Foreign Housing Guaranty Program: Financial Condition Is Poor and Goals
Are Not Achieved (Chapter Report, 06/02/95, GAO/NSIAD-95-108).
Since 1961, the Agency for International Development's Housing Guaranty
Program has guarantied more than $2.7 billion in loans in 44 countries
for home construction, mortgages, home improvements, urban
infrastructure, and other shelter projects. A fundamental program goal
is to increase housing for low-income families in developing countries
by motivating local institutions to provide investment capital and other
resources. However, Congress should consider terminating the program
because it has failed to spur private-sector investment in low-income
housing in developing countries, its benefits often go to higher-income
persons, and its loan defaults may ultimately cost the U.S. government
as much as $1 billion. Moreover, program assistance has gone
increasingly to creditworthy developing nations that have ready access
to international financing. GAO summarized this report in testimony
before Congress; see: Foreign Housing Guaranty Program: Goals Are Not
Achieved and Financial Condition Is Poor, by Frank C. Conahan, Senior
Defense and International Affairs Advisor to the Comptroller General,
before the Subcommittee on International Economic Policy and Trade,
House Committee on International Relations. GAO/T-NSIAD-95-181, June 28,
1995 (11 pages).
--------------------------- Indexing Terms -----------------------------
REPORTNUM: NSIAD-95-108
TITLE: Foreign Housing Guaranty Program: Financial Condition Is
Poor and Goals Are Not Achieved
DATE: 06/02/95
SUBJECT: Technical assistance
Housing programs
Loan defaults
Foreign loans
Investments abroad
International economic relations
Foreign economic assistance
Foreign economic development credit
Developing countries
Foreign aid programs
IDENTIFIER: AID Housing Guaranty Program
**************************************************************************
* 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. *
* *
* A printed copy of this report may be obtained from the GAO Document *
* Distribution Facility by calling (202) 512-6000, by faxing your *
* request to (301) 258-4066, or by writing to P.O. Box 6015, *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time. *
**************************************************************************
Cover
================================================================ COVER
Report to Congressional Requesters
June 1995
FOREIGN HOUSING GUARANTY PROGRAM -
FINANCIAL CONDITION IS POOR AND
GOALS ARE NOT ACHIEVED
GAO/NSIAD-95-108
Foreign Housing Guaranty Program
(711059)
Abbreviations
=============================================================== ABBREV
OMB - Office of Management and Budget
RHUDO - Regional Housing and Urban Development Office
USAID - U.S. Agency for International Development
Letter
=============================================================== LETTER
B-258693
June 2, 1995
The Honorable Toby Roth
Chairman
The Honorable Sam Gejdenson
Ranking Minority Member
Subcommittee on International
Economic Policy and Trade
The Honorable Doug Bereuter
Chairman, Subcommittee on Asia and
the Pacific
Committee on International Relations
House of Representatives
As you requested, we reviewed the U.S. Agency for International
Development's (USAID) use of Housing Guaranty Program funds.
Specifically, our review focused on the program's (1) evolution, (2)
financial condition, and (3) impact. This report suggests that the
Congress may wish to terminate the program. If the Congress does not
terminate the program, the report also contains recommendations to
the USAID Administrator intended to strengthen the financial
condition of the program and improve its management.
Unless you publicly announce its contents earlier, we plan no further
distribution of this report until 15 days from its issue date. At
that time, we will send copies to other appropriate congressional
committees, the Administrator of the U.S. Agency for International
Development, and the Director of the Office of Management and Budget.
Copies will also be made available to others upon request.
This report was prepared under the direction of Joseph E. Kelley,
Director-in-Charge, International Affairs Issues. If you have any
questions concerning this report, please call him at (202) 512-4128.
Other major contributors to this report are listed in appendix III.
Henry L. Hinton, Jr.
Assistant Comptroller General
EXECUTIVE SUMMARY
============================================================ Chapter 0
PURPOSE
---------------------------------------------------------- Chapter 0:1
The Foreign Assistance Act of 1961, as amended (22 U.S.C. 2151 et.
seq.), authorizes the U.S. Agency for International Development
(USAID) to guaranty loans made by U.S. investors to borrowers in
developing countries for shelter-related projects. With this
authority, USAID operates the Housing Guaranty Program. A long-run
goal of this program is to increase shelter for low-income families
in developing countries by stimulating local credit institutions to
provide the necessary investment capital and other resources. Since
1961, USAID has guarantied over $2.7 billion in loans (valued at $5
billion in 1995 dollars) in 44 countries for home construction,
mortgages, home improvements, urban infrastructure, and other
shelter-related projects.
Members of the House Committee on International Relations have become
concerned about the use of Housing Guaranty Program funds and the
program's true cost to the U.S. government. At their request, GAO
reviewed the program's (1) evolution, (2) financial condition, and
(3) impact.
BACKGROUND
---------------------------------------------------------- Chapter 0:2
USAID's Office of Environment and Urban Programs is responsible for
managing the Housing Guaranty Program and related technical
assistance programs, while USAID's regional bureaus are responsible
for authorizing them. Overseas, the programs are implemented
primarily by USAID's Regional Housing and Urban Development Offices.
RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3
Over the past 34 years, the focus and geographic scope of the Housing
Guaranty Program have changed significantly. Program focus has
expanded beyond home construction and mortgage financing to include
urban infrastructure financing and governmental policy reform. Due
to changes in international economic conditions and U.S. foreign and
budget policies, more creditworthy and advanced developing countries
are now the most likely recipients of loan guaranties. In addition
to loan guaranties, USAID relies on technical assistance to implement
the program.
The Housing Guaranty Program is in serious financial condition
because program fees have not been sufficient to cover the cost of
this program. In 1995 dollars,\1 program losses due to loan defaults
have already cost the U.S. government over $540 million,\2 and GAO
estimates that the cost of future defaults is likely to be an
additional $600 million. Although USAID program officials believe
that almost all defaulted borrowers will ultimately repay their
debts, GAO estimates that USAID is likely to recover only about $200
million. As a result, the total U.S. government cost associated
with the program's existing portfolio is likely to be about
$1 billion.\3 This amount does not include several hundred million
dollars in technical assistance that USAID has provided to implement
the program but has not reported as a Housing Guaranty Program cost.
The legislated ceiling on guaranties\4 did not effectively limit the
U.S. government's overall investment in the program because
defaulted loans, once repaid by USAID, are not counted against the
ceiling.
Although the Housing Guaranty Program has contributed to shelter
sector reforms in many participating countries, it has not stimulated
increased private investment in low-income shelter, one of the
program's long-run objectives. Further, in nearly every country
visited for this review, GAO observed program-financed shelter
projects that were outside the reach of the poorer families that the
program is supposed to target. USAID documents also reflected this
problem. For example, in India a USAID study indicated that 36
percent of program-financed home loans went to families with
above-median incomes, and in Tunisia one sampling showed that 17
percent of eligible beneficiaries had incomes above the median.
USAID's performance indicators do not adequately measure progress in
achieving these objectives.
Since fiscal year 1992, the Congress has authorized new loan
guaranties under this program and, as required by the Credit Reform
Act, appropriated about $50 million to cover probable loan default
costs associated with them. However, the Congress has done so
without full knowledge of the U.S. government's liabilities from
earlier guaranties and the other problems GAO has outlined in this
report.
--------------------
\1 Unless otherwise noted, all figures are presented in 1995 dollars.
\2 $383 million in then-year dollars.
\3 GAO's cost projections are rough and may vary over time depending
upon market expectations for USAID's guaranty portfolio.
\4 This $2.6 billion ceiling applies only to loan guaranties
authorized before fiscal year 1992.
PRINCIPAL FINDINGS
---------------------------------------------------------- Chapter 0:4
HOUSING GUARANTY PROGRAM HAS
CHANGED SIGNIFICANTLY
-------------------------------------------------------- Chapter 0:4.1
Since the Housing Guaranty Program was introduced in the early 1960s,
it has changed significantly in terms of the (1) profile of
participating countries, (2) type of projects funded, and (3)
implementation strategy. Historically, assistance under this program
went primarily to countries in Latin America. Foreign policy
considerations, U.S. credit reform, and the 1980s debt crisis have
led to a program shift toward more creditworthy and advanced
countries in other regions of the world. Some program activities,
such as those in Israel and Portugal, were mandated by the Congress
for specific foreign policy reasons. Due to the Federal Credit
Reform Act of 1990, which requires appropriated funds to be set aside
to cover probable loan defaults, USAID has been unable to maintain
its historical volume of loan guaranties to higher risk borrowers
with its existing budget authority. Recent programs have focused on
countries such as Indonesia, India, Tunisia, and Thailand, which have
relatively favorable credit ratings and, therefore, require smaller
reserves for probable default.
With respect to the types of projects this program has funded, the
program's loans were initially limited to financing housing built by
U.S. contractors and local firms. Now, program loans are used to
finance a wide variety of shelter-related projects that affect the
urban environment in developing countries. Such projects have
included urban renewal projects, sewage systems, water treatment
facilities, and other infrastructure projects.
USAID's initial implementation strategy was to demonstrate the
feasibility of building low-cost housing projects that could be
replicated by local entrepreneurs and institutions. However, more
recently, the goal has been policy reform in the recipient country.
USAID has offered loan guaranties as incentives for governments to
adopt more effective laws and policies on shelter and urban
development. USAID also provides technical assistance as part of its
efforts to achieve desired policy reform and institutional
development objectives. The cost of this closely-linked assistance,
totaling $471 million\5 for USAID's ongoing projects, has been funded
through USAID's regional bureaus.
--------------------
\5 Amount is not adjusted for inflation.
HOUSING GUARANTY PROGRAM
LOSES MILLIONS ANNUALLY
-------------------------------------------------------- Chapter 0:4.2
Although the Congress specifically authorized USAID to charge fees to
cover the cost of its Housing Guaranty Program, the program has been
losing millions of dollars annually. Fee revenues have been adequate
to cover administrative costs but not loan default costs. As a
result, the total net cost of the program to the U.S. government is
likely to be about
$1 billion. Based on audited financial statements and other
financial information, GAO determined that since 1970 the program has
received $417 million\6 in appropriations and has borrowed $125
million from the U.S. Treasury to finance losses due to loan
defaults. The cost to the U.S. government of future loan defaults
on USAID's portfolio is likely to be an additional $600 million.
USAID has only enough reserves to cover about $50 million in defaults
for loans authorized since fiscal year 1992.\7
Although program officials expect to recover nearly the full amount
of all loan defaults, GAO estimates that USAID will ultimately
recover only about $200 million.
In addition to these costs, USAID has committed several hundred
million dollars in technical assistance to program participants. The
cost of this assistance, although not reflected in the program's
budget or financial statements, should be considered a cost of the
program.
To date, USAID has been unable to collect much of the debts owed by
defaulted borrowers and has continued to extend loan guaranties to
countries in spite of repeated debt rescheduling. According to the
program's financial statements, uncollected debt has more than
doubled from $171 million in 1989 to $409 million in 1994.\8
Deferring repayment through multiple rescheduling of uncollected
debts is common, and some debts have been rescheduled up to 10 times.
By the end of fiscal year 1994, USAID had recovered only 5 percent of
all rescheduled debt.
The $2.6 billion legislated ceiling on pre-fiscal year 1992 loan
guaranties did not effectively limit the cost of this program to the
U.S. government. The ceiling limited only the amount of guaranties
outstanding at any one time. Thus, if a borrower defaulted on a
guarantied loan and USAID made a principal payment to the lender, the
outstanding guaranty was reduced. This allowed USAID to issue new
guaranties under the ceiling, even though the borrower had not repaid
(and may never repay) USAID for its default costs. Thus, USAID
continued guarantying loans while over $400 million in uncollected
debt accumulated, and the total amount of the U.S. government
resources committed grew beyond the $2.6 billion ceiling. As of
March 1995, USAID planned to guaranty $193 million more in loans that
were authorized under this ceiling and are likely to result in over
$50 million in default costs.
--------------------
\6 $258 million in then-year dollars.
\7 Under the Credit Reform Act, USAID and the Congress must budget
up-front for the estimated cost of probable defaults for guarantied
loans authorized since fiscal year 1992. Therefore, the Congress has
appropriated funds for the probable defaults of new loan guaranties
under this program.
\8 This total does not include debts that USAID has written off as
uncollectible. Amounts are not adjusted for inflation.
IMPACT ON LOW-INCOME SHELTER
INVESTMENT IS NOT EVIDENT
-------------------------------------------------------- Chapter 0:4.3
Although the Housing Guaranty Program has contributed to shelter
sector reform in developing countries, it has not achieved one of its
fundamental, long-run objectives. The program has not led to an
increase in the availability of domestic capital to finance shelter
projects benefiting poor families, particularly from private sources.
The Foreign Assistance Act states that the program should demonstrate
to local entrepreneurs and institutions the feasibility of investing
in low-income shelter projects, thereby stimulating them to provide
increased investment capital and other resources. USAID-sponsored
projects have contributed to increased private managerial and
technical involvement in the construction of state-financed shelter
projects. However, USAID documents indicate that private credit
institutions have not invested their own capital in similar projects
as a result of this program. Instead, financing has come from the
limited government resources of these countries.
In addition, the Foreign Assistance Act requires that at least 90
percent of the guaranties be issued for housing suitable for families
with below-median incomes. However, USAID does not routinely ensure
that the projects it finances actually benefit such families. In
many cases, USAID relies on unvalidated reports from the borrowers to
monitor the use of loan funds. In the countries visited for this
review, GAO readily found instances where housing financed by the
program did not appear to be benefiting low-income families. For
example, GAO's review of the program in India revealed that some home
improvement loans went to the above-median income employees of a
participating bank. In some cases, USAID documents GAO reviewed
confirmed that a significant portion of eligible program
beneficiaries probably had above-median incomes. In other cases in
the countries GAO visited, USAID studies acknowledged the potential
for this problem.
USAID has not designed its performance indicators to measure the
Housing Guaranty Program's progress in achieving the goals of
stimulating private investments and providing shelter to below-median
income families. Instead, USAID generally measures a variety of
large-scale shelter sector changes that are not directly attributable
to the program.
MATTERS FOR CONGRESSIONAL
CONSIDERATION
---------------------------------------------------------- Chapter 0:5
The Congress may wish to deauthorize guaranties for undisbursed
loans, where feasible, and terminate the Housing Guaranty Program
because (1) the program is now primarily benefiting borrowers in more
creditworthy and advanced developing countries that have access to
comparable loans from other international lenders; (2) the program
annually costs millions of dollars more than anticipated; (3) loans
previously guarantied under the pre-1992 legislated ceiling continue
to be disbursed, even though USAID has not collected over $400
million in defaulted debt; and (4) there is no evidence that the
program has measurably increased the availability of private domestic
capital for low-income shelter.
Other actions are warranted if the Congress believes that the goals
of (1) stimulating domestic investment in the recipient countries
(particularly private investment) and (2) targeting shelter
assistance to below-median income families are possible to achieve
and are of sufficient priority. Before authorizing additional loan
guaranties, the Congress may wish to require USAID to submit a
comprehensive plan to the appropriate congressional committees on how
it plans to achieve the stated program goals, reduce losses, and
return the program to a viable financial condition.
RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:6
If the Congress does not terminate the Housing Guaranty Program, GAO
recommends that the USAID Administrator take the following actions to
(1) minimize the financial impact of the Housing Guaranty Program on
the foreign assistance budget and the U.S. government budget deficit
and (2) bring the program in line with the objectives of the Foreign
Assistance Act:
Withhold future loan disbursements and related technical assistance
from borrowers that have repeatedly rescheduled debt repayments
to USAID.
Increase program revenues by adopting a fee structure designed to
offset a larger portion of the program's costs.
Ensure that performance indicators measure the extent to which (1)
local investors replicate the program's low-income shelter
projects using private sources of long-term financing and (2)
project benefits accrue to the below-median income target
population.
AGENCY COMMENTS AND GAO'S
EVALUATION
---------------------------------------------------------- Chapter 0:7
In commenting on a draft of this report, USAID strongly disagreed
with GAO's findings, arguing that the Housing Guaranty Program was a
cost-effective method for pursuing development assistance and that
the program had achieved the goals established by the Congress.
USAID criticized the report for failing to adequately consider
alternative program goals and congressional directions. USAID also
asserted that the report was based on a flawed financial analysis and
a distorted portrayal of program costs because of the use of 1995
dollars. Furthermore, USAID maintained that GAO did not adequately
consider evidence that the program had successfully stimulated
private investment and reached below-median income target
populations. USAID suggested that termination of this program and
deauthorization of guaranties might interfere with legally binding
agreements and congressionally supported activities. USAID did not
address the report's specific recommendations, since it contended
that they were based upon incorrect information and a flawed
analysis.
As a result of these comments, GAO modified its matters for
congressional consideration and recommendations to (1) recognize that
it may not be feasible to deauthorize all loan guaranties and (2)
indicate that USAID should increase its guaranty fees to cover more
of its costs. However, GAO's evaluation of these comments revealed
that USAID had not credibly supported any of its broad and strongly
worded criticisms. USAID did not identify any specific legislated
program goals that GAO overlooked or any legal provisions that
contradict the goals GAO focused on. USAID's criticism of GAO's
financial analysis is unfounded and even disregarded its own most
recent audited financial statements regarding probable losses from
loan defaults. GAO's use of 1995 dollars reflects the true cost of
the program more accurately than the use of then-year dollars,
especially for costs incurred over many years. GAO denominated
figures in both 1995 and then-year dollars, except in cases, such as
the discussion of legislated ceilings, where it was not meaningful or
possible to use 1995 figures. Furthermore, none of the additional
data USAID provided (most of which GAO had already reviewed)
undermined GAO's conclusions.
INTRODUCTION
============================================================ Chapter 1
The Foreign Assistance Act of 1961, as amended, provides the
framework for the Housing Guaranty Program (originally called the
Housing Investment Guaranty Program). Since 1961, the Agency for
International Development's (USAID) Office of Environment and Urban
Programs, which is responsible for administering the program, has
guarantied $2.7 billion in loans (valued at $5 billion in 1995
dollars) to 44 developing countries. The projects funded by the
loans were intended to develop the countries' construction
capabilities and encourage the countries' credit institutions to make
domestic capital and other resources available for low-cost shelter
programs.
LEGAL FRAMEWORK
---------------------------------------------------------- Chapter 1:1
The U.S. government has provided a 100-percent guaranty on long-term
loans from U.S. lenders to borrowers in developing countries to
finance housing and shelter-related projects. If a borrower defaults
on a loan payment, the U.S. government pays any principal, interest,
and fees due to the lender. USAID then attempts to collect these
funds from the defaulted borrower.
The Foreign Assistance Act places certain restrictions on the amount
and use of these guaranties. For example, the face value of
guaranties issued may not exceed $25 million per fiscal year in any
country, and the average face value of all guaranties issued in a
fiscal year may not exceed $15 million. Also, at least 90 percent of
the loan funds must be used for projects suitable for the
below-median income population. The Foreign Assistance Act also
includes provisions regarding allowable interest rates, assessment of
fees, coordination with other U.S. foreign assistance, and other
administrative matters.
LOAN GUARANTY PROCESS
---------------------------------------------------------- Chapter 1:2
USAID works in cooperation with host country officials to determine
which activities will be financed through the Housing Guaranty
Program. Borrowers have included government ministries, central
banks, national housing banks, housing development corporations, and
private financial institutions, such as cooperatives or savings and
loan associations. U.S. lenders have included banks, insurance
companies, and other financial institutions and investors. Because
of the U.S. government guaranty, lenders bear virtually no
risk--similar to investing in U.S. Treasury bonds. Lenders base
loan interest rates, which can be either fixed or variable, on
prevailing commercial rates plus a premium. The borrower and the
lender negotiate the terms of the financing, subject to approval by
USAID. Typically, Housing Guaranty Program loans are extended for 30
years, with a 10-year grace period on the repayment of the loan
principal. Loans to governmental borrowers have a host country
guaranty of repayment to USAID. Loans to private borrowers are
backed by equivalent collateral but are not always guarantied by the
host government. To pay for program operations, USAID charges
borrowers an initial disbursement fee of 1 percent of the loan
principal, an annual fee of 0.5 percent of the outstanding principal,
and a fee for each day that a payment is late.
ORGANIZATIONS WITH
RESPONSIBILITY FOR SHELTER
PROGRAMS
---------------------------------------------------------- Chapter 1:3
USAID's Office of Environment and Urban Programs is responsible for
administering the Housing Guaranty Program, including managing the
loan portfolio and associated technical assistance. USAID's regional
bureaus are responsible for authorizing both the loans and the
technical assistance programs. Overseas, individual programs are
implemented by USAID's seven Regional Housing and Urban Development
Offices (RHUDO), located throughout the developing world with
assistance, when necessary, from USAID's Regional Economic
Development Support Offices and USAID missions. Figure 1.1 shows the
locations of the major offices that administer the Housing Guaranty
Program worldwide.
Figure 1.1: Locations of
Housing Guaranty Program
Offices Worldwide
(See figure in printed
edition.)
Source: USAID.
The Office of Environment and Urban Programs and the RHUDOs employ
about 60 direct-hire staff members and 70 contractors to implement
this program.
OBJECTIVES, SCOPE, AND
METHODOLOGY
---------------------------------------------------------- Chapter 1:4
At the request of Representatives Toby Roth, Doug Bereuter, and Sam
Gejdenson, members of the House Committee on International Relations,
we conducted a comprehensive review of the Housing Guaranty Program's
(1) evolution, (2) financial condition, and (3) impact. We performed
our work at USAID offices in Washington, D.C., Chile, Ecuador, India,
Indonesia, Morocco, Poland, and Tunisia. Programs in these countries
reflected a wide variety of shelter-related projects and program
objectives. We interviewed officials in USAID's Office of
Environment and Urban Programs of the Bureau for Global Programs,
Field Support, and Research; USAID Regional Bureaus; and other
organizations with responsibilities related to the Housing Guaranty
Program. We also met with representatives of private voluntary
organizations and other USAID contractors involved in USAID's shelter
programs in the United States and overseas. We also interviewed
representatives of three multilateral organizations-- the World Bank,
the United Nations Development Program, and the United Nations Center
for Human Settlement (commonly known as Habitat). We reviewed the
legislative history of the Foreign Assistance Act and examined USAID
records, including program documentation and evaluations.
To evaluate the program's financial condition, we reviewed the
program's financial records, including audited financial statements
for fiscal years 1989 through 1993, examined budget documents for
fiscal years 1961 through 1994, and reviewed prior reports on the
Housing Guaranty Program prepared by USAID's Office of the Inspector
General and our office. We also interviewed officials of the USAID
Office of Inspector General; Riggs National Bank in Washington, D.C.
(which acts as the program's transfer agent); the accounting firm
that conducted the most recent financial audit of the Housing
Guaranty Program; and the Office of Management and Budget (OMB).
To estimate the U.S. government's probable costs due to nonpayment
of USAID's existing loan guaranty portfolio, we applied an updated
version of our method for estimating the cost of country risk
associated with loan guaranties outstanding.\1 We developed our
country risk ratings by statistically averaging two contemporaneous,
professionally-accepted ratings--Euromoney and the Institutional
Investor--and then transformed those ratings into a country risk cost
using econometric estimates based on data from the secondary market.
Cost estimates for USAID's loan guaranty portfolio using this method
(as well as comparable methods) will change over time depending upon
market expectations for this particular group of guaranties. Our
method involves three issues that could qualify our cost estimates.
Those issues are whether (1) our sample size of debt traded on the
emerging market was large enough to capture the market's real
behavior; (2) the price relationship we estimated for these traded
instruments could be extended to proxy the prices of similar debt
instruments held by the private sector but not traded; and (3) debt
owed the private sector--from which our country risk cost estimates
were derived--is as likely to be repaid in the long run as sovereign
debt owed the U.S. government.
USAID's records did not allow us to categorize the use of guarantied
loan funds by project type. Also, due to the unique nature of the
Housing Guaranty Program in Israel, we specifically excluded it from
the scope of our assessment of USAID's program management.
We performed our work from January 1994 through February 1995 in
accordance with generally accepted government auditing standards.
--------------------
\1 This method is presented in detail in Credit Reform: U.S. Needs
Better Method for Estimating Cost of Foreign Loans and Guarantees
(GAO/NSIAD/GGD-95-31, Dec. 19, 1994).
HOUSING GUARANTY PROGRAM HAS
CHANGED SIGNIFICANTLY
============================================================ Chapter 2
The Housing Guaranty Program has changed considerably in scope and
content since its inception in 1961. Though initially concentrated
in Latin American countries, the program now operates in developing
countries in Africa, Asia, and the Middle East and increasingly in
more creditworthy and advanced developing countries. Similarly, the
program's focus has expanded from basic housing construction to a
variety of projects that address the broader needs of the urban
environment. The program's strategy has shifted from simply
demonstrating the feasibility of constructing low-income housing to
reforming shelter sector policy in borrowing countries. In addition,
the provision of technical assistance has become a fundamental part
of the program.
ADVANCED DEVELOPING COUNTRIES
ARE LIKELY LOAN RECIPIENTS
---------------------------------------------------------- Chapter 2:1
The debt crisis of the 1980s, U.S. credit reform, and foreign policy
considerations have influenced USAID to change the type of countries
likely to participate in the Housing Guaranty Program. The result
has been an increase in guarantied loans authorized for more
creditworthy and advanced developing countries and for countries
specified by the Congress.
Due in part to the worldwide recession of the mid-1980s and the
resulting debt crisis, many borrowing countries in Latin America and
Africa began to default on payments on their large portfolio of
international debt and became high credit risks.
In 1990, the Congress passed the Credit Reform Act, which required
each government department or agency to recognize its potential
financial liability for each loan authorized and disbursed after
fiscal year 1991. Prior to the act, USAID, like other agencies,
treated a guaranty as having no cost unless and until the borrower
defaulted. Under the Credit Reform Act, USAID must set aside a
reserve to cover the estimated future cost of loans guarantied,
including default costs. The amount of the reserve is based on an
OMB-approved assessment of the borrower's creditworthiness, which
reflects the probability that the borrower will default: the greater
the credit risk, the higher the subsidy required. Disbursement of
loans is restricted by the extent to which appropriations are
available for the reserve.\1 In response, USAID has ceased to
authorize new loan guaranties to many of the less creditworthy
countries of Latin America and Africa that had historically benefited
from the program. Although the Credit Reform Act has changed the
financing for this program and influenced USAID's selection of
borrowers, the program has continued to operate under the same
authorizing legislation and guidelines.
To cope with the budgetary constraints imposed by this legislation,
USAID has begun to concentrate its new loan programs in more
creditworthy countries, for which smaller reserves are required.
Since Credit Reform measures were enacted, USAID has authorized new
guaranties for countries such as India, Indonesia, Tunisia, and
Thailand, which have relatively favorable credit ratings. Because of
their relatively low credit risk, such countries have ready access to
capital from international credit markets.
Foreign policy considerations have also caused USAID to devote a
large portion of its guarantied loans to several more developed
countries. For example, a program in Portugal was mandated by the
Congress in the 1980s as part of an economic support package in
exchange for maintaining U.S. military base rights in the Azores.
In 1990, $400 million in guaranties were authorized for Israel in
response to its request for help in coping with the housing needs of
Soviet immigrants. More recently, loan guaranties for Poland, the
Czech Republic, and Hungary have been authorized as part of a larger
U.S. effort to encourage democracy and promote free market systems
in countries of Eastern Europe and the former Soviet Union.
Figures 2.1 and 2.2 and table 2.1 show the geographic breakdown of
borrowing, by region and by country, in terms of dollars lent since
the Housing Guaranty Program's inception.
Figure 2.1: Regional
Distribution of Housing
Guaranty Program Loans (Fiscal
Years 1964-94) (In millions of
1995 dollars)
(See figure in printed
edition.)
\a Israel, Portugal, and Poland.
Note: Shaded areas reflect total volume of guarantied loans
contracted in respective regions during 5-year intervals. No loans
were contracted prior to fiscal year 1964.
Source: Our analysis of USAID data.
Table 2.1
Value of Loans Guarantied by USAID Under
Housing Guaranty Program by Country (as
of Sept. 30, 1994)
(Dollars in millions)
Value of
Face value of loans
loans guarantied in
Country/regional bank guarantied 1995 dollars
---------------------------------------- ------------- -------------
Belize $2.0 $2.9
Mauritius 4.0 5.4
Ethiopia 1.5 5.5
Guyana 1.6 6.7
Paraguay 4.0 9.3
Poland 10.0 10.2
Interamerican Savings and Loan Bank 6.0 11.9
Barbados 10.0 14.9
Botswana 9.9 15.9
Senegal 5.0 20.5
Republic of China 4.8 21.2
Iran 7.5 23.6
Pakistan 25.0 26.2
Zaire 10.0 33.5
Philippines 35.0 36.9
Thailand 15.0 37.0
Guatemala 17.7 45.1
Lebanon 30.0 45.5
Morocco 43.5 48.8
Mexico 10.8 50.2
Bolivia 28.4 59.4
Nicaragua 16.0 60.0
Dominican Republic 17.0 66.1
Jordan 55.9 66.3
El Salvador 25.9 71.0
Costa Rica 40.0 71.5
Sri Lanka 60.0 83.4
Zimbabwe 65.0 88.6
Kenya 50.0 90.7
Ecuador 63.2 105.0
Colombia 26.9 118.8
Honduras 68.1 119.4
Cote d'Ivoire 70.9 125.1
Indonesia 120.0 133.5
Argentina 39.7 154.9
Chile 79.7 173.0
Panama 87.3 177.0
Tunisia 108.0 183.5
India 155.0 189.5
Venezuela 51.4 199.7
Portugal 140.0 215.5
Jamaica 123.3 223.3
Peru 103.7 226.0
Korea 95.0 251.9
Central American Bank for Economic 139.9 267.1
Integration
Israel 600.0 959.5
======================================================================
Total $2,683.6 $4,950.9
----------------------------------------------------------------------
Source: Our analysis of Housing Guaranty Program Financial
Summaries.
Figure 2.2: Countries That
Have Received Housing Guaranty
Program Loan Funds
(See figure in printed
edition.)
Note: Does not include loans
to two regional banks serving
multiple Central American
countries.
(See figure in printed
edition.)
Source: Our analysis of USAID
data.
(See figure in printed
edition.)
--------------------
\1 USAID has obtained over $50 million in appropriations since fiscal
year 1992 to set up such reserves for new loan guaranties.
SCOPE OF PROGRAM HAS EXTENDED
BEYOND HOUSING PROJECTS
---------------------------------------------------------- Chapter 2:2
The Housing Guaranty Program originally provided financing for U.S.
construction firms to build housing projects in Latin America to
demonstrate advanced methods of construction and finance that could
be replicated in the borrowing country. In 1965, the Congress
reoriented the program toward "institution building"--financing
demonstration projects developed by local institutions so that they
could gain direct experience and continue to build low-cost homes.
In 1975, the Congress changed the authorizing legislation to require
that at least 90 percent of future guaranties be issued for housing
suitable for families with incomes below the median income of the
borrowing country.\2 The Congress also directed that the Housing
Guaranty Program complement other development assistance programs and
that housing projects demonstrate the "feasibility and suitability of
particular kinds of housing or financial or other institutional
arrangements." As a result of these legislative changes, the focus
was on financing low-cost, minimum standard housing projects designed
to be affordable to the poor.
In 1985, the word "housing" was replaced with the word "shelter" in
the Foreign Assistance Act, and the scope of the program was expanded
to include "essential urban development services." USAID interprets
the broader language of the legislation as including any
shelter-related project that affects the greater urban environment,
including infrastructure projects.
USAID currently guaranties loans in developing countries for the
following categories of projects:
finance projects for housing and infrastructure development, such
as potable water, sewerage, electricity, and roads, designed to
be affordable to below-median income families;
institution-building projects to develop or enhance housing finance
systems, such as mortgage banking networks;
urban environment projects designed to manage urban growth and
improve municipal administration; and
projects that enhance local governance and municipal management to
foster broad-based participation in the governance of urban
communities.
The following are examples of the wide variety of uses of guarantied
loan funds we observed in the countries we visited.
In Ecuador and India, loans financed mortgages and home improvement
loans, aimed at involving private financial institutions in
long-term shelter financing.
In Indonesia, loans financed the construction of drainage ditches,
public toilets, roads and walkways, and water supply and
treatment facilities, aimed at promoting the development and
management of environmental infrastructure by local governments.
In Tunisia, loans financed serviced housing sites and sewer
extensions to slum dwellers, aimed at stimulating the production
of more affordable sanitary shelter.
Figures 2.3 through 2.8 are pictures of some projects financed
through the Housing Guaranty Program in the countries we visited.
Figure 2.3: New Housing With
Mortgages Financed by Housing
Guaranty Program in Carapungo,
Ecuador
(See figure in printed
edition.)
Source: Ecuadorian Housing Bank.
Figure 2.4: Housing With
Construction Financed by
Housing Guaranty Program in
Parral, Chile
(See figure in printed
edition.)
Source: GAO.
Figure 2.5: Owner-Improved
Housing With Original
Construction Financed by
Housing Guaranty Program in
Parral, Chile
(See figure in printed
edition.)
Source: GAO.
Figure 2.6: Housing and
Infrastructure Financed by
Housing Guaranty Program in El
Mourouj, Tunisia
(See figure in printed
edition.)
Source: GAO.
Figure 2.7: Housing Built on
Sites With Infrastructure
Financed by Housing Guaranty
Program in Larache, Morocco
(See figure in printed
edition.)
Source: GAO.
Figure 2.8: Footpath and
Drainage Ditch Financed by
Housing Guaranty Program in
Kecamatan, Indonesia
(See figure in printed
edition.)
Source: GAO.
--------------------
\2 The International Development and Food Assistance Act of 1975.
CURRENT STRATEGY FOCUSES ON
POLICY REFORM AND TECHNICAL
ASSISTANCE
---------------------------------------------------------- Chapter 2:3
For many years, USAID used the Housing Guaranty Program primarily to
develop demonstration projects that USAID hoped would be replicated
by local entrepreneurs and institutions. However, USAID has begun to
use the program primarily to leverage policy reforms that are
intended to have positive long-term effects on the borrowing
countries' shelter sectors. According to USAID officials, the loan
guaranties have served as incentives for recipient countries to enact
USAID-recommended reforms. In some cases, specific reforms are
preconditions for receiving loans. In addition, according to RHUDO
officials we spoke to, the loans afford USAID day-to-day influence on
policymakers in the host country: the loans "buy USAID a seat at the
table" where shelter policy is made.
Policy reform objectives associated with the Housing Guaranty Program
have included
in Chile, enacting laws to establish a mortgage-backed securities
market intended to facilitate investment in low-income mortgages
by pension funds and other private investors;
in Ecuador, adopting an accounting unit for housing finance indexed
to inflation (called the constant value unit), intended to
facilitate private investment in housing in a climate of high
inflation;
in Indonesia, strengthening local governments' ability to develop
urban infrastructure by allowing the governments to retain 100
percent of property tax revenues collected and by reducing the
central government's direct involvement; and
in Tunisia, converting a central government agency from a direct
producer of individual housing lots to a wholesaler of land to
private developers.
TECHNICAL ASSISTANCE IS
INTEGRAL TO PROGRAM
-------------------------------------------------------- Chapter 2:3.1
Consistent with its policy reform strategy, USAID has increasingly
relied on technical assistance to implement the Housing Guaranty
Program.\3 USAID is providing technical assistance, through
contracts, for all of its ongoing projects at a cost of $471
million.\4 According to program officials, USAID's regional bureaus
provide the funding for this technical assistance from the
Development Assistance, Economic Support Fund, and other program
accounts.
In the countries we visited, technical assistance was provided as
part of USAID's efforts to achieve desired policy reform and
institutional development objectives.
In Indonesia, USAID-funded technical advisors in four government
agencies helped implement decentralization policies intended to
lead to more efficient development of urban infrastructure
projects by local governments.
In India, a USAID-funded seminar resulted in the development of a
government action plan to introduce fundamental urban policy
changes intended to facilitate private sector participation in
infrastructure development.
In Poland, USAID-funded seminars and consulting services, including
a program to train employees of nine banks in mortgage
underwriting, enabled local financial institutions to originate
sound mortgages.
In Ecuador, USAID-funded technical assistance helped the city of
Quito's Municipal Water Company find ways to reduce its
construction costs and improve water service to the city's
marginal neighborhoods.
In Morocco, extensive technical assistance, including funding for
consultants, computer equipment, and vehicles, was provided to
the city of Tetouan to help it improve municipal management of
urban infrastructure, including project planning, construction,
and cost recovery.
--------------------
\3 Technical assistance consists of consultants, training, equipment,
facilities, and other assistance provided to the borrower and other
institutions participating in the program.
\4 Amount is not adjusted for inflation.
HOUSING GUARANTY PROGRAM IS IN
SERIOUS FINANCIAL CONDITION
============================================================ Chapter 3
The Housing Guaranty Program is not self-sustaining. Although the
Congress specifically authorized USAID to charge borrowers fees to
cover costs, the cost to the U.S. government of the existing loan
guaranty portfolio is likely to total about $1 billion (in 1995
dollars).\1 The Housing Guaranty Program has already received $542
million in appropriations and U.S. Treasury borrowings to finance
earlier losses resulting from loan defaults. In addition, we
estimate the cost of future loan defaults to be about $600 million
more. Projected program revenues are not adequate to cover
administrative and default costs. Although program officials expect
to recover most of the losses from defaulted borrowers, we estimate
USAID will ultimately recover only about $200 million. These costs
do not include several hundred million dollars in technical
assistance that USAID has provided to implement the program but has
not reported as a program cost. Moreover, the legislated ceiling on
loan guaranties was not effective at limiting the U.S. government's
financial investment in this program, since it allowed USAID to issue
new guaranties even though defaulted borrowers had not repaid their
outstanding debts.
--------------------
\1 Unless otherwise noted, all figures are presented in 1995 dollars.
PROGRAM HAS ALREADY INCURRED
SIGNIFICANT COSTS FOR LOAN
DEFAULTS
---------------------------------------------------------- Chapter 3:1
The President's annual budgets for fiscal years 1971 through 1975
indicate that the Housing Guaranty Program was originally intended to
be self-sustaining. The documents stated that "consistent with the
intent of the Congress that the housing programs be totally
self-supporting, the costs of administration, program evaluation and
development, and claims investigations are paid from fee income."
However, fee income has not been adequate to cover all program costs,
and USAID has required $542 million in inflation-adjusted
appropriations and U.S. Treasury loans. Although historical
financial information for this program is incomplete, we determined,
based on a review of audited financial statements and budgetary
documents, that from fiscal years 1970 to 1994, the program had
experienced at least $542 million in losses due to loan defaults.
USAID financed these losses with $258 million in appropriations ($417
million in 1995 dollars) and $125 million in loans from the U.S.
Treasury.\2
USAID's reported program costs do not include several hundred million
dollars that it has committed for technical assistance to implement
this program, as described in chapter 2. USAID does not report the
cost of this assistance as a Housing Guaranty Program expense in its
budget or financial statements, because the cost is funded from other
USAID program accounts. However, we believe that much of this
assistance amounts to closely-linked cross subsidies to program
participants, and including them in reports on the cost of the
program is reasonable.\3 Although USAID's ongoing technical
assistance is currently estimated to cost $471 million,\4 historical
cost information was not available. Therefore, the total cost of
technical assistance associated with this program is unknown, though
it is probably much higher than this amount.
--------------------
\2 A more precise accounting of total losses was not available
because financial records prior to fiscal year 1989 are incomplete
and unreliable. Prior to 1968, an undisclosed amount of
administrative expenses were paid from USAID's general appropriation
instead of from program revenues. In fiscal year 1970, the Congress
provided the program a reserve fund of $50 million and another $40
million in 1985 to cover expenses associated with borrower defaults.
Having expended these reserves, from fiscal years 1987 to 1992 the
program borrowed, but has not yet repaid, $125 million from the U.S.
Treasury, and from fiscal years 1992 through 1994, USAID received
$168 million in appropriations to fund the program's annual
shortfalls. Adjusted for inflation, these figures represent a total
of $542 million that USAID required to finance program losses.
\3 We developed two criteria for defining a closely-linked cross
subsidy program. They are that (1) the appropriation for the cross
subsidy program directs assistance to the credit program participants
and (2) the government's commitment of the two programs is obligated
concurrently. For a more detailed discussion of closely-linked cross
subsidies, see Credit Reform: Case-by-Case Assessment Advisable in
Evaluating Coverage and Compliance (GAO/AIMD-94-57, July 28,1994).
\4 Amount is not adjusted for inflation.
PROJECTED FUTURE DEFAULT COSTS
TOTAL ABOUT $600 MILLION
---------------------------------------------------------- Chapter 3:2
Our analysis of the program's outstanding loan guaranty portfolio and
the risk profile of current borrowers indicates that over the
remaining life of the portfolio USAID will likely have to make
principal and interest payments now valued at $600 million on behalf
of borrowers that do not fully repay their loans.\5 These cost
projections are rough and will vary over time depending upon market
expectations for the portfolio. According to the program's fiscal
year 1993 audited financial statements, USAID estimates this
liability at over $700 million.
USAID has no reserves to fund this liability, except for about $50
million in appropriations it obtained to finance projected default
costs associated with loan guaranties authorized after the Credit
Reform Act took effect. The remaining default costs will be financed
with budget authority established for such unfunded liabilities under
the Credit Reform Act.\6
USAID's unfunded liability includes costs associated with $193
million in loan guaranties it authorized before the Credit Reform Act
took effect but had not contracted or disbursed as of March 1995.
Some of these loans are to countries with relatively poor credit risk
ratings, such as Ecuador and Jamaica. If USAID deauthorizes these
loan guaranties instead of disbursing the loans as currently planned,
future default costs could be reduced by over $50 million.
--------------------
\5 To determine the probable default costs associated with USAID's
pre-Credit Reform loan portfolio, we applied an updated version of
our method for estimating the cost of country risk associated with
loan guaranties outstanding. Our country risk ratings were developed
by statistically averaging two contemporaneous, professionally
accepted ratings--Euromoney and the Institutional Investor. Our
method transformed country risk ratings into country risk cost using
econometric estimates on data derived from the secondary market.
\6 The Credit Reform Act requires that loans disbursed before fiscal
year 1992 be budgeted and accounted for on a cash basis. The act
provides permanent, indefinite budget authority to fund defaults from
such loans to the extent that the defaults are not covered by funds
available in the credit liquidating account from prior year
surpluses.
FUTURE FEE REVENUE WILL BE
INADEQUATE TO COVER COSTS
---------------------------------------------------------- Chapter 3:3
Based on our analysis of the Housing Guaranty Program portfolio,
projected revenues will continue to be inadequate to cover future
defaults and other program costs. Presented in present value terms,
we estimate that from fiscal years 1995 through 2024, USAID's program
expenses for the current portfolio are likely to total about $800
million: about $600 million in default costs, about $80 million in
administrative costs,\7 and $125 million to repay funds it still owes
to the U.S. Treasury to cover earlier program losses. However,
USAID is likely to collect less than $100 million in fees during this
period.\8
The Foreign Assistance Act authorizes USAID to charge borrowers fees
to cover the cost of this program. However, despite the significant
losses incurred and anticipated for this program, USAID has not taken
any steps to increase fee revenues. The program's existing fee
structure has remained virtually unchanged since fiscal year 1972.
Specifically, on new loans borrowers are assessed a 1-percent
initiation fee and an annual fee of 0.5 percent of the outstanding
loan balance, regardless of credit risk.
In its audit report on the program's fiscal year 1991 financial
statements, USAID's Inspector General recommended that USAID managers
consider implementing a risk-based fee structure, and USAID has
funded two consultant studies to examine this issue. One study
showed that an increase in annual fees to 1.5 percent of the
outstanding loan principal for fiscal years 1971 through 1990 would
have eliminated the need for $165 million (nearly $200 million in
1995 dollars) in appropriations and U.S. Treasury borrowing during
that period. Another study indicated that the program's fees are in
the middle range of similar fees charged by other international
lenders. Both studies stated that an increase in fees could
negatively affect the willingness of countries to participate in the
program, although few other lenders offer the attractive 30-year
repayment terms that the Housing Guaranty Program does. According to
one study, charging higher fees for certain borrowers could be
politically embarrassing for USAID.
Our assessment of the program's financial condition indicates a
necessity to raise fees in order to make the program as
self-sustaining as possible. Even with higher fees, USAID-guarantied
loans would likely remain attractive to many developing countries
because of their low interest rates, lengthy repayment terms, foreign
exchange value, accompanying technical assistance, and the increased
financial credibility associated with receiving U.S. government
credit. Also, as one of USAID's consultant studies indicates, USAID
could raise its fees and still remain competitive with other
international lenders.
--------------------
\7 Administrative costs were assumed to be a constant percentage of
the probable outstanding loan balance for the next 10 years and were
estimated based on fiscal year 1994 appropriations figures. We
further assumed that after 10 years, the administrative costs
associated with the existing portfolio would be negligible.
\8 Fee income includes initiation fees, annual fees, and USAID late
fees. Late fees (charged at the loans contracted rate of interest)
were assumed to be a constant percentage of the outstanding loan
balance and were estimated based on actual fiscal year 1993 figures.
This estimate assumes the same probable default rate for annual fees
as we used for principal and interest payments.
USAID IS UNLIKELY TO RECOVER A
GROWING AMOUNT OF DEFAULTED
DEBT
---------------------------------------------------------- Chapter 3:4
Our review of USAID's financial records indicates that the amount of
debt that defaulted borrowers owe USAID has been escalating in recent
years. Figure 3.1 shows that, not adjusted for inflation,
uncollected debt more than doubled, from $171 million in fiscal year
1989 to $409 million in fiscal year 1994. This debt consists of
defaulted principal, interest, and fees. Based on our analysis of
USAID's portfolio of rescheduled debt, we estimate that USAID can
ultimately expect to recover only about one-half of this debt, or
about $200 million. According to the program's fiscal year 1993
audited financial statements, the most recent available, USAID is
likely to collect only about $30 million of this debt (see table 3.1
for a comparison of USAID's and our analyses of program cost). In
addition to this defaulted debt, USAID has incurred over $100 million
in other unrecoverable program costs over the life of the program.\9
Figure 3.1: Cumulative
Uncollected Housing Guaranty
Program Debt Owed to USAID
(Fiscal Years 1989-94) (Dollars
in millions)
(See figure in printed
edition.)
Note: Amounts are not adjusted for inflation.
Source: USAID Office of the Inspector General Audit Reports and our
analysis of Housing Guaranty Program Financial Summaries.
Table 3.1
Comparison of USAID and GAO Analyses of
Program Cost to the U.S. Government
(Dollars in millions)
USAID GAO analysis
Fiscal financial (1995
Program funding year(s) records dollars)
-------------------------------- -------- ------------ ------------
Appropriation for reserve 1970 $50 $185
Appropriation for reserve 1985 40 55
Net U.S. Treasury borrowings 1987-91 125 125
Appropriations for 1992-94 168 177
administrative and default
costs on pre-Credit Reform
guaranties
Appropriations for future 1992-94 49 50
default costs of post-Credit
Reform guaranties
Appropriations for default costs 1995- 707 575
of pre-Credit Reform 2025
guaranties, net of recoveries
Estimated recoveries from about 1995- (29) (200)
$400 million of defaulted loans 2025
outstanding as of September 30,
1994
======================================================================
Total $1,110 $967
----------------------------------------------------------------------
Housing Guaranty Program officials believe that USAID is likely to
recover nearly all of this debt. This belief is based on their
observation that over the past 4 years many of the defaulted
borrowers in Latin America have undergone fundamental economic
changes, resulting in increasing creditworthiness. This economic
improvement, they believe, is unlikely to be reversed, and the
long-term trend for these countries is positive. The officials did
not provide any specific estimates of probable defaults. Our review
indicates that USAID's forecast is unreasonably optimistic; it is
essentially based on intuition rather than sound analysis. In our
December 1994 report on credit reform, we showed that the probability
of default for most of the Latin American countries with the greatest
Housing Guaranty Program debt, despite recent economic improvements,
still exceeded 50 percent.
Instead of immediately repaying defaulted debt to USAID, borrowing
countries have increasingly rescheduled their debts to the United
States through bilateral agreements under the Paris Club.\10 The
borrowing countries agree to repay this debt along with other debt
they owe to the U.S. government, with interest, over an extended
period of time. The rescheduled debt includes all outstanding debt
payments a country has defaulted on or is in imminent default on.
Multiple rescheduling of rescheduled debt to defer repayment is
common; in fact, one borrower has rescheduled its debt 10 times. As
a result, by September 30, 1994, USAID had collected only about 5
percent of all rescheduled debt.
USAID has, in effect, provided additional benefits to borrowers that
rescheduled their debts instead of paying them immediately. For
example, we identified 12 cases since 1985 in which USAID authorized
new disbursements of Housing Guaranty Program loans to countries that
had recently rescheduled (or were in the process of rescheduling)
debt to USAID.\11 USAID provided new guarantied loans for four
countries (Ecuador, Jamaica, Jordan, and Panama) despite repeated
rescheduling of old debt. According to USAID officials, USAID waives
initiation and annual fees on rescheduled loans to encourage
repayment of the original loan amount. This has amounted to a waiver
of over $11 million ($10.6 million in then-year dollars) in fees for
fiscal years 1987 through 1994.
Figure 3.2 shows the distribution of uncollected debt by geographic
region (see app. I for outstanding guarantied loan balances and
unpaid debts by country).
Figure 3.2: Geographic
Distribution of Uncollected
Debt Owed to USAID (as of Sept.
30, 1994)
(See figure in printed
edition.)
Source: Our analysis of Housing Guaranty Program Financial
Summaries.
--------------------
\9 Unrecoverable costs include default costs USAID has "written off"
($39 million in then-year dollars), including default costs for a
loan to the Imperial Government of Iran and 31 pre-1973 loans to
nonsovereign borrowers with no host government guaranty of repayment.
USAID has been making principal and interest payments on these loans
for many years but has recovered virtually none of its costs and does
not expect to recover them. A precise accounting of all
unrecoverable costs was not possible, because financial records
provided to us for periods prior to fiscal year 1989 are incomplete
and unreliable.
\10 When foreign countries are unable to meet their external debt
obligations owed to governments, the Paris Club is the mechanism the
United States and other official creditors use to reschedule the
debt. Paris Club meetings are organized by the French Finance
Ministry. The Department of State represents the U.S. government in
Paris Club negotiations.
\11 This occurred for Bolivia, Costa Rica, Chile, Ecuador, Jamaica,
Jordan, and Panama.
LEGISLATED CEILING DID NOT
LIMIT FINANCIAL COMMITMENT
---------------------------------------------------------- Chapter 3:5
The legislated ceiling on loan guaranties has not been effective at
limiting the U.S. government's financial commitment under this
program. The Foreign Assistance Act authorized USAID to issue
guaranties for up to about $2.6 billion in loans before fiscal year
1992.\12 When loan principal was repaid, USAID reduced the liability
it recorded under the ceiling, allowing it to issue a corresponding
amount of new loan guaranties. This occurred even when USAID was
required, as guarantor, to make a payment on behalf of a borrower.
These amounts paid by USAID are no longer guarantied loans but a
program asset due from a borrower and are therefore no longer counted
against the loan guaranty ceiling. Thus, USAID continued to issue
new loan guaranties while the total amount of uncollected debt grew.
As a result, by fiscal year 1995, the total amount of U.S.
government resources committed by USAID had grown to $2.8 billion,\13
despite the authorized ceiling of $2.6 billion. If USAID had been
required to include uncollected debt under the ceiling, it would have
been unable to issue $200 million in loan guaranties, reducing the
probable cost of loan defaults by about $50 million.
--------------------
\12 Beginning in fiscal year 1992, new loan guaranties were not
subject to this authorized ceiling, but were required to be
authorized individually in accordance with the provisions of the
Credit Reform Act.
\13 This figure reflects the face value (not the 1995 value) of all
guarantied loans outstanding and uncollected debt. It does not
include $355 million in loan guaranties authorized since September
30, 1992, which are subject to the provisions of the Credit Reform
Act and do not fall under the Foreign Assistance Act ceiling.
PROGRAM'S IMPACT ON LOW-INCOME
SHELTER INVESTMENT IS NOT EVIDENT
============================================================ Chapter 4
The Housing Guaranty Program and related technical assistance have
contributed to many improvements in the shelter sectors of
participating countries. However, USAID has not achieved a major
long-run objective of the program as envisioned in the Foreign
Assistance Act: to increase private investment in shelter for poor
families. The program has not convinced local entrepreneurs and
private institutions that investing their capital in lower income
shelter is financially viable. In addition, USAID's implementation
of the program has not ensured that at least 90 percent of guarantied
loan funds were spent on projects directly benefiting families below
the median income level, the program's target population. Finally,
USAID has not designed its performance indicators to measure progress
toward the program's objectives.
PROGRAM HAS NOT STIMULATED
INVESTMENT
---------------------------------------------------------- Chapter 4:1
Despite over 30 years in operation, the Housing Guaranty Program has
not increased private investment in low-income shelter. According to
the Foreign Assistance Act, the program was intended "to increase the
availability of domestic financing for low cost shelter by
demonstrating to local entrepreneurs and institutions that providing
this shelter is financially viable." In the seven participating
countries we visited, the program has helped bring about some
improvements in these countries' shelter markets. However, the
program has not stimulated local credit institutions to make private
investment capital available for low-income shelter projects.
The program can be credited with helping countries increase
participation by private institutions in state-financed shelter
production. However, these institutions have not increased their own
investment in the low-income shelter sector. The following examples
illustrate the limited involvement of private sector investment in
the Housing Guaranty Program.
In India, the Housing Guaranty Program was aimed at increasing
private sector financing of low-income housing. To this end,
the program has helped India's National Housing Bank finance
low-cost mortgages and home improvement loans through private
housing finance corporations. However, the housing finance
corporations have not invested their own capital in these types
of projects. Thus, a 1994 evaluation of the India program,
conducted by a USAID contractor, indicated that USAID had not
met the goal of increasing the availability of private housing
finance for low income households, but rather established
another conduit for distributing public resources.
In Ecuador, the government dominates the formal development of
low-income housing. The most recent Housing Guaranty Program
effort in that country was aimed at involving commercial banks,
credit unions, and savings and loans in low-cost housing
development by providing below market rate investment capital
through the state-run Ecuadorian Housing Bank. Officials of
several of these institutions told us that although their
involvement in the USAID program was financially satisfactory,
they did not plan to invest their own capital in similar
projects.
In Chile, the government provides families with direct subsidies
and loans to purchase homes, but the demand for affordable
housing has exceeded the supply, and many families cannot take
advantage of the government financing. Two of USAID's recent
programs in Chile provided short-term construction financing to
three cooperatives to encourage them to satisfy some of this
housing demand. In addition to Housing Guaranty Program
financing, the cooperatives have been able to obtain additional
short-term, market rate construction loans from commercial
banks. However, when the homes were completed, the government
of Chile reimbursed the cooperatives most of the development
costs through subsidies to the home buyers. Thus, while this
program has helped enhance the availability of housing, the
long-term financing for the housing is still provided by the
government, not the private sector.
In Tunisia, a recent Housing Guaranty Program provided a loan for
the country's Housing Bank to increase private production of
affordable housing and developable land. The loan was
contingent upon a series of policy and institutional reforms.
USAID has indicated that the program's initial $15 million
investment led to millions of dollars of additional lending by
the Housing Bank for low-income mortgages. However, the capital
for this lending was not raised from private credit
institutions, rather from a 1-percent tax on salaries and from
previous loan repayments. According to a 1993 evaluation of
this program, conducted by a USAID contractor, the low return on
investments in this area makes it impossible to raise financing
capital from alternative sources. According to USAID, the
Housing Bank raised about $75 million in short-term financing
from private sources to fund initial project construction.
However, the long-term financing of these projects was provided
through the government-sponsored mortgage program.
In Indonesia, infrastructure projects had been the domain of the
central government, which planned, financed, and constructed
them. Since 1988, in part with the encouragement and assistance
provided by USAID through the Housing Guaranty Program, the
government of Indonesia has pursued decentralization, allowing
municipalities to raise funds locally and develop infrastructure
projects. With seed money from the program, the government of
Indonesia established a Regional Development Account to lend
funds to local governments for slum improvements and other
shelter-related infrastructure projects. However, this program
has not increased the availability of private financing for
low-cost shelter because the investment capital for these
projects was provided by the government of Indonesia at
below-market rates and not from private credit markets.
Evaluations of the Housing Guaranty Program in Indonesia,
conducted by USAID contractors in 1992 and 1993, indicated that
more attention was needed on providing local governments access
to private sector financing.
By failing to stimulate private investment and instead relying on
host government financing and subsidies, the program's impact is
limited. Given the financial constraints on developing country
governments, they have been unable to provide adequate resources to
finance the shelter their poor populations need. As the Director of
USAID's Office of Environment and Urban Programs has acknowledged,
"most existing housing subsidies . . . do not result in programs
of sufficient order of magnitude to really have an impact on the
shelter sector."
Even though it has not been able to convince local entrepreneurs to
invest in low-income shelter, USAID has repeatedly guarantied loans
in the same countries. As table 4.1 shows, in some of the countries
we visited, USAID has been providing guarantied loans for over 25
years.
Table 4.1
Loan Guaranty History in Selected
Countries
(In millions of 1995 dollars)
Total
value
of
Year of loans
first Number guaran
Country loan of loans tied
------------------------------------------ -------- -------- ------
Chile 1968 8 $173.0
Ecuador 1968 7 105.0
India 1983 10 189.5
Indonesia 1989 5 133.5
Morocco 1985 5 48.8
Poland 1994 1 10.2
Tunisia 1964 12 183.5
----------------------------------------------------------------------
Source: Our analysis of Housing Guaranty Program Financial
Summaries.
According to USAID officials, the Housing Guaranty Program and
related technical assistance have contributed to innovations and
greater efficiency of the host governments in providing shelter.
They cite as an example a demonstration project through which USAID
convinced the government of Morocco of the efficacy of rehabilitating
slums rather than bulldozing them. They told us that the program has
also helped countries establish more reasonable building codes and
land tenure policies intended to facilitate the construction and
upgrading of low-income dwellings. However, USAID has not
demonstrated to what extent these innovations have led to greater
long-term investment by the private sector in low-income shelter
developments.
USAID hopes to make progress in marshalling private investment
resources in the future. For example, projects being initiated in
Chile and Ecuador, if successful, could allow financial institutions
to raise capital for shelter projects by selling mortgage-backed
bonds to private investors, such as pension funds and insurance
companies. Also, efforts are underway in India to enable housing
institutions to access private credit markets for long-term
financing. However, our review indicated that these programs were in
their preliminary stages, and USAID has not yet demonstrated whether
this is a viable option for financing low-income shelter projects,
which are traditionally considered higher risk investments.
USAID DOES NOT ALWAYS ENSURE
ACCESS TO PROJECTS BY LOWER
INCOME GROUP
---------------------------------------------------------- Chapter 4:2
The Foreign Assistance Act requires that at least 90 percent of the
guaranties issued under the Housing Guaranty Program be used to
finance shelter suitable for families with incomes below the
country's median income. However, our review indicated that USAID
does not always ensure that its projects are fully accessible by the
below-median income target population.
USAID contends that as long as the projects funded by the program are
priced to be affordable to below-median income families, they are
considered suitable, regardless of whether those families actually
benefit from the projects. Many of our project site visits and
extensive review of USAID documents in the countries we visited,
including a review of USAID consultant studies, showed that factors
other than price have sometimes adversely affected below-median
income families' access to program loan funds.
In India, we found that at least one bank participating in the
program extended loans to its own employees, who had
above-median incomes. One such employee we visited obtained a
home improvement loan to build an addition, where an extended
family member conducted aerobics classes. A RHUDO-sponsored
survey of beneficiaries indicated that about 36 percent of the
loans financed through the Housing Guaranty Program went to
families with above-median incomes.\1
In Ecuador, a nonprofit agency told us that many higher income
people had purchased and upgraded housing financed through the
Housing Guaranty Program. A 1989 USAID-sponsored evaluation of
the Ecuador program also expressed this concern.
In Indonesia, several program-financed projects we visited appeared
to be serving people above the country's median income. This
included a pay toilet in Bali that served tourists (shown in
fig. 4.2) and small-scale infrastructure projects in higher
income neighborhoods. In a 1992 evaluation of the Indonesia
program, a USAID consultant noted that USAID had not taken
adequate measures to ensure that at least 90 percent of the
beneficiaries were below the median income, leading to
"conflicting speculation as to whether these investments are
reaching the poor."
In Morocco, program-financed housing built for lower income
families to relocate from slums (shown in fig. 4.3) had not
been sold to those families. Instead, at the time of our visit,
the municipality that built them was holding some of them vacant
and intended to sell them to other buyers at a profit to raise
funds for other municipal projects.
In Tunisia, a RHUDO-sponsored survey of beneficiaries indicated
that about 17 percent of one sampling of families that USAID
deemed eligible to benefit from the program had above-median
incomes.\2
Figures 4.1 through 4.3 show examples of program-financed projects we
visited that did not benefit below-median income families.
Figure 4.1: House Partially
Financed by the Housing
Guaranty Program for
Above-Median Income Family in
Bangalore, India
(See figure in printed
edition.)
Source: GAO.
Figure 4.2: Pay Toilet for
Tourists Financed by the
Housing Guaranty Program in
Bali, Indonesia
(See figure in printed
edition.)
Source: GAO.
Figure 4.3: Partially
Unoccupied Housing Financed by
Housing Guaranty Program in
Tetouan, Morocco
(See figure in printed
edition.)
Source: GAO.
We determined that, in several countries we visited, RHUDO and USAID
missions did not routinely visit projects or review case files to
monitor beneficiary selection and ensure that programs benefited
below-median income groups. Instead, they often relied on reports
from borrowers, such as certifications attesting that mortgages or
home improvement loans were affordable to the below-median income
target groups.
If USAID is to ensure that host country institutions follow its lead
and develop shelter projects that benefit below-median income
families, USAID must first ensure that its demonstration projects
reach that target population.
--------------------
\1 According to USAID officials in India, these were acceptable
results because participating institutions made additional loans to
families with below-median incomes for which they did not receive
Housing Guaranty Program financing. However, such loans were larger
than USAID determined would be affordable to a family below the
median income.
\2 According to RHUDO officials in Tunisia, these results were
acceptable because the government of Tunisia finances other loans to
families below the median income.
PERFORMANCE INDICATORS DO NOT
MEASURE PROGRESS TOWARD PROGRAM
OBJECTIVES
---------------------------------------------------------- Chapter 4:3
Since 1991, USAID has been developing performance indicators to
measure the Housing Guaranty Program's impacts on the shelter sectors
of developing countries. However, these indicators do not focus on
the effectiveness of the program in persuading local investors to
finance low-income shelter projects. Nor do they measure the extent
to which the program is serving its low-income target population.
Instead, they measure broad changes in the shelter sector, such as in
infrastructure expenditures per capita and the house price to income
ratio, and associate them with USAID's programs.
Given the many factors outside of USAID's control that affect the
overall shelter sector in a country, direct correlations between
individual Housing Guaranty Program efforts and broad sector changes
are unreliable. Domestic conditions, such as the level of social,
economic, and political stability, are significant factors affecting
a country's shelter sector that could be reflected in USAID's
indicators. Also, the effects on the overall shelter sector of
parallel initiatives are difficult to independently monitor. USAID's
assistance may be relatively minor in relation to a country's own
investments and the assistance provided by other bilateral or
multilateral donors. For example, in 1993 USAID-guarantied loans
represented only
7 percent of the capital in Indonesia's Regional Development Account
for shelter-related infrastructure investment. In Poland, USAID's
$25 million loan guaranty program is part of a $400 million housing
finance program with the World Bank and the European Bank for
Reconstruction and Development. USAID officials have also observed
that the performance indicators are of limited use in measuring the
direct results of its program.
USAID can develop ways to measure progress toward its objectives on a
smaller, more realistic scale. For example, through surveys and
other evaluation methodologies, USAID can measure the direct impact
of its demonstration projects and policy reform on the investment
patterns of a sampling of private sector institutions.
CONCLUSIONS, MATTERS FOR
CONGRESSIONAL CONSIDERATION, AND
RECOMMENDATIONS
============================================================ Chapter 5
Throughout the Housing Guaranty Program's 30-year history, the
Congress and USAID have modified the program to try to improve its
effectiveness. After initially focusing primarily on Latin American
countries, USAID shifted to other regions of the world and then to
more creditworthy and advanced countries. USAID has also altered its
strategy by increasingly incorporating policy reform objectives and
providing several hundred million dollars in technical assistance to
complement the guarantied loans. However, despite these
modifications, the Housing Guaranty Program has not met congressional
expectations, as established in the Foreign Assistance Act.
Although the Congress authorized USAID to assess borrowers fees to
cover program expenses, the program's total cost to the U.S.
government is likely to be about $1 billion (in 1995 dollars). Loan
defaults (especially in Latin America, where the program had been
concentrated) have been high, and program costs have escalated. The
program has incurred at least $542 million in default costs. Future
default costs are likely to total about $600 million. To limit the
U.S. government's liability under the program, the Congress
established a $2.6 billion ceiling on loan guaranties; however, loan
guaranties and debts to USAID resulting from this program have
totaled about $2.8 billion.
While the Housing Guaranty Program has sponsored demonstration
projects and helped persuade recipient countries to adopt policy
reforms, it has not succeeded in convincing local entrepreneurs and
institutions to increase private investments in lower income housing
projects. Thus, the program's anticipated long-term impact on the
availability of capital for lower income shelter is not evident.
Furthermore, USAID has not ensured that the actual shelter projects
financed by this program necessarily benefit below-median income
families. USAID has not designed its indicators to measure the
direct impact of this program on private sector investment and on
below-median income families.
MATTERS FOR CONGRESSIONAL
CONSIDERATION
---------------------------------------------------------- Chapter 5:1
The Congress may wish to deauthorize guaranties for undisbursed
loans, where feasible, and terminate the Housing Guaranty Program
because (1) the program is now primarily benefiting borrowers in more
creditworthy and advanced developing countries that have access to
comparable loans from other international lenders; (2) the program
annually costs millions of dollars more than anticipated; (3) loans
previously guarantied under the pre-1992 legislated ceiling continue
to be disbursed, even though USAID has not collected over $400
million in defaulted debt; and (4) there is no evidence that the
program has measurably increased the availability of private domestic
capital for low-income shelter.
Other actions are warranted if the Congress believes that the goals
of (1) stimulating domestic investment in the recipient countries
(particularly private investment) and (2) targeting shelter
assistance to below-median income families are possible to achieve
and are of sufficient priority. Before authorizing additional loan
guaranties, the Congress may wish to require USAID to submit a
comprehensive plan to the appropriate congressional committees on how
it plans to achieve the stated program goals, reduce losses, and
return the program to a viable financial condition.
RECOMMENDATIONS
---------------------------------------------------------- Chapter 5:2
If the Congress does not terminate the Housing Guaranty Program, we
recommend that the USAID Administrator take the following actions to
(1) minimize the financial impact of the Housing Guaranty Program on
the foreign assistance budget and the U.S. government budget deficit
and (2) bring the program in line with the objectives of the Foreign
Assistance Act:
Withhold future loan disbursements and related technical assistance
from borrowers that have repeatedly rescheduled debt repayments
to USAID.
Increase program revenues by adopting a fee structure designed to
offset a larger portion of the program's costs.
Ensure that performance indicators measure the extent to which (1)
local investors replicate the program's low-income shelter
projects using private sources of long-term financing and (2)
project benefits accrue to the below-median income target
population.
AGENCY COMMENTS AND OUR
EVALUATION
---------------------------------------------------------- Chapter 5:3
The report has been revised, where appropriate, to reflect USAID's
concerns. The comments and our evaluation of them appear in
appendix II.
In commenting on a draft of this report, USAID strongly disagreed
with our findings, arguing that the Housing Guaranty Program was a
cost-effective method for pursuing development assistance and that
the program had achieved the goals established by the Congress.
USAID criticized the report for failing to adequately consider
alternative program goals and congressional directions. USAID also
asserted that the report was based on a flawed financial analysis and
a distorted portrayal of program costs because of the use of 1995
dollars. Furthermore, USAID maintained that we did not adequately
consider evidence that the program had successfully stimulated
private investment and reached below-median income target
populations. USAID suggested that termination of this program and
deauthorization of guaranties might interfere with legally binding
agreements and congressionally supported activities. USAID did not
address the report's specific recommendations, since it contended
that they were based upon incorrect information and a flawed
analysis.
As a result of these comments, we modified our matters for
congressional consideration and recommendations to (1) recognize that
it may not be feasible to deauthorize all loan guaranties and (2)
indicate that USAID should increase its guaranty fees to cover more
of its costs. However, our evaluation of these comments revealed
that USAID had not credibly supported any of its broad and strongly
worded criticisms. USAID did not identify any specific legislated
program goals that we overlooked or any legal provisions that
contradict the goals we focused on. USAID's criticism of our
financial analysis is unfounded and even disregards its own most
recent audited financial statements regarding probable losses from
loan defaults. Our use of 1995 dollars reflects the true cost of the
program more accurately than the use of then-year dollars, especially
for costs incurred over many years. We denominated figures in both
1995 and then-year dollars, except in cases, such as the discussion
of legislated ceilings, where it was not meaningful or possible to
use 1995 figures. Furthermore, the additional data USAID provided
(most of which we had already reviewed) were misleading and did not
undermine our conclusions.
HOUSING GUARANTY PROGRAM DEBTS AS
OF
SEPTEMBER 30, 1994
=========================================================== Appendix I
This appendix lists the countries and regional banks with outstanding
guarantied loan balances and debts to the U.S. Agency for
International Development (USAID) under the Housing Guaranty Program
as of September 30, 1994.
Country/regional bank Loan balance Debt to USAID
------------------------------ ------------------ ------------------
Botswana $9,565,714 $134,169
Ethiopia 0 1,878,010
Cote d'Ivoire 50,319,092 31,853,835
Kenya 43,206,224 3,996,380
Senegal 0 4,718,552
Mauritius 4,000,000 0
Zaire 2,384,974 18,392,795
Zimbabwe 65,000,000 0
======================================================================
Subtotal 174,476,004 60,973,741
India 144,000,000 0
Indonesia 120,000,000 0
Korea 53,321,406 0
Pakistan 25,000,000 0
Philippines 35,000,000 0
Sri Lanka 56,010,000 0
Thailand 9,750,000 0
======================================================================
Subtotal 443,081,406 0
Poland 10,000,000 0
======================================================================
Subtotal 10,000,000 0
Israel 520,500,536 0
Jordan 55,858,833 1,827,825
Lebanon 28,428,904 0
Morocco 43,459,684 0
Portugal 100,000,000 0
Tunisia 93,728,115 0
======================================================================
Subtotal 841,976,072 1,827,825
Argentina 4,148,611 42,451,775
Barbados 7,600,000 0
Belize 1,950,000 0
Bolivia 17,445,882 19,272,776
Central American Bank for 109,062,899 153
Economic Integration
Colombia 0 10,914
Chile 62,821,782 437,036
Costa Rica 30,983,229 4,295,015
Dominican Republic 737,257 6,241,941
Ecuador 54,438,827 12,619,483
El Salvador 13,770,659 3,478,405
Guatemala 10,000,000 0
Guyana 0 1,303,527
Honduras 58,793,468 21,749,898
Interamerican Savings and Loan 5,365,967 0
Bank
Jamaica 100,327,044 29,564,332
Nicaragua 2,620,496 22,284,013
Panama 69,322,973 27,520,864
Paraguay 3,036,032 0
Peru 72,060,131 154,888,134
Venezuela 2,388,688 140,784
======================================================================
Subtotal 626,873,945 346,259,050
======================================================================
Total $2,096,407,428 $409,060,616
----------------------------------------------------------------------
Note: Totals may not add due to rounding.
Source: Housing Guaranty Program Financial Summary
(See figure in printed edition.)Appendix II
COMMENTS FROM THE U.S. AGENCY FOR
INTERNATIONAL DEVELOPMENT
=========================================================== Appendix I
See pp. 7-8 and 53-54.
See comment 1.
(See figure in printed edition.)
See pp. 7-8 and 53-54.
(See figure in printed edition.)
See comment 2.
See comment 2.
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
(See figure in printed edition.)
See comment 3.
See comments 1 and 4.
See comment 4.
(See figure in printed edition.)
See comment 5.
See comment 6 and p. 40.
See comment 7.
(See figure in printed edition.)
See comment 8.
See comment 9.
See comment 10 and
pp. 7 and 53.
See comment 11.
(See figure in printed edition.)
(See figure in printed edition.)
See comment 12 and
p. 19.
See comments 12 and 13.
(See figure in printed edition.)
See comments 14 and 15.
See comments 14 and 15.
See comment 12.
See comment 15.
(See figure in printed edition.)
See comment 14.
(See figure in printed edition.)
See comment 16.
(See figure in printed edition.)
(See figure in printed edition.)
See comment 19.
See comment 19.
The following are GAO's comments on USAID's letter dated May 8, 1995.
GAO COMMENTS
--------------------------------------------------------- Appendix I:1
1. In nearly every case, the report shows monetary amounts in both
nominal and 1995 dollars. We did not convert amounts to 1995 dollars
in cases where it was not meaningful to do so or where doing so would
distort the presentation of the financial situation; this includes
our discussion of the legislated ceiling. USAID's claim that it may
ultimately collect 27 percent more than the program has cost (see p.
66, section 1(b)(4) of USAID's letter) reveals the flaws associated
with using nominal instead of 1995 dollars. USAID's calculation does
not take into account over $150 million in interest that the U.S.
government has foregone by using its funds to pay default costs for
this program--a real cost that has impacted the federal deficit.
Furthermore, portraying costs in only nominal dollars would inflate
our projection of future default costs to nearly $1.2 billion, which
would be misleading because it would not reflect the decreased
purchasing power of the dollar over time.
2. Section 221 of the Foreign Assistance Act sets forth the policy
statement for the Housing Guaranty Program. It declares that the
long-run goal of the program "should be to develop domestic
construction capabilities and to stimulate local credit institutions
to make available domestic capital and other management and
technological resources required for effective low-cost shelter
programs and policies." (See
22 U.S.C. 2181.) In 1978, the Congress amended sections 221 and 222
of the act to rewrite the policy statement to consolidate separate
housing guaranty programs into one program. The legislative history
indicates that "[t]he new policy statement does not represent a
change in focus of the program, but merely updates and clarifies the
purposes of the program . . . that the focus of the program should
be on improving the shelter facilities of the poor." (See H. Rpt.
No. 1087, 95th Cong., 2d Sess. 28 (1978).) The Committee included
language in the report reiterating that the long-run goal of the
program as set forth in section 221 remained unchanged. (See H.
Rpt. No. 1087 at 29.)
USAID suggests that there is a "disconnect" between the long-run goal
in section 221 and the activities emphasized in section 222.
According to USAID, the new activities shifted the focus away from
formal credit institutions and toward basic infrastructure needs.
However, the language of the Foreign Assistance Act, along with the
legislative history accompanying the changes made to sections 221 and
222 in 1978, indicates that, at the time, the Congress saw no
inherent contradiction between these two sections of the act.
Section 222 of the act sets forth the means "[t]o carry out the
policy of section 221." (See
22 U.S.C. 2182(a).) It identifies the types of activities that the
Congress intended to be emphasized in the program, including slum
upgrading, community facilities and services, and institution
building. These activities are not, in and of themselves, the
"long-run goals" of the program but simply the types of activities
the Congress intended to be emphasized to achieve the program's
goals. As noted above, the legislative history shows that in
prescribing the activities the Congress intended to retain the
long-run focus reflected in section 221.
USAID believes that the activities in section 222 will not, within
the time periods we considered, achieve the long-run goals of section
221. Nevertheless, some more positive indicators of progress should
be expected after 30 years (and 17 years since the enactment of
section 222(b)). We believe that USAID is accountable for either
meeting the long-run goals of the program or, alternatively,
reporting to the Congress within a reasonable period of time its
inability to meet them.
3. USAID's reference to two "separate and distinct" Housing Guaranty
Programs--pre- and post-Credit Reform--is an artificial distinction,
which obscures the program's overall financial condition and the
total U.S. government liability. While the financing of the program
has changed because of Credit Reform budgeting rules for new loan
guaranties beginning in fiscal year 1992, the operation of the
program has continued under the same authorizing legislation and
guidelines. In fact, as we point out, USAID continues to authorize
disbursement of pre-Credit Reform guarantied loans without any
reserve for defaults. While it is true that the post-Credit Reform
portion of the portfolio has experienced no defaults or write-offs,
the loans in this category represent a very small portion of the
program's portfolio and are currently in a grace period in which no
principal is due.
4. Contrary to USAID's assertion, neither the Office of Management
and Budget (OMB) nor the Congressional Budget Office have ever
challenged the validity of the method upon which we based our
projection of default costs. In fact, OMB and other executive
agencies recognized the validity of increasing the use of market
price to derive better assessments of default costs when assessing
sovereign risk--a key component of our methodology. However, OMB
indicated that estimating default costs for 167 countries using this
method would be overly burdensome. Moreover, our calculations are
comparable to those presented in USAID's most recent audited fiscal
year 1993 financial statements; both sets of calculations indicate
that the cost of the program is about $1 billion. USAID
mischaracterizes our projection of default costs and the one
contained in its own financial statements as "worst case scenarios."
In fact, both projections represent likely costs, based on two
independent assessments of country credit risks.
5. USAID dismisses our observations about the legislated ceiling,
indicating that this ceiling is no longer applicable to the program.
However, this ceiling does still apply to those loan guaranties
authorized prior to fiscal year 1992.
6. We recognize that, as USAID indicates, defaults and debt
rescheduling under this program occur within the larger macroeconomic
context of debt rescheduling that USAID describes. We continue to
believe that the program's poor financial condition has been
aggravated by (1) repeated rescheduling by program borrowers, (2)
USAID's provision of additional loan guaranties to rescheduled
borrowers, and (3) the waiver of administrative fees on rescheduled
debts. However, we have revised the language in our report relating
to USAID's influence on a country's decision to reschedule its debts.
7. USAID asserts that it has collected 95.6 percent of payments due
from rescheduled borrowers. However, this is not because these
borrowers have been repaying their debt to USAID. In fact, most of
these borrowers have rescheduled their payments on uncollected debt,
an average of five times each so far, repeatedly deferring repayment.
Thus, only about 5 percent of the debt has technically come due and
has been repaid.
8. Although, as USAID indicates, loans to previously rescheduled
borrowers amount to about 1 percent of all guarantied loans, the
rescheduled borrowers involved in these cases now collectively owe
USAID about $96 million, or 23 percent of all uncollected debt.
9. USAID claims that the program was never intended to be
self-sustaining. We do not assert that the program was intended to
be self-sustaining. However, as we indicated in our report, on a
number of occasions, presidential budget documents represented the
program to the Congress as being totally self-sustaining "consistent
with the intent of the Congress." Such representations were made even
up to 5 years after USAID first received a substantial reserve
account to cover missed loan payments. Furthermore, USAID's
assertion that it is owed and intends to collect 27 percent more than
the government's contribution seriously underestimates the cost of
this program to the U.S. government. It does not consider the
foregone interest on U.S. funds as we have done by converting dollar
amounts to 1995 dollars where appropriate (see comment 1).
10. The Credit Reform Act does not prohibit the recovery of
administrative costs through fees. Rather, it allows fees to be
applied against probable default costs, thereby reducing the
necessary subsidy appropriation and administrative costs to be funded
separately through appropriations. Thus, the act does not negate
USAID's authorization under the Foreign Assistance Act to collect
fees to cover program expenses. We acknowledge that it may be
difficult to recover all program costs and still charge affordable
fees, although, as our report indicates, USAID could raise its fees
and still be competitive with other international donors. We have
revised our report to recommend that program fees be increased to
cover a greater portion of program costs.
11. We defined a closely-linked cross subsidy as cash or assistance
provided to participants of a subsidized credit program that
decreases the likelihood that borrowers will default by either
increasing the likelihood of income for the borrowers or decreasing
the borrowers' costs. In the countries we visited, USAID's technical
assistance, assuming it achieved its objectives, clearly met this
definition. The World Bank recognizes the financial value of
technical assistance and often includes the cost in the principal
amount of its loans to developing countries.
12. USAID contends that its program overwhelmingly reaches the
targeted poor population in those countries where it operates.
However, USAID does not provide credible evidence to support its
contention. As we indicate in chapter 4, after analyzing USAID's
documentation, including that described in USAID's comments, we found
that the mechanisms USAID employed did not ensure and in some cases
were not intended to ensure that 90 percent of the program funds are
used to finance projects that benefit below-median income families.
13. USAID cites our earlier reports to support its position that it
has reached the target population. While we have in the past, as we
have in this report, acknowledged some of USAID's successes under
this program, we have also pointed out serious shortcomings,
including the potential inadequacy of the program's reserves and
difficulties reaching the target population. None of our earlier
findings contradict our current conclusions. Interestingly, USAID
also points to recent favorable statements we made on USAID's housing
sector reform project in Russia. While, as we reported, the program
achieved its reform objectives, it did so without the use of a
Housing Guaranty Program loan.
14. We believe the fact that host governments have invested in
projects intended to benefit below-median income families does not
relieve USAID from its responsibility for ensuring that the projects
it finances benefit the target population.
15. In several cases, USAID presented inaccurate and misleading
information to support its criticism of our report. For example,
with regard to the pay toilet for tourists in Indonesia, our
interviews with local workers at this site revealed that they did not
use this facility, despite USAID's claims to the contrary. USAID
also inaccurately indicates that only below-median income families
are eligible for Chile's subsidy program, when, in fact, there are no
income restrictions. We were unable to verify USAID's claim that 100
percent of the beneficiaries in the most recent program in Chile had
incomes below the median, because these funds were disbursed after
our field work in Chile was completed. At the time of our visit,
however, USAID officials indicated that they did not intend to check
family income for this program.
16. USAID provides statistics and other information contending that
its programs have stimulated private investment we did not consider.
We had already reviewed virtually all of the data USAID provided and
found that it is not germane to the conclusions of our report. For
example, USAID cites statistics on investments in India, Indonesia,
and Ecuador which benefit essentially upper-income not low-income
families. USAID also describes a $10 million Housing Guaranty
Program investment in Chile that has produced over $45 million worth
of housing. As we explain in our report, this does not represent an
increase in long-term private investment for housing but rather
private construction of housing largely financed by the government of
Chile. In addition, USAID refers to shelter sector developments that
cannot be directly attributed to Housing Guaranty Program activities.
For example, increased private shelter construction in Chile is
driven by that country's government-funded housing subsidy program,
for which USAID cannot take credit.
17. We revised our report to reflect information provided on
short-term construction financing raised from private sources in
Tunisia.
18. USAID criticizes our review for overlooking private investment
by homeowners. We acknowledge that formal investment in shelter can
have a multiplier effect on homeowner consumption and the economy in
general, as USAID's data illustrate. In fact, the program is
predicated on this fact. However, we focused on investment from
credit institutions and entrepreneurs, as this is the type of
investment the program was specifically intended to stimulate,
according to the Foreign Assistance Act. Such investment would
sustain and expand the multiplier effect USAID describes. Also,
since USAID has not ensured that 90 percent of the program
beneficiaries have below-median incomes, it likewise cannot ensure
that all of the homeowner investments it cites benefit the target
population.
19. USAID generally acknowledges that improvements are needed in its
performance measurement system but criticizes our review for having
overlooked pertinent information on project level indicators and
monitoring activities. Our review was specifically aimed at
reviewing all of USAID's monitoring and evaluation activities, and,
accordingly, we conducted an exhaustive review of the agency's
records on this subject. In addition, despite our extensive file
review, we specifically requested the program staff to provide any
pertinent documents we may have missed. We found that the data USAID
provided generally did not demonstrate achievement of the program's
private investment goal or the assurance that projects suitable for
below-median income families actually benefited those families.
MAJOR CONTRIBUTORS TO THIS REPORT
========================================================= Appendix III
NATIONAL SECURITY AND
INTERNATIONAL AFFAIRS DIVISION,
WASHINGTON, D.C.
------------------------------------------------------- Appendix III:1
Benjamin F. Nelson
Ronald A. Kushner
Caroline V. Harper
James B. Michels
EUROPEAN OFFICE
------------------------------------------------------- Appendix III:2
Bettye J. Caton
Neyla Arnas
FAR EAST OFFICE
------------------------------------------------------- Appendix III:3
Kathleen M. Monahan
Kimberly M. Gianopoulos
OFFICE OF GENERAL COUNSEL,
WASHINGTON, D.C.
------------------------------------------------------- Appendix III:4
Mark C. Speight
ACCOUNTING AND INFORMATION
MANAGEMENT DIVISION,
WASHINGTON, D.C.
------------------------------------------------------- Appendix III:5
Roger R. Stoltz
RELATED GAO PRODUCTS
============================================================ Chapter 1
General Counsel Opinion (GAO/OGC-93-2, Nov. 3, 1992).
Israel: U.S. Guaranties for Immigrant Absorption (GAO/NSIAD-92-119,
Feb. 12, 1992).
Foreign Assistance: The Solanda Housing Guaranty Project in Ecuador
(GAO/NSIAD-86-120, May 1986).
AID's Management of the Housing Guaranty Program (GAO/NSIAD-84-75,
Apr. 25, 1984).
Agency for International Development's Housing Investment Guaranty
Program (ID-78-44, Sept. 6, 1978).
The Challenge of Meeting Shelter Needs in Less Developed Countries
(ID-77-39, Nov. 4, 1977).
Low-Income Groups Not Helped by Agency for International
Development's Housing Investment Guaranty Program (B-171526, Nov.
25, 1974).