Federal Housing Administration: Decline in the Agency's Market	 
Share Was Associated with Product and Process Developments of	 
Other Mortgage Market Participants (29-JUN-07, GAO-07-645).	 
                                                                 
The Federal Housing Administration (FHA) historically has been an
important participant in the mortgage market, which includes	 
loans that carry government insurance or guarantees (such as	 
FHA-insured mortgages) and those that do not (conventional	 
mortgages). The conventional market comprises prime loans for the
most creditworthy borrowers and subprime loans for borrowers with
impaired credit. Reduced demand for FHA-insured mortgages--which 
are used primarily by borrowers who would have difficulty	 
obtaining conventional prime loans--has raised questions about	 
the agency's role in and ability to adapt to the mortgage market.
This report discusses (1) trends in FHA's share of the market for
home purchase mortgages from 1996 through 2005, and how they	 
compared with the trends for other market segments; and (2)	 
factors associated with the trends in FHA's market share and the 
implications of these trends for homebuyers and FHA. To address  
these objectives, GAO analyzed FHA and Home Mortgage Disclosure  
Act (HMDA) data and interviewed officials from FHA and other	 
mortgage institutions.						 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-07-645 					        
    ACCNO:   A71690						        
  TITLE:     Federal Housing Administration: Decline in the Agency's  
Market Share Was Associated with Product and Process Developments
of Other Mortgage Market Participants				 
     DATE:   06/29/2007 
  SUBJECT:   Comparative analysis				 
	     Cost sharing (finance)				 
	     Foreclosures					 
	     Housing						 
	     Housing programs					 
	     Income statistics					 
	     Interest rates					 
	     Lending institutions				 
	     Minorities 					 
	     Mortgage interest rates				 
	     Mortgage loans					 
	     Mortgage programs					 
	     Credit ratings					 

******************************************************************
** 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-07-645

   

     * [1]Results in Brief
     * [2]Background
     * [3]FHA's Market Share Declined from 1996 through 2005, While th

          * [4]FHA's Market Share Decreased While Conventional Market Share
          * [5]Especially among Black, Hispanic, and Lower-Income Borrowers

               * [6]Racial Submarkets
               * [7]Income Submarkets
               * [8]FHA-Eligible Submarket
               * [9]Census Tracts Groupings Characterized by Population
                 Characte

     * [10]The Decline in FHA's Market Share Was Associated with Severa

          * [11]FHA's Process Inefficiencies and Product Restrictions Have B
          * [12]Developments in Conventional Market Products and Processes O
          * [13]As FHA Lost Market Share, Many Subprime Borrowers Obtained L

     * [14]Observations
     * [15]Agency Comments and Our Evaluation
     * [16]Appendix I: Objectives, Scope, and Methodology

          * [17]Analysis of Market Share Trends

               * [18]Analysis of Factors Associated with the Trends in FHA's
                 Mark
               * [19]Analysis of Borrower and Loan Characteristics

     * [20]Appendix II: Data on Market Share Trends in the Mortgage Mar
     * [21]Appendix III: Data on Selected Borrower and Loan Characteris
     * [22]Appendix IV: Comments from the Department of Housing and Urb
     * [23]Appendix V: GAO Contact and Staff Acknowledgments

          * [24]GAO Contact
          * [25]Acknowledgments

               * [26]Order by Mail or Phone

Report to Congressional Requesters

United States Government Accountability Office

GAO

June 2007

FEDERAL HOUSING ADMINISTRATION

Decline in the Agency's Market Share Was Associated with Product and
Process Developments of Other Mortgage Market Participants

GAO-07-645

Contents

Letter 1

Results in Brief 4
Background 6
FHA's Market Share Declined from 1996 through 2005, While the Conventional
Market Share Increased, Especially among Minority and Lower-Income
Borrowers 8
The Decline in FHA's Market Share Was Associated with Several Factors and
Has Been Accompanied by Higher Costs for Certain Conventional Borrowers
and Increased Credit Risk for FHA 19
Observations 28
Agency Comments and Our Evaluation 29
Appendix I Objectives, Scope, and Methodology 30
Appendix II Data on Market Share Trends in the Mortgage Market and
Selected Submarkets from 1996 through 2005 35
Appendix III Data on Selected Borrower and Loan Characteristics for FHA,
Prime, and Subprime Loans, 1996 through 2005 42
Appendix IV Comments from the Department of Housing and Urban Development
46
Appendix V GAO Contact and Staff Acknowledgments 48

Tables

Table 1: Market Shares for Home Purchase Loans, 1996-2005 35
Table 2: State-by-State FHA Market Shares and Loan Counts for Home
Purchase Loans, 1996-2005 36
Table 3: Market Shares for Home Purchase Loans in Selected Submarkets,
1996-2005 39
Table 4: Market Shares for Home Purchase and Refinance Loans, 1996-2005 41
Table 5: Market Shares for Refinance Loans, 1996-2005 41
Table 6: Selected FHA Borrower and Loan Characteristics, 1996-2005 42
Table 7: Selected Prime Borrower and Loan Characteristics, 1996-2005 43
Table 8: Selected Subprime Borrower and Loan Characteristics, 1996-2005 44

Figures

Figure 1: FHA and Other Market Participants' Shares of the Home Purchase
Mortgage Market, 1996-2005 10
Figure 2: FHA, Prime, Subprime, and GSE Shares of the Minority Submarket
for Home Purchase Mortgages, 1996-2005 12
Figure 3: FHA, Prime, Subprime, and GSE Shares of the Lower-Income
Submarket for Home Purchase Mortgages, 1996-2005 13
Figure 4: Changes in Market Shares for FHA-Insured, Prime, and Subprime
Loans in Census Tracts with Different Race and Income Characteristics,
1996 through 2005 16
Figure 5: FHA's Market Share in Census Tract Groupings with Different
Median Credit Scores, 1996-2005 18

Abbreviations

APR Annual Percentage Rate

ARM adjustable rate mortgage
FFIEC Federal Financial Institutions Examination Council
FHA Federal Housing Administration
GSE government-sponsored enterprises
HMDA Home Mortgage Disclosure Act
HUD Department of Housing and Urban Development
MBA Mortgage Bankers Association
OFHEO Office of Federal Housing Enterprise Oversight
RHS Rural Housing Service
SFDW Single-Family Data Warehouse
TOTAL Technology Open to Approved Lenders
VA Veterans Administration

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.

United States Government Accountability Office
Washington, DC 20548

June 29, 2007

The Honorable Richard Shelby
Ranking Member, Committee on Banking, Housing, and Urban Affairs
United States Senate 

The Honorable Wayne Allard
United States Senate 

Through its single-family mortgage insurance programs, the Department of
Housing and Urban Development's (HUD) Federal Housing Administration (FHA)
insures private lenders against losses from defaults on mortgages that
meet FHA criteria. FHA historically has been an important participant in
the market for home purchase mortgages, which grew from 3.1 million loans
in 1996 to 4.7 million loans in 2005.^1 FHA in the past has played a
particularly large role among minority, lower-income, and first-time
homebuyers and generally is thought to promote stability in the market by
helping to ensure the availability of mortgage credit in areas that may be
underserved by the private sector during economic downturns. However,
demand for FHA-insured mortgages has dropped sharply in recent years,
raising questions about the agency's role in and ability to adapt to a
changing mortgage market. To help FHA adapt to market changes, in 2006 HUD
submitted a legislative proposal to Congress that, among other things,
would raise FHA's loan limits, give the agency flexibility to set
insurance premiums based on the credit risk of borrowers, and reduce
down-payment requirements from the current 3 percent to potentially zero.
Our analysis of the major elements of this proposal are contained in a
companion report that we are issuing today.^212Through its single-family
mortgage insurance programs, the Department of Housing and Urban
Development's (HUD) Federal Housing Administration (FHA) insures private
lenders against losses from defaults on mortgages that meet FHA criteria.
FHA historically has been an important participant in the market for home
purchase mortgages, which grew from 3.1 million loans in 1996 to 4.7
million loans in 2005. FHA in the past has played a particularly large
role among minority, lower-income, and first-time homebuyers and generally
is thought to promote stability in the market by helping to ensure the
availability of mortgage credit in areas that may be underserved by the
private sector during economic downturns. However, demand for FHA-insured
mortgages has dropped sharply in recent years, raising questions about the
agency's role in and ability to adapt to a changing mortgage market. To
help FHA adapt to market changes, in 2006 HUD submitted a legislative
proposal to Congress that, among other things, would raise FHA's loan
limits, give the agency flexibility to set insurance premiums based on the
credit risk of borrowers, and reduce down-payment requirements from the
current 3 percent to potentially zero. Our analysis of the major elements
of this proposal are contained in a companion report that we are issuing
today.

The different parts of the mortgage market are defined by the types of
mortgage institutions that serve them and the credit quality of the
borrowers. The conventional market, comprising mortgages that do not carry
government insurance or guarantees, has prime and subprime The different
parts of the mortgage market are defined by the types of mortgage
institutions that serve them and the credit quality of the borrowers. The
conventional market, comprising mortgages that do not carry government
insurance or guarantees, has prime and subprime segments.^3 Prime
borrowers typically have strong credit scores and obtain the most
competitive interest rates and mortgage terms.^4 In contrast, subprime
borrowers typically have blemished credit and lower credit scores, may
have difficulty providing income documentation, and generally pay higher
interest rates and fees than prime borrowers. Mortgages purchased by two
government-sponsored enterprises (GSE), Fannie Mae and Freddie Mac,
comprise another, albeit overlapping, market segment. The GSEs purchase
(primarily conventional prime) loans and pool them to create securities
sold to investors. The GSEs have goals directed at financing housing for
lower-income families and in underserved areas. FHA is a major part of the
market segment comprising loans with government insurance or guarantees,
which primarily serves borrowers who would have difficulty obtaining
conventional prime mortgages.^5 The borrowers are allowed to make very low
down payments and generally pay interest rates that are competitive with
prime mortgages but also pay fees or premiums to cover the cost of the
guaranty.

^1Home purchase mortgages do not include mortgages for refinancing
existing loans.

^2See GAO, Federal Housing Administration: Modernization Proposals Would
Have Program and Budget Implications and Require Continued Improvements in
Risk Management, [27]GAO-07-708 (Washington, D.C.: June 29, 2007).

To provide insights into FHA's role in the mortgage market, this report
discusses (1) trends in FHA's share of the market for home purchase
mortgages and selected submarkets from 1996 through 2005, and how they
compared with the trends for the prime, subprime, and GSE market segments;
and (2) the major factors associated with the trends in FHA's market share
and the implications of these trends for homebuyers and FHA.^6 In
addition, appendix III of this report provides information on selected
borrower and loan characteristics of FHA mortgages and mortgages in the
prime and subprime market segments.

^3There is no uniform definition across the lending industry for what
characterizes a loan as subprime. Subprime loans are generally given to
borrowers with credit scores that are below a certain threshold, but that
threshold can vary according to the policies of the individual lender.

^4Credit scores, which assign a numeric value to a borrower's credit
history, have become a common tool for assessing loan applications.

^5The insurance or guarantees protect lenders against losses from loan
defaults. The Department of Veterans Affairs and the Department of
Agriculture administer the other two federal programs that guarantee
single-family mortgages. Conventional loans with low down payments also
may require mortgage insurance, which borrowers purchase from private
companies.

^6We calculated market shares in terms of numbers of loans. We use the
term "submarkets" to mean subsets of the home purchase mortgage market
defined by various borrower, loan, and census tract characteristics.

To analyze trends in the overall market for home purchase mortgages, we
compiled and analyzed loan data for 1996 through 2005 collected under the
Home Mortgage Disclosure Act (HMDA). HMDA requires lending institutions to
collect and publicly disclose information about housing loans and
applications for such loans. HMDA data capture about 80 percent of the
mortgage loans funded each year, according to estimates by the Board of
Governors of the Federal Reserve System (Federal Reserve), and are one of
the most comprehensive sources of information on mortgage lending. HMDA
data have a number of limitations that affected our analysis. More
specifically, the data (1) understate the number of loans purchased by the
GSEs, (2) do not include a precise indicator for subprime loans, and (3)
do not distinguish first mortgages from "piggyback" loans (i.e., the
junior lien in a pair of loans used to finance the same property) for most
of the period we examined. While we acknowledge these limitations, we used
HMDA data to evaluate long-term market share trends rather than to provide
precise annual figures for each market segment, including the GSE segment.
According to Freddie Mac, Fannie Mae, and Federal Reserve officials, our
use of HMDA data was appropriate for this purpose. We identified subprime
loans by merging the data with a HUD-maintained list of lenders that
specialize in subprime lending. We identified piggyback loans using a data
matching process based on an algorithm developed by the Federal Reserve
and excluded these loans from our analysis. To analyze trends in various
submarkets, we incorporated additional data from FHA, the Census Bureau,
the Office of Federal Housing Enterprise Oversight, and TransUnion.^7 To
determine the factors associated with the trends in FHA's market share and
the potential implications of these trends, we analyzed HMDA data,
information from HUD's Single-Family Data Warehouse (SFDW), the Mortgage
Bankers Association's (MBA) National Delinquency Survey, and summary
statistics provided by FHA and contained in prior studies from databases
maintained by LoanPerformance.^8 We also reviewed literature, analyzed
agency documents, and interviewed FHA officials, mortgage industry
participants, and academic researchers. To determine borrower and loan
characteristics for different market segments, we analyzed HMDA data,
information from SFDW and the Federal Housing Finance Board, and summary
LoanPerformance data. Appendix I contains additional information on our
scope and methodology. We conducted our work in Washington, D.C., from
September 2006 through May 2007 in accordance with generally accepted
government auditing standards.

^7TransUnion is one of the three main consumer credit reporting agencies.

^8LoanPerformance is a private firm that maintains databases containing
detailed information submitted by participating lenders and third parties
(e.g., securities issuers and dealers) on millions of mortgages. SFDW
contains detailed information on the borrower and loan characteristics of
the mortgages FHA insures.

Results in Brief

FHA's share of the market for home purchase mortgages declined
substantially from 1996 through 2005, most significantly among minority
borrowers who accounted for a growing share of subprime loans in that
period. More specifically, FHA's market share in terms of numbers of loans
fell from 19 percent in 1996 to 6 percent in 2005, with almost all of the
decline occurring since 2001. In contrast,

           o the market share for prime loans was relatively stable over the
           10-year period, growing from 73 to 76 percent;
           o the market share for subprime loans grew nearly every year,
           rising from 2 percent to 15 percent overall, with particularly
           large increases since 2001; and
           o the market share of the housing GSEs--essentially a subset of
           the prime market--rose 3 percentage points overall (to roughly 30
           percent in 2005), with nearly all of the growth occurring before
           2002.

           During the 10-year period, the same general pattern of declining
           market share for FHA and increasing market share for conventional
           loans held true in submarkets where FHA traditionally has played a
           major role. For example, among minorities, FHA's market share fell
           25 percentage points (from 32 to 7 percent), while conventional
           prime and subprime shares rose 6 and 24 percentage points,
           respectively. The drop in FHA's market share was particularly
           large--35 percentage points--among Hispanic borrowers. Among
           lower-income (i.e., low- and moderate-income) borrowers, FHA's
           market share fell 16 percentage points (from 26 to 10 percent),
           while the prime and subprime shares grew 7 and 14 percentage
           points, respectively.^9 Consistent with these trends, in
           geographic areas with higher proportions of minority and
           lower-income borrowers, FHA lost substantial market share while
           subprime lending grew dramatically. The same pattern was also
           evident in areas with relatively low median credit scores and
           where median home prices rose to at least 75 percent of FHA's loan
           limit during the 10-year period.
			  
^9We defined lower-income borrowers as those with incomes less than 120
percent of the area median income.			  

           The decline in FHA's market share was associated with several
           factors and has been accompanied by higher ultimate costs for
           certain conventional borrowers and a worsening in indicators of
           credit risk among FHA borrowers. More specifically, (1) FHA's
           product restrictions and lack of process improvements relative to
           the conventional market and (2) product innovations and expanded
           loan origination and funding channels in the conventional
           market--coupled with interest rate and house price
           changes--provided conditions that favored conventional mortgages
           over FHA products. For example, mortgage industry officials with
           whom we spoke cited FHA's administrative requirements and loan
           limits as factors that limited the attractiveness of FHA-insured
           mortgages. Additionally, historically low interest rates and
           rising house prices increased demand for loan products offered by
           the conventional market (especially subprime lenders), which
           featured flexible payment and interest options that allowed
           borrowers to qualify for mortgages despite the appreciations in
           home values. Most subprime borrowers opted for adjustable rate
           products (in 2005, more than 75 percent of subprime loans were
           adjustable rate), having been attracted by their "affordability"
           features, such as lower initial payments and interest rates. In
           contrast to FHA-insured loans, the majority of subprime loans had
           higher ultimate costs, in part because their initial interest
           rates could increase 3 percentage points in as little as 2 years
           and two-thirds featured prepayment penalties, which can deter
           borrowers from refinancing into lower-cost products. Subprime
           loans also have experienced relatively high rates of default and
           foreclosure, adding to concerns about their long-term cost to
           borrowers. Certain factors associated with the decline in FHA's
           market share also have negatively affected the financial
           performance of FHA's insurance program. More specifically, as
           conventional lenders expanded their presence in traditional FHA
           submarkets through the development of new products and use of
           automated underwriting tools, FHA experienced adverse
           selection--that is, conventional providers identified and approved
           relatively lower-risk borrowers, leaving relatively higher-risk
           borrowers for FHA.

           While our report does not make recommendations, we make
           observations about how developments in the different segments of
           the mortgage market could affect FHA's market share in the future.
           The relatively poor performance of subprime mortgages in recent
           months and a contraction of this market segment could shift market
           share to FHA. But the size of this shift depends partly on the
           efforts of conventional mortgage providers, including the GSEs, to
           offer viable alternatives to subprime borrowers. Notwithstanding
           the actions of conventional providers, FHA could be a vehicle to
           provide lower-priced and more sustainable mortgage options for
           some borrowers who are considering or struggling to maintain
           higher-priced subprime loans. However, careful assessment and
           management of the risks associated with serving these borrowers
           would be necessary to avoid exacerbating problems in the financial
           performance of FHA's insurance program.

           We provided HUD with a draft of this report. HUD commented that we
           produced a straightforward, well-researched report on the reasons
           for the recent decline in FHA's market share. HUD also noted that
           additional product and pricing flexibility would help FHA to
           continue serving lower-income and minority households. We discuss
           HUD's comments in the agency comments section, and reproduced its
           written comments in appendix IV.
			  
			  Background

           Congress established FHA in 1934 under the National Housing Act
           (P.L. 73-479) to broaden homeownership, protect and sustain
           lending institutions, and stimulate employment in the building
           industry. Over time, FHA came to play a major role in extending
           mortgage credit to first-time homebuyers and historically
           underserved borrowers such as minority and lower-income families.
           For example, in 2005, slightly less than 80 percent of FHA
           borrowers were first-time homebuyers, more than 80 percent had
           lower incomes, and approximately 30 percent were minorities. (See
           app. III for additional information on the borrower and loan
           characteristics of FHA-insured, prime, and subprime mortgages.)
           FHA currently insures a variety of mortgages for home purchases,
           construction and rehabilitation, and refinancing, with its most
           popular program--Section 203(b)--offering 15- and 30-year
           mortgages for single-family dwellings.

           Generally, borrowers are required to purchase mortgage insurance
           when the loan-to-value (LTV) ratio (the amount of the mortgage
           loan divided by the value of the home) exceeds 80 percent. FHA is
           a government mortgage insurer in a market that also includes
           private insurers. Private mortgage insurance policies provide
           lenders coverage on a portion (generally 20 to 30 percent) of the
           mortgage balance. However, borrowers who have difficulty meeting
           down-payment and credit score requirements for conventional loans
           may find it easier to qualify for a loan with FHA insurance, which
           covers 100 percent of the value of the loan. FHA-insured borrowers
           are required to make minimum cash investments of 3 percent, which
           may come from the borrowers' own funds or from certain third-party
           sources. Borrowers are permitted to finance their mortgage
           insurance premiums and some closing costs, which can create an
           effective LTV ratio of close to 100 percent for some FHA-insured
           loans. In fiscal year 2006, the agency insured almost 426,000
           mortgages, representing about $55 billion in mortgage insurance.

           Congress has set limits on the size of the loans that may be
           insured by FHA. The limit for an FHA-insured mortgage is 95
           percent of the local median home price, not to exceed 87 percent
           or fall below 48 percent of the Freddie Mac conforming loan limit,
           which was $417,000 in 2006. Therefore, in 2006, FHA loan limits
           fell between a floor in low-cost areas of $200,160 and a ceiling
           in high-cost areas of $362,790. Eighty-two percent of counties
           nationwide had loan limits set at the low-cost floor, while 3
           percent had limits set at the high-cost ceiling. The remaining 15
           percent of counties had limits set between the floor and ceiling,
           based on local median house prices.

           FHA determines the expected cost of its insurance program, known
           as the credit subsidy cost, by estimating the program's future
           performance.^10 FHA's mortgage insurance program is currently a
           negative subsidy program, meaning that the present value of
           estimated cash inflows to FHA's Mutual Mortgage Insurance Fund
           (Fund) exceeds the present value of estimated cash outflows. The
           economic value, or net worth, of the Fund depends on the relative
           size of cash outflows and inflows over time. Cash flows out of the
           Fund for payments associated with claims on defaulted loans and
           refunds of up-front premiums on prepaid mortgages. To cover these
           outflows, FHA receives cash inflows from borrowers' insurance
           premiums 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.

           A number of different private-sector and government institutions
           participate in the mortgage market. Along with FHA, the Department
           of Veterans Affairs' (VA) Loan Guaranty Service and the Department
           of Agriculture's Rural Housing Service (RHS) administer federal
           government programs that insure or guarantee single-family
           mortgages made by private lenders. Private lenders that loan
           borrowers funds for home purchase and refinance mortgages often
           work with mortgage brokers, independent contractors that originate
           the loan products of multiple lenders.^11 Fannie Mae and Freddie
           Mac are housing GSEs that purchase primarily prime conventional
           mortgages from lenders across the country, financing their
           purchases through borrowing or by issuing securities backed by the
           mortgages. Since 1994, HUD has set affordable housing goals for
           the housing GSEs and has adjusted the goals upward every few
           years. One goal is that at least 55 percent of the mortgages
           purchased by the GSEs must be made to families whose incomes are
           no greater than the area median income. The other two major goals
           concern the percentage of mortgages to borrowers residing in
           lower-income communities and certain high-minority neighborhoods
           (38 percent) and the percentage of borrowers with very low incomes
           and those with low incomes who live in low-income areas (25
           percent).^12
			  
^10Pursuant to the Federal Credit Reform Act of 1990, HUD must annually
estimate the credit subsidy costs for its loan insurance programs. Credit
subsidy costs are the net present value of estimated payments it makes
less the estimated amounts it receives, excluding administrative costs.

^11Mortgage origination involves such functions as accepting loan
applications and obtaining employment verifications and credit reports on
the borrowers. It is distinct from mortgage underwriting, which refers to
a risk analysis that uses information collected during the origination
process to decide whether to approve a loan.

^12For purposes of the GSE goals, lower-incomes neighborhoods are those
with a median income less than or equal to 90 percent of the area median
income and high-minority neighborhoods are those with at least a 30
percent minority population and a median income less than 120 percent of
the area median. Low- and very-low-income borrowers are defined as those
with incomes less than 80 percent and 60 percent of the area median
income, respectively.

           FHA's Market Share Declined from 1996 through 2005, While the
			  Conventional Market Share Increased, Especially among Minority
			  and Lower-Income Borrowers

           FHA's share of the market for home purchase mortgages in terms of
           numbers of loans declined 13 percentage points from 1996 through
           2005, while the prime share increased slightly and the subprime
           share grew 13 percentage points. Although the decline in FHA's
           market share was broad-based, FHA experienced particularly sharp
           decreases in submarkets where it traditionally has had a strong
           presence, such as among minority and lower-income borrowers.
           Consistent with these trends, in geographic areas with higher
           concentrations of these borrowers, FHA lost substantial market
           share while the subprime share grew dramatically. The same pattern
           held true in areas with relatively low median credit scores and
           where median home prices rose to at least 75 percent of FHA's loan
           limit during the 10-year period.
			  
			  FHA's Market Share Decreased While Conventional Market Share,
			  Particularly for Subprime Loans, Grew

           From 1996 through 2005, FHA's share of the home purchase mortgage
           market declined while the conventional share increased.^13 As
           shown in figure 1, FHA's market share fell from almost 19 percent
           (about 583,000 loans) in 1996 to about 6 percent (about 295,000
           loans) in 2005, with almost all of the decline occurring after
           2001.^14 Although FHA's market share has fluctuated over time,
           during the past two decades it has generally been over 10 percent.
			  
^13During the 10-year period, VA and RHS had small market shares. VA's
market share fell from 6 to 2 percent and RHS's market share remained at
or near 0.4 percent.

^14From 1996 through 2005, FHA's share of the overall mortgage market
(home purchase and refinance loans combined) declined from 12 to 4
percent, while the prime share fell from 78 to 77 percent and the subprime
share increased from 5 to 19 percent. Appendix II provides additional
information on FHA and other market segments' share of the home purchase,
refinance, and overall mortgage markets.

Figure 1: FHA and Other Market Participants' Shares of the Home Purchase
Mortgage Market, 1996-2005

Note: We calculated market shares based on numbers of loans and, to the
extent possible, excluded piggyback loans. Data for the GSEs do not
include loans originated and purchased in different years or all of the
loans sold to intermediaries before being purchased by the GSEs.

Over the 10-year period, the market share for conventional mortgages rose
from almost 75 percent (about 2.3 million loans in 1996) to about 91
percent (about 4.2 million loans in 2005), with much of the increase due
to growth in subprime lending. More specifically, prime market share
increased from 73 percent to 76 percent overall, falling somewhat from
1996 through 2000 but then increasing about 5 percentage points after
2000. Subprime market share increased substantially over the 10-year
period, from 2 percent to 15 percent, with most of the increase occurring
after 2001 (growing from 5 percent in 2001 to 15 percent in 2005). From
1996 through 2005, the market share for the GSEs (essentially a subset of
the conventional prime market) increased 3 percentage points overall (to
roughly 30 percent in 2005), growing about 13 percentage points from 1996
through 2002 but falling 9 percentage points thereafter.

Especially among Black, Hispanic, and Lower-Income Borrowers, FHA's Market Share
Declined Sharply and the Subprime Share Increased

From 1996 through 2005, FHA lost market share in certain key submarkets,
especially among minority and lower-income borrowers, as well as among
borrowers with mortgages within FHA's loan limits. At the same time, the
market share for conventional mortgages, particularly subprime loans, grew
in these submarkets. This trend also held true in census tracts with high
concentrations of low-income and minority households, relatively low
median credit scores, and median home prices within FHA's loan limits.
Mirroring the trend in the overall home purchase mortgage market, FHA's
loss of market share in these submarkets primarily occurred after 2001.
(See app. II for details on the trends in specific submarkets for each
market segment)

  Racial Submarkets

FHA traditionally has played a major role among minority borrowers.
However, over the 10-year period, FHA's share of this submarket fell
substantially. Specifically, as shown in figure 2, FHA's market share
dropped 25 percentage points (from 32 to 7 percent) among minority
borrowers, but declined most sharply among black and Hispanic borrowers
(by 27 and 35 percentage points, respectively). FHA's market share among
white borrowers decreased from 16 percent to 7 percent during the 10-year
period.

Figure 2: FHA, Prime, Subprime, and GSE Shares of the Minority Submarket
for Home Purchase Mortgages, 1996-2005

Note: We calculated market shares based on numbers of loans and, to the
extent possible, excluded piggyback loans. Data for the GSEs do not
include loans originated and purchased in different years or all of the
loans sold to intermediaries before being purchased by the GSEs.

In contrast to FHA, prime market share increased about 6 percentage points
(from 59 to 65 percent) among minority borrowers and 5 percentage points
(from almost 77 to just over 81 percent) among white borrowers from 1996
through 2005. Over the same period, subprime market share increased 24
percentage points (from 2 to 26 percent) among minorities, but especially
among black and Hispanic borrowers (29 percentage points for each group).
Subprime market share among white borrowers increased from 1 to 9 percent
from 1996 through 2005. GSE market share among minority borrowers
ultimately did not change substantially, beginning and ending the period
at roughly 20 percent. From 1996 through 2002, GSE market share among
minority borrowers increased 11 percentage points (to roughly 32 percent),
but fell by about the same amount thereafter. GSE market share among white
borrowers increased 7 percentage points over the 10-year period, to
roughly 35 percent in 2005.

  Income Submarkets

Lower-income (i.e., low- and moderate-income) borrowers historically have
relied heavily on FHA products, but FHA's market share dropped in this
submarket as well.^15 From 1996 through 2005, FHA's market share decreased
among borrowers of all income levels, but, as shown in figure 3,
particularly among lower-income borrowers, where FHA's share declined 16
percentage points (from 26 percent to 10 percent). From 1996 through 2005,
FHA's market share among upper-income borrowers fell from 9 to 2 percent.

Figure 3: FHA, Prime, Subprime, and GSE Shares of the Lower-Income
Submarket for Home Purchase Mortgages, 1996-2005

Note: We calculated market shares based on numbers of loans and, to the
extent possible, excluded piggyback loans. Data for the GSEs do not
include loans originated and purchased in different years or all of the
loans sold to intermediaries before being purchased by the GSEs.

^15We defined low income as less than 80 percent of the median income for
the census tract, moderate income as at least 80 percent but less than 120
percent, and upper income as 120 percent and above.

Over the 10-year period, prime market share increased from 65 to 72
percent among lower-income borrowers and remained consistently high (above
80 percent) among upper-income borrowers. At the same time, subprime
market share increased 14 percentage points (from 1 to 15 percent) among
lower-income borrowers and 12 percentage points (from 2 to 14 percent)
among upper-income borrowers. GSE market share increased 8 percentage
points among lower-income borrowers (to roughly 32 percent in 2005), and
remained at approximately 30 percent for upper-income borrowers.

  FHA-Eligible Submarket

As previously noted, Congress has set limits on the size of FHA-insured
loans. Although loans that fall within these limits comprise what has been
called the FHA-eligible submarket, from 1996 through 2005 FHA's share in
this submarket declined more sharply than in the overall home purchase
mortgage market. Specifically, it decreased 16 percentage points (from 25
to 9 percent), with a steep decline occurring after 2001 when its market
share was 24 percent. FHA's market share among minority borrowers in this
submarket also fell dramatically (from 39 percent in 1996 to 10 percent in
2005), as did its share of loans to lower-income borrowers (from 28
percent in 1996 to 11 percent in 2005).

While FHA's market share declined in the FHA-eligible submarket, the prime
and subprime market shares grew. Overall, the prime share in this
submarket increased modestly (from 67 percent to 73 percent) from 1996 to
2005. In contrast, the subprime share increased 14 percentage points (from
1 percent to 15 percent). The GSE share in this submarket increased 15
percentage points from 1996 to 2002 (to roughly 40 percent) but fell to
about 33 percent as of 2005.

  Census Tracts Groupings Characterized by Population Characteristics and House
  Prices

To further analyze mortgage market trends, we examined FHA, prime, and
subprime market shares in various census tract groupings.^16 Specifically,
we looked at census tracts grouped based on (1) race and income
characteristics, (2) median credit score, and (3) median home price in
relation to FHA loan limits. For the credit score analysis, we limited our
analysis to census tracts where FHA's market share averaged at least 5
percent from 1996 through 1998 (representing about 75 percent of the
census tracts nationwide).^17

16Census tracts are small, relatively permanent statistical subdivisions
of a county. They usually have between 2,500 and 8,000 persons and, when
first delineated, are designed to be homogeneous with respect to
population characteristics, economic status, and living conditions. The
spatial size of census tracts varies widely depending on the density of
settlement.

From 1996 through 2005, FHA lost market share in approximately 90 percent
of the census tracts we included in our analysis. As shown in figure 4,
the losses occurred primarily in census tracts with both medium to high
concentrations of minorities and low to moderate median incomes.^18 At the
same time, the market share for conventional loans increased for this
group of census tracts, especially for subprime loans. This was
particularly evident in census tracts with both the highest concentrations
of minorities and low median incomes, where FHA's market share fell 31
percentage points and subprime market share increased 28 percentage
points.

^17We took this approach because our analysis examined changes in FHA's
market share during a 10-year period when the trend in FHA's share was
downward. By using the 5 percent threshold, our analysis excluded census
tracts where FHA's market share started and ended the period at zero and
census tracts where FHA's market share was sporadic and on average very
small near the beginning of the period.

^18We defined low-, medium-, and high-minority census tracts as those with
minority populations of less than 20 percent, 20 to 49 percent, and 50
percent or more, respectively. We defined low-, moderate-, and
upper-income census tracts as those with median incomes that were less
than 80 percent, at least 80 percent but less than 120 percent, and 120
percent and above, respectively, of the median income for the associated
metropolitan statistical area.

Figure 4: Changes in Market Shares for FHA-Insured, Prime, and Subprime
Loans in Census Tracts with Different Race and Income Characteristics,
1996 through 2005

Note: We calculated market shares based on numbers of loans and, to the
extent possible, excluded piggyback loans.

From 1996 through 2005, FHA's market share declined in all of the census
tract groupings we defined based on the median credit score of mortgage
borrowers. However, FHA's share declined most sharply in census tracts
with relatively low median credit scores.^19 Specifically, FHA's market
share dropped 26 percentage points (from about 37 percent in 1996 to 11
percent in 2005) in census tracts with median credit scores in the bottom
quarter of the credit score distribution for all census tracts included in
our analysis, with the steepest decline occurring after 2001 (see fig. 5).
At the same time, the prime and subprime market shares in these census
tracts increased. More specifically, the prime share increased 12
percentage points (from about 51 percent in 1996 to 63 percent in 2005),
and the subprime share increased 21 percentage points (from about 2
percent in 1996 to 23 percent in 2005). In census tracts with median
credit scores between the bottom quarter and the top quarter of the
distribution, FHA's market share fell 16 percentage points over the
10-year period, while the prime and subprime shares rose 10 and 12
percentage points, respectively. Finally, in census tracts with median
credit scores in the top quarter of the distribution, FHA's market share
decreased 12 percentage points, while the prime and subprime shares
increased 9 and 7 percentage points, respectively.

^19We obtained credit score data from TransUnion's TrenData database.
TransUnion depersonalized and aggregated the data from consumer credit
reports.

Figure 5: FHA's Market Share in Census Tract Groupings with Different
Median Credit Scores, 1996-2005

Note: We limited our analysis to census tracts where FHA's market share
averaged at least 5 percent from 1996 through 1998. We calculated market
shares based on numbers of loans and, to the extent possible, excluded
piggyback loans. The credit score information is from TransUnion and
represents the median credit score of mortgage borrowers in each census
tract as of December 31, 2004.

From 1996 through 2005, FHA lost market share in census tracts where
median home prices were above and below FHA's loan limits (which FHA
adjusts annually within statutory caps), but experienced the greatest
losses in census tracts where median home prices appreciated to at least
75 percent of FHA's loan limit during the 10-year period. In census tracts
where the median home price was less than 75 percent of the FHA loan limit
each year, FHA's market share dropped 24 percentage points (from 36
percent in 1996 to 12 percent in 2005). In census tracts where the median
home price was at least 75 percent but less than 125 percent of the FHA
loan limit each year, FHA's market share fell more modestly--13 percentage
points (from 18 percent in 1996 to 5 percent in 2005). However, FHA lost
almost all of its market share in areas where the median home price was
less than 75 percent of the FHA loan limit early in the 10-year period but
grew to at least 75 percent of the loan limit later in the period.
Specifically, FHA's market share in these areas fell 38 percentage
points--from 41 percent in 1996 to 3 percent in 2005.

At the same time, conventional market share grew in the census tract
groupings we examined. For example, in census tracts where median home
prices were less than 75 percent of FHA's loan limit each year, the prime
share increased 11 percentage points (from 53 percent in 1996 to 64
percent in 2005) and the subprime share increased 18 percentage points
(from 2 percent in 1996 to 20 percent in 2005). The prime and subprime
market shares increased even more in census tracts where median home
prices rose from less than 75 percent of the FHA loan limit to at least 75
percent of the limit during the 10-year period. Specifically, prime and
subprime shares increased 22 and 25 percentage points, respectively.

The Decline in FHA's Market Share Was Associated with Several Factors and Has
Been Accompanied by Higher Costs for Certain Conventional Borrowers and
Increased Credit Risk for FHA

During the period from 1996 through 2005, a combination of factors created
conditions that favored conventional mortgages over FHA products. These
factors included (1) FHA's product restrictions and lack of process
improvements relative to the conventional market and (2) product
innovations and expanded loan origination and funding channels in the
conventional market. Most subprime mortgages, which grew in popularity as
FHA's market share declined, had higher ultimate costs, in part because
their initial interest rates could reset to higher rates. In addition,
subprime mortgages performed worse than prime and FHA-insured loans. As
FHA's market share fell, certain factors associated with this decline also
contributed to a worsening in indicators of credit risk among FHA
borrowers.

FHA's Process Inefficiencies and Product Restrictions Have Been Linked to the
Decline in Use of FHA-Insured Mortgages

Lenders, mortgage industry groups, and consumer advocacy groups have cited
FHA administrative requirements as a factor that contributed to the
decline in the agency's market share over the 10-year period we examined.
According to mortgage industry officials we interviewed, processing
FHA-insured loans was more time consuming, labor intensive, and costly
than processing conventional mortgages. For example, instead of having
lenders submit all loan information electronically, FHA required lenders
to send loan case files to FHA for review before the loans could be
approved for insurance. If the review found a problem with the case file,
FHA would mail the file back to the lender, who in turn would make the
needed corrections and mail the file back to FHA. Additionally, in
contrast to conventional market requirements, FHA's appraisal process
required that minor property repairs, such as cracked window panes, be
corrected prior to loan closing. According to the MBA, some FHA lenders
have reported substantially higher processing times and origination costs
for FHA-insured loans than for conventional loans. In contrast with FHA,
conventional loan processing became increasingly streamlined and less
costly through the use of information technology and the Internet.
According to mortgage industry officials with whom we spoke, FHA's more
cumbersome processes made FHA's products less attractive than conventional
products, particularly in competitive housing markets where it is
important to be able to close on a home quickly.

However, in 2006, FHA made several administrative changes, such as
allowing higher-performing lenders to approve FHA insurance without a
prior review by FHA and simplifying its appraisal process. FHA and
mortgage industry officials with whom we spoke said that these changes
have increased the efficiency of loan and insurance processing, making FHA
products more attractive and, therefore, more likely to be used.^20

FHA and mortgage industry officials with whom we spoke also cited FHA's
loan limits as a factor that contributed to the decline in FHA's market
share. In some areas of the country, particularly in parts of California
and the Northeast, median home prices have been substantially higher than
FHA's maximum loan limits, reducing the agency's ability to serve
borrowers in those markets. For example, the 2005 loan limit in high-cost
areas was $312,895 for one-unit properties, while the median home price
was about $399,000 in Boston, Massachusetts; about $432,000 in Newark, New
Jersey; $500,000 in Salinas, California; and about $646,000 in San
Francisco, California.

Some mortgage industry officials also pointed to other product
restrictions as a reason why FHA loans have been less competitive than
conventional loans. Many borrowers either cannot or do not want to make a
down payment, and in recent years members of the conventional mortgage
market (such as private mortgage insurers, the housing GSEs, and large
private lenders) have been increasingly active in supporting low and
no-down-payment mortgages. For example, the GSEs introduced
no-down-payment mortgage products in 2000. In contrast, FHA does not offer
a zero-down-payment product, which some lenders and industry observers
have cited as a major factor underlying the decline in FHA's market share.
(However, as previously noted, FHA allows borrowers to finance their
up-front insurance premium and some closing costs; as a result, an
FHA-insured loan could equal nearly 100 percent of the property's value or
sales price.)

^20For additional information about changes to FHA's administrative
processes, see [28]GAO-07-708 .

Developments in Conventional Market Products and Processes Occurred as FHA's
Market Share Declined

During the 10-year period we examined, several developments associated
with FHA's declining market share occurred in the conventional market.
First, the conventional market offered products that increased consumer
choices for borrowers, including those who may have previously chosen an
FHA-insured loan. These products, in combination with historically low
interest rates, made it easier for homebuyers to purchase homes in a
period of strong house price appreciation. For example, to serve the
lower-income and minority populations targeted by their affordable housing
goals, the GSEs developed products featuring underwriting criteria that
allowed for higher risks, such as Freddie Mac's Home Possible(R) Mortgage,
which allows qualified borrowers to make no down payment. As the GSEs
worked to meet their goals, their market share among lower-income and
minority borrowers grew over much of the 10-year period we examined, while
FHA's fell. More specifically, the GSE market share among lower-income
borrowers grew nearly 14 percentage points from 1996 through 2002, while
FHA's share dropped 3 percentage points over that period. During the same
time frame, GSE market share among minority borrowers grew 11 percentage
points, while FHA's fell 8 percentage points. Consistent with these
observations, research by An and Bostic (2006) found a significant
negative relationship between the change in the GSE and FHA shares of the
overall mortgage market from 1996 through 2000.^21 However, as previously
noted, the market shares for both FHA and the GSEs ultimately declined
after 2002.

Other products offered by conventional mortgage providers--interest-only
loans, no- and low-documentation mortgages, piggyback loans, and hybrid
adjustable rate mortgages (ARM)--also became popular, especially during
the subprime market's rapid growth after 2001, because they featured
flexible payment and interest options that increased initial
affordability.^22 For example, borrowers were attracted to hybrid ARMs
because they could qualify on the basis of an interest rate at or near the
initial rate rather than the higher reset rate. These nontraditional
products came to represent a sizeable part of the subprime market after
2001. For example, according to data reported by an investment bank, from
the first quarter of 2002 to the third quarter of 2005, the percentage of
subprime mortgages that were interest-only loans increased from zero to 29
percent and the percentage that were no- and low-documentation loans
increased from 30 to 41 percent. Over the same period, the proportion of
subprime mortgages with piggyback loans, which are often used to avoid the
need for mortgage insurance, increased from 2 to 33 percent.^23
Additionally, from 2002 through 2005, the percentage of subprime mortgages
that were ARMs grew from 68 to 73 percent, with hybrid ARMs accounting for
the majority of these loans.^24 In contrast, FHA (which, as previously
discussed, lost substantial market share in submarkets where subprime
lending grew dramatically) does not offer interest-only or no- and
low-documentation products and did not begin insuring hybrid ARMs until
2004.

^21X. An and R. Bostic,"GSE Activity, FHA Feedback and Implications of the
Efficacy of the Affordable Housing Goals," forthcoming in Journal of Real
Estate Finance and Economics.

A second development in the conventional market was advances in
underwriting technology that allowed conventional mortgage providers to
process loan applications more quickly and consistently than in the past
and broaden their customer base. For example, to help assess the default
risk of borrowers, the mortgage industry increasingly used mortgage
scoring and automated underwriting systems. Mortgage scoring is a
technology-based tool that relies on the statistical analysis of millions
of previously originated mortgage loans to determine how key attributes
such as the borrower's credit history, property characteristics, and terms
of the mortgage affect future loan performance. FHA implemented its own
mortgage scoring tool, called the Technology Open to Approved Lenders
(TOTAL) scorecard, in 2004. However, in prior work we found that the way
FHA developed TOTAL may limit the scorecard's effectiveness.^25 To the
extent that conventional mortgage providers were better able than FHA to
use scoring tools, lower-risk borrowers in FHA's traditional market
segment may have migrated toward conventional products, contributing to
the decline in FHA's market share.

^22Interest-only loans allow borrowers to defer the principal payments for
some period and hybrid ARMs allow borrowers to pay a lower interest rate
for a specified time, usually between 2 and 5 years, before the loan
resets to the fully indexed interest rate (i.e., a rate that is comprised
of an adjustable rate index plus the lender's margin). Piggyback loans are
simultaneous second mortgages that allow borrowers to make little or no
down payment. No- and low-documentation loans allow for less detailed
proof of income or assets than lenders traditionally require.

^23UBS Mortgage Strategist, 2005--Good or Bad for Vintage Subprime? (Jan.
31, 2006), 33. UBS analyzed data from LoanPerformance's TrueStandings
Securities subprime database.

^24Figures are from FHA-provided summaries of information from
LoanPerformance's TrueStandings subprime database.

A third development was an increase in mortgage originations through third
parties such as loan correspondents and mortgage brokers, particularly in
the subprime market.^26 This trend has been associated with the decline in
FHA's market share because these mortgage originators primarily market
non-FHA products. According to data reported by the trade publication
Inside B&C Lending, loan correspondents and mortgage brokers increased
their share of subprime loan originations from 66 percent in 2003 to 81
percent in 2005. In contrast, just 27 percent of FHA-insured mortgages in
2005 were originated by loan correspondents and mortgage brokers.
According to the National Association of Mortgage Brokers, many mortgage
brokers do not offer FHA products because they find the financial and
audit requirements for participation in FHA programs cost-prohibitive.^27

A fourth development in the conventional market was the growth in private
mortgage securitization (the bundling of mortgage loans into bond-like
securities that can be bought and sold on the secondary market),
particularly for subprime loans. Securitization allowed lenders to sell
loans from their portfolios, transferring credit risk to investors, and
use the proceeds to make more loans. According to recent testimony by a
senior official from the Federal Deposit Insurance Corporation, many
lenders would not have found subprime mortgages attractive absent the
funding and credit-risk transfer features available through
securitization.^28 At the same time, these securities were attractive to
different types of investors. The combination of higher interest rates and
higher risks for subprime loans facilitated the division of mortgage
securities into risk tranches, which offer investors different risk and
reward options. According to data reported by Inside B&C Lending, from
1999 through 2005, subprime securitization rates--that is, the dollar
amount of securitized loans divided by the dollar amount of loan
originations--rose from less than 40 percent to about 80 percent. In
addition, the dollar volume of subprime loan securitizations increased
from $61 billion in 1999 to nearly $508 billion in 2005.

^25For additional information on FHA's scorecard, see GAO, Mortgage
Financing: HUD Could Realize Additional Benefits from its Mortgage
Scorecard, [29]GAO-06-435 (Washington, D.C.: Apr. 13, 2006).

^26The term loan correspondent originally referred to lenders that
originated, underwrote, and closed loans in their names (usually funding
them with short-term lines of credit from banks) and then immediately sold
the loans to other lenders. Today, the term is sometimes used synonymously
with mortgage broker. Mortgage brokers originate loans for other lenders
but seldom underwrite or close loans in their own names.

^27FHA requires each of its loan correspondent firms (which include
mortgage brokers) to have an annual audited financial statement and retain
a minimum $63,000 net worth.

As FHA Lost Market Share, Many Subprime Borrowers Obtained Loans with High
Ultimate Costs and Credit Characteristics among FHA Borrowers Worsened

As a result of developments in the conventional market, including lower
interest rates, more homebuyers--especially minority and lower-income
families--were able to obtain conventional loans, but many of these loans
had high ultimate costs. As previously discussed, much of the increase in
mortgages to minorities and lower-income borrowers was due to the growth
in subprime lending, and many of these loans offered lower initial costs
through their interest-only features and low introductory interest rates.
However, these mortgages became more costly as the interest rates on many
of these loans reset to higher rates, typically 2 to 3 percentage points
higher in a relatively short time period. A common subprime mortgage
product is a 2/28 hybrid ARM, which features a fixed interest rate for 2
years, followed by a series of resets up to a fully indexed adjustable
rate for the remaining 28 years of the loan.^29 Consider the example of a
borrower who took out a $166,000 2/28 loan in 2003 with an initial
interest rate of 7.5 percent and a first interest rate reset of 2.5
percentage points. During the first 2 years of the loan, the borrower's
monthly payment was $1,161. But after the first interest rate reset, the
borrower's monthly payment grew to $1,446, a $285 or 25 percent
increase.^30 Additional resets up to the fully indexed interest
rate--which can be as much as 6 percentage points higher than the initial
interest rate--would push the borrower's payments even higher. In contrast
to the subprime market, the large majority of FHA-insured loans are
fixed-rate mortgages. For example, fixed-rate loans accounted for 92
percent of FHA-insured mortgages made in 2005. Additionally, for
FHA-insured hybrid ARMs, the allowable interest rate adjustments after the
initial fixed-rate period are comparatively lower--1 percentage point for
3-year ARMs and 2 percentage points for 5-, 7-, and 10-year ARMs.^31

28Testimony from Sandra L. Thompson, Director, Division of Supervision and
Consumer Protection, Federal Deposit Insurance Corporation, entitled
Mortgage Market Turmoil: Causes and Consequences, before the Committee on
Banking, Housing and Urban Affairs U.S. Senate (March 22, 2007).

^29The fully indexed interest rate comprises an adjustable interest rate
index, such as the Federal Home Loan Bank of San Francisco Cost of Funds
Index, plus the lender's margin.

^30We based the loan amount and initial interest rate in this example on
average values for subprime loans made in 2003.

Reflecting in part the generally lower credit scores of subprime
borrowers, subprime mortgages are more likely than prime or FHA loans to
be what the Federal Reserve has designated "high-priced" loans. HMDA data
for the 2 most recent years available (2004 and 2005) include an indicator
for such loans. This indicator is based on a loan's annual percentage rate
(APR), which represents the cost of credit to the consumer by capturing
the contract interest rate on a loan, the points and fees that a consumer
pays, and other finance charges such as mortgage insurance premiums. Loans
with APRs at least 3 percentage points higher than the rate on Treasury
securities of comparable maturity are considered high-priced. Our analysis
of 2005 HMDA data indicates that approximately 90 percent of the loans we
had identified as subprime were high-priced. In contrast, less than 2
percent of FHA-insured loans made that year were high-priced.

Highly leveraged and weaker credit borrowers--the typical subprime
borrowers who have obtained nontraditional mortgage products such as
hybrid ARMs--are the most vulnerable to payment shocks.^32 Although
borrowers could avoid mortgage resets by refinancing to fixed-rate
mortgages, many of these borrowers face challenges to refinancing their
subprime loans. For example, about two-thirds of subprime loans originated
in 2005 had prepayment penalties--a substantially higher proportion than
in other market segments. FHA, for instance, does not permit prepayment
penalties on the loans it insures. Prepayment penalties generally last
from 2 to 4 years from the mortgage origination date and can amount to 4
to 5 percent of the original loan amount. They can make it expensive to
refinance because borrowers must pay the penalty if they wish to pay off
the original loan before the prepayment period expires. In addition,
subprime borrowers who made little or no down payment and live in areas
that experienced home price depreciation may not have sufficient equity to
refinance.

^31FHA does not offer 2-year hybrid ARMs.

^32For additional information about how some nontraditional mortgage
products create the potential for payment shock, see GAO, Alternative
Mortgage Products: Impact on Defaults Remains Unclear, but Disclosure of
Risks to Borrowers Could be Improved, [30]GAO-06-1021 (Washington, D.C.:
Sept. 19, 2006).

Borrowers who obtained subprime mortgages have experienced relatively high
rates of default (i.e., more than 90 days past due) and foreclosure (i.e.,
in any stage of the foreclosure process). According to MBA, as of December
31, 2006, the cumulative default and foreclosure rates for all subprime
mortgages were 7.78 and 4.53 percent, respectively. For subprime ARMs, the
corresponding figures were 9.16 and 5.62 percent. In comparison, as of the
same date, the default and foreclosure rates for FHA-insured loans were
5.78 and 2.19 percent, respectively (6.62 and 2.54 percent for ARMs) and
for prime loans, were 0.86 and 0.50, respectively (1.45 and 0.92 for
ARMs).^33

Some mortgage industry researchers predict that subprime default and
foreclosure rates likely will worsen as the loans age; a substantial
portion of these loans have yet to reach the age when loans tend to
experience the highest rates of default and foreclosure--between 4 and 7
years. Furthermore, because most recent subprime loans have
adjustable-rate features, default and foreclosure rates for ARMs are in
particular danger of increasing as interest rate resets cause monthly
mortgage payments on these loans to rise. A recent study by the director
of research and analytics at First American CoreLogic (one of the largest
private sector providers of mortgage information) illustrates the
potential scope of the problem posed by ARM resets. The study, which
examined 8.37 million ARMs originated in 2004 through 2006, estimated that
1.1 million (13 percent) of these loans would go into foreclosure as they
reset over the next 6 to 7 years.^34

Although the subprime and FHA market segments both serve higher-risk
borrowers, the extent to which subprime borrowers currently at risk of
default would have qualified for FHA-insured loans is not known. Such a
determination would require analysis of detailed, loan-level data for
subprime mortgages. Recently, a number of proposals have been made to help
subprime borrowers at risk of foreclosure refinance into lower-cost fixed
rate mortgages. For example, in April 2007, Freddie Mac announced plans to
purchase $20 billion in mortgages that would refinance troubled subprime
loans. Fannie Mae announced a similar initiative that same month.

^33The default and foreclosure rates for loans reported in the December
31, 2006, MBA National Delinquency Survey are computed using the total
number of loans as the base. The rapid growth in subprime loans in the
last 2 years has increased the base for the default and foreclosure rate
computations for these loans. All other things being equal, the growth in
the base would lead to lower default and foreclosure rates. Given that
many subprime loans are relatively new, the cumulative default and
foreclosure rates for subprime loans are likely to worsen as the newer
loans age.

^34First American CoreLogic, Inc., Mortgage Payment Reset: The Issue and
the Impact (Santa Ana, Calif.: Mar. 19, 2007).

Certain factors associated with FHA's decline in market share also
contributed to a worsening in indicators of credit risk among FHA
borrowers. More specifically, as conventional lenders expanded their
presence in traditional FHA submarkets through the development of new
products and use of automated underwriting tools, FHA experienced adverse
selection--that is, conventional providers identified and approved
relatively lower-risk borrowers, leaving relatively higher-risk borrowers
for FHA. According to analysis by FHA, FHA's loan portfolio is becoming
riskier in terms of the proportions of loans with high LTV,
payment-to-income, and debt-to-income ratios.^35 (Lenders use these ratios
to assess the creditworthiness of borrowers.) For instance, FHA's analysis
indicated that the proportion of loans with effective LTV ratios over 97
percent rose from about 40 percent in 1999 to almost 60 percent in 2005.
The higher the LTV ratio, the less equity borrowers have in their homes
and the more likely it is that they may default on mortgage obligations.
As we reported in November 2005, the substantial portion of FHA-insured
loans with down-payment assistance do not perform as well as loans without
such assistance, due partly to homebuyers having less equity in the
transaction.^36 The changes in borrower characteristics have contributed
to a decline in FHA's financial performance. In recent years, the credit
subsidy rate for FHA's single-family mortgage insurance program has
approached zero (the point at which estimated cash outflows equal
estimated cash inflows). Furthermore, FHA has estimated that, absent
program changes, the program for the first time would require a positive
subsidy (i.e., appropriations) in fiscal year 2008. Therefore, it has been
changes in the credit quality, rather than the volume, of loans FHA
insures that have had the most significant implications for FHA.

^35The payment-to-income ratio is a borrower's expected monthly housing
expenses as a percentage of monthly income. The debt-to-income ratio is a
borrower's expected monthly expenses on housing and other long-term debt
as a percentage of monthly income.

^36GAO, Mortgage Financing: Additional Actions Needed to Manage Risks of
FHA-insured Loans with Down Payment Assistance, [31]GAO-06-24 (Washington,
D.C.: Nov. 9, 2005).

Observations

Our analysis shows that in 2005 FHA was a much smaller part of the market
for home purchase mortgages than it was just a few years earlier. Given
FHA's history of serving minority and lower-income homebuyers, the
agency's sharp drop-off in market share among these populations is
particularly notable. Furthermore, the growth in low- and no-down-payment
mortgages offered by conventional lenders has made FHA's product offerings
less distinct. These trends raise questions about FHA's ability to fulfill
its traditional role and operate successfully in a changing and
competitive mortgage market. However, consistent with FHA's mission,
substantial proportions of recent FHA borrowers are minorities and
lower-income families, including many first-time homebuyers. Additionally,
in the event of an economic downturn, FHA could help ensure the flow of
mortgage credit to areas that private sector market participants may be
reluctant to serve. Furthermore, recent developments in the subprime
market may result in an increase in FHA's role in the mortgage market. For
example, relatively high default and foreclosure rates for subprime
mortgages and a contraction of this market segment could shift market
share to FHA. The extent to which this occurs will depend partly on the
efforts of conventional mortgage providers, including Freddie Mac and
Fannie Mae, to provide alternatives to subprime borrowers. As our report
noted, the GSEs have played a larger role among traditional FHA homebuyers
and recently have proposed steps that would provide additional mortgage
choices to many borrowers who obtained subprime loans.

Although further analysis would be required to determine how many subprime
borrowers at risk of default would qualify for FHA-insured mortgages, FHA
could be a vehicle to provide lower-priced and more sustainable mortgage
options for some borrowers who are considering or struggling to maintain
higher-priced subprime loans. FHA's recent efforts to modernize its
products and processes might facilitate any expansion of the agency's role
by increasing its operational efficiency and flexibility. However,
attracting subprime borrowers to FHA could also have costs, as some of
these borrowers may pose relatively high insurance risks. Careful
assessment and management of these risks would be necessary to avoid
exacerbating problems in the financial performance of FHA's insurance
program.

Agency Comments and Our Evaluation

We provided HUD with a draft of this report. HUD provided comments in a
letter from the Assistant Secretary for Housing-Federal Housing
Commissioner (see app. IV). HUD stated that we produced a straightforward,
well-researched report on the reasons for the recent decline in FHA's
market share.

HUD also provided observations about the homebuyers FHA serves and the
shift of some traditional FHA borrowers to subprime mortgage products that
have the potential to become more costly. Additionally, HUD noted that
additional flexibility, new mortgage insurance products, and risk-based
pricing would help FHA to continue providing lower-income and minority
households with homeownership opportunities at lower risk to themselves
and with manageable risk to FHA's insurance fund.

We are sending copies of this report to the Chairman, Senate Committee on
Banking, Housing, and Urban Affairs; Chairman and Ranking Member,
Subcommittee on Housing and Transportation, Senate Committee on Banking,
Housing, and Urban Affairs; Chairman and Ranking Member, House Committee
on Financial Services; and Chairman and Ranking Member, Subcommittee on
Housing and Community Opportunity, House Committee on Financial Services.
We will also send copies to the Secretary of Housing and Urban Development
and to other interested parties and make copies available to others upon
request. In addition, the report will be made available at no charge on
the GAO Web site at http://www.gao.gov.

Please contact me at (202) 512-8678 or [email protected] if you or your staff
have any questions about this report. Contact points for our Offices of
Congressional Relations and Public Affairs may be found on the last page
of this report. Key contributors to this report are listed in appendix V.

William B. Shear
Director, Financial Markets and Community Investment

Appendix I: Objectives, Scope, and Methodology

Our objectives were to determine (1) trends in the Federal Housing
Administration's (FHA) share of the market for home purchase mortgages and
selected submarkets from 1996 through 2005, and how they compared with the
trends for the prime, subprime, and government-sponsored enterprises (GSE)
market segments; and (2) the major factors associated with the trends in
FHA's market share and the potential implications of these trends for
homebuyers and FHA. To supplement this analysis, we also developed
information on the borrower and loan characteristics of FHA-insured
mortgages and mortgages in the prime and subprime market segments from
1996 through 2005 (see app. III).

Analysis of Market Share Trends

To analyze trends in the overall market for home purchase mortgages, we
compiled and analyzed loan data for 1996 through 2005 collected under the
Home Mortgage Disclosure Act (HMDA).^1 HMDA data are compiled and
published by the Federal Financial Institutions Examination Council
(FFIEC).^2 HMDA requires lending institutions to collect and publicly
disclose information about housing loans and applications for such loans.
This information includes, among other things, the market participant or
segment (conventional, FHA, Veterans Administration (VA), Rural Housing
Service (RHS), GSE), loan amount, property type, census tract and certain
tract characteristics, and loan applicant characteristics such as race,
gender, and income. HMDA data capture about 80 percent of the mortgage
loans funded each year, according to estimates by the Federal Reserve, and
are one of the most comprehensive source of information on mortgage
lending. In general, we limited our analysis to home purchase loans
originated for owner-occupied, one-to-four family and manufactured homes.
To the extent possible, we identified piggyback loans (i.e., the junior
lien in a pair of loans used to finance the same property) using a
data-matching process based on an algorithm developed by the Federal
Reserve and excluded these loans from our analysis.^3 As a result, our
analysis focuses on first liens only. Because the HMDA data do not contain
an indicator for subprime loans, we identified subprime loans by merging
the data with a list--maintained by the Department of Housing and Urban
Development (HUD)--of lenders that specialize in subprime lending. HUD's
Office of Policy Development and Research compiles this list annually by
analyzing HMDA data (e.g., lenders with lower origination rates and a
large share of refinance loans are more likely to be subprime lenders) and
contacting lenders directly. We designated conventional loans that were
not attributable to subprime lenders as prime loans. However, because
subprime specialists may originate some prime loans and nonsubprime
specialists may originate some subprime loans, any analysis that uses this
list will misclassify mortgages to some extent. Finally, HMDA data do not
capture all of the loans purchased by the GSEs. According to GSE and
Federal Reserve officials, HMDA data do not capture all of the loans the
GSEs purchase, including (1) many loans initially sold to intermediaries
(e.g., bank affiliates) and subsequently to the GSEs and (2) loans
originated and purchased in different years. While we acknowledge these
limitations, we used HMDA data to evaluate long-term market share trends
rather than to provide precise annual figures for each market segment,
including the GSE segment. According to Freddie Mac, Fannie Mae, and
Federal Reserve officials, our use of HMDA data was appropriate for this
purpose. Our analysis should be interpreted with these limitations in
mind.

^1HMDA requires lending institutions to collect and publicly disclose
information about housing loans and applications for such loans, including
the loan type and amount, property type, and borrower characteristics
(such as ethnicity, race, sex, and income). These data are the most
comprehensive source of information on mortgage lending.

^2FFIEC is a formal interagency body empowered to prescribe uniform
principles, standards, and report forms for the federal examination of
financial institutions by the Board of Governors of the Federal Reserve
System, the Federal Deposit Insurance Corporation, the National Credit
Union Administration, the Office of the Comptroller of the Currency, and
the Federal Home Loan Bank Board, and to make recommendations to promote
uniformity in the supervision of financial institutions.

To analyze trends in various submarkets, we incorporated additional data
from FHA, the Census Bureau, the Office of Federal Housing Enterprise
Oversight (OFHEO), and TransUnion (one of the three main consumer credit
reporting agencies). More specifically, from FHA we obtained annual
nationwide data on FHA's loan limits for single-family properties. From
the 2000 Decennial Census, we obtained information on the median house
price for each census tract. From OFHEO, we obtained their annual house
price appreciation index for all metropolitan statistical areas. Finally,
from TransUnion, we obtained median credit scores for mortgage borrowers
in each census tract nationwide as of December 31, 2004. We analyzed a
number of submarkets defined by borrower race, borrower income, loan
amount relative to FHA's loan limits, income and minority composition of
the census tract in which the property was located, and whether the
property was owner-occupied. We defined borrower race based on categories
in the HMDA data. Prior to 2004, HMDA data included Hispanic as a race
category but beginning in 2004 also included Hispanic as an ethnicity
variable. For 2004 and 2005, we classified borrowers of Hispanic ethnicity
as Hispanic race. We created borrower income categories using median
family incomes calculated by HUD each year for metropolitan and
nonmetropolitan areas. We defined borrowers with incomes of less than 80
percent of the area median income as low income, those with incomes of at
least 80 percent but less than 120 percent of the area median as moderate
income, and those with incomes of at least 120 percent of area median as
upper income. Finally, we determined the FHA-eligible submarket by
identifying loans with dollar amounts that fell within the relevant FHA
loan limit.

^3More specifically, we identified pairs of loans with identical
characteristics, including lender; property location; loan purpose; and
applicant race, gender, and income. For each matched pair of loans, we
designated (within certain parameters) the loan with the smaller dollar
value as the piggyback loan.

For our analysis of census tract groupings, we limited our examination to
census tracts where FHA's market share averaged at least 5 percent from
1996 through 1998. (We took this approach because our analysis examined
changes in FHA's market share during a 10-year period when the trend in
FHA's share was downward.) Therefore, our analysis excluded census tracts
where FHA's market share started and ended the period at zero and census
tracts where FHA's market share was sporadic and on average very small
near the beginning of the period. We used Census Bureau files relating
1990 census tract definitions to 2000 census tract definitions to provide
consistent geographic areas over the time period of our analysis. The
large majority of the census tracts were the same in both 1990 and 2000.
In many cases, however, 1990 census tracts were split into more than one
2000 census tract. In those cases, we aggregated the affected 2000 census
tracts to the corresponding 1990 tract definitions and used the 1990
tracts as the unit of analysis for the entire 1996 through 2005 period. In
other cases, two or more 1990 census tracts were combined to form one 2000
census tract. In those instances, we aggregated the affected 1990 tracts
that corresponded to the 2000 tract definitions and used the 2000 tracts
as the unit of analysis over the entire period.

We grouped the census tracts according to the percentage of the population
that was minority, median income, median credit score, and median home
price in relation to FHA loan limits. We defined low-, medium-, and
high-minority census tracts as those with minority populations of less
than 20 percent, 20 to 49 percent, and more than 50 percent, respectively.
We defined low-, moderate-, and upper-income census tracts as those with
median incomes that were less than 80 percent, at least 80 percent but
less than 120 percent, and 120 percent and above, respectively, of the
median income for the associated metropolitan statistical area. We also
grouped census tracts based on the TransUnion median credit score for
mortgage borrowers as of December 31, 2004. We categorized census tracts
into three groups: those with median credit scores in the bottom quarter
of the credit score distribution for all census tracts included in our
analysis, those with median scores between the bottom and top quarter, and
those with median scores in the top quarter. Finally, we created three
census tract groupings based on whether the median home price was below 75
percent of the applicable FHA loan limit each year, 75 to 125 percent of
the FHA limit each year, or below 75 percent of the loan limit at the
beginning of the 10-year period but at or above 75 percent of the limit
later in the period.

We assessed the reliability of the data we used by reviewing existing
information about the quality of the data, performing electronic data
testing to detect errors in completeness and reasonableness, and
interviewing Freddie Mac, Fannie Mae, FHA, and Federal Reserve officials
knowledgeable about the data. We determined that the data were
sufficiently reliable for the purposes of this report.

Analysis of Factors Associated with the Trends in FHA's Market Share and the
Implications of These Trends

To analyze factors associated with the trends in FHA's market share and
the implications of these trends, we used information from: the analysis
described in the previous section, HMDA data, HUD's Single-Family Data
Warehouse (SFDW), summary statistics provided by FHA and contained in
prior studies from databases maintained by LoanPerformance, the Mortgage
Bankers Association's (MBA) National Delinquency Survey for the fourth
quarter of 2006, and other published industry data. In order to assess the
reliability of the data we used, we reviewed related documentation and
interviewed officials familiar with the data. In addition, for the HMDA
and SFDW data, we performed internal checks to determine the extent to
which the data fields were populated and the reasonableness of the values
contained in the fields. We concluded that the data were sufficiently
reliable for the purposes of this report. We also reviewed relevant
academic literature and government and industry studies, including
internal FHA analysis of SFDW data.

In addition to our data analysis, we interviewed representatives of four
FHA lenders (Countrywide Financial, Wells Fargo, Bank of America, and
Lenders One--a mortgage cooperative representing 87 independent mortgage
bankers). We also interviewed officials from Fannie Mae, Freddie Mac, and
four private mortgage insurance companies--AIG United Guaranty, Genworth
Financial, Mortgage Guaranty Insurance Corporation, and PMI Mortgage
Insurance Company. Additionally, we interviewed representatives of six
mortgage and real estate industry groups--MBA, National Association of
Realtors, Mortgage Insurance Companies of America, National Association of
Home Builders, National Association of Mortgage Brokers, and American
Financial Services Association. We also spoke with representatives of the
following consumer advocacy groups: Center for Responsible Lending,
Consumer Action, Consumer Federation of America, National Association of
Consumer Advocates, National Community Reinvestment Coalition, National
Consumer Law Center, and National Council of La Raza. Finally, we
interviewed officials from FHA and HUD's Office of Policy Development and
Research.

Analysis of Borrower and Loan Characteristics

To determine the percentages of FHA, prime, and subprime home purchase
loans with certain borrower and loan characteristics each year from 1996
through 2005, we analyzed information from HMDA data, SFDW, the Federal
Housing Finance Board, and summary LoanPerformance data. The borrower and
loan characteristics were race, income, loan type (fixed or adjustable
rate), and presence of prepayment penalty. We also determined the average
interest rate at mortgage origination, loan amount, and median credit
score for FHA-insured loans and loans in the other market segments. In
order to assess the reliability of the data we used, we reviewed existing
information about the data quality and discussed the data with
knowledgeable officials to ensure that we interpreted the information
correctly. For the HMDA and SFDW data, we also performed electronic
testing to assess the reasonableness and completeness of the information.
We concluded that the data were sufficiently reliable for the purposes of
our report.

We conducted this work in Washington, D.C. from September 2006 through May
2007 in accordance with generally accepted government auditing standards.

Appendix II: Data on Market Share Trends in the Mortgage Market and
Selected Submarkets from 1996 through 2005

This appendix contains the results of our analysis using Home Mortgage
Disclosure Act (HMDA) and Federal Housing Administration (FHA) data for
calendar years 1996 through 2005. Tables 1 through 3 provide information
on home purchase mortgages. More specifically, table 1 contains market
shares for FHA and other market participants and segments over the 10-year
period. Table 2 contains FHA market shares and numbers of mortgages in
each state. Table 3 contains market shares in selected submarkets for FHA
and other market participants and segments. Table 4 provides market shares
for home purchase and refinance loans combined. Finally, table 5 provides
market shares for refinance loans.

Table 1: Market Shares for Home Purchase Loans, 1996-2005

Market shares
              1996      1997      1998      1999      2000      2001      2002      2003      2004      2005 
Prime        73.2%     71.8%     72.8%     71.5%     71.2%     71.1%     73.0%     74.0%     75.7%     76.4% 
Subprime       1.6       2.6       3.7       4.5       5.9       5.0       6.6       9.3      12.0      14.5 
FHA           18.8      19.6      17.6      19.0      18.6      19.3      16.5      13.1       9.2       6.3 
VA             6.0       5.4       5.3       4.5       3.9       4.2       3.6       3.1       2.6       2.4 
RHS            0.4       0.5       0.6       0.5       0.4       0.4       0.4       0.5       0.4       0.4 
GSE           26.8      26.6      32.6      29.2      30.1      35.1      39.3      38.2      35.1      30.4 
Number of loans
Prime    2,270,445 2,338,558 2,793,403 2,914,713 2,825,668 2,852,557 2,928,394 3,174,378 3,400,642 3,557,981 
Subprime    48,778    85,721   143,395   181,779   235,044   199,875   265,000   397,833   540,352   675,389 
FHA        582,781   637,354   675,737   774,259   737,914   776,380   660,726   561,582   413,754   294,777 
VA         187,424   176,768   201,973   183,875   156,767   168,365   142,673   134,759   117,961   109,873 
RHS         13,122    16,804    21,961    19,716    14,304    16,906    16,600    21,773    19,696    18,471 
GSE        831,869   865,428 1,250,629 1,188,913 1,193,653 1,408,297 1,576,259 1,639,462 1,577,474 1,415,366 

Source: GAO analysis of HMDA data.

Note: We calculated market shares based on numbers of loans and, to the
extent possible, excluded piggyback loans. The prime, subprime, FHA, VA,
and RHS market shares add to 100 percent (figures used in this table were
rounded to the nearest tenth of a percent). The GSE market segment is
primarily a subset of the prime market segment. Data for the GSEs do not
include loans originated and purchased in different years or all of the
loans sold to intermediaries before being purchased by the GSEs.

Table 2: State-by-State FHA Market Shares and Loan Counts for Home
Purchase Loans, 1996-2005

State            1996    1997    1998    1999    2000    2001    2002    2003    2004    2005 
Market shares                                                                                 
Alabama         12.8%   12.6%   11.9%   13.9%   16.8%   19.7%   19.2%   16.3%   13.0%   10.4% 
Alaska           27.5    37.0    31.0    30.6    29.0    29.2    27.6    24.7    20.5    14.1 
Arizona          22.3    22.1    20.1    21.7    22.6    23.1    20.5    15.8     8.6     2.8 
Arkansas         22.9    22.4    19.4    20.3    20.5    23.9    22.9    20.8    16.1    13.4 
California       23.9    24.1    21.8    21.5    18.8    16.8    11.0     6.1     1.9     0.6 
Colorado         25.5    27.0    22.2    25.3    24.1    24.7    26.7    24.1    15.5     8.9 
Connecticut      19.1    19.2    17.9    18.8    19.0    19.1    15.1    12.2     9.3     7.3 
Delaware         16.3    17.2    18.7    17.0    18.4    20.0    16.5    11.8     8.3     5.5 
District of                                                                                   
Columbia         24.3    22.3    21.9    22.1    20.7    17.3    12.6     6.8     4.3     0.9 
Florida          16.0    17.9    17.5    17.7    17.0    16.7    12.7     9.8     6.2     3.2 
Georgia          17.5    18.4    15.8    21.1    21.0    23.4    22.2    18.2    14.6    10.7 
Hawaii            9.1    13.1    12.6    12.9     9.0     7.3     5.2     3.3     1.3     1.2 
Idaho            25.3    23.5    21.5    23.9    23.4    25.8    22.4    20.0    13.5     8.8 
Illinois         17.8    19.9    15.6    18.6    16.8    17.6    15.5    12.2     7.9     4.9 
Indiana          18.7    21.1    18.4    20.0    21.9    27.6    24.3    21.3    17.0    13.6 
Iowa             11.6    13.8    13.0    13.5    12.7    14.4    11.5     9.8     7.9     6.3 
Kansas           15.2    15.1    14.8    16.3    15.9    18.0    16.5    14.1    12.1     9.7 
Kentucky         14.8    14.6    13.9    15.1    15.9    17.6    17.1    15.3    12.5     9.9 
Louisiana        19.0    18.5    18.9    21.5    21.4    22.4    20.0    15.6    13.1     9.4 
Maine            20.5    20.9    18.2    18.4    15.8    15.0    12.9     7.0     7.5     5.3 
Maryland         33.7    35.6    34.1    32.7    31.6    30.4    25.5    16.6     9.5     3.4 
Massachusetts     9.8    12.1    10.0    11.0    10.8    10.6     8.1     5.8     3.0     1.6 
Michigan         15.3    18.0    15.2    17.0    16.4    18.0    16.4    12.8    10.0     7.5 
Minnesota        26.5    24.7    18.1    19.0    17.2    18.1    15.4    11.9     8.4     6.1 
Mississippi      21.8    18.6    16.4    15.5    18.0    22.9    21.2    18.0    13.9    12.1 
Missouri         21.5    20.7    18.2    19.6    19.7    20.7    18.2    14.4    11.6     9.0 
Montana          22.1    20.2    18.0    18.2    20.2    23.7    22.1    18.5    16.8    12.5 
Nebraska         21.4    25.4    23.5    25.8    26.7    25.8    25.1    19.1    13.0    10.6 
Nevada           25.7    27.8    22.8    27.2    23.6    20.2    18.1    13.1     5.0     2.4 
New Hampshire    20.7    20.6    17.9    16.1    14.7    13.7    10.1     7.6     3.7     2.2 
New Jersey       16.4    17.6    16.0    17.9    17.1    16.0    12.2    10.1     6.1     3.7 
New Mexico       15.0    14.9    14.2    18.4    19.8    23.9    24.5    20.5    15.9    11.3 
New York         16.3    16.8    16.1    16.2    15.2    14.8    11.6     9.6     6.5     5.2 
North            12.3    12.9    10.5    13.0    15.1    17.3    16.3    13.3    10.7     8.4 
Carolina                                                                                      
North Dakota     25.7    26.6    28.8    27.4    23.8    27.6    22.1    18.5    18.2    15.9 
Ohio             14.8    17.4    16.1    17.5    18.6    20.4    17.7    14.7    11.8     9.5 
Oklahoma         25.4    24.2    25.3    25.4    22.9    25.4    23.4    22.1    17.2    14.8 
Oregon           12.7    12.5    10.3    13.9    15.1    17.1    14.3    11.3     6.4     3.3 
Pennsylvania     15.1    16.8    15.6    16.1    16.2    17.5    14.1    10.6     7.9     5.9 
Rhode Island     21.7    25.1    22.5    24.4    23.5    22.7    17.3    13.1     7.7     3.4 
South             8.4     8.3     6.0     8.2    10.0    11.3    10.3     8.8     7.7     5.6 
Carolina                                                                                      
South Dakota     20.4    21.0    19.5    19.2    17.6    18.4    17.4    17.1    13.5     8.7 
Tennessee        25.6    24.0    21.5    22.5    21.5    24.7    21.9    18.8    13.9    10.2 
Texas            20.9    21.1    20.5    21.7    20.9    22.8    22.3    20.4    18.0    13.3 
Utah             27.3    27.4    22.0    26.5    29.9    33.0    29.6    26.3    19.3    12.2 
Vermont           4.6     8.1     7.2     7.3     7.6     8.2     5.6     4.9     3.4     2.2 
Virginia         24.7    24.0    21.9    22.7    22.8    22.0    17.8    13.1     7.4     4.4 
Washington       17.2    16.6    12.4    15.5    15.6    17.9    15.6    12.6     7.6     4.1 
West Virginia     9.2     7.5     8.1     9.2    10.3    12.3    12.1     8.7     7.3     6.5 
Wisconsin         5.1     6.6     5.6     6.8     7.0     8.6     7.4     6.1     4.9     4.3 
Wyoming          22.0    22.2    18.2    20.1    18.0    18.0    15.1    12.2    10.3     7.8 
Number of loans
State            1996    1997    1998    1999    2000    2001    2002    2003    2004    2005 
Alabama         6,321   6,976   7,358   8,892   9,682  10,444   9,504   9,122   7,731   6,602 
Alaska          1,380   2,092   2,111   1,819   1,805   2,043   2,007   2,102   2,246   1,567 
Arizona        18,510  18,091  20,041  23,248  23,051  25,711  22,984  19,509  11,942   4,213 
Arkansas        5,115   5,638   5,297   6,065   5,960   7,117   6,310   6,354   5,225   4,629 
California     77,824  89,847  96,097 104,377  92,884  80,582  58,350  33,652  10,786   3,226 
Colorado       19,302  20,986  21,691  26,302  25,603  25,672  25,527  23,320  15,587   9,222 
Connecticut     7,433   7,717   8,827  10,189   9,494   9,794   7,923   6,397   5,153   3,970 
Delaware        1,576   1,850   2,364   2,221   2,297   2,433   2,064   1,508   1,114     804 
District of     1,078   1,194   1,565   1,799   1,673   1,385   1,013     579     411      86 
Columbia                                                                                      
Florida        35,857  41,886  49,058  53,671  51,341  52,985  40,258  33,967  22,818  12,703 
Georgia        18,095  20,908  21,269  30,766  27,943  31,741  28,307  26,148  22,582  17,705 
Hawaii            547     891     993   1,187     879     673     619     494     190     171 
Idaho           3,456   3,795   4,226   4,997   4,659   5,377   4,541   4,831   3,570   2,733 
Illinois       26,411  28,586  25,976  34,091  30,830  32,595  28,867  24,257  16,281  10,367 
Indiana        13,508  15,265  15,179  17,886  18,993  24,199  20,549  19,154  15,835  13,024 
Iowa            2,900   3,763   4,027   4,287   3,808   4,570   3,622   3,391   2,884   2,452 
Kansas          4,094   4,446   5,136   6,067   5,304   6,327   5,440   5,008   4,501   3,838 
Kentucky        5,200   5,714   6,411   6,940   7,041   7,723   7,107   7,435   6,205   5,131 
Louisiana       7,799   7,578   9,249  10,195   9,469   9,670   8,655   7,488   6,433   4,871 
Maine           1,742   2,198   2,311   2,458   2,001   1,878   1,631     782   1,188     827 
Maryland       21,739  24,431  27,897  28,920  28,269  29,088  24,441  16,135   9,799   3,444 
Massachusetts   6,977   8,931   8,404   9,676   8,713   8,481   6,418   4,754   2,804   1,387 
Michigan       20,381  20,956  22,149  26,331  24,120  25,320  22,341  18,291  14,356  10,212 
Minnesota      17,164  15,323  14,595  15,606  13,848  15,169  13,084  10,685   7,203   5,174 
Mississippi     5,049   5,100   5,168   4,809   5,215   6,153   5,261   5,065   4,080   3,690 
Missouri       14,228  13,840  13,974  15,472  15,035  15,907  13,254  11,592  10,119   8,224 
Montana         1,253   1,464   1,587   1,666   1,694   2,071   1,961   1,824   1,866   1,399 
Nebraska        2,983   3,765   4,197   4,572   4,745   4,389   4,624   3,918   2,762   2,238 
Nevada          8,974   9,938   8,989  11,988  10,601   9,780   8,759   7,581   3,150   1,602 
New Hampshire   2,511   2,912   3,044   2,901   2,636   2,328   1,688   1,264     731     429 
New Jersey     14,661  16,137  17,393  21,025  19,046  17,702  13,944  12,283   7,623   4,585 
New Mexico      3,018   3,086   3,271   4,008   4,079   5,316   5,701   5,194   4,270   3,084 
New York       20,500  22,090  24,313  26,064  23,589  23,055  18,735  16,360  11,670   9,241 
North          12,802  14,620  13,420  17,822  18,476  20,646  17,945  16,005  13,933  12,040 
Carolina                                                                                      
North Dakota    1,260   1,394   1,853   1,628   1,248   1,531   1,315   1,252   1,331   1,257 
Ohio           20,332  23,584  25,098  27,867  28,442  31,935  27,317  24,047  19,059  15,552 
Oklahoma        8,611   8,712  10,864  11,274   9,403  10,308   9,205   9,232   7,554   7,046 
Oregon          4,869   5,685   5,569   7,213   7,353   8,794   7,338   6,311   3,747   2,211 
Pennsylvania   17,466  19,698  20,623  22,761  22,384  24,285  19,799  15,820  11,812   9,266 
Rhode Island    2,104   2,494   2,705   3,315   3,069   3,085   2,466   1,743   1,086     481 
South           4,338   4,368   3,610   5,058   5,538   6,081   5,189   5,105   4,793   3,733 
Carolina                                                                                      
South Dakota    1,249   1,467   1,632   1,619   1,416   1,553   1,400   1,456   1,290     888 
Tennessee      18,026  17,125  17,163  18,547  17,554  19,305  16,283  15,291  12,623  10,088 
Texas          47,836  49,485  58,722  66,753  66,412  73,528  69,465  64,823  58,715  45,779 
Utah            8,348   8,313   7,119   9,230  10,074  11,281   9,782  10,034   8,538   6,501 
Vermont           176     355     373     425     436     488     289     280     218     145 
Virginia       21,282  22,184  24,346  27,480  27,966  29,486  23,927  18,576  11,300   6,496 
Washington     11,843  14,672  12,553  15,840  15,030  17,740  15,726  14,553   8,995   5,298 
West Virginia     983     922   1,150   1,292   1,466   1,821   1,736   1,338   1,273   1,188 
Wisconsin       2,774   3,741   3,642   4,487   4,425   5,847   5,233   4,526   3,659   3,347 
Wyoming           896   1,141   1,128   1,153     915     988     822     746     713     611 
Total         582,781 637,354 675,737 774,259 737,914 776,380 660,726 561,582 413,754 294,777 

Source: GAO analysis of HMDA data.

Note: We calculated market shares based on numbers of loans and, to the
extent possible, excluded piggyback loans.

Table 3: Market Shares for Home Purchase Loans in Selected Submarkets,
1996-2005

              1996      1997      1998      1999      2000      2001      2002      2003      2004      2005 
Minority borrowers
Market shares
Prime        58.6%     56.1%     56.5%     55.9%     56.5%     57.1%     60.2%     62.1%     63.8%     64.7% 
Subprime       2.0       3.5       5.4       6.8       8.2       7.2      10.5      16.1      22.1      26.0 
FHA           31.6      33.5      31.4      32.0      31.0      31.1      25.3      18.4      11.3       6.9 
VA             7.5       6.5       6.2       5.0       4.2       4.4       3.7       3.2       2.6       2.3 
RHS            0.4       0.4       0.4       0.3       0.2       0.3       0.3       0.3       0.2       0.2 
GSE           21.0      19.8      23.3      20.8      22.5      25.7      31.7      31.1      25.5      20.3 
Number of loans
Prime      360,310   364,069   415,696   475,415   486,901   479,712   545,749   627,808   575,271   753,007 
Subprime    12,259    23,003    40,079    57,711    70,455    60,502    95,162   162,540   199,081   302,296 
FHA        193,975   217,379   231,473   272,433   266,912   261,088   229,657   185,532   101,724    80,038 
VA          46,008    42,204    45,784    42,716    35,991    36,832    33,875    31,969    23,097    26,254 
RHS          2,213     2,552     3,149     2,681     1,982     2,261     2,485     2,943     1,880     2,046 
GSE        129,130   128,841   171,718   176,680   193,784   216,195   287,896   314,573   229,635   236,521 
Lower-income borrowers
Market Shares
Prime        65.0%     62.9%     64.4%     63.0%     62.4%     62.3%     65.5%     67.6%     70.1%     71.7% 
Subprime       1.4       2.5       3.9       5.0       6.7       5.3       6.8       9.4      12.7      15.0 
FHA           26.0      27.5      24.7      26.0      25.7      27.1      23.1      18.6      13.4       9.8 
VA             6.9       6.3       6.0       5.2       4.5       4.7       4.0       3.6       3.1       2.9 
RHS            0.7       0.8       0.9       0.8       0.6       0.7       0.6       0.8       0.7       0.7 
GSE           23.8      23.4      28.9      26.7      28.2      32.2      38.4      38.5      35.9      31.9 
Number of loans
Prime    1,185,411 1,210,961 1,469,715 1,569,101 1,486,038 1,498,980 1,592,975 1,749,681 1,868,304 1,828,021 
Subprime    25,490    47,337    87,860   125,000   160,032   126,461   165,128   243,937   338,614   382,837 
FHA        474,582   528,562   564,317   647,179   612,534   651,833   561,894   480,480   356,448   248,863 
VA         125,811   120,664   137,619   128,740   107,526   113,866    96,576    92,749    82,616    73,229 
RHS         12,576    16,207    21,208    18,836    13,595    15,657    14,896    20,301    18,484    17,208 
GSE        433,235   451,031   659,117   664,874   670,876   774,874   932,713   996,578   955,888   812,593 
FHA-eligible loans
Market shares
Prime        66.9%     65.2%     66.7%     65.4%     64.9%     65.2%     67.7%     68.8%     71.0%     72.5% 
Subprime       1.4       2.6       4.0       5.1       7.0       5.4       7.0       9.6      12.7      15.2 
FHA           25.4      26.2      23.3      24.3      23.6      24.4      21.1      17.7      12.8       9.2 
VA             5.7       5.2       5.2       4.6       4.0       4.4       3.7       3.2       2.9       2.5 
RHS            0.6       0.7       0.8       0.6       0.5       0.5       0.5       0.7       0.6       0.6 
GSE           24.0      24.0      29.5      27.0      28.4      33.4      39.0      38.5      37.4      32.8 
Number of loans
Prime    1,428,428 1,522,418 1,872,251 1,977,376 1,870,710 2,038,154 2,082,838 2,093,435 2,228,157 2,260,564 
Subprime    30,685    60,067   112,755   152,900   201,604   170,390   214,870   293,109   397,608   475,661 
FHA        541,744   612,373   653,048   736,229   680,059   763,584   649,852   537,877   402,342   286,470 
VA         122,170   121,717   146,434   138,893   115,075   138,048   113,959    97,958    90,380    79,195 
RHS         12,685    16,633    21,761    19,169    13,874    16,620    16,389    21,109    19,151    17,835 
GSE        512,969   560,485   826,462   817,983   819,521 1,042,809 1,201,354 1,171,130 1,173,893 1,023,192 
Investor properties
Market shares
Prime        94.3%     94.6%     93.7%     94.9%     95.4%     94.3%     95.5%     92.8%     90.9%     90.4% 
Subprime       2.6       4.4       5.7       4.7       4.0       3.7       4.4       7.1       9.1       9.5 
FHA            2.8       0.7       0.5       0.4       0.4       1.4       0.1       0.1       0.0       0.0 
VA             0.3       0.3       0.1       0.1       0.1       0.5       0.0       0.0       0.0       0.0 
RHS            0.0       0.0       0.0       0.0       0.0       0.0       0.0       0.0       0.0       0.0 
GSE           19.1      21.6      26.7      28.1      34.8      40.0      45.9      41.3      34.2      29.2 
Number of loans
Prime      369,504   446,786   522,788   595,192   638,724   698,132   845,224 1,033,234 1,355,352 1,647,122 
Subprime    10,128    20,922    31,730    29,176    26,806    27,420    38,740    79,040   135,134   173,424 
FHA         10,914     3,324     2,704     2,402     2,876    10,378       618       936       394       230 
VA           1,264     1,302       760       588       958     3,946       402       324       188       192 
RHS            104        70       134        76       108        66        66        54        68        72 
GSE         74,788   102,140   149,234   176,248   233,034   295,728   406,028   459,400   510,338   532,204 

Source: GAO analysis of HMDA and HUD data.

Note: We calculated market shares based on numbers of loans and, to the
extent possible, excluded piggyback loans. The prime, subprime, FHA, VA,
and RHS market shares add to 100 percent (figures used in this table were
rounded to the nearest tenth of a percent). The GSE market segment is
primarily a subset of the prime market segment. Data for the GSEs do not
include loans originated and purchased in different years or all of the
loans sold to intermediaries before being purchased by the GSEs.

Table 4: Market Shares for Home Purchase and Refinance Loans, 1996-2005

             1996  1997  1998  1999  2000  2001  2002  2003  2004  2005 
Prime    78.0% 74.4% 80.1% 75.2% 71.8% 81.4% 83.7% 84.5% 75.8% 76.6% 
Subprime   5.4   9.3   8.4  10.9  12.4   7.9   8.5   9.8  18.0  18.5 
FHA       12.3  12.5   8.5  11.0  12.9   8.7   6.3   4.4   4.7   3.5 
VA         4.0   3.5   2.7   2.6   2.7   1.9   1.4   1.2   1.3   1.2 
RHS        0.3   0.3   0.2   0.3   0.2   0.2   0.1   0.1   0.2   0.2 
GSE       28.9  27.0  38.4  31.1  26.9  40.3  46.0  49.9  36.1  30.1 

Source: GAO analysis of HMDA data.

Note: We calculated market shares based on numbers of loans and, to the
extent possible, excluded piggyback loans. The prime, subprime, FHA, VA,
and RHS market shares add to 100 percent. (figures used in this table were
rounded to the nearest tenth of a percent). The GSE market segment is
primarily a subset of the prime market segment. Data for the GSEs do not
include loans originated and purchased in different years or all of the
loans sold to intermediaries before being purchased by the GSEs.

Table 5: Market Shares for Refinance Loans, 1996-2005

             1996  1997  1998  1999  2000  2001  2002  2003  2004  2005 
Prime    85.7% 78.3% 85.4% 79.5% 73.1% 87.9% 88.8% 88.2% 75.8% 76.8% 
Subprime  11.4  19.2  11.8  18.2  25.0   9.7   9.4  10.0  22.6  22.0 
FHA        2.0   1.8   1.9   1.9   1.7   2.0   1.4   1.4   1.3   1.0 
VA         0.9   0.6   0.9   0.4   0.2   0.4   0.4   0.5   0.2   0.2 
RHS        0.1   0.0   0.0   0.0   0.0   0.0   0.0   0.0   0.0   0.0 
GSE       32.2  27.5  42.5  33.4  20.8  43.5  49.3  54.1  36.9  29.9 

Source: GAO analysis of HMDA data.

Note: We calculated market shares based on numbers of loans and, to the
extent possible, excluded piggyback loans. The prime, subprime, FHA, VA,
and RHS market shares add to 100 percent (figures used in this table were
rounded to the nearest tenth of a percent). The GSE market segment is
primarily a subset of the prime market segment. Data for the GSEs do not
include loans originated and purchased in different years or all of the
loans sold to intermediaries before being purchased by the GSEs.

Appendix III: Data on Selected Borrower and Loan Characteristics for FHA,
Prime, and Subprime Loans, 1996 through 2005

This appendix contains the results of our analysis of Home Mortgage
Disclosure Act (HMDA) and Single-Family Data Warehouse (SFDW) data,
information from the Federal Housing Finance Board, and summary
LoanPerformance data. Specifically, tables 6, 7, and 8 contain information
on selected borrower and loan characteristics for Federal Housing
Administration (FHA)-insured, prime, and subprime loans. For prime and
subprime loans, data were not available from the sources we used for the
entire period we examined (1996 through 2005).

Table 6: Selected FHA Borrower and Loan Characteristics, 1996-2005

                   1996    1997    1998     1999     2000     2001     2002     2003     2004     2005 
Percentage of                                                                                          
loans to all                                                                                           
minority                                                                                               
borrowers^a       33.8%   34.8%   35.3%    36.9%    38.7%    36.6%    37.2%    34.9%    31.6%    29.3% 
Percentage of                                                                                          
loans to                                                                                               
Hispanic                                                                                               
borrowers^a       14.4%   15.1%   15.8%    16.5%    17.6%    17.5%    18.0%    16.9%    14.7%          
Percentage of                                                           FHA                            
loans to black                                                       Market                            
borrowers^a                                                           Share                            
                                                                      13.2%                            
                                                                        FHA                            
                                                                     Market                            
                                                                      Share                            
                                                                        FHA                            
                                                                     Market                            
                                                                      Share                            
                                                                        FHA                            
                                                                     Market                            
                                                                      Share                            
                                                                        FHA                            
                                                                     Market                            
                                                                      Share                            
                                                                        FHA                            
                                                                     Market                            
                                                                      Share                            
                                                                        FHA                            
                                                                     Market                            
                  13.8%   13.9%   13.6%    14.0%    14.8%    13.3%    Share    12.4%    13.5%    13.2% 
Percentage of                                                                                          
loans to                                                                                               
low-income                                                                                             
borrowers^a       44.4%   47.0%   48.5%    49.0%    48.1%    50.0%    53.4%    54.1%    55.4%    53.1% 
Percentage of                                                                                          
loans to                                                                                               
moderate-income                                                                                        
borrowers^a       37.1%   35.9%   35.0%    34.6%    34.9%    33.9%    31.6%    31.4%    30.7%    31.3% 
Average loan                                                                                           
amount^a        $85,683 $88,559 $91,279 $100,095 $104,406 $111,720 $117,796 $123,334 $122,840 $123,197 
Average                                                                                                
borrower credit                                                                                        
score^b                     660     662      653      639      647      645      647      640      640 
Average initial                                                                                        
interest rate                                                                                          
(ARMs)^b           6.6%    6.4%    6.0%     6.3%     7.1%     5.7%     5.2%     4.3%     4.5%     4.9% 
Average                                                                                                
interest rate                                                                                          
(fixed-rate                                                                                            
mortgages)^b       7.9%    7.8%    7.2%     7.5%     8.3%     7.3%     6.9%     6.0%     6.1%     5.9% 
Percentage of                                                                                          
first liens                                                                                            
that are ARMs^b   25.5%   32.3%    6.4%     6.6%     7.8%     2.9%     8.5%     7.1%    12.0%     7.6% 
Percentage of                                                                                          
first liens                                                                                            
with prepayment                                                                                        
penalties^a          0%      0%      0%       0%       0%       0%       0%       0%       0%       0% 

Source: GAO analysis of HMDA and HUD data.

aData from HMDA.

bData from SFDW.

Note: FHA does not allow prepayment penalties on the loans it insures.

Table 7: Selected Prime Borrower and Loan Characteristics, 1996-2005

                    1996     1997     1998     1999     2000     2001     2002     2003     2004     2005 
Percentage of                                                                                             
loans to all                                                                                              
minority                                                                                                  
borrowers^a        16.4%    16.3%    15.9%    17.7%    19.4%    19.4%    21.2%    21.7%    21.5%    23.9% 
Percentage of                                                                                             
loans to                                                                                                  
Hispanic                                                                                                  
borrowers^a         4.8%     4.5%     4.5%     5.1%     6.0%     6.5%     7.1%     7.5%     8.5%    10.2% 
Percentage of                                                                                             
loans to black                                                                                            
borrowers^a         4.6%     4.6%     4.3%     4.4%     4.5%     4.0%     4.1%     4.6%     5.0%     5.8% 
Percentage of                                                                                             
loans to                                                                                                  
low-income                                                                                                
borrowers^a        25.0%    25.0%    25.7%    27.5%    26.7%    26.3%    27.5%    28.1%    28.0%    25.5% 
Percentage of                                                                                             
loans to                                                                                                  
moderate-income                                                                                           
borrowers^a        27.2%    26.7%    26.9%    26.3%    25.9%    26.2%    26.9%    27.0%    26.9%    25.9% 
Average loan                                                                                              
amount^a        $114,315 $120,439 $127,858 $137,545 $147,814 $158,335 $178,201 $187,080 $212,528 $235,258 
Average                                                                                                   
borrower credit                                                                                           
score^b                                                                    716      719      721      723 
Average                                                                                                   
interest rate                                                                                             
at origination                                                                                            
(ARMs)^c            6.9%     6.8%     6.4%     6.5%     7.0%     6.3%     5.6%     5.0%     5.2%     5.5% 
Average                                                                                                   
interest rate                                                                                             
(fixed-rate                                                                                               
loans)^c            7.8%     7.7%     7.1%     7.3%     8.1%     7.0%     6.7%     5.8%     6.0%     6.0% 
Percentage of                                                                                             
first-liens                                                                                               
that are ARMs^d                                                          13.1%    11.8%    24.5%    22.5% 
Percentage of                                                                                             
first-liens                                                                                               
with prepayment                                                                                           
penalties^d                                                               0.5%     0.3%     4.0%     6.0% 

Source: GAO analysis of HMDA data; information from Freddie Mac, Fannie
Mae, and the Federal Housing Finance Board; and FHA-provided summaries of
information in LoanPerformance's TrueStandings Servicing prime database.

aData from HMDA.

bData from Freddie Mac and Fannie Mae.

cData from the Federal Housing Finance Board.

dData from FHA-provided summaries of information in LoanPerformance's
TrueStandings Servicing prime database.

Table 8: Selected Subprime Borrower and Loan Characteristics, 1996-2005

                    1996     1997     1998     1999    2000     2001     2002     2003     2004     2005 
Percentage of                                                                                            
loans to all                                                                                             
minority                                                                                                 
borrowers^a        27.8%    30.5%    33.2%    39.6%   36.5%    36.0%    42.2%    47.2%    47.2%    53.5% 
Percentage of                                                                                            
loans to                                                                                                 
Hispanic                                                                                                 
borrowers^a         7.7%     8.4%     9.1%    10.1%   10.2%    11.9%    17.1%    20.9%    23.0%    28.4% 
Percentage of                                                                                            
loans to black                                                                                           
borrowers^a         9.3%    12.8%    15.9%    17.7%   17.1%    16.6%    15.6%    15.4%    17.5%    18.6% 
Percentage of                                                                                            
loans to                                                                                                 
low-income                                                                                               
borrowers^a        24.5%    27.5%    32.3%    40.0%   40.1%    33.9%    31.8%    29.9%    30.9%    27.1% 
Percentage of                                                                                            
loans to                                                                                                 
moderate-income                                                                                          
borrowers^a        27.8%    27.7%    29.0%    28.8%   28.0%    29.4%    30.5%    31.5%    31.8%    29.6% 
Average loan                                                                                             
amount^a        $122,478 $118,189 $110,638 $101,337 $94,758 $117,110 $150,050 $178,049 $204,234 $233,901 
Average                                                                                                  
borrower credit                                                                                          
score^b                                                                   611      620      622      622 
Average                                                                                                  
interest rate                                                                                            
at origination                                                                                           
(ARMs)^c            8.7%     9.5%     9.7%     9.9%   10.6%     9.6%     8.5%     7.5%     6.9%          
Average                                                                                                  
interest rate                                                                                            
(fixed-rate                                                                                              
loans)^c           10.1%    10.0%     9.6%    10.4%   11.3%     9.8%     8.6%     7.5%     7.3%          
Percentage of                                                                                            
first liens                                                                                              
that are ARMs^c                                                         67.6%    62.3%    71.9%    72.6% 
Percentage of                                                                                            
first liens                                                                                              
with prepayment                                                                                          
penalties^c                                                             64.1%    61.6%    59.9%    66.2% 

Source: GAO analysis of HMDA data, information from UBS, and data from the
Federal Reserve Bank of St. Louis.

aData from HMDA.

bData from UBS Mortgage Strategist (November 29, 2005). UBS analyzed data
from LoanPerformance's TrueStandings Securities subprime database.

cData from "The Evolution of Subprime Lending," by Anthony
Pennington-Cross and Souphala Chomsisengphet, Federal Reserve Bank of St.
Louis Review, January/February 2006, Vol. 88 (No. 1), pp. 42, 44, 50, 53.
The paper presents analysis of data from LoanPerformance's TrueStandings
Securities subprime database. This database consists primarily of the
least risky (A-) grade of subprime loans. Therefore, the information we
cite from the paper, including average interest rates, is not
representative of riskier grades of subprime loans. The analysis is
reprinted with the permission of the Federal Reserve Bank of St. Louis.

Appendix IV: Comments from the Department of Housing and Urban Development

Appendix V: GAO Contact and Staff Acknowledgments

GAO Contact

William Shear (202) 512-8678 or [email protected]

Acknowledgments

In addition, Steve Westley (Assistant Director), Triana Bash, Steve Brown,
John McGrail, Jeff Miller, Marc Molino, Barbara Roesmann, Richard Vagnoni,
and Jim Vitarello made key contributions to this report.

(250320)

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 GAO Reports and Testimony

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site ( [32]www.gao.gov ). Each weekday, GAO posts
newly released reports, testimony, and correspondence on its Web site. To
have GAO e-mail you a list of newly posted products every afternoon, go to
[33]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, Waste, and Abuse in Federal Programs

Contact:

Web site: [34]www.gao.gov/fraudnet/fraudnet.htm
E-mail: [35][email protected]
Automated answering system: (800) 424-5454 or (202) 512-7470

Congressional Relations

Gloria Jarmon, Managing Director, [36][email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Washington,
D.C. 20548

Public Affairs

Paul Anderson, Managing Director, [37][email protected] (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548

[38]www.gao.gov/cgi-bin/getrpt?GAO-07-645 .

To view the full product, including the scope
and methodology, click on the link above.

For more information, contact William B. Shear at (202) 512-8678 or
[email protected].

Highlights of [39]GAO-07-645 , a report to congressional requesters

June 2007

FEDERAL HOUSING ADMINISTRATION

Decline in the Agency's Market Share Was Associated with Product and
Process Developments of Other Mortgage Market Participants

The Federal Housing Administration (FHA) historically has been an
important participant in the mortgage market, which includes loans that
carry government insurance or guarantees (such as FHA-insured mortgages)
and those that do not (conventional mortgages). The conventional market
comprises prime loans for the most creditworthy borrowers and subprime
loans for borrowers with impaired credit. Reduced demand for FHA-insured
mortgages--which are used primarily by borrowers who would have difficulty
obtaining conventional prime loans--has raised questions about the
agency's role in and ability to adapt to the mortgage market.

This report discusses (1) trends in FHA's share of the market for home
purchase mortgages from 1996 through 2005, and how they compared with the
trends for other market segments; and (2) factors associated with the
trends in FHA's market share and the implications of these trends for
homebuyers and FHA. To address these objectives, GAO analyzed FHA and Home
Mortgage Disclosure Act (HMDA) data and interviewed officials from FHA and
other mortgage institutions.

From 1996 through 2005, FHA's share of the market for home purchase
mortgages in terms of numbers of loans declined 13 percentage points (from
19 to 6 percent), while the prime and subprime shares grew 3 and 13
percentage points, respectively (see figure). The agency experienced a
sharp decrease among populations where it traditionally has had a strong
presence. For example, FHA's market share dropped 25 percentage points
(from 32 to 7 percent) among minority borrowers and 16 percentage points
(from 26 to 10 percent) among low- and moderate-income borrowers. At the
same time, subprime market share among these groups rose dramatically.

The decline in FHA's market share was associated with a number of factors
and has been accompanied by higher ultimate costs for certain conventional
borrowers and a worsening in indicators of credit risk among FHA
borrowers. More specifically, (1) FHA's product restrictions and lack of
process improvements relative to the conventional market and (2) product
innovations and expanded loan origination and funding channels in the
conventional market--coupled with interest rate and house price
changes--provided conditions that favored conventional over FHA-insured
mortgages. In contrast to FHA-insured loans, the majority of conventional
subprime loans had higher ultimate costs to borrowers, partly because
their initial low interest rates could increase substantially in a short
period of time.

Relatively high default and foreclosure rates for subprime mortgages and a
contraction of this market segment could shift market share to FHA. The
extent to which this occurs will depend partly on the ability of FHA and
other market participants to offer mortgage alternatives to borrowers
considering or struggling to maintain higher-priced subprime loans.

Market Shares for Home Purchase Loans, 1996-2005

References

Visible links
  27. http://www.gao.gov/cgi-bin/getrpt?GAO-07-708
  28. http://www.gao.gov/cgi-bin/getrpt?GAO-07-708
  29. http://www.gao.gov/cgi-bin/getrpt?GAO-06-435
  30. http://www.gao.gov/cgi-bin/getrpt?GAO-06-1021
  31. http://www.gao.gov/cgi-bin/getrpt?GAO-06-24
  32. http://www.gao.gov/
  33. http://www.gao.gov/
  34. http://www.gao.gov/fraudnet/fraudnet.htm
  35. mailto:[email protected]
  36. mailto:[email protected]
  37. mailto:[email protected]
  38. http://www.gao.gov/cgi-bin/getrpt?GAO-07-645
  39. http://www.gao.gov/cgi-bin/getrpt?GAO-07-645
*** End of document. ***