Homeownership: Potential Effects of Reducing FHA's Insurance Coverage for
Home Mortgages (Chapter Report, 05/01/97, GAO/RCED-97-93).
Pursuant to a congressional request, GAO provided information on the
impact of a reduction in the Federal Housing Administration's (FHA)
mortgage insurance coverage on: (1) home mortgage lenders, the home
mortgage market, and the types of borrowers served by FHA; (2) the
financial condition of FHA's insurance fund; and (3) the Department of
Housing and Urban Development's (HUD) Government National Mortgage
Association.
GAO noted that: (1) if FHA's insurance coverage is reduced and lenders
become responsible for the risk associated with the uninsured portion of
loans, lenders will likely make fewer and more costly FHA loans; (2) the
general consensus of a HUD-sponsored lender focus group was that the
number of FHA-insured loans would fall by 28 percent and that interest
rates would increase by one-quarter to one-half of a percent; (3)
although some decrease in volume and increase in interest rates would be
likely, GAO's analyses indicate that the changes would likely not be as
great as those that the focus group predicted; (4) nevertheless, any
reduction in the volume of loans and increase in interest rates is
likely to disproportionately affect higher-risk borrowers, low-income,
first time, and minority borrowers and those individuals purchasing
older homes, the types of borrowers frequently served by FHA; (5)
although uncertainty is associated with any forecast, the federal
government would likely improve the financial health of the Fund, by
lowering its exposure to financial losses, if FHA's insurance coverage
were reduced, according to GAO's analyses; (6) this would likely occur
in part because FHA would be liable for only a portion of the losses on
loans that go to foreclosure and therefore would suffer lower financial
losses than it would under full insurance coverage; (7) decreasing FHA's
insurance coverage would have an even greater positive impact on the
Fund's capital reserve ratio if future economic conditions are worse
than the conditions assumed in GAO's baseline scenario; (8) reducing
FHA's insurance coverage might shift some losses from FHA to Ginnie Mae;
(9) reducing FHA's insurance coverage might increase costs for Ginnie
Mae if more lenders were unable to make payments to investors because
they could not shoulder their portion of the losses on defaulted
FHA-insured loans; and (10) however, if lenders respond to a reduction
in FHA's insurance coverage by taking steps to maintain their financial
position, such as targeting FHA-insured loans away from the riskiest
borrowers and increasing interest rates, the impact on Ginnie Mae's
losses would likely be lessened.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: RCED-97-93
TITLE: Homeownership: Potential Effects of Reducing FHA's
Insurance Coverage for Home Mortgages
DATE: 05/01/97
SUBJECT: Lending institutions
Mortgage interest rates
Mortgage loans
Mortgage-backed securities
Mortgage programs
Federal aid for housing
Mortgage protection insurance
Loan defaults
Economic analysis
Losses
IDENTIFIER: FHA Single-Family Mortgage Insurance Program
Mutual Mortgage Insurance Fund
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO report. Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved. Major **
** divisions and subdivisions of the text, such as Chapters, **
** Sections, and Appendixes, are identified by double and **
** single lines. The numbers on the right end of these lines **
** indicate the position of each of the subsections in the **
** document outline. These numbers do NOT correspond with the **
** page numbers of the printed 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. **
** **
** A printed copy of this report may be obtained from the GAO **
** Document Distribution Center. For further details, please **
** send an e-mail message to: **
** **
** **
** **
** with the message 'info' in the body. **
******************************************************************
Cover
================================================================ COVER
Report to the Chairman, Subcommittee on Housing and Community
Opportunity, Committee on Banking and Financial Services, House of
Representatives
May 1997
HOMEOWNERSHIP - POTENTIAL EFFECTS
OF REDUCING FHA'S INSURANCE
COVERAGE FOR HOME MORTGAGES
GAO/RCED-97-93
Homeownership
(365641)
Abbreviations
=============================================================== ABBREV
FHA - Federal Housing Administration
GAO - General Accounting Office
GinnieMae - Government National Mortgage Association
HUD - Department of Housing and Urban Affairs
LTV - loan-to-value (ratio)
MBA - Mortgage Bankers Association
SAMS - Single-Family Accounting Management System
VA - Department of Veterans Affairs
Letter
=============================================================== LETTER
B-276382
May 1, 1997
The Honorable Rick A. Lazio
Chairman, Subcommittee on Housing
and Community Opportunity
Committee on Banking
and Financial Services
House of Representatives
Dear Mr. Chairman:
This report responds to your request that we analyze and present
certain information on the implications of limiting the portion of
single-family home mortgages insured by the Department of Housing and
Urban Development's (HUD) Federal Housing Administration (FHA).
Specifically, the report discusses the impact of a reduction in FHA's
insurance coverage on (1) home mortgage lenders, the home mortgage
market, and the types of borrowers served by FHA; (2) the financial
condition of FHA's insurance fund; and (3) HUD's Government National
Mortgage Association.
As agreed with your office, unless you publicly announce its contents
earlier, we plan no further distribution of this report until 30 days
from the date of this letter. At that time we will send copies of
this report to the Secretaries of Housing and Urban Development and
Veterans Affairs. We will make copies available to others on
request.
Please call me at (202) 512-7631 if you or your staff have any
questions. Major contributors to this report are listed in appendix
V.
Sincerely yours,
Judy A. England-Joseph
Director, Housing and Community
Development Issues
EXECUTIVE SUMMARY
============================================================ Chapter 0
PURPOSE
---------------------------------------------------------- Chapter 0:1
Through its Federal Housing Administration (FHA), the Department of
Housing and Urban Development (HUD) insures private lenders against
nearly all losses resulting from foreclosures on single-family homes
insured under HUD's Mutual Mortgage Insurance Fund (Fund). Although
FHA has always received enough in premiums from borrowers and other
revenues to more than cover these losses, losses totaled about $12.8
billion in 1994 dollars, or about $24,400 for each foreclosed and
subsequently sold single-family home over the 19-year period ending
in 1993. The Department of Veterans Affairs (VA) also operates a
single-family mortgage guaranty program. However, unlike FHA, VA
covers only 25 to 50 percent of the original loan amount against
losses incurred when borrowers default on loans, leaving lenders
responsible for any remaining losses. Virtually all FHA and VA
single-family mortgages are pooled into securities that are then sold
to investors with the help of HUD's Government National Mortgage
Association (Ginnie Mae).
Concerned about the financial health of the Fund and its exposure to
losses when borrowers default on the loans it has insured, the
Chairman of the Subcommittee on Housing and Community Opportunity,
House Committee on Banking and Financial Services, asked GAO to
address the following three questions about the implications of
limiting the insured portion of future FHA-insured loans to that used
by VA: (1) How are lenders and the market expected to react to an
increased risk of loss to lenders and how will this affect FHA's
borrowers? (2) What potential financial impact will reducing FHA's
coverage have on the Fund under different economic conditions? (3)
What are the potential impacts of reducing FHA's insurance coverage
on Ginnie Mae?
BACKGROUND
---------------------------------------------------------- Chapter 0:2
FHA and VA are the principal providers of federally backed mortgage
insurance. Together, they insured 46 percent of all the insured
mortgages originated in 1995; FHA insured 32 percent and VA insured
14 percent.\1 FHA-insured single-family mortgages in the Fund were
valued at about $337 billion as of September 30, 1996. Although
eligibility for VA's program is limited to U.S. veterans and their
families and to certain active duty military personnel, FHA's program
may be used by any household that meets financial underwriting and
loan amount requirements. However, a primary goal of FHA is to
assist households who may be underserved by the private market.
Home mortgage lenders usually require mortgage insurance when a home
buyer has a down payment of less than 20 percent of the value of the
home because defaults are more likely on these loans than on loans
with greater down payments. If an FHA-insured mortgage goes to
foreclosure, FHA pays the lender for virtually all of the losses
associated with the property and then almost always takes possession
of and subsequently sells the property. VA, on the other hand, has
the following two options if a mortgage it has insured goes to
foreclosure: (1) paying the lender the insurance due and leaving the
property with the lender or (2) purchasing the property from the
lender and reselling it. For example, if the cost of paying the
insurance claim is less than the estimated costs of taking possession
of the property and reselling it, VA will choose the former--this is
referred to as a VA no-bid. If VA chooses the "no-bid" option, the
limit on the amount that VA guarantees against loss is from 25 to 50
percent of the mortgage's balance, depending on the original loan
amount. The partial VA guaranty allows VA the option of leaving a
foreclosed property with the lender. Similarly, a reduction in FHA's
insurance coverage could result in FHA leaving foreclosed properties
with lenders and paying only the insured portion of the mortgage's
balance.
FHA's primary single-family mortgage program currently requires no
federal funds to operate. The Fund, which supports this program, is
required by law to contain sufficient reserves and funding to cover
the estimated future payment of claims on foreclosed mortgages and
other costs. Cash flows into the Fund from insurance premiums and
from the sale of foreclosed property. According to Price
Waterhouse's most recent study of the Fund's financial soundness
issued in February 1997, at the end of fiscal year 1996, the Fund had
an estimated economic value/reserves of about $9.4 billion and an
estimated capital reserve ratio of 2.54 percent--a ratio greater than
the 2-percent capital reserve ratio target set by the Congress for
fiscal year 2000.\2 The reserves in the Fund have always been more
than enough to cover the expenses incurred. However, if the Fund
were to deplete its reserves, the U.S. Treasury would have to
directly cover lenders' claims and other costs of the program, by
law. On the other hand, VA's program is not required to be
self-sustaining; VA's program receives an appropriation of federal
funds each year.
In 1995, virtually all FHA- and VA-insured mortgages were pooled into
securities that were then sold to investors with the help of Ginnie
Mae, which, like FHA, is a part of HUD. Under Ginnie Mae's program,
issuers of securities backed by pooled mortgages sell the securities
to investors. Issuers can be mortgage bankers, savings institutions,
or other financial intermediaries. The investors that buy the
securities are guaranteed by Ginnie Mae to receive timely principal
and interest payments, regardless of the performance of the mortgages
backing the securities or the issuer's performance. Ginnie Mae's net
revenues (after expenses) totaled $515 million in fiscal year 1996.
Ginnie Mae derived its revenues from interest on U.S. securities and
fees collected from lenders to cover its costs and offset its future
payments of claims under the guaranty representing its major sources
of revenues.
The actual impact of a reduction in insurance coverage on the
financial condition of the Fund will depend on future economic
conditions. Since uncertainty always exists when forecasting future
economic conditions, especially over a long period of time, GAO
prepared estimates of the financial impact on the Fund under three
different economic scenarios (a baseline, an optimistic, and a
pessimistic scenario), assuming a different rate of appreciation in
the price of homes over the next 30 years for each economic
scenario.\3
--------------------
\1 The remaining insured mortgages were insured by private companies,
and a very small portion (less than 1 percent) was insured by the
Department of Agriculture's Rural Housing Service. Although VA
actually guarantees rather than insures mortgages, this report uses
the term "mortgage insurance" to refer to the guaranty provided by VA
as well as the insurance provided by FHA, the Rural Housing Service,
and private mortgage insurers.
\2 The capital reserve ratio is the economic value/reserves of the
Fund expressed as a percentage of the outstanding principal balance
of FHA-insured loans (insurance-in-force). The economic
value/reserves of the Fund are the current assets available to the
Fund, plus the net present value of all future cash inflows and
outflows expected to result from mortgages insured under the Fund.
\3 Although future economic conditions are uncertain, GAO placed
greater emphasis in this report on its baseline economic scenario
because it assumes slightly lower house price appreciation than the
rates forecasted by DRI/McGraw-Hill, a private economic forecasting
company that GAO used in developing its economic scenarios. GAO's
baseline scenario assumes house price appreciation rates to be 1
percentage point lower than DRI's forecast. GAO's optimistic and
pessimistic economic scenarios assume a house price appreciation of 2
percentage points higher or lower than the baseline, respectively.
RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3
If FHA's insurance coverage is reduced and lenders become responsible
for the risk associated with the uninsured portion of loans, lenders
will likely make fewer and more costly FHA loans. The general
consensus of a HUD-sponsored lender focus group was that the number
of FHA-insured loans would fall by about 28 percent and that interest
rates would increase by one-quarter to one-half of a percent.\4
Although some decrease in volume and increase in interest rates would
be likely, GAO's analyses indicate that the changes would likely not
be as great as those that the focus group predicted. This is because
the higher revenues earned by lenders from a one-quarter percent
increase in interest rates would likely produce more than enough
revenue to lenders to cover any losses that they would incur if FHA's
insurance coverage were reduced. Moreover, according to VA
officials, there was no discernible reduction in the volume of VA
loans after the Department's current policy of limiting losses on
each property to the guarantee amount--referred to as the "no-bid"
policy--went into effect in 1985. Nevertheless, any reduction in the
volume of loans and increase in interest rates is likely to
disproportionately affect higher-risk borrowers--low-income,
first-time, and minority borrowers and those individuals purchasing
older homes--the types of borrowers frequently served by FHA.
Although uncertainty is associated with any forecast, the federal
government would likely improve the financial health of the Fund--by
lowering its exposure to financial losses--if FHA's insurance
coverage were reduced, according to GAO's analyses. Specifically,
the ratio of the capital reserves/economic value in the Fund to its
outstanding insurance liabilities--the capital reserve ratio--would
likely increase over time if FHA provided only the partial coverage
currently used by VA. This would likely occur in part because FHA
would be liable for only a portion of the losses on loans that go to
foreclosure and therefore would suffer lower financial losses than it
would under full insurance coverage. The losses thus avoided by
partial insurance coverage would more than offset the premium income
lost from a reduction in the volume of FHA loans. Decreasing FHA's
insurance coverage would have an even greater positive impact on the
Fund's capital reserve ratio if future economic conditions are worse
than the conditions assumed in GAO's baseline scenario. This would
likely occur because FHA would face a greater number of insurance
claims under adverse economic conditions and, in turn, the reduction
in claims payments with partial insurance coverage rather than with
full coverage would be greater.
Reducing FHA's insurance coverage might shift some losses from FHA to
Ginnie Mae. When lenders who have issued securities backed by pools
of FHA- and VA-insured mortgages are unable to pay investors in these
securities, Ginnie Mae steps in and provides the investor with
payments. Reducing FHA's insurance coverage might increase costs for
Ginnie Mae if more lenders were unable to make payments to investors
because they could not shoulder their portion of the losses on
defaulted FHA-insured loans. However, if lenders respond to a
reduction in FHA's insurance coverage by taking steps to maintain
their financial position, such as targeting FHA-insured loans away
from the riskiest borrowers and increasing interest rates, the impact
on Ginnie Mae's losses would likely be lessened.
--------------------
\4 In October 1995, HUD held a lender focus group to identify
lenders' anticipated responses to a reduction in FHA's insurance
coverage. The lenders represented in the focus group were
responsible for 35 to 40 percent of all FHA-insured loans in 1994.
GAO'S ANALYSIS
---------------------------------------------------------- Chapter 0:4
LENDERS ARE LIKELY TO MAKE
FEWER AND MORE COSTLY LOANS
TO HIGHER-RISK BORROWERS IF
FHA'S COVERAGE IS REDUCED
-------------------------------------------------------- Chapter 0:4.1
Lenders are likely to reduce the number of FHA-insured loans made
each year by tightening underwriting standards and increasing
interest rates if FHA reduces the portion of losses it will cover on
loans that go to foreclosure. The lender focus group convened by HUD
indicated that an interest rate increase of one-quarter to one-half a
percentage point for FHA-insured loans would be necessary to
compensate them for the additional risk they would face.\5
According to GAO's analysis, an interest rate increase of one-quarter
percentage point--the lowest figure in the range estimated by the
focus group--would produce more than enough revenue to cover lenders'
estimated losses, even under the most adverse economic conditions
experienced by FHA in the last 20 years. The focus group also
projected that FHA's volume of loans would fall by 28 percent if its
insurance coverage were reduced. However, VA officials told GAO that
there was no discernible reduction in the volume of VA loans after
VA's current no-bid policy went into effect in 1985. If lenders do
not increase interest rates as much as the focus group predicted or
if the rate increase is targeted to higher-risk borrowers, as is
likely, the volume of loans made by FHA following a coverage
reduction would likely fall by less than 28 percent.
Regardless of the size of the reduction in FHA's loan volume,
reducing FHA's insurance coverage would likely have a greater effect
on low-income, first-time, and minority borrowers, as well as on
individuals purchasing older homes, than on other borrowers with
better access to private mortgage credit. Lenders indicated that
they would probably tighten underwriting standards for FHA-insured
loans to target them away from borrowers in higher-risk categories.
GAO's analysis indicates that, all else held equal, the loss rate
and/or rate of foreclosure for minorities, buildings with multiple
living units, loans written at higher interest rates, older
properties, and smaller loans are higher, suggesting that those are
the types of loans that lenders would likely refrain from making.\6
Reducing FHA's insurance coverage would also lessen FHA's ability to
stabilize local housing markets during regional economic downturns.
--------------------
\5 An interest rate increase of one-quarter to one-half percent on a
$100,000, 30-year loan, originated at 8 percent, would increase
monthly mortgage payments by $17 to $35, respectively.
\6 Buildings with multiple living units are homes containing two to
four living units, such as duplexes. Provided that they contain no
more than four units, these homes are eligible for mortgage insurance
through FHA's primary program.
REDUCING FHA'S INSURANCE
COVERAGE WOULD LIKELY
INCREASE THE FUND'S CAPITAL
RESERVE RATIO
-------------------------------------------------------- Chapter 0:4.2
The capital reserve ratio of the Fund would likely increase if
insurance coverage levels were reduced. For each partial coverage
scenario that GAO examined, the estimated capital reserve ratio for
loans made by FHA in fiscal year 1995 is at least 25 percent higher
than GAO's estimate of that ratio under full coverage. This higher
ratio would occur because FHA would cover a smaller percentage of the
losses on each foreclosed loan and because lenders would likely make
fewer FHA-insured loans to higher-risk borrowers, thus exposing FHA
to less risk than with full coverage. The capital reserve ratio is
likely to increase regardless of whether FHA implements a partial
coverage schedule identical to VA's or uses one of the other partial
coverage options that GAO examined.\7
The impact of reducing FHA's insurance coverage on the estimated
economic value/reserves of the loans that FHA will insure in the
future is less certain than the impact of reducing FHA's insurance
coverage on the capital reserve ratio, assuming baseline economic
conditions. The economic value of FHA loans is important for various
reasons. First, the estimated economic value represents capital
reserves that, when divided by the outstanding principal amount of
FHA loans, equal the estimated capital reserve ratio. Second, the
economic value provides an estimate of how profitable these loans are
for FHA, which is important because estimated increases in economic
value due to legislative changes allow additional mandatory spending
authorizations to be made, other revenues to be reduced, or projected
deficits in the federal budget to be reduced.
If FHA's insurance coverage is reduced, the change in the economic
value of FHA loans will depend largely on the way that lenders
respond. The smaller the reduction in FHA-insured loans that lenders
make and the greater their ability to target FHA loans to less-risky
borrowers, the greater the positive effect on the economic value of
the new loans insured. GAO's estimates under baseline economic
conditions of the economic value of the loans made by FHA in fiscal
year 1995, assuming partial insurance coverage, range from about $29
million less than to $150 million greater than GAO's full coverage
estimate of $536 million.
Under GAO's pessimistic economic scenario, which assumes house price
rates of appreciation of 2 percentage points lower than GAO's
baseline economic scenario, reducing FHA's insurance coverage could
have a much greater positive effect on FHA's economic value and the
resulting capital reserve ratio. GAO estimates that a reduction in
FHA's insurance coverage could increase economic value by as much as
$150 million under baseline economic conditions, compared with an
increase in economic value of as much as $199 million under
pessimistic economic conditions.
--------------------
\7 In addition to analyzing the impact of reducing FHA's insurance
coverage to VA's current level of coverage, GAO also examined two
other alternatives--(1) the impact of retaining full FHA insurance
coverage for first-time home buyers and VA coverage for all repeat
buyers and (2) reducing FHA's insurance coverage for all new FHA
loans but to a level slightly higher than VA's.
REDUCING FHA'S INSURANCE
COVERAGE MIGHT SHIFT SOME
LOSSES TO GINNIE MAE
-------------------------------------------------------- Chapter 0:4.3
Reducing FHA's insurance coverage might shift some losses from FHA to
Ginnie Mae. Ginnie Mae is ultimately responsible for ensuring that
investors receive timely principal and interest payments on
securities backed by pools of FHA and VA mortgages. Ginnie Mae's
costs increased after changes in VA's "no-bid" procedures took effect
in the 1980s, in part because lenders were left to cover a portion of
the losses on VA-guaranteed loans more often than they had in the
past, causing some lenders to default.\8 When lenders default, Ginnie
Mae takes over the lenders' entire portfolio of FHA and VA loans that
have been pooled into securities and becomes responsible for making
timely payments to investors of any shortfalls in principal and
interest payments. In addition, Ginnie Mae incurs costs for paying
taxes and insurance, as well as costs associated with acquiring,
managing, and disposing of portfolio properties.
The extent to which Ginnie Mae's losses might increase depends
largely on lenders' responses to a reduction in FHA's insurance
coverage. If lenders respond by very carefully targeting FHA-insured
loans away from the most-risky borrowers, fewer loans will default,
and lenders will be in a better position to cover their share of the
losses for the FHA-insured loans that default anyway. The increase
in losses for Ginnie Mae will then be lower than if lenders did not
target loans as carefully. Similarly, if lenders increase interest
rates on FHA-insured loans, they will have more funds for covering
their share of the losses on the loans that default. Fewer losses
will then be shifted to Ginnie Mae as a result of a reduction in
FHA's insurance coverage.
Reducing FHA's insurance coverage would not likely affect Ginnie
Mae's ability to provide lenders of FHA and VA loans with liquidity
(cash), since Ginnie Mae's securities would continue to maintain the
U.S. Government's full faith and credit guarantee.
--------------------
\8 In 1984, the Congress enacted provisions of the Deficit Reduction
Act to, among other things, reduce the losses associated with
foreclosures on VA home loans. The act required VA, in deciding
whether to pay the VA guaranty on defaulted loans or acquire the
property, to limit its loss to the amount of the guaranty. The act
also required VA, in deciding whether to acquire a property, to
consider post-acquisition costs that were previously excluded from
VA's decision. After the implementation of these provisions in 1985,
the number of VA no-bids increased substantially--rising from 2
percent of all VA foreclosures in 1982 to 24 percent in 1988.
AGENCY COMMENTS
---------------------------------------------------------- Chapter 0:5
GAO provided HUD and VA with a draft of this report for their review
and comment. GAO received written comments on the draft report from
HUD. (See app. IV.) In addition, GAO received comments on the draft
report from Ginnie Mae's Executive Assistant, which are discussed at
the end of chapter 4. VA's Acting Under Secretary for Benefits
provided GAO with a memorandum that clarified changes to information
contained in the report and updated figures with more recent
information. Where appropriate, GAO incorporated VA's clarifications
into the report.
HUD disagreed with some of the conclusions reached by GAO. For
example, HUD stated that it believes, on the basis of its analysis,
which assumed a slightly higher rate of appreciation in house prices
than that assumed in GAO's baseline and a 28-percent reduction in
loan volume, that reducing FHA's insurance coverage will decrease the
economic value of the Fund. HUD also stated that GAO overstates the
private sector's ability to accept the transfer of risk. HUD said
that the increase in price and rationing that will occur is
understated by GAO and that eliminating the catastrophic coverage
feature (insurance against all losses) of FHA insurance could result
in losses that quickly overwhelm the resources of lenders. HUD also
stated that GAO gives inadequate attention to the potential that
FHA's lenders will withdraw from the marketplace in times of economic
downturns under a partial insurance model.
GAO agrees with HUD that under certain combinations of assumptions
about the rate of appreciation in house prices and reductions in loan
volume, the estimated economic value of FHA loans would be less than
under full insurance coverage. In general, the smaller the reduction
in loan volume, the higher the economic value contributed by the
remaining loans insured by FHA, according to GAO's analysis. Under
its baseline scenario,\9 GAO estimates a small ($29 million) decline
in economic value from changing to partial insurance if the loan
volume drops by 28 percent and lenders only loosely target their
application denials toward high-risk categories of borrowers.
Furthermore, under its optimistic economic scenario, which assumes
higher-than-expected rates of house price appreciation, GAO estimates
that the economic value of FHA's fiscal year 1995 loans will likely
decline by about $67 million to $94 million below the value achieved
under full insurance if the volume of loans is reduced by 28 percent.
However, GAO believes that a 28-percent reduction in the volume of
loans may not be likely and that such a fall in the volume of loans
is even more unlikely if house prices increase at a rate greater than
forecasted. Moreover, while both GAO's and HUD's analyses estimate
that economic value would likely be reduced under certain
assumptions, both analyses estimate that under these same
assumptions, the capital ratio will be higher under partial insurance
than under full insurance.
HUD believes that GAO is understating the likely increase in price
(interest rates) resulting from partial insurance. GAO disagrees.
GAO's analysis of the adequacy of the additional revenues that would
be earned by lenders from a one-quarter-percent increase in interest
rates to cover additional losses resulting from partial insurance is
based on the most adverse economic conditions experienced by FHA in
the last 20 years. While it is difficult to quantify the volatility
of losses, GAO's analysis suggests that lenders would likely earn
more than enough revenues to cover losses from a reduction in FHA's
insurance coverage even under extremely adverse circumstances. Under
GAO's baseline economic scenario, this increase in interest rates
would cover losses several times over. As a result, GAO believes
that it is unlikely that interest rates would increase by as much as
one-quarter percentage point in a competitive marketplace. While GAO
agrees that catastrophic losses may overwhelm the resources of some
lenders, it does not believe that such losses would jeopardize FHA.
During the lender focus group meeting, FHA lenders still concluded
that making FHA loans under partial insurance is profitable and that
72 percent of FHA's loan volume would be retained.
In response to HUD's comment on the potential that FHA's lenders
would withdraw from the marketplace in times of economic downturn
under a partial insurance model, GAO has added information describing
the role of FHA as a countercyclical force in the market during times
of economic hardship and the adverse impact that partial insurance
could have on this role. GAO's detailed responses and HUD's entire
comments appear in appendix IV and are summarized and discussed at
the end of chapters 2 through 4.
--------------------
\9 As in previous reports, GAO continues to use a slightly lower rate
of appreciation in house prices than forecasted (1 percentage point)
in its baseline estimate to be conservative in assessing the impact
of reducing FHA's insurance coverage on the capital reserve ratio,
which is a measure of how well the Fund can withstand difficult
economic conditions.
INTRODUCTION
============================================================ Chapter 1
Mortgage insurance helps home buyers with limited down payment funds
to obtain mortgages. It is generally used when a borrower makes a
down payment of less than 20 percent of the value of the home--when
the mortgage has a loan-to-value (LTV) ratio greater than 80 percent.
Most lenders require mortgage insurance for these loans because they
are more likely to default than loans with lower LTV ratios. If a
loan with mortgage insurance defaults, the lender may foreclose on
the loan and collect all or a portion of the losses from the insurer.
Losses generally include the unpaid principal balance and delinquent
interest due on the loan, legal expenses incurred during foreclosure,
the expense of maintaining the home, and any advances that the lender
made to pay taxes or insurance. The primary mortgage insurers are
the Department of Housing and Urban Development's (HUD) Federal
Housing Administration (FHA), the U.S. Department of Veterans
Affairs (VA), and private mortgage insurance companies.\10 Although
FHA operates a number of single-family mortgage insurance programs,
its primary program is the Section 203(b) Single-Family Mortgage
Insurance program.
--------------------
\10 For additional information on the insurance offered and borrowers
served by FHA, VA, and private mortgage insurers, see our August 1996
report entitled Homeownership: FHA's Role in Helping People Obtain
Home Mortgages (GAO/RCED-96-123).
FHA'S INSURANCE COVERAGE IS
HIGHER THAN THAT OF OTHER
INSURERS
---------------------------------------------------------- Chapter 1:1
While FHA's Section 203(b) program protects lenders against nearly
100 percent of the loss associated with a foreclosed mortgage, VA and
private mortgage insurers limit their coverage to a portion of the
mortgage's balance. The amount that VA guarantees against loss is
established by legislation and has periodically increased.
Furthermore, VA's guaranty depends on the loan's original amount, as
shown in table 1.1 below:
Table 1.1
VA's Partial Guarantee Schedule
Loan's original amount VA's guaranty
---------------------------------- ----------------------------------
Less than or equal to $45,000 50% of the loan
Greater than $45,000, $22,500
but not more than $56,250
Greater than $56,250, 40% of the loan or $36,000,
but not more than $144,000 whichever is less
Greater than $144,000 25% of the loan or $50,750,
whichever is less
----------------------------------------------------------------------
When a loan that it has guaranteed goes to foreclosure, VA chooses
between (1) paying the amount of insurance due on the property and
leaving it with the lender or (2) purchasing the property from the
lender, taking possession of it, and reselling it. VA selects the
option that is likely to be more financially advantageous.
Specifically, if the guarantee amount, as calculated according to the
guidelines shown in table 1.1, is less than the net expected costs of
taking possession of the property and reselling it, VA is required by
law to choose the former option--this is generally referred to as a
VA "no-bid" because VA limits its losses by not acquiring the
property.\11 Fewer than 12 percent of VA's foreclosures during fiscal
years 1990 through 1996 were "no-bids," resulting in losses to
lenders. The "no-bid" policy allows VA the option of not bidding on
a foreclosed property, leaving the property with the lender, and
limiting VA's losses to the guaranty amount.
For private mortgage insurers, the type and amount of coverage
selected by the lender determine how much the private mortgage
insurer will pay if the borrower defaults and the lender must
foreclose. Typically, this amount is limited to 20 percent to 30
percent of the losses but can go as high as 35 percent.
--------------------
\11 When determining whether to pay the guaranty amount or acquire
the property, VA calculates the costs of acquiring and reselling the
property by considering the following two primary factors: (1) the
total indebtedness on the property (the outstanding mortgage balance,
interest, and allowable charges) and (2) the net value of the
property (the appraised fair market value minus the administrative
costs of holding and reselling the property, currently estimated at
15.11 percent of the fair market value). If the total indebtedness
minus the guaranty amount (i.e., the unguaranteed portion) is greater
than the net value of the property, VA considers it as a "no-bid"
case and will not acquire the property.
FHA IS THE BIGGEST FEDERAL
INSURER, BUT PRIVATE MORTGAGE
INSURERS SERVE THE MOST HOME
BUYERS
---------------------------------------------------------- Chapter 1:2
FHA serves more homeowners than VA, but private mortgage insurers
account for most of the mortgage insurance market, according to data
on loans insured in 1995.\12
From 1985 through 1995, FHA's share of all loans insured each year
has fluctuated from a low of 29 percent in 1992 to a high of 51
percent in 1987. VA's share during this period stayed at 13 to 20
percent. Private mortgage insurers' share during the same period
fluctuated from a low of 29 percent in 1987 to a high of 57 percent
in 1992. The relative market shares of FHA, VA, and private mortgage
insurers are shown in figure 1.1.
Figure 1.1: All Insured
Mortgages, by Insurer, 1985-95
\ a
(See figure in printed
edition.)
\a This figure does not include data on the small number of loans
insured through the Department of Agriculture's Rural Housing
Service.
Source: GAO's analysis of HUD's data.
--------------------
\12 The Department of Agriculture's Rural Housing Service also
insures a small number of mortgages each year.
CUSTOMERS TARGETED BY FHA AND
VA
---------------------------------------------------------- Chapter 1:3
Although a primary goal of FHA is to assist borrowers who may be
underserved by the private market, any household who takes out a loan
no greater than FHA's maximum loan amount and who meets other
financial qualifications is eligible to obtain FHA's mortgage
insurance.\13 VA's insurance is available only to U.S. veterans,
their families, and certain active duty military personnel. Like FHA
insurance, private mortgage insurance is available to any household
who meets the insurer's loan guidelines and financial qualifications.
--------------------
\13 The maximum loan amount permitted under FHA's program for
single-family homes in the highest cost areas in the continental
United States is currently set at $160,950.
FHA'S PROGRAM IS FINANCIALLY
SELF-SUFFICIENT, BUT VA'S
PROGRAM IS NOT
---------------------------------------------------------- Chapter 1:4
Unlike VA, FHA requires no federal funds to carry out its primary
single-family mortgage insurance program. Borrowers who obtain
FHA-insured mortgage loans pay insurance premiums, which are
deposited into HUD's Mutual Mortgage Insurance Fund (Fund).
According to studies conducted by GAO and Price Waterhouse, the
Fund's reserves exceed those needed to meet the legislatively
prescribed capital reserve ratio.\14 Price Waterhouse reported that
the Fund had an estimated capital reserve ratio of 2.54 percent at
the end of fiscal year 1996--a ratio greater than the 2-percent
target set in 1990 by the Congress for fiscal year 2000. Capital
reserve ratio estimates are computed by presenting the net present
value of all future cash inflows and outflows expected to result from
the mortgages insured by FHA (the economic value of the Fund) as of
the end of a fiscal year, plus accumulated reserves, as a percentage
of the total insurance-in-force at that time. Price Waterhouse
estimated that by 2000, this ratio would be 3.57 percent. In
addition to the capital reserve ratio, Price Waterhouse estimated
that the economic value of the loans insured by FHA would increase
from $9.4 billion at the end of fiscal year 1996 to $14.8 billion by
the end of fiscal year 2000. Economic value provides an estimate of
the profitability of FHA loans, which is important because estimated
increases in economic value due to legislative changes allow
additional mandatory spending authorizations to be made, other
revenues to be reduced, or projected deficits in the federal budget
to be reduced.
Although the most recent estimate of the Fund's capital reserve ratio
exceeds the 2-percent legislative target, the Fund experienced
substantial losses during the 1980s. This occurred primarily because
foreclosure rates on single-family homes supported by the Fund were
high in economically stressed regions. To help place the Fund on a
financially sound basis, legislative reforms such as requiring FHA
borrowers to pay more in insurance premiums were made in 1990.
VA's program requires an annual federal appropriation because
premiums collected from borrowers of VA-insured mortgages do not
cover the estimated future losses for these mortgages. In 1996, VA
received a credit subsidy of $697 million for covering costs
associated with its mortgage insurance operations.
--------------------
\14 Mortgage Financing: FHA Has Achieved Its Home Mortgage Capital
Reserve Target (GAO/RCED-96-50, Apr. 12, 1996) and An Actuarial
Review for Fiscal Year 1996 of the Federal Housing Administration's
Mutual Mortgage Insurance Fund: Final Report, Price Waterhouse LLP
(Feb. 14, 1997).
GINNIE MAE CONNECTS FEDERALLY
INSURED MORTGAGES WITH
SECONDARY MARKET INVESTORS
---------------------------------------------------------- Chapter 1:5
Almost all mortgages insured by FHA and VA are pooled into securities
that are then sold to investors with the help of HUD's Government
National Mortgage Association (Ginnie Mae). Issuers of Ginnie Mae
securities can be mortgage bankers, savings institutions, or other
financial intermediaries. The investors that buy the securities are
guaranteed by Ginnie Mae to receive timely principal and interest
payments, regardless of the performance of the mortgages backing the
securities or the issuer's performance. At the end of fiscal year
1996, Ginnie Mae had outstanding guarantees of mortgage-backed
securities totaling $533 billion.
OBJECTIVES, SCOPE, AND
METHODOLOGY
---------------------------------------------------------- Chapter 1:6
To obtain information on the impact of reducing FHA's insurance
coverage, the Chairman, Subcommittee on Housing and Community
Opportunity, House Committee on Banking and Financial Services, asked
us to address the following three questions: (1) How are lenders and
the market expected to react to an increased risk of loss to lenders
and how will this affect FHA's borrowers? (2) What potential
financial impact will reducing FHA's coverage have on the Fund under
different economic conditions? (3) What are the potential impacts of
reducing FHA's insurance coverage on Ginnie Mae?
To identify lenders' anticipated responses to a reduction in FHA's
insurance coverage, we reviewed Price Waterhouse's summary of the
results of a focus group held in October 1995 by HUD and interviewed
officials from four mortgage-lending institutions. The lenders
represented in the focus group were responsible for approximately 35
to 40 percent of all FHA-insured loans in 1994. The four lenders
that we interviewed were among the largest FHA lenders in four
regions. One of these lenders is the largest FHA lender nationwide,
and 1 was among the 14 lenders in HUD's focus group.
We analyzed the reasonableness of the interest rate increase and loan
volume decrease predicted by the lenders that participated in HUD's
focus group. Specifically, we analyzed the group's estimated
interest rate increase by comparing the increase in lenders' revenues
likely to result from the rate increase with the increase in lenders'
losses likely to result from a reduction in FHA's insurance coverage.
Appendix I presents a more detailed discussion of this analysis. We
analyzed the volume decrease predicted by the focus group by
considering the following factors: (1) the potential effect of an
interest rate increase on the number of FHA-insured loans made, (2)
the likelihood that lenders would target an interest rate increase
toward higher-risk borrowers, and (3) the impact of tighter
underwriting standards on FHA's volume of loans in the long run.
To obtain information on the types of borrowers that would most
likely be affected by an insurance coverage reduction, we obtained
information from VA and FHA on their mortgage insurance programs and
the types of borrowers they currently serve. We also used HUD's data
on foreclosed properties and on the demographic characteristics of
FHA's borrowers to conduct a regression analysis. Specifically, we
analyzed the effects of several variables on total loss rates and
losses to the lender.\15 Appendix II presents a more detailed
discussion of this analysis.
To estimate the impact that reducing FHA's insurance coverage to VA
levels would have on the estimated capital reserve ratio of the Fund,
we used an econometric model of FHA's loan terminations and a model
of the cash flows associated with FHA mortgages, both of which had
been developed for a previous GAO report.\16
Specifically, we used these models to estimate the impact of reducing
FHA's insurance coverage on the estimated economic value and the
resulting capital reserve ratio for loans insured by FHA in fiscal
year 1995 under three different economic scenarios. Our baseline
economic scenario assumes that house price appreciation rates are 1
percentage point lower than those in forecasts prepared by
DRI/McGraw-Hill, a private economic-forecasting company. Our
pessimistic economic scenario assumes that house price appreciation
rates are 2 percentage points lower than our baseline. Our
optimistic economic scenario assumes that house price appreciation
rates are 2 percentage points higher than our baseline. Although
future economic conditions are uncertain, to be conservative in
assessing the impact of reducing FHA's insurance coverage on the
capital reserve ratio and to be consistent with past analyses that we
have made of the actuarial condition of FHA's Fund, we placed greater
emphasis on the estimates prepared under the baseline economic
scenario. Our estimates of the impact of reducing FHA's insurance
coverage on the Fund's capital reserve ratio are based on the
assumption that FHA's current premium structure would not change
following a reduction in FHA's insurance coverage. Furthermore,
because lenders' responses to a reduction in insurance coverage play
an important role in the impact of a reduction on FHA's Fund, we also
prepared estimates assuming several different FHA loan volume
reduction scenarios.
Our econometric analysis estimated the historical relationships
between the probability of a loan's foreclosure and prepayment and
key explanatory factors, such as the borrower's equity and the
interest rate. Using our estimates of these relationships, we
developed forecasts of future loan performance to estimate the
economic net worth and resulting capital reserve ratio for FHA loans
originated in fiscal year 1995 under the three economic scenarios.
We selected loans made in fiscal year 1995 because it was the most
recent year for which complete data were available and because any
change in FHA's insurance coverage would apply only to new
FHA-insured loans and not to FHA's entire portfolio of existing
loans. The econometric model and equations used to forecast the
impact on the fiscal year 1995 loans were developed on the basis of
historical data on the performance of FHA loans originated from
fiscal year 1975 through fiscal year 1994.
We looked primarily at the impact of reducing FHA's coverage to VA's
levels. However, we also looked at the impact of two alternative
scenarios for reducing FHA's coverage: (1) retaining full coverage
for first-time home buyers but reducing coverage to VA's levels for
repeat buyers and (2) imposing a graduated coverage schedule that
provides slightly higher coverage levels than VA's program for loans
of low and moderate size--60 percent coverage for the lowest loan
amounts, declining to 30 percent for the highest loan amounts. These
scenarios are discussed in greater detail in appendix III.
We obtained data from FHA's property disposition database--the
Single-Family Accounting Management System (SAMS)--so that we could
examine the division of foreclosed loan losses between FHA and
mortgage lenders if FHA were to change from full to partial
insurance. We calculated the loss rate and determined what fraction
of the loss rate would be paid by FHA, recognizing that the remainder
would be paid by the lender, under a partial insurance program like
VA's and assuming a no-bid type policy in effect for FHA. Appendix
II presents a detailed explanation of this analysis.
To obtain information on the potential impacts on Ginnie Mae, we met
with Ginnie Mae officials and reviewed an analysis conducted by Price
Waterhouse for them on the likely results of reducing FHA's insurance
coverage.
Our analysis of the implications of limiting the insured portion of
future FHA-insured loans to that used by VA did not include assessing
the implications of such a change on FHA's staffing levels.
We performed our work from May 1996 through March 1997 in accordance
with generally accepted government auditing standards. HUD's
comments on a draft of this report and our responses to them are
included in chapters 2 through 4 and in appendix IV.
--------------------
\15 We defined loss rate as the loss (or profit) on a foreclosed
property divided by the acquisition cost. We defined acquisition
cost as the claim paid by FHA plus any extra costs incurred for FHA's
acquisition of the property. We defined loss to the lender as any
loss in excess of the amount that would be covered under VA's
guarantee limits.
\16 See our April 1996 report entitled Mortgage Financing: FHA Has
Achieved Its Home Mortgage Capital Reserve Target (GAO/RCED-96-50).
IF FHA'S INSURANCE COVERAGE IS
REDUCED, LENDERS MAY MAKE FEWER
AND HIGHER INTEREST LOANS TO SOME
TYPES OF BORROWERS
============================================================ Chapter 2
Since its inception in 1934, FHA's single-family mortgage insurance
program has protected private lenders against nearly all losses
resulting from foreclosures. Losses incurred by FHA have averaged
about $24,400 for each foreclosed and subsequently sold defaulted
single-family property.\17 These losses were offset by insurance
premiums paid by FHA's borrowers, not by funds from the U.S.
Treasury. For home loans made by private lenders that are guaranteed
by other federal agencies such as VA or made in the conventional
market and insured by private mortgage insurers, lenders are not
fully protected against losses.
A reduction in FHA's insurance coverage would increase lenders'
vulnerability to risks associated with the uninsured portion of
loans. Some FHA lenders and other mortgage-lending industry
representatives believe that a reduction in FHA's insurance coverage
will reduce the number of FHA mortgage loans made and increase
interest rates on them as lenders try to compensate for losses on
loans that go to foreclosure. While such market responses by lenders
are expected, we question whether the volume of FHA-insured loans
would decline as much and whether interest rates would increase as
much as anticipated by FHA's lenders. VA officials told us that
there was no discernible reduction in the volume of VA loans after
its current partial insurance policy went into effect in 1985. While
it is difficult to predict precisely such market responses, any
reduction in the volume of loans and increase in interest rates would
probably be borne disproportionately by low-income, first-time, and
minority homeowners and those individuals purchasing older homes
because these are the types of borrowers that are frequently served
by FHA and that many housing advocates believe are not being
completely served by the private market. For such borrowers, FHA
represents a significant source of mortgage insurance. For this
reason, reducing FHA's insurance coverage may make it more difficult
and, in some cases, more costly for some potential home buyers to
purchase a home. While some potential home buyers may be ruled out,
others may have to delay their purchase of a home or reduce the value
of the home purchased. Any reduction in FHA's insurance coverage
would shift risk to originators of FHA mortgages, causing possible
changes in the lending industry, such as mergers of small FHA
lenders. Reducing FHA's insurance coverage could also lessen FHA's
ability to stabilize local housing markets when regional economies
decline.
--------------------
\17 During the 19-year period ending September 30, 1993, FHA incurred
losses totaling about $12.8 billion in 1994 dollars following the
foreclosure and subsequent sale of about 525,000 defaulted
single-family housing loans that FHA had insured.
MORTGAGE-LENDING INDUSTRY
OFFICIALS ANTICIPATE A
REDUCTION IN THE VOLUME OF
LOANS AND AN INCREASE IN
INTEREST RATES
---------------------------------------------------------- Chapter 2:1
To assess the impact that an FHA partial guarantee would have on the
mortgage markets and borrowers, FHA sponsored a meeting of 14 FHA
lenders in October 1995 to obtain their views. Price Waterhouse was
retained by FHA and Ginnie Mae to gather the lenders' responses.
According to an FHA official, collectively, the 14 lenders accounted
for 35 to 40 percent of all FHA loans made nationwide. The lenders
indicated that if FHA implements a partial insurance guarantee, they
would (1) charge FHA's borrowers between one-quarter to one-half
percent more in interest and (2) establish stricter underwriting
standards that would limit the number of higher-risk borrowers who
would qualify for an FHA-insured loan. The lenders believe that
these two market responses would be necessary to offset losses they
might incur if a borrower defaulted on a loan and to compensate them
for the additional risk.
The lenders estimated that these two market responses would cause the
number of FHA-insured loans they make to decline by 28 percent under
normal economic conditions and more in the event of an economic or
natural disaster in a particular geographical area.\18 These lenders
concluded that 18 percent of the 28-percent volume reduction would be
due to higher-risk borrowers who, because of the higher costs and
stricter underwriting standards, would not be able to obtain an
FHA-insured loan, while the remaining 10-percent reduction would be
due to low-risk borrowers who would opt to obtain privately insured
mortgages at a lower cost.
The four FHA lenders we contacted agreed with the lenders convened by
FHA that some borrowers would not be able to obtain an FHA loan, but
none of the four offered estimates of specific percentage reductions
in volume of loans. Those four lenders, who are among the largest
FHA lenders in California, Colorado, Virginia, and Maryland, noted
that they would also pass on costs they may incur from a reduction in
FHA's insurance coverage to the consumer in the form of higher
interest rates and that they would set higher underwriting standards
that would exclude higher-risk individuals from qualifying for an FHA
loan. These lenders said they would determine an appropriate
interest rate to charge borrowers by considering factors such as the
risk of default, amount of the loan, value and location of the
property, and amount of funds they would no longer receive from FHA
if borrowers defaulted on loans. The lenders believe that because of
these changes, many high-risk people would not be able to acquire a
home because they would be either unable to afford the loan or unable
to qualify for the loan under the more restrictive underwriting
standards.
FHA lenders at the October 1995 meeting and other mortgage-lending
industry representatives also pointed out other adverse impacts that
reducing FHA's insurance coverage could have on them and the mortgage
market. The lenders predicted that a partial FHA insurance guarantee
would make some lenders less willing to make FHA loans and that some
medium-sized lenders would stop making FHA loans because they would
be unable to afford the increased risk of losses. Only one of the
four lenders we spoke to indicated that he might stop making FHA
loans because of a limit on FHA's insurance coverage.
An executive with the Mortgage Bankers Association (MBA) said that
large FHA lenders will not pull out of the FHA market if FHA's
coverage is reduced but that smaller lenders would leave the program
because they would not be able to afford the additional risk
associated with a partial guarantee. In addition, the executive
stated that the existing insurance program is working well and is
serving many borrowers who would not be served by the conventional
market. Reducing the program's insurance coverage, the executive
believes, would make lenders less inclined to make loans to some low-
and moderate-income borrowers, first-time home buyers, and households
in inner cities.
--------------------
\18 According to Price Waterhouse, the 28-percent decrease in FHA
loans was a general consensus of the 14 FHA lenders.
LENDERS' ESTIMATES OF THE
REDUCTION IN FHA'S VOLUME OF
LOANS AND INCREASE IN FHA'S
INTEREST RATES MAY BE
OVERSTATED
---------------------------------------------------------- Chapter 2:2
While we cannot estimate the actual loan volume and interest rate
that may prevail if FHA's insurance coverage is reduced, our analysis
of FHA's loan volume and interest rates under normal economic
conditions shows that lenders may provide more loans at lower costs
than predicted by those FHA lenders in the focus group if FHA reduces
its insurance guarantee. As previously stated, FHA's lenders
estimated that they will lose about 28 percent of their business if
FHA loans were partially insured. This estimate is based on the
following three assumptions: (1) as a result of the increased risk
faced by lenders, interest rates charged to FHA's borrowers will rise
by one-quarter to one-half percent; (2) the increase in interest
rates and a tightening of underwriting standards will result in a
loss of 18 percent of FHA's high-risk borrowers; and (3) the increase
in interest rates will result in a further loss of 10 percent of
FHA's least-risky borrowers.
Our analysis shows that these assumptions may overstate the extent to
which FHA's loan volume may decrease and interest rates may increase.
This is because the higher revenues earned by lenders from a
one-quarter to one-half percent increase in interest rates may exceed
the losses that lenders would incur if FHA's insurance coverage were
limited to the VA levels. This increase in lenders' net revenues
could prompt some lenders to moderate any interest rate increases,
resulting in a loan volume above that projected by the lenders.
Finally, if FHA were to respond to a reduction in risk related to
lower insurance coverage by lowering FHA's insurance premiums, this
reduction in cost to FHA's borrowers would partly offset any interest
rate increase by lenders. This reduction in borrowers' costs would
also mitigate the effect of the lower-risk borrowers lost by FHA.
An increase in interest rates of one-quarter of a percent would
produce enough revenue to lenders to cover losses, even under the
most adverse economic conditions experienced by FHA in the last 20
years. We chose two criteria to determine the most adverse economic
conditions. We examined Price Waterhouse's 1995 actuarial study to
determine which fiscal year in the last 20 had the highest rate of
lifetime foreclosures and HUD's A43 database to determine which
fiscal year had the highest loss rate per foreclosed loan. We found
that the highest rate of lifetime foreclosures was 22 percent for FHA
loans written in fiscal year 1981 and the highest loss rate per
foreclosed loan was 45 percent, for FHA loans written in fiscal year
1982. If 1981's foreclosure rate were repeated and loss rates per
foreclosed loan were slightly higher than 1982's experience, the
one-quarter percent increase in interest would still generate more
income than a lender would lose in increased claims.
Furthermore, if the increase in interest rates charged to borrowers
were less than the one-quarter to one-half percent assumed by the
lender's focus group, fewer low-risk borrowers may opt to leave FHA's
applicant pool. In addition, lenders would have the capacity and the
incentive to target interest rate increases to higher-risk borrowers,
such as those making low down payments. The exclusion of some
higher-risk borrowers would reduce the lenders' expected losses and
their incentive to increase rates.
While tighter underwriting criteria will lead to some reduction in
FHA's business among higher-risk borrowers, some of that decrease may
be temporary. Persons attempting to buy homes with low down payments
or high payment-to-income ratios may be prevented from using the FHA
program if tighter standards are imposed. However, those people
could return to the home purchase market at a later time when their
income and/or savings have increased. At that point, the decline in
volume would at least be partially replaced by an increase in more
credit-worthy applicants.
VA'S VIEWS ON LIMITED INSURANCE
COVERAGE DIFFER FROM LENDERS'
VIEWS
---------------------------------------------------------- Chapter 2:3
VA operates a home loan insurance program that, depending on the
original amount of the loan, guarantees from 25 to 50 percent of the
loan. VA officials told us that there was no discernible reduction
in the volume of VA loans after the current no-bid policy went into
effect in 1985.\19 They also told us that despite repeated
predictions from lenders and others at the time, they did not
experience an exodus by VA's lenders because of the no-bid policy.
VA officials noted that some of their lenders perceive that rates for
VA loans are higher than those for FHA loans because of VA's no-bid
policy. However, VA officials said they do not know if this is true
because of the lack of supporting data.
VA officials also pointed out that on a typical foreclosure, the VA
guarantee is adequate to cover most losses and that VA and the lender
share in a substantial portion of the losses only for foreclosures
that involve catastrophic losses. VA had no estimates of the savings
that it realized because of its partial guarantee (no-bids) but told
us they thought that the savings were relatively small. VA's data
show that lenders sustained losses in less than 16 percent of VA's
foreclosures during fiscal years 1992 to 1994.
--------------------
\19 VA's no-bid policy is an option that limits losses to VA by
allowing VA to take back a property or leave it with the lender,
depending on which action is more in VA's financial interest. VA
decides which option to follow after estimating and comparing the
cost of taking possession of and reselling a foreclosed property with
the cost of leaving the property with the lender and paying the
lender the guaranteed portion of the mortgaged loan. When a no-bid
occurs, the lender is responsible for losses incurred above those
guaranteed by VA.
LIMITING FHA'S INSURANCE
COVERAGE MAY CAUSE CHANGES IN
THE LENDING INDUSTRY
---------------------------------------------------------- Chapter 2:4
Limiting FHA's insurance coverage would change the distribution of
risk among market participants. Originators of VA's guaranteed loans
or loans made without government guarantees are initially exposed to
the risk of catastrophic losses (losses that exceed the amount
insured). In the market for nongovernment-insured loans, the
catastrophic risk is often sold to the government-sponsored
enterprises--the Federal National Mortgage Association and the
Federal Home Loan Mortgage Corporation. Because FHA's single family
program provides 100-percent insurance coverage, FHA's lenders are
not exposed to these risks. Any reduction in FHA's coverage would
entail shifting this risk to the originators of FHA mortgages.
The lending industry could respond to this new distribution of risk
in several ways. Lenders could sell this risk to other financial
firms as occurs in the conventional mortgage market. Small,
geographically concentrated lenders would be exposed to greater risk
than would geographically diversified lenders, as the smaller lenders
may experience losses from regional as well as national downturns.
Small lenders may merge or be absorbed by larger lenders so as to
geographically diversify, or may leave the FHA lending market. If
lenders continued to hold this catastrophic risk, their regulators or
shareholders might require capital to be held against this risk,
decreasing the profitability of lending.
REDUCING FHA'S INSURANCE
COVERAGE MAY MAKE IT MORE
DIFFICULT FOR SOME BORROWERS TO
PURCHASE A HOME
---------------------------------------------------------- Chapter 2:5
FHA plays a significant role in the single-family housing market by
providing higher-risk borrowers with insured loans and stabilizing
housing markets in areas that are affected by natural disasters or
economic distress. For example, FHA insured 35 percent of the
insured home purchase loans made in 1994. However, FHA fulfills an
even larger role in some specific market segments, particularly
low-income home buyers, minorities, central city residents, and
borrowers qualifying for loans with high loan-to-value (LTV),
housing-expense-to-income, or debt-to-income ratios.\20 The
recipients of these loans are typically higher-risk borrowers that do
not qualify for a conventional home loan with private mortgage
insurance. While about a third of the loans that FHA insured in 1995
might have qualified for conventional mortgages, the other two-thirds
probably would not have qualified, on the basis of maximum LTV and
qualifying ratios of the loans that FHA insured.\21 Without FHA's
100-percent insurance coverage on losses, private lenders anticipate
that they would most likely serve fewer FHA high-risk borrowers and
would raise interest rates on FHA loans. While we do anticipate that
lenders would increase interest rates and reduce lending to
higher-risk borrowers, as discussed previously, our analysis shows
that the volume of loans may not decrease and interest rates may not
increase as much as anticipated by FHA's lenders. For these reasons,
reducing FHA's insurance coverage will make it more difficult and, in
some cases, more costly for some potential home buyers to obtain home
mortgages.
--------------------
\20 The LTV ratio is the amount of the loan divided by the property's
appraised value. FHA's way of calculating LTV is different from the
private sector's or VA's, which results in a slightly lower LTV
ratio. The total debt-to-income ratio compares all of the borrower's
long-term debt payments, including his/her housing expenses, with
his/her income.
\21 For a full discussion of the role that FHA plays in the housing
market, see Homeownership: FHA's Role in Helping People Obtain Home
Mortgages (GAO/RCED-96-123, Aug. 13, 1996).
FHA IS A SIGNIFICANT INSURER
FOR LOW-INCOME, MINORITY,
AND FIRST-TIME HOME BUYERS
-------------------------------------------------------- Chapter 2:5.1
Borrowers with FHA-insured mortgages are more likely to have lower
incomes, be first-time home buyers, or be minorities than borrowers
with privately insured loans. While FHA insured about 35 percent of
all insured mortgages used to purchase homes in 1994, it insured 46
percent of the insured home purchase mortgages made to low-income
borrowers and 48 percent of those made to minorities. Furthermore,
mortgages for low-income and minority home buyers constituted a
greater portion of FHA's 1994 business than they did for private
mortgage insurance or VA, as shown in figure 2.1. In addition, about
67 percent of all 1994 FHA home purchase borrowers were first-time
home buyers.
Figure 2.1: Proportion of Home
Purchase Loans Insured in 1994
for Low-Income Borrowers and
Minorities
(See figure in printed
edition.)
Source: GAO's analysis of data from the Home Mortgage Disclosure Act
and the Mortgage Insurance Companies of America.
FHA also plays a primary role in providing central cities with
mortgage loans. According to an October 1995 FHA study of 1993 data
from the Home Mortgage Disclosure Act, 46 percent of the FHA-insured
loans were for properties located in central cities compared with 38
percent for private insurers.\22 In terms of market share in central
cities, FHA accounted for 35 percent of the FHA eligible market in
1993.\23
FHA also plays an important role in insuring high LTV ratio loans.
FHA insured 43 percent of all the home purchase loans made in 1994
with an LTV ratio of at least 90 percent. Private mortgage insurers
insured 37 percent and VA guaranteed 19 percent. Furthermore, FHA
and VA loans account for almost all the loans made with LTV ratios
greater than 95 percent. In 1994, 65 percent of FHA's home purchase
loans and 86 percent of VA's loans had LTV ratios of at least 95
percent, compared with only 8 percent of the private mortgage
insurance loans with such LTV ratios.
--------------------
\22 An Analysis of FHA's Single-Family Insurance Program, Department
of Housing and Urban Development (Oct. 1995).
\23 "FHA eligible" loans are conventional loans below FHA's maximum
loan amount for each metropolitan area.
MOST FHA-INSURED HOME LOANS
WOULD MOST LIKELY NOT
QUALIFY FOR PRIVATE MORTGAGE
INSURANCE
-------------------------------------------------------- Chapter 2:5.2
On the basis of the LTV and qualifying ratios of their FHA-insured
mortgages and the maximum ratios generally permitted by private
mortgage insurers, about 66 percent of the loans insured by FHA in
1995 would probably not have qualified for private mortgage
insurance.\24 These loans had LTV ratios above the private mortgage
maximum of 97 percent, had housing expense-to-income ratios above the
private mortgage insurance maximum of 33 percent, or had total
debt-to-income ratios above the private mortgage insurance maximum of
38 percent.\25
The number of first-time or low-income FHA borrowers who would not
qualify for private mortgage insurance is even higher. Looking
exclusively at first-time home buyers who took out FHA insurance in
1995, 77 percent would not have met the private mortgage insurers'
LTV and qualifying ratio standards. An even greater portion (85
percent) of the low-income borrowers would not have met the private
mortgage insurance companies' standards.
In addition, FHA's higher debt-to-income and LTV ratios means that
FHA serves higher- risk borrowers who would most likely not qualify
for conventional home loans. FHA is able to serve higher-risk
borrowers because it insures lenders against almost 100 percent of
the losses associated with a foreclosed loan. FHA, therefore,
assumes 100 percent of the risk of loss on a loan it insures.
Private lenders reduce the risk of loss on conventional loans by
maintaining stricter underwriting standards and by requiring private
mortgage insurance that protects them against a portion of the loss.
--------------------
\24 However, some of the borrowers obtaining those loans might have
been able to obtain privately insured loans by increasing their down
payments or buying lower-priced homes.
\25 Not every borrower who meets the LTV and qualifying ratio
guidelines used by private mortgage insurers automatically qualifies
for private mortgage insurance because lenders and insurers consider
additional factors, such as credit history, during the underwriting
process. Similarly, borrowers who fail to meet all three of the
private mortgage insurance guidelines may occasionally qualify for
private mortgage insurance because of compensating factors considered
during the underwriting process.
FHA INSURES HOUSING MARKET
STABILITY
-------------------------------------------------------- Chapter 2:5.3
FHA has been instrumental in providing housing market stability in
some areas where private mortgage companies have decreased their
single-family insurance activities because of a change in the areas'
economy or problems from a natural disaster. Except for FHA's loan
limit,\26 the terms under which FHA and VA mortgage insurance is
available such as the maximum LTV ratio do not generally vary across
different geographic locations, according to the FHA program's
guidelines. However, private mortgage insurance companies may change
the conditions under which they will provide new insurance in a
particular geographic area to reflect the increased risk of losses in
an area experiencing economic hardship. By tightening the terms of
the insurance they would provide, private mortgage insurance
companies have, in the past, decreased their share of the market in
economically stressed regions of the country.
Private mortgage insurance companies accounted for about two-thirds
of the insured market in 1984. However, changing economic and market
conditions that occurred during the early and mid-1980s, including
severe regional recessions and attendant declines in property values
in some energy-producing states, resulted in decreasing loan activity
by private insurers in some single-family markets. For example, in
the five energy-producing states of Alaska, Colorado, Louisiana,
Oklahoma, and Texas, the private mortgage insurance companies' market
share declined from 41 percent in 1984 to 9 percent in 1987. In
contrast, FHA increased its market share from 1984 through 1987,
including in the aforementioned states. FHA's lenders and other
mortgage industry officials indicated, however, that partial
insurance would compromise FHA's mission to stabilize distressed
communities. This is because the lenders' increased exposure to risk
would make them disinclined to provide riskier borrowers with home
loans in areas affected by economic distress or by natural disaster.
--------------------
\26 FHA's loan limit may differ among geographic areas to reflect
differentials in the cost of housing.
SOME FHA LOANS ARE MORE
RISKY THAN OTHERS
-------------------------------------------------------- Chapter 2:5.4
To determine which FHA borrowers represent the greatest risk to
lenders and therefore are more likely to be denied a loan as a result
of tighter underwriting standards applied by lenders, we analyzed the
actual loss experience on foreclosed properties acquired and then
sold from fiscal year 1992 through fiscal year 1994.\27 We found that
five variables had consistent and significant associations with the
losses that lenders would sustain if FHA's insurance coverage were to
be reduced to VA's level. Those variables were (1) minority
borrowers, (2) buildings with multiple living units (characterized as
investor loans), (3) loans written at higher interest rates, (4)
older properties, and (5) smaller loans. Loans with these
characteristics have higher loss rates, all else held equal, and
expose lenders to a greater risk of loss.\28 Therefore, lenders may
disproportionately reduce their lending in these risk categories.
While our analysis did not show that first-time home buyers have
higher loss rates and expose lenders to a greater risk of loss, FHA's
lenders convened at the 1995 focus group considered that many such
borrowers would be priced out of the market. Our analysis of factors
contributing to higher loss rates is explained in greater detail in
appendix II.
--------------------
\27 Data were gathered and linked from three FHA databases--SAMS
database of foreclosed properties, FHA's A43 (financial
characteristics), and the F42 (demographic characteristics) database.
\28 Previous work by GAO, Price Waterhouse, and others indicates that
these variables influence the probability of foreclosure in the same
way that they influence loss rates.
AGENCY COMMENTS AND OUR
EVALUATION
---------------------------------------------------------- Chapter 2:6
HUD stated that our report does not adequately discuss the impact
that reducing FHA's insurance coverage would have on FHA's ability to
stabilize regional housing markets during periods of economic
distress. We agree that FHA has historically played a role in
stabilizing local housing markets affected by economic stress.
Information has been added to the report reflecting FHA's role in
market stabilization and the possible effect that a reduction in
FHA's insurance coverage could have on diminishing FHA's ability to
continue to stabilize distressed communities.
HUD also stated that our report does not adequately discuss the
impact that reducing FHA's insurance coverage would have on the
structure of the lending industry. We agree that reducing FHA's
insurance coverage may change the types and numbers of firms making
FHA loans and the capital structure of those lenders. While a full
analysis of the structure of the mortgage lending industry is beyond
the scope of this report, we have added to this chapter a discussion
of possible changes in the structure of the mortgage- lending
industry.
HUD disagreed with our finding that reducing FHA's insurance coverage
may reduce the number of FHA loans by less than 28 percent. HUD
pointed out that because lenders consider the volatility of losses as
well as the expected level of losses, lenders may seek to reduce the
volatility of losses even if the revenue from higher interest rates
covers the expected level of losses. We agree that lenders consider
volatility in their decision making, and that lenders would have an
incentive to tighten underwriting if FHA insurance were limited.
However, our expectation that lending will be reduced by less than 28
percent is not based on an assumption that underwriting will not be
tightened. Rather, it is predicated on our analysis that suggests
that with a one-quarter percent rise in interest rates, lenders would
likely earn more than enough revenues to cover losses from a
reduction in FHA's insurance coverage even under extremely adverse
circumstances. If interest rates increase by less than one-quarter
of a percent, fewer borrowers would be priced out of FHA's insurance.
Our expectation is also predicated on the fact that borrowers
excluded from the FHA program may reapply and qualify at a later
date.
We disagree with HUD's comment that it is inappropriate to use VA's
experience (including the VA no-bid procedure) as a basis for
predicting lenders' response to a reduction in FHA's insurance
coverage. HUD pointed out that VA has always operated a partial
insurance program and that for this reason, it is inappropriate to
use the example of changes in VA's no-bid procedure as a predictor of
the impact that a reduction in FHA's insurance coverage would have on
lenders' participation.
While it is true that VA has always provided only partial insurance
coverage, as discussed in this report, the 1984 changes to VA's
no-bid process, required by the 1984 Deficit Reduction Act, had a
substantial impact on the number of foreclosed loans for which VA
paid the total guarantee and left the property with the
lender---rising from less than 2 percent of all VA foreclosures
during fiscal year 1982 to 24 percent in fiscal year 1988.
Therefore, lenders' response to changes in the VA program is likely
to be indicative of their response to a change in the FHA program.
HUD also stated that the VA program is very limited and serves
borrowers who are very different from the type served by FHA (i.e.,
borrowers who are concentrated geographically, with stable
employment, higher incomes, and subject to military discipline) and
for this reason the two programs are not comparable. While the VA
program is available to some active military personnel, veterans, and
their families, currently about 80 percent of VA's borrowers are
veterans and not subject to military discipline. Also, as pointed
out in chapter 2, FHA's and VA's programs are similar in that a large
proportion of the home buyers served by both programs are low-income
borrowers and minorities and both play an important role in insuring
high LTV loans. Finally, as discussed in our earlier report, VA's
and FHA's programs serve similar geographic markets.\29
--------------------
\29 For a full discussion of FHA's and VA's role in the housing
market, see Homeownership: FHA's Role in Helping People Obtain Home
Mortgages (GAO/RCED-96-123, Aug. 13, 1996).
REDUCING FHA'S INSURANCE COVERAGE
WOULD LIKELY IMPROVE THE FINANCIAL
HEALTH OF FHA'S FUND
============================================================ Chapter 3
Reducing FHA's insurance coverage to the level permitted for VA home
loans would likely reduce the Fund's exposure to financial losses,
thereby improving its financial health. However, the actual impact
of reducing FHA's insurance coverage on the economic value/reserves,
the resulting capital ratio of loans insured by FHA in fiscal year
1995, and the validity of our estimates would depend on future
economic conditions and on how lenders would respond to a reduction
in FHA's insurance coverage. We estimate that reducing FHA's
insurance coverage would likely result in increasing the capital
reserve ratio for the loans insured by FHA in fiscal year 1995 above
the ratio that results from the current policy of full insurance
coverage. A higher capital reserve ratio will result under reduced
insurance coverage even if FHA's volume of loans declines by 28
percent as predicted by FHA's lenders. This occurs because the
positive financial effect of the following factors more than offsets
the loss of premium income on loans that are not insured by the Fund
when the volume of loans is reduced. First, by reducing FHA's
insurance coverage to VA's level, FHA would be liable for only a
portion of the losses on loans that are foreclosed. Our analysis
shows that FHA's losses on loans that go to foreclosure would be
reduced by an estimated 16 percent if FHA were to go to a partial
insurance coverage like VA's, thereby increasing the estimated
economic value/reserves of the loans it insures. Second, when FHA's
volume of loans is reduced, which would likely occur if FHA's
mortgage insurance coverage were reduced, the capital reserve ratio
increases because, on average, the loans that would be excluded from
FHA's insurance would be higher-risk loans for which higher
foreclosure and loss rates would be expected.
Our estimates also show that at various house price appreciation
rates--a key factor influencing future economic value/reserves and
resulting capital reserve ratios--the capital reserve ratio increases
when moving from full mortgage insurance to partial insurance. In
fact, if house price appreciation rates are 2 percentage points lower
than those used in our estimate, reducing FHA's insurance coverage
may have a much greater positive effect on the capital reserve ratio
of the loans insured by FHA than on the capital reserve ratio in our
baseline estimate.
Although our analysis indicates that reducing FHA's insurance
coverage is likely to increase the Fund's capital reserve ratio, the
impact on a key component of the ratio--the estimated economic
value/reserves of the Fund--is less certain. This is because the
effect of insurance coverage reductions on the economic value of the
Fund depends largely on the way in which lenders respond in excluding
loans from FHA insurance. This is important because our analysis
indicates that the smaller the reduction in the volume of FHA loans,
the higher the economic value contributed by the remaining loans
insured by FHA in fiscal year 1995. Our estimates of the impact of
reducing FHA's insurance coverage on the Fund's capital reserves are
based on the assumption that FHA's premium structure would not change
following a reduction in FHA's insurance coverage.
Alternative approaches to reducing FHA's insurance coverage that we
assessed--retaining full FHA insurance coverage for first-time
home-buyers but reducing FHA's coverage for repeat buyers to VA's
current level and reducing FHA insurance coverage to a level higher
than VA's--also increased the capital reserve ratio above the full
insurance ratio but not by as much as the VA limitation on insurance
coverage did. (See app. III for a more detailed discussion of these
analyses.)
OUR APPROACH TO FORECASTING
FHA'S CAPITAL RESERVE RATIO
UNDER PARTIAL INSURANCE
COVERAGE
---------------------------------------------------------- Chapter 3:1
The actual impact of reducing FHA's insurance coverage on the
economic value/reserves, the resulting capital ratio of loans insured
by FHA in fiscal year 1995, and the validity of our estimates would
depend on future economic conditions and on how lenders would respond
to a reduction in FHA's insurance coverage. To estimate the
financial impact on FHA from a reduction in insurance coverage, we
prepared estimates of economic value and resulting capital ratios
under three different economic scenarios--baseline, optimistic, and
pessimistic--assuming a different rate of appreciation in house
prices over the next 30 years for each economic scenario.
Although future rates of appreciation in house prices are uncertain,
to be conservative in assessing the impact of reducing FHA's
insurance coverage on the capital reserve ratio of FHA's Fund and to
be consistent with past analyses that we have made of the actuarial
condition of the Fund, we placed greater emphasis on our mid-range
baseline economic scenario, which assumes slightly lower house price
appreciation rates (1 percentage point annually) than the rates
forecasted by DRI/McGraw-Hill, a private economic-forecasting
company. Our optimistic and pessimistic economic scenarios assume
that price appreciation rates for houses will be 2 percentage points
higher or lower than those in our baseline. Our estimates of
economic value and resulting capital ratios under reduced insurance
coverage were compared with our estimates for these measures under
full insurance coverage to determine if the Fund's financial health
would be improved. Appendix II contains a detailed description of
how we estimated the decrease in the average loss rate per loan to be
employed under partial FHA insurance coverage.
While our estimates of the impact of reducing FHA's insurance
coverage on the Fund's capital reserves are based on the assumption
that FHA's premium structure would not change, the financial
improvement in the Fund resulting from a reduction in insurance
coverage may result in a lowering of FHA's premium charged to
borrowers. If that occurs, the reduced insurance premium may at
least partially offset any increase in the Fund's capital reserves.
Because the financial outcome of the loans made by FHA in fiscal year
1995 is sensitive to the volume of loans made, our three economic
scenarios included estimates that we made under three loan volume
assumptions--no change in the volume of loans, a 28-percent reduction
in volume as predicted by FHA lenders, and a midpoint reduction of 14
percent. Furthermore, because FHA's lender focus group did not
discuss how higher-risk and lower-risk borrowers would be identified
for exclusion from the program, we used two different assumptions
regarding lenders' ability to screen-out borrowers under each loan
volume reduction. Under one assumption, which we call
"loose-targeting," lenders would tighten underwriting standards to a
limited extent so that there would be some tendency for higher-risk
borrowers to be denied loans. Under the other assumption
"tight-targeting" we assume a stronger tendency for denials to be
concentrated on higher-risk applicants. Appendix II describes in
detail how we implemented those assumptions.
Our estimates of the financial impact of reducing FHA's insurance
coverage to VA's levels under our baseline, optimistic, and
pessimistic scenarios follow. We present our estimates for the
alternative coverage reduction scenarios in appendix III.
OUR BASELINE ESTIMATES OF
INCREASES IN THE FUND'S CAPITAL
RATIO ATTRIBUTABLE TO PARTIAL
INSURANCE
---------------------------------------------------------- Chapter 3:2
As shown in table 3.1, all of our baseline estimates of capital
reserve ratios for loans made by FHA in fiscal year 1995 under VA's
level of insurance coverage are higher than our estimated ratio of
1.48 percent under full insurance coverage remaining in effect.\30
Increases in the capital reserve ratio above those resulting from
full insurance coverage range from 0.41 to 0.51 percentage points
under the volume reduction scenarios.\31
Table 3.1
GAO's Estimates of the Impact of
Reducing Insurance Coverage to VA's
Levels on Loans Insured by FHA in Fiscal
Year 1995, Under Base Case Economic
Conditions
(Dollars in millions)
Initial
Volume Volume Dollar Estimated capital
reduction reduction volume economic ratio\a
Level of coverage (percent) technique of loans value (percent)
------------------- ------------ ------------ ------------ ------------ ------------
Full insurance 0 N/A $36,200 $536 1.48
coverage
VA insurance 0 N/A 36,200 686 1.89
coverage levels
14 Loose 31,100 588 1.89
targeting 31,700 615 1.94
Tight
targeting
28 Loose 26,400 507 1.92
targeting 27,300 543 1.99
Tight
targeting
-----------------------------------------------------------------------------------------
Legend
N/A = not applicable
\a "Initial Capital Ratio" refers to the economic value of the loans
insured by FHA in 1995 as a percentage of the total original loan
amounts for these loans.
Source: GAO's analysis.
Although our analysis indicates that reducing FHA's insurance
coverage is likely to increase the Fund's capital reserve ratio, the
impact on a key component of the ratio--the estimated economic value
of the Fund--is less certain. This is because the effect of
insurance coverage reductions on the economic value of the Fund
depends largely on the way in which lenders would respond in
excluding loans from FHA's insurance.
The size and nature of the reduction in the volume of the loans that
would accompany a reduction in FHA's insurance coverage to VA's
levels are important factors in determining the reduction's impact on
the economic value and resulting capital ratio of FHA's 1995 loans.
If insurance coverage on FHA's 1995 loans were reduced to VA's levels
with no change in the volume of loans insured, the estimated economic
value of the loans would be $686 million--substantially greater than
our estimate assuming no coverage reduction ($536 million). This is
because of the reduction in claim payments that would be expected
from lowered insurance coverage.
However, FHA's lenders estimate that 28-percent fewer FHA-insured
loans would be made each year if a partial insurance program like
VA's were established for FHA. With a 28-percent volume reduction,
we estimate that the economic value of FHA's 1995 loans would be in
the range of $507 million to $543 million--not substantially
different from the $536 million-estimated economic value of the 1995
loans if full insurance coverage remained in effect.\32 The reduction
in lending volume under less than full insurance coverage scenarios
partially offsets the increase in economic value stemming from
reduced claim payments, as some of the loans that would not be
insured would have had positive expected cash flows.
Although a reduction in FHA's insurance coverage to VA's levels would
cause some volume reduction, as noted earlier, we question whether
the reduction would be as high as the FHA lender estimate of 28
percent. This is important because our analysis indicates that the
smaller the reduction in the volume of FHA's loans, the higher the
estimated economic value of FHA's loans. For example, assuming a
14-percent volume reduction, we estimate that the economic value of
the FHA-insured loans made in 1995 would be in the range of $588
million to $615 million if the insurance coverage were reduced to
VA's levels, compared with the $507 million to $543 million estimated
economic value under a 28-percent reduction.
--------------------
\30 The 2-percent capital reserve target, prescribed by legislation,
applies to the entire portfolio of FHA insured loans in the MMI Fund.
HUD does not have to meet the 2-percent requirement on a sub-group of
loans insured in a single year. The initial capital reserve ratio
for loans written in a single year can be lower than 2-percent, and
the loans can still contribute to an improvement in the Fund's
financial health. This is due to the fact that, while the economic
value of a group of loans written in 1 year, if estimated accurately,
will not change over time, the volume of loans outstanding (the
denominator) will decline. Thus, over time, the capital ratio for
loans written in any given year will increase. In our analysis, what
determines whether loans written in a given year contribute more to
the Fund's financial health is whether the estimated capital ratio is
greater than it would be with FHA's full insurance coverage in place.
This, in fact, is the case under all our scenarios and options.
\31 We used the full mortgage principal to calculate capital reserve
ratios in this report to maintain consistency with estimates in our
previous reports. However, if FHA were to cover less than 100
percent of the mortgage's balance, it would be possible to calculate
a capital reserve ratio using the smaller insured portion of mortgage
principal rather than the full mortgage principal as the denominator.
If the capital reserve ratio were calculated using only the insured
portion of the mortgage balance under partial insurance, the ratios
would increase above those we estimate, as the denominator in the
ratio calculation would be smaller.
\32 Our estimated change in economic value varies by about $30
million, depending on how precisely lenders can target tighter
underwriting standards to high-risk borrowers. Lenders may target
their loan rejections even more or less precisely than we have
assumed in our analysis, which could lead to values outside our
estimated range.
OUR ESTIMATES UNDER OTHER
ECONOMIC SCENARIOS
---------------------------------------------------------- Chapter 3:3
If future price appreciation for houses is lower (pessimistic case
scenario) than predicted under our base case scenario, reducing FHA's
insurance coverage to VA's levels could have a much more positive
effect on economic value and the resulting capital reserve ratio than
on the capital reserve ratio in our baseline scenario. As shown in
table 3.2, if a VA-like partial guarantee had been in place for FHA
in fiscal year 1995, the estimated capital ratio and economic value
of the loans made that year under our pessimistic case scenario would
be much higher under partial insurance than they would be with full
insurance coverage.\33 The effect of reduced coverage is larger under
our pessimistic scenario because in that scenario, our model projects
higher claims in future years so that limited coverage would save
more claim payments.
Table 3.2
Estimates of the Impact of Reducing
Insurance Coverage to VA's Levels on
Loans Insured by FHA in Fiscal Year
1995, Under Pessimistic Economic
Conditions
(Dollars in millions)
Initial
Volume Volume Dollar Estimated capital
reduction reduction volume of economic ratio\a
Level of coverage (percent) technique loans value (percent)
------------------- ------------ ------------ ------------ ------------ ------------
Full insurance 0 N/A $36,200 $245 0.68
coverage
VA insurance 0 N/A 36,200 444 1.22
coverage levels
14 Loose 31,100 383 1.23
targeting 31,700 406 1.28
Tight
targeting
28 Loose 26,400 336 1.27
targeting 27,300 366 1.34
Tight
targeting
-----------------------------------------------------------------------------------------
Legend
N/A = not applicable
\a "Initial Capital Ratio" refers to the estimated economic value of
the loans insured by FHA in 1995 as a percentage of the total
original loan amounts for these loans.
Source: GAO's analysis.
If the future price appreciation for houses is higher (optimistic
case scenario) than predicted under our base case scenario, reducing
FHA's insurance coverage to VA's levels could have a much smaller
positive impact on the capital reserve ratio than in our baseline
scenario as shown in table 3.3. The impact on economic value depends
on the extent of the volume reduction accompanying the limitation in
insurance coverage. If the reduction in lending volume is as large
as 28 percent, the economic value declines below that resulting in
our baseline scenario, while a 14-percent reduction in lending leads
to a small increase in economic value. However, it is likely that
higher house price appreciation rates would be associated with
smaller reductions in lending volume because lenders' expected losses
would be lower.
Table 3.3
Estimates of the Impact of Reducing
Insurance Coverage to VA's Levels on
Loans Insured by FHA in Fiscal Year
1995, Under Optimistic Economic
Conditions
(Dollars in millions)
Initial
Volume Volume Dollar Estimated capital
reduction reduction volume of economic ratio\a
Level of coverage (percent) technique loans value (percent)
------------------- ------------ ------------ ------------ ------------ ------------
Full insurance 0 N/A $36,200 $717 1.98
coverage
VA insurance 0 N/A 36,200 837 2.31
coverage levels
14 Loose 31,100 721 2.32
targeting 31,700 747 2.36
Tight
targeting
28 Loose 26,400 623 2.36
targeting 27,300 650 2.38
Tight
targeting
-----------------------------------------------------------------------------------------
Legend
N/A = not applicable
\a "Initial Capital Ratio" refers to the estimated economic value of
the loans insured by FHA in 1995 as a percentage of the total
original loan amounts for these loans.
Source: GAO analysis.
--------------------
\33 Under our optimistic scenario, the capital ratio increases from
1.98 percent to 2.31 percent, if reduced insurance coverage results
in no reduction in the volume of loans.
CONCLUSIONS
---------------------------------------------------------- Chapter 3:4
If FHA's insurance coverage were reduced to VA's insurance coverage
levels, FHA would be insuring better-quality loans with less loss
exposure. As a result, FHA's capital reserve ratio would increase.
While the Fund's financial health has improved and is projected to
continue to improve at least in the near term, reducing FHA's
insurance coverage would enhance the Fund's ability to maintain
financial self-sufficiency in an uncertain future. Forecasting
economic value and resulting capital ratios to determine whether FHA
will have the funds it needs to cover its losses over the 30-year
life of an FHA mortgage is uncertain. Loan performance and,
therefore, capital ratios, will depend on a number of economic and
other factors, particularly on the actual rate of appreciation in
house prices over that period. This uncertainty was demonstrated
during the 1980s when the Fund experienced substantial losses
requiring legislative reforms to help improve its financial health.
However, reducing FHA's insurance coverage does pose trade-offs
affecting lenders, borrowers, and FHA's role. Private lenders will
likely react by raising interest rates and making fewer higher-risk
FHA loans although not to the extent predicted by FHA's lenders.
Borrowers most likely affected by such reactions by lenders would be
low-income, first-time, and minority home buyers and those
individuals purchasing older homes--the type of households for which
FHA has historically been the primary lender. While the denial of
home loans for some of these borrowers may be temporary, posing
delays until their income and/or savings have increased, for other
borrowers it may mean losing their only opportunity to become home
owners. Finally, partial FHA insurance could result in changes in
the lending industry, particularly among smaller FHA lenders and
diminish the federal role in stabilizing markets.
AGENCY COMMENTS AND OUR
EVALUATION
---------------------------------------------------------- Chapter 3:5
HUD stated that our report used overly conservative assumptions
concerning house price appreciation rates and that using more
optimistic assumptions would show that limits on FHA insurance would
reduce the economic value to FHA of originated loans. As pointed out
in our report, estimates of economic value and changes in economic
value are sensitive to the underlying assumptions used in the
forecast of future economic conditions and are subject to
uncertainty. We disagree with HUD's statement that our assumptions
are overly conservative. Our choice of baseline economic conditions
is consistent with our approach in previous reports and conservative
in terms of assessing the impact of reducing FHA's insurance coverage
on the capital reserve ratio of FHA's Fund. We also disagree with
HUD's conclusion that economic value would be reduced by a limitation
in insurance coverage under optimistic assumptions concerning house
price appreciation. The direction of the change in economic value
depends in large part on lenders' behavior, which is difficult to
predict. This is true for both the baseline and optimistic scenarios
that we analyzed. Using a higher rate of house price appreciation
than that assumed in our baseline scenario, we find that economic
value is reduced by $67 million to $94 million if loan volumes are
reduced by 28 percent. However, we believe that optimistic economic
conditions are likely to lead to smaller reductions in volume than
would pessimistic conditions. Thus, the 28-percent volume reduction
under optimistic conditions is even less likely than under the
baseline. If the volume of loans is reduced by 14 percent under
optimistic economic conditions, the economic value increases by $4
million to $30 million above that resulting under full insurance.
REDUCING FHA'S INSURANCE COVERAGE
MAY TRANSFER SOME FINANCIAL LOSSES
TO THE GOVERNMENT NATIONAL
MORTGAGE ASSOCIATION
============================================================ Chapter 4
Ginnie Mae's full guaranty of the timely payment of principal and
interest to investors in its securities was established through
legislation that, unless changed, would remain in place no matter
what level of insurance coverage is provided by FHA for the
single-family loans it insures. Specifically, Ginnie Mae's
securities would continue to maintain the U.S. government's
full-faith-and-credit guaranty, and Ginnie Mae's ability to provide
lenders of FHA and VA loans with liquidity (cash or assets readily
convertible into cash) would not be impaired. However, because of
this full guaranty, Ginnie Mae officials believe that reducing FHA's
insurance coverage would result in transferring some of the losses
that would have been incurred by FHA under full insurance to Ginnie
Mae. The amount of such additional losses that may be incurred by
Ginnie Mae, however, is not clear. The amount will depend on how
successful lenders are in reducing their risk of loss and the
likelihood of default if lenders make fewer higher-risk loans and
charge higher interest rates.
Ginnie Mae's net revenues (after expenses) totaled $515 million in
fiscal year 1996. The major source of Ginnie Mae's revenues was
interest from U. S. securities and fees collected from lenders to
cover its costs and offset its future payments of claims under the
guaranty. Ginnie Mae incurs losses when its approved issuers
(mortgage bankers, savings institutions, and other financial
intermediaries) default on their payments to investors in Ginnie Mae
securities. In the past, losses to Ginnie Mae's issuers on partially
insured VA loans was one of several factors contributing to issuers'
defaults. How lenders respond to a reduction in FHA's insurance
coverage will largely determine the financial impact on Ginnie Mae.
If FHA's lenders react to a reduction in coverage by making fewer
higher-risk loans and increasing mortgage interest rates, lenders'
risk of loss and possible default may be reduced. Ginnie Mae's costs
could rise or fall depending on how well lenders target the reduced
volume of loans toward the highest-risk borrowers and the extent to
which revenues from increased interest rates are adequate to cover
these losses. On the other hand, any reduction in the volume of
loans resulting from a reduction in FHA's insurance coverage would
also result in a proportionate reduction in Ginnie Mae's guaranty
fees and net revenues. Since a reduction in FHA's insurance coverage
could result in FHA's not taking possession of a foreclosed property,
Ginnie Mae would also see an increase in its mortgage and property
holdings.
REDUCING FHA'S INSURANCE
COVERAGE MAY CONTRIBUTE TO SOME
ISSUERS' DEFAULTS AND RESULTING
LOSSES FOR GINNIE MAE
---------------------------------------------------------- Chapter 4:1
When a Ginnie Mae issuer defaults in making timely payments of
principal and interest to investors, Ginnie Mae makes the payments
and takes over the issuer's entire portfolio of FHA and VA mortgage
loans that stand behind securities that Ginnie Mae has guaranteed.
When Ginnie Mae takes over such portfolios, it exposes itself to
greater financial losses because it must manage these portfolios by
(1) collecting principal and interest payments from borrowers, (2)
making payments to owners of the securities, and (3) awarding
servicing contracts to firms that manage the portfolios.
Ginnie Mae officials pointed out that VA's partial insurance coverage
has contributed to increasing the likelihood that issuers will
default. VA currently guarantees a portion of the loan for the
mortgage against loss, and issuers are responsible for losses
incurred above those guaranteed by VA. If the Congress directed FHA
to move to a partial insurance program similar to VA's, the move
would likely expose the entire Ginnie Mae portfolio of newly
securitized loans to losses from this new risk, according to Ginnie
Mae officials. Ginnie Mae officials believe that the likelihood of
losses to issuers from no-bids would increase substantially,
especially during periods of economic decline.
To estimate the financial impact from partial FHA insurance, Ginnie
Mae directed its consultant, Price Waterhouse, to explicitly simulate
the implementation of a partial insurance program for FHA similar to
that currently used in VA's guaranty program.\34
In doing so, Price Waterhouse assumed, under economic conditions
similar to those in our base case, a decline in FHA's annual loan
volume of 28 percent on the basis of the consensus reached by FHA's
lender focus group discussed earlier.
The results of the model showed that issuers' increased exposure to
losses along with a decrease in guaranty fees resulting from a lower
volume of loans would have a negative financial impact on Ginnie Mae.
However, the model's results also indicated that Ginnie Mae would
remain sound and that its ability to provide the U.S. housing market
with liquidity would not be impaired. The severity of the financial
impact on Ginnie Mae would be curtailed, however, as lenders'
redistributed insurance activity away from the riskiest borrowers in
order to minimize the impact of an FHA partial guaranty on the
lenders' financial resources. Under normal economic conditions,
Price Waterhouse estimated that Ginnie Mae's guaranty fee revenues
would decline by approximately 10 percent, on average, primarily
owing to the 28-percent reduction in the volume of loans projected by
the industry focus group. As a result, Ginnie Mae's net revenues
would decline by from 2 to 7 percent annually. In present-value
terms, the cumulative decline in net revenues over the period 1997
through 2007 is estimated at $287 million, or 6.5 percent. At the
same time, however, the results show that advances against defaulted
mortgage-backed security pools would show little change primarily
because of the offsetting effect of the issuers' response.\35
Specifically, according to Price Waterhouse, Ginnie Mae's losses on
defaulted mortgage pools are curtailed as lenders respond to the
reduced insurance coverage by redirecting lending away from risky
borrowers. According to the model's results, the most dramatic
effect on Ginnie Mae would be on its holdings of mortgages and
properties. Since a partial insurance program could allow FHA the
option of not taking possession of a foreclosed property but instead
leaving it with the lender, Ginnie Mae could see a dramatic increase
in the property it acquires from defaulted lenders. Price Waterhouse
estimated that Ginnie Mae would see a five-fold increase in the
dollar volume of mortgages and property holdings because of partial
FHA insurance coverage. The Price Waterhouse analysis also
indicated, however, that Ginnie Mae would remain financially sound,
even under a partial insurance coverage, as indicated by the fact
that Ginnie Mae's U.S. Treasury holdings (reserves for future losses
invested in U.S. Treasury securities) would continue to follow a
smooth path upward. The study's results show that the implementation
of partial insurance coverage results in only a small decline in
Ginnie Mae's U.S. Treasury holdings, compared with such holdings
under full FHA insurance coverage, primarily because of the decreased
volume of loans. Since Ginnie Mae's reserves would continue to grow
in relation to securities outstanding, Ginnie Mae's financial
soundness and ability to provide the U.S. housing market with
liquidity should not be impaired, according to the Price Waterhouse
study.
While the estimated impacts reported above provide an indication of
the potential financial impact on Ginnie Mae, it should be noted that
a number of personal and financial variables can lead to defaults by
issuers and subsequent losses for Ginnie Mae. We did not have the
opportunity during this review to completely evaluate the
reasonableness of Ginnie Mae's model or the assumptions behind the
analysis. Although the results seem reasonable, we question some of
the assumptions used. As noted in chapter 2, we question whether the
lenders' reduction in loan volume will approach the 28-percent
reduction assumed in the analysis. Similarly, while the analysis
assumes that lenders will increase interest rates in response to a
partial FHA insurance program, the fact that the increased interest
revenue would reduce the likelihood of lenders' defaults was not
incorporated into the model. Both of these factors would tend to
lessen the financial impact on Ginnie Mae.
--------------------
\34 This model, which combines a series of econometric models with a
financial model, is designed to enable Ginnie Mae's management to
simulate the impacts of changes in a wide variety of economic,
financial, policy, and programmatic variables on Ginnie Mae's
financial condition.
\35 Advances against mortgage-backed security pools consist of
principal and interest payments made by Ginnie Mae to security
investors to make up for those payments not made by defaulted Ginnie
Mae issuers.
DURING THE 1980S, VA'S
NO-BIDS WERE ONE OF SEVERAL
FACTORS LEADING TO ISSUERS'
DEFAULTS
-------------------------------------------------------- Chapter 4:1.1
During the 1980s, Ginnie Mae, while remaining profitable, began to
experience increased defaults by issuers that exposed it to greater
financial losses. As pointed out in our June 1993 report, four
factors contributed to weakening Ginnie Mae's issuers' financial
health and increasing defaults.\36 Changes in VA's home loan guaranty
program, which resulted in lenders' becoming responsible for losses
above the VA guarantee, were among these factors. Other factors were
economic distress and a resulting decline in regional real estate
markets, a flawed FHA multifamily coinsurance program design, and
issuers' mismanagement of funds. Of the four factors that
contributed to increasing issuers' defaults, Ginnie Mae officials
considered declining regional real estate markets to be the major
factor.
When borrowers are unable to repay their mortgages, the issuers who
pooled these mortgages become responsible for making payments to
investors. If Ginnie Mae's issuers sustain losses brought on by
events such as those discussed above, some issuers may be financially
weakened, causing them also to default. In honoring its guaranties
to investors and acquiring these issuers' portfolios of defaulted
mortgages, Ginnie Mae replenishes funds in certain accounts and hires
firms to temporarily administer the portfolios.
In 1984, the Congress enacted provisions of the Deficit Reduction Act
to, among other things, reduce losses associated with foreclosures on
VA's home loans. The act required VA, in deciding whether to pay the
total VA guaranty on defaulted loans or acquire the property, to
limit its estimated loss to the amount of the guaranty. In making
its decision on whether to acquire a property, the act required VA to
consider post-acquisition costs that were previously excluded from
VA's decision. This increased the number of foreclosed properties
for which expected losses exceeded the guaranty amount. After these
provisions were implemented in 1985, the number of VA's no-bids
increased substantially--rising from 6 percent of all VA foreclosures
in 1984 to 21 percent in 1987. When VA leaves properties with
issuers, the issuers are responsible for losses incurred above those
guaranteed by VA. As issuers' resources are reduced by such losses,
the probability that they will default increases. When issuers
default, Ginnie Mae is responsible for the portion of the losses not
guaranteed by VA.
The number of defaulted Ginnie Mae issuers rose from 5 in 1983 to a
peak of 19 in 1989, declined back to 6 in 1992, and has averaged 4
per year since then. In total, Ginnie Mae acquired portfolios of
defaulted mortgages valued at about $20 billion from its issuers from
fiscal year 1987 through fiscal 1996; over half of this amount, $11.5
billion, occurred in 1989.
--------------------
\36 See Government National Mortgage Association: Greater Staffing
Flexibility Needed to Improve Management (GAO/RCED-93-100, June 30,
1993).
GINNIE MAE'S EXPOSURE TO LOSSES
DEPENDS ON HOW LENDERS RESPOND
TO A REDUCTION IN FHA'S
INSURANCE COVERAGE
---------------------------------------------------------- Chapter 4:2
The extent to which a reduction in FHA's insurance coverage will
increase Ginnie Mae's exposure to financial losses depends on how
lenders react. Annually, about 72 percent of Ginnie Mae's newly
guaranteed securities are made up of almost 100-percent FHA-insured
mortgages, and the remaining 35 percent is partially guaranteed
against loss to the lender by VA. At present, only VA loans, with
their partial guaranty, are candidates for a possible no-bid.
Providing partial insurance coverage on FHA loans would expose the
entire pool of newly securitized FHA loans to the possibility of a
no-bid, thus increasing the lender's and Ginnie Mae's risk exposure.
However, as noted in chapter 2, FHA's lenders will most likely
respond to this increase in risk by making fewer higher-risk loans
and by increasing interest rates for mortgages. Making fewer
higher-risk loans would reduce the likelihood of borrowers' defaults,
thus reducing lenders' risk. According to Price Waterhouse's
analysis, "the severity of the FHA partial guaranty's negative
financial impact on Ginnie Mae would be curtailed by lenders'
redistribution of insurance activity away from the riskiest
borrowers." At the same time, increasing interest rates may produce
the additional revenue needed to cover any increased losses resulting
from a reduction in FHA's insurance coverage--a factor not considered
in the Price Waterhouse analysis. Our analysis in appendix I shows
that increasing interest rates by an extra one-quarter percent would
produce enough revenue to lenders to cover losses even under adverse
economic conditions, thus reducing substantially the likelihood that
issuers would default.
AGENCY COMMENTS AND OUR
EVALUATION
---------------------------------------------------------- Chapter 4:3
In commenting on a draft of this report, Ginnie Mae officials stated
that the discussion of the impact on Ginnie Mae was not well
supported by the analytical results from the memorandum that Ginnie
Mae's contractor provided us with. In addition, they stated that we
should have included the results of an "extreme economic stress"
level test conducted by Price Waterhouse for Ginnie Mae--rates of
house price depreciation greater than those experienced during the
Great Depression of 1929-33.
While Ginnie Mae's contractor provided us with considerable
quantitative material, we limited our discussion in this report to
the material contained in the contractor's memorandum that best
summarized the impact on Ginnie Mae from a reduction in FHA's
insurance coverage. As such, we believe that the information
contained in this report accurately summarizes and reflects the
results contained in the memorandum from Ginnie Mae's contractor. We
chose not to include Ginnie Mae's extreme economic stress scenario
because (1) it is very unlikely that such a severe event would occur
and (2) if it were to occur, the impacts would be so severe for FHA's
lenders (under either full or partial insurance coverage) that we
question the usefulness of the forecasts, especially since such an
economic event is far more severe than any recent historical
conditions encountered.
Ginnie Mae also provided suggested clarifications to the report that
we have incorporated where appropriate.
DESCRIPTION OF INTEREST RATE
CHANGE ANALYSES
=========================================================== Appendix I
Chapter 2 described the potential lender responses and impacts on
potential borrowers of a reduction in the Federal Housing
Administration's (FHA) insurance coverage. This appendix provides a
detailed description of the analyses we conducted of potential
interest rate changes resulting from a reduction in FHA's insurance
coverage. Appendix II describes our analysis of impacts on potential
borrowers.
INTEREST RATE CHANGES RESULTING
FROM A REDUCTION IN FHA'S
INSURANCE COVERAGE
--------------------------------------------------------- Appendix I:1
We determined that lenders would be unlikely to increase interest
rates by as much as the lender focus group convened by the Department
of Housing and Urban Development (HUD) estimated. The focus group
estimated that lenders would raise interest rates on FHA-insured
loans by one-quarter to one-half of a percent to cover the losses
they would become responsible for if FHA reduced its insurance
coverage. As shown in table I.1, an increase of one-quarter of a
percent would produce enough revenue to lenders to cover losses, even
under adverse economic conditions. The adverse conditions we
examined are based on foreclosure rates and balances of loans
originated in fiscal year 1981, which had the highest rate of
foreclosure of any year in the last 20, and a rapid rate of
prepayment. The highest loss rate for any origination year was 45
percent for loans originated in fiscal year 1982. Our analysis of
loans originated in that year indicates that lenders' losses would
have been 7.7 percent, if a partial guarantee had been in place. The
rate of losses to lenders used in the following analysis is 9
percent, which is higher than the 7.7-percent losses that we estimate
lenders would have experienced for 1982 loans.
Table I.1
Estimated Net Revenue to Lenders if They
Had Increased Interest Rates by One-
Quarter of a Percent in Response to an
Insurance Coverage Reduction for FHA's
1981 Loans
As percentage of all 1981 originations Per $1 million in originations
-------------------------------------------------------- ------------------------------
Loans Extra losses
resulting in Loans that Loans still Remaining resulting from Extra interest
Year claims prepay active principal claims rate revenue
------------- ------------ ------------ ------------ -------------- -------------- --------------
1981 0.10 0.17 100.00 100.00 $95 $ 2,493
1982 1.59 0.42 99.73 99.67 1,494 2,435
1983 3.50 6.93 97.73 99.31 3,283 2,167
1984 2.88 4.15 87.30 98.89 2,692 1,984
1985 2.72 4.83 80.27 98.42 2,531 1,789
1986 2.33 14.08 72.72 97.88 2,159 1,378
1987 2.18 12.13 56.30 97.28 2,008 1,021
1988 1.84 4.10 41.99 96.59 1,679 870
1989 1.17 2.61 36.05 95.81 1,057 773
1990 0.83 2.28 32.27 94.93 741 692
1991 0.60 2.13 29.16 93.94 533 621
1992 0.45 2.91 26.43 92.81 399 535
1993 0.34 2.94 23.07 91.53 297 453
1994 0.26 2.62 19.78 90.08 224 381
1995 0.21 1.03 16.91 88.44 174 347
Present value ($9,901) $10,105
of cash
flows
Net revenue $204
-------------------------------------------------------------------------------------------------------
Source: GAO's analysis of data from Price Waterhouse's MMI Fund
analysis for fiscal year 1995.
The first column in table I.1 shows the policy year, ranging from the
1st to the 15th year of a group of mortgages totaling $1 million in
unpaid principal at the time the loans were made. The second, third,
and fourth columns show the fraction of loans that terminate in
claims and prepayments and the fraction of remaining loans at the end
of each year. We calculated these columns from data in Price
Waterhouse's 1995 Actuarial Study and refer to FHA mortgages made in
fiscal year 1981. The next column shows the reduction in principal
from amortization on the standard 30-year fixed-rate mortgages
calculated at an interest rate of 13.24 percent (the average rate on
30-year, fixed-rate FHA mortgages in 1981). The losses to lenders
from a limitation on FHA's insurance coverage, per $1 million of loan
originations, would equal the following: $1 million in originations
times the percent of the original balance remaining each year after
amortization; times the fraction of loans terminating in a claim in
each year; times 1.05, the ratio of acquisition costs to the
outstanding loan balance; times the 9-percent loss rate assumed for
this analysis. The last column shows the extra revenue that would
accrue to lenders from charging an extra one-quarter of a percent per
year, calculated as one-quarter of a percent times the remaining
mortgage balance. We show the present value of the additional claim
costs and additional interest revenue at the bottom of the table
(using 13.24 percent as the discount rate). The difference in
present values demonstrates that the additional revenues generated by
an interest rate increase of one-quarter would have been more than
sufficient to cover the additional claim costs resulting from a limit
in FHA insurance, even under the highly adverse circumstances of
fiscal year 1981 and a 9- percent loss rate to lenders.
LOSS RATE ESTIMATES AND
VOLUME-REDUCTION SCENARIOS
========================================================== Appendix II
This appendix provides a detailed description of our analysis of
FHA's loss rates. We first describe the regression analysis that we
conducted to determine which financial and demographic variables are
associated with higher loss rates. We then describe our methodology
for targeting loans with high loss rates for volume-reduction
scenarios.
IMPACT OF CHARACTERISTICS OF
BORROWERS, LOANS, AND
PROPERTIES ON LOSS RATES
-------------------------------------------------------- Appendix II:1
To determine what types of households would be most affected by a
reduction in FHA's insurance coverage, we analyzed FHA's data on
foreclosed properties acquired and sold by FHA during fiscal years
1992-94. We gathered and linked data from three FHA databases--the
Single-Family Accounting Management System (SAMS, for data on
foreclosed properties), A43 (for accounting data), and F42 (for
demographics on borrowers). Using these data, we analyzed the
effects of several variables on losses. We conducted regression
analyses using two dependent variables--the total loss rate and the
loss to the lender. The total loss rate was defined as the loss (or
profit) on a foreclosed property divided by the acquisition cost (the
claim paid by FHA plus any extra legal or other costs incurred by FHA
for acquisition of the property). The loss to the lender was defined
as any loss in excess of the amount that would be covered at the
Department of Veterans Affairs' (VA) guarantee limits.
We found that certain variables had consistent and significant
associations with losses that lenders would sustain, as shown in
tables II.1 and II.2. Other things equal, loans on buildings with
multiple living units, loans to minority borrowers, loans written at
higher interest rates, loans for older properties, and smaller loans
had higher loss rates.\37
The results in table II.1 are based on a merger of three HUD files,
including the F42 database, which represents a sample of loans. The
results in table II.2 use data from HUD's A43 database only, which is
a population of mortgages originated.
Table II.1
Impact of Characteristics of Borrowers,
Loans, and Properties on Loss Rates
(Income and loan size in thousands)
Total loss rate Lenders' loss rate
---------------------- ----------------------
Coefficien Significan Coefficien Significan
Variable t ce\a t ce\a
---------------------- ---------- ---------- ---------- ----------
Intercept 0.1126 -0.1608
Black 0.0878 ** 0.0643 **
Hispanic 0.0332 ** 0.0257 **
Native American 0.0416 0.0252
Asian and Pacific 0.0023 0.0082
Island
Unknown race 0.0517 ** 0.0597 **
Female 0.0060 0.0055 *
Borrower age -0.0000 0.0002
Multiple living units 0.1809 ** 0.1743 **
Income -0.0000 -0.0000
First-time-buyer 0.0066 * 0.0036
Interest rate 0.0339 ** 0.0197 **
Loan size -0.0020 ** -0.0003 **
Loan-to-value 0.0078 0.0011
No appraisal -0.0040 -0.0088
Building age 0.0030 ** 0.0021 **
R-squared 0.244 0.216
Observations 19,974
----------------------------------------------------------------------
\a Two asterisks indicate significance at the 99-percent confidence
level, and one asterisk indicates significance at the 90-percent
level.
Source: GAO's analysis of HUD's data.
Table II.2
Impact of Selected Characteristics of
Borrowers, Loans, and Properties on Loss
Rates
(Loan size in thousands)
Total loss rate Lenders' loss rate
---------------------- ----------------------
Coefficien Significan Coefficien Significan
Variable t ce\a t ce\a
---------------------- ---------- ---------- ---------- ----------
Intercept 0.6608 0.1927
Black 0.1049 ** 0.0812 **
Hispanic 0.0450 ** 0.0341 **
Native American 0.0436 ** 0.0407 **
Asian and Pacific 0.0027 0.0085
Island
Unknown race 0.0290 ** 0.0182 **
Multiple living units 0.3067 ** 0.2536 **
Loan size -0.0036 ** -0.0014 **
Loan-to-value 0.0097 ** 0.0048 **
No appraisal -0.0094 * 0.0021
R-squared 0.154 0.106
Observations 118,529
----------------------------------------------------------------------
\a Two asterisks indicate significance at the 99-percent confidence
level, and one asterisk indicates significance at the 90-percent
level.
Source: GAO's analysis of HUD's data.
--------------------
\37 Income and loan size were adjusted to 1995 constant dollars.
DETERMINATION OF LOSS RATES AND
VOLUME-REDUCTION SCENARIOS
-------------------------------------------------------- Appendix II:2
To estimate the impact of a reduction in FHA's insurance coverage to
VA's levels on the economic value and resulting capital reserve ratio
of FHA-insured loans, we adjusted the econometric model of FHA's home
loans developed previously.\38 We made two adjustments to reflect the
estimated decrease in the average loss rate per loan and the
reduction in the volume of FHA loans that are likely from an
insurance coverage reduction.
We applied VA's guaranty limits to the losses experienced on FHA
loans during fiscal years 1992-94 and determined what portion of the
losses would have been paid by FHA and what portion would have been
borne by the lender if VA limits had been in effect for those loans.
We estimated that lowering FHA's insurance coverage on these
foreclosed loans would have lowered FHA's average loss rate by 17
percent (from 36 to 30 percent) and losses by about $780 million over
the 3-year period.
Regarding the adjustments that we made to our model to reflect
reductions in FHA's loan volume, FHA's lender focus group indicated
that FHA's loan volume would be reduced by 28 percent--18 percent
because of higher costs and tighter underwriting standards targeted
at higher-risk borrowers plus 10 percent because higher FHA costs
would cause lower- risk borrowers to obtain lower-cost private
mortgage insurance. However, the lender focus groups did not discuss
how these changes would be implemented or how higher- and lower-risk
borrowers would be identified. Hence, we estimated our loan volume
reduction scenarios using two different assumptions regarding (1)
lenders' ability to screen out higher-risk borrowers and (2) which
lower-risk borrowers would seek private insurance. Under what we
call the "loose-targeting" scenario, we assumed that lenders would
tighten underwriting standards to a limited extent so that there was
some tendency for higher-risk borrowers to be denied loans. Under
our "tight-targeting" scenario, we assumed a stronger tendency for
denials to be concentrated among higher-risk applicants. Under both
scenarios, we assumed that lenders would screen out loans to
properties with two to four living units, as these loans have high
foreclosure and loss rates. These loans constituted 4 percent of the
loans insured by FHA in fiscal year 1995. For the remaining loans,
our loose-targeting scenario assumed that a further 14 percent of
FHA's fiscal year 1995 insured loans would be screened out by tighter
underwriting standards. The loans excluded would be loans written at
loan-to-value (LTV) ratios of 95 percent or above and would be spread
randomly through the lower half of the loan size distribution, as
small loans have higher foreclosure and loss rates. Under tight
targeting, we assumed that the smallest 14 percent of the loans with
95 percent or higher LTV ratios would not be insured.
We then eliminated 10 percent of fiscal year 1995's insured loans to
account for lower-risk borrowers who may seek private insurance
rather than FHA insurance because of costs. These loans were taken
from the population of loans originated with LTV ratios of less than
95 percent, selected randomly from the top half of the size
distribution.
While we chose to focus on the LTV, loan's size, and number of units
in a property as our determinants of risk, lenders could use a
variety of approaches to target FHA-insured loans away from borrowers
associated with greater risks of loss. Our analysis of FHA's data on
loan losses, as discussed above, indicates that in addition to the
loan's size, its LTV ratio, and the number of units in the property,
the age of the property, race of the borrower, and interest rate on
the mortgage are associated with higher foreclosure and/or loss rates
on mortgages. However, while lenders may know the age of the
property before making a loan, FHA's A43 database does not record the
age of the property, so we could not use this variable to eliminate
loans. We did not choose loans for volume reduction on the basis of
the borrower's race.\39 Lastly, interest rates are driven largely by
macroeconomic factors. For these reasons, we focused on the loan's
size, the LTV ratio, and the number of units in the property to
determine which loans carried the most risk and should be excluded
from reduced-volume scenarios. Lenders could use additional
variables to target risk even more closely.
--------------------
\38 See Mortgage Financing: FHA Has Achieved Its Home Mortgage
Capital Reserve Target (GAO/RCED-96-50, Apr. 12, 1996).
\39 Legislation prohibits discrimination based on race. Some
analysts maintain that, though illegal, some lending decisions are
influenced by race. The extent to which this happens is a subject of
controversy and beyond the scope of this report.
OTHER APPROACHES WOULD RESULT IN
MORE MODEST CHANGES TO THE FUND'S
CAPITAL RATIO
========================================================= Appendix III
While chapter 3 focused primarily on the financial impact on FHA's
capital reserves of imposing insurance coverage limitations similar
to those used in VA's loan guarantee program, we also estimated the
impact of implementing two alternative methods of reducing FHA's
insurance coverage. The primary objectives of the proposals we
reviewed are to reduce FHA's liability while attempting to provide a
higher level of coverage for those current FHA customers who may be
underserved by the conventional market, such as first-time home
buyers and borrowers with low and moderate incomes. Specifically, we
estimated the impact of reducing FHA's insurance coverage to VA's
levels for all repeat buyers while retaining FHA's existing
100-percent coverage for first-time home buyers. We also estimated
the impact of imposing a graduated coverage schedule like VA's,
adjusted to provide slightly higher coverage levels for loans of low
and moderate size--60-percent coverage for the lowest loan amounts,
declining to 30 percent for high loan amounts.
While, according to our analysis, these proposals had less of an
impact than using VA's coverage levels on the estimated economic
value of the Fund compared with the current full insurance policy,
they do increase FHA's capital reserve ratio. For the first-time
home buyer alternative, we estimated the economic value and the
capital reserve ratio under the assumptions that there would be no
reduction in the volume of loans and that the volume of loans for
repeat home buyers would be reduced by 14 percent, while insurance
coverage would remain unchanged for first-time buyers.\40 We estimate
that this would increase economic value by $96 million to $118
million over our baseline's full- insurance estimate and increase the
capital reserve ratio by about 0.3 percent.
Since loans of lower dollar amount have higher losses, we also
examined the alternative of imposing insurance coverage limits that
are similar to VA's but with slightly higher coverage levels for
lower loan amounts. Specifically, insurance coverage was set at 5
percentage points higher than VA's for loans from $90,000 to $144,000
and the same as VA's for loans above $144,000. For loans less than
$90,000, the insurance coverage was set 10 percentage points higher.
We produced estimates assuming no reduction in loan volume and a
14-percent reduction in the volume of loans. Under these scenarios,
we estimate an increase in economic value above our baseline's full
insurance coverage scenario of $90 million if loan volumes are not
reduced but estimate an increase in economic value of $9 million to
$20 million if loan volumes are reduced 14 percent. Capital reserve
ratios increase by about one-quarter of a percent. Our estimates
under these two alternatives are shown in table III.1.
Table III.1
Estimates of the Impact of Alternative
Insurance Coverage Reduction Approaches
on Loans Insured by FHA in Fiscal Year
1995, Under Base Case Economic
Conditions
(Dollars in millions)
Initial
Volume Volume Dollar Estimated capital
reduction reduction volume economic ratio\a
Insurance coverage (percent) technique of loans value (percent)
------------------- ------------ ------------ ------------ ------------ ------------
Full insurance 0 N/A $36,200 $536 1.48
coverage
Retain full 0 N/A 36,200 654 1.80
insurance coverage
for first-time
home buyers
14 Loose 35,200 632 1.81
targeting 35,000 639 1.82
Tight
targeting
Reduce insurance 0 N/A 36,200 626 1.73
coverage to levels
higher than VA's
14 Loose 31,100 545 1.75
targeting 31,700 556 1.76
Tight
targeting
-----------------------------------------------------------------------------------------
Legend
N/A = not applicable
\a "Initial Capital Ratio" refers to the economic value of the loans
insured by FHA in 1995 as a percentage of the total original loan
amounts for these loans.
Source: GAO's analysis.
(See figure in printed edition.)Appendix IV
--------------------
\40 Because less risk is transferred to lenders under this and the
following alternative, we reasoned that any reduction in the volume
of loans would be less than that indicated by FHA's lender focus
group.
COMMENTS FROM THE DEPARTMENT OF
HOUSING AND URBAN DEVELOPMENT
========================================================= Appendix III
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
The following are GAO's comments on the Department of Housing and
Urban Development's letter dated April 4, 1997.
GAO'S COMMENTS
------------------------------------------------------- Appendix III:1
1. We believe that our report provides adequate information in
chapters 1 and 2 on the relationship among FHA, VA, private lenders,
and Ginnie Mae and the role that each plays in the U.S. housing
market, especially in terms of the type of borrowers that each
serves. Our report also provides information on the impacts that
partial insurance would have on FHA, Ginnie Mae, private lenders, and
potential home buyers. Also, we believe that our report provides
reasonable estimates, on the basis of our analysis, of the impact on
lenders and high-risk borrowers of a change to a partial insurance
program for FHA. We agree that FHA's role in stabilizing housing
markets during periods of economic distress will be lessened by a
reduction in FHA's insurance coverage; hence, information has been
added to reflect this in the report.
2. We believe that the report provides reasonable estimates of the
range of possible changes in economic value that may result from a
limitation of FHA's insurance coverage. We agree with FHA that the
estimated effect of limiting coverage depends on assumptions about
economic conditions, and we clearly indicate circumstances in which
economic value is estimated to decrease as a result of a limitation
on insurance coverage. One of our scenarios in table 3.1 (baseline
forecast with a 28-percent reduction in the volume of loans) and two
of our scenarios in table 3.3 (optimistic forecast with a 28-percent
reduction in volume of loans) show declining economic value. Several
other scenarios show changes that are positive but small in
magnitude.
3. Our expectation that the volume of loans would likely decline by
less than 28 percent is not based on a belief that lenders who are
able to cover their expected risk will have no incentive to reduce
their loan originations. Our expectation rests on two premises. The
first premise is that the estimate made by FHA's lender focus group
of a one-quarter to one-half percentage point increase in interest
rates is overstated. We believe that the estimate is overstated
because such an increase would cover expected losses 5 to 10 times
over and would more than cover losses under the most adverse
conditions experienced by FHA over the last 20 years. If lenders
raise interest rates by a smaller amount than was predicted by the
focus group, fewer potential borrowers would be priced out of the FHA
program. The second premise is that borrowers who are excluded from
the FHA program because lenders require higher down payments or
higher income may return to the housing market at a later date. For
at least some of these borrowers, savings and/or income will have
increased, allowing them to qualify for mortgages under more
stringent underwriting standards. While we agree with FHA's lender
focus group that issues of risk and the volatility of risk would
cause FHA lenders to tighten underwriting standards, we believe that
some of the resulting decrease in FHA's insured lending will be
temporary, not permanent.
4. Text has been added to chapter 2, which discusses the transfer of
risk for non-FHA loans. The added text discusses the possibility
that limiting FHA's insurance coverage could change the distribution
of risk among market participants and the fact that any reduction in
FHA's coverage would entail shifting this risk to the originators of
FHA mortgages.
5. We believe that the 1984 changes that were made in VA's
procedures for determining when to acquire a foreclosed property or
leave it with the lender provide valuable insight into the potential
response of FHA's lenders to a change in their risk exposure.
However, HUD believes that it is inappropriate to use VA's experience
(including the VA no-bid procedures) as a basis for predicting
lenders' response to a reduction in FHA's insurance coverage for two
reasons. First, VA has always operated a partial insurance program,
and for this reason, the 1984 changes in VA's no-bid procedure would
not have as major an impact on lenders' participation as a reduction
in FHA's insurance coverage. Second, HUD stated that VA's program is
very limited and serves a very different type of borrower from the
type served by FHA. We disagree on both points. While VA has always
provided only partial insurance coverage, as discussed in this
report, the 1984 changes to VA's no-bid process required by the 1984
Deficit Reduction Act had a substantial impact on the level of risk
assumed by VA's lenders. The number of foreclosed loans in which VA
paid the total guarantee and left the property with the lender
increased substantially after the no-bid policy went into effect in
fiscal year 1985. While VA's program is available to members of the
armed forces, veterans and their families currently make up about 80
percent of VA's borrowers. In addition, as pointed out in chapter 2
of this report, FHA's and VA's programs are very similar in terms of
the types of borrowers, LTV ratios, and geographic markets served.
6. We agree that one cannot eliminate the risk of catastrophic loss,
and our report does not suggest that this may occur. Our report
states that reducing FHA's insurance coverage will shift this risk to
lenders, who will then respond by increasing interest rates and
making fewer higher-risk loans. Lenders may sell this risk to other
firms. Even after considering catastrophic losses, HUD's lender
focus group concluded that making FHA loans under partial insurance
is profitable and that 72 percent of FHA volume of loans would be
retained. The report also states that this increase in the risk of
loss to lenders could lead to increased financial losses for Ginnie
Mae.
7. Text that discusses the issue of capital requirements has been
added to chapter 2.
8. We added information to the section of chapter 2 to explain that
because other factors, such as credit history, are considered during
the underwriting process, not every borrower who meets the private
mortgage insurers' LTV and qualifying ratio guidelines may
necessarily be eligible for private mortgage insurance. Similarly,
we note that private mortgage insurance might still be provided to a
borrower who fails to meet all of the LTV and qualifying ratio
guidelines if compensating factors are deemed sufficient during the
underwriting process. This information, although useful for
understanding the types of borrowers likely to be most affected by an
insurance coverage reduction, is not the basis for our critique of
the 28-percent volume reduction estimate. As discussed in chapter 2,
we believe that the 28-percent estimate is overstated for two
reasons. First, the 28 percent estimate is predicated upon an
interest rate increase of one-quarter to one-half percentage point,
which our analysis indicates is more than is likely to occur. Also,
potential borrowers who are excluded by tightened underwriting
criteria may qualify for a loan at a later date--a possibility not
considered by the lender focus group convened by FHA.
9. As noted in chapter 3, our analyses of the impact of a coverage
reduction on the capital reserve ratio and economic value of FHA's
fiscal year 1995 loans assumes, as did HUD's analysis, that FHA's
premium structure would remain unchanged. Our analysis is largely
based on the results of the lender focus group sponsored by HUD,
which assumed a constant premium structure. Our report explains in
chapters 2 and 3 that reducing FHA's insurance coverage might cause
decisionmakers to lower premiums. Our report also explains that if
this occurs, the reduced insurance premium may at least partially
offset any increase in the Fund's capital reserve.
10. HUD stated that we did not adequately discuss the impact that
reducing FHA's insurance coverage would have on FHA's ability to
stabilize regional housing markets during periods of economic
distress. We agree. Information has been added to the report
reflecting FHA's role in market stabilization and the possible
diminishing effect that a reduction in FHA's insurance coverage could
have on FHA's ability to continue to stabilize distressed
communities.
MAJOR CONTRIBUTORS TO THIS REPORT
=========================================================== Appendix V
RESOURCES, COMMUNITY, AND ECONOMIC
DEVELOPMENT DIVISION, WASHINGTON,
D.C.
Leslie Black-Plumeau
Jay R. Cherlow
Austin Kelly
Robert S. Procaccini
Phillis Riley
Patrick L. Valentine
RELATED GAO PRODUCTS
============================================================ Chapter 1
Homeownership: FHA's Role in Helping People Obtain Home Mortgages
(GAO/RCED- 96-123, Aug. 13, 1996).
Mortgage Financing: FHA Has Achieved Its Home Mortgage Capital
Reserve Target (GAO/RCED-96-50, Apr. 12, 1996).
Rural Housing: Opportunities Exist for Cost Savings and Management
Improvement (GAO/RCED-96-11, Nov. 16, 1995).
Homeownership: Mixed Results and High Costs Raise Concerns About
HUD's Mortgage Assignment Program (GAO/RCED-96-2, Oct. 18, 1995).
Property Disposition: Information on HUD's Acquisition and
Disposition of Single-Family Properties (GAO/RCED-95-144FS, July 24,
1995).
Rural Housing: Shift to Guaranteed Program Can Benefit Borrowers and
Reduce Government's Exposure (GAO/RCED/AIMD-95-63, Dec. 21, 1994).
Mortgage Financing: Financial Health of FHA's Home Mortgage
Insurance Program Has Improved (GAO/RCED-95-20, Oct. 18, 1994).
Rural Development: Patchwork of Federal Programs Needs to Be
Reappraised (GAO/RCED-94-165, July 28, 1994).
Homeownership: Appropriations Made to Finance VA's Housing Program
May Be Overestimated (GAO/RCED-93-173, Sept. 8, 1993).
Homeownership: Actuarial Soundness of FHA's Single-Family Mortgage
Insurance Program (GAO/T-RCED-93-64, July 27, 1993).
Government National Mortgage Association: Greater Staffing
Flexibility Needed to Improve Management (GAO/RCED-93-100, June 30,
1993).
Homeownership: Loan Policy Changes Made to Strengthen FHA's Mortgage
Insurance Program (GAO/RCED-91-61, Mar. 1, 1991).
*** End of document. ***