Housing Enterprises: Potential Impacts of Severing Government Sponsorship
(Chapter Report, 05/13/96, GAO/GGD-96-120).
Pursuant to a legislative requirement and a congressional request, GAO
examined the potential effects of privatizing the Federal National
Mortgage Association (Fannie Mae) and the Federal National Mortgage
Association (Freddie Mac).
GAO noted that: (1) the privatization of Fannie Mae and Freddie Mac
would have a major impact on both the secondary and primary mortgage
markets; (2) if the two government-sponsored enterprises lost the
benefits of their federal charter, their costs would increase because
they would be responsible for paying Securities and Exchange Commission
registration fees on their securities or state and local taxes; (3) by
eliminating or reducing the implied federal guarantee on mortgage-backed
securities (MBS) and debt, the enterprises' borrowing costs would
increase from 30 to 106 basis points if the perceived federal guarantee
were completely eliminated; (4) these increased costs would be passed to
the homebuyer and the average mortgage interest rate would increase by
15 to 35 basis points; (5) eliminating the cost advantages of federal
sponsorship could spur more competition and retain liquidity in the
secondary market because firms would find it profitable to purchase and
securitize conforming mortgages; (6) privatization could stabilize the
securities market and prevent it from experiencing regional disparity;
(7) privatization would pose an adverse threat to the enterprises'
financial performance because they would be more dependent on their
strategic business decisions and total quality management; (8) low and
moderate-income borrowers would be most impacted by the enterprises'
privatization because the federal programs providing credit to these
groups would be eliminated; and (9) alternative initiatives should be
studied to limit the risk to taxpayers.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: GGD-96-120
TITLE: Housing Enterprises: Potential Impacts of Severing
Government Sponsorship
DATE: 05/13/96
SUBJECT: Government sponsored enterprises
Federal aid for housing
Federal agency reorganization
Mortgage programs
Federal corporations
Privatization
Mortgage-backed securities
Economic analysis
Mortgage loans
Government guaranteed loans
IDENTIFIER: Federal Home Loan Bank System
******************************************************************
** 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 Congressional Requesters
May 1996
HOUSING ENTERPRISES - POTENTIAL
IMPACTS OF SEVERING GOVERNMENT
SPONSORSHIP
GAO/GGD-96-120
Housing Enterprises
(233420)
Abbreviations
=============================================================== ABBREV
CBO - Congressional Budget Office
CMO - Collateralized Mortgage Obligation
CRA - Community Reinvestment Act
FDIC - Federal Deposit Insurance Corporation
FHA - Federal Housing Administration
FHFB - Federal Housing Finance Board
FHLB - Federal Home Loan Bank System
FIRREA - Financial Institutions Reform, Recovery, and Enforcement
Act of 1989
FSLIC - Federal Savings and Loan Insurance Corporation
GAO - General Accounting Office
GSE - Government-Sponsored Enterprise
HMDA - Home Mortgage Disclosure Act
HUD - Department of Housing and Urban Development
MBS - Mortgage-Backed Security
OFHEO - Office of Federal Housing Enterprise Oversight
OMB - Office of Management and Budget
REMIC - Real Estate Mortgage Investment Conduit
RTC - Resolution Trust Corporation
SEC - Securities and Exchange Commission
UNBOG - Underwriting Barriers Outreach Group
VA - Veterans Administration
WACC - weighted average cost of capital
Letter
=============================================================== LETTER
B-259964
May 13, 1996
The Honorable Alfonse M. D'Amato
Chairman
The Honorable Paul S. Sarbanes
Ranking Minority Member
Committee on Banking, Housing, and
Urban Affairs
United States Senate
The Honorable James A. Leach
Chairman
The Honorable Henry B. Gonzalez
Ranking Minority Member
Committee on Banking and Financial Services
House of Representatives
This report responds to the questions concerning the implications of
privatizing the Federal National Mortgage Association and the Federal
Home Loan Mortgage Corporation contained in Sec. 1355 of the Housing
and Community Development Act of 1992 (P.L. 102-550).
We are sending copies of this report to the Secretary of Housing and
Urban Development, the Secretary of the Treasury, the Director of the
Congressional Budget Office, the Director of the Office of Federal
Housing Enterprise Oversight, the Federal Home Loan Mortgage
Corporation, and the Federal National Mortgage Association.
Major contributors to this report are listed in the appendix. If you
have any questions about this report, please call me on (202)
512-8678.
James L. Bothwell
Director, Financial Institutions
and Markets Issues
EXECUTIVE SUMMARY
============================================================ Chapter 0
PURPOSE
---------------------------------------------------------- Chapter 0:1
The Federal National Mortgage Association (Fannie Mae) and the
Federal Home Loan Mortgage Corporation (Freddie Mac) are
government-sponsored enterprises with about $1.4 trillion in combined
obligations as of December 1995. In response to growing concern
about the potential risk that these obligations represent to
taxpayers and increasing questions about the continued need for their
government-sponsored status, Congress directed GAO, the Department of
the Treasury, the Congressional Budget Office (CBO), and the
Department of Housing and Urban Development (HUD) to each study the
effects of privatizing Fannie Mae and Freddie Mac by repealing their
federal charters, eliminating any federal sponsorship, and allowing
the enterprises to operate as fully private corporations.
This report responds to GAO's study requirement, contained in the
Housing and Community Development Act of 1992, by assessing the
potential effects of privatization on (1) the enterprises; (2)
residential mortgage markets in general; and (3) housing finance,
homeownership, and housing affordability for very low-, low-, and
moderate-income families and residents of underserved areas in
particular. It also responds to the request of the House Committee
on Banking and Financial Services, Subcommittee on Capital Markets,
Securities, and Government Sponsored Enterprises, that GAO identify
and discuss alternative policy options that Congress could consider
to limit the enterprises' potential risk to taxpayers or increase
their social benefits.
BACKGROUND
---------------------------------------------------------- Chapter 0:2
Congress established and chartered the enterprises as
government-sponsored, privately owned and operated corporations to
enhance the availability of mortgage credit across the nation during
both good and bad economic times. The enterprises accomplish this
mission by borrowing funds in the capital markets and using these
funds to purchase mortgages from lenders (banks, thrifts, and
mortgage bankers) across the country, who can then make additional
loans to borrowers in the primary mortgage market. The enterprises
retain some of the mortgages they purchase in their own portfolios.
Most of the mortgages, however, are pooled to create mortgage-backed
securities (MBS) that are sold to investors in the secondary mortgage
market. The enterprises charge fees for guaranteeing the timely
payment of principal and interest on the MBS backed by the mortgage
pools.
As of December 1995, Fannie Mae had $513 billion in MBS obligations
outstanding, $299 billion in debt obligations, and $253 billion in
retained mortgage holdings. Freddie Mac had $459 billion in MBS
obligations outstanding, $119 billion in debt obligations, and $107
billion in retained mortgages. Of the total $3.9 trillion in U.S.
residential mortgage debt as of September 1995, about 25 percent was
in enterprise MBS, and about 9 percent was in enterprise-retained
portfolio.
The enterprises' federal charters grant each of them explicit
benefits, which GAO assumed they would lose if privatized. These
explicit benefits include (1) exemption from state and local
corporate income taxes, (2) exemption from registering their
securities with the Securities and Exchange Commission (SEC), (3)
$2.25 billion conditional lines of credit with the Treasury
Department, and (4) use of the Federal Reserve as a transfer agent.
The most important benefit that the enterprises receive from their
government-sponsored status, however, is an implicit one stemming
from investors' perception that the federal government would not
allow the enterprises to default on their obligations. While the
enterprises' charters state that their obligations must include a
statement that they are not guaranteed by the United States, it seems
clear that the enterprises' federal ties cause creditors to believe
that their investments are safe. For example, the enterprises can
borrow at rates that are only slightly above Treasury borrowing rates
and can continue to borrow in the capital markets even if performing
poorly. Although part of this perceived federal guarantee could also
be due to the very size of the enterprises, GAO assumed that an
implied federal guarantee would be substantially reduced, if not
eliminated, on any debt and MBS the enterprises issue after they
became fully private corporations.
As federally sponsored corporations, the enterprises are also
required to operate under certain restrictions that GAO assumed would
not exist if they were privatized. These restrictions include (1)
confining their operations to the secondary mortgage market; (2)
limits on the maximum size of mortgages they can purchase (mortgages
that meet the enterprises' underwriting standards and are within this
limit, currently $207,000 on single-unit residences, are called
"conforming loans" and mortgages above this limit "jumbo loans"); (3)
an obligation to be active in the secondary market across the country
at all times; (4) regulations requiring them to meet certain
numerical goals regarding purchases of mortgages to very low-, low-,
and moderate-income borrowers, and borrowers in central cities and
other underserved areas; and (5) compliance with capital requirements
and safety and soundness regulations issued by the Office of Federal
Housing Enterprise Oversight.
During the 1980s, several large financial companies became
increasingly important issuers of "private-label" MBS. These firms
did not compete directly with the enterprises, but rather purchased
and securitized nonconforming or jumbo mortgages. As of September
1995, private-label MBS accounted for about 13 percent of total MBS
outstanding, or about 6 percent of total residential debt.
To achieve the objectives of this report, GAO reviewed academic,
professional, and business literature on the role of the enterprises
in the mortgage market. GAO also interviewed representatives of the
enterprises, other market participants, and individuals with
expertise in mortgage markets. GAO participated with CBO, HUD, and
the Treasury Department in commissioning five academic studies on
different aspects of privatization to provide a common source of
information. These studies are to be published separately by HUD.
GAO drew upon these studies as it deemed appropriate. GAO also made
estimates, based on various assumptions and the best available data,
to quantify the potential impact of privatization, noting that such
estimates are inherently imprecise because of the difficulty in
predicting responses by the enterprises and other market participants
to a major change such as privatization.
RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3
Fannie Mae and Freddie Mac have played critical roles in establishing
and maintaining a nationwide secondary mortgage market, enhancing the
affordability and availability of housing finance, and increasing
efficiency through greater standardization of mortgage products and
processes. In part, their success has been due to the special nature
of their government-sponsored status. Their government sponsorship,
however, has also created a large potential risk to taxpayers if the
federal government should ever have to assist them in meeting their
financial obligations. Thus, altering or repealing their federal
charters involves both potential benefits and risks as well as
trade-offs among competing policy objectives that need to be
considered and weighed carefully.
Although there are inherent difficulties in attempting to predict
privatization's effects, privatization could undoubtedly have a major
impact on the two enterprises and on both the secondary and primary
mortgage markets. Losing the explicit benefits contained in their
federal charters, for example, would increase the enterprises' costs
because they would no longer be exempted from paying either SEC
registration fees on their securities or state and local income
taxes. A much more substantial cost impact would result as
privatization raised their cost of funds by eliminating or
substantially reducing investors' perception of an implied federal
guarantee on their MBS and debt. For example, GAO's analysis of
available evidence indicated that the enterprises' borrowing costs
would increase and that the increase could range from about 30 basis
points to as much as 106 basis points if the perceived guarantee were
completely eliminated by privatization.\1
To some extent, these increased costs would be passed along to
homebuyers in the form of increased mortgage interest rates. GAO's
analysis indicated that average interest rates on mortgages below the
conforming limit would likely increase by about 15 to 35 basis
points.
Eliminating the cost advantages of federal sponsorship would also
likely increase competition in the secondary market if firms that
currently purchase and securitize nonconforming and jumbo mortgages
find it profitable to purchase and securitize conforming mortgages.
The expected increased participation of these firms and perhaps a few
new entrants, along with the continued participation of the
enterprises, could be sufficient to keep the secondary mortgage
market efficient and liquid enough to prevent the appearance of any
significant regional disparities in the availability and cost of
mortgage credit that are not related to differences in risk.
By increasing their costs and competition, privatization would likely
have an adverse impact on the enterprises' profits, stock values, and
market shares. For example, the enterprises believe that
privatization would largely eliminate profits from holding mortgages
in portfolio, thus making them strictly issuers of MBS.
Privatization would, however, presumably allow the enterprises to
enter other lines of business by removing their charter restrictions.
The ultimate impact of privatization on their financial performance
would thus depend upon such other factors as their strategic business
decisions and the quality of management.
Privatization could have a relatively greater impact on the
availability and affordability of mortgage credit for very low-,
low-, and moderate-income borrowers and borrowers in underserved
areas, because it would likely eliminate one of the federal
mechanisms for channeling residential mortgage credit to these
borrowers and areas. Both enterprises have initiated substantial
efforts to help meet their numerical goals for purchasing such loans,
although it is too soon to reliably estimate the effects of these
efforts on housing affordability. Other existing housing policy
mechanisms or new policy initiatives could be used to help offset any
undue adverse impact of privatization on such residential lending.
As possible alternatives to privatization, GAO identified other
policy options that could be used to limit the enterprises' risk to
taxpayers or increase their social benefits. Like privatization,
each of these options involves both benefits and risks and trade-offs
among competing policy goals that need to be considered and weighed
carefully.
--------------------
\1 A basis point is equal to one one-hundredth of a percent.
GAO'S ANALYSIS
---------------------------------------------------------- Chapter 0:4
PRIVATIZATION WOULD INCREASE
ENTERPRISES' COSTS AND
AFFECT THEIR ACTIVITIES
-------------------------------------------------------- Chapter 0:4.1
There are substantial economic benefits associated with the
enterprises' federal sponsorship that presumably would be reduced, if
not entirely eliminated, if Fannie Mae and Freddie Mac were to be
privatized. GAO estimated that the value of these benefits in 1995
ranged from about $2.2 billion to $8.3 billion on a beforetax basis
and from about $1.6 billion to $5.9 billion on an aftertax basis.
Between 80 percent and 95 percent of these estimated benefits
resulted from lower funding costs for enterprise debt and MBS
attributable to the perception of an implied federal guarantee.
Thus, the enterprises' funding costs would increase as privatization
substantially reduced, or even removed, investors' perception of an
implied federal guarantee on their obligations. On the basis of
available evidence, GAO estimated that the increase in the
enterprises' borrowing costs could range from about 30 basis points
to as much as 106 basis points if the perceived federal guarantee
were completely eliminated. Although harder to estimate, the
enterprises' funding costs for MBS could rise anywhere from 5 to 35
basis points in the absence of a perceived federal guarantee.
By repealing their federal charters, privatization would also
presumably raise the enterprises' costs by eliminating their explicit
charter-based benefits. On the basis of 1995 information, for
example, GAO estimated that elimination of the enterprises'
exemptions from state and local income taxes and SEC registration
fees could raise their costs by $300 to $400 million on an aftertax
basis.
In addition to eliminating or reducing the financial benefits derived
from federal sponsorship, privatization would also presumably remove
the charter-based restrictions on the enterprises' business
activities. Although it is difficult to predict how the enterprises
would respond, GAO believes it reasonable to assume that they would
most likely move into areas that complement their existing business,
such as securitizing jumbo mortgages or other financial assets, or
providing private mortgage insurance.
By increasing their costs and reducing their competitive advantage in
the secondary mortgage market, privatization would likely have an
adverse impact on the enterprises' market shares, profits, and stock
values. For example, the enterprises expect that their potential
profits from holding mortgages in portfolio could be substantially
reduced, if not eliminated, without their funding advantage.
However, the ability to enter new business lines could also bring new
opportunities for profit. Thus, GAO notes that it is especially
difficult to predict the ultimate impact of privatization on the
enterprises' financial condition because it will depend so much upon
such other factors as their strategic business decisions and the
quality of their management.
PRIVATIZATION WOULD INCREASE
MORTGAGE INTEREST RATES AND
AFFECT BEHAVIOR IN THE
RESIDENTIAL MORTGAGE MARKETS
-------------------------------------------------------- Chapter 0:4.2
To some extent, the enterprises are likely to pass along the funding
cost increases resulting from privatization to homebuyers in the form
of higher mortgage interest rates. Specifically, GAO estimated,
partly on the basis of one of the commissioned papers prepared for
this project, that interest rates on single-family, fixed-rate,
conforming mortgages would rise on average by about 15 to 35 basis
points. For $2 trillion in outstanding conventional conforming
fixed-rate mortgages, this would amount to about $3 billion to $7
billion in additional interest costs incurred by households. For a
typical mortgage of about $100,000, this would amount to an increased
monthly payment of between $10 and $25.
The increased cost of funds resulting from privatization should also
allow other firms to compete with the enterprises in securitizing
conforming mortgages. With this increased competition, the
enterprises would likely require MBS guarantee fees that more fully
reflect the risk of the underlying mortgages than they do currently.
As a result, mortgage interest rates would likely rise relatively
more for borrowers making smaller down payments because such
borrowers have historically had higher default rates than borrowers
making larger down payments.
GAO believes that the likely increased participation of these other
firms, along with the continued participation of the
enterprises--even if at a reduced level--could be sufficient to keep
the secondary mortgage market efficient and liquid enough to prevent
the appearance of any significant regional disparities in the
availability and cost of mortgage credit. Evidence on whether the
enterprises have offset cyclical downturns is inconclusive.
Privatization could increase the extent to which regional differences
in risk are reflected in mortgage costs if the enterprises base their
MBS guarantee fees on such risk factors.
In addition to increased competition in the secondary market for
conforming loans, privatization would likely lead to increased
competition in the secondary market for jumbo loans to the extent the
enterprises moved into this area. The effects of privatization would
not necessarily be limited to the secondary market, but could also
affect the primary market. For example, the increased profit
potential of mortgages resulting from the expected rise in mortgage
interest rates could induce some banks and thrifts to hold more of
the mortgages they originate in portfolio rather than to sell them in
the secondary market. To offset the interest rate risk associated
with fixed-rate mortgages, these banks and thrifts could also be
induced to originate more variable rate mortgages. Such mortgages
are not sold in the secondary market as frequently.
PRIVATIZATION WOULD REMOVE
ONE METHOD OF CHANNELING
MORTGAGE CREDIT
-------------------------------------------------------- Chapter 0:4.3
By eliminating the enterprises' federal charters, privatization would
presumably remove one of the available methods used for channeling
mortgage credit to targeted groups of borrowers. In particular, HUD
recently established specific numerical goals for the enterprises in
funding mortgages for very low-, low-, and moderate-income households
and borrowers in underserved areas. For example, in 1996, 40 percent
of each enterprise's purchases are required to serve low- and
moderate-income households. Such regulatory requirements, authorized
by the Federal Housing Enterprise Financial Safety and Soundness Act
of 1992, could conceivably be either partly or entirely eliminated by
privatization. Both enterprises have initiated substantial efforts
to help meet their numerical goals for purchasing loans serving
targeted households, but it is too soon to reliably estimate the
effects of these efforts on housing affordability or the impact they
may have on residential mortgage lending by banks and thrifts to
targeted groups.
According to one of the commissioned studies prepared for this
project, the expected increase in mortgage interest rates resulting
from privatization would also reduce homeownership opportunities for
targeted borrowers and, in particular, would delay homeownership for
households with low but rising incomes. This study also concluded
that interest rates on mortgages used to fund multifamily rental
housing, and therefore rental housing affordability, should not be
affected by privatization.
If Congress chooses to privatize the enterprises, it is possible that
many targeted borrowers would turn to other governmentally assisted
loan sources, such as mortgages insured by the Federal Housing
Administration and securitized through the Government National
Mortgage Association. Congress could also use other existing
mechanisms, such as the Community Reinvestment Act, or new policy
initiatives to help offset any undue impact of privatization on
mortgage financing for targeted borrowers.
POLICY OPTIONS TO REDUCE
TAXPAYERS' RISK OR INCREASE
SOCIAL BENEFITS
-------------------------------------------------------- Chapter 0:4.4
By eliminating the enterprises' federal charters, privatization
should reduce the potential risk that their obligations pose to
taxpayers. Should Congress choose to privatize, GAO believes that it
would be essential to achieve a clear and deliberate break between
the enterprises and the government to substantially reduce or
eliminate investors' perceptions about the implied federal guarantee.
GAO also believes that an effective approach for achieving this
change in investors' perception would be to adopt a structure that
would separate the financial obligations of the previously
government-sponsored enterprises from the obligations of the newly
privatized ones. One such option, under congressional consideration
in the proposed privatization of the Student Loan Marketing
Association, is to have a holding company with one subsidiary
conducting old business and another subsidiary conducting new
business. Other options include having the federal government assume
responsibility for old enterprise obligations or breaking each
enterprise up into several much smaller firms. Each of these two
options could, however, create short-term disruptions in the
financial markets. The break up-option, for example, could reduce
liquidity and impair efficiency in the secondary mortgage market.
GAO also identified alternative policy options short of privatization
that could limit to some extent the taxpayers' risk associated with
the enterprises or increase the social benefits the enterprises
provide. As with privatization, these options would also require
Congress to consider trade-offs between benefits and risks as well as
trade-offs among competing policy goals.
For example, one alternative could be lowering the conforming loan
limit. This would further restrict the range of the market in which
the enterprises would be allowed to operate. Because it would likely
reduce the amount of activity engaged in by the enterprises, it could
reduce the overall level of taxpayers' risk. However, for those
borrowers who would be shifted out of the conforming market and into
the redefined jumbo market, mortgage interest rates would likely
rise. There could be some offset to this increase, however, to the
extent the expanded jumbo market improved the ability of other firms
to regionally diversify their private-label MBS.
As another alternative, the federal government could require
compensation from the enterprises in return for the risk exposure
that their activities generate for the government and the taxpayer.
This compensation could take the form of either user fees based on
debt and MBS issuance or a percentage of income from operations.
Each approach would have pros and cons. For example, while a fee on
the enterprises could raise federal revenues to compensate for
taxpayers' risk, it would also be likely to increase mortgage
interest rates because the fee would increase the cost of funds. To
the extent it reduced the enterprises' funding advantage, it could
also increase the likelihood that other firms would securitize
conforming mortgages.
These and other alternatives are discussed in greater detail in
chapter 5.
RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:5
GAO is making no recommendations.
ENTERPRISE COMMENTS AND GAO'S
EVALUATION
---------------------------------------------------------- Chapter 0:6
GAO received comments from both Fannie Mae and Freddie Mac on a draft
of this report. Officials from both enterprises stated that GAO's
estimates of the benefits associated with government sponsorship were
overstated, suggested that GAO provide an overall estimate of
enterprises' benefits to the mortgage market, and asked that GAO
indicate that most of the benefits were passed on to homebuyers.
Both enterprises emphasized their importance in maintaining liquidity
in the conventional mortgage market, and they suggested that GAO was
not accounting sufficiently for the possibility of reduced liquidity
should privatization occur. Freddie Mac also suggested that
privatization would increase regional disparities and cyclical
variations in the availability and cost of mortgage credit. Both
enterprises asserted that they compete vigorously and that there is
no evidence that they have any market power. Fannie Mae officials
suggested that GAO should have concluded that their efforts in
promoting home ownership among targeted groups had been successful.
On the basis of the enterprises' technical comments, GAO reduced its
estimate of the benefits associated with government sponsorship by
lowering its upper range for the enterprises' funding advantage on
debt from 120 basis points to 106 basis points. GAO also included,
as a measure of the enterprises' benefits to the mortgage market, an
overall estimate of $3 billion to $7 billion in increased mortgage
costs that would be incurred by households if privatization occurred.
GAO also pointed out that it had no way to quantify how much of the
benefits associated with government sponsorship are passed on to
homebuyers. GAO agreed that the possible effects of privatization on
liquidity both in terms of cost and cyclical availability are
important and clarified its text in this regard. GAO reiterated that
it cannot determine whether the enterprises do or do not have market
power. In addition, GAO continues to believe that, while the
enterprises have made efforts to provide additional funding to
targeted groups, it is still too early to measure a clear effect.
INTRODUCTION
============================================================ Chapter 1
The Federal National Mortgage Association (Fannie Mae) and Federal
Home Loan Mortgage Corporation (Freddie Mac)(referred to in this
report jointly as the enterprises) are government-sponsored
enterprises that play important roles in federal support of home
ownership and America's housing finance system.\1
The primary role of the enterprises is to ensure that mortgage funds
are available to home buyers in all regions of the country at all
times. Congress has asked us to study the desirability and
feasibility of repealing the federal charters of the enterprises,
eliminating any federal sponsorship of the enterprises, and allowing
the enterprises to continue to operate as fully private entities.\2
--------------------
\1 The enterprises are only part of the federal support for home
ownership; the most significant additional federal support comes from
provisions of the tax code and activities of other federal entities.
The tax code supports home ownership by permitting taxpayers, who
itemize, to deduct their mortgage interest payments from their
adjusted gross income. In addition taxpayers are permitted to defer
payment of tax on capital gains on the sale of their houses and may
exclude, one time, up to $125,000 in gain on sale when they are 55 or
older. Various federal agencies support home ownership. For
example, the Federal Housing Administration (FHA) and Department of
Veterans Affairs (VA) lower ownership costs by insuring mortgages
with favorable terms for qualified individuals. In addition, the
Federal Home Loan Bank System lends to mortgage lenders so they can
originate and fund mortgages.
\2 A fully private entity would have no special ties to the federal
government. It would not have special market restrictions, operating
advantages, or exemptions from certain taxes or fees. In addition, a
fully private entity should have no special ties to the government
that would lead financial market participants to perceive an implied
federal guarantee of the entity's financial obligations. We define
privatization as an action that severs all special ties and
restrictions as well as eliminating or substantially reducing
investors' perception of an implied federal guarantee.
BACKGROUND
---------------------------------------------------------- Chapter 1:1
The enterprises help ensure that mortgage funds are available to home
buyers by buying mortgages from mortgage originators, such as savings
and loans (thrifts), commercial banks, and mortgage bankers. The
enterprises hold some of these mortgages in portfolio as direct
investment on their own books and issue debt and equity securities to
finance these holdings.
Most mortgages that the enterprises buy from mortgage originators are
"securitized"--that is, the enterprises package them into mortgage
pools to support mortgage-backed securities (MBS). These mortgage
pools receive the interest and principal payments from the mortgages
in the pools and pass them on to the investors who purchased the MBS.
The enterprises guarantee the timely payment of principal and
interest payments from the mortgages in the pools to the investors
and administer the payments.\3 In September 1995, Fannie Mae and
Freddie Mac either owned in portfolio or guaranteed about $1.3
trillion of the $3.9 trillion of outstanding residential mortgages in
the United States.\4
The enterprises are government-sponsored in that they operate under
federal charters that convey certain benefits, impose certain
restrictions, and permit the enterprises to earn a profit while
serving public policy purposes, such as providing liquidity\5 to
mortgage markets. In 1992, Congress expanded the enterprises' public
purpose by requiring annual goals that are to be set, monitored, and
enforced by the Department of Housing and Urban Development for the
purchase of mortgages on housing purchased by very low-, low-, and
moderate-income and other households that are underserved by the
residential mortgage market.\6 The enterprises' charters exempt them
from certain fees and taxes paid by other private sector firms. At
the same time, the charters restrict the enterprises to buying
mortgages that do not exceed a set dollar amount, known as the
conforming loan limit.\7
A major factor that enhances the enterprises' profitability is the
financial market's perception that there exists an implied federal
guarantee of their debt and other obligations (i.e., a perception
that the federal government would act to ensure that the enterprises
will always be able to meet their financial obligations on their debt
and MBS guarantees). Investors perceive that this implied guarantee
decreases the risk that the enterprises will ever fail to meet their
financial responsibilities. Consequently, this perception lowers the
enterprises' borrowing costs because investors are willing to accept
lower expected returns on enterprise debt than they would for private
firms without government ties. Likewise, funding costs on MBS are
also lowered by this perception.\8 Their lower funding costs allow
the enterprises to increase their purchases and give them a cost
advantage over competitors. This perception of a federal guarantee
remains even though the laws chartering the enterprises contain
explicit language stating that there is no such guarantee.
The perception of the implied guarantee is based on special federal
ties to the enterprises, including government-sponsored status, each
enterprise's $2.25 billion conditional line of credit with the
Treasury Department, and a belief that the federal government would
consider such large institutions too big to fail.\9
The federal charter also provides several explicit provisions that
lower operating costs for the enterprises. For example, certain fees
paid by other corporations to the Securities and Exchange Commission
(SEC) are not levied against the enterprises since the enterprises do
not need to register their issuances with the SEC. They are also
exempt from state and local income taxes. In addition, they can use
the Federal Reserve's electronic payments system for transactions.
These privileges, plus each enterprise's $2.25 billion conditional
line of credit with the Treasury, reinforce the market's perception
that the government will not let the enterprises fail. Given the
lower funding costs created by this perception and the lower
operating costs created by certain privileges and exemptions, the
enterprises have cost advantages over any potential direct
competitor.\10
--------------------
\3 The MBS are considered to be "off book" or "off the balance sheet"
of the enterprises. The enterprises collect fees on MBS for carrying
out these functions.
\4 Year-end 1995 data on total residential mortgage holdings by all
financial sectors were not available.
\5 A market is more liquid if investors can buy and sell large
amounts of holdings without affecting the prices of the traded
securities.
\6 Federal Housing Enterprises Financial Safety and Soundness Act of
1992 ("1992 Act") Pub. L. No. 102-550, Title XIII.
Housing-related goals are contained in subpart B of Title XIII.
\7 The conforming loan limit depends on how many housing units are
financed by a single residential mortgage loan. Currently the
conforming loan limit on a single-unit residence is $207,000.
\8 Investors will accept lower expected returns on enterprise MBS,
just as for enterprise debt, because of the perception of an implied
federal guarantee. This in turn lowers the cost of funding mortgages
through issuance of MBS.
\9 "Too big to fail" is a common way to express the idea that the
failure of certain institutions that could have economywide
consequences that could require the federal government to forestall
normal bankruptcy and liquidation processes. This intervention
decision would not depend on a formal prior declaration by the
federal government that it would bail out a large failing
institution. In the past, the federal government has intervened to
prevent the failure of New York City, Lockheed, and Chrysler. It
also intervened in the bankruptcy of Penn Central and created
Conrail. In each case the failing institution had no guarantee, but
federal interests were considered important enough to intervene in
the normal bankruptcy resolution. See Guidelines for Rescuing Large
Failing Firms and Municipalities (GAO/GGD-84-34, Mar. 29, 1984).
\10 See FNMA and FHLMC: Benefits Derived From Federal Ties
(GAO/GGD-96-98R, March 25, 1996).
MORTGAGES ARE FUNDED BY PRIMARY
AND SECONDARY MARKETS
---------------------------------------------------------- Chapter 1:2
The mortgage market is made up of primary and secondary parts; and
many institutions serve several roles within the overall market, as
shown in tables 1.1 and 1.2. Consequently, institutions sell to, buy
from, and compete with each other. As shown in table 1.2, the
enterprises function as conduits and guaranteeing agencies in the
secondary mortgage market.
Table 1.1
The Roles, Functions, and Participants
in the Primary Mortgage Market
Role Function Participants
------------------------------ ------------------ ------------------
Home buyers Apply for mortgage -Individual
consumers
Originators Receive -Depositories such
applications and as banks or
make lending savings and loans
decisions -Mortgage bankers
Mortgage insurers If the home buyer -Federal Housing
defaults, provide Administration
compensation to -Veterans Affairs
originators who Department
keep the mortgages -Private mortgage
as investments insurance
companies
----------------------------------------------------------------------
Source: GAO analysis based on The Revolution in Real Estate Finance
by Anthony Downs (Brookings Institution, 1985).
Table 1.2
The Roles and Functions of Participants
in the Secondary Mortgage Market
Role Function Participants
------------------------------ ------------------ ------------------
Originators Sell mortgages to -Depositories
investors and -Mortgage bankers
conduits that
create MBS
Investors Buy mortgages and -Depositories
MBS -Housing
enterprises
-Other financial
institutions such
as life insurers
and pension funds
-Individuals
Conduits Buy mortgages sold -Freddie Mac
by originators and -Fannie Mae
others, and create -Private conduits
and manage MBS
Mortgage insurers Compensate -Federal Housing
investors if the Administration
home buyers -Veterans Affairs
default Department
-Private mortgage
insurance
Companies
Guaranteeing agencies Guarantee timely -Fannie Mae
payment of -Freddie Mac
principal and -Government
interest to National Mortgage
investors from Association
mortgages in the -Private conduits
pools
----------------------------------------------------------------------
Source: GAO analysis based on The Revolution in Real Estate Finance.
In the primary market, the home buyer applies to an originator for a
mortgage. The originator can be a depository, such as a bank or
thrift, or a mortgage banker. Traditionally, depositories originated
mortgages and held them as direct investments in portfolio on their
books. Their profits from holding mortgages were the difference
between interest earned from the mortgages and their costs of funds,
primarily interest paid to depositors after adjusting for other
expenses.
Mortgage bankers originate mortgages for immediate resale in the
secondary market. They earn profits primarily from two sources. The
first source is fees charged to originate mortgages and profits from
the sale of mortgages (losses can also result from such sales). The
second source is fees investors pay to mortgage bankers for
"servicing" mortgages--collecting and processing mortgage payments.
In recent years many depositories have also acted like mortgage
bankers in that they originate mortgages and sell them to investors
rather than hold them on their books.
Mortgage insurers improve the liquidity\11 of the market by
compensating investors for losses caused by mortgage defaults--
losses created when the net sales price of the house after
foreclosure does not cover the outstanding balance on the mortgage.
This compensation reduces risks and makes the market more liquid.
FHA and VA are the primary federal government insurers. The private
mortgage insurance companies provide insurance for conventional
mortgages--that is, mortgages not backed by the federal government.
The secondary mortgage market channels mortgages from originators to
investors. The Government National Mortgage Association (Ginnie
Mae), the enterprises and other private companies, acting as
conduits, create mortgage pools and MBS that are sold to investors.
From a pool of mortgages, the MBS investors receive their
proportional shares of interest and principal flows.
Private-label MBS are created by fully private
(nongovernment-sponsored) conduits. As of September 1995,
private-label MBS totalled about 13 percent of outstanding MBS. The
mortgages that these private conduits securitize either exceed the
enterprises' conforming loan limit--$207,000 on one-unit,
single-family properties--or do not meet the enterprises'
underwriting standards.\12 A loan whose underwriting standards do not
meet the standards of either enterprise or exceeds the conforming
loan limit is called a nonconforming loan. A loan that exceeds the
conforming loan limit is called a jumbo loan.
Guarantees on MBS enhance the liquidity of the secondary market.
Ginnie Mae guarantees timely payment of principal and interest for
mortgage pools of FHA and VA insured mortgages for a fee. The
enterprises and private-label conduits guarantee timely payment of
principal and interest on conventional mortgages in pools backing
their MBS. The guarantees are an enhancement that reduces the risk
that any given mortgage will not be paid on a timely basis.
Private-label conduits generally use risk-based guarantee fees, which
are based on the expected incremental cost of guaranteeing a
particular level of credit risk exposure for the investor. For
example, the conduits charge lower fees on mortgages with large down
payments (i.e., mortgages with low loan-to-value ratios) than on
loans with small down payments. The enterprises said that their
mortgage commitment policies move them partially, but not fully,
toward a risk-based fee structure.
Private-label conduits may enhance the liquidity of their MBS with
other credit enhancements. Private mortgage insurance is a common
form of credit enhancement to reduce risk. Another common private
label credit enhancement is over-collateralization. This means that
the principal amount of the mortgages backing the MBS exceeds the
dollar amount of the MBS shares sold to investors. Other forms of
credit enhancement for the private-label MBS include bank letters of
credit,\13 corporate guarantees,\14 and private insurance of mortgage
pools.
The cash flows to investors generated by interest and principal
payments from any pool may vary over time. As interest rates fall,
households tend to prepay principal more quickly, which is called
prepayment risk. As interest rates increase, prepayments tend to
slow down and cash flows to investors decline, since homeowners are
refinancing or selling their houses more slowly. This tendency is
called extension risk.
--------------------
\11 The provision of market liquidity is an important function of
mortgage insurers and secondary market conduits.
\12 Underwriting standards are used by the enterprises to determine
which mortgages they will buy as investments or place into mortgage
pools. The standards limit the risk that the mortgages will create
losses for the pools or the enterprises. The standards are meant to
ensure the buyer has the ability to pay; the buyer is creditworthy
and is likely to meet scheduled payments; and, in the event of a
default, the value of the house and mortgage insurance limits any
losses.
\13 A letter of credit is a guarantee by a bank to pay the principal
and interest due on an MBS if the conduit fails to do so. It is a
form of insurance for the investors.
\14 A corporate guarantee is a guarantee by the corporate parent of
the conduit to pay the principal and interest due on an MBS if the
conduit fails to do so. It is a form of insurance for the investors.
MULTICLASS SECURITIES HAVE
IMPROVED LIQUIDITY OF THE
SECONDARY MORTGAGE MARKET
-------------------------------------------------------- Chapter 1:2.1
To address the need for more predictable cash flows, the enterprises
and private-label conduits issue multiclass mortgage securities
called collateralized mortgage obligations (CMOs) and Real Estate
Mortgage Investment Conduits (REMICS). These multiclass securities
can help investors better manage prepayment and extension risks by
creating from the same mortgage pool several securities that receive
different parts of the pool's interest and principal payments.
Investors more concerned about variations in cash flows over time can
buy the classes that pay off more quickly or have a fixed payment
period. Investors more willing to undertake prepayment and extension
risks can buy classes with payments that vary with interest rates.
The expected return on classes with prepayment and extension risks
exceeds the expected return on classes without such risk. Multiclass
securities can also redistribute credit risk so that one class can be
designed to absorb all or much of the credit risk in return for a
higher expected return. Because multiclass securities bring new
investors who wish to avoid unpredictable cash flows into the market,
they improve the market's liquidity and help ensure continuing
funding for home mortgages. The new investors that multiclass
securities have attracted include banks, thrifts, pension funds,
insurance companies, and other financial institutions as well as
individuals who originate, buy, hold, or sell whole mortgages.
MORE THAN 45 PERCENT OF ALL
RESIDENTIAL MORTGAGE DEBT IN
SEPTEMBER 1995 WAS
SECURITIZED AND HELD IN MBS
-------------------------------------------------------- Chapter 1:2.2
Tables 1.3 and 1.4 show the different sectors of the housing finance
system. The total of all residential mortgage debt in September 1995
was $3.9 trillion. About 6 percent was held by the enterprises in
portfolio, 33 percent was held in portfolio by wholly private
financial institutions, 45 percent was securitized and held by
various types of investors in MBS, 1 percent was held by the federal
government or related agencies,\15 and the rest was held by
individuals and other investors.
Commercial banks and thrifts were significant holders of whole
mortgages and MBS. In September 1995, banks held about 33 percent of
whole mortgages and savings and loans held about 26 percent of all
whole mortgages. At year-end 1995, commercial banks held 20 percent
of all MBS and thrifts held about 10 percent of MBS. Other major
investors included life insurance companies and mutual funds.
Table 1.3
Residential Mortgage Debt Outstanding as
of September, 1995
Value (in
millions) Percent of total
------------------------------ ------------------ ------------------
Housing enterprises (in portfolio)
----------------------------------------------------------------------
Fannie Mae $182,229 4.71%
Freddie Mac 42,678 1.10
======================================================================
(Total) 224,907 5.81
Financial institutions
Commercial banks 705,844 18.25
Savings institutions 552,144 14.28
Life insurance companies 32,607 .84
======================================================================
(Total) 1,290,595 33.37
Federally related agencies
----------------------------------------------------------------------
Ginnie Mae 2 0.0
Farmers Home Administration 24,471 .63
FHA and VA 9,535 .25
Resolution Trust Corporation 3,719 .10
Federal Deposit Insurance 1,340 .03
Corp.
Federal Land Banks 1,656 .04
======================================================================
(Total) 40,723 1.05
Mortgage pools
----------------------------------------------------------------------
Ginnie Mae 463,654 11.99
Freddie Mac 503,457 13.02
Fannie Mae 559,585 14.47
Private mortgage conduits 228,616 5.91
Farmers Home Administration 2 0.00
======================================================================
(Total) 1,755,314 45.39
Individuals and others\a
----------------------------------------------------------------------
======================================================================
(Total) 556,108 14.38
======================================================================
Grand Total 3,867,647 100.00
----------------------------------------------------------------------
\a Others include mortgage companies, real estate investment trusts,
state and local credit agencies, state and local retirement funds,
noninsured pension funds, credit unions, and finance companies.
Source: Federal Reserve Bulletin, March 1996.
Table 1.4
Percent of All Mortgage-Backed Security
Holdings by Investor Type as of Year-
End 1995
Percent
of total
Investor type MBS
------------------------------------------------------------ --------
FDIC commercial banks 20.02%
FDIC savings banks 2.67
OTS regulated S&Ls 9.64
Federal credit unions 1.12
Federal Home Loan Banks 2.24
Public pension funds 7.97
Private pension funds 3.66
Life insurance companies 14.88
Mutual funds 5.04
Private individuals 1.56
Real Estate Investment Trusts .45
Foreign investors 9.68
Dealer inventory 5.30
Housing enterprises' portfolios 7.88
All other investors 7.88
======================================================================
Total 100.00
----------------------------------------------------------------------
Source: Inside Mortgage Securities, March 1, 1996.
--------------------
\15 Federally related agencies are corporations and GSEs with federal
ties, such as FHA, VA and the Federal Deposit Insurance Corporation.
THE ENTERPRISES HAVE EVOLVED
WITH CHANGES IN ECONOMIC
CONDITIONS AND PUBLIC POLICY
---------------------------------------------------------- Chapter 1:3
The enterprises have evolved since Congress created Fannie Mae to
remedy the housing market effects of the Great Depression of the
1930s and Freddie Mac was created in 1970. Modifications in their
charters have occurred as the result of changing economic conditions
and government policies.
THE GREAT DEPRESSION LED TO
FEDERAL ASSISTANCE TO
HOUSING MARKETS
-------------------------------------------------------- Chapter 1:3.1
In the wake of the Great Depression of the 1930s, the federal
government took steps to revive the economy, stabilize financial
markets, and ensure mortgage markets were liquid. The government's
response concentrated on the savings and loan industry, which was
then the backbone of the housing finance system. Congress created a
thrift regulator to ensure the safety and soundness of the thrift
industry; to bolster consumer confidence and keep deposits flowing
into the thrifts, it created a deposit insurance system. In
addition, Congress created the Federal Home Loan Banks, which
borrowed in capital markets\16 and made loans to thrifts so that they
could continue to fund and originate mortgages. To further support
housing, Congress created FHA, which insured mortgages originated by
private financial institutions and reduced credit risk for investors.
Congress also authorized the establishment of private mortgage
associations to create a secondary market for mortgages. Because
private mortgage associations did not develop, Congress chartered
Fannie Mae in 1938 as a government-held association to buy and hold
mortgages insured by FHA. Later it was authorized to purchase
VA-insured mortgages. In its early years, Fannie Mae was part of the
Reconstruction Finance Corporation and subject to the regulation of
the Federal Housing Administration. Modifications in Fannie Mae's
structure occurred during the post-War period without changing its
fundamental mission.
In the early post-World War II period, Congress articulated Fannie
Mae's purposes as
-- providing liquidity and special assistance for selected housing
types,
-- supporting the mortgage market, and
-- stabilizing the economy.
Fannie Mae's mortgage purchases increased substantially during most
of the 1950s. During the late 1950s through the mid-1960s, Fannie
Mae sold mortgages when other sources of credit were readily
available or purchased mortgages when credit was tight. After 1968,
Fannie Mae's, and later Freddie Mac's, portfolios grew.
--------------------
\16 Capital markets provide long-term funding through debt and stock
issuances.
VOLATILE INTEREST RATES AND
INCREASING HOUSE PRICES
AFFECTED THE HOUSING
MORTGAGE MARKETS
-------------------------------------------------------- Chapter 1:3.2
In the Housing and Urban Development Act of 1968,\17 Congress split
Fannie Mae into two components. One component, Ginnie Mae, remained
in HUD to provide support to FHA, VA, and special assistance
programs. The other component was the government-sponsored,
privately owned, for-profit Federal National Mortgage Association,
which was to be concerned exclusively with attracting funding into
residential mortgages. Thus, the newly private, yet
government-sponsored, Fannie Mae continued to provide a secondary
residential mortgage market and was governed by a board of directors
dominated by its private sector owners with a minority of its members
(5 of 18) appointed by the president. Fannie Mae was regulated by
the Department of Housing and Urban Development in terms of capital
requirements and approval of new mortgage acquisition programs.
Ginnie Mae and Fannie Mae operated differently. Ginnie Mae did not
purchase mortgages. Instead, it "guaranteed the timely payment of
principal and interest" from pools of FHA- and VA-insured mortgages
originated by mortgage bankers and other financial institutions. In
contrast, Fannie Mae operated as a large portfolio investor. It
bought mortgages from originators and financed these investments by
selling debt and equity in the financial markets.
Congress permitted Fannie Mae to develop a secondary market for
conventional loans to counter periodic scarcities of mortgage credit
in different regions of the country during different parts of the
business cycle. Consequently, Fannie Mae helped counter a scarcity
of mortgage credit during the late 1960s and early 1970s, when
interest rates paid by thrifts and other depository institutions were
capped--sometimes below market levels. In response to these
below-market rates, depositors withdrew funds and looked for higher
returns elsewhere. As funds were withdrawn, thrifts were unable to
originate or fund mortgages. At the same time, other originators
such as mortgage bankers were able to originate mortgages at market
rates and sell them to Fannie Mae. Since Fannie Mae did not have an
interest rate cap, it could raise funds at market rates and thus
continue to purchase mortgages at current market rates from all
originators.
Congress chartered Freddie Mac in 1970 in reaction to the loss of
deposits in the savings and loan industry that was curtailing that
industry's ability to fund and originate home mortgages.\18
Its creation ensured that the savings and loan industry had access to
funds to continue to fund mortgages. Freddie Mac was first owned by
the Federal Home Loan Bank Board, which regulated savings and loans,
helped fund their operations through the Federal Home Loan Banks,
provided deposit insurance to the thrifts through the Federal Savings
and Loan Insurance Corporation, and liquidated insolvent thrifts.
Freddie Mac mostly securitized the mortgages that it purchased and
guaranteed timely interest and principal payments from the resulting
mortgage pools.
Originally, the enterprises and FHA had identical conforming loan
limits for mortgages they could purchase or guarantee. In 1974,
Congress raised the conforming loan limit for both enterprises above
FHA's limit. Consequently, Fannie Mae and Freddie Mac could buy an
increasing share of mortgages that were not provided by, or
guaranteed by, the federal government.
In 1981, Congress created a formula for adjusting the conforming loan
limit to account for the effects of inflation on house values. A
three-tiered secondary mortgage market evolved in the late 1980s.
Ginnie Mae primarily served a tier of lower value FHA and VA
mortgages. The enterprises primarily served a middle tier of larger
mortgages. The private-label conduits served a tier of jumbo--loans
with principal amounts that exceeded the conforming limit--and other
conventional, nonconforming mortgages.
In the early 1980s, Fannie Mae and Freddie Mac experienced different
financial results as short-term interest rates increased. Fannie Mae
held mortgages in portfolio and funded them with short-term debt. As
rates increased, Fannie Mae had to issue new short-term debt at
higher rates to replace existing short-term debt that came due.
Because interest earned on the old mortgages in portfolio was less
than interest expenses on the newly issued debt, Fannie Mae
experienced total losses of about $277 million between 1981 and 1984.
In response to Fannie Mae's financial problems, the federal
government provided limited tax relief and regulatory forbearance in
the form of relaxed capital requirements.
Unlike Fannie Mae, Freddie Mac held few mortgages in portfolio and
issued little debt to fund mortgage holdings. Rather, it created MBS
and sold them to investors. Consequently the investors and not
Freddie Mac bore the risks of changing interest rates. To avoid
future losses from interest rate changes, Fannie Mae partially
adopted Freddie Mac's strategy of issuing MBS and passing interest
rate risk to investors.
--------------------
\17 Pub. L. No. 90-448
\18 Federal Home Loan Mortgage Corporation Act, 12 U.S.C. �1451 et
seq., originally enacted as Title III of the Emergency Home Finance
Act of 1970.
MORTGAGE BANKERS AND THE
ENTERPRISES GAINED
IMPORTANCE IN MORTGAGE
MARKETS AS MANY THRIFTS
FAILED
-------------------------------------------------------- Chapter 1:3.3
The unexpected increase in interest rates in 1979 through 1981 that
created problems for Fannie Mae also contributed to the failure of
many thrifts in the 1980s. As interest rates rose, many thrifts
became unprofitable, and some thrifts hoping to regain profitability
undertook risky investments as their losses grew. In many of these
cases, such actions accelerated and increased losses to the thrift
deposit insurance fund, the Federal Savings and Loan Insurance
Corporation (FSLIC). At the same time, FSLIC did not have the
resources to close all insolvent thrifts. As the weakened thrifts
deteriorated further, closure costs continued to increase.
In 1989, Congress abolished the Federal Home Loan Bank Board and
dispersed its functions to other agencies. The Office of Thrift
Supervision became the regulator of federally chartered savings and
loans. Freddie Mac became a government-sponsored enterprise owned by
private investors. Deposit insurance for thrifts went to the Savings
Association Insurance Fund under the Federal Deposit Insurance
Corporation (FDIC), and the Resolution Trust Corporation (RTC) was
created to close and liquidate insolvent thrifts that were still open
when RTC was created.
As thrifts failed and the thrift industry's originations and holdings
of mortgages decreased, mortgage bankers originated more mortgages
and mortgage conduits increased their issuance of MBS. As shown in
figure 1.1, the importance of mortgage bankers as originators
increased as that of thrifts decreased. In 1982, thrifts originated
35.9 percent of all mortgages on 1-4 family units (commonly called
single-family units); however, their
Figure 1.1: Single-Family
Mortgage Originations
(See figure in printed
edition.)
share had dropped by 1994 to 15.9 percent. In 1982, the mortgage
bankers originated 28.9 percent of all mortgages for single-family
units, and by 1994, their share had increased to 52.8 percent.
Mortgage originations were no longer strongly tied to the thrift
industry.
Not only did thrifts become less prominent as originators, they also
held less mortgage debt directly in portfolio. As shown in figure
1.2, the conduits and especially the enterprises became an
increasingly important mechanism for channeling residential mortgage
funds. In 1982, thrifts held in portfolio 36.6 percent of all
outstanding mortgages on single-family units (their holdings of MBS
were not reported). By 1994, thrifts held directly, in portfolio,
only 14.3 percent of outstanding mortgages on single-family units.
Much of this shrinkage of direct mortgage holdings was accounted for
by the growth of the enterprises' activities. By 1994, the
enterprises held in portfolio 6.7 percent of all mortgages on
single-family units, and their MBS represented 29.5 percent of the
outstanding single-family unit mortgages. However, the thrifts
continued to hold mortgages indirectly since they held MBS created by
the enterprises and other conduits.
Figure 1.2: Single-Family
Mortgages Outstanding (Dollars
in Millions)
(See figure in printed
edition.)
CONGRESS ALTERED THE HOUSING
ENTERPRISES' CHARTERS TO
EXPAND PUBLIC POLICY
PURPOSES AND LIMIT RISKS
-------------------------------------------------------- Chapter 1:3.4
The effect of the 1992 Act, in combination with the GSE-related
provisions in the Financial Institutions Reform, Recovery, and
Enforcement Act of 1989 (FIRREA), was to make the charters of the
enterprises substantially the same. Provisions of the enterprises'
charters, which remain in force today, include the following broad
public policy purposes:
-- provide stability in the secondary market for residential
mortgages;
-- respond appropriately to private capital markets;
-- provide ongoing assistance to the secondary market for
residential mortgages (including activities relating to
mortgages for very low-, low-, and moderate-income households
involving a reasonable economic return that may be less than the
return earned on other activities) by increasing the liquidity
of mortgage investments and improving the distribution of
investment capital available for mortgage financing; and
-- promote access to mortgage markets throughout the nation
(including central cities, rural areas, and underserved areas)
by increasing the liquidity of mortgage investments and
improving the distribution of investment capital available for
residential mortgage financing.
In the 1992 Act, Congress created the Office of Federal Housing
Enterprise Oversight (OFHEO). OFHEO was to regulate the enterprises
for safety and soundness and set capital standards for the
enterprises, and HUD was authorized to establish, monitor, and
enforce mortgage purchasing goals for the enterprises.
THE ENTERPRISES' CHARTERS
CONFER BENEFITS AND IMPOSE
RESTRICTIONS
---------------------------------------------------------- Chapter 1:4
Although the enterprises are privately owned and for profit, their
charters impose restrictions and confer benefits that affect their
ability to make profits. The enterprises are specifically authorized
to deal in conventional residential mortgages under the conforming
loan limit; other kinds of business are not so authorized. Other
restrictions include goals set by the Secretary of HUD for the dollar
volume of mortgages that the enterprises must purchase from very
low-, low-, and moderate-income households and underserved rural and
urban areas.
The benefits provided to each enterprise include
-- a $2.25 billion conditional line of credit with the U.S.
Treasury;
-- an exemption from paying state and local corporate income taxes;
-- an exemption from registering their securities with the
Securities and Exchange Commission (SEC), which means they do
not pay SEC fees; and
-- the ability to use the Federal Reserve as a transfer agent,
which enhances the enterprises' operating efficiency.
HOUSING ENTERPRISES HAVE
GENERALLY DONE WELL FINANCIALLY
---------------------------------------------------------- Chapter 1:5
Although operating with the restrictions and benefits established by
the government, the enterprises have been consistently profitable
since the mid-1980s. (See table 1.5.)
Table 1.5
1995 Financial Performance of the
Housing Enterprises
(Dollars in billions)
Fannie Mae Freddie Mac
------------------------------ ------------------ ------------------
Total assets $316.55 $137.18
Mortgage holdings\a 252.59 107.41
MBS outstanding\a 513.23 459.05
Stockholder equity 10.96 5.86
Net income 2.14 1.09
Return on equity as of year 19.53% 18.60%
end
Equity as a percentage of 1.32% .98%
total assets plus MBS
outstanding at year end
----------------------------------------------------------------------
\a MBS outstanding excludes MBS held in portfolio, and mortgage
holdings include MBS held in portfolio.
Sources: Fannie Mae Investor/Analysts Report for fourth quarter 1995
and Freddie Mac news release dated January 16, 1996.
As of year-end 1995, Fannie Mae's assets exceeded $316 billion, and
Freddie Mac's $137 billion. In addition, Fannie Mae's outstanding
mortgage holdings exceeded $252 billion, and Freddie Mac's exceeded
$107 billion. Although Freddie Mac historically retained relatively
fewer mortgages than Fannie Mae, Freddie Mac has in recent years
increased its share of mortgages held in portfolio. At year-end
1995, Fannie Mae mortgage holdings were about 80 percent of its total
assets and 23 percent of its total mortgage servicing portfolio--the
sum of mortgage holdings and MBS outstanding. Freddie Mac's mortgage
holdings were about 78 percent of its total assets and 19 percent of
its total servicing portfolio. In 1995, Fannie Mae's return on
equity was 19.53 percent and Freddie Mac's was 18.60 percent. In
1995, Fannie Mae's equity ratio (equity divided by the sum of total
assets and MBS outstanding) was 1.32 percent, and Freddie Mac's was
.98 percent.
THE ENTERPRISES ARE FACED WITH
FOUR MAJOR TYPES OF RISK
---------------------------------------------------------- Chapter 1:6
While earning profits, Fannie Mae and Freddie Mac must deal with four
major types of risk: business, interest rate, credit, and management
risk.
BUSINESS RISK
-------------------------------------------------------- Chapter 1:6.1
Business risk is the possibility of financial loss due to conditions
within the market or markets in which a firm operates. Because the
enterprises serve the secondary market for conforming mortgages,
their financial health depends on the factors that create a healthy
secondary market for such mortgages. If profits decline or risks
increase in this limited market, the enterprises cannot avoid
associated problems by exiting their current market and entering new
markets.
INTEREST RATE RISK
-------------------------------------------------------- Chapter 1:6.2
Interest rate risk is the possibility of financial loss due to
changes in market interest rates. Movements in market interest rates
can affect interest expenses, interest earnings, prepayments by
homeowners, and the value of assets and liabilities on the balance
sheet.
Rising market interest rates increase interest expenses as debt turns
over and decrease the value of existing assets that are paying a
below-market rate. As discussed earlier, Fannie Mae experienced this
problem in the early 1980s as its interest expenses increased and
interest earnings on its existing pool of mortgages were relatively
constant.
When market interest rates decline, homeowners tend to prepay
mortgages more quickly, resulting in a decrease of the net average
interest rate received by the enterprises on mortgages held in
portfolio. The net rate decreases even if new lower rate mortgages
are bought by the enterprises as long as the interest rate paid on
outstanding debt does not change. At the same time, if prepaid
mortgages are not replaced with new lower rate mortgages, the
enterprises' outstanding debt balance could exceed their mortgage
balances. Whether or not the enterprises replace prepaid mortgages
with new lower rate mortgages, they face interest rate risk.
The enterprises limit interest rate risk in several ways. First, the
enterprises avoid interest rate risk by passing it to investors when
they create MBS. Second, both enterprises limit interest rate risk
by issuing callable bonds that can be paid off early if rates fall.
By calling the bonds and issuing new debt as interest rates fall, the
enterprises curtail interest rate expenses. Conversely, if rates
increase, the enterprises continue to pay below-market rates on their
existing bonds. Callable bonds are one example of how the
enterprises manage their liabilities to hedge interest rate risk
associated with their asset holdings. The enterprises have also
developed other methods, including certain derivative products, to
control their interest expenses as the economy varies.
CREDIT RISK
-------------------------------------------------------- Chapter 1:6.3
Credit risk is the possibility of financial loss resulting from
default by homeowners on housing assets that have lost value. Credit
risk on mortgages is the possibility that mortgages will go into
default, and the net recoveries from selling the property and
collecting private mortgage insurance will not cover outstanding
balances. This risk occurs when the enterprises hold mortgages in
portfolio and when they guarantee principal and interest payments to
investors in their MBS.
Primary determinants of credit risk are the homeowner's payment
burden, the homeowner's creditworthiness, the size of the down
payment, and the existence of private mortgage insurance. The first
three factors affect whether the applicant can and will make timely
mortgage payments. The size of the down payment and the existence of
private mortgage insurance (PMI) affect the size of any loss in the
event of a default. A larger down payment and PMI increase the
likelihood that the house can be sold after foreclosure for an amount
that is sufficient to recover the outstanding mortgage balance.
MANAGEMENT RISK
-------------------------------------------------------- Chapter 1:6.4
Management risk is the possibility of financial loss resulting from a
management mistake that can threaten the company's viability.
Careful oversight by the company's board, stockholders, financial
markets, and regulators can help ensure that management risk is
adequately controlled.
OBJECTIVES, SCOPE, AND
METHODOLOGY
---------------------------------------------------------- Chapter 1:7
Section 1355 of the 1992 Act mandated us, the Congressional Budget
Office (CBO), the Department of Housing and Urban Development (HUD),
and the Department of the Treasury to separately study and report on
"the desirability and feasibility of repealing the federal charters
of the Federal National Mortgage Association and the Federal Home
Loan Mortgage Corporation, eliminating any federal sponsorship of the
enterprises, and allowing the institutions to continue to operate as
fully private entities." This report is our response to that mandate.
We and the other agencies were directed to examine the effects of
privatization on
-- the requirements imposed upon and costs to the enterprises,
-- the cost of capital to the enterprises,
-- housing affordability and availability and the cost of
homeownership,
-- the level of secondary mortgage market competition subsequently
available in the private sector,
-- whether increased amounts of capital would be necessary for the
enterprises to continue operation,
-- the secondary market for residential loans and the liquidity of
such loans, and
-- any other factors each of the agencies deemed appropriate.
In addition to the legislative mandate, we had discussions with staff
of the Subcommittee on Capital Markets, Securities, and Government
Sponsored Enterprises of the House Committee on Banking and Financial
Services. In these discussions staff asked us to evaluate
alternative policies other than privatizing the enterprises.
To respond to the mandate and the Subcommittee staff request, we
developed a list of the economic behaviors most likely to be affected
by privatization, assessed how well such adjustments can be
quantified, and analyzed the probable outcomes resulting from
privatization.
The results of our analysis are presented in this report in terms of
three principal objectives. These objectives were to assess the
potential effects of privatization on (1) the enterprises; (2)
residential mortgage markets in general; and (3) housing finance,
homeownership, and housing affordability for very low-, low-, and
moderate-income families and residents of underserved areas in
particular.\19 In addition, we identified and analyzed, in response
to the Subcommittee's subsequent request, four policy alternatives
that Congress could consider to limit the enterprises' potential risk
to taxpayers or increase their social benefits.
To determine how the enterprises and housing finance markets would
react to privatization of the enterprises, we reviewed academic,
professional, and business literature on the role of the enterprises
in mortgage markets. This review identified several ways the
enterprises could be affected by privatization and how markets may
evolve and change. We interviewed market participants, such as
mortgage bankers, private mortgage insurers, mortgage security
underwriters, bond rating agencies, and private-label mortgage
conduits, to gain their insights into how the market might perform if
the enterprises were to be privatized. We interviewed additional
individuals with expertise in mortgage markets, including analysts at
the Federal Reserve Board, and current and former HUD staff members.
We also interviewed representatives from the enterprises to obtain
their perspectives on the effects of privatization.
We participated with the Congressional Budget Office (CBO), HUD, and
the Treasury in commissioning five studies on different aspects of
privatization. The authors of these studies presented findings at
seminars attended by representatives of the four agencies and the
enterprises as well as discussants who were invited to provide
comments. We had extensive interactions with the authors, both
within and outside of the seminars, to evaluate their methodologies
and results as needed. We did not, however, verify their data.
We used the studies, the material discussed at the seminars, and
comments prepared by the discussants and the enterprises as an
additional source of information in preparing this report. The
studies and written comments by discussants and Fannie Mae will be
published by HUD in Studies on Privatizing Fannie Mae and Freddie Mac
(forthcoming May 1996) and do not necessarily reflect the opinions of
GAO or the other agencies.\20
We also relied on data and information contained in annual and
investor analyst reports over the past 6 years published by the
enterprises, information statements and prospectuses provided by the
enterprises and private-label conduits, studies and statistical
tabulations provided by the enterprises, and other information
provided by parties we interviewed. We obtained documentation and
evaluated the data and information as needed, but we did not verify
these data. We conducted our work in Washington, D.C., from March
1994 through December 1995 in accordance with generally accepted
government auditing standards.
In addition, we provided copies of the draft of this report to the
Chairmen of Fannie Mae and Freddie Mac. On April 26, 1996, we met
separately with senior enterprise officials, which included senior
vice-presidents from each agency, and they provided oral comments,
which are presented and discussed on pages 49-53, 70-77, and 94-97.
One Freddie Mac official said that we relied on work performed by
others that we did not verify, and therefore we should make clear
when estimates by others were used. We have clarified how we
evaluated and relied upon the five studies on different aspects of
privatization as well as the data and information supplied by the
enterprises and others.
--------------------
\19 Our second and third objectives overlap. We analyzed single-
family mortgages to address the second objective, market effects.
Single-family mortgages finance 1-4 unit residences. We analyze
single-family housing initiatives undertaken by the enterprises to
fulfill social goals as part of the third objective. We also
analyzed multifamily housing as part of the third objective.
\20 Freddie Mac did not provide written comments on these studies.
PRIVATIZATION WOULD LIKELY
INCREASE ENTERPRISE COSTS AND
CHANGE THE ENTERPRISES' OPERATING
STRATEGIES AND ACTIVITIES
============================================================ Chapter 2
Assuming that privatization eliminates the perception by investors of
an implied federal guarantee of the enterprises' financial
obligations as well as explicit charter benefits, the enterprises'
overall annual costs would increase substantially. Based on 1995
financial statements and operations of the enterprises, total cost
increases on a pretax basis could have been in the range of $2.2
billion to $8.3 billion. The largest increase, probably in the range
of $1.3 billion to $4.4 billion, would likely have been in an expense
that has represented in recent years more than two-thirds of total
expenses of each of the enterprises--the interest the enterprises pay
on their debt securities. Without the perception of an implied
federal guarantee, investors would likely require higher interest
rates on the enterprises' debt securities to make up for the
perceived increase in risk. For the same reason, the enterprises
would also have higher funding costs on the MBS they issue. In
addition, increased overhead and operating expenses would result from
the elimination of the enterprises' exemption from SEC registration
requirements and state and local corporate income taxes.
The increased costs would likely lead the enterprises to change their
operating strategies and activities so that they would probably
resemble more closely the strategies of private-label conduits. In
addition, they could enter into new lines of business, both within
and outside of the housing finance industry.
On the other hand, if the markets' perception about the implied
guarantee does not change, or changes very little, the effect of
privatization on the enterprises' costs would be limited largely to
expenses related to SEC registration requirements and state and local
corporate income taxes. The primary effect of privatization in this
case may be the enterprises' increased opportunities to enter new
lines of business. Therefore, the effect of privatization largely
depends on the markets' perception of the riskiness of the
enterprises' debt securities and MBS following privatization.
INTEREST INCOME AND EXPENSES
DOMINATE THE ENTERPRISES'
FINANCES
---------------------------------------------------------- Chapter 2:1
As shown in the 1995 income statements of the enterprises (see table
2.1 for summary information), interest income and expenses dominated
the enterprises' finances. In 1995, the enterprises' major sources
of income were interest earned on mortgages retained in portfolio and
guarantee fees on MBS.\1 Interest income provided 94.7 percent of
Fannie Mae's total revenue and 88.2 percent of Freddie Mac's.
Interest income was relatively higher at Fannie Mae because it
retained in portfolio a relatively large proportion of mortgages it
has bought--33 percent, compared to Freddie Mac's retention of 18
percent (see tables 2.2 and 2.3.)
Expenses were dominated by interest paid on debt securities. As
shown in table 2.1 (under Total interest expenses), this expense
represented 81.0 percent of Fannie Mae's total revenues and 73.5
percent of Freddie Mac's. Debt securities are the enterprises'
primary source of financing. For this reason, the enterprises'
funding costs are driven primarily by the interest paid on debt
securities.
Table 2.1
Income Statements for the Housing
Enterprises in 1995
(Dollars in Millions)
Percent
of Percent
total of total
Dollars revenues Dollars revenues
--------------------------- --------- -------- --------- ---------
Revenues
----------------------------------------------------------------------
Interest income $21,070.8 94.70 $8,393.0 88.17
Guarantee and 1,085.7 4.88 $1,087.0 11.42
management fees
Other income 92.5 0 .42 39.0 0.41
======================================================================
Total revenues $22,249.0 100.0 $9,519.0 100.0
Expenses
----------------------------------------------------------------------
Interest expenses ($18,023. 81.01 ($6,997.0 73.51
4) )
Other expenses (1,230.2) 5.53 (936.0) 9.83
Income before taxes and $2,995.4 13.46 $1,586.0 16.66
extraordinary items
(revenues minus
interest and other
expenses)
Provision for federal ($839.8) 3.77 ($495.0) 5.20
income taxes
Extraordinary losses ($11.4) 0.05 $0.0 0.00
Net income $2,144.2 9.64 $1,091.0 11.46
----------------------------------------------------------------------
Source: Preliminary 1995 financial data from Fannie Mae and Freddie
Mac.
Table 2.2
Balance Sheets for the Housing
Enterprises in 1995
(Dollars in Millions)
Percent Percent
of total of total
Assets Dollars assets Dollars assets
------------------------------ -------- -------- -------- --------
Mortgage holdings\a $252,588 79.8% $107,411 78.3%
Other assets 63,962 20.2 29,770 21.7
======================================================================
Total assets $316,550 100.0 $137,181 100.0
Liabilities
----------------------------------------------------------------------
Liabilities and reserves $305,591 96.5 $131,318 95.7
Stockholders' equity 10,959 3.5 5,863 4.3
======================================================================
Total liabilities and equity $316,550 100.0 $137,181 100.0
----------------------------------------------------------------------
\a Mortgage-backed securities exclude MBS held in portfolio, and
mortgage holdings include MBS held in portfolio.
Source: Preliminary 1995 financial data from Fannie Mae and Freddie
Mac.
Table 2.3
Total Mortgage Servicing Portfolio
(Dollars in Millions)
Percent Percent
Dollars of total Dollars of total
------------------------------ -------- -------- -------- --------
Mortgage holdings $252,588 33.0% $107,411 17.8%
Mortgage-backed securities\a 513,230 67.0 495,045 82.2
======================================================================
Total mortgage servicing $765,818 100.0 $602,456 100.0
portfolio
----------------------------------------------------------------------
\a Off balance sheet item.
\b Mortgage-backed securities exclude MBS held in portfolio, and
mortgage holdings include MBS held in portfolio.
Source: Preliminary 1995 financial data from Fannie Mae and Freddie
Mac.
--------------------
\1 Interest and principal flowing through the MBS mortgage pools are
revenues to the mortgage pools, which are legally separate entities
from the enterprises. These flows are not reported on the
enterprises' income statements, but the fees paid by the pools to the
enterprises are reported.
LARGEST ENTERPRISE BENEFITS
FLOW FROM MARKET PERCEPTION OF
IMPLIED GUARANTEE
---------------------------------------------------------- Chapter 2:2
Due to the importance of interest income and expense in the financial
condition of the enterprises, the most important advantage of the
enterprises' government-sponsored status is the perception of
financial market participants that the federal government is likely
to act to ensure that the enterprises will meet their debt and MBS
obligations. The perceived federal guarantee lowers the enterprises'
funding costs in two primary ways. First, it decreases perceived
risk for investors in the enterprises' debt and MBS; this lowers the
funding costs that the enterprises must pay. Consequently, the
enterprises pay interest rates on their debt that are above the rates
that the Treasury pays and below the rates paid by highly rated
financial corporations on similar debt.
The second way that the perceived federal guarantee lowers the
enterprises' funding costs is that it decreases the extent to which
the enterprises must fund themselves with relatively more expensive
equity capital--the difference between assets and liabilities.
Equity serves as a financial cushion that can absorb financial losses
in bad years. Investors in a corporation's debt require this cushion
because it can help ensure the continued operation of the company
when downturns occur. The amount of equity a firm needs to maintain
a high debt rating depends on financial risk; if risk is relatively
high, equity must be correspondingly high. Because the perceived
federal guarantee lowers investors' perceived financial risk, the
enterprises are able to hold less equity and fund more of their
operations through issuing debt securities, compared to potential
private competitors.
A further advantage of government sponsorship is that bond rating
agencies and bank regulators consider the enterprises issuers of
low-risk debt on the basis of their perceived government ties. This
ensures that the enterprises' debt securities and MBS can be bought
and held by a large class of investors that must invest in high-grade
securities. These investors include banks, insurance companies, and
other regulated institutions, which provide a ready and consistent
outlet for enterprise debt and MBS.
The last funding advantage is that most investors realize that the
very size of the enterprise ensures a ready market for reselling
enterprise debt securities and MBS. Government sponsorship does not
in itself guarantee large size. However, the combination of a
multibillion-dollar mortgage market, the financial cost advantage
arising from the perception of government backing, and the fact that
only two organizations have been granted these advantages contribute
to the enterprises' size. This marketability or liquidity further
lowers the enterprises' funding costs since investors know they can
readily resell the securities if they need cash quickly.
Consequently, investors do not require higher interest rates on
enterprise debt issuances due to the risk that they cannot be resold
in a liquid market. The large size of the enterprises' operations
may also lower their average operating costs per MBS or per mortgage
due to economies of scale.\2
Other benefits that derive directly or indirectly from the federal
charters and lower the enterprises' operating costs include
-- a conditional line of credit of up to $2.25 billion available
for each enterprise from the U.S. Treasury at Treasury's
discretion;
-- exemption from registering securities or paying fees to SEC;
-- ability to issue debt and MBS that the Federal Reserve, other
bank regulators, and bond rating agencies consider high-quality,
low-risk paper; and
-- use of the Federal Reserve as a transfer agent, which enhances
operating efficiency.
The combined funding and operating cost advantages, along with any
additional efficiencies arising from sound management practices, help
ensure that the enterprises are the lowest cost participants in the
secondary conforming mortgage market. In effect, the advantages
flowing from government sponsorship make it difficult if not
impossible for other companies to compete in the secondary market for
conforming loans.
--------------------
\2 Here economies of scale refer to the cost of creating and
administering MBS; the cost impact for the enterprises is in addition
to the funding advantage for MBS because investors require a lower
yield on enterprise MBS due to federal ties.
ELIMINATION OF THE PERCEPTION
OF AN IMPLIED GUARANTEE WOULD
INCREASE ENTERPRISES' FUNDING
COSTS
---------------------------------------------------------- Chapter 2:3
Assuming that privatization causes the market to no longer perceive
an implied guarantee by the government or perceived it to be
substantially weakened, the market would in turn likely demand a
higher payment on debt and MBS.\3 We used the 1995 financial
statements and operations of the enterprises to estimate the dollar
benefits of government sponsorship in funding costs on debt
securities and MBS.\4 The extent to which the savings of sponsorship
flow to the enterprises, borrowers, and investors is unknown; we
discuss impacts on borrowers in chapter 3. We estimated, using
conservative measures of the enterprises' funding advantages
resulting from government sponsorship\5 , that the total benefit in
reduced interest costs the enterprises paid in 1995 on debt
securities was in the range of $893 million to $1.3 billion, with the
amount depending upon how the enterprises would treat cost increases
resulting from privatization on their federal tax returns.\6 We
estimate the total combined benefit in funding costs the enterprises
received on MBS was in the range of $343 million to $486 million. We
also estimated, using higher measures of the enterprises' funding
advantages, that the total benefit on debt was in the range of $3.2
billion to $4.4 billion and was in the range of $2.4 billion to $3.4
billion on MBS.\7
In one of the studies done for this project, Ambrose and Warga\8
estimated the current funding advantage of government sponsorship for
enterprise fixed-rate debt. They used two approaches. In their
first approach, they estimated how much lower the enterprises'
average current interest rate is than the average interest rate on
similar debt\9 issued by their potential competitors. On the basis
of yield data, they estimated that the enterprises paid on average
about 0.37 percent less on noncallable debt and about .63 percent
less on callable debt from 1985 to 1994.\10 They also made estimates
for the more current 1991 to 1994 time period and using different A,
double-A, and triple-A rated corporations as benchmarks. The
estimated funding advantage on callable debt for the 1991 to 1994
period ranged from .8 to 1.06 percent. The enterprises' interest
rates, however, were higher than rates on U.S. Treasury debt. This
difference suggests uncertainty in the market's perception that the
government is likely to rescue the enterprises if they failed.
In the second approach, Ambrose and Warga evaluated how differences
in cash flows and returns over time between debt and equity issued by
the enterprises and other borrowers may have affected cost of capital
differentials.\11 They concluded, on the basis of this approach, that
if the enterprises had to issue debt with characteristics similar to
debt issued by potential A-rated competitors, their cost of funds
would have increased by about 1.5 percentage points.
Ambrose and Warga also compared average differences in investor
yields between enterprise and private-label multiclass MBS.
Enterprise MBS had average yields that were .27 to .37 percentage
points lower than private-label MBS.\12
--------------------
\3 If the enterprises had to pay higher interest rates on their debt,
their income from issuing debt to purchase mortgages would be reduced
as long as mortgage interest rates remained unchanged. Likewise, if
the enterprises had to pay investors higher returns on their MBS,
their income from issuing MBS to fund mortgages they purchase would
also be reduced. The enterprises collect fees on MBS that represent
part of the difference between payments by borrowers and the share of
those payments to MBS investors.
\4 FNMA and FHLMC: Benefits Derived From Federal Ties
(GAO/GGD-96-98R, Mar. 25, 1996).
\5 For this calculation, we used 30 and 5 basis points as the funding
advantage on debt and MBS, respectively, based on Congressional
Budget Office, Reducing the Deficit: Spending and Revenue Options
(Feb. 1995) pp. 318-319. CBO stated that the enterprises' funding
advantage on debt was at least 30 basis points and on MBS was at
least 5 basis points
\6 Funding costs are deductible for federal corporate income tax
purposes. It should be noted that the (higher) estimated pretax
value of financial benefits is consistent with a scenario in which
the extra costs to the enterprises resulting from repeal of benefits
would be passed through entirely to homebuyers with no corresponding
loss in each enterprise's corporate income. The estimated aftertax
value is consistent with a case in which the enterprises would not be
able to pass through any extra costs to homebuyers. As a result,
deductibility of these extra costs would directly lower corporate
income.
\7 The higher range of our dollar estimates of the funding advantage
resulting from government sponsorship assumed a 1.06 percentage point
funding advantage on debt and .35 percentage points on MBS.
\8 These data are reported in a study commissioned by HUD. The study
was conducted by Brent Ambrose and Arthur Warga and was entitled
"Implications of Privatization: The Costs to FNMA and FHLMC."
\9 Their estimates are based on bond yields in which they control for
differences in bond characteristics such as maturity and age. The
estimates reported here are based on a rating of AA to represent the
credit quality of debt issued by potential competitors. Double-A is
a rating assigned by Standard and Poors, a major bond rating agency.
It indicates quality of the debt and the likelihood that a bond
issuer will make principal and interest payments on schedule. The
best rating is AAA. Yield is the interest rate received on a
security if the investor holds it until maturity.
\10 The enterprises rely heavily on callable debt to finance their
retained mortgage portfolios. Noncallable debt is debt that must be
paid on schedule by the issuer and cannot be prepaid. Callable debt,
in contrast, can be paid off at the issuer's discretion. Callable
debt is advantageous when interest rates decline below the interest
paid on such debt. As of year-end 1994, over half of Fannie Mae'
long-term debt and over 80 percent of Freddie Mac's were callable or
had downward rate adjustment features.
\11 The approach is based on a model called the capital asset pricing
model. The approach compares how one financial instrument's returns
fluctuate, in terms of both direction and volatility, compared to
benchmark securities chosen by the analyst. The authors chose
securities issued by highly rated depository institutions as their
benchmark securities.
\12 Ambrose and Warga discussed estimation problems in estimating
enterprise funding advantages because the debt and MBS instruments
issued by the enterprises have features different from those issued
by private issuers. In particular, they report only average yield
differences for MBS, because they were unable to control for the
large differences in features between enterprise and private label
MBS. The MBS estimates were based on collateralized mortgage
obligations, which were discussed in chapter 1.
ANALYSTS EXPECT THAT
PRIVATIZATION WOULD PROBABLY
REQUIRE AN INCREASE IN THE
ENTERPRISES' EQUITY CAPITAL
LEVELS
---------------------------------------------------------- Chapter 2:4
Because the perception of an implied federal guarantee lowers the
perceived risk of the enterprises' debt securities and MBS, investors
accept lower yields on all enterprise securities and permit the
enterprises to operate with less equity than they would otherwise
require. A good measure of equity adequacy is the ratio of equity to
all assets--the sum of book assets and MBS. (Generally, the larger
the ratio, the less the likelihood that operating losses will result
in the failure of the entity.) In 1995, Fannie Mae's ratio of equity
to all assets was 1.3 percent, and Freddie Mac's was 0.9 percent.
(Freddie Mac's lower ratio reflects the fact that less equity needs
to be held against the risks of MBS.) These equity ratios are
generally lower than ratios maintained by other financial
institutions that deal in mortgages and MBS.
As of December 1995, OFHEO required the enterprises to meet two
different minimum equity ratios: the minimum ratio of equity to
retained assets and the minimum ratio of equity to off-balance sheet
assets. The minimum ratio of equity to retained assets, which
includes mortgages held in portfolio, was 2.5 percent; the minimum
ratio to off-balance sheet assets, which includes MBS, was .45
percent. The enterprises' current ratios satisfy the minimums set by
OFHEO.
Analysts at the enterprises, mortgage market analysts at rating
agencies, and private-label conduits told us high bond ratings are
desirable since they indicate the firm imposes lower risks and
investors will permit lower risk firms to pay lower interest rates on
their debt. However, to obtain such ratings as fully private firms,
these analysts generally told us that the enterprises would probably
have to increase their equity levels.
PRIVATIZATION WOULD ELIMINATE
DIRECT BENEFITS
---------------------------------------------------------- Chapter 2:5
Privatization would eliminate the direct benefits conveyed by the
enterprises' federal charters. The most significant of the direct
charter-based benefits is probably the exemption from state and local
corporate income taxes. If the enterprises had paid state and local
corporate income taxes at an average rate of 8 percent in 1995 and if
no other costs, capital levels, or operating strategies had changed,
we estimated that this would have resulted in a combined increase in
expenses for the enterprises in the range of $367 million to $256
million, again depending upon the enterprises' treatment of the
increases in their federal tax returns.\13
Expenses related to SEC registration fees, which the enterprises
would also have to pay if privatized, would also be significant. If
the enterprises had been required to register with SEC and pay fees
in 1995 and if no other costs, capital levels, or operating
strategies had changed, registration would likely have cost the
enterprises SEC's statutory fee of 3.4 basis points on each dollar of
long-term debt, MBS, and CMO issued. The combined increase in
expenses for the enterprises would have been in the range of $102
million to $72 million.\14
The enterprises do not currently have to obtain ratings on their
debt, MBS, and equity issuances from private rating firms. If they
were privatized, they would need to obtain such ratings. We
understand that rating fees average about 3 basis points (.03
percent) on issuances but are subject to substantial discounts for
large issuers. Our calculations, however, do not include an estimate
of the amount of fees that the enterprises might have to pay if
privatized.
Determining the cost advantage of using the Federal Reserve as a
transfer agent is difficult. However, using the Federal Reserve
could make enterprise securities more liquid and convenient
investments than they would be otherwise. Such convenience could
also lower MBS issuance costs. Just as with rating fees, we did not
estimate such costs.\15
These cost increases resulting from privatization would also likely
have an adverse impact on the enterprises' market shares, profits,
and stock values. The magnitude of the effect would depend on the
magnitude of the cost increase. In addition, certain expenditures,
such as those for compensation, could decline.
--------------------
\13 FNMA and FHLMC: Benefits Derived From Federal Ties
(GAO/GGD-96-98R, Mar. 25, 1996). The 8 percent average state
corporate income tax rate is based on a Congressional Research
Service report entitled "Unfunded Mandates and State Taxation of the
Income of Fannie Mae, Freddie Mac, and Sallie Mae: Implications for
D.C. Finances," written by Dennis Zimmerman (Sept. 8, 1995).
\14 FNMA and FHLMC: Benefits Derived From Federal Ties
(GAO/GGD-96-98R, Mar. 25, 1996). We excluded short-term debt
issuances by the enterprises because SEC officials told us that such
debt could be defined as commercial paper and not be subject to SEC
registration fees. A basis point is one one-hundredth of a
percentage point.
\15 We did not have data to estimate rating fees, because the
discounts given by rating agencies are proprietary information. In
addition, the enterprises pay fees to the Federal Reserve for being a
transfer agent, and we do not have a basis to estimate the potential
enterprise cost advantage that may be above and beyond these fees.
PRIVATIZATION COULD CHANGE THE
ENTERPRISES' OPERATING AND
MARKETING STRATEGIES
---------------------------------------------------------- Chapter 2:6
The combination of potentially higher funding costs, increases in
other expenses, and opportunities to expand into new business areas
associated with privatization could alter the enterprises' operating
strategies. The enterprises have noted that removal of their
benefits and restrictions would lead them to change their operating
strategies. An important determinant in the extent and type of
behavioral change would be the effect of privatization on the
enterprises' funding costs. For example, if debt costs increase
substantially as a result of privatization but MBS funding costs and
mortgage interest rates go up by lesser amounts, the enterprises
would have strong incentives to change both the amount of mortgages
they fund and the way they fund mortgages. They might decide to hold
fewer mortgages in portfolio and fund a larger proportion of
mortgages by issuing MBS. This possibility is discussed in more
detail in chapter 3.
The markets' perception of increased credit risk of enterprise
securities could also lead the enterprises to change the terms under
which they securitize mortgages. The MBS issued by the enterprises
could come to more closely resemble those issued by private-label
conduits. In addition, the elimination of charter restrictions would
provide the enterprises with expanded opportunity in the areas of
nonconforming mortgages and nonmortgage securitization as well as
areas related to secondary mortgage market lending.
THE ENTERPRISES' MBS MIGHT
MORE CLOSELY RESEMBLE
PRIVATE- LABEL MBS
-------------------------------------------------------- Chapter 2:6.1
If the markets perceived a decline in the creditworthiness of the
enterprises as a result of privatization, one response the
enterprises could choose would be to alter their MBS to more closely
resemble those now issued by private-label conduits. Under the
current structure, the enterprises insure the creditworthiness of
their MBS. Without the benefit of the market perception of an
implied federal guarantee of creditworthiness, investors could
require the enterprises to deal more directly with credit risks in
their MBS.\16
The funding mechanism of current private label conduits in the jumbo
market provides some information about how MBS might be structured
with privatization.\17 The enterprises provide credit enhancement for
their MBS by requiring mortgage insurance on mortgages with
loan-to-value ratios above 80 percent\18 and fully insuring the
remaining credit risk on most mortgages. The private-label conduits
issue multiple-class MBS in which part of the credit risk is passed
onto investors. Providing credit enhancements that limit the credit
risk to investors is important to the marketability and liquidity of
the MBS.\19
--------------------
\16 The funding advantage the enterprises have over private-label
issuers of MBS is not directly observable because private-label MBS
have contract features that differ from the enterprises' features.
For example, private-label MBS pass credit risk onto at least one
class of investors while enterprises MBS retain all the credit risk.
\17 Private-label conduits securitize mortgages they purchase; they
do not fund mortgages out of retained portfolio.
\18 The enterprises are not allowed by statute to purchase mortgages
with loan-to-value ratios above 80 percent without mortgage insurance
or another form of credit enhancement. The enterprises, however,
often require mortgage insurance beyond the minimum statutory level.
\19 The funding mechanisms of current private-label conduits,
including their credit enhancements, are discussed in chapter 1.
THE ENTERPRISES COULD ENTER
NEW MARKETS AND OTHER
INDUSTRIES
-------------------------------------------------------- Chapter 2:6.2
Without their current charter restrictions, the enterprises would be
allowed to enter the current jumbo mortgage market. They would also
be allowed to engage in business activities that complement their
existing businesses--for example, the proprietary information
technology developed by the enterprises could lead to nonmortgage
securitization and provision of automated financial transactions
services.
THE ENTERPRISES AND OTHER
CONDUITS COULD
PARTICIPATE IN RELATED
REAL ESTATE ACTIVITIES
------------------------------------------------------ Chapter 2:6.2.1
The residential mortgage market consists of a vertical stream of
entities beginning with home buyers and mortgage originators and
continuing with mortgage underwriters, insurers, conduits, and
investors. Privatization would allow the enterprises to enter
different vertical segments of the housing finance system, such as
origination and mortgage insurance, that their charters now prevent
them from entering.
The enterprises compile extensive information on housing and mortgage
markets, including home sales prices, housing ownership turnover, and
flows of mortgage credit. Currently, private- label conduits, their
mortgage banking subsidiaries, and other large mortgage banking
businesses are developing products such as real estate appraisal
services. The enterprises, private- label conduits, and many
mortgage banking businesses have developed expertise in hedging
interest rate risks associated with providing mortgage commitments
before funding. The enterprises have also developed this expertise
as it applies to funding long-term, fixed-rate mortgage products.
One interesting possibility is that the enterprises, private- label
conduits, large mortgage bankers, and other industry participants
might vertically integrate\20 or form networks, including firms
specializing in different vertical stages of the process, to provide
residential mortgage credit. If this were to occur, the resulting
entities might develop large capacities for information retrieval and
distribution to effectively compete in the mortgage markets as well
as expertise in financial and risk management. These capacities
could create synergies in related real estate activities and in
nonhousing financial markets. Under privatization, the enterprises
would not face the restrictions in their current charters that now
prevent them from supplying these alternative services.
--------------------
\20 Vertical integration by a business firm occurs when the firm
combines with another firm operating in a different vertical segment
of an industry through acquisition or merger.
EFFECT OF PRIVATIZATION WOULD
LARGELY DEPEND ON MARKET
PERCEPTIONS
---------------------------------------------------------- Chapter 2:7
In our analysis of the likely effects of privatization on the
enterprises, we assumed that privatization would result in the
reduction or elimination of the perception of an implied federal
guarantee. While it appears that eliminating the benefits,
restrictions, and obligations associated with the enterprises'
federal charters would be likely to at least reduce the markets'
perception of the implied guarantee, we recognize the uncertainty
inherent in any attempt to predict the behavior of financial markets.
To the extent that the markets do not perceive that the ties between
the enterprises and the federal government are broken, the
enterprises' funding advantage may remain. In case little change in
the funding advantage occurs, the primary effects of privatization
would be to (1) raise some operating costs by eliminating the tax and
SEC registration-related benefits that flow directly from the
charter, and (2) free the enterprises to do business in new areas.
In such a case, the enterprises could become even larger and generate
even greater potential risk to the government should the government
feel the need to rescue a failing enterprise that was "too big to
fail."
CONCLUSION
---------------------------------------------------------- Chapter 2:8
The effect of privatization on the enterprises is difficult to
predict. First, it is always difficult to predict with much
precision how an organization will respond to changes in its
environment whether from higher tax liabilities, higher interest
costs, or reduced restrictions on its actions. Second, the most
important effects depend on changes in market perceptions and the
subsequent effect of those perceptions on the funding costs the
enterprises would face. If the markets perceive the privatized
enterprises' securities as being riskier than the
government-sponsored enterprises' securities, they are likely to
demand higher returns to pay for the greater perceived risk. This
could cause the enterprises' funding costs to rise significantly.
The markets would also likely insist on greater capital to maintain a
given credit rating. These increased funding costs and any resulting
changes in enterprise behavior could bring about substantial change
in the overall mortgage market. The enterprises could alter their
behavior in a number of areas, including the amount of mortgage
financing they do, the way they finance mortgages, and the way they
deal with credit risk in their MBS. The potential effect under this
scenario also depends on responses of other participants in the
housing finance market, as discussed in the next chapter.
On the other hand, if market perceptions do not change, and interest
costs do not rise, the primary cost increases from privatization
would come from SEC registration fees and state and local taxes. In
this case, the cost increases that the enterprises would face may be
minor in relation to the potential profitability from their increased
business opportunities. Changes in the operating and marketing
strategies of the enterprises--whatever the specific changes might
be--could also affect behaviors of other industry participants.
ENTERPRISE COMMENTS AND OUR
EVALUATION
---------------------------------------------------------- Chapter 2:9
In oral comments on our draft report, Fannie Mae and Freddie Mac
officials disagreed with our analysis of the financial benefits that
government sponsorship provides to the enterprises and what they
perceived as an implication that the benefits are derived by the
enterprises rather than homebuyers. Fannie Mae officials said that
the draft report did not provide sufficient context for the estimated
range of financial benefits that government sponsorship provides to
the enterprises. For example, the officials said the draft report
implied that the benefits are derived by the enterprises rather than
homebuyers. Generally, they did not think that it is meaningful to
discuss charter-based benefits without discussing their restrictions,
obligations, and what is passed thorough to borrowers. Although
Fannie Mae officials also said that it is possible to estimate the
value of government sponsorship, they said a more appropriate
analysis would require a specific identification of who benefits from
government sponsorship. Because the Fannie Mae officials believe
that homebuyers are the primary beneficiaries of the financial
benefits, they said we should have estimated the total value of lower
mortgage interest rates to the American public. In addition, Fannie
Mae officials said that the enterprises do not pay MBS yields;
rather, they only guarantee the timely payment of MBS principal and
interest in exchange for a guarantee fee. Therefore, the officials
said Fannie Mae does not incur funding costs on MBS, so it would not
incur additional costs of 5 to 35 basis points in the event of
privatization.
Freddie Mac officials also said that our estimates of the benefits
associated with government sponsorship are high and that any
financial benefits flow to homebuyers in the form of lower mortgage
interest rates. The Freddie Mac officials further stated that we
should use the aftertax estimates since a portion of any financial
benefits is returned to the federal government in the form of income
taxes. In addition, the Freddie Mac officials said that
privatization would not eliminate the perception of the federal
government's implied guarantee to support the housing finance system.
The officials said that the implied guarantee would remain because
the federal government has supported a stable, low-cost housing
system for 60 years, and the market would still believe that the
government would take necessary steps to protect that system,
including providing emergency financial support to significant
participants in the mortgage finance system.
Officials from both enterprises argued that the Ambrose and Warga
study has fundamental flaws and that we should not have relied on it.
Fannie Mae officials said that the high-end estimate of $8.9 billion
in 1995 is excessive on its face because the enterprises had a
combined beforetax income of only $4.6 billion that year: the
estimated cost savings to the enterprises was about twice their
beforetax profits. Moreover, the officials said that there were
problems with Ambrose and Warga's analysis of rate of return data.
Although the Fannie Mae officials acknowledged that Ambrose and
Warga's study also found differences when using yield data, they said
the report had so many flaws we should not use any of its findings to
estimate the enterprises' funding advantage.
Freddie Mac officials also said that the Ambrose and Warga study
suffers from substantial limitations and questioned our using it as a
basis for estimating the enterprises' funding advantage. The
following summarizes their concerns regarding our use of the Ambrose
and Warga study:
-- They believe the weighted average cost of capital methodology is
flawed and should not be relied upon in setting the top of our
range of the funding advantage on debt.
-- One official said Ambrose and Warga relied on debt return data
from 1991 to 1994 to estimate that enterprise funding costs
would rise 100 to 200 basis points in the event of
privatization. However, the officials said that a similar
analysis performed for the years 1985 to 1994 would have shown
no difference in returns between enterprise bonds and bonds
issued by other borrowers.
-- One official said that bond yields interact over time, an
econometric problem called serial correlation, and that this
invalidates Ambrose and Warga's estimates.
Fannie Mae officials commented that the draft report's discussion of
the enterprises' capital adequacy was misleading. Contrary to a
statement in the draft report, they said that the ratio of equity
capital to assets is not a good measure of the enterprises' capital
because they are a unique institution that faces risks that are
different from depository institutions' risks. In particular, the
enterprises can hold relatively less capital against MBS since it
presents lower risks than other types of assets. Moreover, Fannie
Mae officials said the draft report failed to mention that OFHEO is
developing risk-based capital standards to ensure the safety and
soundness of the enterprises. These standards are intended to ensure
that the enterprises will have adequate capital to protect against
interest rate risks and other types of risks.
We do not believe there is sufficient evidence to conclude that all
of the benefits derived from government sponsorship flow through to
homebuyers, an issue we address more completely in chapter 3. We
have concluded, however, that if the enterprises were fully
privatized and the perceived guarantee were reduced or eliminated,
their funding costs would increase for both MBS and debt. Although
in a strict accounting sense, the enterprises charge guarantee fees
for guaranteeing the timely payment of principal and interest, the
fees the market is willing to bear depend, in part, on how much
higher mortgage interest rates are than the yield investors will
accept for investing in MBS. Because of the perception of an implied
guarantee, the market is willing to pay higher guarantee fees or
accept a lower yield on GSE MBS than on private-label MBS.
The use of beforetax or aftertax measures of the benefits derived
from government sponsorship and therefore of the potential costs of
privatization is spelled out in the report. The use of an aftertax
measure is consistent with a case in which the enterprises would not
be able to pass through any extra costs to homebuyers. Therefore,
the use of an aftertax measure appears inconsistent with the
enterprises' view that existing benefits flow through to homebuyers
and that eliminating those benefits would harm borrowers. On the
issue of whether it is even feasible to eliminate the perception of
the implied guarantee, we do not take a position. We assume that
privatization would reduce, if not eliminate, investors' perception,
but we acknowledge the possibility that it may not occur and we
discuss the implications in chapter 3.
We did not base our estimates on the Ambrose and Warga study in its
entirety. Rather, we relied on selected analyses from the study,
after satisfying ourselves that those analyses were methodologically
sound and appropriate for our use. For example, we relied on part of
the study to calculate our ranges for the funding advantage on debt
and MBS. The study is technical; therefore, use of their results
required some technical judgments. In their first approach to
analyzing interest rate spreads on debt, they make estimates using
both yield and rate of return data. In prior written comments on the
study (see p. 36), Fannie Mae objected to use of return data,
largely because it measures both investor returns that are expected
upon purchase and unanticipated changes in the value of the bond. We
used the results from the yield rather than the return data, because
yields are a better measure of expected returns at the time an
investor buys a bond.
Because bond characteristics differ between bonds issued by the
enterprises and other issuers, and bond yields interact with one
another over time (serial correlation), disentangling these effects
can be difficult. Ambrose and Warga recognized how difficult their
task was and qualified their results on the basis of the statistical
complexities. We relied on their results for the mean yield spread
between enterprise and others' debt based on their approach using
yield data in which they controlled for differences in bond
characteristics such as maturity and age. They recognized that
interactions between bond yields over time create serial correlation,
a criticism cited by Freddie Mac. We recognize this problem but we
also recognize that serial correlation affects only the precision of
the estimates. The estimated mean yield spreads, which we relied on,
are not biased. Because such estimates lack precision, however, we
used a wide range for the funding advantage on debt.
In their second approach, Ambrose and Warga use a weighted average
cost of capital (WACC) approach to estimate cost of capital
differences. Although we initially employed estimates derived using
this approach, we have decided to base our estimates on the more
straightforward approach based on yields. In our draft report the
upper-range of our estimate for the funding advantage on debt was 120
basis points. We revised this upper-range to 106 basis points.
Overall, we do not believe that the statistical estimation problems
with the WACC approach or the acknowledged limits of the return-based
approach provide sufficient basis to discard the authors' results
that were based on yield data.
The Ambrose and Warga estimates based on yield data were higher for
the 1991 to 1994 period than for the 1985 to 1990 period. In our
view, this most likely reflects imprecision associated with such
estimates, changes in the funding strategies of the enterprises,
and/or changes in financial markets. We believe it reaffirms our
position that a wide range of possible outcomes should be associated
with privatization.
Finally, analyzing the capital adequacy of the enterprises is a
complicated and largely unanswered question. Our understanding based
on past government studies, discussions with financial market
analysts, and regulators is that each enterprise would likely require
greater capital for its current activities if they were privatized.
OFHEO is developing risk-based capital standards to help ensure the
safety and soundness of the enterprises. If these standards require
the enterprises to increase their capital levels, enterprise funding
costs and mortgage interest rates could be affected.
PRIVATIZATION COULD INCREASE
MORTGAGE INTEREST RATES AND CHANGE
BEHAVIOR IN THE RESIDENTIAL
MORTGAGE MARKETS
============================================================ Chapter 3
The exact effects of privatization on the residential mortgage
markets cannot be determined with certainty, in part because of
difficulty in knowing how the financial markets would respond to
privatization. Our analysis of the effects of privatization on the
residential mortgage markets is based on the assumption that
privatization would eliminate or substantially reduce the perception
of an implied federal guarantee of the enterprises' financial
obligations and increase the enterprises' costs (as discussed in ch.
2). Under this assumption, privatization would likely lead to an
increase in mortgage interest rates. Privatization would also likely
lead to changes in behavior in the mortgage markets, particularly
increased competition in the secondary mortgage market. The
enterprises' higher cost of funds would likely allow private conduits
to compete with the enterprises in purchasing conforming mortgages.
In purchasing mortgages, the enterprises may be unable to fully pass
their increased funding and other costs to borrowers, since mortgages
with other sources of funding would be available to borrowers. The
enterprises would also be likely to charge fees more fully risk-based
than their current fees; this would cause increases in mortgage
interest rates to be greater for borrowers making smaller down
payments. In addition, mortgage interest rates could fluctuate more
than they have with the demand for mortgage credit. Due to the size
and sophistication of the mortgage finance market, significant
regional variations in interest rates seem an unlikely result of
privatization.
MORTGAGE INTEREST RATES ON
SINGLE-FAMILY HOUSING COULD
INCREASE
---------------------------------------------------------- Chapter 3:1
It is widely accepted that the enterprises, through portfolio
investments and securitization, have generated many benefits to
mortgage borrowers. These benefits include the reduction of regional
disparities in interest rates and mortgage availability, spurring of
innovations in mortgage standardization and transaction technology,
and lowering of mortgage interest rates. The markets' perception of
the implied federal guarantee on the enterprises financial
obligations plays an important--although not singular--role in
enabling the enterprises' to lower mortgage interest rates, in that
the perception lowers the enterprises' cost of funds. For this
reason, the effect of privatization on mortgage interest rates
depends critically on the extent to which privatization changes the
market's perception of the likelihood the government would come to
the enterprises' rescue.
If privatization caused the market to change its perception of an
implied tie with the government or substantially weakened it,
investors are likely to demand a higher payment for the perceived
increase in risk. The resulting higher cost of funds would lead to
higher mortgage interest rates as the enterprises attempt to maintain
their profits. However, the enterprises may not be able to fully
pass on the higher cost of funds, because competition could increase
in the conforming mortgage market.
In a competitive market, cost savings, such as those realized by the
special advantages granted to the enterprises, tend to flow through
to consumers, in this case residential mortgage borrowers. When
competition is limited, businesses can exercise what is often called
market power. When such market power is exercised, cost savings are
less likely to fully flow through to consumers, and businesses can
realize higher profits. Such profits can accrue to stockholders,
managers, employees, and others who provide goods and services to
businesses possessing the market power. In this respect,
privatization poses complicated policy questions. The fact that
government sponsorship ensures the dominance of two chartered
enterprises in the securitization of conventional, conforming
mortgages produces some benefits such as greater market liquidity,
but it may also produce costs due to lessened competition.
If the enterprises currently possess and exercise market power,
increasing effective competition would tend to cause more of the
benefits of government sponsorship to flow through to borrowers. The
extent of market power, however, is difficult to determine for a
number of reasons. For our purposes, the most important difficulty
is defining the relevant product market when alternative distribution
systems deliver similar, yet differentiated products.\1
For example, the enterprises state that the share of residential
mortgages they have funded--about 30 percent--is too small to convey
market power, so the benefits of government sponsorship flow through
to borrowers. The study commissioned for this project to analyze the
effect of privatization on the mortgage market defined the relevant
market for purposes of determining market power as conventional,
conforming mortgages securitized in the secondary mortgage market.\2
The resulting duopoly--a market served by two suppliers--and other
characteristics of the secondary market (for example, its offering of
a fairly standardized product) led them to conclude that the
enterprises "tacitly collude" and earn above average profits.\3 They
contend that government sponsorship introduces inefficiencies that
privatization could eliminate.\4
Because of insufficient statistical evidence, we do not know whether
a broad or narrow product market definition is appropriate in
determining the market power of the enterprises. Therefore, we
cannot determine the enterprises' market power or the potential
benefits resulting from increased competition. If, under the current
structure, the enterprises are not exercising market power and are
passing most of the benefits from government sponsorship on to
mortgage borrowers, increased competition may have little effect on
mitigating the increase in mortgage interest rates in the conforming
loan market that could result from privatization. However, if
government sponsorship creates market power for the enterprises,
conforming interest rates in the current environment may incorporate
to some extent the extra profits resulting from the market power of
the enterprises. Under this scenario, any increased competition
resulting from privatization could provide the potential benefit of
putting downward pressure on conforming mortgage rates.
The likely increase in average mortgage interest rates is the
broadest, most important market effect of privatization. The results
of our analysis indicated that privatization could increase interest
rates on fixed-rate, single-family housing mortgages below the
conforming loan limits within an average range of about 15 to 35
basis points.\5 Assuming that the interest rate increase does not
cause a decline in house prices, the monthly payments of a borrower
with a $100,000 thirty-year, fixed-rate mortgage would increase by
$10 to $25.\6 We use a $100,000 thirty-year fixed-rate mortgage to
illustrate the increase in monthly payments because the average
conventional, conforming loan amount for mortgages purchased by the
enterprises is about $100,000. For $2 trillion in outstanding
conventional conforming fixed-rate mortgages, the aggregate annual
increase in mortgage payments would be in the neighborhood of $3
billion to $7 billion.\7
Our estimate of the likely effect of privatization on fixed-rate,
single-family mortgage rates is based on a multipart analysis. For a
preliminary estimate of how much interest rates might rise with
privatization, we first sought to determine the interest rate spread
between conforming mortgages (those purchased by the enterprises) and
jumbo mortgages (those purchased by private-label conduits).
Realizing that the interest rate differential is influenced by some
factors specific to government sponsorship (which we assumed would be
eliminated through privatization) and some that are not, we sought to
adjust the estimated spread, accounting for specific factors
unrelated to the enterprises' government sponsorship. The results of
this work indicated that it would be reasonable to estimate that the
conforming jumbo interest rate spread would be about 20 to 40 basis
points. We next considered the need for one upward adjustment to
account for the possibility of reduced liquidity and three downward
adjustments. The three downward adjustments we considered were to
account for (1) the geographic concentration of existing jumbo
mortgages, which currently increases credit risk; (2) the possibility
that the volatility of loan collateral for jumbo mortgages may exceed
that of conforming mortgages; and (3) the likelihood of increased
competition and operational efficiencies in the conforming and jumbo
markets that could result from privatization. On the basis of this
analysis, we estimate that privatization would probably increase
average interest rates by about 15 to 35 basis points.
--------------------
\1 The enterprises fund mortgages in the predominantly conventional,
conforming mortgage market. The secondary market for conventional,
conforming mortgages represents the narrowest reasonable
interpretation of the relevant product market. With this definition,
the market is basically what is known as a duopoly (i.e., all
production is by two business firms), and each participant has market
power. Depositories fund mortgages in the primary market. The
conforming mortgage market (without a distinction between primary and
secondary activity) represents the broadest reasonable interpretation
of the relevant product market. With this definition, the
enterprises have a smaller market share and more competitors are
present. The participants in this defined market provide different
services and differentiated products. For example, depositories are
more likely than the enterprises to lend in local mortgage markets
where they have superior information. In addition, the depositories
are more likely to fund variable rate mortgages. If one were to
accept this broad market definition, a finding of market power would
be unlikely.
\2 Benjamin Hermalin and Dwight Jaffee, "The Privatization of Fannie
Mae and Freddie Mac: Implications for Mortgage Industry Structure,"
published by HUD in Studies on Privatizing Fannie Mae and Freddie Mac
(forthcoming May 1996).
\3 Tacit collusion is a term used by economists. Tacit collusion
does not mean that the enterprises are directly colluding with one
another, an act that is illegal. It means that the enterprises are
behaving (e.g., price policies) in a way that is consistent with a
situation in which they were directly colluding. Hermalin and Jaffee
argue for their narrow market definition and the finding of tacit
collusion partially on the basis of the high observed profit rates of
the enterprises and indirect statistical evidence found in John L.
Goodman, Jr., and S. Wayne Passmore, "Market Power and the Pricing
of Mortgage Securitization," Finance and Economics Discussion Paper,
Federal Reserve Board (March 1992).
\4 Hermalin and Jaffee also consider whether government sponsorship,
by increasing the resulting size of the enterprises, creates
spillover benefits in the form of increased liquidity. They
conclude, on the basis of their observations of private-label
conduits and the development of credit enhancements by underwriters
and credit rating agencies, that such spillover benefits from
government sponsorship are low.
\5 We think that privatization would likely cause an increase in
interest rates on variable-rate mortgages, but such increases would
not exceed and probably be less than the increase on fixed-rate
mortgages. Variable-rate and fixed-rate mortgages are
differentiated, competitive products. The benefits to borrowers from
government sponsorship and securitization are greater for fixed-rate
than for variable-rate mortgages.
\6 It is possible that a permanent increase in interest rates above
the trend that would evolve without privatization could force the
sellers to discount current housing prices so that buyers could
afford to pay mortgages issued at higher interest rates. In this
case, existing owners of housing would realize a capital loss.
However, such a decline in house prices and loan amounts could
partially mitigate the increase in mortgage payments generated from
the increase in interest rates.
\7 A Freddie Mac official told us that conventional, conforming,
single-family, fixed-rate mortgage debt equals about $2 trillion.
THE SPREAD OF INTEREST RATES
ON CONFORMING AND JUMBO
MORTGAGES IS A GROSS MEASURE
OF PRIVATIZATION'S IMPACT ON
MORTGAGE RATES
-------------------------------------------------------- Chapter 3:1.1
Our primary information sources for the gross measure of the impact
of privatization on mortgage interest rates included Freddie Mac and
the Federal Housing Finance Board. Freddie Mac officials provided us
with the interest rate spread between jumbo and conventional mortgage
rates for 30-year, fixed-rate mortgages from their Primary Mortgage
Market Survey for selected years between 1986 and 1995. The Survey
asks mortgage lenders their current commitment rates for a loan with
an 80 percent loan-to-value ratio on a monthly basis. Spreads were
in the 35 to 55 basis points range in 1988, 1989, 1990, and 1992.
Lower spreads ranging from 20 to 25 basis points occurred in 1986,
1993, and 1995.
We also analyzed the interest rate spread for the years 1990 through
1994 using the Federal Housing Finance Board's (FHFB) survey, Rates &
Terms on Conventional Home Mortgages. The survey collects interest
rates monthly on a sample of closed loans. We relied on spreads
reported for fixed-rate loans. Average spreads were 18, 9, 11, and
negative 2 basis points in 1990 through 1993, respectively. Reported
spreads continued to be negative in most months in 1994.
The Freddie Mac and FHFB data differ in certain respects. The
Freddie Mac data do not provide information on mortgage interest
rates for borrowers meeting any specific underwriting standard except
for loan-to-value ratio. The FHFB survey reports average loan
amount, loan-to-value ratio, and term; these averages are generally
similar between conforming and jumbo loans.
ADJUSTING SPREAD
DIFFERENCES TO ACCOUNT
FOR DIFFERENCES IN
MORTGAGE CHARACTERISTICS
------------------------------------------------------ Chapter 3:1.1.1
To estimate the interest rate differential created exclusively by the
enterprises' government sponsorship, we turned to a study
commissioned for this project. This study analyzed the interest rate
spread between conforming and jumbo mortgages by using individual
loan level data. For the years 1989 through 1993, the statistical
analysis standardized for many individual loan characteristics such
as location and loan-to-value ratio.\8 The results indicated interest
rate spreads of about 40 basis points in California and 30 to 35
basis points in the other states studied for 1989 through 1991. The
results for 1992 and 1993 found smaller spreads (in the 25 basis
point range), and the results for California were similar to those in
other states. For the last two quarters of 1993, the results
indicated interest rate spreads of about 20 basis points. The
study's findings were similar to those of two previous studies
employing the same methodology that found spreads in the 30 basis
point range.\9
The authors concluded, on the basis of the results over the entire
period, that single-family, fixed-rate jumbo loans had interest rates
that were generally 25 to 40 basis points higher than single-family,
fixed-rate conforming loan rates, holding other characteristics
constant. They concluded that a lowering of the conforming loan
limit would likely result in an increase in mortgage interest rates
in the lower part of the 25 to 40 basis point range for affected
mortgages (i.e., those shifting from conforming to jumbo status),
because liquidity in the jumbo market could increase from such
expansion. The authors did not reach a numeric conclusion for the
effects of privatization, largely because they did not know how much
liquidity would be affected by privatization.\10
Primarily on the basis of the results of the commissioned study and
the other two studies employing similar methodology, and recognizing
that the estimated spreads were volatile, we used 20 to 40 basis
points as the estimated average spread between conforming and jumbo
mortgages.\11 This estimate served as our initial baseline
approximation of how much interest rates would rise with
privatization.
--------------------
\8 Robert F. Cotterman and James E. Pearce, "The Effects of the
Activities of the Federal National Mortgage Association and the
Federal Home Loan Mortgage Corporation on Conventional Fixed Rate
Mortgage Yields," published by HUD in Studies on Privatizing Fannie
Mae and Freddie Mac (forthcoming May 1996).
\9 Patric H. Hendershott and James D. Shilling, "The Impact of
Agencies on Conventional Fixed-Rate Mortgage Yields," Journal of Real
Estate Finance and Economics (June 1989) pp. 101-115; and ICF
Incorporated, "Effects of the Conforming Loan Limit on Mortgage
Markets," Final Report prepared for Office of Policy Development and
Research, U.S. Department of Housing and Urban Development (March
1990).
\10 The authors presented an earlier version of their paper at a
seminar attended by representatives of the four agencies (GAO, CBO,
HUD, and Treasury) and the enterprises. The earlier version
concluded that the spread between conforming and jumbo mortgages was
in the 20 to 35 basis point range, which was their estimate of the
effects of privatization. The final version emphasized the
uncertainty associated with estimating the impact of privatization on
interest rates. Specifically, they state that liquidity effects
could cause interest rate changes to be above or below the 25 to 40
basis point range.
\11 The bottom of our range is 20 basis points because the seminar
draft of the commissioned paper includes it and the other two similar
studies found spreads in the 30 basis point range, the midpoint of
our 20 to 40 basis point spread. We did not rely heavily on the
other data, because those data series do not adjust for loan
characteristics.
RATIONALE FOR THE UPWARD
ADJUSTMENT TO ACCOUNT FOR
POSSIBILITY OF REDUCED
LIQUIDITY
------------------------------------------------------ Chapter 3:1.1.2
As mentioned earlier, we considered four adjustments to the 20 to 40
basis point range--one upward and three downward--in determining the
likely effect of privatization on mortgage interest rates. The
upward adjustment was to account for the possibility of reduced
liquidity. Officials of both enterprises emphasized the importance
of this factor, but they also acknowledged the difficulties in
measuring the liquidity effect. Officials from Freddie Mac stated
that liquidity in a privatized market would tend to decrease most
when mortgage originations were at their highest levels.\12 We
acknowledge that such an effect could result; however, it is our
understanding that liquidity in the jumbo market over the past decade
has generally been sufficient.\13 Because the private-label conduits
would likely expand and compete with the enterprises in the (current)
conforming and jumbo markets with privatization, the share of
conventional mortgages that would be securitized with privatization
would likely exceed the current share of jumbo mortgages securitized.
Such a development would contribute to a higher level of liquidity in
the conventional market than exists now in the jumbo market. In
summary, there is no convincing evidence that the upward adjustment
for reduced liquidity should be significant.
--------------------
\12 We generally agree with part of their analytical claim as it
relates to periods with relatively large demands for residential
mortgage credit. However, we question the broader claim that market
liquidity would be substantially reduced with privatization. In
addition, it should be noted that Hermalin and Jaffee concluded that
market liquidity would be unlikely to suffer while Cotterman and
Pearce concluded that liquidity could suffer from privatization.
\13 A Fannie Mae official told us, based on his conversations with
traders, that there have been three temporary disruptions in
liquidity in the jumbo market. We discuss this evidence later in
this chapter. We also note that the liquidity impact in a privatized
market, when demand increases, may depend on whether the increase in
mortgage originations is primarily from borrowing for home purchase
(purchase money mortgages) or for refinancing existing mortgages. If
the primary source is the latter, MBS investors may demand more new
MBS to replace MBS that experience more prepayments during times of
heavy refinancings.
RATIONALE FOR A DOWNWARD
ADJUSTMENT FOR POTENTIAL
GAIN IN REGIONAL
DIVERSIFICATION OF CREDIT
RISK
------------------------------------------------------ Chapter 3:1.1.3
One of the general benefits from mortgage securitization that helps
lower interest rates is regional diversification of credit risk. A
limiting factor for the private-label conduits that securitize jumbo
mortgages is that these loans tend to be concentrated in the
northeast region of the country and the state of California. We
discussed the impact of this factor with private-label issuers and
credit rating agencies. One way they quantified this limiting factor
was by relating it to the level of over-collateralization used for
credit enhancement. The general consensus was that if a pool of
jumbo mortgages that was geographically diversified could be backed
by collateral equal to 103 percent of the security issue, a jumbo
mortgage pool with similar characteristics but without such
geographic diversification would require 106 to 108 percent
collateral. Since such limits to diversification are not present in
the conforming market and would not be present with privatization,
the observed spread should be adjusted downward.
We could not reach a precise statistical estimate of what the
downward adjustment for regional diversification should be, but the
information on over-collateralization supports a downward adjustment.
In addition, the observed difference between the estimated interest
rate spread between conforming and jumbo mortgages in California and
other states for 1986 through 1991 is suggestive of a 5 to 10 basis
point adjustment; the higher estimated spread in California is
consistent with the large concentration of jumbo mortgages in that
state. We adjusted the spread downward by 5 basis points to account
for regional diversification.
RATIONALE FOR A DOWNWARD
ADJUSTMENT FOR GREATER
VOLATILITY OF LOAN
COLLATERAL FOR JUMBO
MORTGAGES
------------------------------------------------------ Chapter 3:1.1.4
The conforming jumbo spread may require a downward adjustment due to
the possibility that the volatility of loan collateral for jumbo
mortgages may exceed that on conforming mortgages. Borrowers are
more likely to default on their mortgage payments if the market value
of their residences, the collateral for the loan, falls below the
outstanding principal balance on their mortgage loans. One reason
why default is more likely on mortgages with relatively high
loan-to-value ratios is that relatively small local housing market
downturns can trigger default. For any given loan-to-value ratio of
a mortgage at the time of origination and the more volatile the price
of the residence, the greater the probability of default.
We obtained statistical evidence indicating that during the housing
market downturn in the state of California, the percentage decline in
house prices was greater for higher priced houses (that is, those
with jumbo mortgages) than houses with values below the conforming
loan limit. On the basis of our discussions with credit rating
agencies, we understand that this is factored into the credit
enhancement and pricing of jumbo, private-label MBS. Therefore, a
downward adjustment in the estimated conforming jumbo spread, even if
the estimate controls for the loan-to-value ratio, may be warranted.
However, there is no convincing evidence that the downward adjustment
should be significant over the period when interest rate spreads were
estimated.
RATIONALE FOR A DOWNWARD
ADJUSTMENT FOR INCREASED
COMPETITION RESULTING
FROM PRIVATIZATION
------------------------------------------------------ Chapter 3:1.1.5
Privatization would abolish charter restrictions on the enterprises
that limit their ability to diversify into other markets and, more
importantly, to vertically integrate throughout the different
segments of this market, such as residential mortgages, to realize
potential efficiencies. Privatization would also likely lead to
entry into the current conforming market by existing private-label
conduits and other potential entrants. These private label entities
could better realize economies of large-scale securitization with
privatization.\14
We have already addressed how competitive factors could affect how
much the benefits of government sponsorship are passed on to
residential mortgage borrowers. Generally, these factors are
reflected in the interest rate spread between conforming and jumbo
mortgages, because interest rates in the conforming market are
currently affected by government sponsorship. However, these
potential improvements in operational efficiencies, resulting from
increased competition, are not reflected in this interest rate
spread, because interest rates in the jumbo market are not currently
affected by the potential efficiencies that could result from
privatization. Therefore, there is a rationale for a downward
adjustment. However, there is no convincing evidence that the
downward adjustment should be significant.
--------------------
\14 The enterprises could potentially lose some scale economies.
ADJUSTMENT RESULTS:
PRIVATIZATION WOULD PROBABLY
INCREASE AVERAGE INTEREST
RATES BY ABOUT 15 TO 35
BASIS POINTS
-------------------------------------------------------- Chapter 3:1.2
From the studies we analyzed, it appears that a reasonable estimate
of the conforming jumbo interest rate spread is currently about 20 to
40 basis points. Of the adjustments that need to be made to account
for differences between the two markets, the most important appears
to be the downward one for the potential gain in regional
diversification of credit risk. There is no convincing evidence that
the other adjustments should be significant; we assume that the
upward adjustment for liquidity does not exceed the combination of
the downward adjustments for the higher volatility of jumbo
collateral and the effect of operational efficiencies from increased
competition during most common mortgage market conditions. This
conclusion is largely based on observed liquidity in the jumbo
market, observed substitutions by mortgage borrowers and lenders
between fixed- and variable-rate mortgages, and Hermalin and Jaffee's
analysis of liquidity in the private label market.
Assuming that the sum of the liquidity, house price volatility, and
competition adjustments are a wash or near-wash, the estimated
interest rate spread could be adjusted downward by 5 to 10 basis
points for regional diversification benefits resulting from
privatization. Applying this assumption, we adjusted the estimated
interest rate spread of 20 to 40 basis points downward by 5 basis
points. From this, we concluded that privatization would probably
increase average interest rates within an average range of about 15
to 35 basis points.
INCREASE IN MORTGAGE INTEREST
RATES WOULD BE RELATIVELY
LARGER FOR BORROWERS MAKING
SMALL DOWN PAYMENTS
---------------------------------------------------------- Chapter 3:2
According to the enterprises' officials, the enterprises take account
of credit risk in their treatment of the mortgages they purchase, all
of which must meet their underwriting standards. For example, the
enterprises share some credit risk with private mortgage insurers and
generally require more mortgage insurance on mortgages with
loan-to-value ratios above 85 percent.
Both the enterprises and private-label conduits charge guarantee fees
for insuring the timely payment of principal and interest on their
MBS. The private-label conduits charge risk-based guarantee fees.\15
Although the enterprises have policies consistent with risk-based
fees, both the officials from the enterprises and other mortgage
industry participants told us that the enterprises do not charge fees
that are fully risk-based. Because privatization would likely
increase the number of secondary market competitors and change the
missions of the enterprises, it would probably motivate the
enterprises to implement fully risk-based fee structures. For this
reason, the increase in mortgage interest rates associated with
privatization would likely be relatively higher for borrowers making
small down payments and relatively smaller for borrowers making
larger down payments. As discussed more fully in chapter 4, one of
the negatively affected groups would be first-time homebuyers, who
tend to make relatively small down payments.
--------------------
\15 Risk-based fee structures charge based on the expected
incremental cost of providing a particular level of insurance for
credit risk exposure. For example, with risk-based fees borrowers
making large down payments (i.e., borrowers with low loan-to-value
ratios) will be charged a lower interest rate than borrowers making
small down payments. The enterprises indicated that their mortgage
commitment policies move them partially, but not fully, toward a
risk-based fee structure. We characterize this policy as one where
some cross-subsidy of riskier borrowers from less risky borrowers
occurs. To some degree, the enterprises attribute this policy to
their housing mission and their efforts to help first-time home
buyers. Our understanding is that private-label conduits attempt to
fully implement risk-based fees.
MORTGAGE INTEREST RATES COULD
FLUCTUATE MORE WITH DEMAND FOR
MORTGAGE CREDIT
---------------------------------------------------------- Chapter 3:3
Officials from both enterprises told us that primary and secondary
mortgage market liquidity would suffer with privatization, largely
because of the loss of the perceived guarantee of enterprise MBS. In
addition, the enterprises' increased borrowing costs could sharply
curtail or eliminate portfolio lending by the enterprises. Officials
from Fannie Mae emphasized that this decline in funding from retained
portfolio would reduce liquidity. This could result in less
liquidity generally, for particular mortgage products, or for
specific geographic markets during different parts of the economic
cycle, because the enterprises would not necessarily step into the
market to buy products whose price were falling. Officials from
Freddie Mac emphasized that the impact of privatization could not
only raise the average cost of financial capital to fund mortgages
but could also raise it more in periods of high demand for mortgage
credit.
Neither we nor the enterprises have quantified this liquidity effect
of privatization or estimated how much it would affect the mortgage
interest rate increase.\16 One reason for the liquidity of the
enterprises' securities is that regulatory guidelines governing
concentration of any one issuer's securities in the portfolios of
investors such as insurance companies and depository institutions do
not generally apply to securities issued by the enterprises, because
they are considered relatively low-risk government agency securities.
If privatization eliminates this agency status, many large mortgage
investors, including depositories, would likely have concentration
limits\17 on how much they could invest in each of the now private
conduits' securities. With privatization it is possible that a
relative scarcity of investors willing to accept private credit
enhancements of securities that were no longer perceived to have
government backing could develop during periods with high demand for
mortgage credit. However, as stated earlier, we have found no
statistical evidence that privatization would result in a substantial
reduction of liquidity in the secondary mortgage market. As a
result, mortgage interest rates could fluctuate more than they
currently do with demand for mortgage credit, but the extent of such
additional fluctuations is unknown.
--------------------
\16 Officials from Freddie Mac provided statistics from their primary
mortgage market survey comparing mortgage interest rates (adjusted
for the prepayment option) to yields on Treasury securities over the
past 10 years. The statistics indicated that the 1986 refinance boom
had a large upward impact on mortgage interest rates, but the 1992-93
refinance boom did not. Officials from Freddie Mac attributed this
increased stability to the growth of the enterprises. We attribute
this finding primarily to the development of more complex CMOs in the
financial community. Privatization could reduce activity by the
enterprises, but we question how much it would reduce the potential
benefits of newly created multiclass security products.
\17 Concentration limit is a limit on the extent to which an investor
can hold an individual company's securities. This protects investors
from credit risks imposed by undiversified holdings.
SIGNIFICANT REGIONAL
DISPARITIES IN INTEREST RATES
WOULD BE UNLIKELY TO DEVELOP
---------------------------------------------------------- Chapter 3:4
Before the creation of the enterprises, mortgages were funded by
depositories that primarily served local markets; this created
regional disparities in mortgage interest rates, resulting from
regional differences in the demand for and supply of mortgage credit.
The enterprises established a valuable secondary market mechanism
that enabled financial capital to flow to geographic areas with the
greatest demand for mortgage credit. This free flow of capital
tended to equalize interest rates across regions on mortgages with
similar risk characteristics.
Privatization is not likely to result in a return to a mortgage
market dominated by depositories holding mortgages in portfolio
because of the continuance of existing mechanisms (including the
private-label market) and tools to promote securitization, which the
enterprises fostered. On the other hand, the enterprises' levels of
mortgage funding could decrease, and we cannot be certain of the
extent to which other entities would be likely to "make up" this
decrease in funding. The possibility, with privatization, of a
decline in the level of mortgage funding by the secondary market
raises the question, however, of how much securitization and capital
mobility to fund mortgages are necessary to offset potential regional
interest rate disparities on mortgages with similar risk
characteristics.
To determine the likelihood that privatization would result in
regional interest rate disparities, we sought to determine the
relationship between the growth of the secondary mortgage market and
regional interest rate disparities. First, we analyzed regional
interest rate differentials (the difference between the highest and
the lowest regional mortgage interest rate) based on data for the
years 1980 through 1993 that Freddie Mac officials provided from
their Primary Mortgage Market Survey. It is important to note that
credit risk variables excluded from the data can create part of the
interest rate differentials. The regional interest rate differential
declined from 100 basis points in 1980 to less than 20 basis points
since 1988. This showed that interest rate disparities had lessened
substantially over time. However, the data did not show that the
reduction in regional interest rate disparities was due only to
greater secondary market activity, because other variables could have
influenced regional mortgage interest rates. Nonetheless, Freddie
Mac officials attributed this decline to the growth of secondary
mortgage markets created by the enterprises.
Evidence presented in a study by Jud and Epley\18 using statistics
for the years 1984 through 1987 indicated that after adjustments for
loan characteristic factors that affect interest rate differentials,
no significant regional differences remained in mortgage interest
rates. This evidence is consistent with the hypothesis that the
substantial development of the secondary market, facilitated by
government sponsorship, helped eliminate the regional interest rate
disparities that had existed before 1984. The result that
significant regional disparities were all but eliminated even when
the enterprises were much smaller than they currently are is also
consistent with the idea that the elimination of this disparity did
not require the enterprises to be as large as they are today. This
result, plus the growth and importance of private-label conduits
leads us to the conclusion that significant regional interest rate
disparities on mortgages with similar risk characteristics are not
likely to reappear with privatization.
--------------------
\18 G. Donald Jud and Donald R. Epley, "Regional Differences in
Mortgage Rates: An Updated Examination," Journal of Housing
Economics 1 (June 1991) pp. 127-139. They controlled for the
independent effects of a number of economic variables in measuring
regional differences. The Freddie Mac statistics do not control for
such variables.
PRIVATIZATION IS NOT LIKELY
TO RESULT IN INSUFFICIENT
CAPITAL MOBILITY ACROSS
REGIONS
-------------------------------------------------------- Chapter 3:4.1
Potential regional disparities in interest rates are also relevant to
analyzing the importance of the enterprises' operating "in all
markets at all times." Generally, mortgage lenders may be motivated
to tighten borrowing standards or charge higher fees in local markets
where housing prices are declining. Such behavior is consistent with
risk-based fee structures. Officials from Fannie Mae told us that
their charter and mission require them to operate in all markets at
all times. They said that one benefit of this requirement is that
they serve as a cushion in markets experiencing economic decline. As
an example, they stated that they continued to operate in and serve
the housing market in Texas throughout the economic decline in the
middle 1980s.
If Fannie Mae does not restrict credit to regions undergoing
recessions while other providers of credit do, Fannie Mae purchases
should represent a higher share of mortgage originations in years
when a region is in recession.\19 We received annual data on Fannie
Mae's market shares and housing price index for the years 1980
through 1994 for the states of Texas, Louisiana, Oklahoma, Colorado,
California, and Alaska and the New England region. We agree with
Fannie Mae officials in the statement that many factors affect the
level of participation of Fannie Mae and other lenders in any year.
We analyzed year-to-year correlations between Fannie Mae's share and
the housing price index and found no evidence that Fannie Mae
provides a cushion during downturns.
However, a Fannie Mae official aggregated data across years and said
the results provided evidence that Fannie Mae provides a cushion.
While aggregating statistics across years can be appropriate for
analyzing long-term trends in economic variables such as funding
levels and interest rate spreads, we question how appropriate such
aggregation is for analyzing cyclical trends. On balance, we did not
find sufficient evidence to determine whether or not Fannie Mae
provides a cushion during housing market downturns in specific
regions.\20
The continued market presence of the enterprises in all geographic
markets nationwide has helped to eliminate regional disparities in
mortgage interest rates and may provide a cushion for local housing
markets experiencing an economic downturn. Other financial
institutions that fund mortgages and mortgage insurance include those
that operate in specific geographic areas and base funding decisions,
including decisions on pricing and geographic limitations, on
expected profitability of each product in each geographic market.
Privatization would likely motivate the enterprises to adopt funding
decisions based on criteria more similar to those of other financial
corporations. In addition, the potential increase in secondary
market competition would reinforce this change in business behavior.
Even so, we conclude that significant regional disparities in
mortgage interest rates are unlikely to occur with privatization,
because securitization activity should provide sufficient capital
mobility across regions. Also, we do not think that privatization
would eliminate any substantial stabilizing mechanism for local
housing markets with declining market prices. In large part, this is
because we found little evidence that such mechanisms still require
government sponsorship to function effectively.
--------------------
\19 Here, we are examining evidence as to whether Fannie Mae provides
a cushion during a local housing market downturn. We are not
analyzing whether it is economically or socially desirable for Fannie
Mae to serve such a function.
\20 We do not dispute the statement that the enterprises undertake
policies that meet their mission and charter requirements to operate
in all markets at all times. The analysis here is our attempt to
ascertain the importance of such actions to the housing markets
during specific downturns.
PRIVATIZATION WOULD LIKELY
CHANGE THE BEHAVIOR OF
PARTICIPANTS IN THE PRIMARY AND
SECONDARY MORTGAGE MARKETS
---------------------------------------------------------- Chapter 3:5
Currently, conventional mortgages are funded by the enterprises and
depositories, while private-label conduits operate primarily in the
nonconforming market.\21 Virtually all conventional mortgages were
funded by depositories before the enterprises existed. However, for
a number of reasons, privatization would not likely cause a return to
this earlier environment.
One reason is the existence of private-label conduits, which were in
their infancy in the latter half of the 1980s. Their development is
largely attributable to two related factors: (1) the standardization
and technological innovation spurred by the enterprises and (2) the
general improvement in financial and information technology in the
economy. Private-label conduits, which currently specialize in
nonconforming mortgages (mostly in the jumbo market), accounted for
approximately 18 percent of combined Fannie Mae, Freddie Mac, and
private-label MBS outstanding and 13 percent of total MBS outstanding
as of September 1995.\22
--------------------
\21 The enterprises operate almost exclusively in the conventional
mortgage market. We examine possible implications for
government-insured mortgages, including those guaranteed by the
Government National Mortgage Association (Ginnie Mae), in our
discussion of social goals in chapter 4.
\22 Total MBS outstanding includes Ginnie Mae MBS.
THE ENTERPRISES AND OTHER
CONDUITS ARE LIKELY TO
COMPETE IN BOTH JUMBO AND
CONFORMING MORTGAGE MARKETS
-------------------------------------------------------- Chapter 3:5.1
If privatization were to lead to the enterprises' loss of both their
direct and indirect benefits--especially their funding advantage,
this would allow private-label conduits to operate on a more level
playing field with the enterprises in the conforming market. Because
privatization is likely to remove many, if not all, of the
enterprises' restrictions, the enterprises are likely to take the
opportunity to operate in the current jumbo market along with the
other conduits.\23
Should the enterprises' cost of funds rise from privatization, it is
likely that the overall amount of mortgage funding they provide,
whether out of retained portfolio or as MBS, would decline. However,
if the overall level of mortgage interest rates in the unified
(post-privatization) mortgage market rises, there would be incentives
generated for increased funding by private-label conduits in the
conventional market. If this increased funding occurs, it should
partially offset the enterprises' reduced funding.
To compete successfully in this new privatized market, it may be
necessary for any conduit to be a large organization. First, it
appears that there are financial and technological cost efficiencies
in the securitization process from operating on a large scale.
Second, such conduits would need regionally diversified loan pools to
keep the costs of their risks at a competitive level. Third, there
may be both incentives to and additional advantages from innovation
for firms that are a significant part of the mortgage market. For
example, it may improve efficiency and profitability to vertically
integrate or form networks within the housing finance system. This
could lead to further improvements in technology and advantages from
information sharing. As a result, we would not anticipate that a
large number of major firms would compete in this market.
--------------------
\23 Jumbo mortgages are relatively more prevalent in the state of
California and the northeast region, where housing prices are higher
than in other areas in the country. Therefore, geographic
diversification, which would decrease credit risk, would be
facilitated for any conduits operating in both market segments.
COMPETITION COULD SPUR OR
INHIBIT INNOVATION
-------------------------------------------------------- Chapter 3:5.2
While the possibility of additional competition in the housing
finance market could be a spur to increased innovation, the
possibility that the enterprises could lose their dominant position
may reduce their incentives to innovate. As government-sponsored
enterprises, Fannie Mae and Freddie Mac currently have cost
advantages (mostly funding) over potential rivals in the development
of efficiency generating innovations. In addition, their cost
advantages may have sheltered them from potential competitors in the
secondary market. Because of their market dominance, the financial
returns from developing innovations are likely to accrue to the
enterprises rather than to a multitude of competitors. To the extent
the enterprises' market share declines, privatization could cause the
enterprises to innovate less. The incentives to innovate by other
market participants, however, would increase with privatization. For
example, our discussions with industry participants and experts
indicated that large mortgage bankers would be more likely to develop
automated underwriting, appraisal, and mortgage servicing innovations
if the enterprises were privatized. Because of these offsetting
incentives, the net effect on the overall level of innovation is
impossible to predict.
CREDIT ENHANCEMENT
MECHANISMS COULD BE THE
KEY DETERMINANT OF LEVELS
OF CONDUITS' MORTGAGE
FUNDING
------------------------------------------------------ Chapter 3:5.2.1
An increase in the ability of private-label conduits to diversify
credit risks across a wider range of housing prices and geographic
locations could facilitate their expansion and could be a determining
factor in whether and to what extent these conduits would be able to
replace the expected decline in funding by the enterprises. As with
many financial products, credit enhancement mechanisms, such as pool
insurance and parent guarantees, have evolved over time. To the
extent this evolution takes advantage of enhanced efficiencies, it is
more likely to improve the overall functioning of the mortgage
market. The recent development of private-label MBS has motivated
development of credit enhancement mechanisms by issuers and
underwriters. Privatization could motivate even greater development.
One of the major uncertainties associated with privatization,
however, is how well market participants can develop credit
enhancement mechanisms that can provide the assurances required by a
wide range of mortgage investors. This uncertainty complicates the
task of estimating the growth of private-label conduits with
privatization of the enterprises.
WITH PRIVATIZATION,
DEPOSITORIES WOULD LIKELY
FUND MORE MORTGAGES THROUGH
VARIABLE-RATE LOANS
-------------------------------------------------------- Chapter 3:5.3
Competition between the enterprises and private-label conduits is
unlikely to fully offset the overall reduced availability of
secondary mortgage market financing that would likely result from the
enterprises' increased funding costs. To some extent, the need for
secondary mortgage market financing would also likely be less,
because the increased profit potential of mortgages resulting from
the expected rise in mortgage interest rates could induce some banks
and thrifts to hold more of the mortgages they originate in portfolio
rather than to sell them in the secondary market. To offset the
interest rate risk associated with fixed-rate mortgages, these banks
and thrifts could also be induced to originate more variable-rate
mortgages. Such mortgages are not sold as frequently as fixed-rate
mortgages in the secondary market. If banks and thrifts would hold
more of the mortgages they originate in portfolio, it could lead to
depositories' greater use of Federal Home Loan Bank (FHLB) System
advances.\24
Data show the depositories' increased use of variable rate mortgages.
Before the thrift crisis in the late 1980s, depositories tended to
originate long-term, fixed-rate mortgages funded by short-term
liabilities. About 6 percent of all mortgage holdings by thrifts
were variable rate in 1980. In 1993, about 47 percent of all jumbo
mortgage originations were variable rate; further, about two-thirds
of all mortgage holdings by thrifts and nearly 40 percent by
commercial banks were variable rate as of June 1995.
Unlike fixed-rate mortgages, variable-rate mortgages tend to be
funded by depositories rather than securitized, because they can be
held in portfolio with less interest rate risk. In 1993, less than
half of all jumbo originations--45 percent--were securitized,
compared to nearly 60 percent of conforming mortgage originations.
However, with privatization, to the extent that private-label
conduits would be better able to diversify risks geographically, the
share of mortgages that are securitized is likely to be greater than
that in the current jumbo mortgage market, although possibly smaller
than that currently observed in the conventional market.
--------------------
\24 The Federal Home Loan Bank System is another government-
sponsored enterprise that assists the residential mortgage market.
While privatization of Fannie Mae and Freddie Mac could reduce the
likelihood that the federal government might feel the need to rescue
these organizations, increased use of FHLBank advances could raise
that System's exposure. Our mandate did not ask us to evaluate the
impacts on the FHLBank System. Discussions of the System are
contained in Federal Home Loan Bank System: Reforms Needed to
Promote Its Safety, Soundness, and Effectiveness (GAO/GGD-94-38,
December 1993; GAO/T-GGD-95-244, testimony delivered September 27,
1995); and our correspondence to the Honorable James A. Leach on
proposed legislation entitled The Federal Home Loan Bank System
Modernization Act of 1995 (Oct. 11, 1995).
CONCLUSION
---------------------------------------------------------- Chapter 3:6
Privatization would likely change the behavior of market participants
and increase average interest rates on fixed-rate, single-family
mortgages within an average range of about 15 to 35 basis points.
However, privatization would not mean the end of the secondary
mortgage market, a return to regional disparities in mortgage
interest rates that were not based on differences in risk, or a lack
of mortgage credit in the economy during parts of the business cycle.
It would probably mean that mortgage rates would increase in areas
with higher risks, for houses with higher loan-to-value ratios, and
in periods of high mortgage demand.
ENTERPRISE COMMENTS AND OUR
EVALUATION
---------------------------------------------------------- Chapter 3:7
In oral comments, Fannie Mae and Freddie Mac officials disputed
several issues included in the draft version of this chapter. The
officials said that privatization would result in higher mortgage
interest rates than stated in the draft, and Fannie Mae officials
said they did not fully understand the methodology we used to
estimate the potential mortgage rate increase. Enterprise officials
also disagreed with statements in the draft that they said implied
the housing markets may lack competition and that the enterprises
exercise market power. Moreover, enterprise officials said that
privatization would generate significant regional variations in
mortgage costs, and they disagreed with our contention that there is
no sufficient evidence for concluding that the enterprises provide a
cushion during housing market downturns in specific regions. In
addition, Freddie Mac officials said that the increased use of
adjustable-rate mortgages (a form of variable-rate mortgage) would
result in higher mortgage foreclosure rates.
Fannie Mae officials said that privatization would likely raise
mortgage interest rates more than the 15 to 35 basis points estimated
in the draft report. They said that one reason for this disagreement
is that we did not adequately consider the impact that privatization
would have on the liquidity of the home financing system. On the
basis of discussions with private sector jumbo MBS traders who were
asked to list periods of illiquidity, a Fannie Mae official said that
the traders listed three periods of illiquidity over the past decade.
The traders told the Fannie Mae official that increasing interest
rates in 1994, combined with observed differences in jumbo prepayment
speeds by issuers, led to a period during which pricing existing
jumbo securities became extremely difficult. Because the jumbo
market has experienced such periods of illiquidity, the Fannie Mae
officials said it is not unreasonable to predict that the larger
mortgage market would experience similar illiquid periods and higher
mortgage rates in the event of privatization. In addition, they
thought that greater use of private-label credit enhancements would
result in higher mortgage rates. They did not, however, predict the
potential impact of reduced liquidity on mortgage interest rates.
Freddie Mac officials said that mortgage rates would increase by more
than 15 to 35 basis points; in fact, they predicted an increase of 55
to 86 basis points. The officials said that the spread between
conforming and jumbo rates ranged from 11 to 70 basis points between
1986 and 1996, with a mean spread of 43 basis points. They stated
that several factors resulting from privatization would cause
interest rates to increase by 55 to 86 basis points. For example,
they said that in the event of privatization, private-label issuers
would have to increase the volume of subordinated securities by 500
percent to replace the role of the enterprises. Freddie Mac
estimated that this change alone would add 25 basis points to the
estimated increase in mortgage rates. In addition, they said that
the commercial mortgage market, in which Freddie Mac does not
participate, experiences substantial periods of illiquidity.
Fannie Mae officials also said that we did not clarify our
methodology for estimating the spread between conforming and jumbo
loans prior to adjustments; we estimated a spread of 20 to 40 basis
points before adjustments. The Fannie Mae officials said that the
Cotterman and Pierce paper estimated a spread of 25 to 40 basis
points between conforming and jumbo loans and could not understand
why we used an estimated range of 20 to 40 basis points.
A Fannie Mae official also said that there is no evidence that the
enterprises exercise market power, and that the secondary market for
conforming loans is not a relevant market for analyzing market power.
Therefore, there is no meaningful duopoly consisting of Fannie Mae
and Freddie Mac. The enterprises are participants in the mortgage
financing market along with many other players, such as banks and
insurance companies, that also buy and sell mortgages. Additionally,
the Fannie Mae official stated that there were substantial flaws in
the Hermalin and Jaffee paper which contended that the enterprises
tacitly collude. For example, he said the authors reviewed data only
from 1989 to 1993 when an analysis of 1985 to 1995 would have
produced contrary results. The Fannie Mae official also said that
Hermalin and Jaffee ignored evidence that shows, on a monthly basis,
that the market share data of Fannie Mae and Freddie Mac are quite
volatile. He cited this as evidence that the enterprises do not
engage in tacit collusion.
Freddie Mac officials stated that there is no evidence of a lack of
competition in the mortgage markets. They said there is no basis for
excluding all firms that buy and sell mortgages from the definition
of the relevant market. Further, the Freddie Mac officials stated
that the guarantee fees the enterprises charge for securitization
services have declined since the early 1980s. They said that
declining fees are inconsistent with arguments that the enterprises
exercise market power. The Freddie Mac officials also reemphasized
comments they made on the chapter 2 draft that the financial benefits
of government sponsorship flow to homebuyers in the form of lower
interest rates and are not retained by the enterprises.
Fannie Mae officials also disagreed with an assertion in the draft
report that privatization would not result in significant regional
variations in mortgage interest rates. The officials said the report
acknowledged that privatization would result in risk-based pricing:
for example, homebuyers making relatively low down payments would pay
higher mortgage rates and fees. The Fannie Mae officials said they
could not understand why the draft report did not seem to consider
the possibility that with privatization, specific regions of the
country experiencing economic downturns would also experience
relatively higher mortgage costs. The Fannie Mae officials said that
this "risk premium" would probably become permanent in regions of the
country that are perceived to have volatile home prices. The
officials said this contrasts sharply with the current conforming
mortgage market where lenders nationwide can get the same posted cash
price for loans and homebuyers nationwide have access to the same
rates.
Freddie Mac officials also said that privatization would result in
significant regional variations in mortgage interest rates. For
example, they said that the regional variations observed in today's
jumbo mortgage market would likely be replicated in the larger
mortgage market. Freddie Mac officials also said that evidence from
regions of the country that have suffered economic downturns in
recent years, such as New England, indicate that lenders and
borrowers in these areas experience disparities in the cost and
availability of credit.
Fannie Mae officials also said the draft report ignored substantial
evidence that the enterprises currently provide a substantial
"cushion" to the housing markets regions of the country experiencing
economic downturns. For example, the officials said that the
enterprises' market share increased in such regions during economic
downturns. The officials also found that there was a significant
negative correlation between changes in the housing price index and
Fannie Mae's market share in California and New England between 1984
and 1994. In other words, when housing prices declined in these
areas, Fannie Mae's market share tended to increase, which officials
said demonstrates the regional cushion.
Freddie Mac officials disputed the draft report's analysis of
correlated annual data for 1980 to 1994, by state, on Fannie Mae's
market share and a house price index; this analysis found "no
evidence" that Fannie Mae provided a regional cushion. The officials
said that including early 1980s data ignores substantial changes in
the secondary market that occurred during those years. The officials
said that the data for the early 1980s is skewed because the
enterprises dramatically increased their mortgage purchase volume
during those years, particularly as a result of the introduction of
the Guarantor and Swap program in 1981 and CMOs in 1983. The
officials said that changing the beginning of the sample period from
1980 into 1985 changes the results. The officials stated that such
an adjustment showed, for example, a strong negative correlation
between declining house prices and Freddie Mac's market share in
three states that experienced substantial economic downturns:
California, New York, and Texas.
In addition, Freddie Mac officials said that the expected increase in
adjustable-rate mortgages at the expense of fixed-rate mortgages
would result in more mortgage foreclosures. The Freddie Mac
officials provided data on Freddie Mac purchased mortgages that show
the foreclosure rate on adjustable-rate mortgages between 1990 and
1995 was at least twice the foreclosure rate on fixed-rate mortgages,
even though adjustable-rate mortgages have higher down payment
requirements.
We explained how important the enterprises are to the housing markets
and we analyze the connection between the benefits conferred on that
market through the enterprises and benefits received by households.
We do not, however, believe that we can state how much of the
benefits generated flow to households. Nor can we say exactly how
privatization would affect the housing market.
Even so, we made a change to our draft report to address the
enterprises' concerns that we did not provide an overall measure of
the effects of lower interest rates on the mortgage market as a
whole. Using an estimate provided by Freddie Mac for the outstanding
value of conforming, conventional, fixed-rate mortgages, we
calculated the total benefit as ranging from $3 billion to $7
billion. We also clarified how we derived the spread between jumbo
and conforming fixed-rate mortgages. We also added more precise
language to indicate that we would not expect significant regional
variations in mortgage costs across regions on mortgages with similar
risk characteristics.
We have included information provided by a Fannie Mae official on
temporary disruptions in liquidity in the jumbo market. The official
did not know how serious these disruptions were. We continue to
conclude that the share of conventional mortgages that would be
securitized with privatization would likely exceed the current share
of jumbo mortgages securitized, and such a development would
contribute to a higher level of liquidity in the conventional market
with privatization than exists now in the jumbo market.
Our discussion indicates that we placed more emphasis on studies such
as the commissioned one by Cotterman and Pearce that use individual
loan level data and control for loan characteristics than we did on
other data sources reporting the interest rate spread between jumbo
and conforming mortgages. Officials from neither Fannie Mae nor
Freddie Mac criticized the Cotterman and Pearce study. Freddie Mac
estimated the spread using a data source different from the one they
had originally used and provided to us when we met in the course of
this assignment. Both data sources are based on telephone surveys of
lenders. We cannot determine why the spreads they now report are
larger than those they reported previously, but we continue to rely
primarily on the studies by Cotterman and Pearce and the other two
studies employing similar methodology. Freddie Mac officials
adjusted this spread upward by 25 basis points, on the basis of their
estimate of the effect on rates of an increase in the use of
subordinated securities used to finance many private mortgage pools.
We did not make such an adjustment because, in our view, it is likely
that interest rate spreads between jumbo and conforming mortgages
already reflect the impact of subordinated securities on jumbo
mortgages. Finally, we do not see how the commercial mortgage
market, a market in which loan underwriting decisions and
standardization are very different from the single-family residential
mortgage market, provides reliable information on the level of
liquidity that could result from privatization of the enterprises.
Our draft report did not take a position on whether the enterprises
do or do not have market power, because we could not, from our
analysis of the data, make such a determination. While the
enterprises would like us to conclude that they do not exercise
market power, we continue to conclude that there is insufficient
statistical evidence to reach such a conclusion. Both enterprises
emphasize that they compete vigorously both with each other and with
depository institutions. We think this evidence is insufficient to
conclude an absence of market power, because depository institutions
fund a higher share of variable-rate mortgages while the enterprises
fund relatively more fixed-rate mortgages. These products have
differing characteristics, and their competitive impacts on one
another depend on how highly substitutable they are to borrowers.\25
Both enterprises also criticize the commissioned study by Hermalin
and Jaffee, stating that the study does not consider how monthly
shares of secondary market purchases fluctuate between the
enterprises. Hermalin and Jaffee attributed the stability of annual
shares to their finding that the enterprises are tacitly colluding
duopolists in the (narrowly defined) secondary mortgage market for
conforming loans. In competitive markets the process of rewarding
the relative efficiency of one or more sellers tends to create
instable market shares measured over long periods of time.\26
The evidence the enterprises presented showing market shares that
fluctuated was based on monthly data and we believe it could just
reflect random or seasonal fluctuations in mortgage originations that
affect each enterprise differently (e.g., because the regional
distribution of their mortgages differ). Finally, Freddie Mac
officials argued that the general decline in guarantee fees by both
enterprises since 1985 indicates a competitive market where all of
the benefits to the enterprises flow through to borrowers.\27 The
data provided by Freddie Mac show that fees have declined, but they
do not show whether they are high or low compared with a competitive
market. The competitive process the enterprises have described was
largely in place in the 1980s, when fees were higher. Thus, the
decline in fees reflects either cost changes or an increase in
competition or potential competition. The private-label conduits, in
their infancy in the middle 1980s, may have provided a source of
potential competition.\28 CBO emphasized the possible impact of
potential competition on the enterprises when it stated: "Some
empirical evidence suggests that the GSEs may not have priced their
services at fully competitive levels in the 1980s."\29 Even if there
were evidence of some increased competition from private-label
conduits or other sources, we still do not know whether the market is
competitive enough to cause all or a large part of the benefits from
government sponsorship to flow through to households with mortgages.
After considering the enterprises' comments, we clarified our
discussion to indicate that we do not think privatization would lead
to significant regional disparities in mortgage interest rates that
were not based on risk differences. However, we did not change our
overall conclusion that privatization is not likely to significantly
reduce capital mobility across regions. We analyzed year-to-year
correlations between Fannie Mae's share of originations and their
housing price index in the states of Texas, Louisiana, Oklahoma,
Colorado, California, and Alaska, and in the New England region. A
negative correlation indicates that Fannie Mae could be providing a
cushion in declining markets. When we did the analysis using data
for the 1984 to 1994 period, as suggested by enterprise officials, we
found negative correlations in Texas and Oklahoma and positive
correlations in the remaining areas. We also reanalyzed the data for
the 1984 to 1994 period by estimating correlations between changes in
Fannie Mae's share and changes in the housing price index. In
addition to Texas and Oklahoma, the correlation for Colorado was also
negative. These results are also consistent with our original
conclusion that the evidence is ambiguous.
Finally, we have no evidence on what effects privatization would have
on foreclosure rates. We have no basis to evaluate the various
factors that may be associated with foreclosure rates on
adjustable-rate mortgages purchased by Freddie Mac.
--------------------
\25 See, for example, Dennis W. Carlton and Jeffrey M. Perloff,
Modern Industrial Organization (1990) page 738.
\26 See, for example, Richard A. Posner and Frank H. Easterbrook,
Antitrust: Cases, Economic Notes, and Other Materials (Second
Edition, 1981) p. 339.
\27 Average guarantee fees declined from the early 1980s until 1990,
remained relatively stable in 1991 and 1992, and increased slightly
in 1993 and 1994. Overall, there has been a decline from over 25
basis points to about 22 basis points.
\28 As we state in the report, we think the enterprises have played
critical roles in establishing and maintaining a nationwide secondary
mortgage market. We think that the development of the private-label
market has resulted, in part, from the success of the enterprises.
This development is integral to our assessment of how privatization
would affect the mortgage markets.
\29 CBO, The Federal Home Loan Banks in the Housing Finance System
(July 1993) p. 9.
PRIVATIZATION WOULD LIKELY REMOVE
ONE MECHANISM FOR CHANNELING
RESIDENTIAL MORTGAGE FUNDING TO
TARGETED GROUPS
============================================================ Chapter 4
Privatization would likely remove one of the federal mechanisms for
channeling residential mortgage funding to those borrowers and
geographic areas that lawmakers have deemed worthy of special
consideration (targeted groups). In our review of the enterprises'
activities that were designed to meet their social goal obligations
as established by HUD, we found little definitive evidence of how
housing affordability and homeownership opportunities for targeted
groups would be affected by privatization. The effects on targeted
groups of eliminating the enterprises' social goal obligations are
uncertain for three primary reasons. First, the effects would depend
largely upon whether other federal mechanisms that support housing
affordability and homeownership are maintained or expanded after
privatization and the impacts of those mechanisms. Second, it is
difficult to judge whether and how well the enterprises have achieved
their goals, because 1993 was the first year for which the
enterprises provided HUD the data necessary to monitor the amount of
funding provided to targeted groups under HUD's interim goals, and
the permanent goals HUD has recently promulgated have a new measure
of underserved areas. Third, neither we nor the enterprises were
able to quantify the impacts of the enterprises' social goal efforts
on housing affordability. Assuming that privatization leads to the
elimination of the enterprises' social goal requirements without any
change in other government mechanisms, the likely increase in
mortgage interest rates for single-family housing (the broad market
effects discussed in chapter 3) would make homeownership less
affordable. In particular, the increase in mortgage interest rates
could cause a delay in homeownership, primarily for young households
with low but rising incomes. Because the enterprises play such a
small role in the multifamily housing market, it is unlikely that
privatization would have a significant effect on mortgage interest
rates for multifamily housing or on housing affordability for
residents of such rental housing.
PRIVATIZATION WOULD LIKELY
REMOVE ENTERPRISES' OBLIGATION
TO CHANNEL RESIDENTIAL MORTGAGE
CREDIT TO TARGETED GROUPS
---------------------------------------------------------- Chapter 4:1
Privatization would likely eliminate one of the federal government's
means of channeling residential mortgage credit to borrowers and
geographic areas that lawmakers have designated for special
consideration. More specifically, privatization would likely
eliminate the enterprises' affirmative obligations as set forth by
The Housing and Community Development Act of 1992 (the 1992 Act):
"to facilitate the financing of affordable housing for low- and
moderate-income families in a manner consistent with [the
enterprises] overall public purposes, while maintaining a strong
financial condition and a reasonable economic return."\1
If the enterprises were privatized, HUD's regulation of the
enterprises to achieve social goals would likely have to be
eliminated for the following reasons:
-- The enterprises would have new charters that would eliminate
both privileges and restrictions specific to their housing
finance missions, and the social goals are now an integral part
of this overall organization.
-- If social goal requirements remained, the financial marketplace
might continue to perceive an implied federal guarantee for the
enterprises.
-- If the enterprises continued to face social goal requirements
and the new competitors that entered the secondary market did
not, there would not be a level playing field among the
secondary market entities.
We discussed one option that would continue HUD's social goal
regulation of the enterprises with HUD officials. It would involve
retaining some social goal regulation of the enterprises because of
possible residual advantages they would still have due to the period
of government sponsorship.\2 This issue is related to whether the
enterprises should pay some sort of exit fee (directly or indirectly
in the form of social goal requirements) upon privatization for
benefits received during the period of government sponsorship.
However, based on our discussions with industry participants and
regulators, it seems likely that social goal regulation of the
enterprises by HUD would not continue following privatization. As
discussed in chapter 5, if Congress decides to privatize, it could be
important to convince the markets that links between the enterprises
and the government are broken, in order to change investors'
perceptions about any implied guarantee. It could be harder to
convince the markets if some residual social goals remained for the
privatized entities.
--------------------
\1 Pub. L. No. 102-550 � 1302(7).
\2 In other words, the enterprises have developed technology and
operating systems over an extended period of time during which they
had certain advantages. If the enterprises were privatized, it could
take potential competitors time and resources to effectively catch up
to and effectively compete with the enterprises.
THE IMPACTS ON TARGETED GROUPS
OF THE ENTERPRISES' SOCIAL
GOALS ACTIVITIES ARE DIFFICULT
TO MEASURE
---------------------------------------------------------- Chapter 4:2
In our review of the enterprises' activities to meet social goal
requirements, we found little definitive evidence of how housing
affordability and homeownership opportunities for targeted groups
would be affected by privatization. Fannie Mae has devoted extensive
resources to special programs to meet social goal requirements and
help fulfill its housing mission. Freddie Mac has devoted extensive
resources to pilot programs and related activities, such as its
Underwriting Barriers Outreach Group program, to expand housing
opportunities both generally and for underserved areas and groups.
However, quantification of the enterprises' efforts at the time of
our review was generally a measurement of resource commitments rather
than outcomes.
THE PURPOSES OF THE
ENTERPRISES' SOCIAL GOALS
WERE ESTABLISHED BY THE
FEDERAL HOUSING ENTERPRISE
FINANCIAL SAFETY AND
SOUNDNESS ACT OF 1992
-------------------------------------------------------- Chapter 4:2.1
As discussed in chapter 1, two of the statutory purposes of the
enterprises are
-- to provide ongoing assistance to the secondary market for
residential mortgages (including activities relating to
mortgages on housing for low- and moderate-income families\3
involving a reasonable economic return that may be less than the
return earned on other activities) by increasing the liquidity of
mortgage investments and improving the distribution of investment
capital available for residential mortgage financing; and
-- to promote access to mortgage credit throughout the nation
(including central cities, rural areas, and underserved areas)
by increasing the liquidity of mortgage investments and
improving the distribution of investment capital available for
residential mortgage financing.\4
The 1992 Act required HUD to promulgate rules that set forth goals
for the enterprises to meet in purchasing mortgages made to
designated income groups and in geographic areas defined as
underserved. The motivation for promulgation and enforcement of the
social goals was partially attributed by individuals we interviewed,
to the perception that the enterprises' distribution of conventional,
conforming loan funding going to low- and moderate-income borrowers
was lagging behind the primary mortgage market's funding of such
mortgages. A Federal Reserve Board study using 1992 Home Mortgage
Disclosure Act data supported this perception.\5
The purpose of the goals is to increase the total supply of
residential mortgage funds to targeted borrowers, which in turn could
reduce mortgage costs for such borrowers. The impact on mortgage
costs depends on how much the social goals serve to increase
enterprise funding levels to targeted borrowers and how mortgage
originations by other lenders (namely depository institutions that
undertake portfolio lending and mortgage bankers who originate
federally insured mortgages for Ginnie Mae mortgage pools) are
affected. It is easier to quantify how the social goals affect
enterprise activities than it is to quantify the final market
outcomes of such activities.
--------------------
\3 Household borrowers are defined as low-income if income does not
exceed 80 percent of area median family income. Moderate-income
includes household borrowers with incomes that do not exceed area
median family income. The relevant geographic areas are the
metropolitan area for metropolitan residents and non-metropolitan
counties in the state for rural residents. Households who reside in
rental housing units count toward the goals based upon whether the
rent level in the housing unit is affordable to a very low-, low-, or
moderate-income resident, whichever is relevant.
\4 12 U.S.C. � 1716.
\5 Glenn B. Canner and Wayne Passmore, "Residential Lending to
Low-Income and Minority Families: Evidence from the 1992 HMDA Data,"
Federal Reserve Bulletin (February 1994) pp. 79-108. The authors
state that their results may reflect the relatively high rates of
residential mortgage refinancing in 1992.
EXPECTATIONS REGARDING THE
NATURE OF THE ENTERPRISES'
SOCIAL POLICY-RELATED
ACTIVITIES ARE UNCLEAR
-------------------------------------------------------- Chapter 4:2.2
The broad purposes of the 1992 Act do not answer a number of
questions about legislative expectations of HUD and the enterprises
in their implementation of these social goal requirements. For
example:
-- Should the enterprises' promotion of access to mortgage credit
throughout the nation provide remedies to alleviate possible
imperfections in private mortgage markets such as those created
by racial discrimination? Or, should the enterprises improve
the distribution of investment capital using some different
standards?
-- Should HUD promulgate separate subgoals for central cities and
rural areas, or specify one or more geographic areas that it
considers underserved?\6
--------------------
\6 Section 1334 of the 1992 Act requires HUD to establish "an annual
goal for the purchase by each enterprise of mortgages on housing
located in central cities, rural areas and other underserved areas."
The section further authorizes HUD to establish "separate subgoals
within the goal" under the section but specifies that the subgoals
are not enforceable.
HUD GOALS HAVE CHANGED WITH
SHIFTS IN HUD OBJECTIVES
-------------------------------------------------------- Chapter 4:2.3
The 1992 Act directed HUD to promulgate regulations setting annual
goals for each enterprise for the purchase of mortgages relating to
each of the following three categories:
-- housing for low- and moderate-income families;
-- housing located in central cities, rural areas, and other
underserved areas; and
-- rental and owner-occupied housing for low-income families in
low-income areas and for very low-income families.
These goals were set in part to bolster HUD's monitoring and
enforcement of goals for both enterprises that previously had been
established only for Fannie Mae.\7
The 1992 Act established a transition period of calendar years 1993
and 1994 to allow time for HUD to collect data and implement these
requirements and provided interim annual purchase goals for each
enterprise during the period. Under these goals, 30 percent of the
total number of dwelling units financed by mortgage purchases of each
enterprise during the year were to be from mortgages serving low- and
moderate-income families and likewise 30 percent of dwelling units
were to be for housing located in central cities designated as such
by the Office of Management and Budget (OMB). The amounts were
essentially the same as the percentage goals (known as the "30/30
goals") that had been previously established for Fannie Mae under
HUD's regulations. Authority for the twin 30/30 goals was contained
in the 1968 chartering legislation for Fannie Mae, but they were not
promulgated until 1979. These goals were not based on any analytical
studies, and, as we understand, they were never monitored or
enforced. In addition, the 1992 Act established interim "special
affordable housing goals" for each enterprise to acquire mortgages
serving low-income families in low-income areas and very low-income
families.\8 Under these goals, Fannie Mae was to purchase at least $2
billion in such mortgages during the period, while Freddie Mac was
required to purchase a volume of at least $1.5 billion.\9
According to HUD officials, HUD had originally begun research on
social goal regulations for the enterprises as early as 1989. The
agency's approach to this area, at that time, was to ensure that the
benefits from government sponsorship were equally distributed across
all borrowers. Following passage of the 1992 Act and the beginning
of the Clinton administration in January 1993, this approach shifted
somewhat. HUD's policy became one in which the enterprises should
lead the market for lending to low- and moderate-income and other
underserved borrowers, rather than simply mirroring the primary,
conforming, conventional mortgage market. HUD officials are
presently considering the appropriate scope of this shift. If
mirroring the market means that the enterprises fund a share of
mortgages benefiting a targeted group equal to the share observed in
the overall primary market, "leading the market" could be interpreted
to mean that the enterprises should devote larger shares of their
funding to targeted groups. If social goal regulations were to
require leading rather than mirroring the market, it would be more
likely that housing opportunities and affordability for targeted
borrowers would be improved.
--------------------
\7 See 24 C.F.R. Part 81 (1992). Part 81 set goals for purchases by
Fannie Mae of mortgages by very low-, low-, and moderate- income
families and mortgages for residential properties in central cities.
These goals essentially targeted 30 percent of Fannie Mae's annual
mortgage purchases for low- and moderate-income mortgages and 30
percent for mortgages in central cities. HUD was granted regulatory
authority over Freddie Mac in 1989 under the Financial Institutions
Reform, Recovery, and Enforcement Act of 1989 (FIRREA), Pub. L. No.
101-73 � 731(c). However, prior to the 1992 Act, HUD had not
extended the housing goals to Freddie Mac. Also before the 1992 Act,
HUD had not consistently enforced the housing goals and had not
collected mortgage data sufficient to monitor compliance with the
goals. HUD's new Part 81 goals are discussed later in this report.
\8 Generally, the affordable housing goals define low-income
households as having income not exceeding 80 percent of area median
income and very low-income households as having income not exceeding
60 percent of area median income. Multifamily rental units count
toward the goals based on the affordability of rent levels for low-
and very low-income households.
\9 The special affordable housing goal for 1993 and 1994 was, in
effect, an increase above each enterprise's estimated 1992 purchases
fulfilling the definition. The goal targets lending to low- and very
low-income borrowers and therefore is primarily geared to rental
housing. We were told by HUD officials that HUD relied heavily on
input from Fannie Mae and a number of housing advocacy groups in
developing this goal.
THE PERMANENT GOALS
PROMULGATED BY HUD REQUIRE
THE ENTERPRISES TO MIRROR
THE PRIMARY, CONFORMING
MARKET
-------------------------------------------------------- Chapter 4:2.4
The goals established for the enterprises are based, in part, on the
targeted groups' shares in the primary, conforming, conventional
market. The relevant comparison was the primary market because the
secondary, conforming, conventional market is so dominated by the
enterprises that they would always mirror it. In 1993, HUD published
a notice of proposed housing goals under the 1992 Act that included
interim goals for the enterprises for 1993 and 1994.\10 Final goals
were promulgated on December 1, 1995, effective January 2, 1996.\11
For low- and moderate-income housing, the goals are 40 percent of
mortgage purchases during 1996 and 42 percent yearly during 1997
through 1999. The special affordable housing goals (for mortgages of
low-income families in low-income areas and very low-income families)
are 12 percent of all mortgage purchases in 1996 and 14 percent
yearly during 1997 through 1999.
The underserved area component replaced the old central city
requirement.\12 Purchases are to count toward the goal if the census
tract has median income below 120 percent of median income for the
overall metropolitan area (nonmetropolitan areas in the state if a
rural census tract) and at least 30 percent of the residents are
minority. Purchases also count if census tract median income is
below 90 percent of median income for the overall metropolitan area
and, in rural areas, if census tract median income is below 95
percent of median income for nonmetropolitan areas in the state. For
purchases of mortgages on housing located in underserved areas, the
goals are 21 percent of purchases in 1996 and 24 percent yearly
during 1997 through 1999.
--------------------
\10 58 Fed. Reg. 53048 (Fannie Mae), 53072 (Freddie Mac) (October
13, 1993).
\11 60 Fed. Reg. 61486 (Dec. 1, 1995).
\12 See footnote 7 on Part 81 requirements.
FANNIE MAE'S PRODUCTION
LEVELS EXCEEDED GOALS;
FREDDIE MAC'S PERFORMANCE
WAS MIXED
-------------------------------------------------------- Chapter 4:2.5
HUD estimated the percentage of each enterprise's purchases in 1994
that met the income, special affordable, and underserved area
components in the new final rule (see table 4.1). Fannie Mae's 1994
production levels exceeded the goals set for the remainder of the
decade in the final rule. Freddie Mac's 1994 production exceeded the
underserved areas goal but fell short of the low- and moderate-income
and special affordable goals set for the remainder of the decade.\13
Each enterprise's production toward each goal in 1994 exceeded the
share attained the previous year.\14
Table 4.1
Enterprise Performance in Relation to
Newly Established HUD Requirements
Estimate
d share
1997- of
Fannie Freddi 1996 99 primary
Requirement Mae e Mac Goal Goal market
---------------------------- ------ ------ ------ ------ --------
Low-and moderate-income 46% 37% 40% 42% 48-52%
Special affordable 17 11 12 14 20-23
Underserved areas 29 24 21 24 25-28
----------------------------------------------------------------------
Source: HUD.
--------------------
\13 Fannie Mae met the central city goal in 1994 with 31.5 percent of
its purchases; Freddie Mac did not meet the goal with 25 percent of
purchases from central cities. The special affordable housing goal
for the 2-year period beginning in 1993 had a multifamily and
single-family component. Fannie Mae met each component and the total
requirement. Freddie Mac met the single- family and total
requirement but did not meet the multifamily component.
\14 Fannie Mae officials attributed their increased percentages, in
part, to increased efforts to reach out to targeted borrowers. In
contrast, Freddie Mac officials told us that their improved
performance statistics for 1994 were more affected by changes in
economic conditions than HUD's regulatory oversight. Specifically,
measures of homeownership affordability were favorable at the same
time mortgage refinancings were declining. These events increase the
share of borrowers who are first-time homebuyers, and borrowers who
refinance tend to have above- average incomes.
THE ENTERPRISES DIFFER IN
THEIR APPROACHES TO
IMPLEMENTING SOCIAL GOAL
OBLIGATIONS
-------------------------------------------------------- Chapter 4:2.6
Officials of both enterprises told us that their charters and the
1992 Act are consistent with their mission requirements to be in all
markets at all times. Both enterprises emphasized that their
standard programs are designed to benefit all homebuying
borrowers--including those that are low- and moderate-income,
minority, or underserved area residents, and both have targeted
lending programs to support homeownership and housing affordability
for targeted groups. The enterprises, however, have differing
perceptions of how they should respond in meeting the regulatory
social goals.
FANNIE MAE'S TARGETED
LENDING PROGRAMS
------------------------------------------------------ Chapter 4:2.6.1
Fannie Mae has a number of special programs that are designed to
reach out to central city, low-income, and minority and ethnic group
borrowers who may feel disenfranchised from the housing finance
market and the attainment of homeownership. Fannie Mae officials
stress the importance of their outreach efforts with community groups
in this process. These efforts are reflected in Fannie Mae's strong
support for a central city lending goal, which they argue is legally
required by the 1992 Act. Fannie Mae also has consistently purchased
mortgages supporting multifamily rental housing, which is reflected
in its support for the special affordable housing goal. Fannie Mae
officials generally view the low-income, central city, and special
affordable goals as a reaffirmation, in part, of their housing
finance mission.
Fannie Mae officials told us that their standard business practices,
in addition to their special programs, provided benefits to customers
with characteristics similar to the targeted groups. For example,
because the fees they charge on MBS may not be risk-based, borrowers
who make high down payments may be charged more and those who make
low down payments less than they would be charged if fees were truly
risk-based. Fannie Mae officials said that the general intent of
such a cross-subsidy would be to facilitate first-time homeownership.
They also indicated that this form of cross-subsidy is not
systematically related to borrower income.\15
Fannie Mae officials said that whatever cross-subsidization affected
their targeted lending programs was due to the extra administrative
costs of these programs.\16 About 8 percent of Fannie Mae's 1994
purchases were accounted for by targeted lending programs. They
stated that the benefits for targeted borrowers tend to be textural
compared to explicit subsidy programs where benefits can be more
easily quantified.\17 The expected benefit of many of these outreach
efforts is to bring more households, including future generations,
into the housing finance and homebuying system.
--------------------
\15 In other words, in an analysis of total Fannie Mae purchases the
distribution of loan-to-value ratios among borrowers is similar among
low- to moderate-income and higher income borrowers. We obtained
statistics from HUD for each of the enterprises' total book of
business in 1993 on loan-to-value ratios by income group. The
statistics are consistent with the notion that loan-to-value ratio
was not systematically related to borrower income.
\16 We were told by Fannie Mae officials that Fannie Mae accounting
data are not sufficiently detailed to explicitly compare
administrative costs among lending programs.
\17 An example of a textural benefit for targeted borrowers is
literature produced by Fannie Mae, such as Choosing the Mortgage
That's Right For You, which the enterprise printed in seven different
languages.
FREDDIE MAC'S TARGETED
LENDING PROGRAMS
------------------------------------------------------ Chapter 4:2.6.2
Freddie Mac has a number of pilot programs designed to identify
cost-effective methods to expand housing opportunities. The intent
is to identify such methods for subsequent implementation into
standard Freddie Mac mortgage products. The programs' primary
emphasis is on identifying inefficiencies in mortgage markets that
could result from possible discrimination and arbitrary underwriting
standards. Freddie Mac officials said they generally view HUD's
social goal regulatory enforcement as a monitoring device to verify
that the enterprises are serving all parts of the primary mortgage
market rather than as a device that has a substantial independent
effect on their allocation of mortgage credit.
Freddie Mac officials emphasized the role of their special affordable
targeted lending initiatives as pilot programs meant to identify
cost-efficient methods to expand homeownership opportunities. For
example, Freddie Mac officials emphasized their Underwriting Barriers
Outreach Group (UNBOG) activities in reaching out to prospective
homebuyers and expanding homeownership opportunities. UNBOG created
focus groups comprising members of organizations involved in
community lending issues. They were asked what Freddie Mac
underwriting standards were perceived as being barriers to community
lending, especially in communities that could be considered
underserved. On the basis of the responses of the focus group
participants, Freddie Mac clarified underwriting standards that were
perceived as creating barriers to lending in particular communities.
The clarifications apply to Freddie Mac's standard purchase programs.
This effort appears to be consistent with Freddie Mac's philosophy
that its major mission is to make sure that all parts of the primary
mortgage market are served by its products.
THE IMPACTS ON TARGETED
GROUPS OF THE ENTERPRISES'
SOCIAL GOAL ACTIVITIES ARE
DIFFICULT TO MEASURE
-------------------------------------------------------- Chapter 4:2.7
Fannie Mae is devoting extensive resources to special programs to
meet social goal requirements and help fulfill its housing mission.
Freddie Mac is devoting extensive resources to pilot programs and
related activities, such as its Underwriting Barriers Outreach Group
program, to expand housing opportunities generally and in areas and
to groups that are perceived to be underserved. Privatization would
likely cause a decline in such efforts by the enterprises. However,
neither we nor the enterprises are able to quantify the impacts of
these efforts on housing affordability and homeownership
opportunities among different borrowers. Whatever quantification of
these efforts exists is generally a measurement of resource
commitments and not outcomes, such as the impacts on mortgage
interest rates and housing affordability for targeted groups.
A recent Federal Reserve Board study estimated the amount of credit
risk on lower income and minority borrower mortgages taken on by
different participants in the mortgage market.\18 Although the study
does not measure outcomes related to housing affordability and
homeownership opportunity, it does estimate the supply of one of the
more important inputs, namely the ability and willingness to
undertake credit risk, affecting the supply of mortgage credit. The
authors expected that the enterprises would promote homeownership
among lower income households more than entities such as depository
institutions. They found, however, that depositories take on more of
the total credit risk associated with lower income lending than the
enterprises.\19
From this they concluded: "The difference [in taking on credit risk]
may arise because Fannie Mae and Freddie Mac, unlike depositories,
generally have no interactions with borrowers and are not located in
the neighborhoods where the mortgages are originated; thus they lack
the opportunity to look beyond traditional measures of risk." Thus,
the enterprises, as secondary market participants, may not be as well
situated as a primary lender to effectively distinguish more
creditworthy targeted groups from less creditworthy targeted groups.
There are other reasons why knowing the extent of the enterprises'
resource commitments is not sufficient to allow quantification of the
program outcomes for targeted borrowers. First, it is necessary to
determine how much the social goals serve to increase enterprise
funding levels to targeted borrowers compared to what they would have
been without the goals. On this score, we observe that the
enterprises have increased mortgage funding to targeted groups. Even
if some of this increase is due to other economic factors, the goals
have likely caused part of this expansion. Second, it would be
necessary to determine how mortgage originations by other lenders,
namely depository institutions who undertake portfolio lending and
mortgage bankers who originate federally insured mortgages for Ginnie
Mae mortgage pools, are affected and respond to this change in
funding. On this score, we are uncertain.
--------------------
\18 Glenn Canner and Wayne Passmore, "Credit Risk and the Provision
of Mortgages to Lower-Income and Minority Homebuyers," Federal
Reserve Bulletin (November 1995).
\19 A Fannie Mae official stated that the study (1) ignored the big
role Fannie Mae plays in extending credit to the underserved, (2)
relied on incomplete data, (3) relied on 1994 data even though 1994
was a year with an unusually high share of mortgages that were
variable-rate mortgages funded by depositories, (4) included
federally insured FHA and Veterans Affairs (VA) mortgages, and (5)
did not take into account the credit risk taken by Fannie Mae on
mortgages with private mortgage insurance (PMI). The authors'
response stated that their general finding that depositories take on
more credit risk than the enterprises holds with inclusion or
exclusion of federally insured loans and with private mortgage
insurers or the enterprises taking on the credit risk on mortgages
with PMI.
THE MARKET EFFECTS OF
PRIVATIZATION COULD RESULT IN A
DELAY OF HOMEOWNERSHIP FOR SOME
LOW-INCOME FAMILIES
---------------------------------------------------------- Chapter 4:3
Assuming privatization and no adjustments or change in any federal
mechanism supporting housing affordablity and homeownership, it is
likely that mortgage interest rates could increase by about 15 to 35
basis points on average, with larger increases likely for homebuyers
making relatively small down payments (as discussed in ch. 3). One
of the five studies commissioned to help assess privatization
analyzed the implications of higher mortgage interest rates on
housing affordability and homeownership.\20 The authors developed an
economic model in which underwriting requirements created constraints
(such as minimum downpayments or monthly payment to income ceilings)
that would keep some prospective borrowers from purchasing a home of
the size and value they would be expected to prefer on the basis of
household characteristics and expected future income patterns.
The authors performed a statistical simulation to estimate the impact
of a 50 basis point increase in mortgage interest rates on
homeownership. They used 50 basis points as an upper bound of how
much privatization could affect mortgage rates.\21
The authors estimated that their baseline homeownership rate\22 of
63.6 percent would have been about 1.2 percentage points lower if
mortgage interest rates had been 50 basis points higher as a result
of privatization (see table 4.2). The estimated impacts on minority,
low-to-moderate income, and young households\23 would have been more
pronounced; the respective estimated downward impacts on the
homeownership rate would have been about 2.8 (from a baseline of 43.9
percent), 2.6 (from a baseline of 45.7 percent), and 3.5 percentage
points (from a baseline of 33.7 percent).
Table 4.2
Estimated Impacts of Privatization on
Homeownership Rates
Estimated
homeownership
Baseline (1989) rate
Household group homeownership rate with privatization
------------------------------ ------------------ ------------------
All 63.6% 62.4%
Minority 43.9 41.1
Low-to-moderate 45.7 43.1
income
Head under 30 years 33.7 30.2
old
----------------------------------------------------------------------
Source: Wachter, Follain, et al.
The authors' statistical analysis indicated that the primary impact
of the interest rate increase associated with privatization on
homeownership rates was due to an increase in the relative cost of
homeownership compared with the cost of rental housing. The
remainder was accounted for by the estimated impacts of privatization
on down payment and monthly payment to income constraints associated
with underwriting standards. For example, with higher mortgage
interest rates, more potential homeowners would find that their ratio
of mortgage payments to income would be above current underwriting
standards. Because the authors are comparing the cost of
homeownership to the alternative of renting, the relative cost impact
depends on the authors' analysis of the impacts of privatization on
multifamily housing.
--------------------
\20 Susan Wachter, James Follain, Roberto G. Quercia, Peter
Linneman, and George McCarthy, "Implications of Privatization: The
Attainment of Social Goals," published in Studies on Privatizing
Fannie Mae and Freddie Mac (forthcoming May 1996).
\21 The 50 basis points represent the total impact from privatization
on the mortgage interest rates paid by households from different
income, race and ethnic, and age groups. The 50 basis points is
meant to include the interest rate impact from the expected decline
in targeted lending programs by the enterprises associated with
privatization.
\22 The homeownership rate equals the percentage of households
residing in owner-occupied housing.
\23 Young households are defined in the study as those with a head of
household under 30 years old.
PRIVATIZATION WOULD BE
UNLIKELY TO HAVE A
SIGNIFICANT IMPACT ON
MORTGAGE INTEREST RATES ON
MULTIFAMILY HOUSING
-------------------------------------------------------- Chapter 4:3.1
On the basis of limited statistical information, the authors found
that privatization would not have a significant impact on mortgage
interest rates on multifamily housing.\24 As a result, they concluded
that multifamily housing concerns should not be the basis of policy
decisions on privatization. In addition, this conclusion affected
the authors' estimates of the effects of privatization on
homeownership because the most important variable used to estimate
the homeownership impact is the relative cost of owning versus
renting a housing unit. Their estimates assume that privatization
would increase single-family mortgage interest rates, and therefore
the cost of owning, but not have any significant effect on
multifamily mortgage interest rates, and therefore the cost of
renting. If the cost of rental housing were affected by
privatization proportionally to the cost of owner-occupied housing,
the relative cost of owning versus renting would be unaffected, and
most of the estimated impacts on homeownership rates would have
disappeared.\25 In addition, privatization could cause mortgage
interest rates on single-family rental housing, and thus rental costs
on such housing, to increase.
The primary reason why it appears unlikely that the supply of
multifamily housing would be affected by privatization is that the
enterprises currently finance little multifamily activity and with
privatization are more likely to do less than more.\26 The
enterprises' purchases of multifamily mortgages represent a small
share of their total purchases.\27 Fannie Mae's purchases of
multifamily mortgage purchases accounted for $4.8 billion in 1994,
about 3 percent of total Fannie Mae purchases. Fannie Mae officials
told us that the $4.8 billion was about 11 percent of total 1994
multifamily mortgage originations. They added that Fannie Mae's 1995
multifamily purchases were $6.5 billion, which were about 20 percent
of total multifamily originations. Freddie Mac purchased $913
million in multifamily loans in 1994, less than 1 percent of total
purchases.\28 Unlike the jumbo market, there are no prohibitions or
constraints keeping the enterprises from expanding in this area. It
is their existing and prospective social goals that are motivating
much of the multifamily financing they are currently doing or are
contemplating for the future.\29 Without those goals, they would
probably do less. However, due to their limited role, such a
reduction or withdrawal is not likely to have much effect on either
the supply or the rental cost of multifamily housing.
--------------------
\24 For various reasons, a larger share of multifamily units than
single-family units contribute toward fulfillment of the housing
goals. Renters who reside in multifamily housing units generally
have lower incomes than homeowners; also, the preponderance of
multifamily housing is also greater in central cities than in other
geographic locations. Therefore, the general increase in
requirements promulgated by HUD for 1996 through the remainder of the
decade creates incentives for expanded multifamily purchases by the
enterprises.
\25 An official from Fannie Mae stated at the Wachter and Follain
seminar that lowering multifamily housing financing costs may be the
next frontier for the enterprises to bring additional benefits to the
housing market. If such events occur in the future, households could
benefit from lower rental costs. In the context of the Wachter
analysis, however, the relative cost impact from privatization that
lowers the expected homeownership rate would partially disappear. It
should also be noted that the Wachter analysis assumes that the cost
of multifamily housing (i.e., housing units in structures with five
or more units) determines the cost of rental housing. Over half of
the rental housing stock in the nation is accounted for by
single-family housing (i.e., housing units in structures with one to
four housing units).
\26 Here we are not reaching any conclusion with respect to whether
more financial capital should be committed to multifamily housing.
In the absence of privatization, we assume that social goal
regulation would likely motivate increased funding of multifamily
mortgages by the enterprises in the future.
\27 HUD officials told us that they are developing programs in which
the enterprises and FHA would cooperate in risk-sharing arrangements
for multifamily housing. We do not know whether such cooperative
arrangements would improve housing affordability for lower income
rental housing tenants.
\28 Freddie Mac experienced substantial losses in its multifamily
business in the late 1980s, discontinued new multifamily activity for
a number of years, and did not begin new business until December 1,
1993. Freddie Mac's March 1, 1995 report to HUD states: "Freddie
Mac now underwrites multifamily loans individually and treats them in
a fashion more similar to commercial business loans than to
single-family loans."
\29 A Fannie Mae official said that Fannie Mae has always funded
multifamily mortgages. He said that the social goals change the
focus of this activity toward targeted groups.
STUDY FOUND THAT
PRIVATIZATION COULD DELAY
FIRST-TIME HOMEOWNERSHIP OF
LOW-INCOME HOUSEHOLDS
-------------------------------------------------------- Chapter 4:3.2
The authors also distinguished between the impacts of privatization
on current ownership and when households first become homeowners.\30
Borrowing constraints created by mortgage underwriting standards can
be overcome over time if households save for a downpayment over a
longer period of time. In addition, borrowing constraints tend to be
greater for households with low current income who have relatively
high levels of expected income in the future, because the optimal
house that first-time homebuyers purchase is dictated in part by
their expected future incomes. The authors found that a majority of
the households with low current income had relatively high levels of
expected future income. Therefore, it can be expected that one of
the primary impacts of privatization on homeownership would be to
delay rather than permanently preclude homeownership for the group of
households with low current income and relatively high expected
future income.
Even if privatization's effect on interest rates were only to delay
and not preclude homeownership, such a delay could still have social
costs. Among the many reasons stated by Members of Congress for
providing favorable tax and financial treatment to homeownership is
the belief that owning a home fosters wealth accumulation and family
stability. If so, then attaining homeownership at a younger age by
households with relatively low but rising incomes could help promote
such social goals. Furthermore, the attainment of homeownership by
households with low incomes that are not expected to increase could
include the accrual of wealth accumulation and family stability over
protracted time periods as well as other benefits, such as fostering
stronger community ties among neighborhood residents.
--------------------
\30 They use the terms the "current homeownership rate" and the
"ever-own rate." The latter is a measure of the percentage of
households who at some time become homeowners. Fannie Mae comments
on the Wachter analysis argued that the ever-own rate was a rather
meaningless theoretical construct and policy decisions should be
based on the impacts of privatization on the current homeownership
rate.
THE FEDERAL GOVERNMENT HAS A
NUMBER OF MECHANISMS TO SUPPORT
HOUSING AFFORDABILITY AND
HOMEOWNERSHIP
---------------------------------------------------------- Chapter 4:4
HUD's social goal regulation of the enterprises represents one of a
number of federal government mechanisms that support housing
affordablity and homeownership.\31 Various federal agencies support
homeownership. For example, FHA and VA lower ownership costs by
guaranteeing mortgages with favorable terms for qualified
individuals. Ginnie Mae guarantees timely payment of principal and
interest from mortgage pools of FHA- and VA-insured mortgages. The
Federal Home Loan Bank System lends to mortgage lenders so they can
originate and fund mortgages. Federal financial institution
regulators also have responsibilities under the Community
Reinvestment Act to encourage banks and thrifts to help meet the
credit needs in all areas of their communities, including low- and
moderate-income areas. These regulators also enforce fair lending
laws that prohibit discriminatory lending practices.
--------------------
\31 Our discussion here is on federal programs rather than provisions
in the tax code.
WITH PRIVATIZATION,
ENTERPRISES ARE LIKELY TO
REDUCE OR ELIMINATE SPECIAL
PROGRAMS TO AID TARGETED
GROUPS
-------------------------------------------------------- Chapter 4:4.1
Privatization would likely provide the enterprises with new
incentives, including an altered cost structure and few if any
restrictions on their activities. As discussed in chapter 3, the
resulting secondary market entities would likely operate as conduits
rather than operate directly in the primary market or hold many
mortgages in portfolio. They would also not be likely to develop low
down payment mortgage products or purchase and securitize multifamily
mortgage products.\32
The enterprises' programs aimed at targeted groups, in general, are
more costly than their standard business. Fannie Mae officials told
us that most of their targeted lending products were more costly than
standard mortgage products. For example, in our review of the
enterprises' targeted lending programs, we found that default rates
were substantially higher on purchases through those programs. We
examined Fannie Mae mortgage default and borrower targeting
statistics comparing targeted lending programs and standard business
for mortgages purchased in 1994.\33
The difference in the default rates appears to result from the higher
loan-to-value ratios and the easing of other underwriting
restrictions in the targeted lending programs. This finding is
consistent with preliminary analysis at the Office of Federal Housing
Enterprise Oversight (OFHEO) indicating that enterprise funded loans
with loan to value ratios above 90 percent, in census tracts in which
incomes are below metropolitan area median income, and where more
than 30 percent of residents are minority group members default more
often than other mortgages purchased by the enterprises. As
designed, Fannie Mae's targeted lending programs purchase larger
shares of loans made to low-income, central city, minority, and
first-time homebuyer borrowers compared to standard business.
As a result, privatization is likely to reduce the significant
resources the enterprises are currently expending on these targeted
borrower programs. Because neither we nor the enterprises have been
able to quantify the impact of these efforts, however, it is
difficult to know whether privatization would have a significant
effect on affordability or homeownership opportunities among targeted
groups.
--------------------
\32 This finding is based on our interviews with enterprise officials
and our analysis of private-label conduits in the secondary market.
\33 Fannie Mae's National Housing Impact Division runs the
enterprise's targeted lending programs. Therefore, we use Housing
Impact initiatives and targeted lending programs synonymously.
Purchase money mortgages finance the purchase of a housing unit
(i.e., refinancings are excluded).
NET EFFECT ON SOCIAL GOAL
ATTAINMENT DEPENDS ON
WHETHER OTHER PROGRAMS ARE
MAINTAINED OR NEW PROGRAMS
ARE ESTABLISHED
-------------------------------------------------------- Chapter 4:4.2
The potential impacts of privatization on social goal attainment
depend, in part, on how well targeted borrowers would be able to
obtain financing from depository institutions and primary lenders who
originate FHA-insured loans. The FHA single-family mortgage
insurance program serves many lower income, minority, and central
city borrowers and these loans are securitized in Ginnie Mae MBS. It
is not clear how well these FHA programs serve or could serve
targeted borrowers compared with how the enterprises, without
privatization, would serve similar borrowers. Likewise, the FHA
multifamily insurance program is a possible policy alternative to
multifamily products now being developed by the enterprises.\34
However, the potential increased reliance on FHA and VA programs
resulting from privatization could increase the total risk of these
programs.
If privatization occurred and alternative policy levers could not be
developed for the ensuing secondary market participants, there are
other mechanisms available for achieving such goals. For example,
financial institution regulators could develop new Community
Reinvestment Act requirements that improve the incentives depository
institutions face for originating mortgages to targeted groups that
are sold in the secondary market.
Mortgage bankers are not subject to regulations such as the Community
Reinvestment Act (CRA). Some mortgage bankers have entered into
agreements with HUD concerning the distribution of their mortgage
origination activities. The current social goal regulations motivate
the enterprises to compete for loans originated by mortgage bankers
to designated borrowers. Privatization may sever this tie.
Therefore, if privatization occurred, it may be that some new
mechanism could be created to give mortgage bankers incentives to
originate mortgages to these targeted groups.
--------------------
\34 We did not make a separate examination of FHA operations in the
course of this mandated study, but we have relied upon publicly
available information and interviews with HUD officials and other
policy experts on FHA programs. In addition, we did not examine FHA
risk-sharing programs with the enterprises.
ENTERPRISE COMMENTS AND OUR
EVALUATION
---------------------------------------------------------- Chapter 4:5
In oral comments on a draft version of this chapter, a Fannie Mae
official said that Fannie Mae appreciates the report's recognition of
the commitment that the organization has made to affordable housing
and targeted financing overall. However, he said that the draft
report was inexplicably reluctant to draw unqualified conclusions
about the success of the enterprises' efforts in promoting
homeownership for targeted groups. Additionally, he said that
privatization would result in higher rental costs for occupants of
multifamily residences, and he said that Home Mortgage Disclosure Act
(HMDA) data has substantial limitations for assessing the
enterprises' efforts to promote homeownership among targeted groups.
The Fannie Mae official also said that the draft report neglected to
mention that increased reliance on other federal programs designed to
promote homeownership, such as FHA and VA, will increase the risks of
a taxpayer rescue. Fannie Mae officials also provided technical
comments, which we have incorporated where appropriate. Freddie Mac
officials said that privatization would result in higher rental costs
because owners of single-family rental housing would pass increased
mortgage rates on to their tenants.
The Fannie Mae official said that the draft report ignored
substantial evidence that the enterprise's commitment to the housing
goals has increased homeownership opportunities for targeted groups.
For example, he said that Fannie Mae provided tracking data for the
years 1993 to 1995 that clearly show the enterprise's overall share
of business serving low- and moderate-income groups has increased
consistently. He also said that there are quantifiable measures of
the success of Fannie Mae's efforts to make mortgage financing more
affordable for certain targeted groups; for instance, the use of
higher debt-to-income and loan-to-value ratios means that targeted
groups can more easily qualify for mortgages. Moreover, he said
there is no reason to expect that the social goals would be retained
in any form in the event of privatization. He noted that private
sector conduits that perform similar functions as the enterprises are
not subject to social goal requirements.
The Fannie Mae official also disputed the Wachter and Follain finding
that privatization would not have a significant effect on mortgage
interest rates for multifamily housing. He said that Fannie Mae's
commitment to this market predates the social goals, but its
extensive innovation and outreach efforts would likely be curtailed
in the event of privatization. He said that this would have genuine
effects on capital availability for affordable rental housing
development. In addition, the Fannie Mae official said that the
enterprise would probably respond to privatization by curtailing more
flexible credit tests and higher loan-to-value ratios which the
enterprise currently use to increase its participation in the market
for multifamily housing.
The Fannie Mae official further commented that HMDA data is not a
reliable basis for determining, as the draft report stated, that the
enterprises lagged other mortgage market participants in providing
credit to low- and moderate-income groups. For example, he said that
many such mortgages that the enterprises purchase are not credited by
the HMDA data. He attributed this shortcoming to the fact that only
the first mortgage sale is recorded by HMDA, and some mortgage
lenders are not covered and this can be a mortgage affiliate or other
player that eventually sells the mortgage to Fannie Mae or Freddie
Mac.
The Fannie Mae official also commented that the FHA and VA mortgage
programs and other options we listed could not possibly substitute
for the dollar volume commitments that the enterprises make each year
to purchase low- and moderate-income mortgages. Moreover, he said
that relying on these programs further does not necessarily represent
good public policy because it would shift potential loss liabilities
directly to the federal government and the taxpayers and the options
are not viable. He also said that it is highly speculative to assume
that Congress would enact CRA-type requirements for the enterprises
in the event of privatization.
Freddie Mac officials also said that potential taxpayer risks would
increase with privatization due to increased reliance on the FHA and
VA programs, as well as insured depository institutions. Further,
they stated that privatization would result in depositories
increasing their use of FHLB advances and generating additional
taxpayer risks. The Freddie Mac officials also said that renters
would likely face higher housing costs in the event of full
privatization because the owners of single-family rental properties
would pass increased mortgage costs on to their tenants.
We believe it is still too early to measure the impact of the
enterprises' social goals on the provision of additional housing
finance to targeted groups. For that reason, in the report we
presented information we obtained during this assignment on the
resource commitments the enterprises are making to fund mortgages
serving targeted borrowers. In addition to not being able to draw
unqualified conclusions about the effects of existing programs, it is
even more difficult to predict the effect of privatization largely
because we do not have enough information to predict (1) how
eliminating the enterprises' social goal obligations will interact
with other federal mechanisms, (2) what requirements HUD would have
set in the future without privatization, and (3) the market impact of
eliminating social goals on housing affordability. We do not have a
basis for knowing whether the limited coverage of HMDA biases
estimates of the enterprises' contributions to funding mortgages to
targeted borrowers. The increased reliance on FHA and VA programs
resulting from privatization could increase the total risk of these
programs, although it could also lower their average level of risk if
the enterprises' expanded efforts are taking away the more, rather
than the less, profitable business of these federal insurance
programs. We acknowledge that privatization could cause mortgage
interest rates on single-family rental housing, and thus rental costs
on such housing, to increase.
The report indicates that the enterprises' overall share of business
serving low- and moderate-income groups has increased consistently.
We do not know how much this has increased homeownership
opportunities for targeted groups, although the results from the
commissioned study by Wachter and Follain, as discussed, indicated
that privatization could reduce homeownership opportunities.
The best evidence we have available to assess the enterprises' impact
on mortgage interest rates for multifamily housing is from the
Wachter and Follain study, which concludes that privatization would
not radically alter the current situation.
We do not think it is clear how well other federal programs and other
mortgage providers could fill the void that could result from
privatization. The enterprises fund many mortgages, including those
serving targeted groups. With privatization, some of this activity
would be curtailed. The Federal Reserve Board study on credit risk
referred to in this report suggests that depository institutions may
be able to profitably serve some of these affected borrowers. We do
not know how much extra business FHA programs could be faced with if
the enterprises were privatized.
TRANSITION ISSUES AND ALTERNATIVE
POLICY OPTIONS
============================================================ Chapter 5
Privatization of the enterprises would clearly be a major policy
change. As such, it would require a careful examination of the
benefits and costs and involve difficult policy choices that only
Congress can make. Should Congress decide that privatization is
worth pursuing, there are a number of ways it could structure the
transition to privatization. Each of these has advantages and
disadvantages. For example, an approach designed to be least
disruptive to the mortgage market might leave institutions that were
still perceived as too big to fail. As a result, such an approach
might not fully break the government ties that cause the market to
perceive an implied guarantee. Alternatively, an approach that more
effectively broke those ties by breaking up the privatized
enterprises into smaller companies could reduce some of the potential
benefits from mortgage standardization and maintenance of liquidity
in the market.
Privatization is only one alternative to the status quo. There are
other policy options, short of privatization, that would adjust the
activities or responsibilities of the enterprises in such a way that
the potential public benefits generated by government sponsorship
could potentially increase or the size of enterprise activity or the
riskiness of that activity to the government could be decreased. The
latter could reduce the potential cost should the federal government
ever decide to bail out a failing enterprise. We selected
alternatives from among what appeared to be the most frequently
mentioned in the available literature while attempting to identify a
variety of approaches. The four alternatives we discuss include:
-- lowering or freezing the conforming loan limit,
-- increasing minimum requirements for mortgage insurance coverage,
-- charging the enterprises for the government's risk exposure, and
-- authorizing another government-sponsored enterprise to compete
with Fannie Mae and Freddie Mac.
TRANSITION WOULD INVOLVE
DIFFICULT CHOICES BUT WOULD
HOLD THE KEY TO MAKING
PRIVATIZATION A SUCCESS
---------------------------------------------------------- Chapter 5:1
Should Congress decide to privatize the enterprises, it would be
important to achieve a clear and deliberate elimination of the
special benefits and restrictions the enterprises have under their
current federal charters. To be successful, the legal transition to
privatization would need to be structured to eliminate investor
perceptions of an implied federal guarantee so that other private
companies could compete in the secondary mortgage market on a level
playing field. This perception is key to the fact that the
enterprises are the only two important competitors in the
conventional, conforming secondary mortgage market. Privatization
would be more likely to lead to more secondary market competitors if
the enterprises' special advantages were clearly removed.
SEVERAL POSSIBLE APPROACHES
TO RESTRUCTURING COULD BE
CONSIDERED
-------------------------------------------------------- Chapter 5:1.1
A transition to privatization would have to deal with a number of
trade-offs. First, the number of successful competitors would be
determined in part by the structure of the transition. If Congress
were to create more competitors initially, this could act to reduce
market liquidity and standardization. However, the number of
competitors that ultimately prevail in the secondary market would be
partly limited by market forces, including how much investors value
market liquidity. Second, the newly privatized enterprises must have
the managerial, capital, and other resources necessary for them to be
successful going concerns without preventing entry into the
conforming secondary mortgage market by potential competitors.
Engineering the restructuring necessary for the transition would
require extensive expertise by legal and financial experts. This
engineering would also involve trade-offs among competing objectives
and create policy challenges. Generally, the larger the new
enterprises are, the greater the risks that
-- investors would continue to perceive an implicit federal
guarantee, because the enterprises could be considered too big
to fail and there would be increased potential cost to taxpayers
if the enterprises were rescued by the government; and
-- the enterprises, because of their size and the possible
remaining perception of an implicit federal guarantee, would
exercise market power in business activities outside of the
secondary mortgage market for conventional, conforming
residential mortgages.
One approach would be to make each enterprise a holding company with
two subsidiaries--one subsidiary conducting liquidation of old (that
is, preprivatization) business and the other, conducting new
business. The proposed privatization of the Student Loan Marketing
Association (known as Sallie Mae) contains such a structure.\1
Segregation of securities created under government sponsorship and
new private entity securities would help sever the perceived implied
federal guarantee on post-privatization business, although it could
strengthen the tie on old business. If outstanding debt and MBS
previously issued by the enterprises as government-sponsored entities
were to be segregated, market stability and liquidity are less likely
to be jeopardized, because the liquidating subsidiary's securities
would be more likely to keep their current government-sponsored
status.\2
In addition to this option, the study commissioned by CBO assessed
other possible approaches to restructuring.\3 These included creating
-- two separate privatized companies that receive an allocation of
resources along with government actions to liquidate the
terminating government-sponsored enterprises; and
-- a number of separate privatized companies, which would break up
Fannie Mae and Freddie Mac into smaller operating companies,
followed by restructuring to remove government- sponsorship from
the successor companies.
The first of these options may be more likely than the "old company
new company" approach to prevent the perception of an implied federal
guarantee on new business, because all of the old obligations that
were thought to have the implied guarantee would be liquidated.
However, liquidating such a large amount of existing debt and MBS
could disrupt financial markets. The second option could be the most
conducive to insuring competition and to eliminating the "too big to
fail" perception, because there would presumably be a larger number
of smaller companies created out of the current enterprises.
However, forcing the new companies to be small could reduce
efficiencies associated with standardization and liquidity.
--------------------
\1 The current Sallie Mae proposal includes the notion that the
perception of the implied federal guarantee would remain for the
liquidating old company subsidiary. One possible policy option could
be the granting of an explicit federal guarantee for the liquidating
subsidiary's debt and MBS. A holding company structure with old and
new business subsidiaries would require strict regulation to limit
the capital and other resources from the liquidating old company
subsidiary from transferring to assist the new, privatized entity.
\2 The perception of an implied federal guarantee has affected how
the enterprises fund themselves. As we have noted, after
privatization the enterprises would likely change how they fund
mortgages to create financial instruments that do not require the GSE
form of credit enhancement.
\3 Thomas Stanton, �Restructuring Fannie Mae and Freddie Mac:
Framework and Policy Options � (October 1994).
AN ANALYSIS OF FOUR POLICY
ALTERNATIVES TO INCREASE
BENEFITS AND/OR REDUCE RISK TO
THE AMERICAN PUBLIC
---------------------------------------------------------- Chapter 5:2
To address adjusting the activities or responsibilities of the
enterprises to increase the public benefits and/or reduce the overall
size of enterprise debt or the probability that the government may
have to rescue a failing enterprise, we examined four policy options.
We identified a range of policy alternatives from our examination of
the policy literature. The following four alternatives we chose to
discuss involve trade-offs among competing policy interests, should
not be construed as our proposals, and by no means exhaust the
possible policy alternatives Congress may want to consider. The list
includes
(1) lowering or freezing the conforming loan limit,
(2)increasing minimum requirements for mortgage insurance coverage,\4
(3)charging the enterprises for the government's risk exposure, and
(4)authorizing another government-sponsored enterprise to compete
with Fannie Mae and Freddie Mac.
--------------------
\4 We define mortgage interest rates as including mortgage insurance
payments by the borrower.
LOWERING OR FREEZING THE
CONFORMING LOAN LIMIT
-------------------------------------------------------- Chapter 5:2.1
It appears that a lowering or freezing (i.e., not allowing
inflationary adjustments) of the conforming loan limit would have a
number of effects. First, it could reduce the amount of enterprise
activity without greatly limiting the ability of the enterprises to
diversify risk and thereby should reduce the potential taxpayer risk
in the event of a government bailout. This reduction could be offset
somewhat, because some of the activity that currently fits under the
conforming label but would not fit under the tighter ceiling may end
up in the portfolios of depositories rather than being securitized.
To the extent this occurs, there could be an increase in potential
taxpayer exposure. For example, depositories taking on more credit
risk could raise the risk exposure of the deposit insurance funds.
If the depositories are members of the Federal Home Loan Bank System
that receive additional advances, the potential taxpayer exposure of
this system could increase.
Second, mortgage interest rates for borrowers that would shift from
conforming to jumbo mortgage status would probably increase. There
is currently an interest rate spread between fixed-rate conforming
and jumbo mortgages. The study commissioned by HUD examining this
spread predicted that a 10-percent decline in the conforming loan
limit would likely lead to an increase in mortgage interest rates on
affected mortgages near the lower end of the 25 to 40 basis point
range.\5
Third, there could be a decline in mortgage interest rates for the
remaining jumbo market to the extent that private-label conduits
would choose to expand and become better able to geographically
diversify their funding.\6 The expected decline in mortgage interest
rates would still, however, probably leave jumbo rates above those on
conforming mortgages.
--------------------
\5 Cotterman and Pearce, op. cit., pages 63-64. The authors state
that the expansion in the private-label market might increase the
liquidity of private-label securities, which would put downward
pressure on interest rates in the jumbo market.
\6 This effect contributes to our expectation that potential
contingent liabilities would decline; we would expect more of the
shift in funding activity to accrue to the now more competitive
private-label conduits.
INCREASING MINIMUM
REQUIREMENTS FOR MORTGAGE
INSURANCE COVERAGE
-------------------------------------------------------- Chapter 5:2.2
The enterprises are not allowed to purchase mortgages with
loan-to-value ratios above 80 percent unless the borrower obtains
mortgage insurance. In 1995, the enterprises changed their
underwriting guidelines and now require greater insurance coverage on
mortgages with loan-to-value ratios exceeding 85 percent. If
Congress legislated higher requirements for mortgage insurance
coverage, the enterprises would be exposed to less credit risk.
Simply put, when mortgage defaults occurred, more of the burden would
fall on private mortgage insurers that have no federal ties and less
would fall on the enterprises. The reduced risk taken on by the
enterprises would reduce the likelihood that the enterprises would
need to be bailed out, and the potential risk to the taxpayer would
be reduced as well.\7
Mortgage interest rates would likely increase. If for no other
reason, the capital costs of private mortgage insurers tend to be
higher than the enterprises' costs because private insurers have no
federal ties. Mortgage interest rates would likely increase more for
borrowers making downpayments below any legislated minimum, because
private mortgage insurers charge fees that are more fully risk-based
than the guarantee fees charged by the enterprises.
--------------------
\7 As stated earlier, we define mortgage interest rates as inclusive
of mortgage insurance payments.
ENTERPRISE PAYMENT FOR
TAXPAYER RISK EXPOSURE
-------------------------------------------------------- Chapter 5:2.3
An alternative type of policy approach would charge the enterprises
to compensate, in whole or in part, for the risk exposure that their
activities generate for the government and taxpayers. One such
alternative is a fee, sometimes referred to as a user fee, that could
provide a full or partial offset for the estimated benefits received
from government sponsorship. Levying user fees on the value of
enterprise debt and MBS issuance could be thought to compensate
taxpayers for the possibility that they might be asked someday to
come to the rescue of a failing enterprise.\8 User fees could be
passed onto borrowers in part or in whole and result in higher
interest costs. The net effects would depend on the level of user
fees. User fees on the enterprises could help level the playing
field between the enterprises and private-label conduits and motivate
these conduits to securitize conforming mortgages, because the cost
of funds differential would be reduced.
CBO analyzed the federal revenue consequences of user fees on
enterprise debt and MBS.\9 CBO's revenue projection was based on
estimates indicating that the enterprises probably save more than 30
basis points on their debt and more than 5 basis points on their MBS.
Annual revenue from a user fee equal to half of the dollar amount of
estimated funding cost savings was estimated to be about $700
million. The passing on of part or all of this payment to borrowers
would raise mortgage interest rates.
Determining the correct level of such a fee would be difficult
because of problems associated with measuring the value of the
funding cost savings resulting from investors' perception of an
implied guarantee. Another difficulty is determining the possible
interaction between a user fee and regulatory capital charges.
OFHEO told us that user fees that are set through legislation are a
fairly blunt instrument, while the risk-based capital requirements
that OFHEO is developing could be flexible over time. Both user fees
and capital requirements increase the cost of capital to the
enterprises, which can, in turn, pass on to borrowers some or all of
these costs in the form of higher guarantee fees and interest rates.
If Congress legislated user fees, OFHEO's ability to set capital
charges to manage enterprise risk taking could be affected, because
both actions would increase enterprise costs and could contribute to
higher mortgage interest rates. In other words, user fees and
capital requirements must be viewed in conjunction with one another
to determine cost impacts on the enterprises and residential mortgage
borrowers.
A somewhat different approach to compensating the government for its
risk exposure would attempt to make that exposure more symmetrical
than it is currently. If the government felt the need to rescue a
failing enterprise, clearly it would face "downside" risk. However,
when the result of enterprise risk-taking is additional income, the
government shares only to the same extent it shares with any private
company, that is through increased corporate income tax revenue. One
way to make the government's payoff more symmetrical would be for the
government to receive a greater share of income in good times to make
up for the possibility it will have to come to the rescue if the
enterprises face bad times. The effects of such a payment would
depend on how it was structured. For example, if it were simply a
surtax on corporate income, it could end up being passed on to
borrowers in the mortgage market or be passed back to shareholders.
It could also raise the relative cost of equity capital compared to
debt capital and further reduce the incentives of the enterprises to
hold equity in the absence of safety and soundness regulation.
--------------------
\8 The enterprises view a user fee as a tax on homebuyers rather than
as an offset for the value of benefits provided by the government to
the enterprises. CBO states that a user fee has characteristics of
both a user fee and a tax, and the ambiguity makes it unclear whether
the proceeds should be shown on the revenue side of the budget as
governmental receipts or on the spending side as offsetting
collections. On balance, CBO thinks that the charge seems closer to
a fee for services than a tax.
\9 U.S. Congressional Budget Office, Reducing the Deficit: Spending
and Revenue Options (February 1995) pp. 318-319. CBO defined the
user fee as a cost-of-capital offset fee on Fannie Mae and Freddie
Mac.
AUTHORIZING ANOTHER
GOVERNMENT-SPONSORED
ENTERPRISE TO COMPETE WITH
FANNIE MAE AND FREDDIE MAC
-------------------------------------------------------- Chapter 5:2.4
Privatization would in essence eliminate the enterprises as
government-sponsored entities. The three preceding alternatives to
privatization would either decrease the government's risk exposure
from enterprise activities or compensate the government in whole or
in part for that exposure. A fourth alternative would attempt to
increase the public benefits from enterprise activity by lowering
mortgage rates through increased competition among enterprises. This
alternative would entail authorizing another government-sponsored
enterprise with a similar charter and subject to the same regulatory
requirements to compete with Fannie Mae and Freddie Mac. This could
increase the overall size of enterprise activity in the mortgage
market and, as a result, raise the potential at risk in case of a
government bailout. It could also increase the level of enterprise
risk because entities operating in new markets often have greater
managerial and operations risk than those operating in established
markets. In addition, there could be increased credit risk if the
new entity attempted to establish market share by lowering
underwriting standards. Any other potential effects of a third
competing enterprise would depend on whether Fannie Mae and Freddie
Mac do or do not have market power. If they do not, there is little
in the way of efficiency gain to expect from a new competitive force
in the market. However, to the extent there is market power, a third
competing enterprise could put pressure on the existing enterprises
to lower mortgage rates. In addition, because increased competition
could motivate fuller use of risk-based guarantee fees, it could
reduce the ability of the enterprises to achieve social goals to the
extent attainment requires charging targeted groups less than fully
risk-based fees.
HUD could still set performance measures to attain social goals with
increased competition.\10 The possible decline in profit levels and
increased use of fully risk-based guarantee fees, however, could
lessen (1) HUD's ability to set demanding performance measures to
attain social goals and (2) the ability of the enterprises to
unilaterally cross-subsidize funding activities to help achieve their
missions.
--------------------
\10 As an analogy, federally insured depository institutions operate
in competitive markets and are subject to regulations that they, in
part, respond to by creating cross-subsidies across their business
lines.
MAJOR CONTRIBUTORS TO THIS REPORT
=========================================================== Appendix I
GENERAL GOVERNMENT DIVISION
--------------------------------------------------------- Appendix I:1
Thomas J. McCool, Associate Director
William B. Shear, Assistant Director
Mitchell Rachlis, Senior Economist
Wesley Phillips, Evaluator
OFFICE OF GENERAL COUNSEL
--------------------------------------------------------- Appendix I:2
Paul Thompsen, Attorney
*** End of document. ***