Mortgage Financing: FHA's $7 Billion Reestimate Reflects Higher  
Claims and Changing Loan and Performance Estimates (02-SEP-05,	 
GAO-05-875).							 
                                                                 
The U.S. Department of Housing and Urban Development (HUD),	 
through its Federal Housing Administration (FHA), provides	 
insurance for private lenders against losses on home mortgages.  
FHA's largest insurance program is the Mutual Mortgage Insurance 
Fund (Fund), which currently is self-financed and operates at a  
profit. FHA submitted a "reestimate" of $7 billion for the credit
subsidy and interest for the Fund as of the end of fiscal year	 
2003, reflecting a reduction in estimated profits. Given this	 
substantial reestimate, Congress asked GAO, among other things,  
to determine what factors contributed to the $7 billion 	 
reestimate and the underlying loan performance variables	 
influencing these factors and to assess how the loan performance 
variables underlying the reestimate could impact future estimates
of new loans.							 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-875 					        
    ACCNO:   A35457						        
  TITLE:     Mortgage Financing: FHA's $7 Billion Reestimate Reflects 
Higher Claims and Changing Loan and Performance Estimates	 
     DATE:   09/02/2005 
  SUBJECT:   Federal fund accounts				 
	     Federal funds					 
	     Financial analysis 				 
	     Funds management					 
	     Future budget projections				 
	     Housing programs					 
	     Losses						 
	     Mortgage loans					 
	     Mortgage programs					 
	     Mortgage protection insurance			 
	     Mutual Mortgage Insurance Fund			 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-05-875

United States Government Accountability Office

GAO	Report to the Chairman, Subcommittee on Housing and Community
Opportunity,

           Committee on Financial Services, House of Representatives

September 2005

                                    MORTGAGE
                                   FINANCING

FHA's $7 Billion Reestimate Reflects Higher Claims and Changing Loan Performance
                                   Estimates

                                       a

GAO-05-875

[IMG]

1993 1995 Cohort

Estimate Actual

September 2005

MORTGAGE FINANCING

FHA's $7 Billion Reestimate Reflects Higher Claims and Changing Loan Performance
Estimates

                                 What GAO Found

The $7 billion reestimate was due primarily to an increase in estimated
and actual claims over what FHA previously estimated. For example, actual
claim activity in fiscal year 2003 exceeded estimated claim activity for
2003-by twice as much in some cases-for the majority of loan cohorts.
Prepayments also played a role in the reestimate as they were higher than
previous estimates. In fact, actual prepayment activity during 2003
exceeded estimated prepayment activity for all cohorts. Because of the
additional claims it paid, upfront premiums it refunded, and the annual
premiums it lost, FHA's net cash outflows for the year increased,
contributing to the $7 billion adjustment of the Fund's credit subsidy.

Several recent events may help explain this increase, including changes to
underwriting guidelines, competition from the private sector, and an
increase in the use of down payment assistance. FHA has taken some steps
to tighten underwriting guidelines and better estimate loan performance,
though it is not clear that these steps are sufficient to reverse recent
increases in actual and estimated claims and prepayments or help FHA to
more reliably predict future claim and prepayment activity. Increases in
claim and prepayment activity are likely to continue to add risk to FHA's
portfolio.

Actual Versus Estimated Claim and Prepayment Rates for FY 2003 Claim rate
Prepayment rate

Percentage Percentage 5 50

4 40

3 30

2 20

1 10

00

1997 1999 2001 2003	1993 1995 1997 1999 2001 2003 Cohort

                       Source: GAO analysis of FHA data.

                 United States Government Accountability Office

Contents

  Letter

Results in Brief
Background
The $7 Billion Reestimate Is Significant for Its Size and Direction
The $7 Billion Reestimate Primarily Reflects Higher-Than-Estimated

Claims

The Loan Performance Variables Underlying the $7 Billion Reestimate Will
Likely Affect Future Credit Subsidy Estimates, but Are Being Addressed

The Loan Performance Variables Underlying the Reestimate Could

Affect Estimates of the Fund's Long-Term Viability Conclusions
Recommendations for Executive Action Agency Comments and Our Evaluation

1 3 4 9

13

19

26 31 31 31

Appendixes

Appendix I: Scope and Methodology 34

Appendix II: Data for Figures Used in This Report 36

Appendix III:	Comments from the Department of Housing and Urban
Development 42

Appendix IV: GAO Contact and Staff Acknowledgments 44

Tables	Table 1: Table 2:

Table 3:

Table 4: Table 5: Table 6:

Annual Credit Subsidy Reestimates For the MMI Fund, Fiscal Years 2000-2004
(Figure 3) 36 Amount of the 2003 Reestimate Attributed to the 2001-2003
Cohorts, as a Percentage of the Original Loan Amount, For Single-Family
Loan Guarantee Programs (Figure 4) 36 Original Estimated Credit Subsidy
Rates and Most Recent Reestimated Rates for the FHA and VA Loan Guarantee
Programs, 1992-2004 Cohorts (Figure 5) 36 Primary Factors Contributing to
the Fiscal Year 2003 MMI Credit Subsidy Reestimate (Figure 6) 37 Change in
Future Cash Flow Estimates for the Fund from Fiscal Year 2002 to Fiscal
Year 2003 (Figure 7) 38 Variables Contributing to the $3.9 Billion Change
in Estimated Cash Flows (Figure 8) 38

                                    Contents

Table 7:	Increase in Estimated Net Cash Outflows from Removing the Loss
Mitigation Adjustment Factor, 1992-2003 Cohorts (Figure 9) 38

Table 8: Actual Versus Estimated Conditional Claim Rates for Fiscal Year
2003, 1993-2003 Cohorts (Figure 10) 39 Table 9: Actual Versus Estimated
Conditional Prepayment Rates for Fiscal Year 2003, 1993-2003 Cohorts
(Figure 10) 39 Table 10: Amount of FHA Prepayments During Fiscal Years
2000-2004 (Figure 11) 40 Table 11: Capital Ratio Versus Economic Value of
the MMI Fund, Fiscal Years 2000-2004 (Figure 12) 40 Table 12: Amortized
Insurance-In-Force, Fiscal Years 2000-2004 (Figure 13) 40 Table 13:
Minimum Required Capital Ratio Versus Actual Capital Ratio (Figure 14) 41

Figures	Figure 1: Figure 2: Figure 3:

Figure 4:

Figure 5:

Figure 6: Figure 7: Figure 8: Figure 9:

Calculation of Credit Subsidy for the Fund
Relationship between Actuarial Review and Reestimate
Annual Credit Subsidy Reestimates for the MMI Fund,
Fiscal Years 2000-2004
Amount of the 2003 Reestimate Attributed to the
2001-2003 Cohorts, as a Percentage of the Original Loan
Amount, for Single-Family Loan Guarantee Programs
Original Estimated Credit Subsidy Rates and Most Recent
Reestimated Rates for the FHA and VA Loan Guarantee
Programs, 1992-2004 Cohorts
Primary Factors Contributing to the Fiscal Year 2003 MMI
Credit Subsidy Reestimate
Change in Future Cash Flow Estimates for the Fund from
Fiscal Year 2002 to Fiscal Year 2003
Variables Contributing to the $3.9 Billion Change in
Estimated Cash Flows
Increase in Estimated Cash Outflows from Removing the
Loss Mitigation Adjustment Factor, 1992-2003 Cohorts

                                      6 8

10

11

13 14 15 16 17

19

23

27 29

Figure 10: Actual Versus Estimated Conditional Claim and Prepayment Rates
for Fiscal Year 2003, 1993-2003 Cohorts

Figure 11: Amount of FHA Prepayments during Fiscal Years 2000-2004

Figure 12: Capital Ratio Versus Economic Value of the MMI Fund, Fiscal
Years 2000-2004

Figure 13: Amortized Insurance-In-Force, Fiscal Years 2000-2004

Contents

Figure 14: Minimum Required Capital Ratio Versus Actual Capital Ratio 30

Abbreviations

FCRA Federal Credit Reform Act
FHA Federal Housing Administration
HUD U.S. Department of Housing and Urban Development
OMB U.S. Office of Management and Budget
USDA U.S. Department of Agriculture
VA U.S. Department of Veterans Affairs

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.

A

United States Government Accountability Office Washington, D.C. 20548

September 2, 2005

The Honorable Bob Ney
Chairman
Subcommittee on Housing and Community Opportunity
Committee on Financial Services
House of Representatives

Dear Mr. Chairman:

The Department of Housing and Urban Development (HUD), through its
Federal Housing Administration (FHA), provides insurance for single
family home mortgage loans made by private lenders. During fiscal year
2004, FHA insured 892,591 mortgages, representing $107.7 billion in single
family mortgage insurance. The insurance program is supported by the
Mutual Mortgage Insurance Fund (Fund), which is financed through
insurance premiums that FHA charges its borrowers.1 FHA's mortgage
insurance program is currently a negative subsidy program, meaning that
the Fund is self-financed and operates at a profit.

In 2001 we reported that the Fund had an economic value, or net worth, of
about $15.8 billion (as of the end of fiscal year 1999) and a capital
ratio of
3.20 percent of the unamortized insurance-in-force,2 or the initial amount
of
the mortgages.3 We noted that the minimum required capital ratio of 2
percent, set by Congress in 1990, appeared sufficient to withstand
moderately severe economic downturns that could lead to worse-than
expected loan performance. In 2002, we reported that while loans made

1FHA also provides mortgage insurance for certain single-family programs,
such as condominiums and home equity conversion mortgages, through its
General and Special Risk Insurance Fund. The single-family mortgage
insurance programs supported by the General and Special Risk Insurance
Fund represented about 13 percent of all single-family mortgages that FHA
insured in 2004. The remaining 87 percent were insured through the Mutual
Mortgage Insurance Fund.

2The Omnibus Budget Reconciliation Act of 1990 defined the capital ratio
as the ratio of the Fund's capital, or economic net worth (economic
value), to its unamortized insurance-inforce. However, the act defined
unamortized insurance-in-force as the remaining obligation on outstanding
mortgages-a definition generally understood to apply to amortized
insurance-in-force. HUD has calculated the capital ratio using unamortized
insurance-inforce as it is generally understood-which is the initial
amount of mortgages.

3See Mortgage Financing: FHA's Fund Has Grown, but Options for Drawing on
the Fund Have Uncertain Outcomes, GAO-01-460 (Washington, D.C.: Feb. 28,
2001).

during the 1990s were performing much better than loans made in the 1980s,
performance was somewhat weaker for loans originated during the latter
1990s than for those originated earlier in the decade.4 Our analysis
suggested that changes in FHA's underwriting procedures and in the
conventional mortgage market may have increased the overall riskiness of
FHA's portfolio, potentially affecting the Fund's economic value and its
ability to withstand future economic downturns. Therefore, we cautioned
against concluding that the Fund could withstand specified economic
scenarios. In October 2004, FHA estimated that the Fund had an economic
value of about $22 billion and a capital ratio of 5.5 percent. However,
because of the uncertainty of these measures and recent declines in loan
performance, we continue to believe that caution is warranted.

In recent years, FHA has adjusted its budget estimates to reflect that,
while not requiring subsidy, the performance of FHA-insured loans and the
resulting cash flows were not as strong as previously estimated. Higher
estimated costs caused the program to be less profitable than previously
estimated. Specifically, as of the end of fiscal year 2003, FHA submitted
a "reestimate" of $7 billion for the Fund. Given this substantial
reestimate of program cash flows, you asked us to (1) assess the
significance of the $7 billion reestimate, (2) determine what factors
contributed to the $7 billion reestimate and the underlying loan
performance variables influencing these factors, (3) assess how the loan
performance variables underlying the reestimate could impact future
estimates of new loans, and (4) assess what the reestimate and the
underlying loan performance variables mean for the long-term viability of
the Fund.

To respond to these objectives, we interviewed officials at FHA and staff
from the Office of Management and Budget (OMB). We collected and analyzed
budget data on FHA and comparable loan guarantee programs at the
Department of Veterans Affairs (VA), and Department of Agriculture (USDA)
to determine the significance of FHA's $7 billion credit subsidy
reestimate for the Fund. We interviewed FHA officials and FHA contractors
and collected and analyzed their written information and data to determine
the main factors contributing to the reestimate, the underlying loan
performance variables influencing these factors, and the likelihood these
variables could impact future estimates. We also analyzed FHA and other
data on new loan products and home mortgage industry trends to assess

4See Mortgage Financing: Changes in the Performance of FHA-Insured Loans,
GAO-02773 (Washington, D.C.: July 10, 2002).

what the reestimate and underlying loan performance variables mean for the
long-term viability of the Fund. Details about our scope and methodology
appear at the end of this letter.

We conducted our work from November 2004 through July 2005 in Washington,
D.C., in accordance with generally accepted government auditing standards.

Results in Brief	The $7 billion credit subsidy reestimate for the Fund was
more than twice the size of FHA's other recent reestimates and represented
a greater proportion of the Fund's recent cohorts5 than was the case for
the 2003 reestimates for comparable loan guarantee programs. While the $7
billion reestimate is unusually large, the upward direction of FHA's
recent reestimates in general is of concern because of the increase in
estimated cash outflows they represent. Also, FHA's current credit subsidy
estimates are, with one exception, higher than the original estimates for
all post-1991 cohorts.

Three major factors contributed to the $7 billion reestimate: a change in
the estimated future cash flows of its loans insured through 2003, the
difference between estimated and actual cash flows occurring during fiscal
year 2003, and an interest adjustment. The primary loan performance
variable underlying these factors is unexpectedly high claims. For
example, in 2003 FHA estimated that most cohorts would experience more
claim activity over the course of their 30-year terms than it estimated in
2002, increasing estimated cash outflows by $2.5 billion.
Higher-than-estimated prepayments as well as changing assumptions about
the impact that FHA's loss mitigation efforts could have on claims are
also important variables. For example, the revisions to loss mitigation
assumptions increased estimated cash outflows by $1.7 billion.

The change in expected claims underlying the $7 billion reestimate will
likely affect credit subsidy estimates for future loan cohorts, but the
effect of prepayments is less certain. Several recent policy changes and
trends may help explain the increase in claims, including changes to
underwriting guidelines, competition from the private sector, and an
increase in the use

5A cohort includes those direct loans or loan guarantees of a program for
which a subsidy appropriation is provided in a given year even if the
loans are not disbursed until subsequent years.

of down payment assistance. It appears that these policy changes and
trends will continue to impact claims, and thus they will likely continue
to add risk to FHA's portfolio. Prepayment rates increased significantly
prior to the 2003 reestimate, but it is less likely that the same
conditions that caused the surge in prepayments early in the decade will
be repeated. The revisions to loss mitigation assumptions will also affect
future estimates of subsidy by no longer artificially reducing claims,
though the significance may decline. FHA does not intend to use the same
assumption again given its greater historical experience with loss
mitigation.

Because the loan performance variables underlying the $7 billion
reestimate will likely persist to varying degrees, they are also likely to
affect estimates of the Fund's long-term viability. The capital ratio, a
measure of the Fund's long-term viability, has increased in recent years.
However, if the Fund's economic value declines or is restated at a lower
level than previously estimated, because of higher claims, and if the
insurance-in-force remains steady, because of declining prepayments, then
the capital ratio will decline. Whether the currently estimated 5.5
percent capital ratio or a lower capital ratio is sufficient depends on
the scenarios the Fund is expected to survive while maintaining the
minimum 2 percent reserve. Neither Congress nor HUD has established
criteria to determine how severe a stress the Fund should be able to
withstand.

To more reliably estimate program costs, we recommend that the Secretary
of HUD direct the FHA Commissioner to study and report the impact on the
forecasting ability of its loan performance models of variables that have
been found in other studies to influence credit risk, such as
payment-toincome ratios, credit scores, and the presence of down payment
assistance. We also recommend that when changing the definitions of key
variables, FHA should report the impact of such changes on the forecasting
ability of its loan performance models.

Background 	FHA was established in 1934 under the National Housing Act
(P.L. 73-479) to broaden homeownership, shore up and protect lending
institutions, and stimulate employment in the building industry by
providing mortgage insurance for loans made by private lenders. Generally,
borrowers are required to purchase single-family mortgage insurance when
the value of the mortgage is large relative to the price of the house.
Together, FHA, VA, USDA, and private mortgage insurers provide virtually
all of this insurance. FHA provides insurance for mortgages that finance
the purchase of

properties with one to four housing units, often by low-income, minority,
and first-time homebuyers.

The economic value of the Fund that supports FHA's guarantees depends on
the relative size of cash outflows and inflows over time. Cash flows out
of the Fund from payments associated with claims on defaulted loans and
refunds of up-front premiums on prepaid mortgages.6 To cover these
outflows, FHA receives cash inflows from up-front and annual insurance
premiums from borrowers and net proceeds from recoveries on defaulted
loans. If the Fund were to be exhausted, the U.S. Treasury would have to
cover lenders' claims directly.

The Fund remained relatively healthy from its inception until the 1980s
when claims and losses were substantial, primarily because of high
foreclosure rates in regions experiencing economic stress. These losses
prompted reforms that were enacted as part of the Omnibus Budget
Reconciliation Act of 1990 (P.L. 101-508). The reforms were designed to
place the Fund on an actuarially sound basis and required, among other
things, that it maintain a capital ratio of 2 percent of the
insurance-in-force and that an independent contractor conduct an annual
actuarial review of the Fund to analyze its economic value.

The Federal Credit Reform Act of 1990 (FCRA), enacted as part of the
Omnibus Budget Reconciliation Act of 1990, reformed budgeting methods for
federal credit programs, including FHA's mortgage insurance program. As a
result of FCRA, OMB requires federal credit agencies to report the actual
and estimated lifetime cost to the government of their programs in their
annual budgets. Similarly, federal accounting standards require agencies
to recognize the estimated lifetime costs of their programs in their
financial statements. To determine the expected cost of credit programs,
agencies predict or estimate the future performance of the programs on a
cohort basis. This cost, known as the subsidy cost, is the net present
value7 of estimated payments the government makes less estimated amounts
it

6FHA refunds a portion of the up-front premium based on the time elapsed
since the loan was originated and when a borrower prepays or refinances
their loan.

7Present value is the worth of the future stream of cash inflows and
outflows, as if they had occurred immediately. In calculating present
value, prevailing interest rates provide the basis for converting future
amounts into their "money now" equivalents. Net present value is the
present value of estimated future cash inflows minus the present value of
estimated future cash outflows.

receives over the life of the loan or loan guarantee, excluding
administrative costs. For the Fund, the overall subsidy is currently a
negative cost, meaning that the present value of cash inflows exceeds cash
outflows. Outflows include claims paid on foreclosed properties, refunds
of up-front insurance premiums, and foreclosed property holding costs,
while inflows include insurance premiums and proceeds from the sale of
foreclosed properties, over the life of the loan guarantees (fig. 1).

              Figure 1: Calculation of Credit Subsidy for the Fund

Source: GAO.

FCRA established a special budgetary accounting system to record the
budget information necessary to implement credit reform. For loans and
loan guarantees made during or after fiscal year 1992-the effective date
of credit reform-federal agencies use program and financing accounts to
handle credit transactions.8 The program account is included in budget
totals, receives separate appropriations for the administrative and
subsidy costs of a credit program, and records the budget authority and
outlays for these costs. The program account is used to pay the associated
subsidy cost to the financing account when a direct or guaranteed loan is
disbursed. The financing account, which is nonbudgetary,9 is used to
collect the subsidy cost from the program account, borrow from Treasury to
provide financing for loan disbursements, and record the lifetime cash
flows associated with direct loans or loan guarantees. In 2002, a new
capital reserve account was established for the Mutual Mortgage Insurance

8Liquidating accounts were established to handle credit transactions on a
cash basis for precredit reform loans and loan guarantees.

9Nonbudgetary accounts may appear in the budget document for informational
purposes but are not included in the budget totals for budget authority or
budget outlays.

Fund to maintain reserves for the post-1991 cohorts. In 2003 this new
account started earning interest on Treasury investments, collecting
negative subsidy and downward reestimates from the financing account, and
paying upward reestimates.

Agencies are required to reestimate subsidy costs annually to reflect
actual loan performance and expected changes in estimates of future loan
performance. Annual estimates of a program's expected lifetime subsidy
change from year to year. Beyond changes in estimation methodology, each
additional year provides more historical data on loan performance that may
influence estimates of the amount and timing of future claims and
prepayments. Economic assumptions also change from one year to the next,
including assumptions on interest rates, unemployment, and home prices.
Assumptions about the impact of policy changes also can affect estimates
of subsidy costs-for example, by changing how loans are serviced or the
treatment of foreclosed properties, which potentially influences the
timing and amount of losses.

In accordance with the Omnibus Budget Reconciliation Act of 1990, FHA
contracts with private firms to prepare an annual actuarial review.10
Figure 2 illustrates the relationship between the actuarial review and the
credit subsidy reestimate. For the review, the contractors develop
econometric loan performance models to estimate future claim and
prepayment activity for the loans FHA insures. The contractors also
develop a cash flow model through which they run the output of the loan
performance models. The actuarial cash flow model calculates the net
present value of future cash flows in and out of the Fund to estimate its
economic net worth and capital ratio. FHA also uses the actuarial claim
and prepayment data with its credit subsidy cash flow model to estimate
the net present value of future cash flows for the budget and the ending
balance of the liability for loan guarantees in the financial statements.
At the end of the fiscal year, FHA

10From 1989 to 1998, Price Waterhouse (PricewaterhouseCoopers as of 1998)
performed the actuarial review; from 1999 to 2003, Deloitte & Touche
performed the review; in 2004, Technical Analysis Center, Inc., was
awarded the contract.

uses a "balances approach" to compare the resources in the financing
account to the liability for loan guarantees.11 The difference is the
credit subsidy reestimate.

         Figure 2: Relationship between Actuarial Review and Reestimate

                                  Source: GAO.

In November 1996, FHA implemented a new loss mitigation program that
included a range of options to help homeowners who have defaulted on their
mortgage to either retain their homes or enable FHA to dispose of them in
ways that reduced the costs of foreclosure. The loss mitigation program
has five options: (1) special forbearance, or a repayment agreement
between the lender and borrower to reinstate a loan; (2) loan
modification, which provides borrowers with a permanent reduction in

11Current OMB guidance allows agencies to use either the "traditional
approach" or the "balances approach" to reestimate costs. The traditional
approach uses both actual past and estimated future cash flows to
calculate a revised expected cost. Then the amount of the reestimate is
based on the change in the expected cost. HUD uses the balances approach,
which compares the net resources (cash, other assets, and liabilities) in
the financing account to the total estimated future cash flows. Both
approaches yield simular results. Figure 2 illustrates the balances
approach.

mortgage payment; (3) partial claim, which enables a borrower to get an
interest-free loan from HUD to bring their mortgage payments up to date;
(4) pre-foreclosure sale, which provides borrowers with a transition to
more affordable housing; and (5) deed-in-lieu of foreclosure, an
alternative to foreclosure whereby a borrower voluntarily deeds the
property to HUD and is released from all mortgage obligations.

  The $7 Billion Reestimate Is Significant for Its Size and Direction

The $7 billion credit subsidy reestimate for the Fund was more than twice
the size of other recent FHA reestimates and represented a greater
proportion of the Fund's recent cohorts than other 2003 reestimates for
comparable loan guarantee programs. Both this unusually large reestimate
and the upward direction of FHA's recent reestimates are matters for
concern. Overall, though the Fund still operates at a profit, FHA's
current reestimated credit subsidy rates are higher than FHA originally
estimated for all but one of the 1992 through 2004 cohorts. In comparison,
current reestimated subsidy rates for VA's loan guarantee program are
lower than VA originally estimated for all but one of the 1992 through
2004 cohorts.

The Reestimate Is Large The $7 billion reestimate FHA reported in its 2003
financial statements was Compared with Other by far the largest reestimate
FHA has made in recent years. As figure 3 Recent Reestimates and
illustrates, it was more than twice the size of any other reestimate from

2000 through 2004, indicating that FHA's actual and estimated cash
flowsPrograms have changed substantially.

Figure 3: Annual Credit Subsidy Reestimates for the MMI Fund, Fiscal Years
20002004

Dollars in billions

8

7

6

5

4

3

2

1

0

-1

-2

2000 2001 2002 2003 2004

Fiscal year

Source: GAO analysis of FHA data.

An alternative way of measuring the magnitude of the reestimate is by
comparing it with reestimates for comparable loan guarantee programs.
FHA's 2003 reestimate was also unusually large compared with reestimates
for the same year for VA's and USDA's single-family loan guarantee
programs.12 FHA reestimates the credit subsidy separately for each cohort
of loans that it insures, totaling the separate reestimates into one
overall reestimate for the fiscal year. Loans that FHA insured in 2001
through 2003 accounted for $4.5 billion, or 64 percent, of the total $7
billion reestimate. The $4.5 billion of the reestimate attributed to these
three cohorts of loans

12Because the age composition of these programs' portfolios may differ, we
selected only the three most recent cohorts for our analysis. These three
cohorts represented the majority of FHA's loan portfolio in 2003.

equaled 1.22 percent of their combined total endorsements.13 As figure 4
illustrates, this percentage is more than double that of comparable loan
guarantee programs at VA and USDA.

Figure 4: Amount of the 2003 Reestimate Attributed to the 2001-2003
Cohorts, as a Percentage of the Original Loan Amount, for Single-Family
Loan Guarantee Programs

                                 Percentage 1.5

                                      1.2

FHA has estimated negative credit subsidies for the Fund since 1992, when
credit reform became effective. However, with one exception, current
reestimated subsidy rates for FHA's loan guarantees are less favorable
than originally estimated. Meanwhile, across the country home prices have
been growing faster and more uniformly since 2000 than they grew during
the 1990s and most of the 1980s. This indicates that very few borrowers
would have seen their home values decline to the point at which their
homes were

13Figure 4 is based on data from the fiscal year 2006 Federal Credit
Supplements, which reports $369 billion in Fund loans endorsed
(guaranteed) to date for the fiscal year 20012003 cohorts. According to
the Federal Credit Supplement, 100 percent of Fund loan guarantees are
endorsed in the first year.

                                       .9

                                       .6

                               .3 0.0 FHA VA USDA

  Source: GAO analysis of Federal Credit Supplements, fiscal years 2005-2006.

    FHA's Current Reestimated Subsidy Rates Are Less Favorable Than Its Original
    Estimates

worth less than their mortgage balances, putting them at a greater risk of
foreclosure and causing subsidy rates to worsen. In keeping with the trend
of increasing home prices, current reestimated rates for VA's program are
more favorable than originally estimated. As shown in figure 5, the
original and current subsidy cost estimates for FHA's 1992 through 2004
cohorts were negative, meaning FHA estimated total cash inflows to be
greater than outflows over the life of each cohort. FHA's most recent
reestimates indicate that all but the 1992 cohort will be less profitable
than originally estimated, though FHA is not estimating that these cohorts
will have overall negative cash flows. In comparison, the original subsidy
estimates for VA's 1992 through 2004 cohorts did indicate negative cash
flows, meaning VA estimated that the present value of total cash outflows
would exceed inflows over the life of each cohort. With the exception of
one cohort, VA's reestimated subsidy costs are all lower than originally
estimated, indicating that VA currently estimates that its cohorts will
perform better than originally expected. However, VA estimates that
several cohorts will continue to have overall negative cash flows.

Figure 5: Original Estimated Credit Subsidy Rates and Most Recent
Reestimated Rates for the FHA and VA Loan Guarantee Programs, 1992-2004
Cohorts

FHA VA

                              Rates Rates 2.5 -3.5

1992 1994 1996 1998 2000 2002 2004 Cohort

1992 1994 1996 1998 2000 2002 2004 Cohort

                             Original estimate rate

                             FY2005 reestimate rate

Source: GAO analysis of Federal Credit Supplements, fiscal years
2005-2006.

The $7 Billion The $7 billion reestimate represents the changes in FHA's
estimates of

future loan performance and the change in cash flows stemming from
theReestimate Primarily difference between estimated and actual loan
performance during fiscal Reflects Higher-Than-year 2003. These changes
primarily reflect the impact of higher-than-Estimated Claims estimated
claims, but also reflect the impact of higher-than-estimated

prepayments and a technical change in FHA's calculation of claims. The

reestimate also represents an interest adjustment (fig. 6).

Figure 6: Primary Factors Contributing to the Fiscal Year 2003 MMI Credit
Subsidy Reestimate

                              Dollars in billions

                     Change in estimated future cash flows

    Three Main Factors Contributed to the Reestimate

The largest contributing factor-55 percent-was the $3.9 billion difference
between FHA's fiscal year 2003 estimates of the net present value of
future cash flows and the estimates it made one year earlier. As
previously discussed, FHA estimates the value of expected future cash
flows each year by calculating the present value of anticipated cash
outflows, such as claim payments and premium refunds, and subtracting
inflows, such as insurance premiums and proceeds from the sale of
foreclosed properties. In 2002, FHA estimated that the net present value
of future cash flows for the 1992 through 2002 cohorts was a positive $1.9
billion, meaning that FHA expected cash inflows to exceed cash outflows on
a net present value basis. In 2003, FHA estimated that the net present
value of future cash flows for the 1992 through 2003 cohorts was negative
$2 billion, meaning that FHA expected future cash outflows to exceed
future cash inflows.14 As figure 7 illustrates, the difference between the
two estimates is $3.9 billion.

14Lifetime cash flow estimates continued to be positive, primarily because
of positive cash flows occurring earlier in the life of the cohort.

                              Interest adjustment

Difference between estimated and actual cash flows occurring during FY
2003

                       Source: GAO analysis of FHA data.

Figure 7: Change in Future Cash Flow Estimates for the Fund from Fiscal
Year 2002 to Fiscal Year 2003

Dollars in billions

2.0

1.5

1.0

.5

0 $3.9B

-.5

-1.0

-1.5

-2.0

2002 2003

Year

Source: GAO analysis of FHA financial statements, fiscal years 2002-2003.

The second factor contributing to the $7 billion reestimate-30 percent-
was the $2.1 billion difference between estimated and actual cash flows
occurring during fiscal year 2003. This amount indicates that FHA had $2.1
billion less in cash inflows during 2003 than it had estimated it would
have a year earlier. The final factor contributing to the reestimate (15
percent) was the $1.1 billion of interest on the reestimate. OMB guidance
requires agencies to calculate an interest adjustment on the reestimate.15
In FHA's case, the interest adjustment increased the total reestimate by
$1.1 billion.

15Circular No. A-11, Part 5: Federal Credit Programs, Office of Management
and Budget, June 2002.

    A Change in Estimated Claims Was the Primary Loan Performance Variable
    Behind the $3.9 Billion Change in Estimated Future Cash Flows

Approximately $2.7 billion (70 percent) of the $3.9 billion net change in
FHA's estimate of future cash flows stems from changes in FHA's estimates
of claims and, to a lesser extent, prepayments (fig. 8). That is, FHA
changed its estimate of future loan performance based on its observation
of actual loan performance during 2003 and revised economic assumptions.
In 2003 FHA estimated that, except for the 1993 and 1994 loan cohorts, all
cohorts would experience more claim activity over the course of their
30-year terms-and thus increase FHA's outflows-than estimated in 2002. The
cash flows associated with these claims increased estimated cash outflows
by $2.5 billion, accounting for 92 percent of the $2.7 billion. Increases
in the expected level of prepayments also affected FHA's estimate of
future cash flows. FHA estimated in 2003 that about half of the cohorts
would experience more prepayment activity than it had estimated in 2002.
Because of the increase in estimated prepayments, FHA expected to collect
less premium income and to pay out premium refunds more often, reducing
estimated cash inflows by about $200 million and accounting for 8 percent
of the $2.7 billion.

 Figure 8: Variables Contributing to the $3.9 Billion Change in Estimated Cash
                                     Flows

                       Source: GAO analysis of FHA data.

Another major variable that contributed to the $3.9 billion change in
estimated future cash flows was a technical change in FHA's calculation of

claims that increased the reestimate by $1.7 billion. Specifically, for
estimates prepared during fiscal years 2001 and 2002, FHA used a cash flow
assumption-a loss mitigation adjustment factor-to reduce the claim rates
predicted by the actuarial review and used in the subsidy cash flow model.
FHA had been using this factor in the belief that the historical data used
to estimate claim rates did not include enough years under the loss
mitigation program to adequately reflect the impact of this program-that
is, an expected decline in claims. However, FHA officials stated that in
fiscal year 2003 FHA removed the factor because the historical loan
performance data, which by then included more years of experience with the
loss mitigation program, sufficiently reflected the program's impact. In
addition, FHA noted that its actuarial review was underestimating claims,
making it counterproductive to use a loss mitigation adjustment factor
that further reduced the actuarial claim predictions. Removing the loss
mitigation adjustment factor from the 2003 subsidy cash flow model
increased the reestimate by a total of $1.7 billion, with the greatest
increase related to loans made in the most recent years (fig. 9).

Figure 9: Increase in Estimated Cash Outflows from Removing the Loss
Mitigation Adjustment Factor, 1992-2003 Cohorts

Dollars in millions 500

400

300

200

100

0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Cohort

                       Source: GAO analysis of FHA data.

The above increases in estimated cash outflows are offset by the estimated
additional cash inflows from new loans that FHA insured in 2003.
Specifically, FHA estimated that for loans originating during 2003, future
inflows would exceed future outflows by $1 billion. Several other factors
had much smaller positive or negative impacts on future cash flows. The
net impact of these other factors contributed $500 million to the
reestimate.

    Higher-Than-Estimated Claims and Prepayments Contributed to the $2.1 Billion
    Difference between Estimated and Actual Cash Flows Occurring during 2003

The remaining part of the $7 billion-$2.1 billion-represents the
difference between estimated and actual cash flows occurring during fiscal
year 2003. Certain elements of the difference relate to the 1992 through
2002 cohorts, including $330 million in underestimated claims and
recoveries on claims and $1 billion in overestimated net fees (insurance
premium receipts less premium refunds). The remaining $700 million relates
to cash flow differences associated with the 2003 cohort.

Our analysis of loan performance data found that claims and prepayments
occurring during 2003 exceeded FHA's estimates. As figure 10 illustrates,
actual claim activity in fiscal year 2003 exceeded estimated claim
activity for 2003-by twice as much in some cases-for the majority of loan
cohorts. For example, FHA estimated that about 1.6 percent of all the
loans it insured in 2000 that were in the portfolio at the beginning of
2003 would result in a claim during 2003. However, 4 percent of such loans
actually ended in a claim in 2003. Actual prepayment activity exceeded
estimated prepayment activity for all loan cohorts. For example, FHA
estimated that 14 percent of all the loans it insured in 2001 that were
still in the portfolio at the beginning of 2003 would prepay during 2003.
However, more than 40 percent of such loans actually prepaid during 2003.
Because of the additional claims it paid, up-front premiums it refunded,
and the annual premiums it lost, FHA's cash inflows for the year declined
and resulted in a $2.1 billion upward adjustment of the Fund's credit
subsidy.

 Figure 10: Actual Versus Estimated Conditional Claim and Prepayment Rates for
                      Fiscal Year 2003, 1993-2003 Cohorts

Claim rate Prepayment rate

Percentage Percentage

                                50 40 30 20 10 0

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 1993 1994 1995 1996
1997 1998 1999 2000 2001 2002 2003 Cohort Cohort

                                    Estimate

Actual

                       Source: GAO analysis of FHA data.

Note: The 2003 estimate data are from the 2002 actuarial review. The 2003
actual data are from the 2004 actuarial review.

  The Loan Performance Variables Underlying the $7 Billion Reestimate Will
  Likely Affect Future Credit Subsidy Estimates, but Are Being Addressed

The events behind the change in expected claims underlying the $7 billion
reestimate will likely continue to affect future credit subsidy estimates,
though prepayments may have a smaller effect. The one-time modeling change
caused by removing the loss mitigation adjustment factor should also
continue to have an effect on future estimates, though its significance
may decline.

    Higher Claims Will Likely Continue, but FHA Is Taking Steps to Improve Its
    Estimates

As we have seen, the $7 billion reestimate was largely due to
higher-thanestimated claims. Several recent events may help explain this
increase, including changes to underwriting guidelines, competition from
the private sector, and an increase in the use of down payment assistance.
FHA has taken some steps to tighten underwriting and to better estimate
claims, but it is not clear that these steps will be sufficient to reverse
recent increases in claims or significantly improve future estimates of
claims.

According to FHA, revised underwriting guidelines issued in 1995
represented significant changes that would enhance home-buying
opportunities for a substantial number of borrowers. These changes made it
easier for borrowers to qualify for loans and for higher loan amounts. In
previous work, we noted that these underwriting changes may partly explain
the higher claim rates of the late 1990s.16 FHA officials told us that
since making these changes, FHA's share of first-time homebuyers has
increased by more than 30 percent, and its share of minority homebuyers
has increased by 40 percent. FHA officials noted that these borrowers are
more susceptible to changes in economic conditions and, thus, may be more
likely to default on their mortgages. The officials also noted that, while
this change in the composition of their borrowers had resulted in a
one-time increase in claims, claims have leveled off and should remain
steady at the new level.

To evaluate the impact of the underwriting changes, FHA introduced a
simple variable into its annual actuarial models that captures whether or
not a loan was made after fiscal year 1995. This variable is intended to
capture the one-time impact of the 1995 underwriting changes, not to
capture any adverse trends that might result from changes that accrue over
time, such as increasing competition from the private sector or the
growing prevalence of down payment assistance. If there are adverse
trends, as opposed to only one-time changes, the model will not fully
capture them and, therefore, will likely underestimate future claims. For
example, if loans with down payment assistance have higher claims and if
this category of loans grows over time, then the claim model will
consistently underestimate claims and the model's error will worsen with
time.

16See Mortgage Financing: Changes in the Performance of FHA-Insured Loans,
GAO-02773 (Washington, D.C.: July 10, 2002).

In 2002, we reported that the performance of loans insured during the late
1990s was weaker than the performance of loans originated earlier in the
decade. We noted then that increased competition and changes in the
conventional mortgage market could result in FHA's insuring relatively
more loans that carried greater risk. These issues continue to be
significant. In recent years, private mortgage insurers and conventional
mortgage lenders have increasingly offered products that compete with FHA
for homebuyers who are borrowing more than 95 percent of the value of
their home. In addition, automated underwriting systems and creditscoring
analytic software are believed to be able to more effectively distinguish
low-risk loans for expedited processing. If, by selectively offering these
low down payment loans to better risk borrowers, conventional mortgage
lenders and private mortgage insurers were able to attract lower-risk
borrowers that would have traditionally sought FHAinsured loans, recent
FHA-insured loans with down payments of less than 5 percent may be more
risky on average than they have been historically.

A growing trend that has raised some concerns and may increase claims is
the use of seller-funded down payment assistance for mortgages insured by
FHA. FHA requires borrowers to make a 3 percent contribution toward the
purchase of the property, but that contribution can come indirectly
through borrowers' relatives or nonprofit organizations. Although FHA does
not permit down payment funds to come directly or indirectly from sellers,
it does permit nonprofits that receive contributions from sellers to
provide down payment assistance to homebuyers. Many conventional mortgage
products also permit down payment funds to come from sources other than
the borrower; however, the terms of these mortgage products generally
stipulate that such funds cannot come either directly or indirectly from
an interested or seller-related party. A HUD Office of the Inspector
General evaluation of FHA-insured loans found that loans with down payment
assistance from seller-funded nonprofits had a greater risk of default and
that the percentage of FHA-insured loans with down payment assistance from
seller-funded nonprofits was growing at an increasing rate. As of July
2005, FHA had not revised its policies regarding acceptable sources of
down payment assistance or imposed additional underwriting requirements on
borrowers who obtained down payment assistance from seller-funded
nonprofits. At your request, we are currently conducting a study on down
payment assistance and evaluating the performance of these loans.

A program assessment jointly prepared by OMB and FHA and included with the
2006 President's Budget noted that FHA's loan performance model

is neither accurate nor reliable because it consistently under predicts
claims. For the 2004 actuarial review, FHA worked with a new contractor to
redesign and respecify its loan performance models. FHA continues to work
on improving its new models so that it can more accurately and reliably
predict claims and prepayments. Several factors distinguish the new models
from those used previously. First, the new models use quarterly data,
while the previous models used annual data. In addition, the new models
explicitly address the time lag in claims and the implications of the time
lag for prepayments,17 and allow for a closer correspondence between the
actual and predicted time pattern of claim and prepayment rates. These
changes may improve the models' ability to predict the number and timing
of claims and prepayments. Nonetheless, for the 2004 actuarial review FHA
had to adjust model estimates of claims that will occur in fiscal years
2004 through 2006 for all loan cohorts. By the third quarter of fiscal
year 2004, while FHA was preparing the 2004 actuarial review, FHA realized
that actual claims for the year were outpacing the amount of estimated
claims based on data from the first half of 2004 and earlier. FHA and its
contractors assumed this difference was caused by a temporary deviation
and adjusted the model's projected claim rates to match the recorded claim
counts. Specifically, FHA applied a claim rate multiplier to increase
estimated claim rates by 50 percent for all cohorts for fiscal years 2004
and 2005 and by 25 percent for fiscal year 2006. Because FHA was
responding to what it believed to be a temporary deviation, it did not
apply the multiplier to any years after 2006.

The new models also eliminated some explanatory variables, such as
unemployment rates and payment-to-income ratios, and altered the
definitions of other key variables. For example, the previous models
assumed that borrowers who passed up profitable refinancing opportunities
would experience permanently higher claim rates- sometimes referred to as
burnout-while the new models assume that higher claim rates are a
temporary phenomenon that will last only 2 years. In addition, neither
model incorporates certain variables that have been found to be important
in assessing credit risk, such as credit scores and the source of down
payments. FHA officials are researching these variables currently. FHA
officials told us they will not be including credit scores for

17A loan may be seriously delinquent for several quarters before that
delinquency is resolved. Because it is difficult for a borrower with a
delinquent loan to obtain a new loan in order to refinance, several
quarters may pass during which time a loan has a high probability of
resulting in a claim, because it is delinquent, and has a low probability
of resulting in a prepayment, because the borrower cannot refinance using
conventional channels.

the 2005 actuarial review, though they are considering ways to account for
credit scores in the 2006 actuarial model. For the 2005 model, FHA made
adjustments for the source of down payments by adjusting the loan-tovalue
ratio for seller-funded down payment assistance. On balance, it is not
clear that these changes to the actuarial models will permit FHA to more
reliably estimate claim (or prepayment) activity. In fact, the $7 billion
reestimate was followed a year later with an upward reestimate of $2.3
billion for fiscal year 2004.

Prepayments Had a Smaller While claims may have been the largest driver
behind the reestimate, Impact on the $7 Billion prepayments also had an
impact. As we discussed above, FHA experienced Reestimate and Their a
significant increase in prepayment activity from 2001 through 2004. As

figure 11 illustrates, between 2000 and 2001, the dollar amount ofImpact
on Future Credit prepayments more than doubled, rising from $37 billion to
$82 billion. Subsidy Estimates Is Prepayments reached a total of $190
billion in 2003 and decreased slightlyUncertain to $123 billion in 2004.

Figure 11: Amount of FHA Prepayments during Fiscal Years 2000-2004

Dollars in billions 200

150

100

50

0 2000 2001 2002 2003 2004 Fiscal year

Source: GAO analysis of FHA data.

FHA experienced surges of prepayment activity in the mid-1980s and early
1990s. All three of these time periods coincided with periods of declining
interest rates, though the rates of prepayment are highest in 2003, when
mortgage interest rates reached a 30-year low. As of July 2005, mortgage
interest rates had declined even further from their 2004 level, though by
a much smaller amount compared with the 2002-2003 decline. Because it is
difficult to project future interest rates, it is difficult to project
their impact on future prepayment activity. As we noted previously, house
prices have risen faster in the first part of the decade than they did in
the 1990s or most of the 1980s. Rapid appreciation in housing prices
permits borrowers to refinance using conventional loans, however, it is
uncertain that the upward trend in appreciation will continue.

    One-time Removal of Loss Mitigation Factor Should Continue to Affect Future
    Estimates

The removal of the loss mitigation adjustment factor also had a notable
impact, affecting the cash flow model's calculation of claims and thus
contributing $1.7 billion to the reestimate. FHA does not intend to use
this adjustment again given its greater historical experience with loss
mitigation. That is, FHA expects that the historical data on which loan
performance estimates are based will include and reflect more years of
experience with the loss mitigation program. However, this change in the
assumptions used in estimating loan performance will affect the estimated
subsidy costs of new cohorts because estimates of future cohorts will not
include the loss mitigation adjustment factor, though the significance of
no longer making this adjustment may decline over time. That is, as FHA
estimates of loan performance include more historical experience with loss
mitigation, any positive effect loss mitigation may have would be
reflected in the loan performance variables.

Recent Policy Changes May In recent years, FHA has introduced several
policy changes that may affect Affect Claims and claim activity. Since
2000, FHA has loosened some underwriting Prepayments procedures to
encourage homeownership. For example, FHA increased the

amount of the mortgage payment it will permit relative to borrower income.
Specifically, in April 2005, FHA increased its maximum paymentto-income
ratio from 29 percent to 31 percent and its debt-to-income ratio

from 41 percent to 43 percent.18 By increasing these qualifying ratios,
FHA could offer mortgage insurance to borrowers who would not have
otherwise been approved for a loan. However, borrowers who devote more of
their income to their mortgage payments could have trouble meeting their
payments if they encounter financial trouble. FHA made these changes in
response to recent federal tax cuts, which increased potential borrowers'
buying power. Therefore, FHA noted, the changes should broaden eligibility
without increasing the risk of default.

FHA has also taken steps to tighten some underwriting guidelines. For
example, in 2000 it changed its policies on gift transfers and the types
of assets that may be considered for cash reserves. FHA now requires more
documentation for gift transfers to ensure that the funds are applied
toward the borrower's down payment and come from sources with no interest
in the sale of the property. However, in a recent review of FHA's new
mortgage loan products, we found that FHA does permit nonprofits that
receive contributions from sellers to provide down payment assistance to
borrowers. 19 FHA also now requires lenders to ensure that borrowers'
assets, such as retirement accounts, can be easily converted into cash
before applying them toward cash reserves. This policy change requires
that lenders account for any applicable taxes or withdrawal penalties that
borrowers may incur when converting their assets to cash, potentially
reducing the amount of cash available to these borrowers. In early 2004
FHA introduced the Technology Open to Approved Lenders Mortgage Scorecard.
This tool is used in conjunction with automated underwriting systems to
evaluate the credit risk of borrowers who apply for FHA insured loans. The
introduction of the new mortgage scorecard may help FHA and lenders more
efficiently and effectively identify and evaluate credit risk and,
therefore, may help reduce claims.

FHA has taken measures to enhance the effectiveness of its loss mitigation
program. In 2002, FHA modified some of its loss mitigation options to give

18The payment-to-income ratio, also referred to as the
housing-expense-to-income ratio, examines a borrower's expected monthly
housing expenses as a percentage of the borrower's monthly income. The
debt-to-income ratio looks at a borrower's expected monthly housing
expenses plus long-term debt as a percentage of the borrower's monthly
income. FHA limits the monthly mortgage payment to no more than 31 percent
of monthly gross income (before taxes) and limits the mortgage payment
combined with other debts to no more than 43 percent of income.

19See Mortgage Financing: Actions Needed to Help FHA Manage Risks from New
Mortgage Loan Products, GAO-05-194 (Washington, D.C.: Feb. 11, 2005).

lenders more flexibility to assist borrowers who are unable to make their
monthly payments, help avoid or reduce the time and expense of the
foreclosure process, and enable borrowers to obtain credit again in the
future. FHA believes that the introduction of loss mitigation and changes
made since the program's implementation should reduce losses it incurs
when borrowers default on their loans. FHA also introduced the Accelerated
Claim Disposition demonstration program in 2002 (referred to as the "601"
program) to streamline the claim and property disposition processes with
the goal of reducing losses to the Fund.

FHA has also made some recent policy changes that may affect prepayment
activity. For example, FHA changed its up-front mortgage insurance premium
rules for mortgages endorsed after December 2004. In the past, FHA
refunded a percentage of the up-front premium to borrowers when they
prepaid their loans, typically by refinancing or selling their homes.
Borrowers were entitled to this refund even when they refinanced outside
of FHA. For new loans guaranteed after December 2004, FHA will no longer
refund a percentage of the up-front premium to borrowers who refinance
their mortgages outside of FHA. FHA also shortened the refund schedule of
the up-front premium from 5 to 3 years. These changes could encourage
borrowers to refinance their mortgage with another FHA-insured loan, while
reducing the amount of refunds that FHA pays to borrowers who refinance or
sell their homes. However, these changes may also discourage some
borrowers from choosing to finance their home purchases with an
FHA-insured mortgage. FHA predicts that the changes to its up-front
premium rules will increase cash flows by about $168 million annually.

  The Loan Performance Variables Underlying the Reestimate Could Affect
  Estimates of the Fund's Long-Term Viability

The effect of recent trends on the loan performance variables underlying
the $7 billion reestimate will likely persist to varying degrees and
therefore affect estimates of the Fund's long-term viability. The capital
ratio, a measure of the Fund's long-term viability, has increased in
recent years. However, should the economic value decline or be restated as
lower than previously estimated (due to higher-than-estimated claims), and
should the insurance-in-force remain steady (due to declining
prepayments), then the capital ratio will decline. Whether the currently
estimated 5.5 percent capital ratio or a lower capital ratio is sufficient
to meet federal requirements depends on what conditions the Fund is
expected to survive while maintaining the minimum 2 percent reserve.
Neither the Congress nor HUD has established criteria to determine how
severe of a stress the Fund should be able to withstand.

The Fund is required to maintain a minimum capital ratio (a measure of its
long-term viability) of 2 percent of the insurance-in-force. As figure 12
illustrates, the Fund's capital ratio has been well above 3 percent and
rising since fiscal year 2000. The economic value of the Fund-the sum of
existing capital plus the net present value of expected future cash flows
from existing cohorts-has also been rising for a number of years, though
it declined in fiscal year 2004. However, the Fund's insurance-in-force
declined 20 percent between 2002 and 2004 in response to increased claim
and prepayment activity during those years and a decline in new loan
originations. As the capital ratio is the Fund's economic value divided by
its insurance-in-force, the capital ratio only increased because the
decrease in the insurance-in-force was proportionately larger than the
decrease in the economic value of the Fund.

Figure 12: Capital Ratio Versus Economic Value of the MMI Fund, Fiscal
Years 20002004

Dollars in billions Capital ratio (percentage) 25 6

5 20

4 15

3

10 2

5 1

00 2000 2001 2002 2003 2004 Fiscal year

Economic value of MMI fund

Capital ratio Source: GAO analysis of FHA data.

If the economic value declines or is restated at a lower level than
previously estimated and if the insurance-in-force does not decline (for
example, due to substantial prepayments), then the capital ratio will
decline. As we noted, the events that help explain the increase in claims
underlying the $7 billion reestimate-such as changes in underwriting
guidelines, competition from the private sector, and an increase in the
use of down payment assistance-do not appear to be one-time events and
likely will continue to add risk to FHA's portfolio. For example, the
borrowers FHA has attracted since introducing its 1995 underwriting
changes are more susceptible to economic downturns and, therefore, more
likely to default on their mortgages. Further, despite HUD's Office of the
Inspector General finding that loans with down payment assistance from
seller-funded nonprofits have a greater risk of default, the percentage of
FHA-insured loans with down payment assistance from seller-funded
nonprofits is growing at an increasing rate.

FHA has introduced several policy changes that may help reduce claim
activity, such as requiring lenders to ensure that borrowers' assets can
be easily converted into cash before applying them toward cash reserves
and introducing the TOTAL Mortgage Scorecard to evaluate the credit risk
of borrowers who apply for FHA-insured loans. Despite these changes, it
seems likely that FHA's higher level of claims will continue. Higher claim
rates imply a lower estimated economic value of the Fund.

While prepayment rates increased significantly in the early part of the
decade, it is less likely that the same conditions that caused the surge
in prepayments will be repeated, reducing the impact that prepayments may
have on reducing the insurance-in-force. As we noted above, the three
surges of prepayment activity that FHA experienced coincided with periods
of declining interest rates. The rates of prepayment were highest in 2003,
when mortgage interest rates reached a 30-year low. As figure 13
illustrates, the Fund's amortized insurance-in-force also declined in
fiscal years 2003 and 2004 as prepaying borrowers left the portfolio.
Mortgage interest rates have been even lower in the spring and early
summer of 2005. Even if prepayments slow, should claim activity continue
to be higher and FHA be unable to compete for new borrowers, the Fund's
insurance-inforce may shrink. But if the net effect is that the size of
the portfolio stabilizes or declines only slightly, higher claim activity
could result in a lower capital ratio.

Figure 13: Amortized Insurance-In-Force, Fiscal Years 2000-2004

Amortized insurance-in-force (dollars in billions)

500

450

400

350

300

250

200

150

100

50

0 2000 2001 2002 2003 2004

Fiscal year

Source: GAO analysis of FHA data.

The long-term viability of the Fund depends on both the impact that the
underlying change in loan performance may have on the capital ratio and
the conditions or scenarios under which Congress expects the Fund to
maintain its 2 percent minimum reserve. A lower capital ratio would mean
that the Fund is less able to withstand adverse economic conditions. As
figure 14 illustrates, the Fund's capital ratio has been well above the 2
percent minimum and rising since fiscal year 2000. But whether the
currently estimated 5.5 percent capital ratio or a lower capital ratio is
sufficient depends on what conditions the Fund is expected to survive
while maintaining the minimum 2 percent reserve.

Figure 14: Minimum Required Capital Ratio Versus Actual Capital Ratio

Capital ratio (percentage)

6

5

4

3

2 Required minimum capital ratio

1

0 2000 2001 2002 2003 2004 Fiscal year

Source: GAO analysis of FHA data.

Because economic downturns put downward pressure on house prices and
incomes, they can stress FHA's ability to meet its obligations. Thus, it
is reasonable that measures of the financial soundness of the Fund would
be based on tests of the Fund's ability to withstand recent recessions or
regional economic downturns. The 2004 actuarial review examines four
stress scenarios, none of which are particularly severe. Three of the 4
stress tests examine one source of stress at a time, while one examines
two stresses simultaneously. A severe stress test would examine the
possibility of multiple stresses occurring simultaneously, such as a
decrease in house prices coupled with a decrease in recoveries on the sale
of foreclosed homes and an increase in the dispersion of house price
changes across multiple regions. Neither Congress nor HUD has established
criteria to determine how severe a stress the Fund should be able to
withstand. While the Fund continues to maintain a capital ratio above the
required minimum, we have recommended in the past that HUD develop
criteria that specify the economic conditions the Fund should be able to
withstand and the capital ratios currently consistent with those criteria.
We also recommended that the annual actuarial analysis give more attention
to tests of the Fund's ability to withstand appropriate stresses. Finally,
we

recommended that HUD develop better tools for assessing the impact that
policy changes may have on the volume and riskiness of the loans that FHA
insures.20

Conclusions	There are two important ways that FHA can manage risks to the
Fund and its ability to withstand economic downturns. First, FHA needs to
be able to reliably estimate program costs. To do so, FHA needs to
understand the factors that influence loan performance and, considering
this information, accurately estimate future claims and prepayments and
the resulting cash flows. Without better estimates of loan performance,
FHA cannot reasonably estimate the economic net worth of the Fund or its
capital ratio. Second, even if FHA can better estimate program costs, it
still needs to know what conditions the Fund is expected to endure while
maintaining the minimum 2 percent capital reserve.

  Recommendations for Executive Action

To more reliably estimate program costs, the Secretary of HUD should
direct the FHA Commissioner to study and report in the annual actuarial
review the impact of variables that have been found in other studies to
influence credit risk, such as payment-to-income ratios, credit scores,
and the presence of down payment assistance, on the forecasting ability of
the loan performance models used in FHA's actuarial reviews of the Fund.
FHA also should report in its annual actuarial review the impact of any
changes it makes to key variables, such as the burnout variable, on the
forecasting ability of the loan performance models.

  Agency Comments and Our Evaluation

We provided HUD, VA, and OMB with a draft of this report for their review
and comment. We received written comments from HUD, which are reprinted in
appendix III. We also received technical comments from HUD, which have
been incorporated where appropriate. VA and OMB did not have comments on
the draft.

HUD stated that it agrees with GAO's overall finding that higher than
projected claims were a significant variable underlying the $7 billion
reestimate, and that its 1995 underwriting changes help explain the

20See Mortgage Financing: FHA's Fund Has Grown, but Options for Drawing on
the Fund Have Uncertain Outcomes, GAO-01-460 (Washington, D.C.: Feb. 28,
2001).

increase in claims. HUD also agreed with our description of the steps it
has taken to better estimate claims in its recent actuarial reviews.

HUD raised a concern that our first recommendation would require FHA to
direct its actuarial contractor to include certain variables in its loan
performance models, and that this would compromise the requirement for an
independent actuarial study of the Fund. In response, we recommend instead
that FHA study and report in the annual actuarial review the impact of
such variables on the forecasting ability of the loan performance models.
HUD further noted that its contractor is actively considering the specific
variables that we had recommended FHA include in its annual actuarial
review.

In response to our second recommendation that FHA report in its actuarial
review the impact of any changes it makes to key variables on the
forecasting ability of its loan performance models, HUD noted that the
actuarial reviews and appendices contain full documentation of the models
and justifications for the selection of the included variables and their
definitions. However, we found, for example, that the 2004 actuarial
review did not fully document or justify the change in the definition of
the burnout variable. Specifically, the 2004 actuarial review contained
only a short statement regarding this change, with no accompanying
analysis of its impact on the forecasting ability of FHA's loan
performance models. We therefore continue to recommend that the annual
actuarial review include analyses of the impact of changes made to key
variables on the forecasting ability of the loan performance models.

As agreed with your office, unless you publicly announce the contents of
this report earlier, we plan no further distribution until 30 days from
the report date. At that time, we will send copies of this report to
interested Members of Congress and congressional committees. We will also
send copies to the Secretary of Housing and Urban Development and Director
of the Office of Management and Budget and make copies available to others
upon request. In addition, this report will be available at no charge on
the GAO Web site at http://www.gao.gov.

If you or your staff have any questions about this report, please contact
me at (202) 512-8678 or [email protected]. Contact points for our offices of
Congressional Relations and Public Affairs may be found on the last page
of this report. GAO staff who made key contributions to this report are
listed in appendix IV.

Sincerely yours,

William B. Shear Director, Financial Markets and Community Investment

Appendix I

Scope and Methodology

To assess the significance of the $7 billion reestimate, we interviewed
officials at the Department of Housing and Urban Development's (HUD)
Federal Housing Administration (FHA) and Office of the Inspector General
(OIG) and staff from the Office of Management and Budget (OMB). We
reviewed the fiscal year 2000-2004 audited financial statements for FHA to
compare the size and direction of MMI reestimates over time. We analyzed
data from the fiscal year 2005-2006 Federal Credit Supplements to compare
the size and direction of reestimates, by cohort, among comparable loan
guarantee programs at FHA, the Department of Veterans Affairs, and
Department of Agriculture.

To determine what factors contributed to the $7 billion reestimate and the
underlying loan performance variables influencing these factors, we
collected and analyzed supporting documentation for the reestimate,
including analyses prepared by FHA and work papers prepared by FHA's
financial statement auditor. We collected and analyzed the fiscal year
20002004 actuarial reviews of the MMI Fund and related loan performance
data to examine trends in loan performance and consider the impact that
model changes may have had on estimated subsidy costs. We collected and
analyzed fiscal year 2002-2003 credit subsidy cash flow models used to
calculate the reestimates for those years, to consider the impact of loan
performance on cash flows. We supplemented this analysis by interviewing
the 2003 financial statement auditors, OMB staff, officials in the OIG,
FHA staff, FHA contractors that assist in the preparation of the
reestimate, and the 2004 actuarial review contractors for background
information to verify our findings on the factors and underlying loan
performance variables.

To assess the control procedures governing the loan performance data we
collected, we reviewed the findings of our previous studies in which we
assessed the reliability of data for FHA-insured loans that came from the
same source as the data used in this report. While the data in these
previous reports covered a limited number of loan cohorts, the control
activities we reviewed apply to all cohorts. In 2004 we assessed the
reliability of a random sample of FHA-insured loans from the 1996-1999
cohorts, comparing seven elements of the paper loan file to the electronic
file to determine if they matched, and found no material errors. We also
reviewed several years' worth of FHA financial statement audits and found
no known or suspected problems with the relevant FHA information systems.
From these steps, we concluded these data were sufficiently

Appendix I Scope and Methodology

reliable for our analyses.1 In 2005 we obtained loan performance data on
FHA-insured loans from the 1992, 1994, and 1996 cohorts. To verify this
data, we met with FHA staff involved in generating the sample data set and
discussed data quality procedures with appropriate FHA staff. FHA
officials indicated that their data systems contain data entry checks and
that data submitted by lenders were reviewed by FHA. As part of its annual
financial statement audit, FHA's data system was audited by external
auditors, and no major issues concerning data quality were raised. Based
on these discussions, we determined that the FHA data were sufficiently
reliable for our analyses.2

To assess how the loan performance variables underlying the reestimate
could impact future estimates of new loans, we interviewed FHA officials
and the contractors for the 2004 actuarial review regarding the causes of
the loan performance variables and their impact on future estimates. We
also discussed recent and planned changes to the loan performance models
that may affect FHA's future estimates. We reviewed recent policy changes
that may impact loan performance variables by analyzing relevant policy
changes discussed in recent actuarial reviews and mortgagee letters issued
by FHA through the HUD Web site.

To assess what the reestimate and its underlying loan performance
variables mean for the long-term viability of the Fund, we analyzed FHA
and other data on new loan products and home mortgage industry trends. We
reviewed prior GAO reports describing changes in the home mortgage market
and FHA loan performance and used professional judgment to opine on
whether earlier concerns for the viability of the Fund persist.

1See Home Inspections: Many Buyers Benefit from Inspections, but Mandating
Their Use Is Questionable, GAO-04-462 (Washington, D.C.: April 30, 2004).

2See Mortgage Financing: Actions Needed to Help FHA Manage Risks from New
Mortgage Loan Products, GAO-05-194 (Washington, D.C.: Feb. 11, 2005).

Appendix II

                      Data for Figures Used in This Report

Table 1: Annual Credit Subsidy Reestimates For the MMI Fund, Fiscal Years
20002004 (Figure 3)

           Dollars in millions Fiscal year Credit subsidy reestimate

                                  2000 $3,350

                                  2001 -1,687

                                   2002 1,526

                                   2003 7,029

                                  2004 $2,340

                       Source: GAO analysis of FHA data.

Table 2: Amount of the 2003 Reestimate Attributed to the 2001-2003
Cohorts, as a Percentage of the Original Loan Amount, For Single-Family
Loan Guarantee Programs (Figure 4)

Current reestimate as a percentage of Agencies total disbursements

                                   FHA 1.22%

VA

USDA

Source: GAO analysis of Federal Credit Supplements, fiscal years 2005 and 2006.

Table 3: Original Estimated Credit Subsidy Rates and Most Recent
Reestimated Rates for the FHA and VA Loan Guarantee Programs, 1992-2004
Cohorts (Figure 5)

                               Original subsidy              Fiscal year 2005 
                Cohort               rate (MMI)         reestimate rate (MMI) 
                   FHA                          
                  1992                   -2.60%                        -3.03% 
                  1993                    -2.70                         -2.55 
                  1994                    -2.79                         -1.58 
                  1995                    -1.95                         -0.44 
                  1996                    -2.77                         -0.85 
                  1997                    -2.88                         -1.10 
                  1998                    -2.99                         -1.74 
                  1999                    -2.62                         -1.95 

Appendix II Data for Figures Used in This Report

(Continued From Previous Page)

                               Original subsidy              Fiscal year 2005 
                Cohort               rate (MMI)         reestimate rate (MMI) 
                  2000                    -1.99                         -0.55 
                  2001                    -2.15                         -0.94 
                  2002                    -2.07                         -1.07 
                  2003                    -2.53                         -1.53 
                  2004                    -2.47                         -1.61 
                  2005                    -1.82 
                  2006                    -1.70 

VA

                          1992          2.19%                           1.72% 
                          1993          2.33                             0.31 
                          1994          1.36                            -0.02 
                          1995          1.18                            -0.13 
                          1996          1.56           
                          1997          0.74                            -0.25 
                          1998          0.49                             0.01 
                          1999          0.45                             0.01 

2000 0.68 -0.25 2001 0.29 0.35 2002 0.39 0.27 2003 0.81 0.44 2004 0.50
-0.07 2005 -0.32 2006 -0.32

Source: GAO analysis of Federal Credit Supplements, fiscal years 2005 and
2006.

Table 4: Primary Factors Contributing to the Fiscal Year 2003 MMI Credit
Subsidy Reestimate (Figure 6)

       Dollars in billions                         
        Difference between                         
       estimated and actual   Change in estimated  
      cash flows for FY 2003      future cashflows  Interest on adjustment 
               $2.1                           $3.9                    $1.1 

Source: GAO analysis of FHA data.

Appendix II Data for Figures Used in This Report

Table 5: Change in Future Cash Flow Estimates for the Fund from Fiscal
Year 2002 to Fiscal Year 2003 (Figure 7)

Dollars in millions

                               Fiscal year Amount

                                  2002 $1,864

                                  2003 -2,008

Source: GAO analysis of FHA financial statements, fiscal years 2002-2003.

Table 6: Variables Contributing to the $3.9 Billion Change in Estimated
Cash Flows (Figure 8)

Dollars in billions

                    Loans          Removal of loss      Change in conditional 
           originating in                  mitigation    claim and prepayment 
                     2003    Other         adjustment                   rates 
                     $1.0    -$0.5              -$1.7                   -$2.7 

                       Source: GAO analysis of FHA data.

Table 7: Increase in Estimated Net Cash Outflows from Removing the Loss
Mitigation Adjustment Factor, 1992-2003 Cohorts (Figure 9)

                              Dollars in thousands

                                 Cohort Impact

                                  1992 $10,979

                                  1993 24,126

                                  1994 32,926

                                  1995 21,632

1996 40,131 1997 52,666 1998 108,630 1999 151,683 2000 142,118 2001
349,441 2002 451,067 2003 339,110

                       Source: GAO analysis of FHA data.

Appendix II Data for Figures Used in This Report

Table 8: Actual Versus Estimated Conditional Claim Rates for Fiscal Year
2003, 19932003 Cohorts (Figure 10)

                                     Cohort  Estimated claims   Actual claims 
                                       1993              0.60%          0.48% 
                                       1994               0.58           0.53 
                                       1995               0.96           1.71 
                                       1996               0.93           1.69 
                                       1997               1.08           2.29 
                                       1998               0.95           1.75 
                                       1999               0.92           1.93 
                                       2000               1.58           4.03 
                                       2001               0.87           1.96 
                                       2002               0.32           0.50 
                                       2003               0.01           0.01 
          Source: GAO analysis of FHA data.                    

Table 9: Actual Versus Estimated Conditional Prepayment Rates for Fiscal
Year 2003, 1993-2003 Cohorts (Figure 10)

                Cohort Estimated prepayments Actual prepayments

                               1993 17.10% 34.33%

                                1994 16.32 32.13

                       1995                        17.33             29.91 
                       1996                        17.58             32.45 
                       1997                        18.20             31.33 
                       1998                        17.36             36.54 
                       1999                        16.58             35.80 
                       2000                        22.73             34.67 
                       2001                        14.37             41.63 
                       2002                        7.78              33.81 
                       2003                        1.12               7.00 
         Source: GAO analysis of FHA data.                   

Appendix II Data for Figures Used in This Report

 Table 10: Amount of FHA Prepayments During Fiscal Years 2000-2004 (Figure 11)

Dollars in millions

                             Fiscal year Prepayment

                                  2000 $37,576

                                  2001 82,260

                                  2002 121,154

                                  2003 190,370

                                  2004 123,029

                       Source: GAO analysis of FHA data.

Table 11: Capital Ratio Versus Economic Value of the MMI Fund, Fiscal
Years 20002004 (Figure 12)

Dollars in millions

                  Fiscal year      Economic value               Capital ratio 
                         2000                  $16,962                  3.51% 
                         2001                   18,510                   3.75 
                         2002                   22,636                   4.52 
                         2003                   22,736                   5.21 
                         2004                   21,977                   5.53 

                       Source: GAO analysis of FHA data.

Table 12: Amortized Insurance-In-Force, Fiscal Years 2000-2004 (Figure 13)

Dollars in millions

Fiscal year Amortized insurance in force

2000 $449,867 2001 459,305 2002 466,598 2003 406,619 2004 $372,373

Source: GAO analysis of FHA data.

Appendix II Data for Figures Used in This Report

Table 13: Minimum Required Capital Ratio Versus Actual Capital Ratio (Figure 14)

                                                             Required minimum 
                 Fiscal year          Capital ratio             capital ratio 
                        2000                  3.51%                      0.02 
                        2001                   3.75                      0.02 
                        2002                   4.52                      0.02 
                        2003                   5.21                      0.02 
                        2004                   5.53                      0.02 

                       Source: GAO analysis of FHA data.

Appendix III

Comments from the Department of Housing and Urban Development

Appendix III
Comments from the Department of Housing
and Urban Development

Appendix IV

                     GAO Contact and Staff Acknowledgments	

GAO Contact William B. Shear (202) 512-8678

Staff
In addition to the above, Mathew Scire, Assistant Director, Anne Cangi,
Marcia Carlsen, Emily Chalmers, Austin Kelly, Mamesho Macaulay, Marc

Acknowledgments Molino, and Katherine Trimble made key contributions to
this report.

GAO's Mission
The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of The fastest and easiest way to obtain copies of GAO
documents at no cost

is through GAO's Web site (www.gao.gov). Each weekday, GAO postsGAO
Reports and newly released reports, testimony, and correspondence on its
Web site. To Testimony have GAO e-mail you a list of newly posted products
every afternoon, go to

www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone
The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone:
Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202) 512-6061

  To Report Fraud, Contact:	Waste, and Abuse in Web site:
  www.gao.gov/fraudnet/fraudnet.htm	

E-mail: [email protected] Programs Automated answering system: (800)
424-5454 or (202) 512-7470

Congressional
Gloria Jarmon, Managing Director, [email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125

Relations Washington, D.C. 20548

Public Affairs
Paul Anderson, Managing Director, [email protected] (202) 512-4800 U.S.
Government Accountability Office, 441 G Street NW, Room 7149 Washington,
D.C. 20548
*** End of document. ***