meet the affordable loans. share of GSE resources committed to the special needs and circumstances of the The data in Table A.31 indicate the multifamily purchases falls short of the lower-income portion of the multifamily GSEs’ growing market share during the multifamily proportion prevailing in the market. heavy refinance years of 2001 and 2002. overall mortgage market. HUD estimates Conclusions. While HUD recognizes For example, the GSEs accounted for 62 that newly-mortgaged units in that some segments of the market may percent of the overall single-family- multifamily properties represented be more challenging for the GSEs than owner market that year, and 56–58 almost 14 percent of all (single-family others, the data reported in Tables A.30 percent of the markets covered by the and multifamily) dwelling units and A.31 show that the GSEs have three housing goal categories. While this financed between 1999 and 2002.283 As ample opportunities to purchase goals- improvement is an encouraging trend, shown in Table A.30, multifamily qualifying mortgages. Furthermore, if a there are ample opportunities for the acquisitions represented 9 percent of GSE makes a business decision to not GSEs to continue their improvement. dwelling units financed by the GSEs pursue certain types of goals-qualifying Almost one-half of the goals-qualifying between 1999 and 2002. loans in one segment of the market, they loans originated during 2002 remained The GSEs’ role in the multifamily are free to pursue goals-qualifying available to the GSEs to purchase; there market is significantly smaller than in owner and rental property mortgages in are clearly affordable loans being single-family. As shown in Table A.30, other segments of the market. As market originated that the GSEs can purchase. GSE purchases have accounted for 30 leaders, the GSEs should be looking for Furthermore, the GSEs’ purchases under percent of newly financed multifamily innovative ways to pursue this business. the housing goals are not limited to new units between 1999 and 2002—a market Furthermore, there is evidence that the mortgages that are originated in the share much lower than their 57 percent GSEs can earn reasonable returns on current calendar year. The GSEs can share of the single-family-owner market. their goals business. The Regulatory purchase loans from the substantial, Stated in terms of portfolio shares, Analysis that accompanies this existing stock of affordable loans held in proposed rule provides evidence that lenders’ portfolios, after these loans 283 Based on Table A.30, multifamily properties represented 14.5 percent of total units financed have seasoned and the GSEs have had 284 between 1999 and 2002 (obtained by dividing Abt Associates, op. cit. (August 2002). the opportunity to observe their 7,018,044 multifamily units by 48,270,415 ‘‘Total 285 The problem of secondary market ‘‘adverse payment performance. In fact, based on Market’’ units). Increasing the single-family-owner selection’’ is described in James R. Follain and Fannie Mae’s recent experience, the number in Table A.30 by 2,817,258 to account for Edward J. Szymanoski. ‘‘A Framework for purchase of seasoned loans appears to excluded B&C mortgages increases the ‘‘Total Evaluating Government’s Evolving Role in Market’’ number to 51,087,673 which produces a Multifamily Mortgage Markets,’’ Cityscape: A be one effective strategy for purchasing multifamily share of 13.7 percent. See Appendix D Journal of Policy Development and Research 1(2), goals-qualifying loans. for discussion of the B&C market. 1995.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00135 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24362 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

the GSEs can earn financial returns on who do not sell many of their mortgages 2002 were processed through DU. DU their purchases of goals-qualifying loans to Fannie Mae or Freddie Mac. The evaluated more than 10 million loans in that are only slightly below their return guidelines are also commonly followed 2002, compared with 8 million in 2001. on equity from their normal business. in underwriting ‘‘jumbo’’ mortgages, As of the end of 2002, DU had processed which exceed the maximum principal over 26 million loans since its 2. Qualitative Dimensions of the GSEs’ inception. The GSEs’ systems have also Ability To Lead the Industry amount which can be purchased by the GSEs (the conforming loan limit)—such been adapted for FHA and jumbo loans. This section discusses several mortgages eventually might be sold to Automated underwriting systems are qualitative factors that are indicators of the GSEs, as the principal balance is being further adapted to facilitate risk- the GSEs’ ability to lead the industry in amortized or when the conforming loan based pricing, which enables mortgage affordable lending. It discusses the limit is otherwise increased. Changes lenders to offer each borrow an GSEs’ role in the mortgage market; their that the GSEs have made to their individual rate based on his or her risk. ability, through their underwriting underwriting standards in order to As discussed earlier, concerns about the standards, new programs, and address the unique needs of low-income use of automated underwriting and risk- innovative products, to influence the families were discussed in Section C.4 based pricing include the disparate types of loans made by private lenders; of this Appendix. The GSEs’ market impact on minorities and low-income their development and utilization of influence is one reason these new, more borrowers and the ‘‘black box’’ nature of state-of-the-art technology; the flexible underwriting standards have the score algorithm. competence, expertise and training of spread throughout the market. Because The GSEs are using their state-of-the- their staffs; and their financial the GSEs’ guidelines set the credit art technology in certain ways to help resources. standards against which the mortgage expand homeownership opportunities. a. Role in the Mortgage Market applications of lower-income families For example, Fannie Mae has developed are judged, the enterprises have a Fannie Mae Property GeoCoder a The GSEs have played a dominant profound influence on the rate at which computerized mapping service offered role in the single-family mortgage mortgage funds flow to low- and to lenders, nonprofit organizations, and market. As reported in Section C.3, moderate-income borrowers and state and local governments to help mortgage purchases by the GSEs underserved neighborhoods. them determine whether a property is reached extraordinary levels in 2001 As discussed below, the GSEs’ new located in an area that qualifies for and 2003. Purchases by Fannie Mae automated underwriting systems are Fannie Mae’s community lending stood at $568 billion in 2001 and $848 widely used to originate mortgages in products designed to increase billion in 2002. Freddie Mac’s single- today’s market. As discussed in Sections homeownership and revitalization in family mortgage purchases were $393 C.7 and C.8, the GSEs have started traditionally underserved areas. In billion in 2001 and $475 billion in 2002. adapting their underwriting systems for addition, eFannieMae.com is Fannie The Office of Federal Housing subprime loans and other loans that Mae’s business-to-business web site Enterprise Oversight (OFHEO) estimates have not met their traditional where lenders can access product that the GSEs’ purchased 40 percent of underwriting standards. information and important technology newly-originated conventional tools, view upcoming events, and mortgages in 2001. Total GSE purchases, c. State-of-the-Art Technology receive news about training including loans originated in prior Both GSEs are in the forefront of new opportunities. This site receives on 287 years, amounted to 46 percent of developments in mortgage industry average 80,000 visitors per week. conventional originations in 2001. technology. Automated underwriting Freddie Mac has introduced in recent The dominant position of the GSEs in and online mortgage processing are a years internet-based debt auctions, debt the mortgage market is reinforced by couple of the new technologies that repurchase operations, and debt their relationships with other market have impacted the mortgage market, exchanges. These mechanisms benefit institutions. Commercial banks, mutual expanding homeownership investors by providing more uniform savings banks, and savings and loans are opportunities. This section provides an pricing, greater transparency and faster their competitors as well as their overview of these new technologies and price discovery—all of which makes customers—they compete to the extent the extent of their use. Freddie Mac debt more attractive to they hold mortgages in portfolio, but at Each enterprise released an automated investors and reduces the cost of the same time they sell mortgages to the funding mortgages.288 In addition, underwriting system in 1995—Freddie GSEs. They also buy mortgage-backed Freddie Mac has provided automated Mac’s ‘‘Loan Prospector’’ (LP) and securities, as well as the debt securities tools for lenders to identify and work Fannie Mae’s ‘‘Desktop Underwriter’’ used to finance the GSEs’ portfolios. with borrowers most likely to encounter (DU). During 2001 and 2002, roughly 60 Mortgage bankers sell virtually all of problems making their mortgage percent of all newly-originated their prime conventional conforming payments. EarlyIndicator has become mortgages that Freddie Mac purchased loans to the GSEs. Private mortgage the industry standard for default were processed through LP. Lenders and insurers are closely linked to the GSEs, management technology. It can reduce brokers used LP to evaluate 7.3 million because mortgages purchased by the the consequences of mortgage loan applications in 2001 (almost enterprises that have loan-to-value delinquency for borrowers, servicers double the amount in 2000) and 8.2 ratios in excess of 80 percent are and investors.289 million loans in 2002.286 As of the end normally required to be covered by The GSEs are also expanding of 2002, LP had processed 25 million private mortgage insurance, in homeownership opportunities through loans since its inception. Fannie Mae accordance with the GSEs’ charter acts. the use of the Internet in processing also reports that roughly 60 percent of b. Underwriting Standards for the the loans it purchased during 2001 and 287 Fannie Mae, 2002 Annual Housing Activities Primary Mortgage Market Report, 2003, pp. 10–11. 286 This section is based heavily on ‘‘DU and LP 288 Freddie Mac, 2002 Annual Housing Activities The GSEs’ underwriting guidelines Usage Continues to Rise,’’ in Inside Mortgage Report, 2003, p. 14. are followed by virtually all originators Technology published by Inside Mortgage Finance, 289 Freddie Mac, 2002 Annual Housing Activities of prime mortgages, including lenders January 27, 2003, page 1–2. Report, 2003, p. 52.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00136 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24363

mortgage originations. New online in carrying out their current programs, Freddie Mac. Freddie Mac has shown mortgage originations reached $267.6 conducting basic and applied research similar trends. Freddie Mac’s net billion in the first half of 2002, regarding mortgage markets, developing income was $3.7 billion in 2000 and compared with $97 billion for the first innovative new programs, and rose to $10.1 billion in 2002, an increase six months of 2001. The 2002 six-month undertaking sophisticated analyses that of 320 percent from the previous volume comprised 26.5 percent of the may lead to new programs in the future. year.298 Freddie Mac’s return on equity estimated $1.01 trillion in total The role that the GSEs have played in averaged 23.4 percent over the 1995–99 mortgage originations for the same time spreading the use of technology period—also well above the rates period.290 Freddie Mac made Loan throughout the mortgage market reflects achieved by most financial corporations. Prospector on the Internet service the enormous expertise of their staff. Freddie Mac’s return on common equity available to lenders for their retail The leaders of these corporations exceeded 20 percent in 2001 for the operations. Freddie Mac also adopted frequently testify before Congressional twentieth consecutive year, reaching a the mortgage industry’s XML (extensible committees on a wide range of housing high of 39.2 percent in 2001. Freddie markup language) data standard, which issues, and both GSEs have developed Mac’s total revenues grew to $7.4 billion is integral to streamlining and extensive working relationships with a in 2001, up from $4.5 billion in 2000.299 simplifying Internet-based transactions. broad spectrum of mortgage market Investors in Freddie Mac’s common In addition, Congress enacted legislation participants, including various stock have seen their annual dividends that allows the use of electronic nonprofit groups, academics, and per share increase from $0.68 in 2000 to signature in contracts in 2001, making a government housing authorities. $0.88 in 2002.300 Earnings per diluted completely electronic mortgage common share increased from $4.23 in transaction possible. With the use of e. Financial Strength 2001 to $14.18 in 2002.301 electronic signatures, electronic Fannie Mae. The benefits that accrue Other Indicators. Additional mortgages are expected to improve the to the GSEs because of their GSE status, indicators of the strength of the GSEs mortgage process, further reducing as well as their solid management, have are provided by various rankings of origination and servicing costs. In made them two of the nation’s most American corporations. Business Week October 2000, Freddie Mac purchased profitable businesses. Fannie Mae’s net has reported that among Standard & its first electronic mortgage under the income was $3.9 billion in 1999, $4.4 Poor’s 500 companies in 1999, Fannie new law. billion in 2000, $5.9 billion in 2001, and Mae and Freddie Mac respectively Fannie Mae also offers a variety of $4.6 billion in 2002.294 Fannie Mae’s ranked 49th and 88th in market value, 302 other online tools and applications that return on equity averaged 24.0 percent and 24th and 43rd in total profits. have the potential to make the mortgage over the 1995–99 period—far above the Fannie Mae ranked 30th in market value loan process more cost effective and rates achieved by most financial and 13th in total profits in 2001, while efficient for lenders. For example, Freddie Mac ranked 23rd in annual corporations. Fannie Mae’s return on 303 ‘‘HomeBuyer Funds Finder,’’ a one-stop equity reached 26.1 percent in 2002, an growth revenues from 1991–2001. online resource designed for lenders increase of 3 percent over the previous f. Conclusion About Leading the and other housing professionals, enables year.295 In 2002, Fannie Mae’s core Industry users to access a database of local business earnings grew by 19 percent, In light of these considerations, the housing subsidy programs available for credit losses fell to their lowest level low- and moderate-income borrowers. Secretary has determined that the GSEs since 1983 and taxable equivalent have the ability to lead the industry in In 2002, the HomeBuyer Funds Finder revenues grew by 17 percent.296 web site received over 24,500 hits.291 making mortgage credit available for Fannie Mae’s core business earnings low- and moderate-income families. ‘‘Home Counselor Online’’ provides have increased from 39 cents a share in homeownership counselors with the 1987 to $6.31 in 2002, and dividends H. Factor 6: The Need To Maintain the necessary tools to help consumers per common share have increased from Sound Financial Condition of the GSEs financially prepare to purchase a home. $.96 in 1998 to $1.32 in 2002, an 10 HUD has undertaken a separate, As of February 2002, over 1,200 percent increase over 2001. Although detailed economic analysis of this final counselors representing 542 operating earnings per diluted common rule, which includes consideration of (a) organizations were using Home share decreased from 2001 to 2002 by 292 the financial returns that the GSEs earn Counselor Online. A more complete 21% to $4.53, Fannie Mae has still on low- and moderate-income loans and list of Fannie Mae’s online tool and produced double-digit increases for the (b) the financial safety and soundness applications can be found in its Annual past 16 years in core business earnings implications of the housing goals. Based Housing Activities Report. In 2002, per share, placing them among the best on this economic analysis and reviewed Fannie Mae’s total eBusiness volume of the S&P 500 companies.297 by the Office of Federal Housing was $1.1 trillion, up from $800 billion 293 Enterprise Oversight, HUD concludes in 2000. 294 The 22% decrease in Fannie Mae’s 2002 net that the goals raise minimal, if any, d. Staff Resources income resulted primarily from a $4.508 billion safety and soundness concerns. increase in purchased options expense, which Both Fannie Mae and Freddie Mac are occurred due to an increase in the notional amount 298 well-known throughout the mortgage of purchased options outstanding and the declining Freddie Mac, Consolidated Statements of interest rate environment. Recorded purchased Income, Restated November 21, 2003. industry for the expertise of their staffs options expense for 2001 was only $37 million by 299 Freddie Mac, 2001 Annual Report to comparison. Fannie Mae 2002 Annual Report, 2003, Shareholders, pp. 21–22. 290 Inside Mortgage Finance, ‘‘Online Volume p. 23. 300 Freddie Mac, Consolidated Statements of Comprises One-Fourth of Total Originations in First 295 Fannie Mae, 2002 Annual Report to Income, Restated November 21, 2003. Half ‘02,’’ September 20, 2002, p. 8. Shareholders, ‘‘Financial Highlights.’’ 301 Freddie Mac, Consolidated Statements of 291 Fannie Mae, 2002 Annual Housing Activities 296 Fannie Mae, 2002 Annual Report to Income, Restated November 21, 2003. Report, 2003, p. 12. Shareholders, Financial Highlights and Letter to 302 Business Week, March 27, 2000, p. 197. 292 Fannie Mae, 2002 Annual Housing Activities Shareholders. 303 The ‘‘2002 Fortune 500 Top Performing Report, 2003, p. 11. 297 Fannie Mae, 2002 Annual Report to Companies and Industries.’’ .

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00137 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24364 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

I. Determination of the Low- and age will have a dampening effect on the nation’s housing and mortgage Moderate-Income Housing Goals housing demand. However, the markets remain. The homeownership The annual goal for each GSE’s continued influx of immigrants will rate for African-American and Hispanic purchases of mortgages financing increase the demand for rental housing, households is about 25 percentage housing for low- and moderate-income while those who immigrated during the points below that of white households. families is proposed to be established at 1980s and 1990s will be in the market In addition to low income, barriers to 52 percent of eligible units financed in for owner-occupied housing. homeownership that disproportionately each of calendar years 2005, 53 percent Immigrants and other minorities—who affect minorities and immigrants in 2006, 55 percent in 2007, and 57 accounted for nearly 40 percent of the include: lack of capital for down growth in the nation’s homeownership payment and closing costs; poor credit percent in 2008. This goal will remain rate over the past five years—will be history; lack of access to mainstream in effect thereafter, unless changed by responsible for almost two-thirds of the lenders; little understanding of the the Secretary prior to that time. In growth in the number of new homebuying process; and, continued addition, a low- and moderate-income households over the next ten years. discrimination in housing markets and subgoal of 45 percent in 2005, 46 Non-traditional households have mortgage lending. With respect to the percent in 2006, and 47 percent in both become more important, as overall latter, a recent HUD-sponsored study of 2007 and is proposed for the GSEs’ household formation rates have slowed. discrimination in the rental and owner acquisitions of single-family-owner With later marriages, divorce, and non- markets found that while differential home purchase loans in metropolitan traditional living arrangements, the treatment between minority and white areas. This subgoal is designed to fastest growing household groups have home seekers had declined over the past encourage the GSEs to lead the primary been single-parent and single-person ten years, it continued at an market in offering homeownership households. As these demographic unacceptable level in the year 2000. In opportunities to low- and moderate- factors play out, the overall effect on addition, disparities in mortgage income families. The Secretary’s housing demand will likely be sustained lending continued across the nation in consideration of the six statutory factors growth and an increasingly diverse 2002, when the loan denial rate for that led to the choice of these goals is household population from which to African-American applicants was summarized in this section. draw new renters and homeowners. almost three times that for white 1. Housing Needs and Demographic According to the National Association applicants, even after controlling for Conditions of Homebuilders, annual housing income of the applicant. HUD studies demand will average 1.82 million units also show that African Americans and Affordability Problems. Data from the over the next decade. Hispanics are subject to discriminatory 2000 Census and the American Housing Growth in Single-Family Affordable treatment during the pre-qualification Surveys demonstrate that there are Lending. Many younger, minority and process of applying for a mortgage. substantial housing needs among low- lower-income families did not become Single-Family Mortgage Market. and moderate-income families. Many of homeowners during the 1980s due to Heavy refinancing due to low interest these households are burdened by high the slow growth of earnings, high real rates increased single-family mortgage homeownership costs or rent payments interest rates, and continued house originations to record levels during and will likely continue to face serious price increases. Over the past ten years, 2001–2003. Demographic forces, housing problems, given the dim economic expansion, accompanied by industry outreach, and low interest rates prospects for earnings growth in entry- low interest rates and increased also kept lending for home purchase at level occupations. There is evidence of outreach on the part of the mortgage record levels as well. As noted above, deep and persistent housing problems industry, has improved affordability the potential homeowner population for Americans with the lowest incomes. conditions for these families. As this over the next decade will be highly Recent HUD analysis reveals that in appendix explains, there has been a diverse, as growing demand from 1999, 4.9 million households had ‘‘revolution in affordable lending’’ that immigrants and minorities are expected ‘‘worst case’’ housing needs, defined as has extended homeownership to sustain the home purchase market, as housing costs greater than 50 percent of opportunities to historically our population ages. Single-family household income or severely underserved households. The mortgage housing starts are expected to continue inadequate housing among unassisted industry has offered more customized in the 1.65–1.70 million range over the very-low-income renter households. mortgage products, more flexible next few years. Refinancing of existing Among the 34 million renters in all underwriting, and expanded outreach to mortgages, which accounted for about income categories, 6.3 million (19 low-income and minority borrowers. 65 percent of originations during 2000– percent) had a severe rent burden and Fannie Mae and Freddie Mac have been 2003 is expected to return to more over one million renters (3 percent) a big part of this ‘‘revolution in normal levels. As this Appendix lived in housing that was severely affordable lending.’’ HMDA data suggest explains, the GSEs will continue to play inadequate. that the industry and GSE initiatives are a dominant role in the single-family Demographic Trends. Changing increasing the flow of credit to market and will both impact and be population demographics will result in underserved borrowers. Between 1993 affected by major market developments a need for the primary and secondary and 2002, conventional loans to low- such as the growth in subprime lending mortgage markets to meet nontraditional income and minority families increased and the increasing use automated credit needs, respond to diverse housing at much faster rates than loans to upper- underwriting. preferences and overcome information income and non-minority families. Multifamily Mortgage Market. The and other barriers that many immigrants Thus, the 1990s and the early part of the market for financing of multifamily and minorities face. It is projected that current decade have seen the apartments has grown to record there will be 1.2 million new development of a strong affordable volumes. The favorable long-term households each year over the next lending market. prospects for apartments, combined decade. The aging of the baby-boom Disparities in Housing and Mortgage with record low interest rates, have kept generation and the entry of the baby- Markets. Despite this strong growth in investor demand for apartments strong bust generation into prime home buying affordable lending, serious disparities in and supported property prices. As

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00138 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24365

explained below, Fannie Mae and the automobile. The recently heightened unit credit) for Freddie Mac’s purchases Freddie Mac have been among those attention to these issues may increase of mortgages on large (more than 50 boosting volumes and introducing new the acceptance of multifamily rental units) multifamily properties; (c) programs to serve the multifamily construction to both potential customers changes in the treatment of missing market. The long run outlook for the and their prospective neighbors. data; and (d) a procedure for the use of multifamily rental market is sustained, imputed or proxy rents for determining 2. Past Performance of the GSEs moderate growth, based on favorable goal credit for multifamily mortgages. demographics. The minority population, This section reviews the low- and Fannie Mae’s performance was 51.5 especially Hispanics, provides a moderate-income performance of Fannie percent in 2001 and 51.8 percent in growing source of demand for affordable Mae and Freddie Mac. It first reviews 2002, and Freddie Mac’s performance rental housing. ‘‘Lifestyle renters’’ the GSEs’ performance on the Low- and was 53.2 percent in 2001 and 51.4 (older, middle-income households) are Moderate-Income Goal, then reviews percent in 2002; thus both GSEs also a fast growing segment of the rental findings from Section E.2 regarding the surpassed this higher goal. population. However, provision of GSEs’ purchases of home loans for Counting requirements (a) and (b) affordable housing will continue to historically underserved families and expired at the end of 2003, while (c) and challenge suppliers of multifamily their communities. Finally, it reviews (d) will remain in effect after that. If this rental housing and policy makers at all findings from Section G concerning the counting approach—without the bonus levels of governments. Low incomes GSEs’ presence in owner and rental points and the ‘‘temporary adjustment combined with high housing costs markets. factor’’ had been in effect in 2000 and define a difficult situation for millions 2001, and the GSEs had purchased the a. Housing Goals Performance of renter households. Housing cost same mortgages that they actually did reductions are constrained by high land In the October 2000 rule, the low- and purchase in both years, then Fannie prices and construction costs in many moderate-income goal was set at 50 Mae’s performance would have been markets. Government action—through percent for 2001–03. Effective on 51.3 percent in 2000, 49.2 percent in land use regulation, building codes, and January 1, 2001, several changes in 2001, and 49.0 percent in 2002. Freddie occupancy standards—are major counting requirements came into effect Mac’s performance would have been contributors to those high costs. In for the low- and moderate-income goal, 50.6 percent in 2000, 47.7 percent in addition to fewer regulatory barriers and as follows: (a) ‘‘B.00000000onus points’’ 2001, and 46.5 percent in 2001. Thus, costs, multifamily housing would (double credit) for purchases of both Fannie Mae and Freddie Mac benefit from more favorable public mortgages on small (5–50 unit) would have surpassed the low- and attitudes. Higher density housing is a multifamily properties and, above a moderate-income goal of 50 percent in potentially powerful tool for preserving threshold level, mortgages on 2–4 unit 2000 and fallen short in 2001 and 2002. open space, reducing sprawl, and owner-occupied properties; (b) a (See Figure A.1.) promoting transportation alternatives to ‘‘temporary adjustment factor’’ (1.35 BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00139 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24366 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00140 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.037 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24367

b. Single-Family Affordable Lending respectively. Both GSEs are also market over the past ten years, they Market entering new and challenging fields of have generally lagged the overall The GSEs have played a major role in mortgage finance, such as purchasing conventional conforming market in the single-family mortgage market over subprime mortgages. providing affordable loans to lower- the past ten years. Their purchases of Despite these efforts and the overall income borrowers and underserved single-family-owner mortgages gains in goal performance, the areas. This finding is based on HUD’s accounted for 57 percent of all Department remains concerned about analysis of GSE and HMDA data and on mortgages originated in the single- the GSEs’ support of home lending for numerous studies by academics and family conventional conforming market the lower-income end of the market and research organizations. for first-time homebuyers. The lower- between 1999 and 2002. Their • The GSEs have shown different income shares of the GSEs’ purchases underwriting and purchase guidelines patterns of mortgage purchases. Except are too low, particularly for underserved are market standards, used in all for two years (1999 and 2000), Fannie groups such as minority first-time segments of the mortgage market. The Mae has performed better than Freddie homebuyers. GSEs have worked to improve their Mac since 1993 on all three goals- This appendix included a affordable lending record—they have qualifying categories—low-mod, special comprehensive analysis of the GSEs’ introduced new low-downpayment affordable, and underserved areas. As a products targeted at lower-income performance in funding home purchase mortgages for families and communities result, the percentage of Freddie Mac’s families; they have customized their purchases benefiting historically underwriting standards to recognize the that historically have not been well underserved families and their unique needs of immigrant and minority served by the mortgage market. The neighborhoods has been less than the families; and, they have entered into following findings are offered with corresponding shares of total market numerous partnerships with lenders respect to the GSEs’ acquisitions of originations, while Fannie Mae’s and non-profit groups to reach out to home purchase loans that qualify for the purchases have been somewhat closer to underserved populations. The three housing goals (special affordable enterprises’ role in the mortgage market and underserved areas as well as low- the patterns of originations in the is also reflected in their use of cutting and moderate-income) and their primary market. edge technology, such as the acquisitions of first-time homebuyer • The above patterns can be seen by development of Loan Prospector and loans: the following percentage shares of home Desktop Underwriter, the automated • While Fannie Mae and Freddie Mac purchase loans that qualified for the underwriting systems developed by have both improved their support for three housing goals between 1996 and Freddie Mac and Fannie Mae, the single-family affordable lending 2002:

Special Underserved affordable Low-mod areas (percent) (percent) (percent)

Freddie Mac ...... 12.8 39.8 21.7 Fannie Mae ...... 13.5 41.2 23.5 Market (w/o B&C) ...... 16.0 43.6 25.4

• During 2001 and 2002, Fannie Mae match the low-mod market. During 2001 Figure A.2 for the low- and moderate- improved its performance enough to and 2002, Freddie Mac lagged the income shares for Fannie Mae, Freddie reduce its gap in the special affordable conventional conforming market on all Mac and the market. and underserved areas markets and to three goals-qualifying categories; see BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00141 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24368 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00142 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.038 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24369

• Both Fannie Mae and Freddie lag been leading the single-family-owner three subgoals for home purchase loans the conventional conforming market in market in purchasing loans that qualify that qualify for the three housing goals. funding first-time homebuyers, and by a for the housing goals, although Fannie The low- and moderate-income subgoal rather wide margin. Between 1999 and Mae improved its low-mod and is discussed in Section I.3 below. 2001, first-time homebuyers accounted underserved area performance during c. Overall Market Shares for 27 percent of each GSE’s purchases 2001 and 2002 to approach the market of home loans, compared with 38 in funding special affordable and This appendix also included an percent for home loans originated in the underserved areas loans and to match analysis of the GSEs’ role in the overall conventional conforming market. the market in funding low- and (owner and rental) conventional • The GSEs also account for a very moderate-income loans. Still, there is conforming mortgage market. While small share of the market for important room for both Fannie Mae and Freddie GSE mortgage purchases represented 49 groups such as minority first-time Mac to further improve their percent of total dwelling units financed homebuyers. Considering the total performance in purchasing affordable mortgage market (both government and between 1999 and 2002, they loans at the lower-income end of the conventional loans), it is estimated that represented smaller shares of the three market, particularly in the minority the GSEs purchased only 14 percent of goals-qualifying markets: 42 percent of first-time homebuyer market. Evidence loans originated between 1999 and 2001 housing units financed for low- and for African-American and Hispanic first- suggests that there is a significant moderate-income families; 41 percent of time homebuyers, or one-third of their population of potential homebuyers newly-mortgaged units in underserved share (42 percent) of all home purchase who might respond well to aggressive areas; and 35 percent of units financed loans originated during that period. outreach by the GSEs—immigrants and for the very-low-income and other Considering the conventional minorities, in particular, are expected to families that qualify as special conforming market and the same time be a major source of future homebuyers. affordable. (See Figure A.3.) In other period, it is estimated that the GSEs Furthermore, studies indicate the words, the GSEs accounted for purchased only 31 percent of loans existence of a large untapped pool of approximately 40 percent or less of the originated for African-American and potential homeowners among the rental single-family and multifamily units Hispanic first-time homebuyers, or population. Indeed, the GSEs’ recent financed in the goals-qualifying approximately one-half of their share experience with new outreach and markets. This market share analysis (57 percent) of all home purchase loans affordable housing initiatives is suggests that there is room for the GSEs in that market. important confirmation of this potential. to increase their purchases in these To summarize, the Department’s To move the GSEs into a leadership goals-qualifying markets. analysis suggests that the GSEs have not position, the Department is establishing BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00143 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24370 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00144 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.039 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24371

The market analysis also examined 2001, before falling to 7.3 percent metropolitan areas. Loans in the B&C the GSEs’ presence in the three major during heavy refinancing year of 2002. portion of the subprime market are not property sectors of the mortgage market: The increased purchases of included in these averages. To reach the Single-family owner (a 57 percent share multifamily mortgages by Fannie Mae 45-percent (47 percent) subgoal for 2005 for the GSEs between 1999 and 2002), and Freddie Mac have major (for 2007–08), both GSEs would have to single-family rental (a 27 percent share), implications for the Low- and Moderate- improve their historical and multifamily (a 30 percent share). Income Housing Goal, since a very high performance’Fannie Mae by 0.8 The GSEs have historically played a percentage of multifamily units have percentage points (2.8 percentage minimal role in the market financing rents which are affordable to low- and points) over its average performance of single-family rental properties, which is moderate-income families. However, the 44.2 percent in 2001 and 2002, and an important source of low-income potential of the GSEs to lead the Freddie by 2.4 percentage points (4.4 rental housing. Fannie Mae and Freddie multifamily mortgage industry has not percentage points) over its average Mac have increased their purchases of been fully developed. As reported performance of 42.6 percent during the these mortgages, but their purchases earlier in Table A.30, the GSEs’ same period. totaled only 27 percent of the single- purchases between 1999 and 2002 As explained in Section E.9.f, HUD family rental units that received accounted for only 30 percent of the will be re-benchmarking its median financing between 1999 and 2002. A multifamily units that received incomes for metropolitan areas and non- further increased presence by Fannie financing during this period. Certainly metropolitan counties based on 2000 Mae and Freddie Mac would bring there are ample opportunities and room Census median incomes, and will be lower interest rates and liquidity to this for expansion of the GSEs’ share of the incorporating the effects of the new market, as well as improve their housing multifamily mortgage market. The GSEs’ OMB metropolitan area definitions. goals performance. size and market position between loan HUD projected the effects of these two originators and mortgage investors changes on the low- and moderate- d. The GSEs’ Purchases of Multifamily makes them the logical institutions to income shares of the single-family- Mortgages identify and promote needed owner market for the years 1999–2002. Fannie Mae and, especially, Freddie innovations and to establish standards These estimates will be referred to as Mac have rapidly expanded their that will improve market efficiency. As ‘‘projected data’’ while the 1990-based presence in the multifamily mortgage their role in the multifamily market data reported above will be referred to market in the period since the passage continues to grow, the GSEs will have as ‘‘historical data.’’ The average low- of FHEFSSA. The Senate report on this the knowledge and market presence to mod share of the home purchase market legislation in 1992 referred to the GSEs’ push simultaneously for standardization (without B&C loans) was 43.1 percent activities in the multifamily arena as and for programmatic flexibility to meet based on projected data, as compared ‘‘troubling,’’ citing Freddie Mac’s special needs and circumstances, with with 44.3 percent based on historical September 1990 suspension of its the ultimate goal of increasing the data. Thus, based on projected data, the purchases of new multifamily mortgages availability and reducing the cost of 45-percent (47-percent) subgoal is and criticism of Fannie Mae for financing for affordable and other approximately two (four) percentage ‘‘creaming’’ the market.304 multifamily rental properties. points above the 1999–2002 market Freddie Mac has successfully rebuilt average. Fannie Mae’s average low-mod 3. Ability To Lead the Single-Family- its multifamily acquisition program, as performance between 1999 and 2002 Owner Market: A Low- and Moderate- shown by the increase in its purchases based on the projected data was 41.4 Income Subgoal of multifamily mortgages: From $27 percent, compared with 42.5 percent million in 1992 to $3 billion in 1997 As discussed in Section E, the based on historical data. To reach the and then to approximately $7 billion Department is proposing to establish a 45-percent subgoal for 2005 based on during the next three years (1998 to subgoal of 45 percent for each GSE’s projected data, Fannie Mae would have 2000), before rising further to $11.9 purchases of home purchase loans for to improve its performance by 2.3 billion in 2001 and $13.3 billion in low- and moderate-income families in percentage points over its projected 2002. Multifamily properties accounted the single-family-owner market of average performance of 42.7 percent in for over 9 percent of all dwelling units metropolitan areas for 2005, with the 2001 and 2002, or by 1.4 percentage (both owner and rental) financed by subgoal rising to 46 percent in 2006 and points over its projected 2002 low-mod Freddie Mac during 2000 and 2001, and 47 percent in 2007 and 2008. The performance of 43.6 percent. Freddie for 7 percent during the heavy purpose of this subgoal is to encourage Mac’s average low-mod performance refinancing year of 2002. Concerns the GSEs to improve their acquisitions between 1999 and 2002 based on the regarding Freddie Mac’s multifamily of home purchase loans for lower- projected data was 40.9 percent, capabilities no longer constrain their income families and first-time compared with 42.3 percent based on performance with regard to low- and homebuyers who are expected to enter historical data. To reach the 45-percent moderate-income families. the homeownership market over the subgoal for 2005 based on projected Fannie Mae never withdrew from the next few years. If the GSEs meet this data, Freddie Mac would have to multifamily market, but it has also goal, they will be leading the primary improve its performance by 4.0 stepped up its activities in this area market by approximately one percentage percentage points over its projected substantially, with multifamily point in 2005 and by three percentage average performance of 41.0 percent in purchases rising from $3.0 billion in points in 2007 and 2008, based on the 2001 and 2002, or by 2.9 percentage 1992 to $9.4 billion in 1999, $18.7 income characteristics of home points over its projected 2002 low-mod billion in 2001, and $18.3 billion in purchase loans reported in HMDA. performance of 42.1 percent. 2002. Multifamily units as a share of all Between 1999 and 2002 (2000 and The approach taken is for the GSEs to dwelling units (both owner and rental) 2002), HMDA data show that low- and obtain their leadership position by financed by Fannie Mae varied in the moderate-income families accounted for staged increases in the low-mod 10–13 percent range between 1999 and an average of 44.3 (44.2) percent of subgoal; this will enable the GSEs to single-family-owner loans originated in take new initiatives in a 304 Senate Report 1023–282, May 15, 1992, p. 36. the conventional conforming market of correspondingly staged manner to

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00145 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24372 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

achieve the new subgoal each year. funding goals-qualifying loans during sub-markets such as the one for Thus, the increases in the low-mod the period that they have operated minority first-time homebuyers. subgoal are sequenced so that the GSEs under the current definitions of HUD’s (d) There are ample opportunities for can gain experience as they improve housing goals. Between 1996 and 2002 the GSEs to improve their performance. and move toward the new higher (1999 and 2002), low- and moderate- Low- and moderate-income loans are subgoal targets. income mortgages accounted for 39.8 available for the GSEs to purchase, As explained in Section E.9, the (42.3) percent of Freddie Mac’s which means they can improve their subgoal applies only to the GSEs’ purchases, 41.2 (42.5) percent of Fannie performance and lead the primary purchases in metropolitan areas because Mae’s purchases, and 43.6 (44.3) percent market in purchasing loans for the HMDA-based market benchmark is of primary market originations (without borrowers with less-than-median only available for metropolitan areas. B&C loans). The type of improvement income. Three indicators of this have The Department is also setting subgoals needed to meet this new low-mod already been discussed. First, Sections B for the other two goals-qualifying subgoal was demonstrated by Fannie and C of this appendix and Appendix D categories, as follows: 17 percent for Mae during 2001 and 2002, as Fannie explain that the affordable lending special affordable loans and 33 percent Mae increased its low-mod purchases market has shown an underlying for loans in underserved areas. from 40.8 percent of its single-family- strength over the past few years that is The Department considered the owner business in 2000 to 45.3 percent unlikely to vanish (without a significant following factors when setting the in 2002 (or from 40.1 percent in 2000 to increase in interest rates or a decline in subgoal for low- and moderate-income 43.6 percent in 2002 based on projected the economy). The low-mod share of the loans. data). home purchase market has averaged (a) The GSEs have the ability to lead 43.6 percent since 1996 and annually (c) Disparities in Homeownership and the market. The GSEs have the ability to has ranged from 42.2 percent to 45.2 Credit Access Remain. There remain lead the primary market for single- percent. Second, the market share data troublesome disparities in our housing family-owner loans, which is the reported in Table A.30 of Section G and mortgage markets, even after the ‘‘bread-and-butter’’ of their business. demonstrate that there are newly- ‘‘revolution in affordable lending’’ and They both have substantial experience originated loans available each year for the growth in homeownership that has in this market, which means there are the GSEs to purchase. The GSEs’ taken place since the mid-1990s. The no issues as whether or not the GSEs purchases of single-family owner loans have yet penetrated the market, as there homeownership rate for African- represented 57 percent of all single- are with the single-family rental and American and Hispanic households family-owner loans originated between multifamily markets. Both GSEs have remains 25 percentage points below that 1999 and 2002, compared with 53 not only been operating in the owner of white households. Minority families percent of the low-mod loans that were market for years, they have been the face many barriers in the mortgage originated during this period. Thus, dominant players in that market, market, such as lack of capital for down almost one-half of the low-mod funding 57 percent of the single-family- payment and lack of access to conforming market is not touched by the owner mortgages financed between 1999 mainstream lenders (see above). GSEs. As noted above, the situation is and 2002. As discussed in Section G, Immigrants and minorities are projected even more extreme for special sub- their underwriting guidelines are to account for almost two-thirds of the markets such the minority first-time industry standards and their automated growth in the number of new homebuyer market where the GSEs have mortgage systems are widely used households over the next ten years. As only a minimal presence. Finally, the throughout the mortgage industry. emphasized throughout this Appendix, GSEs’ purchases under the subgoal are Through their new downpayment and changing population demographics will not limited to new mortgages that are subprime products, and their various result in a need for the primary and originated in the current calendar year. partnership initiatives, the GSEs have secondary mortgage markets to meet The GSEs can purchase loans from the shown that they have the capacity to nontraditional credit needs, respond to substantial, existing stock of affordable reach out to lower-income families diverse housing preferences and loans held in lenders’ portfolios, after seeking to buy a home. Both Fannie Mae overcome information and other barriers these loans have seasoned and the GSEs and Freddie Mac have the staff expertise that many immigrants and minorities have had the opportunity to observe and financial resources to make the face. The GSEs have to increase their their payment performance. In fact, extra effort to lead the primary market efforts in helping these families because based on Fannie Mae’s recent in funding single-family-owner so far they have played a surprisingly experience, the purchase of seasoned mortgages for low- and moderate- small role in serving minority first-time loans appears to be one useful strategy income mortgages, as well for special homebuyers. It is estimated that the for purchasing goals-qualifying loans. affordable and undeserved area GSEs accounted for 46.5 percent of all To summarize, although single- mortgages. (both government and conventional) family-owner mortgages comprise the (b) The GSEs have lagged the market. home loans originated between 1999 ‘‘bread-and-butter’’ of the GSEs’ Even though the GSEs have the ability and 2001; however, they accounted for business, evidence presented above to lead the market, they have lagged the only 14.3 percent of home loans demonstrates that the shares of their market under the housing goals. The originated for African-American and loans for low- and moderate-income Department and independent Hispanic first-time homebuyers. Within families lag the corresponding shares for researchers have published numerous the conventional conforming market, it the primary market. For the reasons studies examining whether or not the is estimated that the GSEs purchased given above, the Secretary believes that GSEs have been leading the single- only 20 percent of loans originated for the GSEs can do more to raise the low- family market in terms of funding loans African-American and Hispanic first- and moderate-income shares of their that qualify for the three housing goals. time homebuyers, even though they mortgages on these properties. This can While the GSEs, and particularly Fannie accounted for 57 percent of all home be accomplished by building on various Mae, have significantly improved their purchase loans in that market. A programs that the enterprises have performance over the past two years, subgoal for home purchase loans should already started, including (1) their they have lagged the primary market in increase the GSEs’ efforts in important partnership and outreach efforts, (2)

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00146 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24373

their incorporation of greater flexibility performance in 2000. Freddie Mac’s GSEs to increase their purchases of into their underwriting guidelines, (3) performance is estimated to have been loans that qualify for the Low- and their purchases of CRA loans, and (4) 46.0 percent in 1999, 50.2 percent in Moderate-Income Goal. Examples of their targeting of important markets 2000, 47.0 percent in 2001, and 45.0 specific market segments that would where they have had only a limited percent in 2002—for 2005, Freddie Mac particularly benefit from a more active presence in the past, such as the market would have to increase its performance secondary market have been provided for minority first-time homebuyers. A by 4.9 percentage points over its average throughout this appendix. wide variety of quantitative and (unweighted) performance of 47.1 6. Conclusions qualitative indicators indicate that the percent over these last four years, or by GSEs’ have the resources and financial 1.8 percentage points over its previous Having considered the projected strength to improve their affordable peak performance (50.2 percent in mortgage market serving low- and lending performance enough to lead the 2000). By 2008, Freddie Mac’s moderate-income families, economic, market for low- and moderate-income performance would have to increase by housing and demographic conditions for families. 9.9 percentage points over average 2005–08, and the GSEs’ recent 1999–2002 performance, and by 6.8 performance in purchasing mortgages 4. Size of the Mortgage Market for Low- percentage points over its previous peak for low- and moderate-income families, and Moderate-Income Families performance. However, the low- and the Secretary has determined that the As detailed in Appendix D, the low- moderate-income market is estimated to proposed goals of 52 percent of eligible and moderate-income mortgage market be 51–57 percent. Thus, the GSEs units financed in 2005, 53 percent in accounts for 51 to 57 percent of should be able to improve their 2006, 55 percent in 2007, and 57 dwelling units financed by conventional performance enough to meet these percent in 2008 are feasible. The conforming mortgages. In estimating the proposed goals of 52–57 percent. Secretary is also proposing a subgoal of size of the market, HUD excluded the The objective of HUD’s proposed 45 percent for the GSEs’ purchases of effects of the B&C market. HUD also Low- and Moderate-Income Goal is to single-family-owner home purchase used alternative assumptions about bring the GSEs’ performance to the mortgages in metropolitan areas in 2005, future economic and market upper end of HUD’s market range increasing to 46 percent in 2006 and 47 affordability conditions that were less estimate for this goal (51–57 percent), percent in 2007 and 2008. The Secretary favorable than those that existed over consistent with the statutory criterion has considered the GSEs’ ability to lead the last five years. HUD is well aware of that HUD should consider the GSEs’ the industry as well as the GSEs’ the volatility of mortgage markets and ability to lead the market for each Goal. financial condition. The Secretary has the possible impacts of changes in To enable the GSEs to achieve this determined that the proposed goals and economic conditions on the GSEs’ leadership, the Department is proposing the proposed subgoals are necessary and ability to meet the housing goals. modest increases in the Low- and appropriate. Should conditions change such that the Moderate-Income Goal for 2005 which Appendix B—Departmental goals are no longer reasonable or will increase further, year-by-year Considerations To Establish the Central feasible, the Department has the through 2008, to achieve the ultimate Cities, Rural Areas, and Other authority to revise the goals. objective for the GSEs to lead the market Underserved Areas Goal under a range of foreseeable economic 5. The Low- and Moderate-Income A. Introduction Housing Goal for 2005–2008. circumstances by 2008. Such a program of staged increases is consistent with the 1. Establishment of Goal The proposed Low- and Moderate- statutory requirement that HUD The Federal Housing Enterprises Financial Income Housing Goal is 52 percent of consider the past performance of the eligible units for 2005, 53 percent for Safety and Soundness Act of 1992 GSEs in setting the Goals. Staged annual (FHEFSSA) requires the Secretary to 2006, 55 percent for 2007, and 57 increases in the Low- and Moderate- establish an annual goal for the purchase of percent for 2008. It is recognized that Income Goal will provide the mortgages on housing located in central neither GSE met these proposed goals in enterprises with opportunity to adjust cities, rural areas, and other underserved 2001 and 2002. However, the market for their business models and prudently try areas (the ‘‘Underserved Areas Housing the Low- and Moderate-Income Goal is out business strategies, so as to meet the Goal’’). In establishing this annual housing goal, estimated to be 51–57 percent. Under required 2008 level without the new counting rules (i.e., 2000- Section 1334 of FHEFSSA requires the compromising other business objectives Secretary to consider: Census income re-benchmarking and and requirements. 1. Urban and rural housing needs and the the new OMB metropolitan area Figure A.3 summarizes many of the housing needs of underserved areas; definitions), Fannie Mae’s low- and points made in this section regarding 2. Economic, housing, and demographic moderate-income performance is opportunities for Fannie Mae and conditions; estimated to have been 46.3 percent in Freddie Mac to improve their overall 3. The performance and effort of the 1999, 51.2 percent in 2000, 48.7 percent performance on the Low- and Moderate- enterprises toward achieving the in 2001, and 47.9 percent in 2002—for Income Goal. The GSEs’ purchases Underserved Areas Housing Goal in previous 2005, Fannie Mae would have to years; provided financing for 23,580,594 (or 49 4. The size of the conventional mortgage increase its performance by 3.5 percent) of the 48,270,415 single-family market for central cities, rural areas, and percentage points over its average and multifamily units that were other underserved areas relative to the size of (unweighted) performance of 48.5 financed in the conventional the overall conventional mortgage market; percent over these last four years, or by conforming market between 1999 and 5. The ability of the enterprises to lead the 0.8 percentage point over its previous 2002. However, in the low- and industry in making mortgage credit available peak performance (51.2 percent in moderate-income part of the market, the throughout the United States, including 2000). By 2008, Fannie Mae’s 11,408,692 units that were financed by central cities, rural areas, and other underserved areas; and performance would have to increase by GSE purchases represented only 42 6. The need to maintain the sound 8.5 percentage points over average percent of the 27,158,020 dwelling units financial condition of the enterprises. 1999–2002 performance, and by 5.8 that were financed in the market. Thus, Organization of Appendix. The remainder percentage points over its previous peak there appears to ample room for the of Section A first defines the Underserved

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00147 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24374 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Areas Housing Goal for both metropolitan tracts defined as underserved, and this definition covered 21,587 (47.5 percent) of areas and nonmetropolitan areas. Sections B necessitates an adjustment of the goal level. the 45,406 census tracts in metropolitan and C address the first two factors listed Metropolitan Areas. This rule provides that areas, which included 44.3 percent of the above, focusing on findings from the within metropolitan areas, mortgage population and 33.7 percent of the owner- literature on access to mortgage credit in purchases will count toward the goal when occupied units in metropolitan areas. metropolitan areas (Section B) and in those mortgages finance properties that are The census tracts included in HUD’s nonmetropolitan areas (Section C). Separate located in census tracts where (1) median definition of underserved areas exhibit low discussions are provided for metropolitan income of families in the tract does not rates of mortgage access and distressed and nonmetropolitan (rural) areas because of exceed 90 percent of area (MSA) median socioeconomic conditions. Between 1999 and differences in the underlying markets and the income or (2) minorities comprise 30 percent 2002, the unweighted average mortgage data available to measure them. Section D or more of the residents and median income denial rate in these tracts was 17.5 percent, discusses the past performance of the GSEs of families in the tract does not exceed 120 almost double the average denial rate (9.3 on the Underserved Areas Housing Goal (the percent of area median income. percent) in excluded tracts. The underserved third factor) and Sections E–G report the In this Rule, the underserved census tracts tracts include 75.3 percent of the number of Secretary’s findings for the remaining factors. are defined in terms of the 2000 Census Section H presents the Department’s rather than the 1990 Census. As shown in persons below the poverty line in proposals relating to the definition of Table B.1a, switching to 2000 Census data metropolitan areas. underserved areas in nonmetropolitan areas. and re-specified MSA boundaries as of June BILLING CODE 4210–27–P Section I summarizes the Secretary’s 2003, increases the proportions of rationale for establishing a subgoal for single- underserved census tracts, population, tracts. When reporting analysis of mortgage loan family-owner home purchase mortgages and owner-occupied housing units, and denial, origination, and application rates later in for setting the level for the Underserved population below the poverty line in this appendix, tracts are excluded if there are no Areas Housing Goal. metropolitan areas. The definition now purchase or refinance applications. Tracts are also covers 26,959 (51.3 percent) of the 52,585 excluded if: (1) group quarters constitute more than 2. HUD’s Underserved Areas Housing Goal census tracts in metropolitan areas, which 50 percent of housing units or (2) there are less than HUD’s definition of the geographic areas include 48.7 percent of the population and 15 home purchase applications in the tract and the targeted by this goal is basically the same as tract denial rates equal 0 or 100 percent. Excluded 38.0 percent of the owner-occupied housing tracts account for a small percentage of mortgage 1 that used during 1996–2003. It is divided units in metropolitan areas. The 1990–based loan applications (1.4 percent). These tracts are not into a metropolitan component and a excluded from HUD’s underserved areas if they nonmetropolitan component. However, as 1 This analysis excludes Puerto Rico. In addition, meet the income and minority thresholds. Rather, explained below, switching to 2000 Census tracts are excluded if median income is suppressed the tracts are excluded to remove the effects of geography increases the number of census in the underlying census data. There are 379 such outliers from the analysis.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00148 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24375

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00149 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.040 24376 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

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 definition covers 1,260 (61.4 percent) of the 2000 Rule, one finding stands out from the nonmetropolitan median income. 2,052 counties in nonmetropolitan areas, existing research literature on mortgage In 1995, two important factors influenced which include 51.0 percent of the access for different types of neighborhoods: HUD’s definition of nonmetropolitan population, 50.7 percent of the owner- High-minority and low-income underserved areas—lack of available data for occupied housing units, and 64.3 percent of neighborhoods continue to have higher measuring mortgage availability in rural areas the population below the poverty level in mortgage denial rates and lower mortgage and lenders’ difficulty in operating mortgage non-metropolitan areas. The 1990-based origination rates than other neighborhoods. programs at the census tract level in rural definition covered 1,514 (65.5 percent) of the A neighborhood’s minority composition and areas. Because of these factors, the 1995 Rule 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 which included 54.6 percent of the Nonmetropolitan Areas. In designating underserved portions of rural population, 53.4 percent of the owner- nonmetropolitan areas, mortgage purchases areas. As discussed in a later section, HUD occupied units, and 67.9 percent of the poor 2 count toward the Underserved Areas Housing is now proposing to replace the county-based in non-metropolitan areas. Goal for properties which are located in definition with a tract-based definition. BILLING CODE 4210–37–P 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 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00150 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24377

BILLING CODE 4210–37–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00151 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.041 24378 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

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 BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00152 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24379

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00153 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.042 24380 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

GSE Performance. Table B.1d shows the under 1990 Census geography. percentages in the table are adjusted to increases in the GSEs’ overall goals Corresponding 2001 figures (adjusted to be exclude the effects of the bonus points and performance under the more expansive comparable with the 2000 figures) are 35.7 Freddie Mac’s Temporary Adjustment Factor, geography of the 2000 Census. During 2000, percent and 30.4 percent. The figures for which became applicable in 2001 for scoring Fannie Mae’s performance would have been Freddie Mac are 34.1 percent and 29.2 of loans toward the housing goals.) an estimated 37.5 percent if underserved percent for 2000 performance, and 32.5 areas were defined in terms of 2000 Census percent and 28.2 percent for 2001 BILLING CODE 4210–37–P geography, compared with 31.0 percent performance. (The 2001 housing goals

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00154 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24381

BILLING CODE 4210–37–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00155 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.043 24382 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Goal and Subgoal Levels. The Department access to mortgage credit, with higher relative to the degree to which they reflect proposes to establish the Underserved Areas mortgage denial rates and lower origination the credit quality of potential borrowers (as Housing Goal as 38 percent of eligible units rates for mortgages. Thus, the income and indicated by applicants’ available assets, financed for 2005, 39 percent for 2006 and minority composition of an area is a good credit rating, employment history, etc.). 2007, and 40 percent for 2008. measure of whether that area is being Without fully accounting for the HUD is proposing to establish a subgoal of underserved by the mortgage market. creditworthiness of the borrower, racial 33 percent for the share of each GSE’s total • Research supports a targeted differences in denial rates cannot be single-family-owner mortgage purchases that neighborhood-based definition of attributed to lender bias. Some studies of finance single-family-owner properties underservice. Studies conclude that credit disparities have attempted to control located in underserved census tracts of characteristics of mortgage loan applicants for credit risk factors that might influence a metropolitan areas for 2005, with this and the neighborhood where the property is lender’s decision to approve a loan. subgoal rising to 34 percent for 2006 and 35 located are the major determinants of Boston Fed Study. The best example of percent for 2007 and 2008. In this case, mortgage denial rates and origination rates. accounting for credit risk is the study of subgoal performance for a particular calendar Once these characteristics are accounted for, mortgage denial rates by researchers at the year would be calculated for each GSE by other influences, such as location in a central Federal Reserve Bank of Boston.5 This dividing (a) the number of mortgages city, play only a minor role in explaining landmark study found that racial differentials purchased by the GSE that finance single- disparities in mortgage lending.3 in mortgage denial rates cannot be fully family-owner properties located in explained by differences in credit risk. To 1. Discrimination in the Mortgage and underserved areas (i.e., census tracts) of control for credit risk, the Boston Fed Housing Markets—An Overview metropolitan areas by (b) the number of researchers included 38 borrower and loan mortgages purchased by the GSE that finance The nation’s housing and mortgage markets variables indicated by lenders to be critical single-family-owner properties located in are highly efficient systems, where most to loan decisions. For example, the Boston metropolitan areas. As explained in Section homebuyers can put down relatively small Fed study included a measure of the H, the purpose of this subgoal is to encourage amounts of cash and obtain long-term borrower’s credit history, which is a variable the GSEs to lead the primary market in funding at relatively small spreads above the not included in other studies. The Boston funding mortgages in underserved census lender’s borrowing costs. Unfortunately, this Fed study found that minorities’ higher tracts. highly efficient financing system does not denial rates could not be explained fully by work everywhere or for everyone. Studies income and credit risk factors. The denial B. Consideration of Factors 1 and 2 in have shown that access to credit often rate for African Americans and Hispanics Metropolitan Areas: The Housing Needs of depends on improper evaluation of was 17 percent, compared with 11 percent Underserved Urban Areas and Housing, characteristics of the mortgage applicant and for Whites with similar characteristics. That Economic, and Demographic Conditions in the neighborhood in which the applicant is, African Americans and Hispanics were Underserved Urban Areas wishes to buy. In addition, though racial about 60 percent more likely to be denied This section discusses differential access to discrimination has become less blatant in the credit than Whites, even after controlling for mortgage funding in urban areas and home purchase market, studies have shown credit risk characteristics such as credit summarizes available evidence on that it is still widespread in more subtle history, employment stability, liquid assets, identifying those neighborhoods that have forms. Partly as a result of these factors, the self-employment, age, and family status and historically experienced problems gaining homeownership rate for minorities is composition. Although almost all highly- access to credit. Section B.1 provides an substantially below that of whites. Appendix qualified applicants were approved, overview of the problem of unequal access to A provided an overview of the differential treatment was observed among mortgage funding, focusing on discrimination homeownership gaps and lending disparities borrowers with more marginal qualifications. and other housing problems faced by faced by minorities. This section briefly That is, highly-qualified borrowers of all minority families and the communities reviews evidence on lending discrimination races seemed to be treated equally, but in where they live. Section B.2 examines as well as a recent HUD-sponsored study of cases where there was some flaw in the mortgage access at the neighborhood level discrimination in the housing market. application, white applicants seemed to be and discusses in some detail the rationale for Mortgage Denial Rates. A quick look at given the benefit of the doubt more the Underserved Areas Housing Goal in mortgage denial rates reported by Home frequently than minority applicants. A metropolitan areas. The most thorough Mortgage Disclosure Act (HMDA) data subsequent refinement of the data used by studies available provide strong evidence reveals that in 2002 minority denial rates the Federal Reserve Bank of Boston that low-income and high-minority census were higher than those for white loan confirmed the findings of that study.6 tracts are underserved by the mortgage applicants. For lower-income borrowers, the The Boston Fed study, as well as market. Section B.3 presents recent statistics denial rate for African Americans applying reassessments of that study by other on the credit characteristics and for conventional loans was 2.1 times the researchers, concluded that the effect of socioeconomic characteristics of underserved denial rate for white borrowers, while for borrower race on mortgage rejections persists areas under HUD’s definition. Readers are higher-income borrowers, the denial rate for even after controlling for legitimate referred to the expansive literature on this African Americans was 2.7 times the rate for determinants of lenders’ credit decisions.7 issue, which is reviewed some detail in white borrowers.4 Appendix B of HUD’s 2000 Rule. This Differentials in denial rates, such as those 5 Alicia H. Munnell, Lynn E. Browne, James section focuses on some of the main studies reported above, are frequently used to McEneaney, and Geoffrey M.B. Tootell, ‘‘Mortgage and their findings. demonstrate the problems that minorities Lending in Boston: Interpreting HMDA Data,’’ Three main points are made in this section: face obtaining access to mortgage credit. American Economic Review, March 1996. 6 • Both borrowers and neighborhoods can However, an important question is the degree William C. Hunter, ‘‘The Cultural Affinity be identified as currently being underserved to which variations in denial rates reflect Hypothesis and Mortgage Lending Decisions,’’ WP– 95–8, Federal Reserve Bank of Chicago, 1995. by the nation’s housing and mortgage lender bias against certain kinds of borrowers 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 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00156 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24383

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 profitability 8 and by underwriting standards share of discrimination for Hispanic and 2. Evidence About Access to Credit in Urban that have disparate effects on minority and African American home seekers can still be Neighborhoods—An Overview 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 10 were conducted. 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 13 While the responses were or an apartment. 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 released its third Housing iscrimination Americans and 22 percent of Hispanics As explained in HUD’s 2000 Rule, the Study (HDS) in the sale and rental of perceived discrimination, compared to only viability of neighborhoods—whether urban, housing. The study, entitled Discrimination 13 percent of whites. rural, or suburban—depends on the access of in Metropolitan Housing Markets: National Segregation in Urban Areas. their residents to mortgage capital to Results from Phase I of the Housing Discrimination, while not the only cause, purchase and improve their homes. While Discrimination Study (HDS), was conducted contributes to the pervasive level of neighborhood problems are caused by a wide by the Urban Institute.12 The results of this segregation that persists between African range of factors, including substantial HDS were based on 4,600 paired tests of Americans and Whites in our urban areas. inequalities in the distribution of the nation’s minority and non-minority home seekers The Census Bureau recently released one of income and wealth, there is increasing conducted during 2000 in 23 metropolitan the most exhaustive studies of residential agreement that imperfections in the nation’s areas nationwide. The report showed large segregation ever undertaken, entitled Racial housing and mortgage markets are hastening decreases between 1989 and 2000 in the level and Ethnic Residential Segregation in the the decline of distressed neighborhoods. of discrimination experienced by Hispanics United States: 1980–2000.14 The Census Disparate denial of credit based on and African Americans seeking to buy a Bureau found that the United States was still geographic criteria can lead to disinvestment home. There has also been a modest decrease very much racially divided. While African and neighborhood decline. Discrimination Americans have made modest strides, they and other factors, such as inflexible and restrictive underwriting guidelines, limit 8 Since upfront loan fees are frequently remain the most highly segregated racial determined as a percentage of the loan amount, group. The authors said that residential access to mortgage credit and leave potential lenders are discouraged from making smaller loans segregation likely results from a variety of borrowers in certain areas underserved. in older neighborhoods, because such loans factors, including choices people make about Data on mortgage credit flows are far from generate lower revenue and are less profitable to where they want to live, restrictions on their perfect, and issues regarding the lenders. choices, or lack of information. The fact that identification of areas with inadequate access 9 Traditional underwriting practices may have many mainstream lenders do not operate in to credit are both complex and controversial. excluded some lower income families that are, in segregated areas makes it even more difficult For this reason, it is essential to define fact, creditworthy. Such families tend to pay cash, ’underserved areas’ as accurately as possible leaving them without a credit history. In addition, for minorities to obtain access to reasonable- 15 based on existing data and evidence. There the usual front-end and back-end ratios applied to priced mortgage credit. Section C.8 of applicants’ housing expenditures and other on- are three sets of studies that provide the going costs may be too stringent for lower income 13 How Much Do We Know? Public Awareness of rationale for the Department’s definition of households, who typically pay larger shares of their the Nation’s Fair Housing Laws, prepared for HUD underserved areas: (1) Studies examining income for housing (including rent and utilities) by Martin D. Abravanel and Mary K. Cunningham racial discrimination against individual than higher income households. of the Urban Institute, April 2002. mortgage applicants; (2) studies that test 10 Margery A. Turner and Felicity Skidmore, eds., 14 U.S. Bureau of the Census, August 2002. The whether mortgage redlining exists at the Mortgage Lending Discrimination: A Review of co-authors of the study were John Iceland and neighborhood level; and (3) studies that Existing Evidence, The Urban Institute: Washington, Daniel H. Weinberg. For a summary of the study, support HUD’s targeted approach to DC, June 1999. see ‘‘Residential Segregation Still Prevalent,’’ 11 measuring areas that are underserved by the Margery Austin Turner, All Other Things Being National Mortgage News, January 6, 2003, page 1. mortgage market. In combination, these Equal: A Paired Testing Study of Mortgage Lending 15 See Randall M. Scheessele, Black and White studies provide strong support for the Institutions, The Urban Institute Press, April 2002. Disparities in Subprime Mortgage Refinance 12 Margery Austin Turner, Stephen L. Ross, Lending, Housing Finance Working Paper No. HF– definition of underserved areas chosen by George Galster, and John Yinger, Discrimination in 114, Office of Policy Development and Research, HUD. The main studies of discrimination Metropolitan Housing Markets, The Urban Institute U.S. Department of Housing and Urban against individuals have already been Press, November 2002. Development, April 2002. summarized in Section B.1 above. Thus, this

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00157 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24384 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

section focuses on the neighborhood-based More Comprehensive Tests of the Redlining variables is expanded to include measures studies in (2) and (3). As noted above, this Hypothesis. Recent statistical studies have that act as proxies for neighborhood risk, the brief overview of these studies draws from sought to test the redlining hypothesis by results do not reveal a pattern of redlining.’’22 Appendix B of the 1995 GSE Rule; readers more completely controlling for differences Other Redlining Studies. To highlight the are referred there for a more detailed in neighborhood risk and demand. In these methodological problems of single-equation treatment of earlier studies of the issues studies, the explanatory power of studies of mortgage redlining, Fred Phillips- discussed below. neighborhood race is reduced to the extent Patrick and Clifford Rossi developed a a. Controlling for Neighborhood Risk and that the effects of neighborhood risk and simultaneous equation model of the demand Tests of the Redlining Hypothesis demand are accounted for; thus, they do not and supply of mortgages, which they support claims of racially induced mortgage estimated for the Washington, DC In its deliberations leading up to redlining. Many of these studies find that the metropolitan area.23 Phillips-Patrick and FHEFSSA, Congress was concerned about race of the individual borrower is more Rossi found that the supply of mortgages is geographic redlining—the refusal of lenders important than the racial composition of the negatively associated with the racial to make loans in certain neighborhoods neighborhood. However, these studies cannot composition of the neighborhood, which led regardless of the creditworthiness of reach definitive conclusions about redlining them to conclude that the results of single- individual applicants. During the 1980s and because segregation in inner cities makes it early 1990s, a number of studies using equation models (such as the one estimated difficult to distinguish the impacts of by Holmes and Horvitz) are not reliable HMDA data (such as that reported in Tables geographic redlining from the effects of B.2 and B.3, below) attempted to test for the indicators of redlining or its absence. individual discrimination. The following are However, Phillips-Patrick and Rossi noted existence of mortgage redlining. Consistent two good examples of these studies. with the redlining hypothesis, these studies that even their simultaneous equations model Holmes and Horvitz examined variations does not provide definitive evidence of found lower volumes of loans going to low- in conventional mortgage originations across 16 redlining because important underwriting income and high-minority neighborhoods. census tracts in Houston.20 Their model However, such analyses were criticized variables (such as credit history), which are explaining census-tract variations in omitted from their model, may be correlated because they did not distinguish between mortgage originations included the following demand, risk, and supply effects 17—that is, with neighborhood race. types of explanatory variables: (a) The A few studies of neighborhood redlining they did not determine whether loan volume economic viability of the loan, (b) was low because families in high-minority have attempted to control for the credit characteristics of properties in and residents history of the borrower, which is the main and low-income areas were unable to afford of the tract (e.g., house value, income, age homeownership and therefore were not omitted variable in the redlining studies distribution and education level), (c) reviewed so far. Samuel Myers, Jr. and Tsze applying for mortgage loans, or because measures of demand (e.g., recent movers into borrowers in these areas were more likely to Chan, who studied mortgage rejections in the the tract and change in owner-occupied units state of New Jersey in 1990, developed a default on their mortgage obligations, or between 1980 and 1990), (d) measures of proxy for bad credit based on the reasons that because lenders refused to make loans to credit risk (defaults on government-insured 18 19 lenders give in their HMDA reports for creditworthy borrowers in these areas. loans and change in tract house values denying a loan.24 They found that 70 percent between 1980 and 1990), and (e) the racial of the gap in rejection rates could not be 16 These studies, which were conducted at the composition of the tract, as a test for the explained by differences in Black and white census tract level, typically involved regressing the existence of racial redlining. Most of the borrower characteristics, loan characteristics, number of mortgage originations (relative to the neighborhood risk and demand variables number of properties in the census tract) on neighborhoods or bad credit. Myers and Chan were significant determinants of the flow of characteristics of the census tract including its concluded that the unexplained Black-white conventional loans in Houston. The minority composition. A negative coefficient gap in rejection rates is a result of estimate for the minority composition variable was coefficients of the racial composition variables were insignificant, which led discrimination. With respect to the racial often interpreted as suggesting redlining. For a composition of the census tract, they found discussion of these models, see Eugene Perle, Holmes and Horvitz to conclude that Kathryn Lynch, and Jeffrey Horner, ‘‘Model allegations of redlining in the Houston that Blacks are more likely to be denied loans Specification and Local Mortgage Market market could not be supported. in racially integrated or predominantly-white Behavior,’’ Journal of Housing Research, Volume 4, Schill and Wachter include several neighborhoods than in predominantly-Black Issue 2, 1993, pp. 225–243. individual borrower and neighborhood neighborhoods. They concluded that middle- 17 For critiques of the early HMDA studies, see characteristics to explain mortgage class Blacks seeking to move out of the inner Andrew Holmes and Paul Horvitz, ‘‘Mortgage acceptance rates in Philadelphia and city would face problems of discrimination Redlining: Race, Risk, and Demand,’’ The Journal of 25 21 in the suburbs. Finance, Volume 49, No. 1, March 1994, pp. 81–99; Boston. They found that the applicant race and Michael H. Schill and Susan M. Wachter, ‘‘A variables—whether the applicant was African Tale of Two Cities: Racial and Ethnic Geographic American or Hispanic—showed significant 22 Schill and Wachter, page 271. Munnell, et al. Disparities in Home Mortgage Lending in Boston negative effects on the probability that a loan reached similar conclusions in their study of and Philadelphia,’’ Journal of Housing Research, would be accepted. Schill and Wachter stated Boston. They found that the race of the individual Volume 4, Issue 2, 1993, pp. 245–276. that this finding does not provide evidence mattered, but that once individual characteristics 18 Like early HMDA studies, an analysis of deed of individual race discrimination because were controlled, racial composition of the neighborhood was insignificant. transfer data in Boston found lower rates of applicant race is most likely serving as a 23 Fred J. Phillips-Patrick and Clifford V. Rossi, mortgage activity in minority neighborhoods. The proxy for credit risk variables omitted from discrepancies held even after controlling for ‘‘Statistical Evidence of Mortgage Redlining?’’ A income, house values and other economic and non- their model (e.g., credit history, wealth and Cautionary Tale‘‘, The Journal of Real Estate racial factors that might explain differences in liquid assets). Schill and Wachter find that Research, Volume 11, Number 1, 1996, pp. 13–23. demand and housing market activity. The study when their neighborhood risk proxies are 24 Samuel L. Myers, Jr. and Tsze Chan, ‘‘Racial concluded that ‘‘the housing market and the credit included in the model along with the Discrimination in Housing Markets: Accounting for market together are functioning in a way that has individual loan variables, the percentage of Credit Risk,’’, Social Science Quarterly, Volume 76, hurt African American neighborhoods in the city of the census tract that was African American Number 3, September 1995, pp. 543–561. Boston.’’ Katherine L. Bradbury, Karl E. Case, and became insignificant. Thus, similarly to 25 For another study that uses HMDA data on Constance R. Dunham, ‘‘Geographic Patterns of Holmes and Horvitz, Schill and Wachter reasons for denial to construct a proxy for bad Mortgage Lending in Boston, 1982–1987,’’ New stated that ‘‘once the set of independent credit, see Steven R. Holloway, ‘‘Exploring the England Economic Review, September/October Neighborhood Contingency of Race Discrimination 1989, pp. 3–30. in Mortgage Lending in Columbus, Ohio’’, Annals 19 Using an analytical approach similar to that of 17, No. 2, 1988, pp. 137–163; and ‘‘Financing of the Association of American Geographers, Bradbury, Case, and Dunham, Anne Shlay found Community: Methods for Assessing Residential Volume 88, Number 2, 1998, pp. 252–276. evidence of fewer mortgage loans originated in Credit Disparities, Market Barriers, and Institutional Holloway finds that mortgage denial rates are black census tracts in Chicago and Baltimore. See Reinvestment Performance in the Metropolis,’’ higher for black applicants (particularly those who Anne Shlay, ‘‘Not in That Neighborhood: The Journal of Urban Affairs, Volume 11, No. 3, 1989, are making large loan requests) in all-white Effects of Population and Housing on the pp. 201–223. neighborhoods that in minority neighborhoods, Distribution of Mortgage Finance within the 20 Holmes and Horitz, op. cit. while the reverse is true for white applicants Chicago SMSA,’’ Social Science Research, Volume 21 Schill and Wachter, op. cit. making small loan requests.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00158 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24385

Geoffrey Tootell has authored two papers the high denial rates for applications without minority areas. Empirical research by David on neighborhood redlining based on the PMI in low-income tracts. Ling and Susan Wachter found that recent mortgage rejection data from the Boston Fed Studies of Information Externalities. tract-level sales transaction volume does study.26 Tootell’s studies are important Another group of studies related to redlining significantly contribute to mortgage loan because they include a direct measure of and the credit problems facing low-income acceptance rates in Dade County, Florida, borrower credit history, as well as the other and minority neighborhoods focus on the also consistent with the Lang and Nakamura underwriting, borrower, and neighborhood ‘‘thin’’ mortgage markets in these hypothesis.31 characteristics that are included in the neighborhoods and the implications of Robert Avery, Patricia Beeson, and Mark Boston Fed data base; thus, his work does not lenders not having enough information about Sniderman found significant evidence of have the problem of omitted variables to the the collateral and other characteristics of economies associated with the scale of these neighborhoods. The low numbers of operation of individual lenders in a same extent as previous redlining studies.27 house sales and mortgages originated in low- neighborhood.32 They concluded that ‘‘The Tootell found that lenders in the Boston area income and high-minority neighborhoods inability to exploit these economies of scale did not appear to be redlining neighborhoods result in individual lenders perceiving these is found to explain a substantial portion of based on the racial composition of the census neighborhoods to be more risky. It is argued the higher denial rates observed in low- tract or the average income in the tract. that lenders do not have enough historical income and minority neighborhoods, where Consistent with the Boston Fed and Schill information to project the expected default the markets are generally thin.’’ Low-income and Wachter studies, Tootell found that it is performance of loans in low-income and and minority neighborhoods often suffer the race of the applicant that mostly affects high-minority neighborhoods, which from low transactions volume, and low the mortgage lending decision; the location of increases their uncertainty about investing in transactions volume represents a barrier to the applicant’s property appears to be far less these areas. the availability of mortgage credit by making relevant. However, he did find that the This recent group of studies that focus on mortgage lenders more reluctant to approve decision to require private mortgage economies of scale in the collection of and originate mortgage loans in these areas. insurance (PMI) depends on the racial information about neighborhood b. Geographic Dimensions of Underserved composition of the neighborhood. Tootell characteristics has implications for the Areas—Targeted versus Broad Approaches suggested that, rather than redline identification of underserved areas and themselves, mortgage lenders may rely on understanding the problems of mortgage HUD’s definition of metropolitan underserved areas is a targeted neighborhood private mortgage insurers to screen access in low-income and minority definition, rather than a broad definition that applications from minority neighborhoods. neighborhoods. William Lang and Leonard would encompass entire cities. It also focuses Tootell also noted that this indirect form of Nakamura argue that individual home sale on those neighborhoods experiencing the redlining would increase the price paid by transactions generate information which reduce lenders’ uncertainty about property most severe credit problems, rather than applicants from minority areas that are neighborhoods experiencing only moderate approved by private mortgage insurers. values, resulting in greater availability of 29 difficulty obtaining credit. During the In a 1999 paper, Stephen Ross and Geoffrey mortgage financing. Conversely, appraisals in neighborhoods where transactions occur regulatory process leading to the 1995 rule, Tootell used the Boston Fed data base to take some argued that underserved areas under a closer at both lender redlining and the role infrequently will tend to be more imprecise, resulting in greater uncertainty to lenders this goal should be defined to include all 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. contribute to disparities in the availability of Shear, Berkovec, Dougherty, and Nothaft conducted an analysis of mortgage flows and 26 mortgage credit by neighborhood income and See Geoffrey M. B. Tootell, ‘‘Redlining in application acceptance rates in 32 Boston: Do Mortgage lenders Discriminate Against minority composition. Paul Calem found that, in low-minority tracts, higher mortgage metropolitan areas that supports a targeted Neighborhoods?’’, Questerly Journal of Economics, 33 111, November, 1996, pp. 1049d–1079; and loan approval rates were associated with definition of underserved areas. They ‘‘Discrimination, Redlining, and Private Mortgage recent sales transactions volume, consistent Insurance’’, unpublished manuscript, October 1995. with the Lang and Nakamura hypothesis.30 31 David C. Ling and Susan M. Wachter, 27 Tootell notes that both omitted variables and While this effect was not found in high- ‘‘Information Externalities and Home Mortgage the strong correlation between borrower race and minority tracts, he concludes that Underwriting,’’ Journal of Urban Economics, neighborhood racial composition in segregated ‘‘informational returns to scale’’ contribute to Volume 44, 1998, pp. 317–332. 32 cities have made it difficult for previous studies to disparities in the availability of mortgage Robert B. Avery, Patricia E. Beeson, and Mark distinguish the impacts of geographic redlining S. Sniderman, ‘‘Neighborhood Information and from the effects of individual borrower credit between low-minority and high- Home Mortgage Lending,’’ Journal of Urban discrimination. He can unravel these effects Economics, Volume 45, 1999, pp. 287–310. because he includes a direct measure of credit 29 William W. Lang and Leonard I. Nakamura, ‘‘A 33 William Shear, James Berkovec, Ann history and because over half of minority applicants Model of Redlining,’’ Journal of Urban Economics, Dougherty, and Frank Nothaft, ‘‘Unmet Housing in the Boston Fed data base applied for mortgages, Volume 33, 1993, pp. 223–234. Needs: The Role of Mortgage Markets,’’ Journal of in predominately white areas. 30 Paul S. Calem, ‘‘Mortgage Credit Availability in Housing Economics, Volume 4 , 1996, pp. 291–306. 28 Stephen L. Ross and Geoffrey M. B. Tootell, Low- and Moderate-Income Minority These researchers regressed the number of mortgage ‘‘Redlining, the Community Reinvestment Act, and Neighborhoods: Are Information Externalities originations per 100 properties in the census tract Private Mortgage Insurance’’, unpublished Critical?’’ Journal of Real Estate Finance and on several independent variables that were manuscript, March 1999. Economics, Volume 13, 1996, pp. 71–89. Continued

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00159 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24386 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

found: (a) Low-income census tracts and may be warranted. They also found that the origination rates—once these characteristics tracts with high concentrations of African median income of the census tract had strong are controlled for, other influences such as American and Hispanic families had lower effects on both application and denial rates central city location play only a minor role rates of mortgage applications, originations, for purchase and refinance loans, even after in explaining disparities in mortgage lending. and acceptance rates; and (b) once census other variables were accounted for. Avery, HUD recognizes that the mortgage tract influences were accounted for, central Beeson and Sniderman concluded that a origination and denial rates forming the basis city location had only a minimal effect on tract-level definition is a more effective way for the research mentioned in the preceding credit flows. These authors recognized that it to define underserved areas than using the paragraph, as well as for HUD’s definition of is difficult to interpret their estimated list of OMB-designated central cities as a underserved areas, are the result of the minority effects—the effects may indicate proxy. interaction of individual risk, demand and lender discrimination, supply and demand c. Conclusions From the Economics supply factors that analysts have yet to fully effects not included in their model but Literature About Urban Underserved Areas disentangle and interpret. The need correlated with minority status, or some continues for further research addressing this combination of these factors. Still, they The implications of studies by HUD and others for defining underserved areas can be problem. conclude that income and minority status are summarized briefly. First, the existence of better indicators of areas with special needs 3. Characteristics of HUD’s Underserved large geographic disparities in mortgage than central city location. Areas credit is well documented. Low-income and Avery, Beeson, and Sniderman of the high-minority neighborhoods receive a. Credit Characteristics Federal Reserve Bank of Cleveland substantially less credit than other specifically addressed the issue of HMDA data provide information on the neighborhoods and fit the definition of being underserved areas in the context of the GSE disposition of mortgage loan applications underserved by the nation’s credit markets. legislation.34 Their study examined (originated, approved but not accepted by the Second, researchers are testing models that variations in application rates and denial borrower, denied, withdrawn, or not more fully account for the various risk, completed) in metropolitan areas. HMDA rates for all individuals and census tracts demand, and supply factors that determine data include the census tract location of the included in the 1990 and 1991 HMDA data the flow of credit to urban neighborhoods. base. These authors found that the individual The studies by Holmes and Horvitz, Schill property being financed and the race and applicant’s race exerts a strong influence on and Wachter, and Tootell are examples of income of the loan applicant(s). Therefore, mortgage application and denial rates. this research. Their attempts to test the this is a rich data base for analyzing mortgage African American applicants, in particular, redlining hypothesis show the analytical activity in urban neighborhoods. HUD’s had unexplainably high denial rates. Once insights that can be gained by more rigorous analysis using HMDA data for 2002 shows individual applicant and other neighborhood modeling of this issue. However, the fact that that high-minority and low-income census characteristics were controlled for, overall urban areas are highly segregated means that tracts have both relatively high loan denial rates for purchase and refinance loans the various loan, applicant, and application denial rates and relatively low were only slightly higher in minority census neighborhood characteristics currently being loan origination rates. tracts than non-minority census tracts. For used to explain credit flows are often highly Table B.2 presents mortgage denial and white applicants, on the other hand, denial correlated with each other, which makes it origination rates by the minority composition rates were significantly higher in minority difficult to reach definitive conclusions about and median income of census tracts in tracts. That is, minorities had higher denial the relative importance of any single variable metropolitan areas. Two patterns are clear: rates wherever they attempted to borrow, but such as neighborhood racial composition. • Census tracts with higher percentages of whites faced higher denials when they Thus, their results are inconclusive, and the minority residents have higher mortgage attempt to borrow in minority need continues for further research on the denial rates and lower mortgage origination neighborhoods. In addition, Avery et al. underlying determinants of geographic rates than all-white or substantially-white found that home improvement loans had disparities in mortgage lending.35 tracts. For example, in 2002 the denial rate significantly higher denial rates in minority Finally, much research strongly supports a for census tracts that are over 90 percent neighborhoods. Given the very strong effect targeted definition of underserved areas. minority (20.2 percent) was 2.4 times that for of the individual applicant’s race on denial Studies by Shear, et al. and Avery, Beeson, census tracts with less than 10 percent rates, the authors noted that since minorities and Sniderman conclude that characteristics minority (8.4 percent). tend to live in segregated communities, a of both the applicant and the neighborhood • Census tracts with lower incomes have policy of targeting minority neighborhoods where the property is located are the major higher denial rates and lower origination determinants of mortgage denials and rates than higher income tracts. For example, intended to account for some of the demand and in 2002 mortgage denial rates declined from supply (i.e., credit risk) influences at the census 35 Methodological and econometric challenges 22.7 percent to 6.6 percent as tract income tract level. See also Susan Wharton Gates, ‘‘Defining that researchers will have to deal with are discussed increased from less than 40 percent of area the Underserved,’’ Secondary Mortgage Markets, 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 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00160 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24387

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00161 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.044 24388 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Table B.3 illustrates the interaction area median) group had a denial rate of 6.5 origination rate of only 13.1 loans per 100 between tract minority composition and tract percent and an origination rate of 22.7 loans owner occupants. The other groupings fall income by aggregating the data in Table B.2 per 100 owner occupants in 2002. The high- between these two extremes. into nine minority and income combinations. minority (over 50 percent), low-income BILLING CODE 4210–27–P The low-minority (less than 30 percent (under 90 percent of area median) group had minority), high-income (over 120 percent of a denial rate of 18.3 percent and an

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00162 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24389

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00163 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.045 24390 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

The advantages of HUD’s underserved area definition does not include high-income mortgage credit. The 15.8 percent denial rate definition can be seen by examining the (over 120 percent of area median) census in these neighborhoods in 2002 was almost minority-income combinations highlighted in tracts even if they meet the minority twice the 8.0 percent denial rate in the Table B.3. The sharp differences in denial threshold. The average denial rate (9.9 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 (14.0 include these ‘‘remaining’’ portions of cities. areas that have significantly less success in percent) in underserved areas as defined by Figure B.1 shows that these areas, which receiving mortgage credit. In 2002 HUD. account for approximately 36 percent of the underserved areas had over one and a half Figure B.1 compares underserved and population in central cities, appear to be well times the average denial rate of served areas served areas within central cities and served by the mortgage market. As a whole, (14.0 percent versus 8.9 percent) and three- suburbs. First, Figure B.1 shows that HUD’s they are not experiencing problems obtaining fourths the average origination rate per 100 definition targets central city neighborhoods mortgage credit. owner occupants (16.0 versus 21.4). HUD’s that are experiencing problems obtaining BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00164 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24391

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00165 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.046 24392 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Second, Figure B.1 shows that HUD’s population, are included in HUD’s definition times the poverty rate (5.7 percent) in served definition also targets underserved census of other underserved areas. census tracts. The unemployment rate and tracts in the suburbs as well as in central b. Socioeconomic Characteristics the high-school dropout rate are also higher cities. The average denial rate in underserved in underserved areas. In addition, there are The targeted nature of HUD’s definition suburban areas (13.7 percent) is 1.7 times nearly three times more female-headed that in the remaining served areas of the can be seen from the data presented in Table households with children in underserved suburbs (8.0 percent), and is almost as large B.4, which show that families living in tracts areas (30.0 percent) than in served areas (13.2 as the average denial rate (15.8 percent) in within metropolitan areas that are 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. which account for 34 percent of the suburban census tracts is 18.5 percent, or over three BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00166 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24393

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00167 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.047 24394 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

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 nonmetropolitan income. This is based on geographic goals were designed. The Housing, Economic, and Demographic HUD’s analysis of 1990 census data, which geographic goals, then, are meant to target Conditions in 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 nonmetropolitan county is classified as an census geography, underserved counties addresses the basic question of whether and underserved area if median income of account for 57 percent (8,091 of 14,419) of the extent to which HUD’s definition of families in the county does not exceed 95 the census tracts and 54 percent of the underservice in nonmetropolitan areas percent of the greater of state population in rural areas. By comparison, the effectively targets areas that encompass large nonmetropolitan or national nonmetropolitan definition of metropolitan underserved areas 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. socioeconomic conditions of underserved not exceed 120 percent of the greater of state The purchasing of loans from underserved and served nonmetropolitan areas based on nonmetropolitan or national nonmetropolitan areas by the GSEs is intended to induce HUD’s definition applied at the county level median income. For nonmetropolitan areas greater homeownership among moderate, using Census 2000 data. (A later section the median income component of the low, very low income, and poor families and considers the effects of applying the underserved definition is broader than that minorities. For various reasons, including definition of the census tract level.) Several used for metropolitan areas. While tract creditworthiness and lending discrimination, variables are used to describe area income is compared with area income for these groups experience greater difficulty in demographic and socioeconomic conditions. metropolitan areas, in rural counties income securing loans under fair and reasonable BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00168 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24395

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00169 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.048 24396 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

On the national level, a few key results • While nearly all housing in served and percent in 1997, 27.0 percent in 1998, 26.8 show that the 1995 definition of underservice underserved areas have complete plumbing percent in 1999, and 31.0 percent in 2000; captures a potentially disadvantaged segment and kitchens, the percentage of units with and Freddie Mac’s performance was 26.3 of the population. In examining the minority incomplete facilities in underserved is twice percent in 1997, 26.1 percent in 1998, 27.5 composition, one can see that the percentage the percentage in served areas. percent in 1999, and 29.2 percent in 2000. of African Americans, Hispanics/Latinos, and • Crowded units are a small share of all • In the October 2000 rule, the total minority population is higher in housing in nonmetropolitan areas, but the underserved areas goal was set at 31 percent underserved nonmetropolitan areas as rate is higher for underserved areas: 4.3 vs. for 2001–03. As of January 1, 2001, several compared to served nonmetropolitan areas. 2.3 percent. changes in counting requirements came into Overall, the minority population of Mikesell 37 found using the 1995 American effect for the undeserved areas goal, as underserved areas is 25.8 percent as Housing Survey that while the rate of follows: ‘‘bonus points’’ (double credit) for compared with 9.3 percent in served areas. homeownership in nonmetropolitan areas is purchases of goal-qualifying mortgages on Other supporting results include median higher than metropolitan areas, the quality of small (5–50 unit) multifamily properties and, family income, poverty rate, unemployment housing is lower as compared to above a threshold level, mortgages on 2–4 rate, school dropout rate, and in-migration metropolitan areas. Results based on the 2000 unit owner-occupied properties; a rate. Specifically we find: Census show that the homeownership rate ‘‘temporary adjustment factor’’ (1.20 units • Median income is approximately for nonmetropolitan areas was 74 percent (73 credit, subsequently increased by Congress to $10,000 less in underserved areas than in percent without manufactured homes), and 1.35 units credit) for Freddie Mac’s served areas. This represents an average gap for metropolitan areas it was 64 percent, but purchases of goal-qualifying mortgages on of 25 percent. both metropolitan and nonmetropolitan areas large (more than 50-unit) multifamily • Poverty in underserved areas is twice the had approximately 97.5 percent of units with properties; and eligibility for purchases of rate in served areas (14.5 vs. 7.5 percent). complete plumbing and 99 percent with certain qualifying government-backed loans • Unemployment is 7.3 percent in complete kitchens. to receive goal credit. These changes are underserved areas and 5.2 percent in served explained below. Fannie Mae’s performance areas. D. Factor 3: Previous Performance and Effort was 32.6 percent in 2001 and 32.8 percent in • The school dropout rate is 28.1 percent of the GSEs in Connection With the Central 2002, and Freddie Mac’s performance was in underserved areas and 18.7 percent in Cities, Rural Areas and Other Underserved 31.7 percent in 2001 and 31.9 percent in served areas. Areas Goal 2002, thus both GSEs surpassed this higher goal in both years. This section discusses the • Migration into underserved areas is Section D.1 reports the past performance of October 2000 counting rule changes in detail somewhat lower than in served areas: 7.4 vs. each GSE with regard to the Underserved below, and provides data on what goal 8.0 percent. Areas Housing Goal. Section D.2 then performance would have been in 2001–02 Table B.5 also includes data on examines the role that the GSEs are playing without these changes.39 homeownership rates, housing affordability, in funding single-family mortgages in housing quality, and overcrowding. On underserved urban neighborhoods based on a. Performance on the Underserved Areas several of these dimensions, housing HUD’s analysis of GSE and HMDA data. That Housing Goal in 1996–2002 conditions and needs in underserved areas section also discusses an underserved area HUD’s December 1995 rule specified that are not substantially worse than in served subgoal for home purchase loans. Section D.3 in 1996 at least 21 percent of the number of areas. Although housing quality and concludes this section with an analysis of the units financed by each of the GSEs that were crowding appear to be marginally worse in GSEs’ purchases in rural (nonmetropolitan) eligible to count toward the Underserved underserved areas, homeownership in the areas. Areas Goal should qualify as units in two areas is about the same and owning a The increased coverage of the Underserved properties located in underserved areas, and home actually appears to be more affordable Areas Housing goal due to switching to 2000 at least 24 percent should qualify in 1997– in underserved areas than in served areas. census geography is discussed throughout 2000. HUD’s October 2000 rule made various Specific findings include the following: this section. changes in the goal counting rules, as • Homeownership is slightly higher in discussed below, and increased the underserved than in served nonmetropolitan 1. Past Performance of the GSEs Underserved Areas Goal to 31 percent for counties: 74.3 percent vs. 73.7 percent. This section discusses each GSE’s 2001–03. Removing manufactured homes lowers performance under the Underserved Areas Table B.6 shows performance on the ownership rates slightly, because ownership Housing Goal over the 1996–2002 period.38 underserved areas goal over the 1996–2002 of such homes is relatively high, but this As explained in Appendix A, the data period, based on HUD’s analysis. The table does not affect the basic result. presented are ’official HUD results’ which, in shows that Fannie Mae surpassed the goals • Owner-occupied and rental vacancy some cases, differ from goal performance by 7.1 percentage points and 4.8 percentage rates are both somewhat higher in reported by the GSEs in the Annual Housing points in 1996 and 1997, respectively, while underserved areas. Activities Reports (AHARs) that they submit Freddie Mac surpassed the goals by narrower • Median housing unit values are to the Department. margins, 4.0 and 2.3 percentage points. In significantly lower in underserved areas: The main finding of this section is that 1998 Fannie Mae’s performance fell by 1.8 $67,358 vs. $88,099. both Fannie Mae and Freddie Mac surpassed percentage points, while Freddie Mac’s • The value of a housing affordability the Department’s Underserved Areas Housing performance fell only slightly, by 0.2 index for owner-occupied housing is slightly Goals for each of the seven years during this percentage point. Freddie Mac showed a gain higher in underserved areas.36 On average, period. Specifically: in performance to 27.5 percent in 1999, median income is 1.83 times higher than • The goal was set at 21 percent for 1996; exceeding its previous high by 1.2 percentage income required to qualify to buy a home of Fannie Mae’s performance was 28.1 percent points. Fannie Mae’s performance in 1999 median value in underserved areas. The and Freddie Mac’s performance was 25.0 was 26.8 percent, which, for the first time, comparable factor for served areas is 1.78. percent. slightly lagged Freddie Mac’s performance in • Rental affordability is approximately the • The goal was set at 24 percent for 1997– that year. same in underserved and served areas. 2000. Fannie Mae’s performance was 28.8 BILLING CODE 4210–27–P

36 The purchase affordability index assesses the income needed to qualify for a mortgage on a unit 38 Performance for the 1993–95 period was extent to which a family with the median income with the median value; and an index number discussed in the October 2000 rule. of a given area would be able to afford a housing greater than 100 means that a family with the 39 To separate out the effects of changes in unit that carries the median purchase price of that median income has 20 percent more than the level counting rules that took effect in 2001, this section area. For example, a purchase affordability index of income needed to qualify for a mortgage on a unit number less than 100 means that a family with the with the median value. The rental affordability also compares performance in 2001 to estimated median income would not qualify for a mortgage on index is similarly constructed. performance in 2000 if the 2001 counting rules had a unit with the median value; a purchase 37 J.J. Mikesell, ‘‘Housing Problems across Types been in effect in that year. affordability index equal to 100 means that a family of rural Households’’, Rural Conditions and Trends, with the median income has exactly the level of Volume 9, Number 2, pp. 97–101, 1999.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00170 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24397

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00171 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.049 24398 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Both GSEs exhibited sharp gains in goal difference in the counting rules applicable to 2001–03.41 This factor did not apply to units performance in 2000—Fannie Mae’s the two GSEs that was enacted by Congress; in large multifamily properties in performance increased by 4.2 percentage if the same counting rules were applied to underserved areas whose mortgages were points, to a record level of 31.0 percent, both GSEs, Fannie Mae’s performance would financed by Fannie Mae during this period. while Freddie Mac’s performance increased have exceeded Freddie Mac’s performance by Purchases of certain government-backed somewhat less, by 1.7 percentage points, an even greater margin, and in fact Freddie loans. Prior to 2001, purchases of which also led to a record level of 29.2 Mac would have just attained the goal, at government-backed loans were not taken into percent. Fannie Mae’s performance was 32.6 31.0 percent, in 2002, and fallen short of the account in determining performance on the percent in 2001 and 32.8 percent in 2002; goal in 2001. GSEs’ low- and moderate-income and Freddie Mac’s performance was 31.7 percent b. Changes in the Goal Counting Rules for underserved area housing goals. As discussed in 2001 and 31.9 percent in 2002. However, 2001–03 in Appendix A, the 2000 rule established as discussed below, using consistent eligibility for FHA-insured home equity accounting rules for 2000–02, under one Several changes in the counting rules underlying the calculation of underserved conversion mortgages (HECMs) for method each GSE’s performance in 2001–02 mortgagors in underserved areas, purchases was below its performance in 2000. areas goal performance took effect beginning of mortgages on properties on tribal lands The official figures for underserved areas in 2001. These also applied to the low- and insured under FHA’s Section 248 program or goal performance presented above for 1996– moderate-income goal and are discussed in Appendix A; only brief summaries of those HUD’s Section 184 program, and purchases 2002 are the same as the corresponding 40 figures presented by Freddie Mac in its changes are given here: of mortgages under the Rural Housing Annual Housing Activity Reports to HUD for Bonus points for multifamily and single- Service’s Single Family Housing Guaranteed every year except 1999 and 2002, when there family rental properties. Each qualifying unit Loan Program to count toward the was a difference of 0.1 percentage point. The in a small multifamily property counted as underserved area goal. two units in the numerator in calculating official figures are the same as those c. Effects of Changes in the Counting Rules performance on all of the goals for 2001–03. presented by Fannie Mae in most years, and on Goal Performance And, above a threshold equal to 60 percent differ by 0.1–0.2 percentage point in the Because of the changes in the underserved other years, reflecting minor differences in of the average number of qualifying rental areas goal counting rules that took effect in the application of counting rules. units financed in owner-occupied properties 2001, direct comparisons between official Fannie Mae’s performance on the over the preceding five years, each unit in a goal performance in 2000 and 2001–02 are underserved areas goal surpassed Freddie 2–4 unit owner-occupied property also somewhat of an ‘‘apples-to-oranges Mac’s in every year through 1998. This counted as two units in the numerator in comparison.’’ For this reason, the Department pattern was reversed in 1999, as Freddie Mac calculating goal performance. surpassed Fannie Mae in goal performance Freddie Mac’s Temporary Adjustment has calculated what performance would have Factor. Freddie Mac received a ‘‘Temporary for the first time, though by only 0.7 been in 2000 under the 2001–03 rules; this Adjustment Factor’’ of 1.35 units of credit for percentage point. This improved relative may compared with official performance in each qualifying unit financed in ‘‘large’’ performance of Freddie Mac was due to its 2001–02—an ‘‘apples-to-apples comparison.’’ multifamily properties (i.e., those with 51 or increased purchases of multifamily loans, as HUD has also calculated what performance more units) in the numerator in calculating it re-entered that market, and to increases in would have been in 2001–02 under the 1996– its performance on the housing goals for the goal-qualifying shares of its single-family 2000 rules; this may be compared with mortgage purchases. However, Fannie Mae’s official performance in 2000—an ‘‘oranges-to- 40 performance once again exceeded Freddie Unlike the low- and moderate-income and oranges comparison.’’ These comparisons are Mac’s performance in 2000, 31.0 percent to special affordable goals, there is no exclusion of presented in Table B.7a. units from the denominator for units with missing BILLING CODE 4210–27–P 29.2 percent. Fannie Mae’s official information about the area in which a property is performance also exceeded Freddie Mac’s located. That is, such units are counted in the official performance in 2001–02, despite the denominator, but not in the numerator, in 41 See Congressional Record, December 15, 2000, fact that Freddie Mac benefited from a determining undeserved area goal performance. pp. H12295–96.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00172 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24399

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00173 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.050 24400 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Specifically, Table B.7a shows Freddie Mac. The largest impact of the extent in 2001 than in 2000. That is, 56 performance under the underserved areas counting rule changes on Freddie Mac’s goal percent of Fannie Mae’s small multifamily goal in three ways. Baseline A represents the performance was due to bonus points for units qualified for the underserved areas goal counting rules in effect in 1996–2000. purchases of mortgages on small multifamily in 2000, but this rose to 64 percent in 2001. Baseline B incorporates the one minor properties; this added 1.3 percentage points Freddie Mac financed 50,211 units in small technical change in counting rules pertaining to goal performance in 2001 and 1.0 multifamily properties in 2001 that were to the underserved areas goal’’ eligibility of percentage points in 2002, as shown in Table eligible for the underserved areas goal, an certain government-backed loans for goals B.7. The application of the temporary increase of more than 1500 percent from the credit. Baseline C incorporates in addition to adjustment factor for purchases of mortgages a small base of 2,985 units financed in 2000. that technical change the bonus points and, on large multifamily properties enacted by Small multifamily properties also accounted for Freddie Mac, the temporary adjustment Congress added 0.9 percentage points to goal for a significantly greater share of Freddie factor. Baseline B corresponds to the performance in 2002. Bonus points for Mac’s multifamily business in 2001—16.1 counting approach proposed in this rule to purchase of mortgages on owner-occupied 2– percent of total multifamily units financed, take effect in 2005. Boldface figures under 4 unit rental properties also added 1.1 up from 1.8 percent in 2000. Baseline A for 1999–2000 and under Baseline percentage points to performance. Credit for Within the small multifamily market, there C for 2001–02 indicate official goal purchases of qualifying government-backed was some evidence that Freddie Mac targeted percentages based on the counting rules in loans played a minor role in determining properties in underserved areas to a greater effect in those years’e.g., for Freddie Mac, Freddie Mac’s goal performance. extent in 2001 than in 2000. That is, 61 27.5 percent in 1999, 29.2 percent in 2000, Fannie Mae. The temporary adjustment percent of Freddie Mac’s small multifamily 31.7 percent in 2001, and 31.8 percent in factor which applied to Freddie Mac’s goal units qualified for the underserved areas goal 2002. performance did not apply to Fannie Mae, in 2000; this rose to 86 percent in 2001. Performance on the Underserved Areas thus counting rule changes had less impact Bonus points for single-family rental Goal under 1996–2000 Counting Rules Plus on its performance than on Freddie Mac’s properties. Above a threshold, each unit Technical Changes. If the ‘‘Baseline B’’ performance in 2002. The largest impact of financed in a 2–4 unit property with at least counting approach had been in effect in the counting rule changes on Fannie Mae’s one owner-occupied unit (referred to as 2000–02 and the GSEs’ had purchased the goal performance was due to the application ‘‘OO24s’’ below) that qualified for any of the same mortgages that they actually did housing goals was counted as two units in purchase in those years, Fannie Mae would of bonus points for purchases of mortgages on owner-occupied 2–4 unit rental the denominator (and one unit in the have just matched the underserved areas goal numerator) in calculating goal performance in 2000 and fallen short in 2001–02, while properties, which added 1.8 percentage points to performance, and for purchases of for that goal in 2001–03. The threshold was Freddie Mac would have fallen short of the equal to 60 percent of the average number of goal in all three years, 2000–02. Specifically, mortgages on small multifamily properties, such qualifying units over the previous five Fannie Mae’s performance would have been which added 0.8 percentage point to years. For example, Fannie Mae financed an 31.0 percent in 2000, 30.4 percent in 2001, performance. Credit for purchases of average of 47,100 underserved area units in and 30.1 percent in 2002. Freddie Mac’s qualifying government-backed loans played a these types of properties between 1996 and performance would have been 29.2 percent minor role in determining Fannie Mae’s goal 2000, and 105,946 such units in 2001. Thus in 2000, 28.2 percent in 2001, and 28.4 performance. in 2001 Fannie Mae received 77,688 bonus percent in 2002. d. Bonus Point Incentives for the GSEs’ points in this area in 2001—that is, 105,946 Performance on the Underserved Areas Purchases in Underserved Areas Goal under 2001–2003 Counting Rules. If the minus 60 percent of 47,100. So 183,629 units The Department established ‘‘bonus 2001–03 counting rules had been in effect in were entered in the numerator for these points’’ for 2001–03 to encourage the GSEs to 2000–02 and the GSEs had purchased the properties in calculating underserved area step up their activity in two segments of the same mortgages that they actually did goal performance. mortgage market’the small (5–50 unit) purchase in those years (i.e., abstracting from Single-family rental bonus points thus multifamily mortgage market, and the market any behavioral effects of ‘‘bonus points,’’ for encouraged the GSEs to play a larger role in for mortgages on 2–4 unit properties where example), both GSEs would have surpassed this market, and also to purchase mortgages 1 unit is owner-occupied and 1–3 units are the underserved areas goal in all three years, on such properties in which large shares of and both GSEs’ performance figures would occupied by renters. the units qualify for the housing goals. As for have increased from 2000 to 2002. Bonus points for small multifamily small multifamily bonus points, some Specifically, Fannie Mae’s ‘‘Baseline C’’ properties. Each unit financed in a small evidence on the effects of such bonus points performance would have been 32.3 percent multifamily property that qualified for any of on the GSEs’ operations may be gleaned from in 2000, 32.6 percent in 2001, and 32.8 the housing goals was counted as two units the data provided to HUD by the GSEs for percent in 2002. Freddie Mac’s performance in the denominator (and one unit in the 2001. would have been 31.4 percent in 2000, 31.7 numerator) in calculating goal performance Fannie Mae financed 177,872 units in percent in 2001, and 31.8 percent in 2002. for that goal. OO24s in 2001 that were eligible for the Measured on this consistent basis, then, Fannie Mae financed 37,389 units in small underserved areas goal, an increase of 116 Fannie Mae’s performance increased by 0.3 multifamily properties in 2001 that were percent from the 82,464 units financed in percentage point in 2001 and 0.2 percentage eligible for the underserved areas goal, an 2000. However, Fannie Mae’s total single- point in 2002, and Freddie Mac’s increase of more than 400 percent from the family business increased at approximately performance increased by 0.4 percentage 7,196 units financed in 2000. As explained the same rate as its OO24 business in 2001, point in 2001 and 0.2 percentage point in in Appendix A, small multifamily properties thus the share of its business accounted for 2002. These increases were the effect of also accounted for a greater share of Fannie by OO24s was the same in 2001 as in 2000— increased activity in mortgages eligible to Mae’s multifamily business in 2001—7.4 4 percent. receive bonus points between 2000 and percent of total multifamily units financed, Within the OO24 market, there was no 2001–02. up from 2.5 percent in 2000. However, HUD’s evidence that Fannie Mae targeted affordable Details of Effects of Changes in Counting Housing Goals 2000 Final Rule cited a properties to a greater extent in 2001 than in Rules on Goal Performance in 2001. As Residential Finance Survey finding that 2000. That is, approximately 60 percent of discussed above, counting rule changes that small multifamily properties account for 37 Fannie Mae’s OO24 units qualified for the took effect in 2001 had significant impacts on percent of total units in multifamily underserved area goal in both 2000 and 2001. the performance of both GSEs on the mortgaged properties, thus Fannie Mae is Freddie Mac financed 96,983 units in underserved areas goal in that year—2.4 still less active in this market than in the OO24s in 2001 that were eligible for the percentage points for Fannie Mae, and 3.5 market for large multifamily properties.42 underserved areas goal, an increase of 91 percentage points for Freddie Mac. This Within the small multifamily market, there percent from the 50,868 units financed in section breaks down the effects of these was some evidence that Fannie Mae targeted 2000. However, Freddie Mac’s total single- changes on goal performance for both GSEs; properties in underserved areas to a greater family business increased at approximately results are shown in Table B.7a along with the same rate as its OO24 business in 2001, figures for other years. 42 65 FR 65141 & n. 145 (2000). thus the share of its business accounted for

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00174 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24401

by OO24s was the same in 2001 as in 2000— HUD’s proposal, the basis for the Analysis. HUD used 2000 census data to 3 percent. determination outside of MSAs will change generate underserved area designations for As for Fannie Mae, within the OO24 from counties to census tracts beginning in census tracts as defined for the 2000 census market there was no evidence that Freddie 2005. with 2003 MSA designations. Because Fannie Mac targeted affordable properties to a 2005 Procedure. Relative to the above Mae and Freddie Mac geocoded the greater extent in 2001 than in 2000. That is, procedure, Underserved Areas Housing Goals mortgages they purchased prior to 2003 60 percent of Fannie Mae’s OO24 units performance percentages for loans purchased based on census tract boundaries as qualified for the underserved areas goal in by the GSEs in and after 2005 will be affected established for the 1990 census, GSE both 2000 and 2001. by three factors. First, 2000 census data on mortgages purchased prior to 2003 can be e. Effects of 2000 Census on Scoring of Loans median incomes and minority populations directly identified as being from a served or Toward the Underserved Areas Housing Goal replace 1990 census data. Second, the Office underserved area only where the property is located in a 1990-defined census tract whose Background. Scoring of housing units of Management and Budget in June, 2003, area consists entirely of whole 2000-defined under the Underserved Areas Housing Goal respecified MSA boundaries based on census tracts, or portions of such tracts, is based on decennial census data used to analysis of 2000 census data. Third, the which are all designated either as served or identify underserved areas, as follows: For Department’s proposed re-specification of the properties in MSAs scoring is based on the Underserved Areas goal in terms of census as underserved. In the situation where the area of a 1990-defined census tract includes median income of the census tract where the tracts rather than counties in non- whole 2000-defined census tracts, or portions property is located, the median income of the metropolitan areas will come into effect.45 of such tracts, some of which are served and MSA, and the percentage minority Thus, for properties located outside of MSAs some underserved, HUD calculated an population in the census tract where the the basis of determination for non- ‘‘underservice factor’’ defined as the property is located. For properties located metropolitan areas will be changed for underserved percentage of the 1990-defined outside of MSAs scoring is based on the properties located outside of MSAs to: The median income of the county, the median tract’s population, based on population data median income of the census tract where the 46 income of the non-metropolitan portion of from the 2000 census. These factors were property is located; the median income of the used in estimating underservice percentages the State in which the property is located or non-metropolitan portion of the State in of the non-metropolitan portion of the United for aggregated GSE purchases in and before which the property is located or of the non- 2002 based on the 2000 census. States, whichever has the larger median metropolitan portion of the United States, income, and the percentage minority The resulting underserved areas file was whichever is larger; and the percentage used to re-score loans purchased by the GSEs population in the county where the property minority population in the census tract is located. Thus, scoring loans under the between 1999 and 2002, and was used where the property is located. Underserved Areas Housing Goal requires further in estimating the share of loans decennial census data on median incomes for originated in metropolitan areas that would metropolitan census tracts, MSAs, non- in establishing the metropolitan or non- be eligible to score toward the Underserved metropolitan counties, the non-metropolitan metropolitan location of a property. The MSA Areas Housing Goal, from HMDA data. The portions of States, and the non-metropolitan definitions established by the Office of Management results of the retrospective GSE analysis are and Budget (OMB) in June, 2003 defined MSAs in portion of the United States. The provided in Table B.7b. The results of the New England in terms of counties. GSE–HMDA comparative analysis are determination has been based on 1990 census 44 The procedure used to generate estimated rents presented in the next section. data through 2004, and beginning in 2005 in connection with Low- and Moderate Income and 43 44 will be based on 2000 census data. Under Special Affordable Housing Goals, as mentioned in BILLING CODE 4210–27–P Appendixes A and C, uses similar data series. 43 In New England, MSAs were defined through 45 HUD has deferred application of the 2000 46 8,717 tracts included both served and mid-2003 in terms of Towns rather than Counties, census data and 2003 MSA designations to 2005, underserved area, out of a total of 61,493 tracts that and the portion of a New England county outside pending completion of the present rulemaking could be classified as served or underserved or of any MSA is regarded as equivalent to a county process. assigned an underservice factor.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00175 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24402 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00176 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.051 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24403

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 2002. To rules in place in 2001–2002 (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 discussed subsection 2.a. In subsection 2.b., both GSEs improved their performance, they 2. GSEs’ Mortgage Purchases in Metropolitan 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:

Freddie Mac Fannie Mae Market Year ≤ ≤ (w/o B&C) (percent ) (percent ) (percent≤)

1996–2002 ...... 21.7 23.5 25.4 1999–2002 ...... 22.9 24.0 25.8 2001–2002 ...... 24.1 25.6 25.9

Between 1996 and 2002, 21.7 percent of respectively. However, like Freddie Mac, based on 2000 census tract geography. For Freddie Mac’s purchases financed properties Fannie Mae’s longer-term performance (since the years 1999, 2000, 2001, and 2002, HUD in underserved neighborhoods, compared 1993 or 1996) as well as its recent average used the apportionment technique described with 23.5 percent of Fannie Mae’s purchases performance (1999 to 2002) has consistently above involving ‘‘underservice factors’’ to re- and 25.4 percent of home purchase loans been below market levels. Over the past four allocate 1990-based GSE and HMDA data originated in the conventional conforming years, Fannie Mae performed at 93 percent of into census tracts as defined by the 2000 market (excluding B&C loans). Thus, Freddie the market (24.0 percent for Fannie Mae Census. Mac performed at only 85 percent of the compared with 25.8 percent for the market). The main results are provided in Table B.8, market (21.7 divided by 25.4), while Fannie Still, it is encouraging that Fannie Mae which compares the GSEs to the market Mae performed at 93 percent of the market. significantly improved its performance and using both the 1990 Census geography and Freddie Mac’s recent performance has been closed its gap with the market during the first the 2000 Census geography. Switching to the slightly closer to the market. Over the past two years of HUD’s higher housing goal 2000-based tracts increases the underserved four years (1999 to 2002), Freddie Mac levels. area share of market originations by nearly performed at 89 percent of the market (22.9 Market Comparisons Based on 2000 six percentage points. Between 1999 and percent for Freddie Mac compared with 25.8 Census Geography. As explained in Section 2002, 31.5 percent of home purchase percent for the market), and in 2001 and A.2 of this appendix, HUD will be defining mortgages (without B&C loans) were 2002, the first two years under HUD’s higher underserved areas based on 2000 Census data originated in underserved tracts based on housing goal targets, at 93 percent of the and re-specified metropolitan area market (24.1 percent compared with 25.9 boundaries beginning in 2005, the first year 2000 geography, compared with 25.8 percent percent). (See Tables A.13 to A.16 in covered by the proposed rule. The number of based on 1990 geography—a differential of Appendix A for complete data going back to census tracts in metropolitan areas covered 5.7 percentage points. As also shown in 1993.) by HUD’s definition will increase from Table B.8, the underserved areas share of Fannie Mae has funded underserved areas 21,587 tracts (based on 1990 Census) to Fannie Mae’s purchases rises by 5.5 at a higher level than Freddie Mac, as 26,959 tracts (based on 2000 Census and new percentage points, and the underserved areas indicated above. And during 2001 and 2002, OMB metropolitan area specifications). The share of Freddie Mac’s purchases rises by 5.4 Fannie Mae average performance was only increase in the number of tracts defined as percentage points. Thus, the conclusions slightly below the market. The share of underserved means that both GSE reported above and in Appendix A about the Fannie Mae’s purchases going to underserved performance and the market estimates will be GSEs’ performance relative to the market areas was 24.4 percent in 2001 to 26.7 higher than reported above. This section about remain the same when the analysis is percent in 2002, compared with market provides an analysis of the performance of conducted based on 2000 Census geography. levels of 25.2 percent and 26.4 percent, the GSEs in the single-family-owner market BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00177 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24404 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00178 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.052 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24405

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 ...... 26.1 27.0 31.4 2000 ...... 27.4 29.9 32.9 2001 ...... 27.4 30.8 31.6 2002 ...... 31.7 32.3 32.3 1999–2002 (average) ...... 28.3 29.5 31.5 1996–2002 (estimate) ...... 27.1 29.0 31.1

Between 1999 and 2002, 28.3 percent of they will be leading the primary market by its performance of 32.3 percent in 2002; and Freddie Mac’s purchases and 29.5 percent of about 1.5 (3.5) percentage points, based on Freddie Mac by 3.4 (5.4) percentage points Fannie Mae’s purchases financed properties historical data. This home purchase subgoal over its average performance of 29.6 percent in underserved neighborhoods, compared will encourage the GSEs to provide in 2001 and 2002, and by 1.3 (3.3) percentage with 31.5 percent home purchase loans additional credit and capital to urban points over its performance of 31.7 percent originated in the conventional conforming neighborhoods that historically have not in 2002. Loans in the B&C portion of the market (excluding B&C loans). Thus, Freddie been adequately served by the mortgage subprime market are excluded from the Mac performed at 90 percent of the market industry—but in the future may be the very market average of 31.5 percent for 1999– level, while Fannie Mae performed at 94 neighborhoods where the growing population 2001. percent of the market level—both results of immigrants and minorities choose to live. The subgoal applies only to the GSEs’ similar to those reported above for As detailed in Section I.5 of this appendix, purchases in metropolitan areas because the underserved areas based on 1990 Census there are four specific reasons for HMDA-based market benchmark is only geography. The 2000-based results also show establishing this subgoal: (1) The GSEs have available for metropolitan areas. HMDA data that Fannie Mae has improved its the expertise, resources, and ability to lead for non-metropolitan counties are not reliable performance and matched the primary the single-family-owner market, which is enough to serve as a market benchmark. The market in funding underserved areas during their ‘‘bread and butter’’ business; (2) the Department is also setting home purchase 2002. The share of Fannie Mae’s purchases GSEs have been lagging the primary market subgoals for the other two goals-qualifying going to underserved areas increased from in underserved areas, not leading it; (3) the categories, as explained in Appendices A and 25.7 in 1999 to 32.3 percent in 2002, which GSEs can help reduce troublesome C. placed it at the market level of 32.3 percent. neighborhood disparities in access to However, the 2000-based results show that, mortgage credit; and (4) there are ample b. Characteristics of GSEs’ Purchases of like Freddie Mac, Fannie Mae’s longer-term opportunities for the GSEs to expand their Mortgages on Properties in Metropolitan performance (since 1996) as well as its recent purchases in low-income and high-minority Underserved Areas average performance (1999 to 2001) have neighborhoods. Sections E.9 and G of Several characteristics of loans purchased consistently been below market levels. (Note Appendix A provide additional information in 2002 by the GSEs in metropolitan that the 1996–2002 averages reported above on the opportunities for an enhanced GSE underserved areas are presented in Table B.9. are estimated by adding the following 2000- role in underserved area segments of the As shown, borrowers in underserved areas Census versus 1990-Census differentials home purchase market and on the ability of are more likely than borrowers in served calculated for 1999–2002: 5.4 percentage the GSEs to lead that market. areas to be first-time homebuyers, all female, points for Freddie Mac, 5.5 for Fannie Mae, As discussed above, underserved areas all male and younger than 30. And, as and 5.7 for the market.) accounted for an average of 31.5 percent of expected, they are more likely to have below- Underserved Area Subgoal for Home home purchase loans originated in the median income and to be members of Purchase Loans. The Department is conventional conforming market of minority groups. For example, first-time proposing to establish a subgoal of 33 percent metropolitan areas (computed over 1999– homebuyers make up 8.7 percent of the for each GSE’s acquisitions of home purchase 2002 or over 2001–2002). To reach the GSEs’ mortgage purchases in underserved loans financing single-family-owner proposed 33-percent (35-percent) subgoal for areas and 6.1 percent of their business in properties located in the underserved census 2005 (2007–2008), both GSEs will have to served areas. In underserved areas, 55.1 tracts of metropolitan areas for 2005, with improve their performance—Fannie Mae by percent of borrowers had incomes below the this proposed subgoal rising to 34 percent for 1.9 (3.9) percentage points over its average area median, compared with 36.7 percent of 2006 and 35 percent for 2007–2008. If the performance of 31.1 percent during 2001 and borrowers in served areas. GSEs meet this 2005 (2007–2008) subgoal, 2002, and by 0.7 (2.7) percentage points over BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00179 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24406 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00180 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.053 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24407

Minorities’ share of the GSEs’ mortgage has a role in the accessibility of credit for lenders, and subprime lenders. Industry purchases in underserved areas (33.3 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 (14.3 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 22.7 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 6.6 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 Fannie Mae and Freddie Mac have the specific characteristics of this region, would have to begin to build relationships different purchasing behavior for home leading to limited GSE activity. with the community lenders and provide Additionally, The Urban Institute prepared purchases and refinance loans in served and education/training on how to sell loans a report for HUD that investigated the factors underserved. While Fannie Mae is less likely directly to the GSEs rather than using influencing GSE activity in nonmetropolitan to purchase refinance mortgages in intermediaries. underserved area than served areas and more areas.48 The authors found that Fannie Mae like to purchase home purchase loans in and Freddie Mac have increased their a. Effects of 2000 Census Geography served areas than underserved areas, Freddie lending to nonmetropolitan areas since 1993; In order to compare served and Mac purchase the same proportion of both however, there are still weak areas in terms underserved areas, either in terms of GSE home purchase and refinance loans in served of the percentage of affordable loans being performance or socioeconomic offered.49 They also established that GSE areas as in underserved areas. characteristics, it is first necessary to update underwriting criteria was not a major barrier 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 market is often made up of locally owned and minority population data, to reflect There are numerous studies that have newly available 2000 census data. Table B.10 evaluated the impact of the GSEs’ purchases community banks, manufactured home shows the impact on 2000, 2001, and 2002 on metropolitan areas, but few address the GSE purchases. These are reported for total impact on nonmetropolitan areas; therefore, Development and Research, Volume 5, 2001, pp. GSE purchases and separately for Fannie Mae our understanding of the GSEs and the 219–264. 48 and Freddie Mac. As above, the results also nonmetropolitan markets is very limited. Jeanette Bradley, Noah Sawyer and Kenneth are shown separately for counties that change A study of the GSE market share in Temkin, Factors Influencing GSE Service to Rural Areas. the Urban Institute, prepared for U.S. classification and those that do not. This underserved counties 47 found that location Department of Housing and Urban Development, analysis is limited to nonmetropolitan areas 2002. based on both the pre- and post-June, 2003 47 49 Heather MacDonald, ‘‘Fannie Mae and Freddie Affordable loans are defined as borrowers OMB metropolitan area designations. Mac in Nonmetropolitan Housing Markets: Does earning less than 80 percent the Area Median Space Matter? ’’ Cityscape: A Journal of Policy Income. BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00181 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24408 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00182 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.054 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24409

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 census tracts; and therefore, census tracts net change disguises a somewhat larger shift minority data and 2003 metropolitan area of counties, as about 11.2 percent of currently definitions. were not an operational concept in rural underserved counties are reclassified as areas. Market data on trends in mortgage b. Characteristics of GSEs’ Purchases of lending for metropolitan areas are provided served counties and 4.6 percent of currently Mortgages on Properties in Non-metropolitan by HMDA; however, no comparable data served counties are reclassified as Underserved Areas underserved. source exists for rural mortgage markets. The Comparing underserved and served Nonmetropolitan mortgage purchases made absence of rural market data is a constraint nonmetropolitan areas in Table B.10, it is up 11.9 percent of the GSEs’ total mortgage for evaluating credit gaps in rural mortgage apparent that underserved nonmetropolitan purchases in 2002. Mortgages in underserved lending and for defining underserved areas. counties made up 39.0 percent of the GSEs’ areas make up a larger percentage of One concern is whether the broad business in nonmetropolitan areas.50 nonmetropolitan areas as a whole than do definition overlooks differences in borrower served nonmetropolitan areas, as shown by 50 Underserved areas make up about 56 percent of characteristics in served and underserved the number of counties (1,260 for counties that should be included. Table B.11 underserved (61.4%); 792 for served the census tracts in nonmetropolitan areas and 47 percent of the census tracts in metropolitan areas. compares borrower and loan characteristics (38.6%)). These relationships hold true also 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 and underserved areas. underserved (50.5%); 9.3 million for served mortgages in nonmetropolitan areas (39 percent) (49.5%)), and the population (24.9 million than in metropolitan areas (23 percent). BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00183 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24410 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00184 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.055 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24411

Fannie Mae is slightly more likely and underserved areas and (b) the financial safety • What is the impact on GSE purchasing Freddie Mac is less likely to purchase loans and soundness implications of the housing patterns if underserved areas are defined by for first-time homebuyers in underserved goals. Based on this economic analysis and tract? areas than in served areas. Mortgages to first- reviewed by the Office of Federal Housing • Applying the current criteria for time homebuyers accounted for 5.6 percent Enterprise Oversight, HUD concludes that the identifying underserved areas to tracts would of Fannie Mae’s mortgage purchases in goals raise minimal, if any, safety and result in reclassifying approximately 23 served counties, compared with 5.8 percent soundness concerns. percent of all tracts, with 28 percent of tracts of its purchases in underserved counties. For in served counties being redesignated as Freddie Mac the corresponding figures are H. Defining Nonmetropolitan Underserved underserved and 19 percent of tracts in 4.7 percent in served counties and 5.1 Areas underserved counties being redesignated as percent in underserved counties. 1. Whether To Adopt a Tract-Based served. Overall, roughly the same percentage The GSEs are more likely to purchases Definition of Underserved Areas of families (and population) would be mortgages for high-income borrowers in reclassified. However, because underserved The current county-based definition for underserved than in served counties. tracts are somewhat less densely populated targeting GSE purchases to underserved Surprisingly, borrowers in served counties than served tracts, the corresponding nonmetropolitan areas was adopted in 1995 were more likely to have incomes below the proportions of families that shift from served over alternative narrower definitions, such as median than in underserved counties (37.8 and underserved counties are closer: 25 vs. census tracts, despite the use of census tracts percent compared to 34.5 percent). These 21 percent. findings lend some support to the claim that, in metropolitan areas. In the 1995 Final Rule, in rural underserved counties, the GSEs HUD found the merits of a county-based a. Do Census Tracts Allow a Sharper purchase mortgages for borrowers that system of targeting outweighed a tract-based Delineation of Served and Underserved probably encounter few obstacles in system. Now, with seven years of experience Areas? obtaining mortgage credit. under a county-based system, the release of This section compares the differences in There are similarities and one difference Census 2000 data, and improvements in housing need and economic, demographic, between the types of loans that Fannie Mae information technology and systems, HUD and housing conditions in served and and Freddie Mac purchase in served and can reexamine whether to switch to census underserved nonmetropolitan areas classified underserved counties. The GSEs are similar tracts for defining underserved on, respectively, counties and tracts. in that they are slightly more likely to nonmetropolitan areas. This section Additionally, the ‘‘efficiency’’ with which purchase refinance loans in underserved compares impacts of the potential shift in counties and tracts cover the target counties than in served counties; mortgage definition for both served and underserved populations is compared. That is, does tract- purchases with loan-to-value ratios above 80 populations as determined by tract-based and based targeting do a better job of capturing percent are more likely to be in underserved county-based definitions using a number of lower income households and excluding counties than in served counties; and common industry variables as focal points for higher income households than county-based seasoned mortgage purchases are more likely analysis. targeting? to be in underserved than in served counties. The rationale for choosing counties in 1995 Table B.12 presents several indicators of The GSEs differ in that Fannie Mae is slightly rested primarily on perceived shortcomings socioeconomic and housing condition in more likely and Freddie Mac is less likely to of census tracts.51 In particular, rural lenders served and underserved areas under both a purchase loans for first-time homebuyers in did not perceive their market areas in terms tract-based and a county-based definition. In underserved areas than served areas. of census tracts, but rather, in terms of addition, served and underserved counties counties. Another concern was a perceived E. Factor 4: Size of the Conventional are subdivided into their served and lack of reliability in geocoding 1990 census Conforming Mortgage Market for underserved tract components. This allows a tracts. At the same time, HUD found merit in Underserved Areas closer examination of the population and using a tract-based geography for housing characteristics of the tracts that are HUD estimates that underserved areas nonmetropolitan areas. Because tracts reclassified (i.e., served to underserved or account for 35–40 percent of the encompass more homogeneous populations visa versa) under tract-based targeting. Thus, conventional conforming mortgage market. than counties, they permit more precise area characteristics of housing need and The analysis underlying this estimate is targeting of underserved populations. In housing, economic, and demographic detailed in Appendix D. other words, more homogeneous geographic conditions can be compared, for the F. Factor 5: Ability To Lead the Industry areas increase the potential for targeting the following four groups of tracts: (1) Tracts in GSE mortgage purchases into areas where This factor is the same as the fifth factor served counties that would remain ‘‘served’’ borrowers are more likely to face obstacles classified as tracts; (2) tracts that remain considered under the goal for mortgage and other challenges in securing mortgage purchases on housing for low- and moderate- ‘‘underserved’’; (3) tracts that shift from credit. served to underserved; and (4) tracts that income families. Accordingly, see Section G The criteria used for this analysis include of Appendix A for a discussion of this factor, shift from underserved to served. In addition, the following: as well as Section I.5 of this Appendix, we provide counts of tracts falling into each • Do tracts provide a sharper delineation which describes the home purchase subgoal of these groups. If a tract-based classification of served and underserved areas? which is designed to place the GSEs in a of underserved areas improves geographic Specifically, are underserved leadership role in the underserved market. targeting, the regrouping of tracts would be nonmetropolitan populations more clearly more similar to one another than to the other G. Factor 6: Need To Maintain the Sound differentiated by adopting tracts vs. counties? tracts in their respective counties: e.g., Financial Condition of the Enterprises Could service to the underserved formerly underserved areas that become HUD has undertaken a separate, detailed nonmetropolitan populations be more served should be more similar to tracts that economic analysis of this rule, which comprehensive under tract-based definitions? were and remain served than to underserved includes consideration of (a) the financial (unchanged). returns that the GSEs earn on loans in 51 60 FR 61925–61958 (1995) (Appendix B). BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00185 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24412 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00186 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.056 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24413

Socioeconomic and Demographic underserved areas and the gap between 1 implies that none of the families in need Conditions. Table B.12 shows that in served and underserved areas widens when live in an underserved area, or equivalently, important socioeconomic and demographic tracts are used to classify areas. Most notably, all families in underserved areas are not in characteristics, tract-based targeting would the percent of owner-occupied housing units need. more effectively distinguish underserved switches from being higher in underserved Comparing coverage efficiency for counties populations. Median family income, poverty, than served counties to being significantly and tracts indicates that tracts do a better job; unemployment, school dropout rates, and lower among underserved tracts. With a shift capturing a higher percentage of minority population all exhibit greater to tracts overall ownership drops in nonmetropolitan families whose income falls differences between served and underserved underserved areas, from 74 to 72 percent, and below the applicable income threshold and areas using tracts. For example, the increases in served areas from 74 to 77 excluding more families whose income difference in median income between served percent. In contrast, the homeownership rate exceeds the threshold.52 Overall, the and underserved counties is $9,579, or for tracts located in served counties that efficiency index rises from 22.0 to 27.4 alternatively, between served and would be deemed underserved if judged percent. underserved tracts, the difference is $12,744. separately is only 65 percent. In fact, this rate Given income thresholds that are not far Similarly, there is a 7-percentage point gap is much lower even than underserved tracts in poverty rates (7.5 vs. 14.5 percent poverty) in underserved counties. Shifting these tracts away from median income in most places using counties, which widens to 8.6 from served to underserved largely accounts and the degree of income variation even with percentage points (6.6 vs. 15.3 percent) using for the switching of homeownership rates. census tract boundaries, it should not come tracts. Minority population also is captured Results for other indicators of housing as a great surprise that neither the levels of somewhat better with tracts, with the served/ need and conditions are less clear-cut. No coverage efficiency (22–27 percent) nor underserved gap increasing from 16.5 to 17.3 definitive patterns are apparent for two, improvement produced in applying tracts (5 percentage points. In all cases, the levels of admittedly weak, measures of housing percent) are not more dramatic. Nevertheless, the indicators for underserved areas move in quality—units with complete plumbing and tracts do produce better tracking of lower a direction consistent with targeting lower units with complete kitchen facilities, as well income, very low income, and minority income households and areas with higher as for crowding. Purchase affordability, as families. minority populations. measured by the ratio of median housing b. Does GSE Performance Vary between The 4-way breakdown of served and value to the income necessary to qualify for Served and Underserved Tracts Within underserved counties reveals some a loan for the median valued unit, is higher Underserved Counties? significant differences between the two in underserved areas than in served areas. component groups. In most respects, However, the measure of purchase A similar analytical approach is used to ‘‘underserved tracts’’ (i.e., those meeting the affordability presented here is influenced by examine how a shift to tracts would impact underserved criteria), whether located in an many market and other economic factors, GSE purchases. Having applied income and underserved or served county, are more alike some of which do not relate to housing need. minority thresholds from the 2000 census than they are like served tracts. Using median For example, a low affordability ratio may and updating census tract geography, Table income again to illustrate, the effect of reflect abundant supply, but it may also B.13 compares, respectively, 2000, 2001, and reclassifying areas by tract characteristics is reflect low demand stemming from, e.g., 2002 GSE purchases for served and to put together two groups of underserved limited availability of credit or high interest underserved counties and tracts and also for tracts: tracts that were in previously rates. the served and underserved tracts within underserved counties and are not reclassified Coverage Efficiency. The coverage county boundaries. On net there would be and tracts that were in served counties but efficiency index measures the effect of somewhat more tracts classified as meet the underserved criteria. A new group adopting tract-based targeting. This index can underserved under a tract-based system than of served tracts is similarly formed. In both be used to indicate how well underserved currently: 6,782 vs. 6,414. As noted above, cases, the difference in median incomes of areas encompass populations deemed to be however, 23.1 percent of all tracts are the constituent groups is about $3,500. In underserved (‘‘sensitivity’’) and to exclude reclassified. Moving to tracts also would have contrast, the served and underserved populations that are deemed to be served a significant effect on the relative counties now encompass ‘‘served’’ and (‘‘specificity’’). The index is computed for performance of the GSEs. In 2002, Fannie ‘‘underserved’’ groups of tracts whose median income as the difference in two Mae’s performance would drop 2.1 respective median incomes differ by almost percentages: (1) the proportion of all families percentage points to 35.4 percent, while $11,000. Combined with the fact that a fairly in nonmetropolitan areas that meet the Freddie Mac’s performance would increase large number of tracts are affected overall applicable income threshold who live in by 0.9 percent to 32.7 percent. (i.e., switch), these results support an underserved tracts minus (2) the proportion BILLING CODE 4210–27–P assessment that counties are relatively crude of all families in nonmetropolitan areas that for targeting underserved populations. do not meet the applicable underserved Housing Needs and Conditions. Table B.12 income threshold who live in underserved 52 In areas with 30 percent or greater minority shows that tract-based targeting would areas. This difference can range from 1 population, all families with income in excess of produce modest gains in focusing GSE (perfect) to—1 (bad; perverse). For example, 120 percent of the greater of State or national median income are counted as qualifying as ‘‘in purchases on areas with relatively greater a coverage efficiency index equal to 1 implies need’’ for these computations. Similarly, in areas housing needs and conditions as measured that every family in need is living in an with less than 30 percent minority, those minority by low owner-occupancy, higher vacancy underserved area while there are no families (headed) families with income between 95 and 120 rates, and crowding. For each of these who are not in need living in an underserved percent of the applicable median income are not indicators, measured need increases in area; a coverage efficiency index equal to— classified as ‘‘in need.’’

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00187 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24414 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00188 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.057 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24415

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 crowding, etc. However, served areas would 2. Alternative Definitions of Underservice 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 tracts as the geographic unit. having greatest housing need. Alternative from losing almost half of the families in County Results. The main effect of definitions of underservice are considered as target income ranges without any appreciable lowering the general income threshold from follows: (1) Variations of the current gain in specificity, i.e., shrinking the 95 to 90 to 80 percent of the reference income thresholds; (2) applying only the State proportion of people living in underserved is to roughly halve the number of counties median income level for qualifying counties with incomes above the respective underserved counties and tracts; and (3) and population residing in underserved areas. Under the current definition, 11.6 target levels. Similar patterns are observed establishing different thresholds in for families with below 70 percent of micropolitan and ‘‘outside of core’’ million people reside in underserved areas as opposed to fewer than 10 million in served reference income, below 50 percent of nonmetropolitan areas. In each case the reference income, and families in poverty. objective is to assess how redesignating areas. With a general income threshold of 80 percent, 5.7 million would be left in The second set of comparisons builds on served and underserved areas would affect the first set by lowering the income threshold relative conditions and needs and GSE underserved areas. A 90 percent threshold would produce a shift of approximately half applicable to areas with a relatively high purchasing performance. In distinguishing minority populations (30 percent) from 120 micropolitan and ‘‘outside of core’’ areas, it this amount. In terms of social, economic, demographic, to 110 percent in addition to the general is of interest to determine whether it would threshold. This change further shrinks, albeit, be appropriate to establish different and housing characteristics, lowering the only marginally, the size and population of thresholds for underservice. The overarching income threshold from 95 to 80 percent underserved areas. Minority underserved criterion for evaluating and comparing would have the following notable populations would be smaller and definitions is their ability to serve very low- consequences: socioeconomic and housing conditions income, low-income and moderate-income • Minority population in underserved would be worse. Not surprisingly, coverage households, households in poverty, first-time areas would increase from 12.4 to 20.8 efficiencies and GSE purchase performance homebuyers, minorities, and households in percent with no significant change in served levels also would decline across the board, remote locations.53 areas. although the marginal effects of reducing the In the current definition, areas are • Median income would fall in both served minority income threshold are quite small. classified as underserved if either the and underserved areas with the difference 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 loses 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 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00189 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24416 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

socioeconomic and housing conditions in Census, the 2000 Census includes a larger areas. HUD’s analysis of HMDA data shows underserved areas and below-average in number of census tracts that meet HUD’s that mortgage credit is substantially lower in terms of conditions in served areas. Coverage definition of underserved area. The proposed high-minority and low-income efficiencies for all cohorts would be lower new 38 percent-40 percent goals are neighborhoods and mortgage denial rates are than for the current definition of commensurate with recent market share much higher for residents of these underservice and GSE performance overall estimates of 37–40 percent for 1999–2002, neighborhoods. The economics literature would be approximately 90 percent of the presented in Appendix D. discusses the underlying causes of these current level. In addition, an Underserved Areas Housing disparities in access to mortgage credit, Using the State median income, alone, as Subgoal of 33 percent is proposed for the particularly as related to the roles of the general reference income would reduce GSEs’ acquisitions of single-family-owner discrimination, segregation, ‘‘redlining’’ of the number underserved counties relative to home purchase loans in metropolitan areas in specific neighborhoods, and the barriers the current definition, and, although there 2005, with the proposed subgoal rising to 34 posed by underwriting guidelines that would still be more underserved counties percent in 2006 and 35 percent in both 2007 disadvantage applicants from inner city (1,274 vs. 1,064), the underserved population and 2008. The subgoal is designed to neighborhoods. Studies reviewed in Section actually would become smaller than the encourage the GSEs to lead the primary B of this Appendix found that the racial and served population. The effect of this market in providing mortgage credit in income composition of neighborhoods alternative on differences in housing underserved areas. influence mortgage access even after conditions and needs between served and This section summarizes the Secretary’s accounting for demand and risk factors that underserved areas is generally small and consideration of the six statutory factors that may influence borrowers’ decisions to apply ambiguous, but overall, results in less led to the Underserved Area Housing Goal for loans and lenders’ decisions to make contrast. Consistent with the results for other and the subgoal for home purchase loans in those loans. Therefore, the Secretary alternatives, applying a State median income metropolitan areas. This section discusses concludes that high-minority and low- standard, alone, would result in lower the Secretary’s rationale for defining income neighborhoods in metropolitan areas coverage efficiency across all target groups. underserved areas and it compares the are underserved by the mortgage system. The Census Tract Results. As discussed above, characteristics of such areas and untargeted income and minority composition of an area the adoption of a tract-based system would areas. The section draws heavily from earlier is a good measure of whether that area is result in greater coverage efficiency of sections which have reported findings from being underserved by the mortgage market. underserved populations and sharper HUD’s analyses of mortgage credit needs as 2. Identifying Underserved Portions of distinctions in the socioeconomic, well as findings from other research studies Metropolitan Areas demographic and housing characteristics of investigating access to mortgage credit. served and underserved areas. That is, tracts To identify areas underserved by the more effectively carve out areas that exhibit 1. Housing and Credit Disparities in mortgage market, HUD focused on two characteristics that are associated with Metropolitan Areas traditional measures used in a number of underservice, such as low income, large There are families who are not being studies based on HMDA data: application minority populations and low adequately served by the nation’s housing denial rates and mortgage origination rates homeownership. The converse is true for and mortgage markets. A major HUD-funded per 100 owner-occupied units. Tables B.2 served areas. In analysis at the tract level, study of discrimination in the sales and and B.3 in Section B of this Appendix these patterns tend to be maintained quite rental markets found that while presented detailed data on denial and consistently. A tract-based system would discrimination against minorities was origination rates by the racial composition improve the power to differentiate generally down since 1989, it remained at and median income of census tracts for underserved and served populations. unacceptable levels in 2000. The greatest metropolitan areas. Aggregating this data is According to virtually every indicator of share of discrimination against Hispanic and useful in order to examine denial and socioeconomic, demographic, and housing African American home seekers can still be origination rates for broader groupings of conditions, applying State median income, attributed to being told that units are census tracts:55 alone, with a tract-based geography would unavailable when they are available to whites produce superior differentiation to the and being shown and told about fewer units Minority composi- current county-based definition. In terms of Denial rate than a comparable white home seeker. There tion (percent) Orig. rate coverage efficiency, we again see has also been an upward trend of (percent) improvement with tracts, but not enough to discrimination in the area of geographic offset the loss of eliminating the national steering for African Americans. 0–30 ...... 8.7 19.3 median income threshold. For the Racial disparities in mortgage lending are 30–50 ...... 11.2 19.3 underserved population, for example, also well documented. HUD-sponsored 50–100 ...... 16.3 14.7 coverage efficiency would be 16.9 percent studies of the pre-qualification process with tracts, still below 22 percent under the conclude that African Americans and current definition.54 Denial rate Hispanics faced a significant risk of unequal Tract income (percent) Orig. rate I. Determination of the Underserved Areas treatment when they visit mainstream Housing Goal mortgage lenders. Numerous studies of Less than 90% of HMDA data have shown that mortgage denial The proposed annual goal for each GSE’s AMI ...... 15.6 13.9 rates are substantially higher for African purchases of mortgages financing housing for 90–120% ...... 10.1 18.6 Americans and Hispanics, even after properties located in geographically targeted Greater than 120% 7.1 22.7 controlling for applicant income. And the areas (central cities, rural areas, and other now-famous Boston Fed study found that the underserved areas) is 38 percent of eligible Two points stand out. First, high-minority units financed in 2005, 39 percent in 2006 higher denial rates for minorities remained census tracts have higher denial rates and and 2007, and 40 percent in 2008. The 2008 after controlling for a host of underwriting lower origination rates than low-minority goal will remain in effect in subsequent characteristics, such as the credit record of tracts. Specifically, tracts that are over 50 years, unless changed by the Secretary prior the applicant. Partly as a result of these racial percent minority have nearly twice the denial to that time. The goal of 38 percent for 2005 disparities in the housing and mortgage rate and three-fourths the origination rate of is larger than the goal of 31 percent for 2001– markets, the homeownership rate for tracts that are under 30 percent minority.56 03 mainly because, compared with the 1990 minorities is 25 percentage points below that for whites. 55 There are also neighborhoods that are not Denial rates are computed for mortgage 54 applications without manufactured housing loans. Note that, unlike the other panels in tables 6.3 being adequately served by the nation’s and 6.8, ‘‘underserved population’’ is defined Origination rates equal home purchase and according to the applicable definition. Thus, housing and mortgage industries. The refinance mortgages (without subprime loans) per eliminating the national median income test, existence of substantial neighborhood 100 owner occupants in a census tract. narrows the defined cohort of underserved families. disparities in homeownership and mortgage 56 The differentials in denial rates are due, in part, Despite this, coverage falls. credit is well documented for metropolitan to differing risk characteristics of the prospective

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00190 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24417

Second, census tracts with lower incomes continues to be a good proxy for underserved census tracts in metropolitan areas, covering have higher denial rates and lower areas in metropolitan areas. The income and 49.2 percent of the metropolitan population origination rates than higher income tracts. minority cutoffs produce sharp differentials in 2000. (By contrast, the 1990-based Tracts with income less than 90 percent of in denial and origination rates between definition included 21,587 of the 45,406 area median income have over twice the underserved areas and adequately served census tracts in metropolitan areas, covering denial rate and three-fifths the origination areas. For example, in 2002 the mortgage 44.3 percent of the metropolitan population rate of tracts with income over 120 percent denial rate in underserved areas (14.0 in 1990.) The 2000-based definition includes of area median income. percent) was over one-and-a-half times that 75.7 percent of the population living in In both the 1995 and the 2000 GSE Rules, in adequately served areas (8.9 percent). poverty in metropolitan areas. The HUD’s research determined that These minority population and income unemployment rate in underserved areas is ‘‘underserved areas’’ could best be thresholds apply in the suburbs as well as in characterized in metropolitan areas as census more than twice that in served areas, and central cities. The average denial rate in owner units comprise only 51.6 percent of tracts where: (1) median income of families underserved suburban areas (13.7 percent) is in the tract does not exceed 90 percent of total dwelling units in underserved tracts, 1.7 times that in the remaining served areas versus 75.9 percent of total units in served area (MSA) median income or (2) minorities of the suburbs (8.0 percent), and is almost as comprise 30 percent or more of the residents tracts. As shown in Table B.14, this large as the average denial rate (15.8 percent) definition covers most of the population in and median income of families in the tract in underserved central city tracts. Low- several distressed central cities including does not exceed 120 percent of area median income and high-minority suburban tracts income. The earlier analysis was based on Bridgeport (100 percent), Newark (99 appear to have credit problems similar to 1990 Census data. HUD has now conducted percent), and Detroit (93 percent). The their central city counterparts. Thus HUD the same analysis using 2000 Census data nation’s five largest cities also contain large uses the same definition of underserved areas and has determined that the above definition concentrations of their population in throughout metropolitan areas—there is no need to define such areas differently in underserved areas: New York (68 percent), borrowers in different areas. However, use of denial Los Angeles (72 percent), Chicago (75 rates is supported by the findings in the Boston Fed central cities and in the suburbs. This definition of metropolitan percent), Houston (73 percent), and Phoenix study which found that denial rate differentials (50 percent). persist, even after controlling for risk of the underserved areas based on 2000 Census borrower. See Section B for a review of that study. geography includes 26,316 of the 51,040 BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00191 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24418 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00192 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.058 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24419

3. Identifying Underserved Portions of Tract-based targeting would potentially targeting. Defining underserved families as Nonmetropolitan Areas focus GSE purchases in areas with relatively those in any area whose income was less Based on an exploration of alternative greater housing needs and conditions as than 95 percent of the reference income (or numerical criteria for identifying measured by owner-occupancy, vacancy in areas with a minority population of 30 underserved nonmetropolitan areas using rates, and crowding. For each of these percent or more, families with incomes 2000 census data, HUD has concluded that indicators, measured need increases in below 120 percent of the reference income) the current definition of underservice is underserved areas and the gap between the use of more refined tract geography broad but efficacious and that any narrower served and underserved areas widens when results in a 5 percentage point increase in the tracts are used to classify areas. Most notably, definition of underservice would not serve coverage efficiency index, from 22 to 27 homeownership would be significantly lower congressional intent under FHEFSSA. percent. This reflects two improvements in underserved areas relative to served areas Narrowing the definition of underservice under a tract-based system. Currently, and under a tract system: underserved areas potentially could promote more intense contrary to expectations, homeownership would capture more of the nonmetropolitan purchasing in needier communities, but this actually is slightly greater in underserved ‘‘underserved’’ families (62 vs. 65 percent) seems unlikely. On the contrary, the greatest areas. Driving this reversal is the fact that and fewer ‘‘served’’ families (decreasing from marginal impact on GSE purchasing could be tracts in served counties that would be 40 to 37 percent of families in underserved in the very areas that would be excluded reclassified as underserved tracts have an areas). under the alternatives. ownership rate of just 65 percent, which is Research comparing a tract-based system 4. Past Performance of the GSEs much lower even than in the underserved for defining underserved areas with the tracts in underserved counties, where Goals Performance. In the October 2000 current county-based system, using 2000 ownership is 73 percent. Meanwhile, the rule, the underserved areas goal was set at 31 census data, indicates that a tract-based served tracts in served and underserved percent for 2001–03. Effective on January 1, system would result in more effective counties have the same ownership rate of 77 2001, several changes in counting geographic targeting of GSE purchases. percent, which is significantly higher than in requirements came into effect for the Although the total number of tracts underserved areas. undeserved areas goal, as follows: (a) ‘‘bonus designated as served and underserved areas Two groups of measures of housing points’’ (double credit) for purchases of would change very little, 23 percent of all conditions—housing quality and mortgages on small (5–50 unit) multifamily tracts would be reclassified, reassigning affordability—exhibit less clear-cut results properties and, above a threshold level, approximately equal numbers of families from applying tracts. However, we conclude from served to underserved and from mortgages on 2–4 unit owner-occupied that these results are consistent with the properties; (b) a ‘‘temporary adjustment underserved to served. ambiguous patterns discussed in chapter 4 factor’’ (1.35 units credit) for Freddie Mac’s The main effect of the reclassification is to above and do not undermine the overall purchases of mortgages on large (more than align tracts into more homogeneous and conclusion that basing geographic targeting distinct groups as measured by differences in on tracts would more sharply define areas 50 unit) multifamily properties; and (c) key socioeconomic and demographic with greater housing need and adverse eligibility for purchases of certain qualifying characteristics such as median family housing conditions. government-backed loans to receive goal income, poverty, unemployment, school Not surprisingly, the results from analyzing credit. Under these counting rules, as shown dropouts, and minority population. As a housing, socioeconomic, and demographic in Figure B.2, Fannie Mae’s performance in result of reclassification, underserved areas characteristics are further reinforced in 2001 was 32.6 percent and Freddie Mac’s stand out more as areas of lower income and finding that a tract-based system would better performance was 31.7 percent; thus both economic activity and somewhat larger capture underserved populations and GSEs surpassed the goal of 31 percent. minority populations. exclude served populations from geographic BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00193 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24420 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00194 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.059 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24421

Counting requirements (a) and (b) expired 38.1 percent in 2000, 36.6 percent in 2001, 1990-based definition, which means the GSE- at the end of 2003, while (c) will remain in and 35.9 percent in 2002. Freddie Mac’s market comparisons need to be updated to effect after that. If this counting approach— performance would have been 35.1 percent incorporate tract designations from the 2000 without the bonus points and the ‘‘temporary in 2000, 33.5 percent in 2001, and 33.6 Census. Therefore, for the years 1999, 2000, adjustment factor’’—had been in effect in percent in 2002. 2001, and 2002, HUD used various 2000 and 2001, and the GSEs’ had purchased Single-Family-Owner Home Purchase apportionment techniques to re-allocate the same mortgages that they actually did Mortgages. Sections E.9 of Appendix A and 1990-based GSE and HMDA data into census purchase in both years, then Fannie Mae’s D.2 of this appendix compared the GSEs’ tracts as defined by the 2000 Census. performance would have been 31.0 percent funding of home purchase loans in Switching to the 2000-based tracts increases in 2000, 30.4 percent in 2001, and 30.2 underserved areas with originations by the underserved area share of market percent in 2002. Freddie Mac’s performance lenders in primary market. To take advantage originations by 5.7 percentage points. would have been 29.2 percent in 2000, 28.2 of HMDA and GSE data going back to 1993, Between 1999 and 2002, 31.5 percent of percent in 2001, and 29.4 percent in 2002. the analysis was conducted using 1990 mortgage originations (without B&C loans) Therefore, Fannie Mae would have just Census tract geography. While both GSEs were originated in underserved tracts based matched the underserved areas goal of 30 have improved their performance since 1993, percent in 2000 and fallen short in 2001 and they have both lagged the conventional on 2000 geography, compared with 25.8 2002, while Freddie Mac would have fallen conforming market in providing affordable percent based on 1990 geography. As shown short of the goal in 2000–2002. loans to underserved areas. The 1990-based in Table B.8 of Section D.2, the underserved The above performance figures are for analysis shows that the two GSEs have areas share of each GSE’s purchases also rises underserved areas (census tracts in engaged in very different patterns of by approximately 5.5 percentage points. metropolitan areas and counties in non- funding—Freddie Mac has been much less Thus, conclusions about the GSEs’ metropolitan areas) defined in terms of 1990 likely than Fannie Mae to fund home loans performance relative to the market are similar Census geography. Switching to 2000 Census in underserved neighborhoods. HUD will whether the analysis is conducted in terms data increases the coverage of underserved begin defining underserved areas based on of 2000 Census geography or 1990 Census areas, which increases the share of the GSEs’ 2000 Census geography and new OMB geography. purchases in underserved areas by definitions of metropolitan areas in 2005, the The analysis for home purchase loans approximately 5 percentage points. Based on first year of the proposed rule. As noted based on 2000 Census geography will be 2000 Census geography, and excluding above, the 2000-based definition of summarized here (see Section D.2 of this counting requirements (a) and (b) then underserved areas includes 5,372 more appendix for a similar analysis using 1990- Fannie Mae ’s performance would have been census tracts in metropolitan areas than the based geography):

Market (w/o Year Freddie Mac Fannie Mae B&C) (percent) (percent) (percent)

1999 ...... 26.1 27.0 31.4 2000 ...... 27.4 29.9 32.9 2001 ...... 27.4 30.8 31.6 2002 ...... 31.7 32.3 32.3 1999–2002 (average) ...... 28.3 29.5 31.5 1996–2001 (estimate) ...... 27.1 29.0 31.1

Between 1999 and 2002, 28.3 percent of 5. Ability To Lead the Single-Family-Owner in 2002. Loans in the B&C portion of the Freddie Mac’s purchases and 29.5 percent of Market: A Subgoal for Underserved Areas subprime market are excluded from the Fannie Mae’s purchases financed properties The Secretary believes the GSEs can play market average of 31.5 percent for 1999– in underserved neighborhoods, compared a leadership role in underserved markets. 2001. with 31.5 percent home purchase loans Thus, as discussed in Section D.2, the The subgoal applies only to the GSEs’ originated in the conventional conforming Department is proposing to establish a purchases in metropolitan areas because the HMDA-based market benchmark is only market (excluding B&C loans). Thus, Freddie subgoal of 33 percent for each GSE’s available for metropolitan areas. HMDA data Mac performed at 90 percent of the market acquisitions of home purchase loans for single-family-owner properties located in the for non-metropolitan counties are not reliable level, while Fannie Mae performed at 94 enough to serve as a market benchmark. The percent of the market level—both results underserved census tracts of metropolitan areas in 2005, rising to 34 percent in 2006 Department is also setting home purchase similar to those reported above for and 35 percent in both 2007 and 2008. If the subgoals for the other two goals-qualifying underserved areas based on 1990 Census GSEs meet this subgoal, they will be leading categories, as explained in Appendices A and geography. The 2000-based results also show the primary market by about 1.5 percentage C. that Fannie Mae has improved its points in 2005 and 3.5 percentage points in The approach taken is for the GSEs to performance and matched the primary 2007–2008, based on historical data. As obtain their leadership position by staged market in funding underserved areas during discussed above, underserved areas increases in the underserved areas subgoal; 2002. The share of Fannie Mae’s purchases accounted for an average of 31.5 percent of this will enable the GSEs to take new going to underserved areas increased from home purchase loans originated in the initiatives in a correspondingly staged 27.0 in 1999 to 32.3 percent in 2002, which conventional conforming market of manner to achieve the new subgoal each placed it at the market level. However, the metropolitan areas (computed over 1999– year. Thus, the increases in the underserved 2000-based results show that, like Freddie 2002 or over 2001–2002). To reach the 33- areas subgoal are sequenced so that the GSEs can gain experience as they improve and Mac, Fannie Mae’s longer-term performance percent (35-percent) subgoal for 2005 (2007– move toward the new higher subgoal targets. (since 1996) as well as its recent average 2008), both GSEs would have to improve their performance—Fannie Mae by 1.9 (3.9) Appendix A discusses in some detail the performance (1999 to 2001) has consistently percentage points over its average factors that the Department considered when been below market levels. But, it is performance of 31.1 percent during 2001 and setting the subgoal for low- and moderate- encouraging that Fannie Mae significantly 2002, and by 0.7 (2.7) percentage points over income loans. Several of the considerations improved its performance relative to the its performance of 32.3 percent in 2002; and were general in nature—for example, related market during the first two years of HUD’s Freddie Mac by 3.4 (5.4) percentage points to the GSEs’ overall ability to lead the single- higher housing goal levels. (See Section D.2 over its average performance of 29.6 percent family-owner market—while others were for the method of estimating the 1996–2002 in 2001 and 2002, and by 1.3 (2.3) percentage specific to the low-mod subgoal. Because the average results.) points over its performance of 31.7 percent reader can refer to Appendix A, this

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00195 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24422 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

appendix provides a briefer discussion of the who disproportionately live in underserved the volatility of mortgage markets and the more general factors. The specific areas, are projected to account for almost possible impacts on the GSEs’ ability to meet considerations that led to the subgoal for two-thirds of the growth in the number of the housing goals. Should conditions change underserved areas can be organized around new households over the next ten years. To such that the goals are no longer reasonable the following four topics: meet the diverse and unique needs of these or feasible, the Secretary has the authority to (1) The GSEs have the ability to lead the families, the GSEs must continue adjusting revise the goals. 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 7. The Underserved Areas Housing Goal for market for single-family-owner loans, which these areas and hopefully attract more 2005–2008 is their ‘‘bread-and-butter’’ business. Both mainstream lenders into our inner city The proposed Underserved Areas Housing GSEs have been dominant players in the neighborhoods. Goal for 2005 is 38 percent of eligible home purchase market for years, funding 57 (4) There are ample opportunities for the purchases, rising to 39 percent in 2006 and percent of the single-family-owner mortgages GSEs to improve their performance. 40 percent in 2007 and 2008. Five percent of financed between 1999 and 2002. Through Mortgages are available for the GSEs to the seven percentage point increase in 2005 their many new product offerings and their purchase in underserved areas. They can simply reflects the expanded coverage of various partnership initiatives, the GSEs have improve their performance and lead the HUD’s definition in the 2000 Census tract shown that they have the capacity to operate primary market in purchasing loans in these data. The bonus points for small multifamily in underserved neighborhoods. They also low-income and high-minority properties and owner-occupied 2–4 units, as have the staff expertise and financial neighborhoods. The underserved areas share well as Freddie Mac’s Temporary Adjustment resources to make the extra effort to lead the of the home purchase market has consistently Factor, will no longer be in effect for goal primary market in funding single-family- been around 31 percent since 1995, which counting purposes. It is recognized that owner mortgages in undeserved areas. suggests a degree of underlying strength in neither GSE would have met the 38-percent (2) The GSEs have lagged the market. Even the market. According to the market share target for 2005 in the past three years. Fannie though they have the ability to lead the data reported in Table A.30 of Appendix A, Mae’s performance is projected to have been market, they have not done so, as discussed the GSEs have been purchasing about half of 37.5 percent in 2000, 35.7 percent in 2001, above. The type of improvement needed to new originations in underserved areas, which and 35.0 percent in 2002, under a 2000-based meet this new underserved area subgoal was means there are plenty of purchase underserved area goal. Freddie Mac’s demonstrated by Fannie Mae during 2001 opportunities left for them in the non-GSE performance is projected to have been 34.1 and 2002. During 2001, underserved area portion of that market. In addition, the GSEs’ percent in 2000, 32.5 percent in 2001, and loans declined as a percentage of primary purchases under the subgoal are not limited 32.8 percent in 2002. However, the market market originations (from 32.2 to 30.9 to new mortgages that are originated in the for the Underserved Areas Housing Goal percent), but they increased as a percentage current calendar year. The GSEs can averaged 39 percent between 1999 and 2002. of Fannie Mae’s purchases (from 29.1 to 29.8 purchase loans from the substantial, existing Thus, the GSEs should be able to improve percent); and during 2002, they increased stock of affordable loans held in lenders’ their performance enough to meet these further as a percentage of Fannie Mae’s portfolios, after these loans have seasoned targets of 38 percent-40 percent. purchases (from 29.8 to 32.3 percent), placing and the GSEs have had the opportunity to The objective of HUD’s proposed Fannie Mae at the market level. observe their track record. In fact, both GSEs Underserved Areas Housing Goal is to bring (3) There are disparities among have often purchased seasoned loans that the GSEs’ performance to the upper end of neighborhoods in access to mortgage credit. were used to finance properties in HUD’s market range estimate for this goal There remain troublesome neighborhood underserved areas (see Table A.11 in (35–40 percent), consistent with the statutory disparities in our mortgage markets, even Appendix A). criterion that HUD should consider the GSEs’ after the substantial growth in conventional To summarize, although single-family- ability to lead the market for each Goal. To lending to low-income and minority owner mortgages comprise the ‘‘bread-and- enable the GSEs to achieve this leadership, neighborhoods that accompanied the so- butter’’ of their business, the GSEs have the Department is proposing modest called ‘‘revolution in affordable lending’’. lagged behind the primary market in increases in the Underserved Areas Housing There is growing evidence that inner city financing properties in underserved areas. Goal for 2005 which will increase further neighborhoods are not being adequately For the reasons given above, the Secretary through 2008, to achieve the ultimate served by mainstream lenders. Some have believes that the GSEs can do more to raise objective for the GSEs to lead the market concluded that a dual mortgage market has the share of their home loan purchases in under a range of foreseeable economic developed in our nation’s financing system, underserved areas. This can be accomplished circumstances by 2008. Such a program of with conventional mainstream lenders by building on efforts that the enterprises staged increases is consistent with the serving white families living in the suburbs have already started, including their new statutory requirement that HUD consider the and FHA and subprime lenders serving affordable lending products, their many past performance of the GSEs in setting the minority families concentrated in inner city 57 partnership efforts, their outreach to inner Goals. Staged increases in the Underserved neighborhoods. In addition to the Areas Housing Goal will provide the unavailability of mainstream lenders, city neighborhoods, their incorporation of enterprises with opportunity to adjust their families living in these often highly- greater flexibility into their underwriting business models and prudently try out segregated neighborhoods face many guidelines, and their purchases of CRA loans. business strategies, so as to meet the required additional hurdles, such as lack of cash for A wide variety of quantitative and qualitative 2008 level without compromising other a down payment, credit problems, and indicators indicate that the GSEs’ have the business objectives and requirements. discrimination. Immigrants and minorities, resources and financial strength to improve their affordable lending performance enough The analysis of this section implies that there are many opportunities for Fannie Mae 57 to lead the market in underserved areas. See Dan Immergluck, Stark Differences: The and Freddie Mac to improve their overall Explosion of the Subprime Industry and Racial 6. Size of the Mortgage Market for performance on the Underserved Areas Hypersegmentation in Home Equity Lending, Underserved Areas Woodstock Institute, October 2000; and Daniel Housing Goal. The GSEs provided financing Immergluck and Marti Wiles, Two Steps Back: The As detailed in Appendix D, the market for for 49 percent of the single-family and Dual Mortgage Market, Predatory Lending, and the mortgages in underserved areas is projected multifamily units that were financed in the Undoing of Community Development, Woodstock to account for 35–40 percent of dwelling conventional conforming market between Institute, Chicago, IL, November 1999. For a nationl units financed by conventional conforming 1999 and 2002. However, in the underserved analyses, see the HUD report Unequal Burden: mortgages; in estimating the size of the areas portion of the market, the GSE’s Income and Racial Disparities in Subprime Lending market, HUD used alternative assumptions purchases represented only 41 percent of the in America, April 2000; and Randall M. Scheesele, about future economic and market conditions dwelling units that were financed in the Black and White Disparities in Subprime Mortgage Refinance Lending, Housing Finance Working Paper that were less favorable than those that market. Thus, there appears to be ample No. HF–114, Office of Policy Development and existed over the last five years. Between 1999 room for the GSEs to increase their purchases Research, U.S. Department of Housing and Urban and 2002, the underserved areas market of loans that qualify for the Underserved Development, April 2002. averaged 39 percent. HUD is well aware of Areas Housing Goal. In addition, there are

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00196 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24423

several market segments that would benefit incomes are 50 percent of area median Reports (AHARs) that they submit to the from a greater secondary market role by the income or less, or where at least 40 percent Department. GSEs, and many of these market segments are of the units are affordable to families whose The main finding of this section is that concentrated in underserved areas. incomes are 60 percent of area median both Fannie Mae and Freddie Mac surpassed income or less. the Department’s Special Affordable Housing 8. Conclusions Multifamily Subgoal. HUD has established Goals for each of the seven years during this Having considered the projected mortgage a special affordable subgoal for GSE period. Specifically: 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 GSEs’ recent performance in purchasing mortgage purchases for units qualifying for percent. 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 proposed annual units financed, as for the three housing goals. 2000. Freddie Mac’s performance was 15.2 goal of 38 percent of eligible units financed Both GSEs have consistently surpassed the percent in 1997, 15.9 percent in 1998, 17.2 in, 2005, 39 percent in 2006 and 2007, and multifamily subgoal since its establishment percent in 1999, and 20.7 percent in 2000; 40 percent in 2008 is feasible. The Secretary in 1996. The proposed rule increases the and Fannie Mae’s performance was 17.0 has also proposed a subgoal of 33 percent for subgoal such that, of the total Special percent in 1997, 14.3 percent in 1998, 17.6 the GSEs’ purchases of single-family-owner Affordable mortgage purchases each year, percent in 1999, and 19.2 percent in 2000. mortgages in metropolitan areas, for 2005, each GSE must purchase special affordable • In HUD’s Housing Goals 2000 Final Rule, rising to 34 percent in 2006 and 35 percent multifamily mortgages in dollar amount the special affordable goal was set at 20 in 2007 and 2008. The Secretary has equal to at least 1 percent of its combined percent for 2001–03. As of January 1, 2001, considered the GSEs’ ability to lead the (i.e., single-family and multifamily) annual several changes in counting requirements industry as well as the GSEs’ financial average mortgage purchases over the 2000– took effect for the special affordable goal, as condition. The Secretary has determined that 2002 period. The proposed level of this follows: ‘‘bonus points’’ (double credit) for the proposed goals and subgoals are subgoal is $5.49 billion per year for Fannie purchases of goal-qualifying mortgages on necessary and appropriate. Mae and $3.92 billion per year for Freddie small (5–50 unit) multifamily properties and, above a threshold level, mortgages on 2–4 Appendix C—Departmental Mac. Single-Family-Owner Home Purchase unit owner-occupied properties; a Considerations To Establish the Special ‘‘temporary adjustment factor’’ (1.20 units Affordable Housing Goal Subgoal. The Department proposes to establish a subgoal of 17 percent for the share credit, subsequently increased by Congress to A. Introduction of each GSE’s purchases of single-family- 1.35 units credit) for Freddie Mac’s owner home purchase mortgages that qualify purchases of goal-qualifying mortgages on 1. Establishment of the Goal as special affordable and are originated in large (more than 50-unit) multifamily The Federal Housing Enterprises Financial metropolitan areas in 2005, with the properties; changes in the treatment of Safety and Soundness Act of 1992 proposed subgoal rising to 18 percent in missing data; a procedure for the use of (FHEFSSA) requires the Secretary to 2006, and 19 percent in 2007 and 2008. imputed or proxy rents for determining goal establish a special annual goal designed to credit for multifamily mortgages; and adjust the purchase by each GSE of mortgages B. Consideration of the Factors changes regarding the ‘‘recycling’’ of funds on rental and owner-occupied housing to In considering the factors under FHEFSSA by loan originators. These changes are meet the unaddressed needs of, and to establish the Special Affordable Housing explained below. Fannie Mae’s performance affordable to, low-income families in low- Goal, HUD relied upon data gathered from was 21.6 percent in 2001 and 21.4 percent in income areas and very-low-income families the American Housing Survey through 2000, 2002, and Freddie Mac’s performance was (the Special Affordable Housing Goal). the Census Bureau’s 1991 Residential 22.6 percent in 2001 and 21.4 percent in In establishing the Special Affordable Finance Survey, the 1990 and 2000 Censuses 2002, thus both GSEs surpassed this higher Housing Goal, FHEFSSA requires the of Population and Housing, Home Mortgage goal in both years. This section discusses the Secretary to consider: Disclosure Act (HMDA) data for 1992 October 2000 counting rule changes in detail 1. Data submitted to the Secretary in through 2002, and annual loan-level data and provides data on what goal performance would have been in 2001–02 without these connection with the Special Affordable from the GSEs on their mortgage purchases 2 Housing Goal for previous years; through 2002. Appendix D discusses in detail changes. 2. The performance and efforts of the GSEs how these data resources were used and how In addition, HUD has established a special affordable subgoal for GSE purchases of toward achieving the Special Affordable the size of the conventional conforming multifamily mortgages. This subgoal is Housing Goal in previous years; market for this goal was estimated. expressed in terms of a minimum annual 3. National housing needs of targeted The remainder of Section C discusses the dollar volume of multifamily mortgage families; factors listed above, and Section D provides purchases for units qualifying for the goal, 4. The ability of the GSEs to lead the the Secretary’s rationale for establishing the rather than as a percentage of total units industry in making mortgage credit available Special Affordable Housing Goal. financed, as for the three housing goals. As for low-income and very-low-income Factors 1 and 2. Data submitted to the discussed below, both GSEs surpassed the families; and Secretary in Connection With the Special multifamily subgoal in each of these years. 5. The need to maintain the sound Affordable Housing Goal for Previous Years, financial condition of the enterprises. a. Performance on the Special Affordable and the Performance and Efforts of the Housing Goal in 1996–2002 2. The Goal and Subgoals Enterprises Toward Achieving the Special Affordable Housing Goal in Previous Years HUD’s Housing Goals 1995 Final Rule Special Affordable Housing Goal. The specified that in 1996 at least 12 percent of proposed rule provides that the Special The discussions of these two factors have the number of units financed by each of the Affordable Housing Goal will be 22 percent been combined because they overlap to a GSEs that were eligible to count toward the in 2005, 24 percent in 2006, 26 percent in significant degree. Special Affordable Housing Goal should 2007, and 28 percent in 2008. This section discusses each GSE’s qualify for the goal (that is, be for very low- Units That Count Toward the Goal. Units performance under the Special Affordable income families or low-income families in that count toward the Special Affordable Housing Goal over the 1996–2002 period.1 As low-income areas), and at least 14 percent Housing Goal include units occupied by low- explained in Appendix A, the data presented should qualify in 1997–2000. HUD’s October income owners and renters in low-income are ‘‘official HUD results’’ which, in some areas, and very low-income owners and cases, differ from goal performance reported 2 To separate out the effects of changes in renters. Other low-income rental units in by the GSEs in the Annual Housing Activities counting rules that took effect in 2001, this section multifamily properties count toward the goal also compares performance in 2001 to estimated where at least 20 percent of the units in the 1 Performance for the 1993–95 period was performance in 2000 if the 2001 counting rules had property are affordable to families whose discussed in HUD’s Housing Goals 2000 Final Rule. been in effect in that year.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00197 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24424 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

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 1996–2002 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 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00198 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24425

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00199 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.060 24426 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00200 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.061 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24427

Both GSEs exhibited sharp gains in goal discrepancy did not persist in 2001–02 Adjustment Factor’’ of 1.35 units of credit for performance in 2000—Fannie Mae’s because of a change in counting each qualifying unit financed in ‘‘large’’ performance increased by 1.6 percentage requirements, described below. And for 2002, multifamily properties (i.e., those with 51 or points, to a record level of 19.2 percent, HUD’s official goal performance figure was more units) in the numerator in calculating while Freddie Mac’s performance increased 21.4 percent, somewhat above the figure of special affordable goal performance for 2001– even more, by 3.5 percentage points, which 20.6 percent submitted to the Department by 03.4 This factor did not apply to special also led to a record level of 20.7 percent. Freddie Mac. affordable units in large multifamily Fannie Mae’s performance was 21.6 percent Fannie Mae’s performance on the Special properties whose mortgages were financed by in 2001 and 21.4 percent in 2002; Freddie Affordable Housing Goal surpassed Freddie Fannie Mae during this period. Mac’s performance was 22.6 percent in 2001 Mac’s in 1996–97. This pattern was reversed • Missing data for single-family properties. and 21.4 percent in 2002. However, as in 1998, as Freddie Mac surpassed Fannie The GSEs may exclude loans with missing discussed below, using consistent accounting Mae in goal performance for the first time, borrower income from the denominator if the rules for 2000–02, each GSE’s Special though by only 0.2 percentage point. This property is located in a below-median Affordable Housing Goal performance in improved relative performance of Freddie income census tract, subject to a ceiling of 1 2001 was below its performance in 2000, and Mac was due to its increased purchases of percent of total owner-occupied units in 2002 each enterprise’s performance was multifamily loans, as it re-entered that financed. The enterprises are also allowed to below its 2001 performance level. market, and to increases in the goal- exclude single-family rental units with With regard to the special affordable qualifying shares of its single-family missing rental information from the multifamily subgoal, Fannie Mae’s purchases mortgage purchases. However, Fannie Mae denominator in calculating performance for have exceeded the subgoal by wide margins again surpassed Freddie Mac in special the special affordable goal. • in all years, with performance ranging from affordable goal performance in 1999, 17.6 Missing data and proxy rents for percent to 17.2 percent; Freddie Mac 184 percent of the goal in 1996 to 315 percent multifamily properties. If rent is missing for regained the lead in 2000, 20.7 percent to of the goal in 1999. Fannie Mae’s subgoal was multifamily units, the GSEs may apply 19.2 percent. Freddie Mac’s official more than doubled in the October 2000 rule, ‘‘proxy rents,’’ up to a ceiling of 5 percent of performance also exceeded Fannie Mae’s to a minimum of $2.85 billion in each year total multifamily units financed, in official performance in 2001, but this from 2001 through 2003, but its qualifying determining whether such units qualify for reflected a difference in the counting rules the special affordable goal. If such proxy purchases amounted to $7.36 billion, or 258 applicable to the two GSEs that was enacted percent of the goal, in 2001, and $7.57 rents cannot be estimated, these multifamily by Congress; if the same counting rules were units are excluded from the denominator in billion, or 260 percent of the goal, in 2002. applied to both GSEs, Fannie Mae’s Freddie Mac has also exceeded its special calculating performance under these goals. performance would have exceeded Freddie • Change in ‘‘recycling’’ requirements. affordable multifamily subgoals in every Mac’s performance, by 21.6 percent to 21.1 year, albeit by smaller margins than Fannie Under Section 1333(b)(1)(B) of FHEFSSA, if percent. a GSE acquires a portfolio of mortgages Mae. In 1996 Freddie Mac’s special In 2002, Freddie Mac’s performance on the affordable multifamily mortgage purchases originated in a previous year (that is, special affordable goal was the same as seasoned mortgages) that qualify under the amounted to $1.06 billion, or 107 percent of Fannie Mae’s performance (21.4 percent), the goal. This ratio rose to 122 percent in Special Affordable Housing goal, the seller even though Freddie Mac had the advantage must be ‘‘engaged in a specific program to 1997, and exceeded 200 percent for each year of the Temporary Adjustment Factor, which from 1998 through 2000. Freddie Mac’s use the proceeds of such sales to originate did not apply to performance by Fannie Mae. additional loans that meet such goal’’ and subgoal was more than doubled in the Freddie Mac’s performance would have October 2000 rule, to a minimum of $2.11 in such purchases or refinancings must trailed Fannie Mae’s without this factor, and ‘‘support additional lending for housing that each year from 2001 through 2003, but its in fact Freddie Mac would have only slightly otherwise qualifies under such goal’’ in order qualifying purchases amounted to $4.65 exceeded the goal, at 20.2 percent. to receive credit toward the goal. This has billion, or 220 percent of the goal, in 2001, b. Changes in the Goal Counting Rules for been referred to as the ‘‘recycling and $5.22 billion, or 247 percent of the goal, 2001–03 requirement.’’ The 2000 rule both clarified in 2002. the conditions under which HUD would The official figures for Freddie Mac’s Several changes in the counting rules regard these statutory conditions to be special affordable goal performance underlying the calculation of special satisfied and established certain categories of presented above differ from the affordable goal performance took effect lenders that would be presumed to meet the corresponding figures presented by Freddie beginning in 2001. Most of these also applied to the low- and moderate-income goal and recycling requirements. These included BIF- Mac in its Annual Housing Activity Reports insured and SAIF-insured depository to HUD by 0.1–0.2 percentage point for 1996– are discussed in Appendix A; only brief summaries of those changes are given here: institutions that are regularly in the business 2000, reflecting minor differences in the • of mortgage lending and which are subject to, application of counting rules. The official Bonus points for multifamily and single- family rental properties. Each qualifying unit and have received at least a satisfactory figures for special affordable goal Community Reinvestment Act performance performance by both GSEs are the same as in a small multifamily property counted as two units in the numerator in calculating evaluation rating under specified those submitted by the enterprises for both conditions.5 GSEs for 2001, and for Fannie Mae for 2002. special affordable goal performance on all of However, for 1996–2000, HUD’s official the goals for 2001–03. And, above a threshold c. Effects of Changes in the Counting Rules special affordable goal performance figures equal to 60 percent of the average number of on Goal Performance qualifying rental units financed in owner- for Fannie Mae were approximately 1–3 Because of the changes in special occupied properties over the preceding five percentage points lower than the affordable goal counting rules that took effect years, each qualifying unit in a 2–4 unit corresponding figures reported by the in 2001, direct comparisons between official owner-occupied property also counted as two enterprise. This was due to differences goal performance in 2000 and 2001–02 are units in the numerator in calculating goal between HUD and Fannie Mae in the somewhat of an ‘‘apples-to-oranges performance. application of counting requirements comparison.’’ For this reason, the Department • Freddie Mac’s Temporary Adjustment applicable to purchases of portfolios of has calculated what performance would have Factor. Freddie Mac received a ‘‘Temporary seasoned loans, based on a statutory been in 2000 under the 2001–03 rules; this requirement that the proceeds of such GSE may be compared with official performance purchases by the loan sellers should be not take such steps; rather, Fannie Mae excluded in 2001–02—an ‘‘apples-to-apples ‘‘recycled’’ in order for the GSE to receive such loans from the denominator in making its own 3 calculations of its special affordable goal Special Affordable goal credit. This performance. In 1996–2000 HUD counted all 4 See Congressional Record, December 15, 2000, eligible loans in the denominator, and, in the pp. H12295–96. 3 During 1996–2000 Freddie Mac took steps to absence of measures to verify ‘‘recycling’’ by Fannie 5 The revised requirements are codified at 24 CFR acquire representations and warranties from lenders Mae, did not award credit in the numerator of the 81.14(e)(4). The changes are discussed in detail in to attest that they were ‘‘recycling’’ the proceeds special affordable goal for most of Fannie Mae’s the rule preamble, 68 FR 65074–76 (October 31, from the sales of qualifying loans. Fannie Mae did seasoned mortgage purchases. 2000).

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00201 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24428 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

comparison.’’ HUD has also calculated what compared with official performance in These comparisons are presented in Table performance would have been in 2001–02 2000—an ‘‘oranges-to-oranges comparison.’’ C.2. under the 1996–2000 rules; this may be BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00202 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24429

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00203 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.062 24430 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Specifically, Table C.2 shows performance • Freddie Mac. The largest impact of the the housing goals was counted as two units under the special affordable goal in three counting rule changes on Freddie Mac’s goal in the numerator (and one unit in the ways. Baseline A presents performance under performance was due to the application of denominator) in calculating goal performance the counting rules in effect for 1996–2000. the temporary adjustment factor for for that goal. For example, if a GSE financed Baseline B incorporates the technical changes purchases of mortgages on large multifamily a mortgage on a 40-unit property in which 10 in counting rules—changes in the treatment properties, as enacted by Congress; this of the units qualified for the special of missing data (including use of proxy added 1.4 percentage points to goal affordable goal, 20 units would be entered in rents), and changes in procedures related to performance in 2001, as shown in Table C.2. the numerator and 40 units in the the ‘‘recycling’’ requirement. Baseline C Bonus points for purchases of mortgages on denominator for this property in calculating incorporates in addition to the technical small multifamily properties added 1.1 goal performance. changes the bonus points and, for Freddie percentage points to performance, and bonus Fannie Mae financed 37,449 units in small Mac, the temporary adjustment factor. points for purchase of mortgages on owner- multifamily properties in 2001 that were Baseline B corresponds to the counting occupied 2–4 unit rental properties added 0.7 eligible for the special affordable goal, and approach proposed in this rule to take effect percentage point to performance. The 58,277 such units in 2002—a two-year in 2005. Boldface figures under Baseline A remaining impact (0.2 percentage point) was increase of more than 700 percent from the for 1999–2000 and under Baseline C for due to technical changes in counting rules— 7,196 such units financed in 2000. Small 2001–02 indicate official goal performance primarily, the exclusion of single-family multifamily properties also accounted for a based on the counting rules in effect in those units with missing information from the greater share of Fannie Mae’s multifamily years—e.g., for Freddie Mac, 17.2 percent in denominator in calculating goal performance. business in 2001–02—7.4 percent of total 1999, 20.7 percent in 2000, 22.6 percent in Changes in the Department’s counting rules multifamily units financed in 2001 and 13.2 2001, and 21.4 percent in 2002. related to ‘‘recycling’’ did not play a role in percent in 2002, up from 2.5 percent in 2000. • Performance on the Special Affordable Freddie Mac’s performance on the special However, HUD’s 2000 rule reported Housing Goal under 1996–2000 Counting affordable goal. These same patterns also information from the 1991 Residential Rules Plus Technical Changes. If the generally appeared in 2002. Finance Survey that small multifamily ‘‘Baseline B’’ counting approach had been in • Fannie Mae. The temporary adjustment properties accounted for 37 percent of all effect in 2000–02 and the GSEs’ had factor applied to Freddie Mac’s goal multifamily units, thus Fannie Mae was still purchased the same mortgages that they performance, but not to Fannie Mae’s less active in this market than in the market actually did purchase in those years, Fannie performance, thus counting rule changes had for large multifamily properties. Mae would have surpassed the special less impact on its performance than on Within the small multifamily market, there affordable goal in both 2000 and 2001, but Freddie Mac’s performance in 2001. The was no evidence that Fannie Mae targeted not in 2002, while Freddie Mac would have largest impacts of the counting rule changes affordable properties to a greater extent in surpassed the goal in 2000 but fallen short in on Fannie Mae’s goal performance were due 2001–02 than in 2000. That is, 61 percent of both 2001 and 2002. Specifically, Fannie to the application of bonus points for Fannie Mae’s small multifamily units Mae’s performance would have been 21.4 purchases of mortgages on owner-occupied percent in 2000, 20.2 percent in 2001, and qualified for the special affordable goal in 2–4 unit rental properties, which added 0.9 2000; this fell to 46 percent in 2001 and 52 19.9 percent in 2002. Freddie Mac’s percentage point to performance; bonus performance would have been 21.0 percent percent in 2002. points for purchases of mortgages on small Freddie Mac financed 50,299 units in small in 2000, 19.3 percent in 2001, and 18.6 multifamily properties, which added 0.4 percent in 2002. multifamily properties in 2001 that were • percentage point to performance; and eligible for the special affordable goal and Performance on the Special Affordable technical changes, which added 1.6 Housing Goal under 2001–2003 Counting 43,979 such units in 2002, a two-year percentage points to performance—this increase of more than 1300 percent from the Rules. If the 2001–03 counting rules had been included the change in the Department’s in effect in 2000–02 and the GSEs’ had 2,996 such units financed in 2000. Small rules regarding ‘‘recycling’’ and the exclusion multifamily properties also accounted for a purchased the same mortgages that they of single-family units with missing actually did purchase in that year (i.e., significantly greater share of Freddie Mac’s information from the denominator in multifamily business in 2001–02—16.0 abstracting from any behavioral effects of calculating goal performance.6 The use of ‘‘bonus points,’’ for example), both GSEs percent of total multifamily units financed in proxy rents for multifamily properties played 2001 and 13.2 percent in 2002, up from 1.8 would have substantially surpassed the a minor role in determining Fannie Mae’s special affordable goal in all three years, but percent in 2000. special affordable goal performance. These Within the small multifamily market, there both GSEs’ performance figures would have same patterns also appeared in 2002. deteriorated somewhat from 2000 to 2001 was some evidence that Freddie Mac targeted and also from 2001 to 2002. Specifically, d. Bonus Points for the Special Affordable affordable properties to a greater extent in Fannie Mae’s ‘‘Baseline C’’ performance Housing Goal 2001 than in 2000. That is, 55 percent of would have been 22.2 percent in 2000, 21.6 As discussed above and in Appendix A, Freddie Mac’s small multifamily units percent in 2001, and 21.4 percent in 2002. the Department established ‘‘bonus points’’ qualified for the special affordable goal in Freddie Mac’s performance would have been to encourage the GSEs to step up their 2000; this rose to 73 percent in 2001 and 64 23.4 percent in 2000, 22.6 percent in 2001, activity in 2001–03 in two segments of the percent in 2002. and 21.4 percent in 2002. Measured on this mortgage market—the small (5–50 unit) In summary, then, there is evidence that consistent basis, then, Fannie Mae’s multifamily mortgage market, and the market bonus points for small multifamily properties performance fell by 0.6 percentage point in for mortgages on 2–4 unit properties where had an impact on Fannie Mae’s role in this 2001 and 0.2 percentage point in 2002. 1 unit is owner-occupied and 1–3 units are market in 2001–02 and an even larger impact Freddie Mac’s ‘‘Baseline C’’ performance fell occupied by renters. Bonus points did not on Freddie Mac’s role in this market. In by 0.8 percentage point in 2001 and 1.2 apply to purchases of mortgages for owner- addition, Fannie Mae has announced a percent in 2002. These reductions were occupied 1-unit properties, for investor- program to increase its role in this market primarily due to 2001–02 being years of owned 1–4 unit properties, and for large further in future years.7 heavy refinance activity. (>50-unit) properties, although as also Bonus points for single-family rental Details of Effects of Changes in Counting discussed above, a ‘‘temporary adjustment properties. Above a threshold, each unit Rules on Goal Performance in 2001–02. As factor’’ applied to Freddie Mac’s purchases of financed in a 2–4 unit property with at least discussed above, counting rule changes that qualifying mortgages on large multifamily one owner-occupied unit (referred to as took effect in 2001 had significant impacts on properties. ‘‘OO24s’’ below) that qualified for any of the the performance of both GSEs on the special Bonus points for small multifamily housing goals was counted as two units in affordable goal in 2001—3.0 percentage properties. Each unit financed in a small the numerator (and one unit in the points for Fannie Mae and 3.5 percentage multifamily property that qualified for any of denominator) in calculating goal performance points for Freddie Mac. This section breaks down the effects of these changes on goal 6 Exclusion of loans with missing information had 7 ‘‘Fannie Courting Multifamily Sellers; Small performance for both GSEs; results are shown a greater impact on Fannie Mae’s goal performance Banks Balking,’’ American Banker, January 13, in Table C.2. than on Freddie Mac’s goal performance. 2003, p. 1.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00204 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24431

for that goal in 2001–03. The threshold was based on data for mortgagors’ incomes for counties, and the non-metropolitan portions equal to 60 percent of the average number of owner-occupied units, rents for rental units, of States.8 such qualifying units over the previous five area median incomes, and, for units that are For scoring loans purchased by the GSEs years. For example, Fannie Mae financed an in the low-income but not the very low- year-by-year from 1993 through 2002, area average of 24,780 special affordable units in income range, decennial census data used to median income estimates produced by HUD’s these types of properties between 1996 and determine whether the median income for Economic and Market Analysis Division were 2000, and 55,118 such units in 2001. Thus the area where the property is located is in used. The same median income data series Fannie Mae received 40,250 bonus points in the low-income range. Specifically, for described in Appendix A for the Low- and this area in 2001—that is, 55,118 minus 60 single-family owner-occupied units scoring is Moderate-Income Goal was used. The percent of 24,780. So 95,368 units were based on— determination of low-income areas was based entered in the numerator for these properties • The mortgagors’ income at the time of on 1990 census data. in calculating special affordable goal mortgage origination 2005 Procedure. Relative to the above performance. • The median income of an area specified procedure, scoring of loans purchased by the Fannie Mae financed 176,369 units in in the same way as for the Low- and GSEs in and after 2005 will be affected by OO24s that were eligible for the special Moderate-Income Housing Goal, that is: (i) two factors—first, re-benchmarking of area affordable goal in 2001 and 229,827 such For properties located in Metropolitan median incomes to the 2000 census as units in 2002, a two-year increase of nearly 200 percent from the 77,985 such units Statistical Areas (MSAs) the area is the MSA; described in Appendix A, with a shift from financed in 2000. However, Fannie Mae’s and (ii) for properties located outside of 1990 to 2000 census data for identifying low- total single-family business increased at MSAs, the area is the county or the non- income areas, and second, the Office of approximately the same rate as its OO24 metropolitan portion of the State in which Management and Budget’s June, 2003, re- business in 2001 and 2002, thus the share of the property is located, whichever has the specification of MSA boundaries based on this business accounted for by OO24s was larger median income, as of the year of analysis of 2000 census data.9 the same in 2001–02 as in 2000—4 percent. mortgage origination (which may be for the Analysis. For purposes of specifying the Within the OO24 market, there was no current year or a prior year). level of the Special Affordable Housing Goal, • evidence that Fannie Mae targeted special Also, if the property is located in a the HUD estimates of area median incomes affordable properties to a greater extent in Metropolitan Statistical Area (MSA), the for MSAs, non-metropolitan counties, and 2001–02 than in 2000. That is, approximately determination for purposes of the Special the non-metropolitan parts of States, as 30 percent of Fannie Mae’s OO24 units Affordable Housing Goal involves data on described in Appendix A, were used in qualified for the special affordable goal in median income of the MSA; or if the property conjunction with the data identifying low- each of these three years. is located elsewhere, the median income of income areas based on the 2000 census, to re- Freddie Mac financed 96,204 units in the county or the non-metropolitan portion of score loans purchased by the GSEs between OO24s that were eligible for the special the State in which the property is located, 1999 and 2002. The same data series were affordable goal in 2001 and 146,242 such whichever is larger, as of the most recent used further in estimating the share of loans units in 2002, a two-year increase of nearly decennial census. originated in metropolitan areas that would 200 percent from the 49,993 such units Analogous specifications to those detailed in be eligible to score toward the Special financed in 2000. However, Freddie Mac’s Appendix A for the Low- and Moderate- Affordable Housing Goal, from HMDA data. total single-family business increased at Income Housing Goal are applied in the case The results of the retrospective GSE analysis approximately the same rate as its OO24 of the Special Affordable Housing Goal for are provided in Table C.3. The results of the business between 2000 and 2002, thus the rental units in single-family properties with GSE–HMDA comparative analysis are share of this business accounted for by rent data available (assuming no income data presented in the next section. OO24s was the same in 2002 as in 2000—4 available for actual or prospective tenants), BILLING CODE 4210–27–P percent. for rental units in multifamily properties As for Fannie Mae, within the OO24 where rent data are available, and for rental 8 In New England, MSAs were defined through market there was no evidence that Freddie units in multifamily properties where rent Mac targeted special affordable properties to mid-2003 in terms of Towns rather than Counties, data are not available. and the portion of a New England county outside a greater extent in 2001–02 than in 2000. Thus, scoring loans under the Special of any MSA was regarded as equivalent to a county That is, approximately 36 percent of Freddie Affordable Housing Goal requires a data in establishing the metropolitan or non- Mac’s OO24 units qualified for the special series showing annual median incomes for metropolitan location of a property. The MSA affordable goal in each of these three years. MSAs, non-metropolitan counties, and the definitions established by the Office of Management e. Effects of 2000 Census on Scoring of Loans non-metropolitan portions of states; and Budget (OMB) in June, 2003 defined MSAs in New England in terms of counties. Toward the Special Affordable Housing Goal decennial census data on median incomes for 9 HUD has deferred application of the 2003 MSA Background. Scoring of housing units census tracts; and decennial census data on specification to 2005, pending completion of the under the Special Affordable Housing Goal is median incomes for MSAs, non-metropolitan present rulemaking process.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00205 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24432 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00206 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.063 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24433

Table C.3 shows three sets of estimates for established for 1996–2000 as 0.8 percent of each GSE’s total mortgage purchases over the each GSE, based respectively on the counting the total dollar volume of single-family and 1997–99 period. Thus Fannie Mae’s subgoal rules in place in 2001–2002 (but disregarding multifamily mortgages purchased by the was established as $2.85 billion per year and the bonus points and Temporary Adjustment respective GSE in 1994. Thus Fannie Mae’s Freddie Mac’s as $2.11 billion per year. In subgoal was $1.29 billion per year and Factor), on the addition of 2000 census re- 2001 Fannie Mae exceeded its subgoal by Freddie Mac’s subgoal was $988 million per benchmarking and low-income areas, and $4.51 billion and Freddie Mac exceeded its finally on the further addition of 2003 MSA year during that period. Fannie Mae subgoal by $2.54 billion. In 2002, Fannie Mae specification. surpassed the subgoal by $1.08 billion, $1.90 billion, $2.24 billion, $2.77 billion, and $2.50 exceeded its subgoal by $4.72 billion and F. The GSEs’ Multifamily Special Affordable billion in 1996, 1997, 1998, 1999, and 2000 Freddie Mac exceeded its subgoal by $3.11 Purchases respectively, while Freddie Mac exceeded billion. Those subgoals are also in effect for Since 1996 each GSE has been subject to the subgoal by $18 million, $220 million, 2004. Table C.1 includes figures on subgoal an annual dollar-based subgoal for Special $1.70 billion, $1.27 billion, and $1.41 billion. performance, and they are depicted Affordable multifamily mortgage purchases, The subgoal was established for 2001–03 as graphically in Figure C.2. as discussed above. This subgoal was 1.0 percent of the average annual volume of

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00207 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24434 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00208 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.064 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24435

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 and 28.8 percent in 2001 through this entire period, as did their information on the composition of the GSEs’ with a high of 44.0 percent in 1997 and a low purchases of mortgages financing single- Special Affordable purchases according to of 27.8 percent in 1998. Freddie Mac’s family rental units from 1996 through 2000. area income, unit affordability, tenure of unit multifamily purchases represented 29.4 Both GSEs’ high points for mortgages and property type (single- or multifamily). percent of all purchases qualifying for the financing single-family rental units occurred Tables C.4 and C.5 show that each GSE’s goal in 1996 and 27.0 percent in 2001, with in 2001: Fannie Mae’s purchase percentage reliance on multifamily housing units to a high of 31.5 percent in 1997 and a low of was 17.1 percent while Freddie Mac’s was meet the special affordable goal has been 21.6 percent in 1999. The two GSEs’ 17.2 percent.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00209 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24436 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00210 Fmt 4701 Sfmt 4725 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.065 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24437

Tables C.4 and C.5 also show the allocation the family income and area median income of units qualifying for the goal as related to criteria in the goal definition. Very-low-

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00211 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.066 24438 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

income families (shown in the two leftmost affordable housing in terms of household and conforming market. (See Tables A.13 to A.16 columns in the tables) accounted for 80.8 area income characteristics, both GSEs have in Appendix A.). There were two main percent of Fannie Mae’s units qualifying consistently relied substantially more on findings with respect to the special affordable under the goal in 1996, rising to 83.6 percent low-income characteristics of households category. First, Freddie Mac and Fannie Mae in 2001. For Freddie Mac, very-low-income than low-income characteristics of census have historically lagged depositories and the families accounted for 82.1 percent of units tracts to meet this goal. overall market in providing mortgage funds qualifying under the goal in 1996, rising to for special affordable borrowers. Between h. The GSEs’ Performance Relative to the 84.4 percent in 2001. In contrast, mortgage 1993 and 2002, 11.8 percent of Freddie Mac’s Market purchases from low-income areas (shown in mortgage purchases were for special the first and third columns in the tables) Section E.9 in Appendix A uses HMDA affordable borrowers, 12.7 percent of Fannie accounted for 37.0 percent of Fannie Mae’s data and GSE loan-level data for home Mae’s purchases, 15.4 percent of loans units qualifying under the goal in 1996, purchase mortgages on single-family-owner originated by depositories, and 15.4 percent compared to 35.5 percent in 2001. The properties in metropolitan areas to compare of loans originated in the conventional corresponding percentages for Freddie Mac the GSEs’ performance in special affordable conforming market (without estimated B&C were 35.6 percent in 1996 and 35.5 percent lending to the performance of depositories loans). For the recent years, the GSE-market in 2001. Thus given the definition of special and other lenders in the conventional comparisons are as follows:

Market (w/o Year Feddie Mac Fannie Mae B&C) (percent) (percent) (percent)

1999 ...... 12.8 12.5 17.0 2000 ...... 14.7 13.3 16.8 2001 ...... 14.4 14.9 15.6 2002 ...... 15.8 16.3 16.3 1996–2002 (average) ...... 12.8 13.5 16.0 1999–2002 (average) ...... 14.5 14.4 16.4 2001–2002 (average) ...... 15.1 15.6 16.0

During the period between 1999 and 2002, rental units that were financed in the complements Section C of Appendix A, both GSEs’ performance was at conventional conforming market between which presents detailed analyses of housing approximately 88 percent of the market— 1999 and 2002. The GSEs’ 35-percent share problems and demographic trends for lower- special affordable loans accounted for 14.4 of the special affordable market was two- income families which are relevant to the percent of Fannie Mae’s purchases, 14.5 thirds of their 49-percent share of the overall issue addressed in this part of Appendix C. percent of Freddie Mac’s purchases, and 16.4 market. Even in the owner market, where the Data from the American Housing Survey percent of loans originated in the conforming GSEs account for 57 percent of the market, demonstrate that housing problems and market. their share of the special affordable market needs for affordable housing continue to be Second, while both GSEs have improved was only 49 percent during this period. more pressing in the lowest-income their performance over the past few years, While the GSEs improved their market shares categories than among moderate-income Fannie Mae has been made more progress during 2001 and 2002, this analysis shows families, as established in HUD’s analysis for than Freddie Mac in closing its gap with the that the GSEs have not been leading the the 1995 and 2000 Final Rules. Table C.6 market. During the first two years (2001 and single-family market in purchasing loans that displays figures on several types of housing 2002) of HUD’s new housing goal targets, the qualify for the Special Affordable Goal. There problems—high housing costs relative to average share of Fannie Mae’s purchases is room and ample opportunities for the GSEs income, physical housing defects, and going to special affordable loans was 15.6 to improve their performance in purchasing crowding—for both owners and renters. percent, which was close to the market affordable loans at the lower-income end of Figures are presented for households average of 16.0 percent. The share of Freddie the market. Section C.3 of this appendix experiencing multiple (two or more) of these Mac’s purchases going to special affordable discusses a home purchase subgoal designed problems as well as households experiencing loans was 15.1 percent during this period. to place the GSEs in such a leadership a severe degree of either cost burden or Section G in Appendix A discusses the role position in the special affordable single- physical problems. Housing problems in of the GSEs both in the overall special family-owner market. 2001 continued to be much more frequent for affordable market and in the different the lowest-income groups.10 Incidence of segments (single-family owner, single-family Factor 3. National Housing Needs of Low- problems is shown for households in the rental, and multifamily rental) of the special Income Families in Low-Income Areas and income range covered by the special affordable market. The GSEs’ special Very-Low-Income Families affordable goal, as well as for higher income affordable purchases accounted for 35 This discussion concentrates on very-low- households. percent of all special affordable owner and income families with the greatest needs. It BILLING CODE 4210–27–P

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 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00212 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24439

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00213 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.067 24440 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

This analysis shows that priority problems family-owner properties in metropolitan and 19.9 percent in 2002. Freddie Mac’s of severe cost burden or severely inadequate areas. performance would have been 21.0 percent housing are noticeably concentrated among in 2000, 19.3 percent in 2001, and 18.6 1. Severe Housing Problems renters and owners with incomes below 60 percent in 2002. Fannie Mae would have percent of area median income: 30.5 percent The data presented in Section C.3 surpassed the special affordable goal in both of renter households and 34.9 percent of demonstrate that housing problems and 2000 and 2001 while Freddie Mac would owner households had priority problems. In needs for affordable housing are much more have surpassed the goal in 2000 and fallen contrast, in the next higher income range, up pressing in the lowest-income categories than short in 2001. to 80 percent of area median income, 2.5 among moderate-income families. The high The above performance figures are for the percent of renter households and 7.3 percent incidence of severe problems among the special affordable goal defined in terms of of owner households had priority problems. lowest-income renters reflects severe 1990 Census geography. Switching to 2000 The table demonstrates the significance of shortages of units affordable to those renters. Census data slightly increases the coverage of affordability problems: Sixty-five percent of At incomes below 60 percent of area median, special affordable goal, which increases the very-low-income renter families had rent 34.7 percent of renters and 21.6 percent of special affordable share of the GSEs’ burden over 30 percent of income; 35 percent owners paid more than 50 percent of their purchases by up to one percentage point. had rent burden over 50 percent of income. income for housing. In this same income Based on 2000 Census geography, and range, 65.6 percent of renters and 42.4 Thirteen percent had moderately or severely excluding counting requirements (a) and (b), percent of owners paid more than 30 percent inadequate housing; 6 percent lived in then Fannie Mae ’s performance would have of their income for housing. In addition, 31.5 crowded conditions, defined as more than been 21.7 percent in 2000, 20.1 percent in percent of renters and 23.8 percent of owners one person per room. 2001, and 19.4 percent in 2002. Freddie exhibited ‘‘priority problems’’, meaning Mac’s performance would have been 20.8 Factor 4. The Ability of the Enterprises To housing costs over 50 percent of income or percent in 2000, 19.1 percent in 2001, and Lead the Industry in Making Mortgage Credit severely inadequate housing. 17.8 percent in 2002. Available for Low-Income and Very-Low- Homeownership gaps and other disparities in Income Families the housing and mortgage markets discussed b. Single-Family Market Comparisons in Metropolitan Areas The discussion of the ability of Fannie Mae in Section H of Appendix A also apply to and Freddie Mac to lead the industry in Special Affordable housing and mortgages. The Special Affordable Housing Goal is Section G of Appendix A is relevant to this 2. GSE Performance and the Market designed, in part, to ensure that the GSEs factor—the GSEs’ roles in the owner and maintain a consistent focus on serving the a. The GSEs’ Special Affordable Housing rental markets, their role in establishing very low-income portion of the housing Goals Performance widely-applied underwriting standards, their market where housing needs are greatest. role in the development of new technology In the October 2000 rule, the special Section C compared the GSEs’ performance for mortgage origination, their strong staff affordable goal was set at 20 percent for in special affordable lending to the resources, and their financial strength. 2001–03. Effective on January 1, 2001, performance of depositories and other Additional analyses of the potential ability of several changes in counting requirements lenders in the conventional conforming the enterprises to lead the industry in the came into effect for the special affordable market for single-family home loans. The low- and very-low-income market appears goal, as follows: (a) ‘‘Bonus points’’ (double analysis showed that while both GSEs have below in Section D, which explains the credit) for purchases of mortgages on small improved their performance, they have Department’s rationale for the home purchase (5–50 unit) multifamily properties and, above historically lagged depositories and the subgoal for Special Affordable loans. a threshold level, mortgages on 2–4 unit overall market in providing mortgage funds owner-occupied properties; (b) a ‘‘temporary for very low-income and other special Factor 5. The Need To Maintain the Sound adjustment factor’’ (1.35 unit credit) for affordable borrowers. Between 1999 and Financial Condition of the GSEs Freddie Mac’s purchases of mortgages on 2002, special affordable borrowers accounted HUD has undertaken a separate, detailed large (more than 50 unit) multifamily for 14.4 percent of the home loans purchased economic analysis of this final rule, which properties; (c) changes in the treatment of by Fannie Mae, 14.5 percent of Freddie Mac’s includes consideration of (a) the financial missing data; (d) a procedure for the use of purchases, 16.4 percent of home loans returns that the GSEs earn on special imputed or proxy rents for determining goal originated by depositories, and 16.4 percent affordable loans and (b) the financial safety credit for multifamily mortgages; and (e) of all home loans originated in the and soundness implications of the housing changes regarding the ‘‘recycling’’ of funds conventional conforming market (without goals. Based on this economic analysis, HUD by loan originators. Fannie Mae’s B&C loans). Section C also noted that while concludes that the housing goals in this final performance in 2001 was 21.6 percent and both GSEs have improved their performance rule raise minimal, if any, safety and Freddie Mac’s performance was 22.6 percent, over the past few years, Fannie Mae has soundness concerns. thus both GSEs surpassed this higher goal. made more progress than Freddie Mac in Counting requirements (a) and (b) expired closing its gap with the market. During the C. Determination of the Special Affordable at the end of 2003 while (c)–(e) will remain first two years (2001 and 2002) of HUD’s new Housing Goal in effect after that. If this counting housing goal targets, the average share of Several considerations, many of which are approach—without the bonus points and the Fannie Mae’s purchases going to special reviewed in Appendixes A and B and in ‘‘temporary adjustment factor’’—had been in affordable loans was 15.6 percent, which was previous sections of this Appendix, led to the effect in 2000–2002, and the GSEs’ had close to the market average of 16.0 percent. determination of the Special Affordable purchased the same mortgages that they The share of Freddie Mac’s purchases going Housing Goal, the multifamily special actually did purchase in both years, then to special affordable loans was 15.1 percent affordable subgoal, and the special affordable Fannie Mae’s performance would have been during this period. (See Figure C.3.) subgoal for home purchase loans on single- 21.4 percent in 2000, 20.2 percent in 2001, BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00214 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24441

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00215 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.068 24442 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C the largest difference was in 2002, when GSEs have significantly improved their 3. Ability To Lead the Single-Family Owner Fannie Mae’s performance was 15.8 percent performance, according to numerous studies Market: A Special Affordable Sub Goal with the projected data, compared with 16.3 by the Department and independent percent with the historical data. The 1999– researchers, they have historically lagged the The Secretary believes the GSEs can play 2002 average performance for Freddie Mac primary market in providing funds for a leadership role in the special affordable was 14.1 percent with the projected data, special affordable borrowers (see above GSE- market. Thus, the Department is proposing to versus 14.5 percent with the historical data; market comparisons). The type of establish a subgoal of 17 percent for each the largest difference was also in 2002, when improvement needed to meet this new GSE’s purchases of home purchase loans for Freddie Mac’s performance was 15.1 percent special affordable subgoal was demonstrated special affordable families in the single- with the projected data, compared with 15.8 by Fannie Mae during 2001 and 2002. family-owner market of metropolitan areas percent with the historical data. Thus, the Between 2000 and 2001, special affordable for 2005, rising to 18 percent in 2006, and 19 increases in each GSE’s performance needed loans declined as a percentage of Freddie percent in both 2007 and 2008. The purpose to meet the proposed special affordable home Mac’s purchases (from 14.7 to 14.4 percent) of this subgoal is to encourage the GSEs to purchase subgoal in 2005–08 will be slightly and as a percentage of primary market improve their purchases of mortgages for higher than those noted above. originations (from 16.8 to 15.6 percent), but very-low-income and minority first-time The approach taken is for the GSEs to they increased as a percentage of Fannie homebuyers who are expected to enter the obtain their leadership position by staged Mae’s purchases (from 13.3 to 14.9 percent). housing market over the next few years. If the increases in the special affordable subgoal; During 2002, Fannie Mae further increased GSEs meet this goal, they will be leading the this will enable the GSEs to take new its special affordable share (from 14.9 percent primary market by approximately one-half initiatives in a correspondingly staged tin 2001 to 16.3 percent in 2002), placing it percentage point in 2005 and 2.5 percentage manner to achieve the new subgoal each at the market level. This subgoal is designed points by 2007 and 2008, based on the year. Thus, the increases in the special to encourage Fannie Mae as well as Freddie income characteristics of home purchase affordable subgoal are sequenced so that the Mac to lead the special affordable market. loans reported in HMDA. HMDA data show GSEs can gain experience as they improve (3) Disparities in Homeownership and that special affordable families accounted for and move toward the new higher subgoal Credit Access Remain. There remain an average of 16.4 percent of single-family- targets. troublesome disparities in our housing and owner loans originated in the conventional The subgoal applies only to the GSEs’ mortgage markets, even after the ‘‘revolution conforming market of metropolitan areas purchases in metropolitan areas because the in affordable lending’’ and the growth in between 1999 and 2002—the special HMDA-based market benchmark is only homeownership that has taken place since affordable market share was 16.0 percent for available for metropolitan areas. HMDA data the mid-1990s. The homeownership rate for both the longer 1996–2002 period and the for non-metropolitan counties are not reliable African-American and Hispanic households shorter 2001–2002 period. Loans in the B&C enough to serve as a market benchmark. The remains 25 percentage points below that of portion of the subprime market are not Department is also setting home purchase white households. Minority families face included in these averages. As explained in subgoals for the other two goals-qualifying many barriers in the mortgage market, such Appendix D, HUD also projected special categories, as explained in Appendices A and as lack of capital for down payment and lack affordable shares for the market for 1999 to 2002 using the new 2000 Census geography B. Sections E.9 and G of Appendix A provide of access to mainstream lenders (see above). and the new OMB specifications. For special additional information on the opportunities Immigrants and minorities—many of whose affordable loans, the 1999–2002 market for an enhanced GSE role in the special very-low-income levels will qualify them as average using these projected data was also affordable segment of the home purchase special affordable—are projected to account 16.4 percent. market and on the ability of the GSEs to lead for almost two-thirds of the growth in the To reach the proposed 17-percent subgoal that market. number of new households over the next ten for 2005, both GSEs will have to improve The preamble and Appendix A discuss in years. As emphasized in Appendix A, their performance—Fannie Mae by 2.6 some detail the factors that the Department changing population demographics will percentage points over its average considered when setting the subgoal for low- result in a need for the primary and performance of 14.4 percent between 1999 and moderate-income loans. Several of the secondary mortgage markets to meet and 2002, by 1.4 percentage points over its considerations were general in nature—for nontraditional credit needs, respond to average performance of 15.6 percent during example, related to the GSEs’ overall ability diverse housing preferences, and overcome 2001 and 2002, and by 0.7 percentage point to lead the single-family-owner market— information and other barriers that many over its 16.3 percent performance in 2002; while others were specific to the low-mod immigrants and minorities face. The GSEs and Freddie Mac by 2.5 percentage points subgoal. Because the reader can refer to have to increase their efforts in helping over its average performance of 14.5 percent Appendix A, this appendix provides a briefer special affordable families—but so far they between 1999 and 2002, by 1.9 percentage discussion of the more general factors. The have played a surprisingly small role in points over its average performance of 15.1 specific considerations that led to the subgoal serving minority first-time homebuyers. It is percent during 2001 and 2002, and by 1.2 for special affordable loans can be organized estimated that the GSEs accounted for 46.5 percentage point over its 15.8 percent around the following four topics: percent of all (both government and performance in 2002. By 2007–2008 the (1) The GSEs have the ability to lead the conventional) home loans originated between required increases in subgoal performance market. As discussed in Appendix A, the 1999 and 2001; however, they accounted for over past performance will be 2 percentage GSEs have the ability to lead the primary only 14.3 percent of home loans originated points higher than the increases cited in the market for single-family-owner loans, which for African-American and Hispanic first-time preceding sentence. For example, Fannie is their ‘‘bread-and-butter’’ business. Both homebuyers. A subgoal for special affordable Mae would have to increase its performance GSEs have been dominant players in the home purchase loans should increase the by 2.7 percentage points over its 16.3 percent home purchase market for years, funding 57 GSEs’ efforts in important sub-markets such performance in 2002; and Freddie Mac percent of the single-family-owner mortgages as the one for minority first-time would have to increase its performance by financed between 1999 and 2002. Through homebuyers. 3.2 percentage points over its 15.8 percent their many new product offerings and their (4) There are ample opportunities for the performance in 2002. The special affordable various partnership initiatives, the GSEs have GSEs to improve their performance. Special performances of Fannie Mae and Freddie shown that they have the capacity to reach affordable mortgages are available for the Mac were also projected to take into account out to very-low-income and other special GSEs to purchase, which means they can the new 2000 Census geography and the new affordable borrowers. They also have the staff improve their performance and lead the OMB specifications. On average, the results expertise and financial resources to make the primary market in purchasing loans for these with the new data were similar to the old extra effort to lead the primary market in very-low-income borrowers. Sections B, C, data, but the differential was higher during funding single-family-owner mortgages for and I of Appendix A and Section H of 2002. For home purchase loans, the 1999– special affordable borrowers. Appendix D explain that the special 2002 average performance for Fannie Mae (2) The GSEs have lagged the market. Even affordable lending market has shown an was 14.3 percent with the projected data, though they have the ability to lead the underlying strength over the past few years versus 14.4 percent with the historical data; market, they have not done so. While the that is unlikely to vanish (without a

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00216 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24443

significant increase in interest rates or a the special affordable market averaged 28 Section C compared the GSEs’ role in the decline in the economy). The special percent. HUD is well aware of the volatility overall market with their role in the special affordable share of the home purchase market of mortgage markets and the possible impacts affordable market. The GSEs’ purchases has averaged 16.0 percent since 1996 and on the GSEs’ ability to meet the housing provided financing for 23,580,594 dwelling annually has ranged from 15.0 percent to goals. Should conditions change such that units, which represented 49 percent of the 17.0 percent. Second, the market share data the goals are no longer reasonable or feasible, 48,270,415 single-family and multifamily reported in Table A.30 of Appendix A the Secretary has the authority to revise the units that were financed in the conventional demonstrate that there are newly-originated goals. conforming market between 1999 and 2002. loans available each year for the GSEs to However, in the special affordable part of the 5. The Special Affordable Housing Goal for purchase. The GSEs’ purchases of single- market, the 4,595,201 units that were 2005–2008 family owner loans represented 57 percent of financed by GSE purchases represented only all single-family-owner loans originated The proposed Special Affordable Housing 35 percent of the 13,232,549 dwelling units between 1999 and 2002, compared with 49 Goal for 2005 is 22 percent of eligible that were financed in the market. Thus, there percent of the special affordable loans that purchases, a two percentage point increase appears to ample room for the GSEs to were originated during this period. Thus, half over the current goal of 20 percent, with the improve their performance in the special of the special affordable conforming market proposed goal rising to 24 percent in 2006, affordable market. In addition, there are is not touched by the GSEs. As noted above, 26 percent in 2007, and 28 percent in 2008. several market segments (e.g., first-time the situation is even more extreme for special The bonus points for small multifamily homebuyers) that would benefit from a sub-markets such the minority first-time properties and owner-occupied 2–4 units, as greater secondary market role by the GSEs, homebuyer market where the GSEs have only well as Freddie Mac’s Temporary Adjustment and special affordable borrowers are a minimal presence. Between 1999 and 2001, Factor, will no longer be in effect for goal concentrated these markets. counting purposes. It is recognized that the GSEs purchased only 33 percent of 6. Multifamily Special Affordable Subgoals conventional conforming loans originated for neither GSE would have met the 22-percent target in the past three years. Under the new Based on the GSEs’ past performance on minority first-time homebuyers, even though counting rules, Fannie Mae’s special the special affordable multifamily subgoals, they purchased 57 percent of all home loans affordable performance is estimated to have and on the outlook for the multifamily originated in the conventional conforming been 18.6 percent in 1999, 21.7 percent in mortgage market, HUD is proposing that market during that period. But also 2000, 20.1 percent in 2001, and 19.4 percent these subgoals be retained and increased for important, the GSEs’ purchases under the in 2002—Fannie Mae would have to increase the 2005–2008 period. Unlike the overall subgoal are not limited to new mortgages that its performance in 2005 by 2.0 percentage goals, which are expressed in terms of are originated in the current calendar year. points over its average (unweighted) minimum goal-qualifying percentages of total The GSEs can purchase loans from the performance of 20.0 percent over these last units financed, these subgoals for 2001–03 substantial, existing stock of special four years. By 2008 this increase relative to and in prior years have been expressed in affordable loans held in lenders’ portfolios, average 1999–2002 performance would be 8.0 terms of minimum dollar volumes of goal- after these loans have seasoned and the GSEs percentage points. Freddie Mac’s qualifying multifamily mortgage purchases. have had the opportunity to observe their performance is projected to have been 17.4 Specifically, each GSE’s special affordable payment performance. In fact, based on percent in 1999, 20.8 percent in 2000, 19.1 multifamily subgoal is currently equal to 1.0 Fannie Mae’s recent experience, the purchase percent in 2001, and 17.8 percent in 2002— percent of its average total (single-family plus of seasoned loans appears to be one useful Freddie Mac would have to increase its multifamily) mortgage volume over the 1997– strategy for purchasing goals-qualifying performance in 2005 by 3.2 percentage points 99 period. Under this formulation, in October loans. over its average (unweighted) performance of 2000 the subgoals were set at $2.85 billion To summarize, although single-family- 18.8 percent over these last four years. By per year for Fannie Mae and $2.11 billion per owner mortgages comprise the ‘‘bread-and- 2008 this increase relative to average 1999– year for Freddie Mac, in each of calendar butter’’ of their business, the GSEs have 2002 performance would be 9.2 percentage years 2001 through 2003. These represented lagged behind the primary market in points. As explained in Appendix D, the increases from the goals for 1996–2000, financing special affordable loans. For the Special Affordable market averaged 28 which were $1.29 billion annually for Fannie reasons given above, the Secretary believes percent between 1999 and 2002. Thus, the Mae and $0.99 billion annually for Freddie that the GSEs can do more to raise the special GSEs should be able to improve their Mac. These subgoals are also in effect for affordable shares of the home loans they performance enough to meet the proposed 2004. purchase on single-family-owner properties. targets of 22 percent in 2005, 24 percent in HUD’s Determination. The multifamily This can be accomplished by building on 2006, 26 percent in 2007, and 28 percent in mortgage market and both GSEs’ multifamily efforts that the enterprises have already 2008. transactions volume grew significantly over started, including their new affordable The objective of HUD’s proposed Special the 1993–2001 period, indicating that both lending products aimed at special groups Affordable Goal is to bring the GSEs’ enterprises have provided increasing support such as first-time homebuyers, their many performance to the upper end of HUD’s for the multifamily market, and that they partnership efforts, their outreach to inner market range estimate for this goal (24–28 have the ability to continue to provide city neighborhoods, their incorporation of percent), consistent with the statutory further support for the market. greater flexibility into their underwriting criterion that HUD should consider the GSEs’ Specifically, Fannie Mae’s total eligible guidelines, and their purchases of seasoned ability to lead the market for each Goal. To multifamily mortgage purchase volume CRA loans. A wide variety of quantitative enable the GSEs to achieve this leadership, increased from $4.6 billion in 1993 to $12.5 and qualitative indicators indicate that the the Department is proposing modest billion in 1998, and then jumped sharply to GSEs’ have the resources and financial increases in the Special Affordable Goal for $18.7 billion in 2001 and $18.3 billion in strength to improve their special affordable 2005 which will increase further, year-by- 2002. Its special affordable multifamily performance enough to lead the market. year through 2008, to achieve the ultimate mortgage purchases followed a similar path, objective for the GSEs to lead the market rising from $1.7 billion in 1993 to $3.5 4. Size of the Overall Special Affordable under a range of foreseeable economic billion in 1998 and $4.1 billion in 1999, and Mortgage Market circumstances by 2008. Such a program of also jumping sharply to $7.4 billion in 2001 As detailed in Appendix D, single-family staged increases is consistent with the and $7.6 billion in 2002. As a result of its and multifamily special affordable mortgages statutory requirement that HUD consider the strong performance, Fannie Mae’s purchases are estimated to account for 24–28 percent of past performance of the GSEs in setting the have been at least twice its minimum subgoal the dwelling units financed by conventional Goals. Staged annual increases in the Special in every year since 1997—247 percent of the conforming mortgages; in estimating the size Affordable Goal will provide the enterprises subgoal in that year, 274 percent in 1998, 315 of the market, HUD used alternative with opportunity to adjust their business percent in 1999, 294 percent in 2000, and, assumptions about future economic and models and prudently try out business under the new higher subgoal level, 258 market affordability conditions that were less strategies, so as to meet the required 2008 percent in 2001, and 266 percent in 2002. favorable than those that existed over the level without compromising other business Freddie Mac’s total eligible multifamily past several years. Between 1999 and 2002, objectives and requirements. mortgage purchase volume increased even

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00217 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24444 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

more sharply, from $0.2 billion in 1993 to achievable. Finally, HUD is proposing to model’s parameters. In their comments on $6.6 billion in 1998, and then jumped establish a subgoal of 17 percent for the the 2000 Proposed GSE Rule, both Fannie sharply in 2001 to $11.8 billion and $13.3 GSEs’ purchases of single-family-owner Mae and Freddie Mac stated that HUD’s billion in 2002. Its special affordable mortgages that qualify for the special market share model (outlined in Section B multifamily mortgage purchases followed a affordable goal and are originated in below) was a reasonable approach for similar path, rising from $0.1 billion in 1993 metropolitan areas, for 2005, with this estimating the goals-qualifying (low-mod, to $2.7 billion in 1998, and also jumping subgoal rising to 18 percent in 2006, and 19 special affordable, and underserved areas) sharply to $4.6 billion in 2001 and $5.2 percent in both 2007 and 2008. The Secretary shares of the mortgage market. Freddie Mac billion in 2002. As a result of its strong has considered the GSEs’ ability to lead the stated: performance, Freddie Mac’s purchases have industry as well as the GSEs’ financial We believe the Department takes the also been at least twice its minimum subgoal condition. The Secretary has determined that correct approach in the Proposed Rule by in every year since 1998—272 percent of the the proposed goals, the proposed multifamily examining several different data sets, using subgoal in that year, 229 percent in 1999, 243 subgoals, and the proposed single-family- alternative methodologies, and conducting percent in 2000, and, under the new higher owner subgoals are necessary and sensitivity analysis. We applaud the subgoal level, 220 percent in 2001, and 247 appropriate. Department’s general approach for percent in 2002. addressing the empirical challenges.2 The Special Affordable Housing Appendix D—Estimating the Size of the Similarly, Fannie Mae stated that ‘‘HUD Multifamily Subgoals set forth in this Conventional Conforming Market for has developed a reasonable model for proposed rule are reasonable and appropriate Each Housing Goal assessing the size of the affordable housing 3 based on the Department’s analysis of this A. Introduction market’’. market. The Department’s decision to retain However, both GSEs have criticized HUD’s In establishing the three housing goals, the the multifamily subgoal is based on the fact implementation of its market methodology. Secretary is required to assess, among a that HUD’s analysis indicates that Their major criticisms and HUD’s responses number of factors, the size of the multifamily housing still serves the housing to their criticisms can be found in Section B conventional market for each goal. This needs of lower-income families and families of Appendix D of the 2000 Final Rule. HUD appendix explains HUD’s methodology for in low-income areas to a greater extent than recognizes that there is no single, perfect data single-family housing. By retaining the estimating the size of the conventional set for estimating the size of the affordable multifamily subgoal, the Department ensures market for each of the three housing goals. lending market and that available data bases that the GSEs continue their activity in this Following this overview, Section B on different sectors of the market must be market, and that they achieve at least a summarizes the main components of HUD’s combined in order to implement its market minimum level of special affordable market-share model and identifies those share model (as outlined in Section B below). multifamily mortgage purchases that are parameters that have a large effect on the As this appendix will show, HUD has affordable to lower-income families. The relative market shares. Sections C and D carefully combined various mortgage market Department proposes to establish each GSE’s discuss two particularly important market data bases in a manner which draws on the special affordable multifamily subgoal as 1.0 parameters, the size of the multifamily strength of each in order to implement its percent of its average annual dollar volume market and the share of the single-family market methodology and to arrive at a of total (single-family and multifamily) mortgage market accounted for by single- reasonable range of estimates for the three mortgage purchases over the 2000–2002 family rental properties. Section E provides goals-qualifying shares of the mortgage period. In dollar terms, the Department’s a more systematic presentation of the model’s market. In this appendix, HUD demonstrates proposal is $5.49 billion per year in special equations and main assumptions. Sections F, the robustness of its market estimates by affordable multifamily mortgage purchases G, and H report HUD’s estimates for the Low- reporting the results of numerous sensitivity for Fannie Mae, and $3.92 billion per year in and Moderate-Income Goal, the Underserved analyses that examine a range of assumptions special affordable multifamily mortgage Areas Goal, and the Special Affordable about the relative importance of the rental purchases for Freddie Mac. These subgoals Housing Goal, respectively. and owner markets and the goals-qualifying would be less than actual special affordable In developing this rule, HUD has followed shares of the owner portion of the mortgage multifamily mortgage purchase volume in the same basic approach that it followed in market. 2001 and 2002 for both GSEs; thus the the last two GSE rules. HUD has carefully This appendix reviews in some detail Department believes that they would be reviewed existing information on mortgage HUD’s efforts to combine information from feasible for the 2005–2008 period. activity in order to understand the weakness several mortgage market data bases to obtain of various data sources and has conducted reasonable values for the model’s parameters. 7. Conclusion sensitivity analyses to show the effects of The next section provides an overview of HUD has determined that the Special alternative parameter assumptions. HUD is HUD’s market share model. Affordable Housing Goal in this proposed well aware of uncertainties with some of the B. Overview of HUD’s Market Share rule addresses national housing needs within data and much of this appendix is spent Methodology 4 the income categories specified for this goal, discussing the effects of alternative while accounting for the GSEs’ past assumptions about data parameters and 1. Definition of Market Share presenting the results of an extensive set of performance in purchasing mortgages The size of the market for each housing meeting the needs of very-low-income sensitivity analyses. In an earlier critique of HUD’s market share goal is one of the factors that the Secretary families and low-income families in low- is required to consider when setting the level income areas. HUD has also considered the model, Blackley and Follain (1995, 1996) concluded that conceptually HUD had size of the conventional mortgage market 2 serving very-low-income families and low- chosen a reasonable approach to determining See Freddie Mac, ‘‘Comments on Estimating the Size of the Conventional Conforming Market for income families in low-income areas. the size of the mortgage market that qualifies 1 Each Housing Goal: Appendix III to the Comments Moreover, HUD has considered the GSEs’ for each of the three housing goals. Blackley of the Federal Home Loan Mortgage Corporation on ability to lead the industry as well as their and Follain correctly note that the challenge HUD’s Regulation of the Federal National Mortgage financial condition. HUD has determined lies in getting accurate estimates of the Association (Fannie Mae) and the Federal Home that a Special Affordable Housing Goal of 22 Loan Mortgage Corporation (Freddie Mac)’’, May 8, percent in 2005, 24 percent in 2006, 26 1 Dixie M. Blackley and James R. Follain, ‘‘A 2000, page 1. percent in-2007, and 28 percent in 2008 is Critique of the Methodology Used to Determine 3 See Fannie Mae, ‘‘Fannie Mae’s Comments on both necessary and achievable. HUD has also Affordable Housing Goals for the Government HUD’s Regulation of the Federal National Mortgage determined that a multifamily special Sponsored Housing Enterprises,’’ unpublished Association (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac)’’, May 8, affordable subgoal for 2005–2008 set at 1.0 report prepared for Office of Policy Development and Research, Department of Housing and Urban 2000, page 53. percent of the average of each GSE’s Development, October 1995; and ‘‘HUD’s Market 4 Readers not interested in this overview may respective dollar volume of combined Share Methodology and its Housing Goals for the want to proceed to Section C, which begins the (single-family and multifamily) 1999–2001 Government Sponsored Enterprises,’’ unpublished market analysis by examining the size of the mortgage purchases in is both necessary and paper, March 1996. multifamily market.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00218 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24445

of each housing goal.5 Using the Low- and moderate-income market is defined as a Step 2: Projecting the ‘‘goal percentage’’ for Moderate-Income Housing Goal as an percentage of the conforming market; this each of the above four property types (for example, the market share in a particular percentage approach maintains consistency example, the ‘‘Low- and Moderate-Income year is defined as follows: with the method for computing each GSE’s Goal percentage for single-family owner- Low- and Moderate-Income Share of performance under the Low- and Moderate- occupied properties’’ is the percentage of Market: The number of dwelling units Income Goal (that is, the number of low- and those dwelling units financed by mortgages financed by the primary mortgage market in moderate-income dwelling units financed by in a particular year that are occupied by a particular calendar year that are occupied GSE mortgage purchases relative to the households with incomes below the area by (or affordable to, in the case of rental overall number of dwelling units financed by median). units) families with incomes equal to or less GSE mortgage purchases). Step 3: Multiplying the four percentages in than the area median income divided by the 2. Three-Step Procedure (2) by their corresponding market shares in total number of dwelling units financed in Ideally, computing the low- and moderate- (1), and summing the results to arrive at an the conforming conventional primary estimate of the overall share of dwelling units mortgage market. income market share would be straightforward, consisting of three steps: financed by mortgages that are occupied by There are three important aspects to this Step 1: Projecting the market shares of the low- and moderate-income families. definition. First, the market is defined in four major property types included in the The four property types are analyzed terms of ‘‘dwelling units’’ rather than, for conventional conforming mortgage market, separately because of their differences in example, ‘‘value of mortgages’’ or ‘‘number of i.e.— low- and moderate-income occupancy. properties.’’ Second, the units are ‘‘financed’’ (a) Single-family owner-occupied dwelling Rental properties have substantially higher units rather than the entire stock of all units (SF–O units); percentages of low- and moderate-income mortgaged dwelling units; that is, the market- (b) Rental units in 2–4 unit properties occupants than owner-occupied properties. share concept is based on the mortgage flow where the owner occupies one unit (SF 2–4 This can be seen in the top portion of Table in a particular year, which will be smaller units); 7 D.1, which illustrates Step 3’s basic formula than total outstanding mortgage debt. Third, (c) Rental units in one-to-four unit for calculating the size of the low- and the low- and moderate-income market is investor-owned properties (SF Investor moderate-income market.9 In this example, expressed relative to the overall conforming units); and, low- and moderate-income dwelling units are conventional market, which is the relevant (d) Rental units in multifamily (5 or more estimated to account for 53.9 percent of the market for the GSEs.6 The low- and 8 units) properties (MF units). total number of dwelling units financed in the conforming mortgage market. 5 Sections 1332(b)(4), 1333(a)(2), and 1334(b)(4). 7 The owner of the SF 2–4 property is counted in BILLING CODE 4210–27–P 6 So-called ‘‘jumbo’’ mortgages, greater than (a). $300,700 in 2002 for 1-unit properties, are excluded 8 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 9 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 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00219 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24446 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00220 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.069 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24447

To examine the other housing goals, the uncertainty surrounding estimates of the C. Size of the Conventional Multifamily ‘‘goal percentages’’ in Step 2 would be multifamily mortgage market, it is important Mortgage Market 12 changed and the new ‘‘goal percentages’’ to consider a range of market estimates, as This section provides estimates of (a) the would be multiplied by Step 1’s property Sections F–H do. annual dollar volume of conventional distribution, which remains constant. For Goal Percentages. To derive the goal multifamily mortgage originations and (b) the 10 example, the Underserved Areas Goal percentages for each property type, HUD annual average loan amount per unit would be derived as illustrated in the bottom relied heavily on HMDA, AHS, and POMS financed. The estimates build on research portion of Table D.1. In this example, units data. For single-family-owner originations, reported in the Final Rule on HUD’s eligible under the Underserved Areas Goal Regulation of Fannie Mae and Freddie Mac are estimated to account for 31.4 percent of HMDA provides comprehensive information on borrower incomes and census tract as published in the Federal Register on the total number of dwelling units financed October 31, 2000, especially in Appendix D. 11 locations for metropolitan areas. in the conforming mortgage market. That material from the 2000 Rule will not be Unfortunately, it provides no information on 3. Data Issues repeated here but will be referenced or the incomes of renters living in mortgaged summarized where appropriate. Unfortunately, complete and consistent properties (either single-family or The section uses the information on dollar mortgage data are not readily available for multifamily) or on the rents (and therefore volume of multifamily originations and carrying out the above three steps. A single the affordability) of rental units in mortgaged average loan amounts to estimate the number data set for calculating either the property properties. The AHS, however, does provide of multifamily units financed each year as a shares or the housing goal percentages does a wealth of information on rents and the percentage share of the total (both single- not exist. However, there are several major family and multifamily) number of dwelling data bases that provide a wealth of useful affordability of the outstanding stock of units financed each year; the years covered information on the mortgage market. HUD single-family and multifamily rental include 1991 to 2002. This percentage share, combined information from the following properties. An important issue here concerns called the ‘‘multifamily mix’’, is an important sources: the Home Mortgage Disclosure Act whether rent data for the stock of rental parameter in HUD’s projection model of the (HMDA) reports, the American Housing properties can serve as a proxy for rents on mortgage market for 2005–08. Survey (AHS), HUD’s Survey of Mortgage newly-mortgaged rental properties. During Estimating this ‘‘multifamily mix’’ is Lending Activity (SMLA), Property Owners the 2000 rule-making process, POMS data important because relative to its share of the and Managers Survey (POMS) and the were used to examine the rents of newly- overall housing market, the multifamily Census Bureau’s Residential Finance Survey mortgaged rental properties; thus, the POMS rental sector has disproportionate importance (RFS). In addition, information on the data supplements the AHS data. The data for the housing goals established for Fannie mortgage market was obtained from the base issues as well as other technical issues Mae and Freddie Mac. This is because most Mortgage Bankers Association, Fannie Mae, related to the goal percentages (such as the multifamily rental units are occupied by Freddie Mac and other organizations. need to consider a range of mortgage market households with low or moderate incomes. Property Shares. To derive the property In 2001, for example, Freddie Mac purchased shares, HUD started with forecasts of single- environments) are discussed in Sections F, G, mortgages on approximately 3.5 million family mortgage originations (expressed in and H, which present the market share housing units, of which only 12 percent were dollars). These forecasts, which are available estimates for the Low- and Moderate-Income multifamily rental units. However, of Freddie from the GSEs and industry groups such as Goal, the Underserved Areas Goal, and the Mac’s purchases qualifying as mortgages on the Mortgage Bankers Association, do not Special Affordable Goal, respectively. low- and moderate-income housing, fully 25 provide information on conforming percent of the units financed were mortgages, on owner versus renter mortgages, 4. Conclusions multifamily rental units. Fannie Mae’s or on the number of units financed. Thus, to HUD is using the same basic methodology experience is similar. Ten percent of all estimate the number of single-family units for estimating market shares that it used in housing units on which mortgages were financed in the conforming conventional 1995 and 2000. As demonstrated in the purchased in 2001 were multifamily rental market, HUD had to project certain market remainder of this appendix, HUD has units, but 21 percent of the units with parameters based on its judgment about the attempted to reduce the range of uncertainty qualifying mortgages were multifamily reliability of different data sources. Sections around its market estimates by carefully rentals. D and E report HUD’s findings related to the reviewing all known major mortgage data The methods used in the 2000 Rule for single-family market. sources and by conducting numerous estimating the size of the multifamily Total market originations are obtained by mortgage market and related variables were adding multifamily originations to the single- sensitivity analyses to show the effects of the product of extensive research by HUD family estimate. Because of the wide range of alternative assumptions. Sections C, D, and E and review by interested parties. The estimates available, the size of the report findings related to the property share approach here is first to extend those multifamily mortgage market turned out to be distributions called for in Step 1, while estimates through 2002 using the same one of the most controversial issues raised Sections F, G, and H report findings related methods as in the 2000 Rule, and then to during the initial rule-making process during to the goal-specific market parameters called present alternative methods, along with 1995; this was also an issue that the GSEs for in Step 2. These latter sections also report commentary. focused on in their comments on the 2000 the overall market estimates for each housing proposed rule. Because most renters qualify goal calculated in Step 3. 1. Data Sources under the Low- and Moderate-Income Goal, In considering the levels of the goals, HUD The data sources available for estimating the chosen market size for multifamily can carefully examined past comments by the the size of the multifamily mortgage market have a substantial effect on the overall GSEs and others on the methodology used to are more limited in scope and timeliness estimate of the low- and moderate-income establish the market share for each of the than was the case for the 2000 Rule. Among market (as well as on the estimate of the goals. Based on that thorough evaluation, as the key sources described in detail in the special affordable market). Thus, it is 2000 Rule, the following are now less useful: important to consider estimates of the size of well as HUD’s additional analysis for this Survey of Mortgage Lending Activity. This the multifamily market in some detail, as Proposed Rule, HUD concludes that its basic survey has been discontinued; estimates are Section C does. In addition, given the methodology is a reasonable and valid approach to estimating market shares. As in available only through 1997. Residential Finance Survey: The 1991 10 the past, HUD recognizes the uncertainty This goal will be referred to as the Residential Finance Survey (RFS) is now 13 regarding some of these estimates, which has ‘‘Underserved Areas Goal’’. years out of date. 11 The example in Table D.1 is based on 1990 led the Department to undertake a number of Urban Institute Statistical Model: This Census tract geography. As explained in Section G, sensitivity and other analyses to reduce this model, developed in 1995 and calibrated switching to 2000 Census tract geography uncertainty and also to provide a range of (scheduled for 2005) increases the underserved areas market share by approximately five market estimates (rather than precise point 12 This section is based on analysis by Jack percentage points. estimates) for each of the housing goals. Goodman under contract with the Urban Institute.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00221 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24448 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

using data from 1975–1990, is now even 2. Estimates Based on ‘‘HUD New’’ Despite these limitations, HUD New is one further removed from its calibration period Methodology sound way to estimate the size of the and probably captures current market In the 2000 Rule, HUD developed a new multifamily conventional mortgage market. conditions less well. methodology for estimating aggregate The method requires unavoidable judgment Estimates from the GSEs: As part of their multifamily conventional loan originations. calls on which analysts will differ. However, comments on the proposed 2000 Rule, The method, here labeled ‘‘HUD New’’, was due to the reasonableness of the HUD New Fannie Mae and Freddie Mac shared with developed to make full use of the available approach, the value of maintaining HUD their own estimates of the size of the data, and in particular the four sources listed continuity in estimation methods, and the multifamily mortgage market. above, which encompass most of the fact that no data has become available in the Fortunately, several key sources are multifamily mortgage market. past few years that would argue for available with the timeliness and quality The advantages of HUD New are that it modifying HUD New, it is used here for the comparable to the sources used during provides reasonably complete coverage of the baseline estimate of the size of the development of the 2000 Rule. These sources market, produces those estimates within nine conventional multifamily mortgage market in are: the Home Mortgage Disclosure Act months of the end of the year, generally 2000, 2001 and 2002. (HMDA); activity reports submitted to HUD includes only current originations and avoids The estimates from HUD New are double counting. The main disadvantage of and the Office of Federal Enterprise presented in Table D.2. This table is the HUD New is that it produces a lower bound Oversight (OFHEO) by Fannie Mae and counterpart of Table D.5 in the 2000 Rule. estimate. Some loan originators are missed, Freddie Mac; non-GSE mortgage-backed including pension funds, government entities The historical years have two columns each, security issuance from the Commercial at the federal, state, and local levels, real one for the estimates presented in the 2000 Mortgage Alert database; and multifamily estate investment trusts, and some mortgage Rule and one for estimates independently mortgage activity by life insurance bankers. Also excluded are loans made by produced as part of this research. Footnotes companies, as estimated by the American private individuals and partnerships. In to the table provide more complete Council of Life Insurers (ACLI). For addition to these exclusions, estimates from descriptions of the components. Additional background information on each of these the covered lenders require some judgmental background on the calculations is provided sources, readers are referred to Appendix D adjustments to conform to the definitions and in the 2000 Rule (Appendix D, Section C). of the 2000 Rule. time intervals of HUD New. BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00222 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24449

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00223 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.070 24450 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

The revisions to the historical estimates requirements between 2001 and 2002. This The Flow of Funds method that is described result from both revisions to some of the concern turns out to be unfounded. The in this section will be called ‘‘FoF-based.’’ input data and recalculations. For the years FFIEC actually raised its eligibility Flow of Funds estimates of mortgage debt 1995 through 1998, the revisions are small requirements. The level of assets required by outstanding are based on data from sources for the estimates of total originations. The FFIEC to be reported to HMDA increased of varying accuracy and timeliness. Bank and only one of note is a 5 percent upward from $31 million in 2001 to $32 million in thrift institution holdings, taken from revision to the estimate for 1995, prompted 2002. In addition, the number of HMDA regulatory filings, are by all accounts highly by a recalculation of the entry for life reporters decreased from 7,771 in 2001 to accurate, as are those from the government insurance companies. The revision to 1999 is 7,638 in 2002. sponsored agencies and direct Federal larger, and results mostly from the 3. An Alternative Method government holdings. The private MBS data substitution of the actual HMDA results for and the life insurance company figures, both The HUD New method makes use of all the that year for the projected value used in the taken from Wall Street sources, are also available sources of data on individual 2000 Rule. Surprisingly, the revised estimate thought to be reasonably accurate. Less origination sources in attempting to estimate for 2000 based on complete data for that year accurate are the estimates of loans made by total conventional mortgage originations. only varies slightly from the projection made private individuals and certain institutions, at the time of the 2000 Rule. However, as discussed in the 2000 Rule and for which comprehensive data on loans Most of the historical estimates produced summarized above, unavoidable gaps in outstanding is provided only once every ten in 2000 can be replicated or closely coverage make the resulting HUD New years, through the Residential Finance approximated, including those for Fannie figures lower-bound estimates of actual Survey. Fortunately, the depository and Freddie, CMBS, HMDA, and life originations rather than best ‘‘point’’ institutions, GSEs, and mortgage-backed insurance companies. The replicability of the estimates. In addition, even for those loans CMBS figures is especially heartening, in that are available, certain assumptions must securities account for the bulk of all holdings light of all the selection criteria and hand be made to convert the available data into of mortgage debt (approximately 72 percent, calculations required to generate those estimates corresponding to the desired according to the Flow of Funds estimates for estimates from the CMBS database. (In the definition and time periods. year-end 2001). 2000 Rule, the estimates for Freddie Mac and An alternative to the bottom-up approach The net change in mortgage debt CMBS originations in 1997 appear to have of HUD New avoids some of the data outstanding in any year is the lower bound been switched, and the revised estimates problems. The Federal Reserve’s Flow of on originations. This is because the net make this correction.) Funds accounts provide the most complete change is defined as originations less the sum The revised figures for 1999 and 2000 and timely set of estimates of multifamily of principal repayments and charge offs. indicate that total conventional originations mortgage credit. The Flow of Funds statistics Historically loan originations have exceeded dropped 8 percent in 1999 from 1998’s very refer to net changes in credit outstanding the net change by a considerable margin in strong level and another 13 percent in 2000. rather than gross originations. Specifically, both the multifamily and single-family However, the HUD New estimate indicates balance sheet estimates of mortgage assets of markets. There are several reasons why the that total conventional originations then lenders are used to produce estimated relationship of originations to net change jumped 40 percent in 2001 and further changes in holdings of mortgages over time. differs between the multifamily and single- increased 15 percent in 2002. Judging from An alternative label for the resulting time family sectors, but the basic principles apply Survey of Mortgage Lending Activity series is ‘‘net change in mortgage debt to both sectors. estimates since 1970, the 2002 number is a outstanding.’’ Table D.3 presents the annual estimates new record high. For 2002, most of the The historical relationship between gross from the Flow of Funds. Also shown are the increased volume is due to increases by originations and net change can be used to estimates of multifamily conventional HMDA lenders and life insurance companies. estimate recent origination volume. Separate originations as published in Table D.10 from One possible concern is that the significant information on FHA multifamily activity can the 2000 rule, and FHA originations from increase in the HMDA number in 2002 was be used to convert the total originations to HUD administrative records. caused by the FFIEC relaxing its eligibility estimates of only conventional originations. BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00224 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24451

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00225 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.071 24452 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

The ratio of mortgage originations to net mortgage debt was higher in 1999 and 2000 Applying the 1.5 multiple to 2001’s net change should be positively correlated with than it had been in 1998. change of $48 billion yields a total the proportion of total originations that are Putting all this together, it seems that the originations estimate of $72 billion. refinancings, for which the net change in appropriate ratio of total originations to net Subtracting the $5 billion of FHA business mortgage debt would be expected to be low change to apply to 1999 and 2000 would be results in a conventional originations relative to that on loans taken out in below that of 1998 and of most other years estimate of $67 billion, to which a subjective connection with a property acquisition. (This of the 1990s. Applying a ratio of 1.5 to the confidence band of at least +/-$2 billion is the pattern observed in the single-family net change estimates in 1999 and 2000 appears warranted. mortgage market.) Refinancings, in turn, results in a total originations estimate of As seen in Table D.3, the Flow of Funds would be expected to be prevalent relative to approximately $56 billion. Subtracting the $4 methodology indicates that total purchase loans at times when interest rates billion in FHA originations results in conventional originations decreased 7.5% are low relative to their recent past. estimates of $52 billion for conventional between 2001 and 2002. In 2002, the net The historical evidence generally supports originations in each year. A subjective change in mortgage debt decreased slightly to this expectation regarding the relationship of confidence band around this point estimate $44 billion. Using the 1.5 multiple for 2002’s originations to net lending. As shown in is at least +/¥$2 billion. net change of $44.2 billion yields a total Table D.3, total originations have been Turning to the estimate for 2001, the first originations estimate of $66 billion. highest relative to net change when interest rates have been low relative to their recent thing to note is that net change in mortgage Subtracting $4.5 billion of FHA business past. The ten-year Treasury yield, a common debt jumped to $48 billion from $37 billion results in a conventional originations benchmark for pricing multifamily of the previous two years. The second thing estimate of $62 billion. mortgages, has generally trended down since to note is that interest rates fell by nearly a This Flow of Funds estimate is over $5 1990. The early 1990s were all marked by percentage point in 2001 relative to their past billion less than the estimate from HUD New. high originations relative to net change, and average. For both of these reasons, total This is surprising given that the HUD New these were also years in which interest rates originations in 2001 would be expected to method is supposed to serve as a lower were particularly low relative to their trailing have been higher than in 1999 or 2000. How boundary on the size of the multifamily five-year averages. In 1996 and 1997, by much higher is a subjective judgment, but 1.5 market, while the Flow of Funds method is contrast, originations were less high relative would seem an appropriate multiple to apply designed to produce a higher ‘‘point’’ to net change, and these were years in which to the net change number in 2001. This is the estimate of the actual size of the market. same multiple as in 1999 and 2000, despite interest rates were only slightly lower than 4. Most Likely Range their five-year trailing averages. the added refinancing incentive in 2001. By In estimating conventional originations for the beginning of 2001, there were relatively In the 2000 Rule, estimates of conventional 1999–2002, the 1998 experience is a useful few properties ‘‘at risk’’ of refinancing. Many multifamily loan originations from various benchmark. That year, total originations presumably had refinanced in one of the sources and methods were evaluated in exceeded the net change by about 80 percent, preceding years, and lock-out provisions, determining the most likely range of annual as shown in Table D.3. There was also a big yield maintenance agreements, and other originations. Those estimates were drop in interest rates in 1998 relative to the loan conditions may have kept these summarized in Table D.10 in the 2000 Rule. recent past, providing an incentive for properties from coming in for refinancings. Some of the estimates from that table are refinancings. As shown in the table, interest Also, there may have been some short-run reproduced below, in Table D.4, along with rates rose slightly in 1999 and again in 2000, capacity problems in the multifamily loan updates and estimates from the Flow of presumably diminishing the incentive to origination industry in 2001 that further Funds method. refinance. Nonetheless, the net change in curtailed volume. BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00226 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24453

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00227 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.072 24454 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Both HUD New (column #4 in Table D.4) selected for 2000. Finally, all indicators point purchase originations from refinancings. and FoF-based (column #9) indicate a surge to a substantial pickup in 2001, and the range Other data sources need to be explored to in lending activity in 2001. Some that seems to fit best with those indicators is determine if an adjustment to the FoF-based corroboration of this jump is provided by $65–$69 billion. model is appropriate. other indicators, flawed though they may be. In 2002, the various methods of estimation HMDA has well-documented coverage give a mixed picture. HUD New indicates a 5. Loan Amount per Unit problems with multifamily loans, but it is surge in lending activity in 2002, while the In determining the size of the conventional noteworthy that HMDA-estimated flow of funds method shows a decrease in multifamily mortgage market for purposes of conventional originations stayed in the same lending activity. Other methods also show the GSE rules, the measure of market size is general range ($26 to $31 billion) in 1998– divergent trends. The composite of 1.25 times the annual number of conventionally 2000 before jumping to $36 billion in 2001. HMDA originations plus life insurance financed multifamily rental housing units. The composite of 1.25 times HMDA commitments also shows a significant The number of units is derived by dividing originations plus life insurance increase between 2001 and 2002. On the the aggregate annual originations by an commitments, described in the 2000 Rule other hand, aggregate GSE multifamily estimate of the average loan amount per and updated here in column #5, also follows purchases and securitizations showed a housing unit financed. For this reason, this basic path. Similarly, aggregate GSE slight decrease between 2001 and 2002. FHA accuracy in the estimate of loan amount per multifamily purchases and securitizations originations (not shown) also decreased unit is as important as accuracy in the dollar stayed in the same general level in 1998– slightly in 2002. estimate of aggregate conventional 2000, before jumping in 2001, although this While this is a subjective judgment, 1.5 originations. A 10 percent error in either will trend reflects changes in both market size may not be the appropriate multiple to apply result in a 10 percent error in the estimate and GSE market share. FHA originations (not to net mortgage debt outstanding in the flow of market size. shown) also rose substantially in 2001, but of funds model in 2002. The difference this too may indicate more than just market between the flow of funds estimate and the The 2000 Rule used estimates of loan size trends. HUD estimate cannot be reconciled without amount per unit drawn from various sources. Column #11 of Table D.4 gives the likely adjusting the FOF multiple. Given the low As summarized in Table D.9 of the 2000 Rule ranges of originations for each of the years. interest rates in 2002, and a refinancing boom and the accompanying text, the estimates for These are based on the estimates from all in the single-family mortgage market, it could 1993–1998 were taken from the GSEs and for sources and interpretations of their strengths be that the multifamily market also had a 1999 from CMBS data. ‘‘Unpaid Principal and weaknesses. In 1999, the $4 billion significant amount of refinancing activity. In Balance’’ or UPB—a balance sheet measure upward revision to the HUD New estimate such a case, there could be an increase in the which for current year loan originations will from the preliminary figure reported in the size of the multifamily market without a differ little from the initial loan amount—is 2000 Rule, together with the higher estimate corresponding increase in net mortgage debt used to calculate aggregate originations of produced by the FoF-based method, justify outstanding. A higher multiple would need loans bought or securitized by the GSEs or an upward revision to the $45–$48 range to be applied to the Flow of Funds model to pooled into non-GSE mortgage-backed estimated in the 2000 Rule. The revised range compensate for the increase in multifamily securities. The figures from Table D.9 of the is set at $50–54 billion. In 2000, HUD New refinancings. 2000 Rule are reproduced below in Table (revised and extended version) suggests that Due to data limitations, the above remains D.5, along with updated estimates from all originations were somewhat lower than in a speculation. The largest increase in three sources for 2000, 2001 and 2002. The 1999, but FoF-based has originations holding multifamily volume came from HMDA estimates that are new since the 2000 Rule at $52 billion. Balancing these conflicting reporting lenders. The HMDA data do not appear in italics. indicators, a range of $48–$52 billion is allow for the separation of multifamily BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00228 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24455

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00229 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.073 24456 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Several options are available for methods could be used. Section F will report than any of the years between 1991 and 2002. developing estimates for 2000, 2001 and the results of several sensitivity analyses Sensitivity analyses are conducted to show 2002. The first is to use the UPB (unpaid showing the effects of the different the effects of multifamily mixes less than the principal balance) per unit estimates from the multifamily mortgage estimates (HUD New previous lows of 11 percent in 1992 and GSEs. These estimates, taken from the Fannie versus Flow-of-Funds) and different per unit 2002. Mae and Freddie Mac annual activity reports amounts ($35,000 or $37,275 which is an The multifamily share of the conforming to HUD, are as follows, computed as in the average of the various estimates) on the goals- conventional market (or ‘‘multifamily mix’’) 2000 Rule as a unit-weighted average of the qualifying shares for the year 2002. Under the is utilized below as part of HUD’s analysis of unpaid principal balance (UPB) per various estimates, the multifamily mix the share of units financed each year meeting multifamily unit in Fannie Mae’s and (defined below) for 2002 ends up around 11 each of the housing goals. Following the 2000 Freddie Mac’s portfolios: percent. Rule, the analysis will focus on multifamily 1997 ...... $27,266 mixes of 15 percent and 16.5 percent, which 6. Multifamily Mix During the 1990s seems reasonable given the 1991–2002 1998 ...... 31,041 The section uses the information on dollar estimates reported in Table D.4. While at the 1999 ...... 35,038 volume of multifamily originations (Table low end of the 1992–2002 averages for the 2000 ...... 37,208 D.4) and average loan amounts (Table D.5) to ‘‘likely range’’, a 15 percent mix more readily 2001 ...... 37,258 estimate the number of multifamily units accommodates any uncertainty about the 2002 ...... 39,787 financed each year as a percentage share of data and the estimation process. An The figure for 2002 is approximately 46 the total (both single-family and multifamily) alternative multifamily mix assumption of percent higher than in 1997. Both Fannie number of dwelling units financed each year; 13.5 percent is also considered, as well as Mae and Freddie Mac’s portfolios generate the years covered include 1991 to 2001. This even lower ones in order to fully consider the estimates of between $39,000 and $40,000 for percentage share, called the ‘‘multifamily effects of heavy refinancing environments 2002. mix’’, is reported in the last two columns of such as 2001–03. Several alternative approaches to Table D.4.13 The ‘‘minimum’’ (‘‘maximum’’) multifamily mix figure reflects the low D. Single-Family Owner and Rental estimating loan amount per unit are Mortgage Market Shares available. The first is to base the estimate on (upper) end of the ‘‘likely range’’ of CMBS data, as was done for 1999 in the 2000 multifamily dollar originations, also reported 1. Available Data Rulemaking. As shown in the last column of in Table D.4. Because of the high goals- qualifying shares of multifamily housing, the As explained later, HUD’s market model Table D.5, the estimates of UPB/unit from will also use projections of mortgage this source are somewhat below those of the multifamily mix is an important parameter in HUD’s projection model for the overall originations on single-family (1–4 unit) GSEs and indicate less increase since the properties. Current mortgage origination data late1990s. market; other things equal, a higher multifamily mix (or conversely, a lower share combine mortgage originations for the three In the first 10 months of 2002, CMBS of single-family loans) leads to a higher different types of single-family properties: properties showed a UPB/unit of $37,038, a estimate of goals-qualifying loans in the owner-occupied, one-unit properties (SF–O); nearly 14 percent jump over the previous overall mortgage market. 2–4 unit rental properties (SF 2–4); and 1– year. Although slightly below the UPB/unit Based on the ‘‘likely range’’ of annual 4 unit rental properties owned by investors for the GSEs, the CMBS numbers are closer conventional multifamily origination (SF–Investor). The fact that the goal to the GSE calculations than in previous volume, multifamily units have represented percentages are much higher for the two years. 15.1 percent (the average of the ‘‘minimum’’ rental categories argues strongly for Another approach is to move the 1999 figures) to 16.3 percent (the average of the disaggregating single-family mortgage estimate of UPB/unit forward by some ‘‘maximum’’ figures) of units financed each originations by property type. This section justifiable index. The 2001 estimates use the year between 1991 and 2002. Considering the discusses available data for estimating the change in average rent on multifamily rental mid-points of the ‘‘likely range’’, the relative size of the single-family rental units from the American Housing Survey. multifamily mix averaged 15.7 percent mortgage market. Because AHS data are not available for 2002, during this period. Notice that multifamily The Residential Finance Survey (RFS) and the 2002 estimate uses the consumer price mix is lower during years of heavy HMDA are the data sources for estimating the index for rent of primary residence. Both refinancing when single-family originations relative size of the single-family rental AHS and CPI rent estimates are listed below: dominate the mortgage market; the market. The RFS, provides mortgage multifamily mix was only 13–14 percent origination estimates for each of the three Year Median Mean CPI during 1993, 1998, and 2001, and single-family property types but it is quite approximately 11 percent during 2002.14 As dated, as it includes mortgages originated 1999 ...... $550 $592 177.5 discussed in Sections F–H, the record single- between 1987 and 1991. (An updated version 2001 ...... 590 647 192.1 family originations ($3.3 trillion) during 2003 of the RFS based on the 2000 Census will not 2002 ...... N/A N/A 199.7 likely resulted in a lower multifamily mix be available until the spring of 2004). HMDA divides newly-originated single-family mortgages into two property types: 15 There is some variation between the two 13 1990 is excluded from this calculation because measures. In the AHS, median rent rose 7.3 of the unusually high multifamily mix that year. (1) Owner-occupied originations, which percent over this two-year period, and mean Also, the estimated multifamily mix from the HUD include both SF–O and SF 2–4. rent increased 9.3 percent. Meanwhile, the New Method is also provided for 2002 since it was (2) Non-owner-occupied mortgage CPI showed an increase of 8.2 percent. In greater than the estimate from the Flow of Funds originations, which include SF Investor. 2001, using the AHS produces an estimate of method. 14 The percentage distributions of mortgages $34,000. The CPI yields a smaller estimate for The projection model for 2002 showed the from these data sources are provided in Table following multifamily mixes for 2002: 11.5 percent 2001; applying the 8.2 percent increase from D.6a. (Table D.6b will be discussed below.) the CPI results in a 2001 estimate of $33,200. for the HUD New multifamily estimate ($67.7 billion) if the average loan amount is $35,000 and Because HMDA combines the first two Since the AHS data are unavailable in 2002, 10.9 percent if the average loan amount is $37,275; categories (SF–O and SF 2–4), the the CPI provides a 2002 estimate of 11.0 percent for the top end ($64 billion) of the comparisons between the data bases must approximately $35,000. Flow of Funds multifamily range ($60–64 billion) necessarily focus on the SF investor category. In 2001, the rent-adjusted 1999 estimate if the average loan amount is $35,000 and 10.4 According to 2000 (2001) HMDA data, was in between the estimates from the CMBS percent if the average loan amount is $37,275; 10.7 investors account for 9.4 (9.9 percent) and GSE data, and was a fair estimate of the percent for the mid-point ($62 billion) of the Flow percent of home purchase loans and 7.6 actual size of the market. In 2002, however, of Funds multifamily range if the average loan percent (5.9 percent) of refinance loans.16 the rent-adjusted number is below both the amount is $35,000 and 10.1 percent if the average loan amount is $37,275; and 10.4 percent for the CMBS and GSE calculations. The rent- low end ($60 billion) of the Flow of Funds 15 The data in Table D.6a ignore HMDA loans adjusted number could be underestimating multifamily range if the average loan amount is with ‘‘non-applicable’’ for owner type. the 2002 UPB/unit. Either the CMBS or GSE $35,000 and 9.8 percent if the average loan amount 16 Due to the higher share of refinance mortgages calculations, or an average of the various is $37,275. during 2001, the overall single-family-owner

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00230 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24457

Assuming a 35 percent refinance rate per of 17.3 percent is approximately twice the assuming 8 percent and 12 percent were also HUD’s projection model, the 2000 (2001) HMDA estimates. In their past comments, the considered. As discussed below, HUD’s HMDA data are consistent with an investor GSEs have argued that the HMDA-reported baseline projection of 10 percent is probably share of 8.8 (8.5) percent.17 The RFS estimate SF investor share should be used by HUD. In quite conservative; however, given the its 1995 and 2000 rules, HUD’s baseline uncertainty around the data, it is difficult to model assumed a 10 percent share for the SF draw firm conclusions about the size of the percentage reported by HMDA for 2001 (92.7 investor group—only slightly higher than the single-family investor market, which percent) is larger than that reported for 2000 (91.3 HMDA-based estimates; alternative models necessitates the sensitivity analysis that HUD percent). conducts. The release this spring of the 17 HMDA data for 2002 would yield a slightly assuming a 35 percent refinance rate would be 9.6 updated RFS should clarify this issue. higher investor share; the derived investor share percent if 2002 HMDA data were used. BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00231 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24458 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00232 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.074 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24459

BILLING CODE 4210–27–C applying the following unit-per-mortgage E. HUD’s Market Share Model 2. Analysis of Investor Market Share assumptions: 2.25 units per SF 2–4 property This section integrates findings from the and 1.35 units per SF investor property. Both Blackley and Follain. During the 1995 rule- previous two sections about the size of the figures were derived from the 1991 RFS.20 multifamily mortgage market and the relative making, HUD asked the Urban Institute to Based on these calculations, the percentage analyze the differences between the RFS and distribution of single-family owner and rental distribution of newly-mortgaged single- HMDA investor shares and determine which mortgages into a single model of the mortgage family dwelling units was derived for each of was the more reasonable. The Urban market. The section provides the basic Institute’s analysis of this issue is contained the various estimates of the investor share of equations for HUD’s market share model and in reports by Dixie Blackley and James single-family mortgages (discussed earlier identifies the remaining parameters that must Follain.18 Blackley and Follain provide and reported in Table D.6a). The results are be estimated. reasons why HMDA should be adjusted presented in Table D.6b. Three points should The output of this section is a unit-based upward as well as reasons why the RFS be made about these data. First, notice that distribution for the four property types should be adjusted downward. They find that the ‘‘SF–Rental’’ row highlights the share of discussed in Section B.21 Sections F–H will HMDA may understate the investor share of the single-family mortgage market accounted apply goal percentages to this property single-family mortgages because of ‘‘hidden for by all rental units. distribution in order to determine the size of investors’’ who falsely claim that a property Second, notice that the rental categories the mortgage market for each of the three is owner-occupied in order to more easily represent a larger share of the unit-based housing goals. obtain mortgage financing. RFS may overstate market than they did of the mortgage-based 1. Basic Equations for Determining Units the investor share of the market because units market reported earlier. This, of course, Financed in the Mortgage Market that are temporarily rented while the owner follows directly from applying the loan-per- seeks another buyer may be counted as rental unit expansion factors. The model first estimates the number of units in the RFS, even though rental status Third, notice that the rental share under dwelling units financed by conventional of such units may only be temporary. The HMDA’s unit-based distribution is again conforming mortgage originations for each of RFS’s investor share should be adjusted about one-half of the rental share under the the four property types. It then determines downward in part because the RFS assigns RFS’s distribution. The rental share in HUD’s each property type’s share of the total all vacant properties to the rental group, but 1995 and 2000 Rules and this year’s number of dwelling units financed. some of these are likely intended for the proposed rule is slightly larger than that a. Single-Family Units owner market, especially among one-unit reported by HMDA. The rental share in the This section estimates the number of properties. Blackley and Follain’s analysis of ‘‘Blackley-Follain’’ alternative is slightly single-family units that will be financed in this issue suggests lowering the investor above HUD’s estimate. Rental units account the conventional conforming market, where share from 17.3 percent to about 14–15 for 15.1 percent of all newly financed single- single-family units (SF–UNITS) are defined percent. family units under HUD’s baseline model, as: Finally, Blackley and Follain note that a compared with 13.7 (13.1) percent under a conservative estimate of the SF investor share model based on 2000 (2001) HMDA data. SF–UNITS = SF–O + SF 2–4 + SF– is advisable because of the difficulty of INVESTOR measuring the magnitudes of the various 4. Conclusions First, the dollar volume of conventional effects that they analyzed. In their 1996 This section has reviewed data and conforming single-family mortgages paper, they conclude that 12 percent is a analyses related to determining the rental (CCSFM$) is derived as follows: reasonable estimate of the investor share of share of the single-family mortgage market. (1) CCSFM$ = CONV% * CONF% * SFORIG$ single-family mortgage originations.19 There are two main conclusions: Where: Blackley and Follain caution that uncertainty • While there is uncertainty concerning exists around this estimate because of the relative size of this market, the CONV% = conventional mortgage inadequate data. projections made by HUD in 1995 and 2000 originations as a percent of total appear reasonable and, therefore, will serve mortgage originations; estimated to be 3. Single-Family Market in Terms of Unit 88%.22 Shares as the baseline assumption in the HUD’s market share model for this year’s Proposed CONF% = conforming mortgage originations The market share estimates for the housing Rule. (measured in dollars) as a percent of goals need to be expressed as percentages of • HMDA likely underestimates the single- conventional single-family originations; units rather than as percentages of mortgages. family rental mortgage market. Thus, this forecasted to be 80% by industry. Thus, it is necessary to compare unit-based part of the HMDA data are not considered SFORIG$ = dollar volume of single-family distributions of the single-family mortgage reliable enough to use in computing the one-to-four unit mortgages; $1,700 market under the alternative estimates market shares for the housing goals. Various billion is used here as a starting discussed so far. The mortgage-based sensitivity analyses of the market shares for assumption to reflect market conditions distributions given in Table D.6a were single-family rental properties are conducted during the years 2005–2008.23 While adjusted in two ways. First, the owner- in Sections F, G, and H. These sensitivity occupied HMDA data were disaggregated analyses will include the GSEs’ 21 The property distribution reported in Table D.1 between SF–O and SF 2–4 mortgages by recommended model that assumes investors is an example of the output of the market share assuming that SF 2–4 mortgages account for account for 8 percent of all single-family model. Thus, this section completes Step 1 of the 2.0 percent of all single-family mortgages; mortgages. These sensitivity analyses will three-step procedure outlined above in Section B. according to RFS data, SF 2–4 mortgages 22 According to estimates by the Mortgage show the effects on the overall market represent 2.3 percent of all single-family Bankers Association of America (MBAA), the estimates of the different projections about mortgages so the 2.0 percent assumption may conventional share of the 1–4 family market was the size of the single-family rental market. be slightly conservative. Second, the between 86 and 88 percent of the market from 1993 The upcoming RFS based on the year 2000 resulting mortgage-based distributions were to 1999, with a one-time low of 81 percent in 1994. Census will help clarify issues related to the Calculated from ‘‘1–4 Family Mortgage shifted to unit-based distributions by investor share of the single-family mortgage Originations’’ tables (Table 1—Industry and Table market. At that time, HUD will reconsider its 2—Conventional Loans) from ‘‘MBAA Mortgage and 18 Dixie M. Blackley and James R. Follain, ‘‘A estimates of the investor share of the Market Data,’’ at www.MBAAa.org/marketkdata/ as Critique of the Methodology Used to Determine mortgage market. of July 13, 2000. More recent unpublished estimates Affordable Housing Goals for the Government by MBAA are slightly higher. Sponsored Housing Enterprises,’’ report prepared 23 Single-family mortgage originations of $1,700 for Office of Policy Development and Research, 20 The unit-per-mortgage data from the 1991 RFS billion are similar to Freddie Mac’s projection of Department of Housing and Urban Development, match closely the GSE purchase data for 2001. $1,748 billion for 2005 and Fannie Mae’s projection October 1995; and ‘‘HUD’s Market Share Blackley and Follain show that an adjustment for of $1,675 billion for 2005. As discussed later, Methodology and its Housing Goals for the vacant investor properties would raise the average single-family originations could differ from $1,700 Government Sponsored Enterprises,’’ unpublished units per mortgage to 1.4; however, this increase is billion during the 2005–2008 period that the goals paper, March 1996. so small that it has little effect on the overall market will be in effect. As recent experience shows, 19 Blackley and Follain (1996), p. 20. estimates. Continued

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00233 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24460 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

alternative assumptions will be Substituting these values into (2) yields an Assuming a multifamily mix of 15 percent examined, it must be emphasized that estimate of 8.5 million mortgages. and solving (5b) yields the following: the important concept for deriving the Third, the total number of single-family (5c) MF–UNITS = [0.15/0.85] * SF–UNITS = goal-qualifying market shares is the mortgages is divided among the three single- 0.176 * SF–UNITS = 1.6 million. relative importance of single-family family property types. Using the 88/2/10 c. Total Units Financed versus multifamily mortgage originations percentage distribution for single-family (the ‘‘multifamily mix’’ discussed in mortgages (see Section D), the following The total number of dwelling units Section C) rather than the total dollar results are obtained: financed by the conventional conforming mortgage market (TOTAL) can be expressed volume of single-family originations (3a) SF–OM# = 0.88 * CCSFM# = number of in three useful ways: considered in isolation. owner-occupied, one-unit mortgages = Substituting these values into (1) yields an 7.5 million. (6a) TOTAL = SF–UNITS + MF–UNITS = estimate for the conventional conforming (3b) SF–2–4M# = 0.02 * CCSFM# = number 10.6 million (or more precisely, market (CCSFM$) of $1,197 billion. of owner-occupied, two-to-four unit 10,632,145 units) Second, the number of conventional mortgages = 0.17 million. (6b) TOTAL = SF–O + SF 2–4 + SF– conforming single-family mortgages (3c) SF–INVM# = 0.10 * CCSFM# = number INVESTOR + MF–UNITS (CCSFM#) is derived as follows: of one-to-four unit investor mortgages = (6c) TOTAL = SF–O + SF–RENTAL + MF– (2) CCSFM# = (CCSFM$ * (1–REFI)/ 0.85 million. UNITS Where: PSFLOAN$) + (CCSFM$ * REFI)/ Fourth, the number of dwelling units RSFLOAN$) SF–RENTAL equals SF–2–4 plus SF– financed for the three single-family property INVESTOR Where: types is derived as follows: REFI = the refinance rate, assumed to be 35 (4a) SF–O = SF–OM# + SF–2–4M# = number 2. Dwelling Unit Distributions by Property percent for the baseline.24 of owner-occupied dwelling units Type PSFLOAN$ = the average conventional financed = 7.7 million. The next step is to express the number of conforming purchase mortgage amount (4b) SF 2–4 = 1.25 * SF–2–4M# = number of dwelling units financed for each property for single-family properties; estimated to rental units in 2–4 properties where a type as a percentage of the total number of be $146,000.25 owner occupies one of the units = 0.2 units financed by conventional conforming RSFLOAN$ = the average conventional million.27 mortgage originations.28 conforming refinance mortgage amount (4c) SF–INVESTOR = 1.35 * SF–INVM# = The projections used above in equations for single-family properties; estimated to number of single-family investor (1)–(6) produce the following distributions of be $131,000.26 dwelling units financed = 1.1 million. financed units by property type: Fifth, summing equations 4a–4c gives the market projections often change. For example, the projected number of newly-mortgaged single- % Share MBAA projected $1,246 billion for 2003, while family units (SF–UNITS): their projection for 2003 rose to $1,774 billion in SF–O ...... 72.2 January 2003; of course, actual 2003 mortgage (5) SF–UNITS = SF–O + SF 2–4 + SF– SF 2–4 ...... 2.0 originations were almost double the latter amount. INVESTOR = 9.0 million SF INVESTOR ...... 10.8 (See http://www.MBAAa.org/marketdata/forecasts b. Multifamily Units MF–UNITS ...... 15.0 for January 2003 Mortgage Finance Forecasts.) In its January 22, 2004 forecast, the MBAA projected The number of multifamily dwelling units mortgage originations of $1.9 trillion in 2004 and (MF–UNITS) financed by conventional Total ...... 100.0 approximately $1.7 trillion in 2005 and 2006. conforming multifamily originations is or Section F will report the effects on the market calculated by the following series of SF–O ...... 72.2 estimates of alternative estimates of single-family equations: SF–RENTER ...... 12.8 mortgage originations. MF–UNITS ...... 15.0 24 The model requires an estimated refinance rate (5a) TOTAL = SF–UNITS + MF–UNITS because purchase and refinance loans can have (5b) MF–UNITS = MF–MIX * TOTAL = MF– different shares of goals-qualifying units. In 2003, MIX * (SF–UNITS + MF–UNITS) = [MF– Total ...... 100.0 the refinance rate was over 60 percent. In its MIX/(1—MF–MIX)] * SF–UNITS January 22, 2004 forecast, the MBAA projects 34 Where: Sections C and D discussed alternative percent for 2004 and 22 percent for 2005. Freddie MF–MIX = the ‘‘multifamily mix’’, or the projections for the mix of multifamily Mac projects a 36 percent refinance rate for 2004 percentage of all newly-mortgaged originations and the investor share of single- and a 29 percent rate for 2005, and Fannie Mae family mortgages. Following the 2000 Rule, projects a 48 percent refinance rate for 2004 and 24 dwelling units that are multifamily; as discussed in Section C, alternative this appendix will focus on three multifamily percent for 2005. The baseline model uses a higher mixes (13.5 percent, 15.0 percent, and 16.5 refinance rate of 35 percent because conforming estimates of the multifamily market will percent) but there will also be sensitivity conventional loans tend to refinance at a higher rate be included in the analysis. As explained analysis of other multifamily mix than the overall market. Sensitivity analyses for in Section C above, the baseline model assumptions. Under a 16.5 percent alternative refinance rates are presented in Sections assumes a multifamily mix of 15 percent; F–H. multifamily mix, the newly-mortgaged unit results are also presented in the basic 25 The average 2002 purchase loan amount is distribution would be 70.9 percent for Single- market tables of Sections F–H for a estimated at $135,060 for owner occupied units Family Owner, 12.6 percent for Single- higher (16.5 percent) and a lower (13.5 using 2002 HMDA average loan amounts for single- Family Renter, and 16.5 percent for percent) multifamily mix. In addition, family home purchase loans in metropolitan areas. Multifamily-Units. The analysis in sections further sensitivity analyses are reported A small adjustment is made to this figure to account F-H will focus on goals-qualifying market for a small number of two-to-four and investor in those sections for even lower shares for this property distribution as well properties (see Section D above). This produces an multifamily mixes that could occur as the one presented above for the more average purchase loan size of $133,458 for 2002 during periods of heavy single-family conservative multifamily mix of 15 percent. which is then inflated 3 percent a year for three refinancing activity. years and then rounded to arrive at an estimated The appendix will assume the following $146,000 average loan size for home purchase loans for the investor share of single-family in 2005. environments are used, $146,000 average loan mortgages—8 percent, 10 percent, and 12 26 The average refinance loan amount is estimated amounts are used for both purchase and refinance percent. The middle value (10 percent by averaging the relationship between HMDA loans. This relationship is consistent with the investor share) is used in the above observed relationship in past refinance years such average purchase and refinance loan amounts for calculations and will be considered the 1999 and 2000, which were non-refinance as 1998, 2001, and 2002. environments. Applying this average of 90 percent 27 Based on the RFS, there is an average of 2.25 (refinance loan amount/purchase loan amount) to housing units per mortgage for 2–4 properties. 1.25 28 The share of the mortgage market accounted for the $146,000 average loan amount for purchase is used here because one (i.e., the owner occupant) by owner occupants is (SF–O)/TOTAL; the share of loans gives a rounded estimate of $131,000 for of the 2.25 units is allocated to the SF–O category. the market accounted for by all single-family rental average refinance loan amounts. When refinance The RFS is also the source of the 1.35 used in (4c). units is SF–RENTAL/TOTAL; and so on.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00234 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24461

‘‘baseline’’ projection throughout the Table D.7 reports the unit-based with an investor mortgage share of 12 appendix. However, HUD recognizes the distributions produced by HUD’s market percent) to a high of 75.5 percent uncertainty of projecting origination volume share model for different combinations of (multifamily mix of 13.5 percent coupled in markets such as single-family investor these projections. The effects of the different with an investor mortgage share of 8 percent). properties; therefore, the analysis in Sections projections can best be seen by examining the The owner share under the baseline F–H will also consider market assumptions owner category which varies by 6.6 projection (15 percent mix and 10 percent investor) is 72.2 percent. other than the baseline assumptions. percentage points, from a low of 68.9 percent (multifamily mix of 16.5 percent coupled BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00235 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24462 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00236 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.075 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24463

Comparison with the RFS. The Residential low- and moderate-income percentages to the refinance their existing mortgage.29 The data Finance Survey is the only mortgage data property shares given in Table D.7. This cover conventional mortgages below the source that provides unit-based property section essentially accomplishes Steps 2 and conforming loan limit, which was $300,700 distributions directly comparable to those 3 of the three-step procedure discussed in in 2002. Table D.8 gives the percentage of reported in Table D.7. Based on RFS data for Section B.2. mortgages originated for low- and moderate- 1987 to 1991, HUD estimated that, of total Technical issues and data adjustments income families for the years 1992–2002. dwelling units in properties financed by related to the low- and moderate-income recently acquired conventional conforming Data are presented for home purchase, percentages for owners and renters are mortgages, 56.5 percent were owner- refinance, and all single-family-owner loans. discussed in the first two subsections. Then, occupied units, 17.9 percent were single- The discussion below will often focus on family rental units, and 25.6 percent were estimates of the size of the low- and home purchase loans because they typically moderate-income market are presented along multifamily rental units. Thus, the RFS account for the majority of all single-family- with several sensitivity analyses. Based on presents a much lower owner share than does owner mortgages.30 For each year, a low- and these analyses, HUD concludes that 51–57 HUD’s model. This difference is due mainly moderate-income percentage is also reported to the relatively high level of multifamily percent is a reasonable estimate of the for the conforming market without B&C originations (relative to single-family mortgage market’s low- and moderate-income originations) during the mid- to late-1980s, share for the four years (2005–2008) when loans. which is the period covered by the RFS. As the new goals will be in effect. BILLING CODE 4210–27–P noted earlier, the RFS based on the year 2000 Census should clarify issues related to the 1. Low- and Moderate-Income Percentage for rental segment of the mortgage market when Single-Family-Owner Mortgages 29 it becomes available in the spring of this year HMDA data are expressed in terms of number a. HMDA Data of loans rather than number of units. In addition, (2004). The most important determinant of the HMDA data do not distinguish between owner- F. Size of the Conventional Conforming low- and moderate-income share of the occupied one-unit properties and owner-occupied Mortgage Market Serving Low- and mortgage market is the income distribution of 2–4 properties. This is not a particular problem for this section’s analysis of owner incomes. Moderate-Income Families single-family borrowers. HMDA reports 30 Sensitivity analyses will focus on how the This section estimates the size of the low- annual income data for families who live in results change during a heavy refinancing and moderate-income market by applying metropolitan areas and purchase a home or environment.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00237 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24464 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00238 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.076 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24465

Table D.8 also reports similar data for very- 2002, there have been nine years of strong influenced by the growth of subprime loans, low-income families (that is, families with affordable lending. which are mainly refinance loans. If B&C incomes less than 60 percent of area median Of course, it is recognized that lending loans are excluded from the market income). As discussed in Section H, very- patterns could change with sharp changes in definition, the home purchase and refinance low-income families are the main component interest rates and the economy. However, the percentages are approximately the same in of the special affordable mortgage market. fact that lending to low-income families has 1997 and 1999, as well as in 1995 and 1996. Two trends in the income data should be remained at a high level for nine years (See Table D.8.) Even after excluding all mentioned—one related to the growth in the demonstrates that the market has changed in subprime loans from the market definition in market’s funding of low- and moderate- fundamental ways from the mortgage market 1997 and 1999, the very-low-income and income families during the 1990s (and of the early 1990s. The numerous innovative low-mod shares for refinance loans are only particularly the growth since 1998 which was products and outreach programs that the slightly less (about one percentage point) the last year analyzed in HUD’s 2000 GSE industry has developed to attract lower- than those for home purchase loans. Rule); and the other related to changes in the income families into the homeownership and The year 2000 stands out because of the borrower income distributions for refinance mortgage markets appear to be working and extremely high lower-income shares for and home purchase mortgages. Throughout there is no reason to believe that they will refinance loans. In that year, the low-mod this appendix, ‘‘low- and moderate-income’’ not continue to assist in closing troubling (very-low-income) share of refinance loans will often be referred to as ‘‘low-mod’’. homeownership gaps that exist today. As was 6.8 (4.3) percentage points higher than Recent Trends in the Market Share for explained in Appendix A, the demand for the low-mod (very-low-income) share of Lower Income Borrowers. First, focus on the homeownership on the part of non- home purchase loans; this differential is percentages in Table D.8 for the total (both traditional borrowers, minorities, and reduced to 5.2 (3.2) percent if B&C loans are home purchase and refinance) conforming immigrants should help to maintain activity excluded from the market definition (see market. After averaging about 30 percent in the affordable portion of the mortgage Table D.8). The differential for 2000 is during 1992–93, the percentage of borrowers market. Thus, while economic recession or reduced further to 2.8 (1.5) percent if all with less than area median income jumped higher interest rates would likely reduce the subprime loans (both A-minus and B&C) are to 41.0 percent in 1994, and remained above low- and moderate-income share of mortgage excluded from the market definition (not 40 percent through 2002. Over the eight year originations, there is evidence that the low- reported). While the projection model period, 1994 to 2001, the low-mod share of mod market might not return to the low (explained below) for years 2005–08 will the total market averaged 43.2 percent (or levels of the early 1990s. There is also input low-mod percentages for the entire 42.4 percent if B&C loans are excluded from evidence that the affordable lending market conforming market, the model will exclude the market totals).31 The share of the market increased slightly since 1998, although it is the effects of B&C loans. Sensitivity analyses accounted for by very-low-income borrowers recognized that this could be due to the will also be conducted showing the effects on followed a similar trend, increasing from 6– recent period of historically low interest the overall market estimates of excluding all 7 percent in 1992–93 to about 12 percent in rates. subprime loans as well as other loan 1994 and averaging 13.3 percent during the Refinance Mortgages. In the 2000 Rule, categories such as manufactured housing 1994-to-2002 period (or 12.8 percent if B&C HUD’s market projection model assumed that loans. The projection model will initially assume loans are excluded). low-mod borrowers represented a smaller that refinancing is 35 percent of the single- Next, consider the percentages for home share of refinance mortgages than they do of family mortgage market; this will be followed purchase loans. The share of the home loan home purchase mortgages. However, as by projection models that reflect heavy market accounted for by less-than-median- shown in Table D.8, the income refinance environments. Given the volatility income borrowers increased from 34.4 characteristics of borrowers refinancing of refinance rates from year to year, it is percent in 1992 to 45.3 percent in 2002. mortgages seem to depend on the overall important to conduct sensitivity tests using Within the 1994-to-2002 period, the low-mod level of refinancing in the market. During the different refinance rates. share of the home purchase market averaged refinancing wave of 1992 and 1993, 44.6 percent between 1999 and 2002, refinancing borrowers had much higher b. Manufactured Housing Loans compared with 42.2 percent between 1994 incomes than borrowers purchasing homes. Because manufactured housing loans are and 1998. Similarly, the very-low-income For example, during 1993 low- and such an important source of affordable share of the home purchase market was also moderate-income borrowers accounted for housing, they are included in the mortgage higher during the 1999-to-2002 period than 29.3 percent of refinance mortgages, market definition in this appendix—or at during the 1994-to-1998 period (14.4 percent compared to 38.9 percent of home purchase least that portion of the manufactured versus 12.6 percent). Note that within the borrowers. While this same pattern was housing market located in metropolitan areas more recent period, the low-mod share for exhibited during the two recent refinancing is included, as HMDA doesn’t adequately home purchase loans was particularly high periods (1998 and 2001–2002), the cover non-metropolitan areas. The GSEs have during 1999 (45.2 percent) and 2000 (44.8 differentials were much smaller—during questioned HUD’s including these loans in percent) before falling slightly in 2001 (43.2 2001–2002 (1998), low-mod borrowers its market estimates; therefore, following the percent), only to rebound again in 2002 (45.3 accounted for 42.1 (39.7) percent of refinance same procedure used in the 2000 Rule, this percent). As shown in Table D.8, the low- loans, compared with 44.3 (43.0) percent of Appendix will report the effects of excluding mod shares do not change much when B&C home purchase loans. However, the refinance manufactured home loans from the market home loans are excluded from the market effect was still evident, as can be seen by the estimates. As explained later, the effect of definition; this is because B&C loans are almost seven percentage drop in the low-mod manufactured housing on HUD’s mainly refinance loans. percentage for refinance loans between 2000 metropolitan area market estimate for each of It appears that the affordable lending (a low refinance year) and 2001 (a high the three housing goals is approximately one market is even stronger today than when refinance year). percentage point or less. HUD wrote the 2000 Rule, which covered On the other hand, for recent years As discussed in Appendix A, the market data through 1998. The very-low- characterized by a low level of refinancing, manufactured housing market increased income and low-mod percentages were the low-mod share of refinance mortgages has rapidly during the 1990s, as units placed in higher during 1999 to 2002 than they were been about the same or even greater than that service increased from 174,000 in 1991 to during the earlier period. In addition, when of home purchase mortgages. As shown in 374,000 in 1999. However, due to various HUD wrote the 2000 Rule, there had been Table D.8, there was little difference in the problems in the industry such as lax five years (1994–98) of solid affordable very-low-income and low-mod shares of underwriting and repossessions, volume has lending for lower-income borrowers. Now, refinance and home purchase loans during declined in recent years, falling to 192,000 in with four additional years of data for 1999– 1995 and 1996. In 1997, 1999, and 2000, the 2001 and to 172,000 in 2002. Still, the two lower-income shares (i.e., very-low- affordability of manufactured homes for 31 The annual averages of the goals-qualifying income and low-mod shares) of refinance lower-income families is demonstrated by mortgages reported in this appendix are unweighted mortgages were significantly higher than the their average price of $48,800 in 2001, a averages; for analyses using weighted average see lower-income shares of home purchase loans. fraction of the median price for new Appendix A. To a certain extent, this pattern was ($175,000) and existing ($147,800) homes.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00239 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24466 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Many households live in manufactured family size as measured by the number of estimate population statistics. Affordability housing because they simply cannot afford bedrooms). Thus, the GSEs’ performance calculations were made using 1993–95 area site-built homes, for which the construction under the housing goals is measured in terms median incomes calculated by HUD. costs per square foot are much higher. of the affordability of the rental dwelling POMS Results. The rent affordability Although manufactured home loans cannot units that are financed by mortgages that the estimates from POMS of the affordability of be identified in the HMDA data, Randy GSEs purchase; the income of the occupants newly-mortgaged rental properties are quite Scheessele at HUD identified 21 lenders that of these rental units is not considered in the consistent with the AHS data on the primarily originated manufactured home calculation of goal performance. For this affordability of the rental stock (discussed loans during 2001 and likely account for reason, it is appropriate to base estimates of above). Ninety-six (96) percent of single- most of these loans in the HMDA data for market size on rent affordability data rather family rental properties with new mortgages metropolitan areas.32 HMDA data on home than on renter income data. between 1993 and 1995 were affordable to loans originated by these lenders indicate A rental unit is considered to be low- and moderate-income families, and 56 that:33 ‘‘affordable’’ to low- and moderate-income percent were affordable to very-low-income • A very high percentage of these loans— families, and thus qualifies for the Low- and families. The corresponding percentages for 75 percent in 2001—would qualify for the Moderate-Income Goal, if that unit’s rent is newly-mortgaged multifamily properties are Low- and Moderate-Income Goal, equal to or less than 30 percent of area 96 percent and 51 percent, respectively. • A substantial percentage of these loans— median income. Table D.14 of Appendix D in Thus, these percentages for newly-mortgaged 42 percent in 2001—would qualify for the HUD’s 2000 Rule reported AHS data on the properties from the POMS are similar to Special Affordable Goal, and affordability of the rental housing stock for those from the AHS for the rental stock. As • Almost half of these loans—47 percent in the survey years between 1985 and 1997. The discussed in the next section, the baseline 2001—would qualify for the Underserved 1997 AHS showed that for 1–4 unit projection from HUD’s market share model Areas Goal.34 unsubsidized single-family rental properties, assumes that 90 percent of newly-mortgaged, Thus an enhanced presence in this market by 94 percent of all units and of units single-family rental and multifamily units are the GSEs would benefit many lower-income constructed in the preceding three years had affordable to low- and moderate-income families. It would also contribute to their gross rent (contract rent plus the cost of all families.35 utilities) less than or equal to 30 percent of presence in underserved rural areas, 3. Size of the Low- and Moderate-Income area median income. For multifamily especially in the South. Mortgage Market unsubsidized rental properties, the 2. Low- and Moderate-Income Percentage for corresponding figure was 92 percent. The This section provides estimates of the size Renter Mortgages AHS data for the other survey years were of the low- and moderate-income mortgage Following the 2000 Rule, measures of the similar to the 1997 data. market. Subsection 3.a provides some necessary background by comparing HUD’s rent affordability of the single-family rental b. Property Owners and Managers Survey estimate made during the 2000 rule-making and the multifamily rental markets are (POMS) obtained from the American Housing Survey process with actual experience between 1999 (AHS) and the Property Owners and As discussed in the 2000 GSE Rule, there and 2001. Subsection 3.b presents new Managers Survey (POMS). As explained were concerns about using AHS data on rents estimates of the low-mod market while below, the AHS provides rent information for from the outstanding rental stock to proxy Subsection 3.c reports the sensitivity of the the stock of rental properties while the POMS rents for newly mortgaged rental units. HUD new estimates to changes in assumptions provides rent information for flow of investigated that issue further using the about economic and mortgage market mortgages financing that stock. As discussed POMS. conditions. POMS Methodology. The affordability of below, the AHS and POMS data provide very a. Actual Market Performance Between 1995 multifamily and single-family rental housing similar estimates of the low- and moderate- and 2002 income share of the rental market. backing mortgages originated in 1993–1995 was calculated using internal Census Bureau Before reporting market projections for the a. American Housing Survey Data files from the American Housing Survey- new goals-setting period (2005–08), this The American Housing Survey does not National Sample (AHS) from 1995 and the section discusses actual market experience include data on mortgages for rental Property Owners and Managers Survey from for 1995 to 2002, as shown in Table D.9.36 properties; rather, it includes data on the 1995–1996. The POMS survey was The 1995 to 1998 market estimates in Table characteristics of the existing rental housing conducted on the same units included in the D.9 were reported by HUD in its 2000 Rule stock and recently completed rental AHS survey, and provides supplemental while the 1999–2002 estimates are new. The properties. Current data on the income of information such as the origination year of 1999–2002 estimates allow a comparison prospective or actual tenants has also not the mortgage loan, if any, recorded against between HUD’s projections and actual market been readily available for rental properties. the property included in the AHS survey. experience. This discussion of the 1995-to- Where such income information is not Monthly housing cost data (including rent 2002 market considers all three housing available, the 1992 GSE Act provides that the and utilities), number of bedrooms, and goals, since the explanations for the rent of a unit can be used to determine the metropolitan area (MSA) location data were differences between the projected and actual affordability of that unit and whether it obtained from the AHS file. market shares are common across the three qualifies for the Low- and Moderate-Income In cases where units in the AHS were not goals. B&C loans are not included in the Goal. A unit qualifies for the Low- and occupied, the AHS typically provides rents, market estimates reported in Table D.9. The Moderate-Income Goal if the rent does not either by obtaining this information from discussion of Table D.9 will first focus on the exceed 30 percent of the local area median property owners or through the use of market estimates for 1995–1997 and 1999– income (with appropriate adjustments for imputation techniques. Estimated monthly 2000, which, because of their relatively low housing costs on vacant units were therefore levels of refinancing, will be referred to as 32 See Randall M. Scheesele, 1998 HMDA calculated as the sum of AHS rent and utility ‘‘home purchase environments’’. The Highlights, op. cit. and ‘‘HUD Subprime and costs estimated using utility allowances discussion will then examine the market Manufactured Home Lender List’’ at http:// published by HUD as part of its regulation of huduseer.org/datasets/manu.html. the GSEs. Observations where neither 35 In 2002, 75 percent of GSE purchases of single- 33 Since most HMDA data are for loans in monthly housing cost nor monthly rent was family rental units and 89 percent of their metropolitan areas and a substantial share of available were omitted, as were observations purchases of multifamily units qualified under the manufactured homes are located outside where MSA could not be determined. Units Low- and Moderate-Income Goal, excluding the metropolitan areas, HMDA data may not accurately with no cash rent and subsidized housing effects of missing data. state the goals-qualifying shares for loans on 36 units were also omitted. Because of the The goals-qualifying shares reported in Table manufactured homes in all areas. D.9 for 1995–2002 are, of course, estimates 34 While many fewer manufactured homes loans shortage of observations with 1995 themselves; even though information is available were identified in the 2002 HMDA data, the loans originations, POMS data on year of mortgage from HMDA and other data sources for most of the showed similar goals-qualifying shares: low-mod origination were utilized to restrict the important model parameters, there are some areas (78.3 percent), special affordable (45.6 percent), and sample to properties mortgaged during 1993– where information is limited, as discussed underserved areas (47.5 percent). 1995. POMS weights were then applied to throughout this appendix.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00240 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24467

estimates for the heavy refinance years of data to exclude B&C loans and to incorporate Underserved Areas in non-metropolitan areas 1998, 2001, and 2002. After that, HUD’s the more expansive definition of will be explained. methods for adjusting the 1995–2001 market BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00241 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24468 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00242 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.077 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24469

HUD’s market projections in the 2000 Rule throughout this appendix, the multifamily 1995–97 and 16–17 percent during 1999– were 50–55 percent for the Low- and mix fell during the heavy refinance years. 2000. Of course, this shift toward single- Moderate-Income Goal, 23–26 percent for the Refinance Years. The goals-qualifying family loans reduces the goals-qualifying Special Affordable Goal, and 29–32 percent percentages for the heavy refinance years shares of the overall market. for the Underserved Areas Goal. Thus, the (1998, 2001 and 2002) are lower than those B&C Mortgages. As discussed in Appendix upper bound figures for the market share for the other years. For example, the low-mod A, the market for subprime mortgages has ranges in the 2000 Rule were lower than market share was 54 percent in 1998 and experienced rapid growth over the past 5–6 actual experience during 1999 and 2000, as 2002 and 55 percent in 2001—both estimates years, rising from an estimated $65 billion in well as for the earlier 1995–97 period—for within HUD’s earlier market share range of 1995 to $174 billion in 2001 and $213 billion the low-mod estimate, 55 percent versus 57– 50–55 percent.39 The special affordable in 2002. Table 9 provides goals-qualifying 59 percent; for the special affordable market share during 1998, 2001, and 2002 market shares that exclude the B&C portion estimate, 26 versus 28–30 percent, and for was 26 percent—which places it at the top of the subprime market; or conversely, that the underserved areas estimate, 32 percent end of HUD’s earlier market range of 23–26 include the A-minus portion of the subprime versus 33–35 percent. percent. The goals-qualifying percentages market. This section explains how these There are three main reasons for the during 1998, 2001, and 2002 are, of course, ‘‘adjusted’’ market shares are calculated from differential between HUD’s earlier estimates lower than those for the ‘‘home purchase’’ ‘‘unadjusted’’ market shares that include B&C (made during 2000 based on HMDA data years of 1995–97 and 1999–2000. For loans, using the year 1999 as an example. through 1998) and the higher goals-qualifying example, the special affordable market share Industry sources estimate that the market shares of recent years. First, of approximately 26 percent in 2001 and subprime market totaled $160 billion in historically low interest rates and strong 2002 was 3–4 percentage points lower than 1999, or 12.5 percent of all mortgages ($1,285 billion) originated that year.41 In terms of economic expansion allowed lower-income the corresponding share in 1999 and 2000. credit risk, this $160 billion includes a wide families to enter the homeownership and There are three main reasons for this. First, range of mortgage types. ‘‘A-minus’’ loans, mortgage market during the mid-to-late the goals-qualifying shares for single-family which represent at least half of the subprime 1990s. Affordable home purchase lending refinance loans fall during heavy refinance years, as middle and upper income borrowers market, make up the least risky category.42 continued during the past four years, at an As discussed in Appendix A, the GSEs are even higher rate than earlier, particularly for dominate that market. On the other hand, in low refinancing years, the goals-qualifying involved in this market both through specific the two borrower-income goals (low-mod and program offerings and through purchases of special affordable). The average low-mod shares of refinance loans can equal or be greater than the goals-qualifying shares of securities backed by subprime loans percentage for home purchase loans during (including B&C loans as well as A-minus 1999–2002 was 44.6 percent, compared with home purchase loans. Second, and related, is the fact that subprime lending, which is loans). The B&C loans experience much 42.2 percent during 1995–98. Similarly, the higher delinquency rates than A-minus average special affordable percentage for characterized by relatively high goals- 43 qualifying shares, accounts for a smaller loans. home purchase loans during 1999–2002 was The procedure for excluding B&C portion of the single-family mortgage market 16.7 percent, compared with 15.1 percent mortgages from estimated ‘‘unadjusted’’ during heavy refinance years. Although they during 1995–98. Thus, the home lending market shares for goals-qualifying loans in market for lower-income borrowers were at a record dollar level ($213 billion) during 2002, subprime originations continued to grow. HUD’s earlier estimates 41 accounted for only 8.6 percent of all single- Estimates of the subprime market for other anticipated smaller shares of new mortgages recent years are as follows (dollar and market being originated for lower-income families. family mortgages originated that year, share): 1995 ($65 billion, 10 percent); 1996 ($96.5 Second, HUD’s projection model in the compared with about 13 percent during 1999 billion, 12.3 percent); 1997 ($125 billion, 15 2000 Rule assumed that refinance loans and 2000. Finally, the high volume of single- percent); 1998 ($150 billion, 10 percent; 1999 ($160 would have lower goals-qualifying family mortgages in a heavy refinance year billion, 12.5 percent); 2001 ($173 billion, 8.5 percentages than home purchase loans; this reduces the share of multifamily rental units. percent); 2002 ($213 billion, 8.6 percent). The assumption was based on the average home- For example, the multifamily share of all uncertainty about what these various estimates include should be emphasized; for example, they purchase-refinance differential between 1992 financed units was less than 14 percent in 1998, 2001, and 2002,40 compared to may include second mortgages and home equity and 1998. As discussed above, this has not loans as well as first mortgages, which are the focus been the case during ‘‘home purchase’’ years multifamily shares of 19 percent during of this analysis. The source for these estimates is such as 1995–97 and 1999–2000. Thus, the Inside Mortgage Finance (various years). projection model underestimates actual percentages for 2000, 2001, and 2002 were 84 42 The one-half assumption for A-minus loans is market experience when the goals-qualifying percent, 89 percent, and 80 percent, respectively. conservative because it probably underestimates shares of refinance loans turn out to be equal Studies of the coverage of HMDA data through 1996 (overestimates) the share of A-minus (B&C) loans. or greater than the goals-qualifying shares of conclude that HMDA covers approximately 85 According to data obtained by the Mortgage percent of the conventional conforming market, home purchase loans.37 This issue will be Information Corporation (see next footnote), 57 which suggests that HUD’s model produces percent of all subprime loans were labeled A-minus addressed further in the sections that present reasonable estimates of single-family-owner loans. (as of September 30, 2000). According to Inside B&C the new market estimates. For analysis of HMDA coverage, see Randall M. Lending, which is published by Inside Mortgage Third, the financing of multifamily Scheesele, HMDA Coverage of the Mortgage Market, Finance, the A-minus share of the subprime market properties continued at strong levels during op. cit. was 61.6 percent in 2000, 70.7 percent in 2001 (see 1999 and 2000. HUD’s baseline model in the 39 As discussed in Section C.6 of this appendix, March 11, 2002 issue), 75 percent in 2002 (see the 2000 Rule assumed a multifamily share of 15 there is some uncertainty about the multifamily mix September 15, 2003 issue), and 82 percent during percent, which was lower than the for the year 2002. The goals-qualifying shares the first nine months of 2003 (see the December 8, approximately 16–17 percent multifamily reported in Table D.9 assume $67.7 billion (the 2003 issue). HUD New estimate) and an average loan amount of 43 share during 1999 and 2000.38 As discussed The Mortgage Information Corporation (MIC) $37,275; this produces a multifamily mix of 10.9 reports the following serious delinquency rates percent. Section C.6 discussed several other (either 90 days past due or in foreclosure) by type 37 The 1995–2002 goals qualifying percentages for multifamily market and average loan amount of subprime loan: 3.36 percent for A-minus; 6.67 single-family mortgages are based on HMDA data estimates sfor 2002, each with a specific percent for B; 9.22 percent for C; and 21.03 percent for all (both home purchase and refinance) multifamily mixes. The low-mod, special for D. The D category accounted for only 2 percent mortgages. Thus, the implicit refinance rate is that affordable, and underserved areas shares for the of subprime loans and of course, is included in the reported by HMDA for conventional conforming other multifamily mixes discussed in Section C.6 ‘‘B&C’’ category referred to in this appendix. By mortgages. are as follows: 11.5 percent (54.4, 26.0, 32.25), 11.3 comparison, MIC reports a seriously delinquent rate 38 The accuracy of a single-family portion of percent (54.3, 25.9, 32.1), 11.0 percent (54.2, 25.8, of 3.63 percent for FHA loans. See MIC, The Market HUD’s model can be tested using HMDA data. The 32.0), 10.7 percent (54.0, 25.7, 31.9), 10.4 percent Pulse, Winter 2001, page 6. Also see ‘‘Subprime number of single-family-owner loans reported to (53.9, 25.6, 31.8), and 10.1 percent (53.8, 25.5, 31.8). Mortgage Delinquencies Inch Higher, Prepayments HMDA for the years 1999–2002 can be compared 40 Although data are not available yet, the Slow During Final Months of 1998’’, Inside MBS & with the corresponding number predicted by HUD’s multifamily share for 2003 will be lower than the ABS: Inside MBS & ABS, March 12, pages 8–11, model. Single-family-owner loans reported to approximately 11 percent in 2002. Senstivity where it is reported that fixed-rate A-minus loans HMDA during 1999 were 87 percent of the number analyses with lower multifamily mixes are provided have delinquency rates similar to high-LTV (over 95 of loans predicted by HUD’s model; comparable below. percent) conventional conforming loans.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00243 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24470 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

1999 combined information from several occurs because the B&C loans that were dropping B&C loans also assumes that the sources. First, the $160 billion estimate for dropped from the analysis had similar low- coverage of B&C and non-B&C loans in the subprime market was multiplied by 79.4 mod and special affordable percentages as HMDA’s metropolitan area data is the same; percent to arrive at an estimate of $127 the overall (both single-family and however, it is likely that HMDA coverage of billion for subprime loans less than the year multifamily) market. For example, the low- non-B&C loans is higher than its coverage of 1999 conforming loan limit of $240,000; the mod share of B&C loans was projected to be B&C loans.45 Despite these caveats, it also 79.4 percent estimate for the conforming 63.0 percent and HUD’s market model appears that reasonably different estimates of market was based on HMDA data for projected the overall low-mod share to be the various market parameters would not mortgages originated by subprime lenders. 58.6 percent. Thus, dropping B&C loans from likely change, in any significant way, the The $127 billion was reduced by one-half to the market totals does not change the overall above estimates of the effects of excluding arrive at an estimate of $63.5 billion for the low-mod share of the market. B&C loans in calculating the goals-qualifying conforming B&C market; with an average The situation is different for the shares of the market. As discussed below, loan amount of $78,801(obtained from Underserved Areas Goal. Underserved areas HUD provides a range of estimates for the HMDA data, as discussed below), the $63.5 account for 47.0 percent of the B&C loans, goals-qualifying market shares to account for billion represented approximately 806,081 which is a higher percentage than the uncertainty related to the various parameters B&C loans originated during 1999 under the underserved area share of the overall market included in its projection model for the conforming loan limit. (34.9 percent). Thus, dropping the B&C loans mortgage market. HMDA data was used to provide an leads to a reduction in the underserved areas Adjustment for Non-Metropolitan Areas. estimate of the portion of these 806,081 B&C market share of 1.0 percentage points, from HUD first estimated the underserved area loans that would qualify for each of the 34.9 percent to 33.9 percent. percentage for 1999–2002 based on single- housing goals. HMDA data does not identify Dropping B&C loans from HUD’s model family-owner parameters for metropolitan subprime loans, much less divide them into changes the mix between rental and owner areas. It was necessary to adjust these their A-minus and B&C components. As units in the final market estimate. Based on metropolitan-based market shares upward to explained in Appendix A, Randall assumptions about the size of the owner and reflect the fact that underserved counties Scheessele in HUD’s Office of Policy rental markets for 1999, HUD’s model account for a much larger portion of non- Development and Research has identified calculates that single-family-owner units metropolitan areas than underserved census almost 200 HMDA reporters that primarily accounted for 71.4 percent of total units tracts do of metropolitan areas. The originate subprime loans. The goals- financed during 1999. Dropping the B&C adjustment averaged about 1.5 percentage qualifying percentages of the loans originated owner loans, as described above, reduces the points; the method for deriving the upward by these subprime lenders in 1999 were as owner percentage of the market by 2.3 adjustment is explained in Section G.3 follows: 63.0 percent qualified for the Low- percentage points to 69.1 percent. Thus, below. and Moderate-Income Goal, 32.5 percent for another way of explaining why the goals- Manufactured Housing Loans. HUD the Special Affordable Goal, and 47.0 percent qualifying market shares are not affected so includes the effects of manufactured housing for the Underserved Areas Goal.44 Applying much by dropping B&C loans is that the loans (at least those financing properties in the goals-qualifying percentages to the rental share of the overall market increases as metropolitan areas) in its market estimates. estimated B&C market total of 806,081 gives the B&C owner units are dropped from the However, sensitivity analyses are conducted the following estimates of B&C loans that market. Since rental units have very high to determine the effects of excluding these qualified for each of the housing goals in goals-qualifying percentages, their increased loans. Excluding these loans from the market 1999: Low- and Moderate Income (507,831), importance in the market partially offsets the definition would reduce the 1995–2001 Special Affordable (261,976), and negative effects on the goals-qualifying shares estimates of the three goals-qualifying market Underserved Areas (378,858). of any reductions in B&C owner loans. In shares by approximately one percentage Adjusting HUD’s model to exclude the B&C fact, this rental mix effect would come into point. Assuming a home purchase market involves subtracting the above four play with any reduction in owner units from environment (1995–97 and 1999–2000) and a figures’ one for the overall B&C market and HUD’s model. constant mix of owner and rental properties, three for B&C loans that qualify for each of Dropping all subprime loans (both A- excluding manufactured housing loans (as the three housing goals ’’ from the minus and B&C) from the market definition well as loans less than $15,000) would corresponding figures estimated by HUD for would lead to similar results for the Low- reduce the goals-qualifying shares reported in the total single-family and multifamily Mod and Special Affordable Goals ’’ little Table D.9 roughly as follows: Low- and market inclusive of B&C loans. HUD’s model change in the market estimates for the Moderate-Income Goal by 1.2 percentage estimates that 10,638,797 single-family and reasons given above (the low-mod estimate points, Special Affordable Goal by 1.0 multifamily units were financed during 1999; falls to 57.8 percent and the special percentage points, and Underserved Areas of these, 6,229,569 (58.6 percent) qualified affordable share falls to 28.9 percent). The Goal by 0.8 percentage point. (The method for the Low- and Moderate-Income Goal, market estimate for the Underserved Areas for calculating these reductions is explained 3,133,701 (29.5 percent) for the Special Goal would fall an additional 1.2 percentage in Section F.3b below.) Dropping Affordable Goal, and 3,711,271 (34.9 percent) points to 32.7 percent (or 2.2 percentage manufactured housing from the market totals for the Underserved Areas Goal. Deducting points lower than the overall estimate of 34.9 would increase the rental share of the the B&C market estimates produces the percent). following adjusted market estimates: a total As discussed in the 2000 Rule, there are 45 Dropping B&C loans in the manner described market of 9,983.276, of which 5,721,738 (58.2 caveats that should be mentioned concerning percent) qualified for the Low- and Moderate- in the text results in the goals-qualifying the above adjustments for the B&C market for percentages for the non-B&C market being Income Goal, 2,871,725 (29.2 percent) for the 1999. The adjustment for B&C loans depends underestimated since HMDA coverage of B&C loans Special Affordable Goal, and 3,332,413 (33.9 on several estimates relating to the 1999 is less than that of non-B&C loans and since B&C percent) for the Underserved Areas Goal. mortgage market, derived from various loans have higher goals-qualifying shares than non- As seen, the low-mod market share B&C loans. For instance, the low-mod shares of the sources. Different estimates of the size of the estimate exclusive of B&C loans (58.2 market reported in Table D.9 underestimate (to an B&C market in 1999 or the goals-qualifying percent) is practically the same as the unknown extent) the low-mod shares of the market shares of the B&C market could lead to original market estimate (58.6 percent), as is inclusive of B&C loans; so reducing the low-mod different estimates of the goals-qualifying also the special affordable market estimate owner shares by dropping B&C loans in the manner shares for the overall market. The goals- described in the text would provide an (29.5 percent versus 29.2 percent). This qualifying shares of the B&C market were underestimate of the low-mod share of the non-B&C based on HMDA data for selected lenders owner market. A study of 1997 HMDA data in 44 The goals-qualifying percentages for subprime that primarily originate subprime loans; since Durham County, North Carolina by the Coalition for lenders are much higher than the percentages (46.3 these lenders are likely originating both A- Responsible Lending (CRL) found that loans by percent, 18.3 percent, and 28.2 percent, mortgage and finance companies are often not respectively) for the overall single-family minus and B&C loans, the goals-qualifying reported to HMDA. For a summary of this study, conventional conforming market in 1999. For percentages used here may not be accurately see ‘‘Renewed Attack on Predatory Subprime further analysis of subprime lenders, see Randall M. measuring the goals-qualifying percentages Lenders’’ in Fair Lending/CRA Compass, June 9, Scheessele, 1998 HMDA Highlights, op. cit. for only B&C loans. The above technique of 1999.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00244 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24471

mortgage market, which would tend to above and throughout this proposed Rule moderate-income percentages are given in increase the goals-qualifying shares and thus pertain only to manufactured housing loans Table D.10. Case 1 projections represent the partially offset the reductions reported above. in metropolitan areas, as measured by loans baseline and intermediate case; it assumes In addition, the estimated reductions in originated by the 21 manufactured housing that investors account for 10 percent of the goals-qualifying shares due to excluding lenders identified by HUD. single-family mortgage market. Case 2 manufactured housing are even lower during b. Estimates of the Low- and Moderate- assumes a lower investor share (8 percent) the heavy refinance years such as 1998 and Income Market based on HMDA data and slightly more 2001. It should also be mentioned that This section provides HUD’s estimates for conservative low- and moderate-income manufactured housing in non-metropolitan the size of the low- and moderate-income percentages for single-family rental and areas is not included in HUD’s analysis due mortgage market that will serve as a proxy for multifamily properties (85 percent). Case 3 to lack of data; including that segment of the the four-year period (2005–2008) when the assumes a higher investor share (12 percent) market would increase the goals-qualifying new housing goals will be in effect. Three consistent with Follain and Blackley’s shares of the overall market. Thus, the alternative sets of projections about property suggestions. analyses of manufactured housing reported shares and rental property low- and BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00245 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24472 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00246 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.078 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24473

Because single-family-owner units account estimates for different low-mod percentages market estimates in Table D.11 exclude B&C for about 70 percent of all newly mortgaged for the owner market as well as for different loans, in the same manner as discussed dwelling units, the low- and moderate- multifamily mix percentages—15.0 percent earlier for the 1995–2001 market estimates. income percentage for owners is the most bracketed by 13.5 percent and 16.5 percent, This is explained further below. important determinant of the total market which are the same multifamily mixes estimate. Thus, Table D.11 provides market assumed in the 2000 Rule. The low-mod

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00247 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24474 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Table D.11 assumes a refinance rate of 35 home purchase or low-refinancing market estimates reflecting heavy refinance percent, which means that the table reflects environments. After presenting these results, environments will be presented. Because of

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00248 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03my04.079 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24475

the increase in single-family mortgages, the reflect the year 2000 environment, which had analyses to determine the effects on the multifamily share of the mortgage market a low-mod home purchase percentage of 45 overall market estimate of different typically falls during a heavy refinance percent combined with a higher low-mod assumptions about the size of that market. As environment; therefore, several sensitivity refinance percentage of 52 percent. Of course, discussed in Section C of this appendix, the analyses using lower multifamily mixes are the 47 percent low-mod share for the overall average multifamily share between 1991 and examined below. single-family-owner market could be 2002 was approximately 16 percent, so 15 In the 2000 Rule, HUD assumed that the consistent with other combinations of low- percent represents a slightly more low-mod share of refinance loans was three mod shares for home purchase and refinance conservative baseline. In addition, in single- percentage points lower than the low-mod loans. In this case, a 47 percent assumption family home purchase (or low refinancing) share of borrowers purchasing a home. for the overall single-family-owner market environments, the multifamily mix has However, as discussed earlier, the low-mod produces an estimate of 59.0 percent for the typically been above 16 percent. Therefore, share of refinance loans has equaled or been low-mod share of the overall (owner and when considering single-family home greater than the low-mod share of home rental) market, excluding B&C loans. purchase environments, it is probably more purchase loans during recent home purchase While both approaches will be discussed appropriate to focus on the top two environments such as 1995–97 or 1999–2000; below, most of the discussion will focus on multifamily mixes (15 percent and 16.5 thus, the assumption of a lower low-mod the first approach. It should be noted that percent) in Table D.11. Still, given the shares for refinance loans is initially dropped several low-mod percentages of the owner uncertainty surrounding the size of the for this analysis but will be reintroduced market are given in Table D.11 to account for multifamily market, it is useful to consider during the sensitivity analysis and during the different perceptions of that market. the effects of lower multifamily mix discussion of heavy refinance environments. Essentially, HUD’s approach throughout this assumptions, even in a home purchase There are two ways to view the single- appendix is to provide several sensitivity environment. Assuming a 13.5 percent family-owner low-mod percentages reported analyses to illustrate the effects of different multifamily mix reduces the overall low-mod in the first column of Table D.11. A first views about the goals-qualifying share of the market estimates by 0.6–0.7 percentage approach would be to view them as single-family-owner market. This approach points compared with a 15 percent mix, and representing low-mod percentages of only the recognizes that there is some uncertainty in by 1.2–1.4 percentage points compared with home purchase market. For example, a low- the data and that there can be different a 16.5 percent mix. For example, when the mod percentage for home purchase loans of viewpoints about the various market low-mod share of the home purchase market 43 percent (as it was say in 1997)—combined definitions and other model parameters. is at 43 percent, the low-mod share of the with the assumption of an equal low-mod Market Estimates. As shown in Table D.11, overall market is 55.3 percent assuming a share for refinance loans (i.e., also 43 the market estimate is: 57–58 percent if the 13.5 percent multifamily mix, compared with percent) and with the other model owner percentage is 45 percent (home 55.9 (56.6) percent assuming a 15 (16.5) assumptions (such as a multifamily mix of 15 purchase share for 1999, 2000, and 2002); percent multifamily mix. The next section percent)—produces an estimate of 55.9 55–57 percent if the owner percentage is 43 examines the effects of multifamily mixes percent for the low-mod share of the overall percent (home purchase share for 1998 and lower than 13.5 percent. (owner and rental) market, excluding B&C 2001); and 54–55 percent if the owner Heavy Refinancing Environments. As loans. Thus, the reader can view Table D.11 percentage is 42 percent (home purchase shown earlier in Table D.11, the low-mod as showing the overall low-mod market average from 1995–97). If the low- and share of the overall market declines when estimate once the reader specifies his or her moderate income percentage for home refinances dominate the market. Compared views about the low-mod share of the single- purchase loans fell to 38 percent—or five with low-mod market shares of 57–59 family home purchase market (given the percentage points from its 1995–2001 average percent during recent home purchase other model assumptions). In this case, if the level of 43 percent—then the overall market environments (1995–97 and 1999–2000), the reader believes that the low-mod share of estimate would be about 52 percent. Thus, 52 low-mod share declined to 54–55 percent refinance loans should be lower than that for percent is consistent with a rather significant during 1998, 2001, and 2002—three years home purchase loans, the reader simply has decline in the low-mod share of the single- where refinancing dominated the single- to multiply the differential amount by 0.35 family home purchase market. If the low-mod family-owner mortgage market. As explained (which is the refinance share of single- percentage for home purchase loans fell earlier, this decline in the low-mod market family-owner loans) and 0.722 (which is the further to 35 percent (or 8 percentage points share during heavy refinancing periods is single-family-owner share of all dwelling below its 1995–2002 average of 43 percent), due to (a) a decline in the low-mod share of units in the baseline model that assumes a 15 the overall market estimate would still be single-family refinance mortgages as middle- percent multifamily mix). For example, approximately 50 percent. Under the baseline and upper-income borrowers dominate the applying the assumption in the 2000 Rule projection, the home purchase percentage refinance market; (b) a decline in the relative that the low-mod share is three percentage can fall as low as 34 percent—about four- importance of the subprime market; and (c) points lower for refinance loans would fifths of the 1995–2002 average—and the a decline in the share of multifamily reduce the overall low-mod share of the low- and moderate-income market share mortgages. For example, during 2001, the market by 0.8 percentage points (3.0 times would still be 49–50 percent. refinance share of low-mod loans fell to 41.8 0.35 times 0.722). In this manner, the reader The market estimates reported in Table percentage points (from about 49 percent can easily adjust the market estimates D.11 for Case 2 and Case 3 bracket those for during 1999 and 2000); the subprime share reported in Table D.11 to incorporate his or Case 1 (the baseline). The smaller single- of the single-family market fell to 8.5 percent her own views about differences in the low- family rental market and lower low- and (from about 13 percent during 1999 and mod share of home purchase and refinance moderate-income percentages for rental 2000); and the multifamily share of the loans. properties result in the Case 2 estimates market fell to 13.4 percent (from about 16 A second approach would be to view the being about one and a half percentage points percent during 1999 and 2000). Similarly low-mod percentages (in the first column of below the Case 1 estimates. Conversely, the during 2002, the low-mod share of refinance Table D.11) as representing low-mod shares higher percentages under Case 3 result in loans was 42.3 percent, the subprime share for the overall single-family-owner market, estimates of the low-mod market of the market was 8.6 percent, and the including both home purchase and refinance approximately two percentage points higher multifamily mix was approximately 11 loans. This approach does not specify than the Case 1 estimates. As discussed in percent. separate low-mod percentages for home Section D, the baseline Case 1 is a reasonable Several assumptions were changed to purchase and refinance loans, but rather approach for estimating the market shares. incorporate a refinance environment into the focuses on the overall single-family-owner Multifamily Mix. The volume of projection model for 2005–08. The refinance environments. Thus, it allows for mortgage multifamily activity is also an important share of single-family mortgages was market environments where the low-mod determinant of the size of the low- and increased to 65 percent, or almost double the share of refinance loans is greater than the moderate-income market. HUD is aware of 35 percent refinance rate assumed in the low-mod share for home purchase loans. For the uncertainty surrounding projections of projection model for a ‘‘home purchase’’ example, a low-mod percentage for single- the multifamily market and consequently environment. The market share for subprime family-owner loans of 47 percent would recognizes the need to conduct sensitivity loans was assumed to be 8.5 percent and the

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00249 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24476 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

multifamily mix, 13.5 percent. The low-mod The various market estimates presented in illustrated using an example of a low-mod share for refinance loans was assumed to be Table D.11 for a home purchase environment percentage of 43 percent for single-family- 39 percent, or four percentage points below and reported above for a refinance owner loans. Again, as explained earlier, this the assumed low-mod share of home environment are not all equally likely. Most 43 percent figure could reflect a mortgage purchase loans (which was set at the 1998 of them equal or exceed 52 percent. In the market environment where home purchase and 2002 level of 43 percent). Under these home purchase environment, estimates and refinance loans had similar low-mod assumptions, the overall low-mod market below 52 percent would require the low-mod percentages (i.e., 43 percent) or a mortgage share (excluding B&C loans) was projected to share of the single-family-owner market for market environment where home purchase be 53.4 percent—or about 1–2 percentage home purchase loans to drop to 36–37 and refinance loans had different low-mod points below the market shares estimated for percent, which would be 6–7 percentage market percentages that together resulted in 1998, 2001, and 2002. If the multifamily mix points below the average. Dropping below 52 a 43 percent average for the single-family- is reduced further to 12 (10) percent, the percent would be more likely in a heavy owner market. market projection falls to 52.7 (51.8) percent. refinance environment, as the actual As Table D.11 shows, a 43 percent low- If the single-family low-mod percentages are estimated market shares during 1998, 2001, mod share for owner mortgages translates reduced to 41 percent (home purchase) and and 2002 were in the 54–55 percent range. into an overall low-mod market share of 55.9 37 percent (refinance), and the multifamily However, sensitivity analyses of a refinance percent. It is assumed that the subprime mix is 12 (10) percent, the overall low-mod environment showed that a 52 percent low- market accounts for 12 percent of all market share falls 51.1 (50.2) percent. Since mod market share was consistent with mortgages originated, which would be $204 refinance environments are characterized by market assumptions more adverse than the billion based on $1,700 billion for the low interest rates, it is unlikely that the low- heavy refinance years of 1998, 2001, and mortgage market. This $204 billion estimate mod share of the home purchase market 2002. for the subprime market is reduced by 20 would fall below 41 percent, given that it has B&C Loans. There are two possible percent to arrive at $163.2 billion for averaged 43 percent over the past eight years. approaches for adjusting for the effects of subprime loans that will be less than the To further examine this issue in the B&C loans in the projection model. First, conforming loan limit. This figure is reduced context of an actual refinance environment, readers could choose a single-family low- by one-half to arrive at $81.6 billion for the the various parameters (e.g., low-mod share mod percentage (that is, one of the conforming B&C market; with an average of home purchase and refinance loans for percentages in the first column in Table D.11) loan amount of $129,899; the $81.6 billion owner and rental properties, the subprime that they believe is adjusted for B&C loans represents 628,180 B&C loans projected to be share of the market, etc.) for the year 2002 and then obtain a rough estimate of the originated under the conforming loan limit. were used except that the multifamily mix overall market estimate from the second to Following the procedure discussed in was lowered from the actual level in 2002. fourth columns corresponding to different Section F.3a, the low-mod share of the During 2002, there was a three percentage multifamily mixes. For instance, if one market exclusive of B&C loans is estimated point differential between the low-mod share believes the appropriate single-family-owner to be 55.9 percent (see Table D.11), which is of home purchase loans (45.3 percent) and percentage adjusted for B&C loans (or only slightly lower than the original refinance loans (42.3 percent). As reported adjusted for any other market sectors that the (unadjusted) estimate of 56.1 percent.48 As earlier, the low-mod share of the 2002 market reader thinks appropriate) is 39 percent, then noted earlier, this occurs because the B&C was estimated to be 54.4 percent assuming a the low-mod market estimate is 52.7 percent loans that were dropped from the analysis multifamily mix of 11.5 percent, and 10.9 assuming a multifamily mix of 15 percent. had similar low-mod percentages as the percent assuming a multifamily mix of 10.9 While intuitively appealing, such an overall (both single-family and multifamily) percent. The multifamily mix for a year such approach would provide inaccurate results, market (58.6 percent for excluded B&C loans as 2003, characterized by single-family as explained next. versus 56.1 percent for the overall, originations of $3.3 trillion, will certainly be Second, readers could choose a single- unadjusted market estimate). The impact of lower than the 11 percent multifamily mix of family-owner percentage directly from dropping B&C loans is larger when the 2002, characterized by $2.5 trillion in single- HMDA data that is unadjusted for B&C loans overall market share for low-mod loans is family originations. Thus, this sensitivity and then rely on HUD’s methodology smaller. If the low-mod share for single- analysis reduces the multifamily mix for the (described below) for excluding the effects of family owners is assumed to be 38 percent, 2002 refinance environment. The low-mod B&C loans. This is the approach taken in dropping B&C loans would reduce the low- shares vary with the multifamily mix as Table D.11. The advantage of the second mod market share by 0.4 percentage points, follows: (53.8 percent low-mod share, 10 approach is that HUD’s methodology makes from 52.5 percent to the 52.1 percent percent multifamily mix); (53.3 percent, 9 the appropriate adjustments to the various reported in Table D.11. Still, dropping B&C percent); (52.9 percent, 8 percent); 52.5 property shares (i.e., the owner versus rental loans from the market totals does not change percent, 7 percent); and (52.1 percent, 6 percentages) that result from excluding the overall low-mod share of the market percent). Thus, under the actual 2002 single-family B&C loans from the analysis. appreciably. assumptions, the low-mod share drops by According to HUD’s methodology, dropping Dropping B&C loans from HUD’s projection about one-half percentage point for each one B&C loans would reduce the various low- model changes the mix between rental and percentage point reduction in the mod market estimates by less than half of a owner units in the final market estimate; multifamily mix.46 The low-share remains percentage point. This minor effect is due to above 52 percent even if the multifamily mix (a) the fact that the low-mod share of B&C 48 falls to 6 percent.47 1999–2002 HMDA data for subprime lenders loans is similar to that of the overall market; were used to provide an estimate of 58.6 percent and (b) the offsetting effects of the increase for the portion of the B&C market that would 46 This analysis assumes the 2002 refinance rate in the rental market share when single-family qualify as low- and moderate-income. Applying the of 62 percent; if the refinance rate is increased to B&C loans are dropped from the market 58.6 percentage to the estimated B&C market total 65–68 percent (current predictions for 2003), then totals. of 628,180 gives an estimate of 367,957 B&C loans the overall low-mod market percentages in this As noted above, if one assumes the single- that would qualify for the Low- and Moderate- sentence would decline by about 0.1 percentage Income Goal. Adjusting HUD’s model to exclude point. If there were a four (five) percentage point family-owner percentages in the first column the B&C market involves subtracting the 628,180 difference between the low-mod shares of home of Table D.11 are unadjusted for B&C loans, B&C loans and the 367,957 B&C low-mod loans purchase and refinance loans, rather than a three then the overall low-mod market estimates from the corresponding figures estimated by HUD percentage point difference as in 2002, then the must be adjusted to exclude these loans. B&C for the total single-family and multifamily market overall low-mod market percentages in this loans were deducted in HUD’s projection inclusive of B&C loans. HUD’s projection model sentence would decline by about 0.5 (1.0) model using the same procedure described estimates that 10,632,145 single-family and percentage point. earlier for the 1995–2002 market estimation multifamily units will be financed and of these, 47 For a given multifamily mix, the low-mod models. The effects of deducting the B&C 5,962,527 (56.1 percent) will qualify for the Low- shares of the market are higher under the and Moderate-Income Goal. Deducting the B&C simulations based on the 2002 environment, as loans from the projection model can be market estimates produces the following adjusted compared with the simulations reported in the market estimates: a total market of 10,003,964 of above paragraph based on the projection model. the various property types were higher during 2002 which 5,594,570 (55.9 percent) will qualify for the The reason for this is that the low-mod shares for than those assumed in the projection model. Low- and Moderate-Income Goal.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00250 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24477

rental units accounted for 29.6 percent of anticipated effects of re-benchmarking remain unchanged. A $400 billion increase total units after dropping B&C loans metropolitan area incomes based on 2000 would reduce the low-mod projected market compared with 27.8 percent before dropping Census data and incorporating the new OMB share by one percentage point. These B&C loans. Since practically all rental units definitions for metropolitan areas. reductions in the low-mod share of the qualify for the low-mod goal, their increased c. Economic Conditions, Market Estimates, mortgage market share occur because the importance in the market partially offsets the and the Feasibility of the Low- and Moderate- multifamily mix is reduced from 15 percent negative effects on the goals-qualifying shares Income Housing Goal to 13.6 percent to 12.5 percent. As explained of any reductions in B&C owner loans. in Section E, the absolute volume of single- During the 2000 rule-making, there was a A similar analysis can be used to family originations (such as the $1,700 concern that the market share estimates and demonstrate the effects of deducting the billion) is not as important as the relative the housing goals failed to recognize the remaining, A-minus portion of the subprime shares of single-family and multifamily rental volatility of housing markets and the market from the market estimates. Of course, units. existence of macroeconomic cycles. There deducting A-minus loans as well as B&C Recent years have been characterized by was particular concern that the market shares loans is equivalent to deducting all subprime record affordability conditions due to low and housing goals were based on a period of loans from the market. In the example given interest rates and economic expansion. Thus, above (43 percent low-mod percentage for economic expansion accompanied by record low interest rates and high housing HUD also examined potential changes in the owners), deducting all subprime loans would market shares under very different further reduce the overall low-mod market affordability. This section discusses these issues, noting that the Secretary can consider macroeconomic environments, including estimate to 55.7 percent. Thus, the periods of recession, high interest rates, and unadjusted low-mod market estimate is 56.1 shifts in economic conditions when evaluating the performance of the GSEs on heavy refinancing (accompanied by low percent, the estimate adjusted for B&C loans interest rates). A recessionary environment is 55.9 percent (reported in Table D.11), and the goals, and noting further that the market share estimates can be examined in terms of would likely be characterized by a reduction the estimate adjusted for all subprime loans in single-family activity (or an increase in the is 55.7 percent. less favorable market conditions than have existed during the 1993 to 2002 period. multifamily share of the market) and a Section F.3.a discussed several caveats reduction in the low-mod shares of the concerning the analysis of subprime loans. It Volatility of Market. Changing economic conditions can affect the validity of HUD’s single-family-owner market. The low- and is not clear what types of loans (e.g., first market estimates as well as the feasibility of moderate-income share of the home purchase versus second mortgages) are included in the the GSEs’ accomplishing the housing goals. market was reduced to 34 percent, or 10.6 subprime market estimates. There is only The volatile nature of the mortgage market in percentage points lower than its 1999–2002 limited data on the borrower characteristics the past few years suggest a degree of average share. Under these rather severe of subprime loans and the extent to which uncertainty around projections of the conditions, the overall market share for the these loans are included in HMDA is not origination market. Large swings in Low- and Moderate-Income Goal would clear. Still, the above analysis demonstrates refinancing, consumers switching between decline to 49.0 (49.8) percent, assuming a that the projection model can incorporate the adjustable-rate mortgages and fixed-rate multifamily mix of 15.0 (16.5) percent. If the effects of dropping B&C loans (or even all mortgages, and increased first-time low-mod share of the owner market were subprime loans) from the final market homebuyer activity due to record low interest reduced more modestly to 37 percent, the estimates. rates, have all characterized the mortgage low-mod share for the overall market would Manufactured Housing Loans. Excluding market during the nineties. These conditions fall to 51.3 percent assuming a multifamily manufactured housing loans (as well as small are beyond the control of the GSEs but they mix of 15.0 percent. (See Table D.11.) loans less than $15,000) reduces the overall would affect their performance on the As explained above, several heavy market estimates reported in Table D.11 by housing goals. A mortgage market dominated refinance environments were simulated. As a one-percentage point. This is estimated as by heavy refinancing on the part of middle- way of examining more extreme refinance follows. First, excluding these loans reduces income homeowners would reduce the GSEs’ environments than 2002, the effects of the unadjusted low-mod percentage for ability to reach a specific target on the Low- reducing the multifamily mix for the 2002 single-family-owner mortgages in and Moderate-Income Goal, for example. A refinance environment were examined. The metropolitan areas by about 1.8 percentage jump in interest rates would reduce the low-mod shares varied with the multifamily points, based on analysis of recent home availability of very-low-income mortgages for mix from 53.8 percent low-mod share with a purchase environments (1995–97 and 1999 the GSEs to purchase. But on the other hand, 10 percent multifamily mix to 52.1 percent and 2000). Multiplying this 1.8 percentage the next few years may be favorable to with a 6 percent multifamily mix. Under the point differential by the property share achieving the goals because of the high actual 2002 market assumptions, the low- (0.722) of single-family-owner units yields refinancing activity in 2001, 2002, and 2003. mod share drops by about one-half 1.3 percentage points, which serves as a A period of low-to-moderate interest rates percentage point for each one percentage proxy for the reduction in the overall low- would sustain affordability levels without point reduction in the multifamily mix.49 mod market share due to dropping causing the rush to refinance seen earlier in manufactured home loans from the market 1998 and 2001–2003. A high percentage of 49 analysis. The actual reduction will be This analysis assumes the 2002 refinance rate potential refinancers have already done so, of 62 percent; if the refinance rate is increased to somewhat less because dropping and are less likely to do so again. However, 65–68 percent (current predictions for 2003), then manufactured home loans will increase the these same predictions were made after the the overall low-mod market percentages in this share of rental units, which increases the 1998 refinance wave, which indicates the sentence would decline by about 0.1 percentage overall low-mod market share, thus partially uncertainty of making predictions about the point. If there were a four (five) percentage point offsetting the 1.3 percent reduction. The net mortgage market. difference between the low-mod shares of home effect is probably a reduction of about one HUD conducted numerous sensitivity purchase and refinance loans, rather than a three percentage point difference as in 2002, then the percentage point. analyses of the market shares, several of The above analysis of the effects of overall low-mod market percentages in this which were described in Section F.3b above. sentence would decline by about 0.5 (1.0) dropping different categories of loans from The starting point of HUD’s estimates is the perecentage point. In addition, due to the the market suggest that 52–58 percent is a projected $1,700 billion in single-family uncertainty surrounding estimates of the investor reasonable range of estimates for the low- and originations. Increasing the single-family share of the single-family mortgage market (see moderate-income market. This range covers mortgage origination forecast while holding Section D), the analysis assumes a constant 10 markets without B&C and allows for market the multifamily origination forecast constant percent share for investors; if the investor share is environments that would be much less is equivalent to reducing the multifamily reduced to 8 percent during a refinance affordable than recent market conditions. The mix. Increasing the single-family projection environment, the estimated low-mod share of the next section presents additional analyses market would fall about one percentage point. This by $200 billion, from $1,700 billion to $1,900 figure is obtained by multiplying the low-mod related to market volatility and affordability billion, would reduce the market share for percentage differential between owner and investor conditions. After that, a one-percentage point the Low- and Moderate-Income Goal by mortgages (about 47 percent) by the resulting downward adjustment is made to the 52–58 approximately 0.6 percentage point, decimal point increase in the share of owner units percent market range to reflect the assuming the other baseline assumptions (.021 as shown in Table D.7).

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00251 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24478 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Affordability Conditions and Market (unweighted average of 1999–2002 e. Conclusions About the Size of Low- and Estimates. As discussed in Appendix A, percentages in Table D.8); the corresponding Moderate-Income Market record low interest rates, a more diverse average with the projected data was 43.4 Based on the above findings as well as socioeconomic group of households seeking percent, yielding a differential of 1.2 numerous sensitivity analyses, HUD homeownership, and affordability initiatives percentage points. For home purchase loans concludes that 51–57 percent is a reasonable of the private sector have encouraged first- in the conventional conforming market, the range of estimates of the mortgage market’s time buyers and low-income borrowers to projected low-mod percentages for each year low- and moderate-income share for the year enter the market since the mid-1990s. A between 1999 and 2002 were as follows (with 2005 and beyond. This range covers much significant increase in interest rates over the historical data from Table D.8 in more adverse economic and market recent levels would reduce the presence of parentheses): 44.4 (45.2) percent for 1999; affordability conditions than have existed low-income families in the mortgage market 44.2 (44.8) percent for 2000; 41.8 (43.2) recently, allows for different assumptions and the availability of low-income mortgages percent for 2001; and 43.3 (45.3) percent for about the multifamily market, and excludes for purchase by the GSEs. As discussed 2002. The differentials between the projected the effects of B&C loans. HUD recognizes that above, the 52–58 percent range for the low- and historical data are larger in 2001 (1.4 shifts in economic conditions and mod market share covers economic and percentage points) and 2002 (2.0 percentage refinancing could increase or decrease the market affordability conditions much less points) than in 1999 (0.8 percentage point) size of the low- and moderate-income market favorable than recent conditions of low and 2000 (0.6 percentage point). For total during that period. interest rates and economic expansion. The (both home purchase and refinance) loans, low-mod share of the single-family home G. Size of the Conventional Conforming the average low-mod share of the purchase market could fall to 38 percent, Market Serving Central Cities, Rural Areas, conventional conforming market based on which is 5.2 percentage points lower than its and Other Underserved Areas historical data was 44.8 percent (unweighted 1995–2002 average level of 43.2 percent, The following discussion presents before the baseline market share for the Low- average of 1999–2002 percentages in Table D.8); the corresponding average with the estimates of the size of the conventional and Moderate-Income Goal would below 52 conforming market for the Central City, Rural percent. projected data was 43.6 percent, again yielding a differential of 1.2 percentage Areas, and other Underserved Areas Goal; Feasibility Determination. As stated in the this housing goal will also be referred to as 2000 Rule, HUD is well aware of the points, with the same pattern exhibited for the annual differentials.51 It appears that the the Underserved Areas Goal. The first three volatility of mortgage markets and the sections, which analyze historical data going possible impacts on the GSEs’ ability to meet low-mod share for single-family-owners in the conventional conforming market will be back to the early 1990’s, necessarily used the housing goals. FHEFSSA allows for 1990 Census geography to define 50 at least one percentage point less due to the changing market conditions. If HUD has set underserved census tracts and underserved re-benchmarking of area median incomes and a goal for a given year and market conditions counties. The first two sections focus on the new OMB definitions of metropolitan change dramatically during or prior to the underserved census tracts in metropolitan areas. year, making it infeasible for the GSE to areas, as Section 1 presents underserved area For the other two property types (single- attain the goal, HUD must determine percentages for different property types while family rental and multifamily), comparisons ‘‘whether (taking into consideration market Section 2 presents market estimates for between projected and historical low-mod and economic conditions and the financial metropolitan areas. Section 3 discusses B&C percentages were made using the GSEs’ data. condition of the enterprise) the achievement loans and rural areas. But as explained in For single-family rental mortgages, the of the housing goal was or is feasible.’’ This Appendix B, HUD will be defining unweighted average of Fannie Mae’s (Freddie provision of FHEFSSA clearly allows for a underserved areas based on 2000 Census finding by HUD that a goal was not feasible Mac’s) low-mod percentage for the years geography beginning in 2005, the first year due to market conditions, and no subsequent 1999 to 2002 was 87.8 (88.1) percent using covered by this proposed rule. Therefore, actions would be taken. As HUD noted in the projected data, compared with 87.7 (88.1) Section 4 repeats much of the analyses in both the 1995 and 2000 GSE Rules, it does percent using the historical data. For Sections 1–3 but in terms of 2000 Census not set the housing goals so that they can be multifamily mortgages, the unweighted geography, rather than 1990 Census met even under the worst of circumstances. average of Fannie Mae’s (Freddie Mac’s) low- geography. Rather, as explained above, HUD has mod percentage for the years 1999 to 2002 conducted numerous sensitivity analyses for was 92.1 (90.3) percent using the projected 1. Underserved Areas Goal Shares by economic and market affordability data, compared with 92.9 (92.6) percent Property Type environments much more adverse than has using the historical data. These comparisons For purposes of the Underserved Areas existed in recent years. If macroeconomic suggest little difference between the Goal, underserved areas in metropolitan conditions change even more dramatically, projected and historical low-mod shares for areas are defined as census tracts with: the levels of the goals can be revised to rental properties. HUD also projected the (a) Tract median income at or below 90 reflect the changed conditions. FHEFSSA overall low-mod goal percentage for each percent of the MSA median income; or and HUD recognize that conditions could GSE. For the overall low-mod goal (b) A minority composition equal to 30 change in ways that require revised (considering all three property types), the percent or more and a tract median income expectations. unweighted average of Fannie Mae’s (Freddie no more than 120 percent of MSA median d. New 2000 Census Data and New OMB Mac’s) low-mod percentage for the years income. Metropolitan Area Definitions 1999 to 2002 was 48.5 (47.1) percent using Owner Mortgages. The first set of numbers Going forward, HUD will be re- the projected data, compared with 49.1 (47.9) in Table D.12 are the percentages of single- benchmarking its median incomes for percent using the historical data. Compared family-owner mortgages that financed metropolitan areas and non-metropolitan with the historical data, the projected data properties located in underserved census counties based on 2000 Census median reduces Fannie Mae’s average low-mod tracts of metropolitan areas between 1992 incomes, and will be incorporating the effects percentage by 0.6 percentage points, and and 2002. There are several interesting of the new OMB metropolitan area Freddie Mac’s by 0.8 percentage point. patterns in these data. During 1999 and 2000, definitions. HUD projected the effects of Based on the above analysis, it appears the 28–30 percent of mortgages (both home these two changes on the low- and moderate- low-mod share of the conventional purchase and refinance loans) financed income shares of the single-family-owner conforming market is about one percentage properties located in these areas; this market for the years 1999–2002. Under the point less when based on projected data, as percentage fell to 25.7 percent in 2001 and historical data, the average low-mod share of compared with historical data. Thus, it seems 25.2 percent in 2002, figures that were the conventional conforming market was 44.6 appropriate to drop the 52–58 percent market slightly below the average (26.8 percent) percent for home purchase loans range to 51–57 percent. between 1994 and 1998. In 1992 and 1993,

50 Section 1336(b)(3)(A). 51 Between 1999 and 2002, the average single- point for Fannie Mae and 1.3 percentage point for family-owner differential between the historical and Freddie Mac. projected low-mod percentages was 1.1 percentage

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00252 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24479

the underserved areas share of single-family- owner mortgages was only 20 percent. BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00253 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24480 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00254 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.080 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24481

In most years, refinance loans are more Considering all (both home purchase and has been in the almost 45 percent range over likely than home purchase loans to finance refinance) loans during recent ‘‘home the past nine years. HMDA data also show properties located in underserved census purchase’’ environments, the underserved that about half of newly-mortgaged tracts. Between 1994 and 2002, 28.5 percent areas share was a high 28–30 percent during multifamily rental units are located in of refinance loans were for properties in 1999–2000, compared with a 27.1 percent underserved areas. underserved areas, compared to 25.6 percent average between 1995 and 1997; excluding of home purchase loans. This refinance- B&C and other (i.e., A-minus) subprime loans 2. Market Estimates for Underserved Areas in home-purchase differential is mostly due to places 1999 on par with the earlier years, Metropolitan Areas the influence of subprime loans. Excluding with only the year 2000 showing a higher In the 2000 GSE Rule, HUD estimated that B&C (all subprime) loans and considering the level of underserved area lending than the market share for underserved areas would same time period, 27.2 (25.6) percent of occurred during 1995–97. These data refinance loans were for properties in indicate that the single-family-owner market be between 29 and 32 percent. This estimate underserved areas, compared to 25.2 (24.8) in underserved areas has remained strong turned out to be below market experience, as percent of home purchase loans. In the year since the 2000 Rule was written. While it is underserved areas accounted for (2000) with the largest differential, excluding recognized that economic and housing approximately 32–35 percent of all mortgages B&C (all subprime) loans reduced the affordability conditions could change and originated in metropolitan areas between refinance-home-purchase differential from reduce the size of the underserved areas 1999 and 2002 (see Table D.9). One reason 8.1 percent to 6.8 (4.9) percent; in this case, market, it appears that the underserved for the underestimation of 1999–2002 a significant differential remained after market has certainly maintained itself at a experience was that the underserved areas excluding B&C (subprime) loans. In the high level over the past four years. share of the single-family-owner market heavy refinance years of 1998, 2001, and Renter Mortgages. The second and third continued to increase during this period of 2002, underserved areas accounted for 25–27 sets of numbers in Table D.12 are the low interest rates. Table D.13 reports HUD’s percent of both home purchase and refinance underserved area percentages for single- new estimates of the market share for loans. family rental mortgages and multifamily underserved areas based on the projection The underserved areas share for home mortgages, respectively. Based on HMDA 52 The estimates in purchase loans has been in the 25–26 range data for single-family, non-owner-occupied model discussed earlier. since 1995, except for 2000 and 2002 when (investor) loans, the underserved area share Table D.13 exclude the effects of B&C loans. it increased to slightly over 27 percent. of newly-mortgaged single-family rental units BILLING CODE 4210–27–P

52 Table D.13 presents estimates for the same combinations of projections used to analyze the Low- and Moderate-Income Goal. Table D.10 in Section F.3 defines Cases 1, 2, and 3; Case 1 (the baseline) projects a 42.5 percent share for single- family rentals and a 48 percent share for multifamily properties while the more conservative Case 2 projects 40 percent and 46 percent, respectively.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00255 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24482 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00256 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.081 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24483

The percentage of single-family-owner at or slightly above 30 percent, which is higher if complete data for non-metropolitan mortgages financing properties in similar to its market share during 1998 (31.0 counties were available. According to underserved areas is the most important percent) but somewhat less than its market HMDA, underserved counties accounted for determinant of the overall market share for share during 2001 (32.6 percent) and 2002 41–45 percent (or 42.7 percent) of all this goal. Therefore, Table D.13 reports (32.0 percent). mortgages originated in non-metropolitan market shares for different single-family- areas between 1999 and 2002. By contrast, 3. Adjustments: B&C Loans, the Rural owner percentages ranging from 30 percent underserved census tracts accounted for Underserved Areas Market, and (2000 level) to 20 percent (1993 level) to 18 approximately 24–33 percent (or 27.4 Manufactured Housing Loans percent. If the single-family-owner percent) of all mortgages originated in percentage for underserved areas is at its B&C Loans. The procedure for dropping metropolitan areas between 1999 and 2002.54 1994–2002 HMDA average of 27 percent, the B&C loans from the projections is the same Assuming that non-metropolitan areas market share estimate is 32–33 percent. The as described in Section F.3.b for the Low- account for 13 percent of all single-family- overall market share for underserved areas and Moderate-Income Goal. The underserved owner mortgages and estimating that the peaks at 35 percent when the single-family- area percentage for B&C loans is 44.5 percent, single-family-owner market for accounts for owner percentage is at its 2000 level of 30 which is much higher than the projected 72 percent of newly-mortgaged dwelling percent. Most of the estimated market shares percentage for the overall market (which units, then the non-metropolitan underserved for the owner percentages that are slightly peaks at 35 percent as indicated in Table area differential of approximately 15 percent below recent experience are in the 30 percent D.13). Thus, dropping B&C loans will reduce would raise the overall market estimate by range. the overall market estimates. Consider the 1.4 percentage point—15 percentage points Unlike the Low- and Moderate-Income case of a single-family-owner percentage of times 0.13 (non-metropolitan area mortgage Goal, the market estimates differ only slightly 27 percent, which yields an overall market market share) times 0.72 (single-family owner as one moves from Case 1 to Case 3 and from estimate for underserved areas of 33.4 mortgage market share). Based on this a 13.5 percent mix to 16.5 percent mix. For percent, including B&C loans. When B&C calculation, if the 15 point differential example, reducing the assumed multifamily loans are excluded from the projection reflected actual market conditions, then the mix from 16.5 percent to 13.5 percent model, the underserved areas market share underserved areas market share estimated reduces the overall market projection for falls by 0.7 percentage points to 32.7 percent, using metropolitan area data should be underserved areas by only about 0.6 which is the figure reported in Table D.13. increased by 1.4 percentage points to account percentage points. This is because the Non-metropolitan Areas. Underserved for the effects of underserved counties in underserved area differentials between owner rural areas are non-metropolitan counties non-metropolitan areas.55 A more and rental properties are not as large as the with: conservative adjustment of 1.25 percentage low- and moderate-income differentials (a) County median income at or below 95 points was made in Table D.13 for the reported earlier. percent of the greater of statewide non- projection model.56 Additional sensitivity analyses were metropolitan median income or nationwide Manufactured Housing Loans. Excluding conducted to reflect the volatility of the non-metropolitan income; or manufactured housing loans (as well as small economy and mortgage market. Recession (b) A minority composition equal to 30 loans less than $15,000) reduces the overall and high interest rate scenarios assumed a percent or more and a county median income underserved area market estimates reported significant drop in the underserved area no more that 120 percent of statewide non- in Table D.13 by less than one percentage percentage for single-family-owner metropolitan median income. point. This is estimated as follows. First, mortgages. The single-family-owner HMDA’s limited coverage of mortgage data excluding these loans reduces the unadjusted percentage can go as low as 24 percent— in non-metropolitan counties makes it underserved areas percentage for single- which is 3 percentage points lower than the impossible to estimate the size of the family-owner mortgages in metropolitan 1994–2002 average of 27 percent—and the mortgage market in rural areas. However, all areas by about 1.2 percentage points, based estimated market share for underserved areas indicators suggest that underserved counties on analysis of recent home purchase remains over 30 percent. In a more severe in non-metropolitan areas comprise a larger environments (1995–97 and 1999 and 2000). case, the overall underserved market share share of the non-metropolitan mortgage Multiplying this 1.2 percentage point would be 28 percent if the single-family- market than the underserved census tracts in differential by the property share of single- owner share fell to 21 percent (its 1992 level), metropolitan areas comprise of the family-owner units (72.2 percent) yields 0.8 which is 8–9 percentage points lower than its metropolitan mortgage market. For instance, percentage points, which serves as a proxy 1999–2000 levels. The heavy refinance underserved counties within rural areas for the reduction in the overall underserved scenarios discussed for the low-mod market include 54 percent of non-metropolitan area market share due to dropping were also projected for the underserved areas homeowners; on the other hand, underserved manufactured home loans from the market market. With a 65 percent refinance rate and census tracts in metropolitan areas account analysis. The actual reduction will be an assumed 24 percent underserved area for only 34 percent of metropolitan somewhat less because dropping percentage for owner mortgages, the homeowners. projection model produced overall market During 1999–2001, 36–39 percent of the 54 estimates that ranged from 32.6 percent These data do not include loans originated by GSEs’ total purchases in non-metropolitan lenders that specialize in manufactured housing (multifamily mix of 13.5 percent) to 31.7 areas were in underserved counties while loans, as well as estimated B&C loans. The averages percent (multifamily mix of 9 percent). 25–30 percent of their purchases in in this and the preceding sentence are annual Lowering the multifamily mix in the heavy metropolitan areas were in underserved unweighted averages. refinance model characterized by year 2002 census tracts. These figures suggest the 55 Mortgage Interest Rate Survey (MIRS) data assumptions produced the following range of market share for underserved counties in reported by the Federal Housing Finance Board estimates for the overall underserved areas rural areas is higher than the market share for separate conventional home purchase loans by their market: 32.1 percent (multifamily mix of 11.0 underserved census tracts in metropolitan metropolitan and non-metropolitan location. The average non-metropolitan share between 1999 and percent) to 31.2 percent (multifamily mix of areas. Thus, using a metropolitan estimate to 8 percent) to 30.7 percent (multifamily mix 2002 was about 13 percent. proxy the overall market for this goal, 56 of 6 percent).53 In the refinance scenarios, the For the 1999–2002 data in Table D.9, the non- including rural areas, is conservative. metropolitan adjustment was calculated by underserved areas market share was typically Between 1999 and 2001, the non- multiplying the actual single-family-owner property metropolitan portion of the Underserved share during a particular year by that year’s 53 During 2002, the underserved areas share was Areas Goal has contributed 1.1 to 1.4 (0.7 to underserved area share for non-metropolitan areas 27.2 percent for home purchase loans and 24.4 1.3) percentage points to Freddie Mac’s by the average metropolitan/non-metropolitan percent for refinance loans, yielding a differential (Fannie Mae’s) performance, compared with differential of 15 percent (see text). The average of 2.8 percentage points. Increasing the differential a goals-counting system that only included differential of 15 percent was used because the to 4 percentage points (by reducing the underserved annual differentials exhibited rather wide variation, area share of refinance loans to 23.2 percent) would metropolitan areas. and given issues about HMDA’s coverage of non- reduce the overall underserved areas market The limited HMDA data available for non- metropolitan areas, the average differential was percentages reported in the text by about 0.6 metropolitan counties also suggest that the used. An adjustment of 1.5 percentage points was percentage point. underserved areas market estimate would be used for the earlier years, 1995 to 1998.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00257 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 24484 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

manufactured home loans will increase the based on past origination activity in The 1990-based analysis that produced the share of rental units, which increases the underserved areas and on scenarios that 30–35 percent range serves as the starting overall underserved areas market share, thus cover a variety of economic and mortgage point for an upward adjustment in the market partially offsetting the 0.8 percent reduction. market conditions. That analysis, which range. The net effect is probably a reduction of included historical data going back to the For the years 1999 to 2002, Table D.14 about three-quarters of a percentage point. early 1990s, necessarily used 1990 Census reports the underserved areas share of the The estimates presented in Table D.13 geography to define underserved census mortgage market for single-family-owner, suggest that 30–35 percent would be a tracts. As explained in Appendix B, HUD investor (non-owner), and multifamily reasonable range for the market estimate for will be defining underserved areas based on properties, with comparisons between 1990- underserved areas based on the projection 2000 Census geography beginning in 2005, based and 2000-based measures of model described earlier and assuming 1990 the first year covered by this proposed rule. underserved areas. HMDA data, which is the Census geography. This range incorporates Appendix B also explains that the number of source of the mortgage data, were reported in market affordability conditions that are more census tracts in metropolitan areas covered terms of 1990 census tracts. For the years adverse than have existed recently and it by HUD’s underserved area definition will 1999 to 2002, HUD used various excludes B&C loans from the market increase from 21,587 tracts (based on 1990 apportionment techniques to re-allocate estimates. As discussed next, switching from Census) to 26,959 tracts (based on 2000 1990-based HMDA mortgage data into census 1990 to 2000 Census geography increases this Census and OMB’s respecification of tracts as defined by the 2000 Census. The market range by five percentage points to 35– metropolitan areas). This increase in the 1990-based underserved area market shares 40 percent. number of tracts defined as underserved reported in Table D.14 are the same data means that the market estimate for the reported earlier in Table D.12, while the 4. 2000-Based Underserved Area Market Geographically Targeted Goal will be higher 2000-based underserved area market shares Shares than the 30–35 percent estimate presented result from re-allocating 1999–2002 HMDA The above analysis has concluded that 30– above. Thus, this section provides a new data into 2000 Census geography. In 35 percent would be a reasonable market range of market estimates for underserved addition, the data are defined in terms of the range for the Geographically Targeted Goal areas defined in terms of 2000 Census data. new OMB metropolitan area definitions.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00258 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24485

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00259 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.082 24486 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

First, consider the market shares for single- compared with 52.3 percent based on 1990 family-owner percentage for underserved family-owner properties in the top portion of geography. areas is at its 1999–2002 HMDA average of Table D.14. In 2002, the underserved area The underserved areas shares based on 33 percent, the market share estimate is 39 percentage for home purchase loans increases 2000 Census geography were estimated for percent. The overall market share for from 27.2 percent (1990-based) to 33.3 the last four years, 1999 to 2002; the underserved areas peaks at approximately 41 following estimates were obtained: 39.0 percent (2000-based), an increase of 6.1 percent when the single-family-owner percentage points; the corresponding percent (1999), 40.4 percent (2000), 37.7 percentage is at its 2000 level of 36 percent. percentages for refinance loans were 24.4 percent (2001), and 37.2 percent (2002). percent (1990-based) and 29.8 percent (2000- These 2000-based market estimates are Most of the estimated market shares for the based), or an increase of 5.4 percentage slightly over five percentage points higher owner percentages that are within four points. Considering total owner loans (i.e., than the 1990-based market estimates for percentage points of recent experience (i.e., both home purchase and refinance owner underserved areas reported in Table D.9: 5.1 the 29–33 percent range) are in the 36–39 loans), the average of the ‘‘Differences’’ percent (1999), 5.2 percent (2000), 5.1 percent range. reported in Table D.14 is 5.6 percentage percent (2001), 5.1 percent (2002), and 5.1 BILLING CODE 4210–27–P points for the conforming market. Between percent (2002).57 This analysis suggests that 1999 and 2001, 32.3 percent of mortgage a reasonable range for the overall market only on the HMDA data reported in Table D.14, originations were originated in underserved share for underserved areas based on 2000 they would have been 48.0% for Case 1, 46.0% for areas based on 2000 geography, compared geography might be 35–40 percent, which is Case 2, and 50.0% for Case 3. However, in with 26.7 percent based on 1990 geography— obtained by simply adding five percentage conducting this 2000-based analysis, HUD also yielding an overall differential of 5.6 points to the 30–35 percent range estimated computed the single-family rental shares for the percentage points. earlier based on 1990-based geography. As GSEs in terms of both the number of mortgages Next, consider the underserved area market discussed next, a 35–40 percent range is (consistent with the HMDA data in Table D.14) and shares reported for single-family rental (or indeed an appropriate estimate of the the number of single-family rental units financed (the concept used in the housing goals calculation). non-owner) and multifamily properties in the underserved area market based on 2000 geography. That analysis showed that the unit-based middle and bottom portions of Table D.14. In underserved area percentage was approximately six 2002, the underserved area percentage for Table D.15 reports the results of the projection model assuming 2000 geography. percentage points higher than the number-of- home purchase non-owner loans increases mortgage-based underserved area percentage. To from 42.1 percent (1990-based) to 48.1 Since Table D.15 has the same interpretation reflect this differential, HUD adjusted the percent (2000-based), an increase of 6.0 as Table D.13, there is no need to provide a percentages in Cases 1–3 by an additional four detailed discussion of it.58 If the single- percentage points; the corresponding percentage points. With respect to multifamily percentages for refinance loans were 45.8 properties, the following assumptions were made percent (1990-based) and 51.2 percent (2000- 57 The differentials reported in Table D.14 for the with respect to underserved areas shares: 58.0% for Case 1, 56.0% for Case 2, and 59.0% for Case 3. If based), or an increase of 5.4 percentage three individual property types tend to be greater than 5.5 percentage points, which raises the these percentages were based only on the HMDA points. Considering total single-family rental question of why the overall differential is only 5.1 data reported in Table D.14, they would have been loans (i.e., home purchase and refinance percentage points. As explained later, the upward 55.0% for Case 1, 53.0% for Case 2, and 55.0% for loans), the 1999–02 average of the adjustment to account for underserved areas in non- Case 3. HUD computed the multifamily ‘‘Differences’’ reported in Table D.14 is 5.3 metropolitan areas is about 0.65 percentage point underserved area shares for the GSEs in terms of percentage points for the single-family rental less using the 2000-based Census data than it was mortgage dollars (consistent with the HMDA data market. The multifamily differentials are using the 1990-based Census data. Table D.14) and the number of multifamily rental 58 slightly higher at approximately 7–8 In addition to adjusting the various single- units financed (the concept used in the housing family-owner parameters upward, the following goals calculation). That analysis showed that the percentage points. Between 1999 and 2002, 2000-based assumptions were made with respect to unit-based underserved area percentage was also 59.8 percent of multifamily originations (on the underserved areas shares of single-family rental approximately six percentage points higher than the a dollar basis) were originated in properties: 52.0% for Case 1, 50.0% for Case 2, and mortgage-dollar-based underserved area percentage; underserved areas based on 2000 geography, 54.0% for Case 3. If these percentages were based thus HUD adjusted the percentages upward.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00260 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24487

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00261 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03MY04.083 24488 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Following the 1990-based analysis in the 8 point differential reflected actual HUD estimates that the special affordable Section G.2, additional sensitivity analyses market conditions, then the underserved market is 24–28 percent of the conventional were conducted to reflect the volatility of the areas market share estimated using conforming market. economy and mortgage market. Recession metropolitan area data should be increased HUD is proposing to establish each GSE’s and high interest rate scenarios assumed a by 0.75 percentage point to account for the special affordable multifamily subgoal as 1.0 significant drop in the underserved area effects of underserved counties in non- percent of its average annual dollar volume percentage for single-family-owner metropolitan areas, based on 2000 geography. of total (single-family and multifamily) mortgages. The single-family-owner A more conservative adjustment of 0.65 mortgage purchases over the 2000–2002 percentage can go as low as 29 percent— percentage points was made in Table D.15, period. In dollar terms, the Department’s which is 3 percentage points lower than the which reports the results of the projection proposal is $5.49 billion per year in special 1994–2002 average of 32 percent and 4 model. affordable multifamily purchases for Fannie Section G.3 reported that excluding percentage points lower than the 1999–2002 Mae, and $3.92 billion for Freddie Mac. The manufactured housing loans (as well as small average of 33 percent—and the estimated multifamily special affordable goal, as well as loans less than $15,000) reduced the overall market share for underserved areas remains the special affordable home purchase about 36 percent. In a more severe case, the underserved area market estimates based on 1990 geography by less than one percentage subgoal, are discussed further in Appendix C. overall underserved market share would be Section F described HUD’s methodology 33–34 percent if the single-family-owner point. Excluding manufactured housing loans for estimating the size of the low- and share fell to 26 percent (its 1992 level), which leads to a similar reduction for the market moderate-income market. Essentially the is 7 percentage points lower than its 1999– estimates based on 2000 geography. same methodology is employed here except 2002 average. In the heavy refinance The estimates presented in Table D.15 that the focus is on the very-low-income scenarios (with their lower multifamily suggest that 35–40 percent is a reasonable market (0–60 percent of Area Median mixes), the underserved areas market share range for the market estimate for underserved was typically around 36–37 percent. areas based on the projection model Income) and that portion of the low-income Non-metropolitan Areas. As explained in described earlier. This range incorporates market (60–80 percent of Area Median Section G.3, in order to account for the much market affordability conditions that are more Income) that is located in low-income census larger coverage of underserved areas in non- adverse than have existed recently and it tracts. Data are not available to estimate the metropolitan areas, 1.25 percent was added excludes B&C loans from the market number of renters with incomes between 60 to the market share based on metropolitan estimates. and 80 percent of Area Median Income who live in projects that meet the tax credit area data, in order to arrive at a nationwide 5. Conclusions estimate of the market share for underserved thresholds. Thus, this part of the Special areas. According to HMDA, underserved Based on the above findings as well as Affordable Housing Goal is not included in counties accounted for 42.7 percent of single- numerous sensitivity analyses, HUD the market estimate. concludes that 35–40 percent is a reasonable family-owner mortgages originated in non- estimate of mortgage market originations that 1. Special Affordable Shares by Property metropolitan areas during the 1999-to-2002 would qualify toward achievement of the Type period, based on 1990 geography. With 2000 Geographically Targeted Goal if purchased by geography and the new tract-based definition The basic approach involves estimating for a GSE. The 35–40 percent range is higher of underserved areas in non-metropolitan each property type the share of dwelling than the market range in the 2000 Rule areas, the market share falls by 2.3 percentage units financed by mortgages that are mainly because it is based on 2000 Census points to 39.6 percent. This 2000-based occupied by very-low-income families or by geography which includes more underserved underserved areas percentage of 39.6 percent low-income families living in low-income census tracts than 1990 Census geography. for non-metropolitan areas is about eight areas. HUD combined mortgage information HUD recognizes that shifts in economic and percentage points less than the comparable from HMDA, the American Housing Survey, 59 housing market conditions could affect the and the Property Owners and Managers percentage for metropolitan areas. This size of this market; however, the market eight-point differential is lower than the 15- Survey in order to estimate these special estimate allows for the possibility that affordable shares. point differential used in the earlier 1990- adverse economic conditions can make based Census analysis. Assuming that non- housing less affordable than it has been in a. Special Affordable Owner Percentages metropolitan areas account for 13 percent of the last few years. In addition, the market HMDA data for the percentage of single- all single-family-owner mortgages and estimate incorporates a range of assumptions family-owners that qualify for the Special estimating that the single-family-owner about the size of the multifamily market and Affordable Goal are reported in Table D.16. market accounts for 72 percent of newly- excludes B&C loans. That table also reports data for the two mortgaged dwelling units, then the non- components of the Special Affordable Goal— H. Size of the Conventional Conforming metropolitan underserved area differential of very-low-income borrowers and low-income Market for the Special Affordable Housing 8 percent would raise the overall market borrowers living in low-income census tracts. estimate by 0.75 percentage point—8 Goal Focusing first on home purchase loans, percentage points times 0.13 (non- This section presents estimates of the HMDA data show that the special affordable metropolitan area mortgage market share) conventional conforming mortgage market for share of the market has followed a pattern times 0.72 (single-family owner mortgage the Special Affordable Housing Goal. The similar to that discussed earlier for the low- market share). Based on this calculation, if special affordable market consists of owner and moderate-income loans. The percentage and rental dwelling units which are occupied of special affordable borrowers increased 59 Between 1999 and 2002, 2000-based by, or affordable to: (a) Very-low-income significantly between 1992 and 1994, from underserved census tracts accounted for 31.4 families; or (b) low-income families in low- 10.4 percent of the conforming market to 12.6 percent (unweighted annual average) of all income census tracts; or (c) low-income mortgages in metropolitan areas. This 1999–02 percent in 1993, and then to 14.1 percent in families in multifamily projects that meet 1994. Between 1995 and 1998, the special average percentage for metropolitan areas is lower minimum income thresholds patterned on that the 33.0 percent reported in previous affordable market was in the 14–16 percent the low-income housing tax credit (LIHTC).60 paragraphs. To be comparable with the non- range, averaging 15.1 percent. Over the past metropolitan data, these metropolitan area data do four years (1999–2002), the special affordable not include loans originated by lenders that 60 There are two LIHTC thresholds: at least 20 share of the home purchase loans has specialize in manufactured housing loans and B&C percent of the units are affordable at 50 percent of averaged 16.7 percent. loans; excluding these loans lowers the underserved AMI or at least 40 percent of the units are affordable areas share. at 60 percent of AMI. BILLING CODE 4210–27–P

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00262 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24489

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00263 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03my04.084 24490 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

Considering all (home purchase and mortgaged rental properties are quite affordable at 60–80 percent of AMI and refinance) loans during recent ‘‘home consistent with the AHS data on the located in low-income areas.63 purchase’’ environments, the special affordability of the rental stock. Fifty-six (56) 2. Size of the Special Affordable Market affordable share averaged 18.8 percent during percent of single-family rental properties 1999–2000, over three percentage points with new mortgages between 1993 and 1995 During the 2000 rule making, HUD more than the 15.4 percent average between were affordable to very-low-income families, estimated a market share for the Special 1995 and 1997. Excluding B&C (all subprime) as was 51 percent of newly-mortgaged Affordable Goal of 23–26 percent. This loans from the analysis reduces this multifamily properties. These percentages for estimate was below market experience, as the special affordable market accounted for 26– differential only slightly to 2.7 (2.4) newly-mortgaged properties from the POMS percentage points. As mentioned earlier, 30 percent of all housing units financed are similar to those reported above from the lending patterns could change with sharp between 1999 and 2002, as well as 26–29 AHS for the rental stock. The baseline changes in the economy, but the fact that percent of units financed between 1995 and there have been several years of strong projection from HUD’s market share model 1998 (see Table D.9). This underestimation affordable lending suggests that the special assumes that 50 percent of newly-mortgaged, was mainly due to the assumption in the affordable market has changed in single-family rental units, and 47 percent of projection model that the special affordable fundamental ways from the mortgage market multifamily units, are affordable to very-low- share of refinance loans was lower than the of the early 1990s. In fact, there appears to income families. special affordable share of home purchase have been a slight increase in this market c. Low-Income Renters in Low-Income Areas loans; and the fact that the special affordable recently, at least during 1999 and 2000. HMDA does not provide data on low- share of the single-family-owner market Except for the three years of heavy increased recently (see above discussion). income renters living in low-income census refinancing (1998, 2001, and 2002), the This section produces new estimates of the tracts. As a substitute, HUD used the POMS special affordable share of the refinance special affordable market. and AHS data. As explained in the 2000 GSE market has recently been higher than the The size of the special affordable market special affordable share of the home purchase Rule, the share of single-family and depends in large part on the size of the market—a pattern discussed in Section F for multifamily rental units affordable to low- multifamily market and on the special low-mod and very-low-income loans. During income renters at 60–80 percent of area affordable percentages of both owners and 1999 (2000), for example, the special median income (AMI) and located in low- renters. Table D.10 gives new market affordable share of the refinance market was income tracts was calculated using the estimates for different combinations of these 19.2 (22.7) percent, compared with 17.3 internal Census Bureau AHS and POMS data factors. As before, Case 2 is slightly more (17.1) percent for the home loan market. The files.61 The POMS data showed that 8.3 conservative than the baseline projections higher special affordable percentages for percent of the 1995, single-family rental (Case 1) mentioned above. For instance, Case refinance loans are reduced or even stock, and 9.3 percent of single-family rental 2 assumes that only 6 percent of rental units eliminated if subprime loans are excluded units receiving financing between 1993 and are affordable to low-income renters living in from the analysis. As shown in Table D.16, 1995, were affordable at the 60–80 percent low-income areas. excluding B&C loans from the data level and were located in low-income census Table D.17 assumes a refinance rate of 35 practically eliminates the refinance-home- tracts. The POMS data also showed that 12.4 percent, which means that the table reflects purchase differential for 1999 and reduces percent of the 1995 multifamily stock, and home purchase or low-refinancing the differential for 2000 to 4.1 percentage 13.5 percent of the multifamily units environments. After presenting these results, points (from 5.6 percentage points). Going receiving financing between 1993 and 1995, market estimates reflecting a heavy refinance further and excluding A-minus loans from were affordable at the 60–80 percent level environment will be presented. In the 2000 the year 2000 data would reduce the and located in low-income census tracts.62 GSE Rule, HUD assumed that the special differential to 2.1 percentage points. HUD’s The baseline analysis below assumes that 8 affordable share of refinance loans was 1.4 projection model excludes B&C loans and percent of the single-family rental units and percentage points lower than the special sensitivity analyses will show the effects on 11.0 percent of multifamily units are affordable share of borrowers purchasing a the overall special affordable market of home. However, as discussed earlier, the excluding all single-family subprime loans. 61 special affordable share of refinance loans b. Very-Low-Income Rental Percentages Affordability was calculated as discussed equaled or was greater than the special earlier in Section F, using AHS monthly housing Table D.14 in Appendix D of the 2000 Rule cost, monthly rent, number of bedrooms, and MSA affordable share of home purchase loans reported the percentages of the single-family location fields. Low-income tracts were identified during home purchase environments such as rental and multifamily stock affordable to using the income characteristics of census tracts 1995–97 or 1999–2000; thus, the assumption very-low-income families. According to the from the 1990 Census of Population, and the census of a lower special affordable shares for AHS, 59 percent of single-family units and 53 tract field on the AHS file was used to assign units refinance loans is initially dropped from the in the AHS survey to low-income tracts and other analysis but will be reintroduced during the percent of multifamily units were affordable tracts. POMS data on year of mortgage origination to very-low-income families in 1997. The sensitivity analysis and the discussion of were utilized to restrict the sample to properties heavy refinancing environments. corresponding average values for the AHS’s mortgaged during 1993–1995. six surveys between 1985 and 1997 were 58 62 During the 1995 rule-making process, HUD BILLING CODE 4210–27–P percent and 47 percent, respectively. As examined the rental housing stock located in low- discussed earlier in Section F, an important income zones of 41 metropolitan areas surveyed as 63 Therefore, combining the assumed very-low- issue concerns whether rent data based on part of the AHS between 1989 and 1993. While the income percentage of 50 percent (47 percent) for the existing rental stock from the AHS can be low-income zones did not exactly coincide with single-family rental (multifamily) units with the used to proxy rents of newly mortgaged low-income tracts, they were the only proxy readily assumed low-income-in-low-income-area available to HUD at that time. Slightly over 13 percentage of 8 percent (11 percent) for single- rental units. HUD’s analysis of POMS data percent of single-family rental units were both family rental (multifamily) units yields the special during the 2000 rule-making process affordable at the 60–80 percent of AMI level and affordable percentage of 58 percent (58 percent) for suggested that it could—estimates from located in low-income zones; almost 16 percent of single-family rental (multifamily) units. This is the POMS of the rent affordability of newly- multifamily units fell into this category. baseline Case 1 in Table D.10.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00264 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24491

BILLING CODE 4210–27–C

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00265 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 EP03my04.085 24492 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules

As shown in Table D.17, the market 8 percent) to 23.9 percent (multifamily mix rather severe conditions, the overall market estimates are: 28–29 percent if the owner of 6 percent).64 share for the Special Affordable Goal would percentage is 17 percent (home purchase The various market estimates presented in decline to 25.1 (23.6) percent, assuming a share for 1999 and 2000); 27–28 percent if Table D.17 for a home purchase environment multifamily mix of 16.5 percent. A the owner percentage is 16 percent (home and reported above for a refinance significant increase in interest rates would purchase share for 1998, 2001, and 2002); environment are not all equally likely. Most also make it more difficult for lower income and 26–27 percent if the owner percentage is of them equal or exceed 25 percent. In the families to afford homeownership and 15 percent (home purchase average from home purchase environment, estimates qualify for mortgages, thus reducing the 1995–97). If the special affordable percentage below 25 percent would require the special special affordable share of the market. But as for home purchase loans fell to 12 percent ’’ affordable share for home purchase loans to noted above, the special affordable share of or by four percentage points below its 1995– drop to 12–13 percent which would be 3–4 the home purchase market could fall to 10 2002 average level of 16 percent ’’ then the percentage points lower than the 1995–2002 percent ’’ almost forty percent below its overall market estimate would be about 25 average for the special affordable share of the seven-year average of 16 percent ’’ before the percent. Thus, 25 percent is consistent with home purchase market. Dropping below 25 market share for the Special Affordable Goal a rather significant decline in the special percent would be more likely in a heavy would fall below 24 percent. affordable share of the single-family home refinance environment, as the actual B&C Loans. The procedure for dropping purchase market. A 25 percent market estimated market shares during 1998, 2001, B&C loans from the projections is the same estimate allows for the possibility that and 2002 were approximately 26 percent. as described in Section F.3.b for the Low- adverse economic and housing affordability However, sensitivity analyses of a refinance and Moderate-Income Goal. The special conditions could keep special affordable environment showed that a 24 percent affordable percentage for B&C loans is 28.0 special affordable market share was families out of the housing market. On the percent, which is similar to the projected consistent with market assumptions percentages for the overall market given in other hand, if the special affordable home significantly more adverse than the heavy Table D.17. Thus, dropping B&C loans (as purchase percentage stays at its recent levels refinance years of 1998, 2001, and 2002. well as all subprime loans) does not (15–17 percent), the market estimate is in the Additional Sensitivity Analyses. appreciably reduce the overall market 27–29 percent range. Additional sensitivity analyses were estimates. Consider the case of a single- Heavy Refinancing Environments. The conducted around the results reported in family-owner percentage of 15 percent, special affordable share of the overall market Table D.17, which reflects a home purchase which yields an overall market estimate for declines when refinances dominate the environment. Assuming that the special Special Affordable Goal of 27.0 percent if market. Section F.3b, which presents the affordable share of the home loan market is B&C loans are included in the analysis. low-mod market estimates, explained the 16 percent, reducing the multifamily mix Dropping B&C loans from the projection assumptions for incorporating a refinance from its baseline of 15 percent to 13.5 (12) model reduces the special affordable market environment into the basic projection model percent would reduce the overall special share by 0.1 percentage points to 26.9, as for 2005–08. Briefly, they are: (1) the affordable market share from 27.7 percent to reported in Table D.15. Dropping all refinance share of single-family mortgages 27.1 (26.4) percent. In this case, increasing subprime loans (A-minus as well as B&C) was increased to 65 percent (from 35 the multifamily mix from 15 percent to 16.5 would reduce the special affordable market percent); the market share for subprime loans percent would increase the special affordable projection to 26.8 percent. reduced to 8.5 percent (from 12 percent); and market share from 27.7 percent to 28.2 Manufactured Housing Loans. Excluding the multifamily mix was initially assumed to percent. manufactured housing loans (as well as small be 13.5 percent (instead of 15 percent or 16.5 As shown in Table D.17, the market loans less than $15,000) reduces the overall percent, which characterize a home purchase estimates under the more conservative Case market estimates reported in Table D.17 by environment). The special affordable share 2 projections are one to one-and-a-half about one percentage point or less. This is for refinance loans was assumed to be 13 percentage points below those under the Case estimated as follows. First, excluding these percent, or two percentage points below the 1 projections. This is due mainly to Case 2’s loans reduces the unadjusted special assumed special affordable share of home lower share of single-family investor affordable percentage for single-family-owner purchase loans (which was set at 15 percent, mortgages (8 percent versus 10 percent in mortgages in metropolitan areas by about 1.5 slightly below the 1998, 2001, and 2002 level Case 1) and its lower affordability and low- percentage points, based on analysis of recent of 16 percent). Under these assumptions, the income-area percentages for rental housing home purchase environments (1995–97 and special affordable market share (excluding (e.g., 53 percent for single-family rental units 1999 and 2000). Multiplying this 1.5 B&C loans) was projected to be 25.4 percent. in Case 2 versus 58 percent in Case 1). percentage point differential by the property If the multifamily mix is reduced further to Recent years have been characterized by share of single-family-owner units (72.2 11 (9) percent, the market projection falls to record low interest rates and strong housing percent) yields 1.1 percentage points, which 24.4 (23.6) percent. If the single-family affordability conditions. Therefore, it was serves as a proxy for the reduction in the special affordable percentages are reduced to important for HUD to examine potential overall special affordable market share due to 14 percent (home purchase) and 12 percent changes in the market shares under more dropping manufactured home loans from the (refinance), and the multifamily mix is 11 (9) adverse market affordability environments market analysis. The actual reduction will be percent, the overall low-mod market share than have existed recently, as well as under somewhat less because dropping falls 23.6 (22.8) percent. As noted in the heavy refinance environments. A heavy manufactured home loans will increase the discussion of the low-mod market, refinance refinance environment has already been share of rental units, which increases the environments are characterized by low discussed so this section focuses on recession overall special affordable market share, thus interest rates; therefore, it is unlikely that the and high-interest-rate scenarios. In the partially offsetting the 1.1 percent reduction. special affordable share of the home purchase recession scenario defined earlier in the low- The net effect is probably a reduction of market would fall below 14 percent during mod analysis (see Section F.3a), the special slightly less than one percentage point. heavy refinance environments, given that it affordable share of the home purchase market Tax Credit Definition. Data are not has averaged almost 16 percent over the past was reduced to 12 (10) percent, or 4 (6) available to measure the increase in market seven years. In addition to these projections, percentage points lower than its 1995–2002 share associated with including low-income a refinance environment characterized by the average share of 16 percent. Under these units located in multifamily buildings that year 2002 market was used to examine how meet threshold standards for the low-income the special affordable market changed under 64 During 2002, the special affordable share was housing tax credit. Currently, the effect on heavy refinancing conditions. Lowering the 15.8 percent for home purchase loans and 14.6 GSE performance under the Special multifamily mix in the heavy refinance percent for refinance loans, yielding a differential Affordable Housing Goal is rather small. For model characterized by year 2002 of 1.2 percentage points. Increasing the differential instance, adding the tax credit condition to 2 percentage points (by reducing the special assumptions produced the following range of increased Fannie Mae’s performance as estimates for the overall special affordable affordable share of refinance loans to 13.8 percent) would reduce the overall special affordable market follows: 0.42 percentage point in 1999 (from market: 25.8 percent (multifamily mix of 11.0 percentages reported in the text by about 0.4 17.20 to 17.62 percent); 0.59 percentage point percent) to 24.7 percent (multifamily mix of percentage point. in 2000 (from 18.64 to 19.23 percent); and

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00266 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2 Federal Register / Vol. 69, No. 85 / Monday, May 3, 2004 / Proposed Rules 24493

0.43 percent point in 2001 (from 19.29 to D.16 in parentheses): 17.5 (17.3) percent for 3. Conclusions 19.72 percent). The increases for Freddie Mac 1999; 17.4 (17.1) percent for 2000; 15.6 (15.8) Sensitivity analyses were conducted for the have been lower (ranging from 0.24 to 0.38 percent for 2001; and 15.8 (16.4) percent for market shares of each property type, for the percentage point during the same period). 2002. While the projected percentages are very-low-income shares of each property New 2000-Based Census Geography and lower in 2001 (0.2 percentage point) and type, and for various assumptions in the New OMB Metropolitan Area Definitions. 2002 (0.6 percentage point), they are higher Going forward, HUD will be re-benchmarking in 1999 (0.2 percentage point) and 2000 (0.3 market projection model. These analyses its median incomes for metropolitan areas percentage point). Given these small suggest that 24–28 percent is a reasonable and non-metropolitan counties based on differences there is no need to changes the estimate of the size of the conventional 2000 Census incomes, will be defining low- market estimates discussed above.65 conforming market for the Special Affordable income census tracts (which are included in Housing Goal. This estimate excludes B&C the definition of special affordable) in terms 65 For the other two property types (single-family loans and allows for the possibility that of the 2000 Census geography, and will be rental and multifamily), comparisons between homeownership will not remain as affordable incorporating the effects of the new OMB projected and historical special affordable as it has over the past five years. In addition, metropolitan area definitions. HUD projected percentages were made using the GSEs’ data. For the estimate covers a range of projections the effects of these three changes on the single-family rental mortgages, the unweighted about the size of the multifamily market. special affordable shares of the market for the average of Fannie Mae’s (Freddie Mac’s) special affordable percentage for the years 1999 to 2002 was [FR Doc. 04–9352 Filed 4–30–04; 8:45 am] years 1999–2002. Under the historical data, 50.2 (51.4) percent using the projected data, BILLING CODE 4210–27–P the average special affordable share of the compared with 48.0 (49.4) percent using the conventional conforming market was 16.7 historical data. For multifamily mortgages, the (16.9) percent for home purchase (total) loans unweighted average of Fannie Mae’s (Freddie special affordable goal (considering all three (see Table D.16); the corresponding average Mac’s) special affordable percentage for the years property types), the unweighted average of Fannie with the projected data was 16.6 (16.9) 1999 to 2002 was 50.4 (45.1) percent using the Mae’s (Freddie Mac’s) special affordable percentage percent. For home purchase loans in the projected data, compared with 53.6 (49.4) percent for the years 1999 to 2002 was 20.0 (18.9) percent using the historical data. These comparisons using the projected data, compared with 20.0 (18.9) conventional conforming market, the suggest little difference between the projected and percent using the historical data. There is little projected special affordable percentages for historical special affordable shares for rental each year between 1999 and 2002 were as properties. HUD also projected the overall special difference in the GSEs’ average special affordable follows (with the historical data from Table affordable percentage for each GSE. For the overall performance between the projected and historical data.

VerDate jul<14>2003 02:21 May 01, 2004 Jkt 203001 PO 00000 Frm 00267 Fmt 4701 Sfmt 4702 E:\FR\FM\03MYP2.SGM 03MYP2