Volume 9, Issue 1 The Quarterly Review First Quarter, 2004 Of

Office of Examinations, Supervision, and Consumer Protection Economic Analysis Division

Interest Rate Sensitivity Falls in First Quarter

Median interest rate sensitivity decreased from Interest Rate Sensitivity Measure 148 basis points in the fourth quarter to 132 basis 450 points in the first quarter 400 of 2004. The fall in sensi- tivity for the thrift industry 350 90th Percentile was due to the decrease in 300 interest rates in the first Special points of interest: 250 quarter. • 30-yr mortgage rate falls Both the median pre– 200 Median • shifts downward in the shock Net Portfolio Value Basis Points 150 first quarter (NPV) ratio and the median post-shock NPV ratios rose 100 • First quarter median interest rate 10th Percentile sensitivity decreases in the first quarter. 50 The number of thrifts 0 • Asset duration falls with post-shock NPV ratios Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 • Comparative Regional Analysis below 4 percent rose to six institutions, up from five • Feature Article on Asset Securitiza- in the previous quarter. ( Continued on page 4) tion Asset

An important financial stitutional investors typi- ing as collateral for the se- Charts in this issue: innovation that has swept cally include mutual funds, curities exists, the degree of through the U.S. financial insurance companies, and homogeneity of the asset, services sector during the pension funds. and the ease of valuing it. Yield Curves 5 past 25 years is asset secu- Some of the most popu- Clearly, the more difficult it ritization. The securitization lar types of loans used for is to value a loan or asset, Mortgage Rates 5 of assets began back in the securitization purposes to- the greater the costs of se- 1970s with the structured day are those involving resi- curitization. financing of mortgage pools. dential mortgages, commer- Today, asset securitiza- Portfolio Mix 5 Asset securitization re- cial mortgages, automobile tion is especially attractive fers to the pooling and loans, and credit-card re- to financial institutions, packaging of homogeneous ceivables. Other types of particularly the larger insti- IRR & NPV Measures 6 loan assets originated by loans that are also securi- tutions, because of its financial and non-financial tized include home equity many benefits. institutions for sale as se- loans, equipment leases, In large measure, the TB13a S-Rating Matrix 7 curities. The cash flows as- and health-care receivables, popularity of securitization sociated with the loan as- to name just a few. is due to recent advances in Comparative Trends in the Four 8 sets serve as the collateral For the most part, any financial engineering and OTS Regions for the issued securities. type of asset can be securi- pressure on financial insti- The purchasers of secu- tized, so long as it is profit- tutions to increase their fee, Aggregate and Regional Appendi- 9-13 ritized assets can be retail able to do so. The profitabil- or noninterest, income and ces investors, commercial ity of a particular asset se- to improve their return on , thrift institutions, or curitization is related to assets. Recent advances in institutional investors. In- how much of the asset serv- (Continued on page 2) Page 2 The Quarterly Review Of Interest Rate Risk

Asset Securitization(continued)

(Continued from page 1) sheet. originate the loans that tors. financial engineering have For example, many are pooled for sale as se- By definition, a pass- facilitated the creation of OTS-regulated institu- curities, the investors who through security, such as the typical structure tions, while originating purchase these securities, a pass-through MBS, is associated with a securi- both fixed-rate and ad- and the borrowers. Fi- one where the principal tized asset from almost justable-rate mortgages, nally, we discuss several and interest payments any type of asset’s cash exhibit a strong tendency of the associated made by borrowers to the flow stream. to retain the ARMs and with the securitization of loan servicer, typically the As such, the advances sell the fixed-rate loans assets. loan originator, are for- in financial engineering into pools of mortgage- warded to a trustee and account for the dramatic backed securities, thus then passed through to increase in the types of reducing their asset dura- What is Asset Securiti- the investors that pur- so-called asset-backed tions. zation? chased the securitities. securities, which are se- Overall, asset securiti- With regard to the five curities that are backed zation provides a financial As noted above, asset parties, loan originators by loan assets other than institution with an impor- securitization refers to the are commercial banks, mortgage loans. tant source of fee income process where interests in thrift institutions, or More specifically, se- and a lower cost source of the cash flows associated other financial institu- curitization allows finan- funding. with loans and other re- tions that originate the cial institutions to get With regard to risk ceivables are pooled, loans. loans they originate off management, financial packaged, underwritten, Loan purchasers are their balance sheets, and institutions, such as com- and sold to investors in usually special purpose in the process, they are mercial banks and thrift the form of securities. vehicles such as trusts able to improve their re- institutions, have two (The following discussion affiliated with the origi- turn on assets in several techniques at their dis- draws from Fabozzi and nating institution or they ways. First, they are able posal to manage their in- Modigliani, Mortgage and can be separate trusts. to raise their return on terest rate risk exposure. Mortgage-Backed Securi- Loan packagers are equity through higher These techniques fall ties Markets, Bhattachrya the underwriters of the noninterest income. into two categories: on- and Fabozzi, Asset- securities. Second, assuming balance-sheet methods Backed Securities, Saun- Guarantors, such as there is no recourse in- and asset securitization. ders, Financial Institutions the federal government volving the loans that are On-balance-sheet meth- Management, and Sinkey, (Ginnie Mae), federal gov- securitized, financial in- ods include such tech- Commercial Finan- ernment–sponsored enter- stitutions can generally niques as maturity cial Management). prises (Fannie Mae and lower their risk-based matching and duration As such, asset securi- Freddie Mac) or commer- capital requirements. Re- matching of assets and tization is a form of what cial banks, provide finan- course refers to the ability liabilities and the use of is known as contingent- cial guarantee insurance of the buyer of a negotia- financial derivatives such commitment banking. or credit enhancement to ble instrument to sell it as interest-rate futures, Contingent-commitment the pool of loans before back to the issuer accord- options, swaps, caps, col- banking refers to the use they are converted to se- ing to the terms and con- lars, and floors to of off-balance sheet activi- curities. Credit enhance- ditions set forth in the interest rate risk. ties by financial institu- ment can take the form of sales contract. Asset securitization, in tions, such as loan com- either a federal or agency Finally, by moving the absence of recourse, mitments, lines of credit, guarantee or a letter of longer duration assets off allows financial institu- and asset securitization, credit. Other forms in- their balance sheets and tions to manage interest in asset/liability manage- clude credit insurance, securitizing them, finan- rate risk by removing long ment. overcollateralization, ex- cial institutions are able duration assets such as For the sake of illus- cess spread, and cash col- to decrease their interest mortgages from their bal- tration, we focus our dis- lateral accounts. rate risk exposure by re- ance sheets. cussion on what is known Finally, investors are ducing the duration gap In what follows, we ex- as a pass-through secu- the individuals or finan- between assets and li- amine what asset securi- rity. There are five parties cial institutions that pur- abilities. This assumes tization is in more detail involved in a typical chase the securities. the financial institutions and discuss the mechan- “pass-through” asset se- Each of the five parties use the proceeds to pur- ics of a typical securitiza- curitization: the loan performs different func- chase assets of shorter tion. We also discuss the originator, the loan pur- tions in the securitization duration or hold it as benefits accruing to the chaser, the loan packager, process. Loan originators cash on the balance financial institutions that a guarantor, and inves- (Continued on page 3) Volume 9, Issue 1 Page 3

Asset Securitization(continued)

(Continued from page 2) Whole loan sales also that they hold in portfolio. tions to immediately rec- require the asset origina- Risk weights associated ognize as earnings the that decide to securitize a tor to find investor groups with the that is present value of excess pool of loans (such as a with investment profiles assumed to exist for dif- servicing. thrift would do with, say, that are consistent with ferent asset categories are In contrast, excess 30-year fixed-rate single- the characteristics of the used in calculating the servicing on whole loans family mortgages) origi- loans that are being sold. required level of risk- can only be recognized as nate the loans and also Second, asset securiti- based capital. earnings over the life of typically service these zation permits risk shar- As a result, financial the loans. loans in order to generate ing and the allocation of institutions can lower Finally, asset securiti- fee income. the asset originator’s bal- their required capital by zation allows financial in- In order to get the ance sheet risk to a vari- purchasing securitized stitutions to lower their loans off the balance ety of investor groups. assets or by securitizing interest rate risk expo- sheet, the loan originator The credit enhancement and then holding the se- sure by reducing the du- will transfer these loans that is usually associated curitized loans (that ration gap between assets to a trust. The trust holds with securitized assets would otherwise be held and liabilities on the bal- these loans as collateral results in wider accep- in portfolio as whole ance sheet. This reduc- for the underwriter, which tance by investors in the loans) and take a much tion in risk exposure is issues the securities and capital markets, leading smaller net risk-based accomplished by securi- then distributes them to to tighter spreads or capital charge. tizing longer maturity as- investors and collects the higher prices on the secu- For example, for sets and removing these initial cash proceeds. ritized assets than the banks and thrift institu- more interest rate sensi- The cash proceeds col- prices on packages of tions, single-family, resi- tive assets from the bal- lected by the underwriter whole loans. dential mortgages have a ance sheet. from the sale of the secu- To the extent that the risk weight of 50 percent, rities are then passed net proceeds from the whereas Fannie Mae and back to the loan origina- sale of securitized assets Freddie Mac pass- Securitization Benefits tor via the trust. are greater than the net through MBSs have a risk to Investors proceeds associated with weight of only 20 percent. loan sales, asset origina- Thus, it is clear that Investors who pur- Securitization Benefits tors are able to achieve financial institutions can chase securitized assets to Issuers liquidity at lower cost. Se- use asset securitization to enjoy the following bene- curitization costs typically take advantage of this fits. First, these securi- There are many bene- include legal fees, invest- risk-weight differential, ties are much more liquid fits of asset securitization ment banking fees, and and by doing so, lower than the assets that serve to asset originators, such distribution and under- their required risk-based as collateral. as commercial banks and writing costs. These must capital. When institutions Second, there is re- thrift institutions. These be subtracted from gross do this, they are said to duced credit risk associ- benefits include broader proceeds. be engaging in what is ated with investing in se- funding sources, the po- Third, financial insti- known as “regulatory curitized assets. This re- tential for lower funding tutions, such as commer- capital .” duction in credit risk is costs, generation of fee cial banks, thrift institu- To some extent, how- due to diversification in income and immediate tions, and insurance com- ever, the adoption of FAS the pool of assets under- recognition of excess ser- panies, can use asset se- 115 has served to make lying the securities and to vicing, and management curitization to manage holding loans on the bal- the private credit en- of interest rate risk. their risk-based capital ance sheet more attrac- hancement or government First, with asset secu- requirements. tive in some circum- insurance guarantee at- ritization, asset origina- In order to comply stances, because loans do tached to the packaged tors are able to access a with the 1988 Basel Ac- not need to be marked to assets. wider array of potential cord, a 2001 interagency market, unlike securitized investors than with loan regulation on recourse assets available for sale. sales. The sale of whole and residuals, and inter- Fourth, financial insti- Securitization Benefits loans as a liquidity source agency guidance on asset tutions can increase ser- to Borrowers is more difficult due to securitization issued in vicing and originating fees greater concern by both 2003, financial institu- by securitizing and selling In general, borrowers originators and investors tions must hold capital loans while retaining ser- will be able to obtain alike with the credit qual- reserves that reflect the vicing. Also, securitization funds from asset origina- ity of the loan package. credit risk of the assets allows financial institu- (Continued on page 4) Page 4 The Quarterly Review Of Interest Rate Risk

Asset Securitization(continued)

(Continued from page 3) association with asset- quality loans or assets. As tors, who then securitize backed a result, institutions will these assets, at a lower and not mortgage-related be left with a concentra- interest rate. securities such as pass- tion of lower quality as- This occurs because through mortgage-backed sets on their balance securitized assets trade at securities or collateralized sheets. a lower spread relative to mortgage obligations. Finally, over-reliance Treasuries. As a result, Concerns with the on securitization as a borrowers can obtain a window dressing of finan- funding source can also lower lending rate. cial statements relate to pose problems for an in- the use of securitization stitution. Without a diver- by financially distressed sified funding base, an Securitization Risks companies to create off- institution could poten- balance sheet financing. tially encounter liquidity Several risks or con- As a result, these compa- problems if it is unable to cerns surround the use of nies are able to show im- securitize its assets asset securitization by provements in their capi- quickly. both financial and nonfi- talization, leverage, and In the next issue of nancial entities. These profitability ratios. These this publication, we will include concerns about improvements are, of examine several of the dif- the window dressing of course, only cosmetic in ferent forms that securiti- financial statements, the nature. zation takes. We will dis- potential for a concentra- Because reduced cuss the characteristics of tion of lower quality as- credit enhancement costs pass-through mortgage- sets kept on the balance and lower due diligence backed securities, collat- sheet, and over-reliance expenses are associated eralized mortgage obliga- on securitization as a with securitizing lower tions, stripped mortgage- funding source. risk assets, concerns backed securities, such as For the most part, arise over the possibility interest-only and princi- these represent concerns that institutions will want pal-only securities, and that are typically raised in to securitize their higher asset-backed securities.¦

Interest Rate Sensitivity Falls in First Quarter (continued)

(Continued from page 1) the end of the first quar- from the fourth quarter. assets in the first quarter, Treasury rates fell for ter from 5.85 percent at Thrift industry earnings down from 0.96 percent all maturities in the first the end of the previous fell three percent to $3.34 in the prior quarter. quarter, except for the quarter. billion in the first quarter, Other non-interest three-month maturity, Thrift profitability was from $3.45 billion in the income rose to 0.94 per- which rose slightly. The lower in the first quarter. prior quarter. cent of average assets decrease for short-term The average return on as- In the first quarter, from 0.46 percent be- and medium-term maturi- sets (ROA) for the indus- total fee income, which tween the fourth and first ties between two years try fell to 1.19 percent includes quarters. Other non- and five years was greater from 1.26 percent in the servicing fee income and interest income can be than for longer-term ma- prior quarter. other fee income, fell to extremely volatile because turities. This decrease in ROA 0.64 percent of average it includes gains and In comparing the for the thrift industry was assets, down from 1.15 losses on assets sold and yield curve to that in the attributed to higher loan percent in the fourth also reflects balance sheet fourth quarter, it was less loss provisions, lower fee quarter. This substantial restructuring activities. In steeply sloped up to the income, and increased drop in total fee income the first quarter, several ten-year maturity point. impairment charges for was due to lower mort- thrifts reported large in- The Freddie Mac con- mortgage servicing rights. gage loan servicing fee creases in income from tract interest rate on com- The first quarter saw income in the first quar- asset sales. mitments for fixed-rate an average net interest ter. The first quarter saw 30-year mortgages de- margin of 288 basis Other fee income fell the ARM share of total creased to 5.52 percent at points, a level unchanged to 0.90 percent of average (Continued on page 5) Volume 9, Issue 1 Page 5

Interest Rates and ARM Market Share

Interest Rates CMT Yield Curves

9.0 6.00

8.0 5.00 7.0 30 Year Mortgage 6.0 4.00

5.0

10 Year CMT Percent 3.00 Percent 4.0

3.0 2.00

2.0 1 Year CMT 1.00 1.0 Maturity Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 September 30, 2003 December 31, 2003 March 31, 2004

ARM Market Share of Originations ARM Share of Thrift Mortgage Portfolios

60% 62%

50% 60%

40% Thrifts 58%

30% 56% Percent Percent 20% 54% ARM Portfolio Percentage All Lenders 10% 52%

0% 50% Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04

Interest Rate Sensitivity Falls in First Quarter (continued) (Continued from page 4) ter. ume of mortgage refinanc- percent at the end of the thrift mortgage originations First-quarter 1-4 family ing activity increased in the first quarter, unchanged rise to 44 percent, up from mortgage originations by first quarter. This occurred from the fourth quarter. 37 percent in the prior thrifts fell to $130.2 billion, as a result of the decline in Refinancing accounted for quarter. down from $143.9 billion interest rates. 37.4 percent of thrift origi- Despite the rise in the in the fourth quarter. Thrifts’ share of all 1-4 nations of single-family share of thrift ARM origina- Total mortgage origi- family originations was mortgages in the first quar- tions, the ARM share of to- nations in the first quarter 22.1 percent in the first ter, up from 25.9 percent tal 1-4 family mortgages were $144.0 billion, down quarter, down from 25.9 in the fourth quarter. held by thrifts in their from $163.9 billion in the percent in the fourth quar- This increase is consis- portfolios fell slightly to fourth quarter. Despite the ter. tent with the refinancing 60.4 percent from 60.6 drop in total mortgage loan The rate of U.S. home activity of all lenders, percent in the prior quar- origination volume, the vol- ownership stood at 68.6 (Continued on page 6) Page 6 The Quarterly Review Of Interest Rate Risk

Duration and NPV Sensitivity Measures

Average Effective Duration Median Pre- and Post-Shock NPV Ratios

2.2 14.0

Assets Pre-Shock 2.0 13.0

1.8 12.0

1.6 11.0 Percent Duration Post-Shock

1.4 10.0 Liabilities

1.2 9.0

1.0 8.0 Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04

Estimated Change in NPV: Estimated Change in NPV: +200bp Rate Change -100bp Rate Change

600 600 Lose NPV Gain NPV Lose NPV Gain NPV 464

400 400 358 Number Number 269 247

200 200 134 107

47 40 19 21 20 0 0 0 0 0 <-40% -40 to -30% -30 to -20% -20 to -10% -10 to 0% 0 to 10% >10% <-10% -10 to 0% 0 to 10% 10 to 20% 20 to 30 30 to 40% >40%

Interest Rate Sensitivity Falls in First Quarter (continued) (Continued from page 5) rate of prepayments of a decrease in the positive the median pre-shock NPV where the rate rose to 53 mortgages held in portfolio. duration gap for the thrift ratio, the median post- percent from 49 percent This lowered the dura- industry as a whole. This shock NPV ratio also rose, between the fourth and tion of mortgages and, reverses the trend of the moving from 11.3 percent first quarters. therefore, total assets du- past two quarters which at the end of the fourth The industry’s effective ration. saw the positive duration quarter to 11.5 percent at duration of assets fell from The industry’s effective gap widen for the industry. the end of the first quarter. 1.92 to 1.74 between the duration of liabilities rose The median pre-shock The number of thrifts fourth and first quarters. slightly from 1.64 to 1.66 NPV ratio for the industry with a post-shock NPV ra- With the decrease in inter- in the first quarter. rose during the first quar- tio below 4 percent rose to est rates in the first quar- The changes in asset ter from 12.9 percent to six institutions from five in ter, the NPV model pre- and liability durations in 13.0 percent. the previous quarter. dicted an increase in the the first quarter produced Along with this rise in (Continued on page 7) Volume 9, Issue 1 Page 7

Interest Rate Risk Measures

Thrifts with Post-Shock NPV Ratios Interest Rate Risk Measures Under 4 Percent Industry Aggregates Last Two Quarters 50 45 NPV as % of PV of % Change % Change 40 Assets in NPV in NPV 35 Dec-03 Mar-04 Dec-03 Mar-04 30 +300 8.09% 7.87% -29% -31% 25 20 +200 9.24% 9.11% -17% -18% Number 15 +100 10.20% 10.12% -7% -7% 10 10 7 8 Base 11.07% 10.80% 0% 0% 6 5 6 3 3 5 1 -100 10.87% 10.92% 3% 2% 0 -200 N/A N/A N/A N/A Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 -300 N/A N/A N/A N/A

Post-Shock NPV Ratio and Post-Shock NPV Ratio and Sensitivity Measure Matrix Sensitivity Measure Matrix December 2003 March 2004

Under 101- 201- Over Under 101- 201- Over Total Total 100bp 200bp 400bp 400bp 100bp 200bp 400bp 400bp Over Over 227 148 158 26 559 258 157 141 17 573 10% 10% 6% to 6% to 79 73 101 15 268 93 79 73 8 253 10% 10% 4% to 4% to 3 7 18 3 31 3 5 21 2 31 6% 6% Below Below 0 1 4 0 5 0 1 3 2 6 4% 4% Total 309 229 281 44 863 Total 354 242 238 29 863

Minimal Moderate Significant High Minimal Moderate Significant High

Interest Rate Sensitivity Falls in First Quarter (continued)

(Continued from page 6) tio below 6 percent rose to 400 basis points in sensi- The percentage of 37 institutions in the first tivity fell to 3.4 percent thrifts with a post-shock quarter, up from 36 in the from 5.1 percent in the NPV ratio over 6 percent prior quarter. prior quarter. remained largely un- The percentage of These results are con- changed between the thrifts with a sensitivity of sistent with the fall in the fourth and first quarters. 200 basis points or less industry’s effective dura- In both the fourth and first increased in the first quar- tion gap and with the fall quarter, these thrifts com- ter, rising to 69.1 percent in its median sensitivity in prised 95.8 percent of the from 62.3 percent in the the first quarter.¦ industry. prior quarter. The number of thrifts In addition, the per- with a post-shock NPV ra- centage of thrifts with over Page 8 The Quarterly Review Of Interest Rate Risk

Comparative Trends in the Four OTS Regions

Median Sensitivity by OTS Region Median Assets Duration by OTS Region

250 2.5

200 2

150 1.5

100 1 Basis Points

50 Effective Duration 0.5

0 0 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04

All Thrifts Northeast Southeast Midwest West All Thrifts Northeast Southeast Midwest West

Median Pre-Shock NPV Ratio by OTS Region Median Post-Shock NPV Ratio by OTS Region

15 15

14 14

13 13

Percent 12 Percent 12

11 11

10 10 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04

All Thrifts Northeast Southeast Midwest West All Thrifts Northeast Southeast Midwest West

Regional Comparisons The Northeast Region dian sensitivity fall by 19.5 All four OTS regions wit- Finally, median post- had the highest median percent, the largest relative nessed a fall in their me- shock NPV ratios also rose sensitivity, at 167 basis decrease of the four re- dian asset durations. in each of the four OTS points at the end of the gions. The Northeast, All OTS regions saw regions in the first quarter. first quarter, while the Southeast, and West Re- their median pre-shock ¦ Midwest Regions had the gions saw their median NPV ratios rise in the first lowest median sensitivity, sensitivities fall by 18.1 quarter. The Northeast Re- at 91 basis points. percent, 11.9 percent, and gion had the highest pre- All OTS regions experi- 8.8 percent, respectively. shock NPV ratio at 13.5 enced a decrease in their The Northeast Region percent, while the West interest rate sensitivity in had the highest median Region had the lowest pre- the first quarter. The Mid- asset duration, at 1.98 at shock NPV ratio at 11.9 west Region saw its me- the end of the first quarter. percent. Volume 9, Issue 1 Page 9

Appendix A — All Thrifts

Sensitivity Measure Distribution All Thrifts Percent of Thrifts 30 Descriptive Statistics Median = 132 Mean = 158 Standard Deviation = 115 Skewness = 1.12 Kurtosis = 1.43 15 Maximum = 750 Minimum = 0 Count = 863

0 0 100 200 300 400 500 600 700 800 Basis Points

Pre-Shock NPV Ratio Distribution Post-Shock NPV Distribution All Thrifts All Thrifts Percent of Thrifts Percent of Thrifts 60 60 Descriptive Statistics Descriptive Statistics Median = 11.54 Median = 12.95 Mean = 13.26 Mean = 14.84 Standard Deviation = 8.58 Standard Deviation = 8.53 Skewness = 5.03 40 Skewness = 4.94 40 Kurtosis = 35.94 Kurtosis = 34.71 Maximum = 97.17 Maximum = 97.28 Minimum = 1.32 Minimum = 2.86 Count = 863 Count = 863

20 20

0 0 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 NPV Ratio (Percent) NPV Ratio (Percent)

Asset Duration Distribution Liabilities Duration Distribution All Thrifts All Thrifts Percent of Thrifts Percent of Thrifts 60 60 Descriptive Statistics Descriptive Statistics Median = 1.74 Median = 1.68 Mean = 1.76 Mean = 1.68 Standard Deviation = 0.7 Standard Deviation = 0.43 40 Skewness = -0.58 40 Skewness = 0.03 Kurtosis = 5.51 Kurtosis = 2.57 Maximum = 4.39 Maximum = 3.94 Minimum = -3.63 Minimum = 0.02 Count = 863 Count = 863

20 20

0 0 -3 -2 -1 0 1 2 3 4 5 6 More -3 -2 -1 0 1 2 3 4 5 6 More Duration Duration Page 10 The Quarterly Review Of Interest Rate Risk

Appendix B — Northeast Region

Sensitivity Measure Distribution Northeast Percent of Thrifts 30 Descriptive Statistics Median = 167 Mean = 180 Standard Deviation = 112 Skewness = 0.95 Kurtosis = 1.92 15 Maximum = 750 Minimum = 0 Count = 267

0 0 100 200 300 400 500 600 700 800 Basis Points

Pre-Shock NPV Ratio Distribution Post-Shock NPV Distribution Northeast Northeast Percent of Thrifts Percent of Thrifts 60 60 Descriptive Statistics Descriptive Statistics Median = 13.48 Median = 12.07 Mean = 15.34 Mean = 13.54 Standard Deviation = 7.67 Standard Deviation = 7.78 Skewness = 4.5 40 Skewness = 4.48 40 Kurtosis = 34.81 Kurtosis = 35.79 Maximum = 89.53 Maximum = 89.39 Minimum = 6.48 Minimum = 1.32 Count = 267 Count = 267

20 20

0 0 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 NPV Ratio (Percent) NPV Ratio (Percent)

Asset Duration Distribution Liabilities Duration Distribution Northeast Northeast Percent of Thrifts Percent of Thrifts 60 60 Descriptive Statistics Descriptive Statistics Median = 1.78 Median = 1.98 Mean = 1.78 Mean = 1.94 Standard Deviation = 0.39 Standard Deviation = 0.7 Skewness = -0.95 40 40 Skewness = -1.23 Kurtosis = 4.44 Kurtosis = 6.62 Maximum = 3.04 Maximum = 3.96 Minimum = 0.02 Minimum = -2.55 Count = 267 Count = 267 20 20

0 0 -3 -2 -1 0 1 2 3 4 5 6 More -3 -2 -1 0 1 2 3 4 5 6 More Duration Duration Volume 9, Issue 1 Page 11

Appendix C — Southeast Region

Sensitivity Measure Distribution Southeast Percent of Thrifts 30

Descriptive Statistics

Median = 133 Mean = 165 Standard Deviation = 119 15 Skewness = 0.92 Kurtosis = 0.38 Maximum = 616 Minimum = 0 Count = 300

0 0 100 200 300 400 500 600 700 800 Basis Points

Pre-Shock NPV Ratio Distribution Post-Shock NPV Distribution Southeast Southeast Percent of Thrifts Percent of Thrifts 60 60 Descriptive Statistics Descriptive Statistics Median = 11.53 Median = 13.04 Mean = 14.43 Mean = 12.77 Standard Deviation = 7.29 Standard Deviation = 7.23 Skewness = 4.86 Skewness = 4.93 40 40 Kurtosis = 40.7 Kurtosis = 39.58 Maximum = 86.17 Maximum = 85.46 Minimum = 2.86 Minimum = 1.33 Count = 300 Count = 300

20 20

0 0 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 NPV Ratio (Percent) NPV Ratio (Percent)

Asset Duration Distribution Liabilities Duration Distribution Southeast Southeast Percent of Thrifts Percent of Thrifts 60 60 Descriptive Statistics Descriptive Statistics Median = 1.75 Median = 1.64 Mean = 1.78 Mean = 1.63 Standard Deviation = 0.63 Standard Deviation = 0.38 40 Skewness = 0.18 40 Skewness = 0.1 Kurtosis = 0.11 Kurtosis = 0.65 Maximum = 3.76 Maximum = 3.16 Minimum = 0.26 Minimum = 0.44 Count = 300 Count = 300

20 20

0 0 -3 -2 -1 0 1 2 3 4 5 6 More -3 -2 -1 0 1 2 3 4 5 6 More Duration Duration Page 12 The Quarterly Review Of Interest Rate Risk

Appendix D — Midwest Region

Sensitivity Measure Distribution Midwest Percent of Thrifts 30 Descriptive Statistics Median = 91 Mean = 125 Standard Deviation = 100 Skewness = 1.61 15 Kurtosis = 2.84 Maximum = 568 Minimum = 4 Count = 203

0 0 100 200 300 400 500 600 700 800 Basis Points

Pre-Shock NPV Ratio Distribution Post-Shock NPV Distribution Midwest Midwest Percent of Thrifts Percent of Thrifts 60 60 Descriptive Statistics Descriptive Statistics Median = 11.54 Median = 12.8 Mean = 13.46 Mean = 14.72 Standard Deviation = 8.1 Standard Deviation = 8.18 Skewness = 4.25 40 Skewness = 4.1 40 Kurtosis = 24.26 Kurtosis = 22.66 Maximum = 67.41 Maximum = 68.11 Minimum = 4.5 Minimum = 5.75 Count = 203 Count = 203

20 20

0 0 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 NPV Ratio (Percent) NPV Ratio (Percent)

Asset Duration Distribution Liabilities Duration Distribution Midwest Midwest Percent of Thrifts Percent of Thrifts 60 60 Descriptive Statistics Descriptive Statistics Median = 1.65 Median = 1.49 Mean = 1.67 Mean = 1.55 Standard Deviation = 0.48 Standard Deviation = 0.72 Skewness = 1.14 Skewness = -1.44 40 40 Kurtosis = 4.04 Kurtosis = 12.94 Maximum = 3.94 Maximum = 3.85 Minimum = 0.4 Minimum = -3.63 Count = 203 Count = 203

20 20

0 0 -3 -2 -1 0 1 2 3 4 5 6 More -3 -2 -1 0 1 2 3 4 5 6 More Duration Duration Volume 9, Issue 1 Page 13

Appendix E — West Region

Sensitivity Measure Distribution West Percent of Thrifts 30 Descriptive Statistics

Median = 103 Mean = 138 Standard Deviation = 124 Skewness = 1.7 15 Kurtosis = 3.65 Maximum = 659 Minimum = 0 Count = 93

0 0 100 200 300 400 500 600 700 800 Basis Points

Pre-Shock NPV Ratio Distribution Post-Shock NPV Distribution West West Percent of Thrifts Percent of Thrifts 60 60 Descriptive Statistics Descriptive Statistics Median = 10.5 Median = 11.9 Mean = 13.61 Mean = 15 Standard Deviation = 14.05 Standard Deviation = 13.91 Skewness = 4.57 Skewness = 4.54 40 Kurtosis = 21.72 40 Kurtosis = 21.53 Maximum = 97.17 Maximum = 97.28 Minimum = 4.46 Minimum = 6.6 Count = 93 Count = 93

20 20

0 0 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 NPV Ratio (Percent) NPV Ratio (Percent)

Asset Duration Distribution Liabilities Duration Distribution West Percent of Thrifts West 60 Percent of Thrifts 60 Descriptive Statistics Descriptive Statistics Median = 1.55 Median = 1.57 Mean = 1.64 Mean = 1.53 Standard Deviation = 0.78 Standard Deviation = 0.5 Count = 267 40 Skewness = 0.93 Kurtosis = 1.76 40 Kurtosis = 0.28 Maximum = 4.39 Maximum = 2.48 Minimum = 0.14 Minimum = 0.03 Count = 93 Count = 93

20 20

0 0 -3 -2 -1 0 1 2 3 4 5 6 More -3 -2 -1 0 1 2 3 4 5 6 More Duration Duration Page 14 The Quarterly Review Of Interest Rate Risk

Glossary

Duration: A first-order approximation of the price Post-Shock NPV Ratio: Equity-to-assets ratio, fol- sensitivity of a financial instrument to changes in yield. lowing an adverse 200 basis point interest rate shock The higher the duration, the greater the instrument’s (assuming a normal interest rate environment), ex- price sensitivity. For example, an asset with a duration of pressed in present value terms (i.e., post-shock NPV di- 1.6 would be predicted to appreciate in value by about vided by post-shock present value of assets). Also re- 1.6 percent for a 1 percent decline in yield. ferred to as the exposure ratio.

Effective Duration: The average rate of price change Pre-Shock NPV Ratio: Equity-to-assets expressed in in a financial instrument over a given discrete range from present value terms (i.e., base case NPV divided by base the current market interest rate (usually, +/-100 basis case present value of assets). points). Sensitivity Measure: The difference between Pre- Estimated Change in NPV: The percentage change shock and Post– shock NPV Ratios (expressed in basis in base case NPV caused by an interest rate shock. points).

Kurtosis: A statistical measure of the tendency of Skewness: A statistical measure of the degree to data to be distributed toward the tails, or ends, of the which a distribution is more spread out on one side than distribution. A normal distribution has a kurtosis statis- the other. A distribution that is symmetric will have a tic of three. skewness statistic of zero.

NPV Model: Measures how six hypothetical changes in interest rates (three successive 100 basis point in- creases and three successive 100 basis point decreases, assuming a normal interest rate environment) affect the estimated market value of a thrift’s net worth.

Economic Analysis Division

Office of Thrift Supervision Prepared by: 1700 G Street, NW Washington, DC 20552 Jonathan D. Jones Economic Analysis Division Phone: 202-906-5729 David Malmquist, Director Email: [email protected] Economic Analysis Division Phone: 202-906-5639 Robert Sutter, IT Specialist, assembled the data Email: [email protected] reported in the Appendices.

We’re on the Web! www.ots.treas.gov/statisticalreleases