Credit Reform: Key Credit Agencies Had Difficulty Making Reasonable Loan
Program Cost Estimates (Letter Report, 01/29/99, GAO/AIMD-99-31).

GAO's report on the government's consolidated financial statements for
fiscal year 1997 (GAO/AIMD-98-127, Mar. 1998) raised significant
concerns about the ability of credit agencies to reasonably estimate
subsidy costs related to the reported $216.6 billion in direct loans and
$712.4 billion in loan guarantees issued by the federal government.
Program managers and Congress depend on this information to make funding
and program decisions involving hundreds of billions of dollars each
year. For some types of credit programs, unreliable information can
affect the availability and the delivery of basic services to taxpayers
because changes in cost estimates may alter the number and amount of
loans available. GAO reviewed the ability of the five key credit
agencies--the Small Business Administration and the Departments of
Education, Housing and Urban Development, Veterans Affairs, and
Agriculture--to reasonably estimate their loan programs, including
whether they used practices identified by the Credit Reform Task Force
as being effective in making these estimates. GAO also reviewed the
status of agencies' efforts to ensure that computer systems used to
estimate the cost of credit programs are  Year 2000 compliant.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  AIMD-99-31
     TITLE:  Credit Reform: Key Credit Agencies Had Difficulty Making 
             Reasonable Loan Program Cost Estimates
      DATE:  01/29/99
   SUBJECT:  Projections
             Financial statement audits
             Loan accounting systems
             Internal controls
             Accounting standards
             Government guaranteed loans
             Data integrity
             Direct loans
IDENTIFIER:  Mutual Mortgage Insurance Fund
             Dept. of Education National Student Loan Data System
             Y2K
             USDA Rural Housing Service Single Family Housing Program
             Farm Operating Loan Program
             VA Direct Loan Guaranty Loan Program
             William D. Ford Federal Direct Loan Program
             SBA Disaster Loan Program
             SBA 7(a) General Business Loan Program
             Federal Family Education Loan Program
             HUD General and Special Risk Insurance Fund
             VA Guaranty and Indemnity Fund
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to Congressional Requesters

January 1999

CREDIT REFORM - KEY CREDIT
AGENCIES HAD DIFFICULTY MAKING
REASONABLE LOAN PROGRAM COST
ESTIMATES

GAO/AIMD-99-31

Credit Reform

(919131)


Abbreviations
=============================================================== ABBREV

  BLM - Budget Loan Model
  FASAB - Federal Accounting Standards Advisory Board
  FCRA - Federal Credit Reform Act
  FHA - Federal Housing Administration
  FSA - Farm Service Agency
  GAAP - Generally Accepted Accounting Principles
  HUD - Department of Housing and Urban Development
  IG - Inspector General
  MMI - Mutual Mortgage Insurance
  NASS - National Agricultural Statistical Service
  NSLDS - National Student Loan Data System
  OMB - Office of Management and Budget
  RD - Rural Development
  SBA - Small Business Administration
  SFFAS - Statement of Federal Financial Accounting Standards
  USDA - Department of Agriculture
  VA - Department of Veterans Affairs

Letter
=============================================================== LETTER


B-281592

January 29, 1999

The Honorable John R.  Kasich
Chairman, Committee on the Budget
House of Representatives

The Honorable Steve Horn
Committee on Government Reform
House of Representatives

Our report on the fiscal year 1997 consolidated financial statements
of the federal government\1 raised major concerns about the ability
of credit agencies' to reasonably estimate, for both financial
statement and budgetary purposes, subsidy costs related to the
reported $216.6 billion in direct loans and $712.4 billion in loan
guarantees issued by the federal government.\2 Providing reasonable
estimates based on reliable data is critical to effective program
stewardship and accountability.  Program managers and the Congress
rely on this information to make funding and programmatic decisions
involving hundreds of billions of dollars annually.  For some types
of credit programs, unreliable information can affect the
availability and the delivery of basic program services to taxpayers
because changes in cost estimates may alter the number and amount of
loans available. 

To gain an understanding of the key issues impeding reasonable
estimates of subsidy costs of credit programs for the five key credit
agencies, the Small Business Administration (SBA) and the Departments
of Education, Housing and Urban Development (HUD), Veterans Affairs
(VA), and Agriculture (USDA),\3 we reviewed these agencies' abilities
to reasonably estimate the cost of their loan programs, including
whether they used practices identified by the Credit Reform Task
Force\4 as being effective in making these estimates.  We also
reviewed the status of agencies' efforts to ensure that computer
systems used to estimate the cost of credit programs are Year 2000
compliant. 


--------------------
\1 Financial Audit:  1997 Consolidated Financial Statements of the
United States Government (GAO/AIMD-98-127, March 31, 1998). 

\2 Fiscal year 1997 financial data was the most recent information
available and, except where noted, is used throughout the report. 

\3 These agencies have the largest domestic federal credit programs
and accounted for 74 percent of the government's outstanding direct
loans and 94 percent of its outstanding guaranteed loans outstanding
as of September 30, 1997. 

\4 The Credit Reform Task Force, formerly known as the Subgroup on
Credit Reform of the Governmentwide Audited Financial Statements Task
Force (established to study accounting and auditing issues related to
credit reform implementation in preparation for the first audit of
the governmentwide consolidated financial statements and to provide
guidance to agencies to resolve these issues), is now a task force of
the Accounting and Auditing Policy Committee sponsored by the Federal
Accounting Standards Advisory Board (FASAB).  This task force
developed a Technical Release, Preparing and Auditing Direct Loan and
Loan Guarantee Subsidies Under the Credit Reform Act, which has been
approved by FASAB and is expected to be issued by the Office of
Management and Budget as authoritative guidance during fiscal year
1999. 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

While federal credit agencies have been required to estimate the cost
of their loan programs in accordance with the requirements of the
Federal Credit Reform Act of 1990 (FCRA) and federal accounting
standards since fiscal years 1992 and 1994, respectively, only two of
the five major credit agencies we reviewed made reasonable loan
program cost estimates in their fiscal year 1997 financial
statements.  In response to our fiscal year 1997 governmentwide
consolidated financial statement report, the Office of Management and
Budget (OMB) directed agencies that did not receive an unqualified
opinion on their financial statements to develop action plans to
address identified financial management weaknesses.  Additionally,
some agencies that received unqualified audit opinions are also
acting to resolve issues with their estimates of loan program costs,
based on recommendations made during financial statement audits. 

The problems agencies faced in making credit subsidy estimates as
required by FCRA and federal accounting standards stemmed largely
from their lack of (1) reliable historical data upon which to base
estimates of future loan performance, (2) adequate systems that have
the capability to track the required information, (3) sound cash flow
models, and/or (4) appropriate policies and procedures for ensuring
the accuracy of data used to generate the estimates.  Additionally,
VA had fundamental problems in maintaining accountability over its
loan portfolio.  These issues, which affected the agencies in varying
degrees, impeded them from making reasonable loan program cost
estimates for fiscal year 1997 for financial statement and/or
budgetary purposes. 

The financial statement audits of all five of these agencies were
instrumental in providing the basis for recommending changes to the
agencies' loan cost estimation processes that could greatly improve
the estimates.  Progress towards addressing cost estimation issues
also varies by agency.  For example, SBA has taken steps to address
the identified weaknesses in its cost estimation process, while USDA
has a number of remaining problems that need resolution.  Until each
agency fully addresses the identified weaknesses in its cost
estimation process, the credibility of loan program cost information
submitted by these agencies for both financial statement and
budgetary purposes will continue to be questionable. 

SBA was one of the two agencies able to make reasonable estimates of
the cost of its loan programs in its fiscal year 1997 financial
statements, primarily because the agency maintained reliable records
of historical loan performance data.  However, for the two programs
we reviewed, SBA made significant errors in initially calculating its
reestimates of loan program costs.  These errors, which were
identified by SBA's independent public accountant, were adjusted for
in SBA's draft financial statements, thereby allowing for an
unqualified audit opinion on those statements.  However, the errors
were not identified in time to be corrected in SBA's fiscal year 1999
budget submission to OMB nor in the President's budget submission to
the Congress.  SBA has recently adopted a number of cost estimation
practices that should help ensure that such errors are detected and
corrected promptly and that SBA's budgetary and financial estimates
of loan program costs are reasonable. 

The Department of Education was able to prepare reasonable credit
program estimates for its fiscal year 1997 financial statements based
on information obtained through a significant data gathering effort
from its guaranty agencies.\5

However, the audited estimates differed significantly from the
estimates based on data from Education's own database, which raises
questions about the validity of Education's database.  Further,
Education used its questionable internal data for its budget
submission to OMB.  Until the agency verifies that its internal
database contains accurate data, Education will be required to expend
an inordinate amount of time and resources gathering the necessary
information to make reasonable loan program cost estimates. 
Education has efforts underway to address the challenges it faces in
preparing reasonable loan cost estimates, including continuing
efforts to assess the accuracy and completeness of its database. 

HUD was unable to provide adequate supporting data for its fiscal
year 1997 financial statement estimates of its loan program costs,
which resulted in a qualified audit opinion\6 from HUD's Inspector
General (IG) on those financial statements.\7 This lack of supporting
data also raises questions about the integrity of loan program cost
information submitted for budgetary purposes.  During fiscal year
1998, HUD focused significant attention on gathering the necessary
data to prepare reasonable estimates of loan program costs.  These
data are currently undergoing an audit which, once completed, will
determine the reasonableness of HUD's revised loan program cost
estimates.  Additionally, HUD has developed an action plan to address
identified financial management issues related to the loan cost
estimation process.  This plan, if fully implemented, should help HUD
prepare reasonable estimates of loan program costs. 

During fiscal year 1997, VA faced significant problems performing
routine accounting for its loan programs including loss of
accountability over certain loans transferred to an outside servicer. 
These problems also hindered VA's ability to make reasonable
financial statement estimates of its credit program costs and
resulted in a qualified audit opinion on VA's fiscal year 1997
financial statements.  Since budgetary estimates ought to be based in
part on accounting data, the reliability of loan program cost
information submitted for budgetary purposes is questionable.  In
addition, we determined that VA initially did not calculate or record
the cost of its guarantee obligations on loans it sold to investors. 
In an attempt to correct this and other errors in its fiscal year
1997 draft financial statements, VA, in consultation with OMB,
estimated and recorded an additional $376 million expense as part of
an aggregate adjustment for future losses and the related liability
for the loans sold between 1992 and 1997.  VA has prepared an action
plan that addresses some of its financial management problems.  The
plan focuses on resolving the fundamental accounting problems for the
loan sales program and the basic data integrity issues related to the
incomplete loan inventory data maintained by the servicer. 

During fiscal year 1997, USDA was unable to make reasonable financial
statement estimates of its loan programs' costs because it had not
maintained the necessary historical data to reasonably estimate
future loan performance and continued to use computer systems that
were not appropriately configured to capture the data necessary to
make such estimates.  These long-standing problems contributed to the
auditor's inability to give an opinion on USDA's fiscal year 1997
consolidated financial statements.  These negative audit findings
also raise concerns over the integrity of USDA's loan program
budgetary data.  Although USDA has developed an action plan to
address deficiencies in its loan estimation process, the plan does
not include several critical components necessary for USDA to resolve
these problems. 

The five key credit agencies also face the challenge of addressing
the Year 2000 problem related to systems used in the loan cost
estimation process.  According to agency officials, for the 10 loan
programs we reviewed, all of the systems that provide key cash flow
data have been identified as "mission-critical" and are either
currently Year 2000 compliant or are scheduled to be compliant by
March 31, 1999.  However, in its Quarterly Report:  Progress on Year
2000 Conversion, as of mid-November 1998, OMB (1) expressed concerns
related to progress on Year 2000 conversion at Education and USDA,
(2) noted that HUD and VA appear to be making satisfactory progress
towards being Year 2000 compliant, and (3) pointed out that SBA was
the first agency to report that all of its mission-critical systems
were Year 2000 compliant.  In previous reports and testimonies we
have also raised issues over the status of Year 2000 conversion
efforts at the five key credit agencies, except for SBA. 
Additionally, according to agency officials, they have all begun
business continuity and contingency planning to help maintain core
business processes in the event of any Year 2000-induced disruptions. 


--------------------
\5 These agencies serve as intermediaries between the government and
the lender.  They are responsible for reviewing student applications
and approving loans, reviewing and paying claims to lenders when
defaults occur, and collecting on defaulted loans. 

\6 A qualified opinion on financial statements is issued when the
auditor believes that a particular aspect of the statements may be
misstated.  The qualification may result from a nonpervasive
departure from federal accounting standards or the inability of an
auditor to obtain audit satisfaction on a component of the financial
statements. 

\7 FHA, a major component of HUD, received an unqualified opinion on
its fiscal year 1997 financial statements prepared using private
sector generally accepted accounting standards (GAAP).  However, the
reported amounts related to FHA's loan programs using federal
accounting standards would be significantly different than those
reported under GAAP. 


   BACKGROUND
------------------------------------------------------------ Letter :2

The federal government uses direct loans and loan guarantees as tools
to achieve numerous program objectives, such as assistance for
housing, farming, education, small businesses, and foreign
governments.  At the end of fiscal year 1997, the Department of the
Treasury reported that the federal government's gross direct loans
outstanding totaled $216.6 billion, and loan guarantees outstanding
totaled $712.4 billion. 

Before enactment of the Federal Credit Reform Act of 1990, credit
programs--like most other federal programs--were recorded in
budgetary accounts on a cash basis.  While this basis reflected cash
flows, it distorted the timing of when costs would actually be
recognized and, thus, distorted the comparability of credit program
costs with other programs intended to achieve similar purposes, such
as grants.  For example, for direct loans, the budget generally
showed budget authority and outlays for loans disbursed that exceeded
repayments received from all past loans in that year.  Therefore, in
the budget a direct loan in the first year of a program was
equivalent to the cost of a grant.  Cash-basis budgetary recording
also suggested a bias in favor of loan guarantees over direct loans. 
Loan guarantees appeared to be free in the short-term because
cash-basis recording did not recognize that some loan guarantees
result in costs due to default of the underlying loans. 

FCRA changed the budgetary treatment of credit programs beginning
with fiscal year 1992 so that their costs could be compared more
appropriately with each other and with the costs of other federal
spending.  FCRA requires that agencies have budget authority to cover
the program's cost to the government in advance, before new direct
loan obligations are incurred and new loan guarantee commitments are
made.  The act therefore requires agencies to estimate the cost of
extending or guaranteeing credit, called the subsidy cost.\8 This
cost is the present value\9 of disbursements--over the life of the
loan--by the government (loan disbursements and other payments) minus
estimated payments to the government (repayments of principal,
payments of interest, other recoveries, and other payments).  For
loan guarantees, the subsidy cost is the present value of cash flows
from estimated payments by the government (for defaults and
delinquencies, interest rate subsidies, and other payments) minus
estimated payments to the government (for loan origination and other
fees, penalties, and recoveries). 

FCRA assigned to OMB the responsibility to coordinate the cost
estimates required by the act.  OMB is authorized to delegate to
lending agencies the authority to estimate costs, based on written
guidelines issued by OMB.  These guidelines are contained in sections
33.1 through 33.12 of OMB Circular No.  A-11, and supporting
exhibits.\10

The Federal Accounting Standards Advisory Board (FASAB)\11 developed
the accounting standard for credit programs, Statement of Federal
Financial Accounting Standards No.  2, Accounting for Direct Loans
and Loan Guarantees (SFFAS No.  2), which became effective with
fiscal year 1994.  This standard, which generally mirrors FCRA,
established guidance for estimating the cost of direct and guaranteed
loan programs, as well as for recording direct loans and the
liability for loan guarantees for financial reporting purposes. 
SFFAS No.  2 states that the actual and expected costs of federal
credit programs should be fully recognized in both budgetary and
financial reporting.  To accomplish this, agencies first predict or
estimate the future performance of direct and guaranteed loans when
preparing their annual budgets.  The data used for these budgetary
estimates are generally reestimated after the fiscal year end to
reflect any changes in actual loan performance since the budget was
prepared, as well as any expected changes in assumptions and future
loan performance.  This reestimated data is then used to report the
cost of the loans disbursed under the direct or guaranteed loan
program as a "Program Cost" on the agencies' Statement of Net Costs
after loans are disbursed. 

Agency management is responsible for accumulating sufficient,
relevant, and reliable data on which to base the estimates.  Further,
SFFAS No.  2 states that agencies should use the historical
experience of the loan programs when estimating future loan
performance.  To accomplish this, agencies use cash flow models based
on various assumptions, often referred to as cash flow assumptions,
such as the number and amount of loans that will default in a given
year--known as the default assumption.  Those assumptions that have
the greatest impact on the credit subsidy vary by program and are
often referred to as key cash flow assumptions. 

Statement on Auditing Standards No.  57 states that auditors should
evaluate the reasonableness of estimates in the context of the
financial statements taken as a whole.  As part of the annual
financial statement audits, agency cash flow models and assumptions
are assessed to determine if management has a reliable basis for its
credit subsidy estimates. 

In 1997, the Credit Reform Task Force of the Accounting and Auditing
Policy Committee\12 was formed in order to address key issues
surrounding the implementation of FCRA and the related federal
accounting standard.  This task force developed a Technical Release,
Preparing and Auditing Direct Loan and Loan Guarantee Subsidies Under
the Credit Reform Act, which has been approved by FASAB and is
expected to be issued by OMB during fiscal year 1999.  This Technical
Release identifies specific practices that, if fully implemented by
credit agencies, will enhance their ability to reasonably estimate
loan program costs.\13 These practices include the following: 

  -- Accumulating sufficient, relevant, and reliable supporting data
     that provides a reliable basis for agencies' estimates of future
     loan performance.  For example, to make reasonable projections
     of future loan defaults, recoveries, prepayments, or other key
     cash flows, agencies should use reliable records of historical
     experience and take into consideration current and forecasted
     economic conditions. 

  -- Conducting periodic comparisons of estimated loan performance to
     actual cash flows in the accounting system.  This comparison
     allows agencies to identify and research significant differences
     and determine whether assumptions related to expected future
     loan performance need to be revised. 

  -- Calculating timely reestimates, based on the most recently
     available data, of the loan program's cost and including the
     reestimates\14 in the current year's financial statements and
     budget submissions.  By performing timely reestimates, agencies
     are including their best estimate of a loan program's cost in
     the agency's financial statements and budget submissions. 

  -- Comparing cash flow models to legislatively mandated program
     requirements to ensure that current cash flow models reasonably
     represent the cash flows of the loan program based on the laws
     and regulations that govern them. 

  -- Coordinating estimates of loan program cost among the budget,
     accounting, and program staff.  These officials should work
     together to ensure that various practices, including those
     described above, are implemented and operating effectively and
     that all key assumptions have been coordinated and reviewed by
     the budget, accounting, and program offices. 

  -- Performing sensitivity analyses to identify which cash flow
     assumptions, such as defaults, recoveries, or prepayments, have
     the greatest impact on the cost of the loan program.  Knowledge
     of these key assumptions provides management with the ability to
     monitor the economic trends that most affect the loan program's
     performance.  These analyses also allow agencies to more
     efficiently focus their efforts on providing support for the key
     assumptions, which need to be documented to pass the test of an
     independent audit. 

  -- Ensuring that agency cash flow models are well organized,
     documented, and, to reduce the chance of errors, require minimal
     data entry.  This documentation should include the rationale for
     using the specific model, the mechanics of the model, including
     formulas and other mathematical functions, and sources of
     supporting data. 

  -- Establishing formal policies and procedures for calculating
     estimates of loan program cost, including a formal review
     process.  Documented policies and procedures, as well as a
     formal review process, are important internal controls that are
     designed to help ensure continuity when there is employee
     turnover and to calculate reasonable, well-supported cost
     estimates. 

During the summer of 1998, in response to our report on the fiscal
year 1997 governmentwide consolidated financial statement audit, OMB
directed agencies that did not receive an unqualified opinion on
their financial statements to develop action plans to address
identified financial management weaknesses.  As a result, three of
the five agencies in our review, HUD, VA, and USDA, prepared action
plans to address, among other things, problems with preparing
reasonable estimates of their loan program costs.  Because SBA and
Education received unqualified opinions on their fiscal year 1997
financial statements, these agencies were not required to, and did
not, prepare formal action plans. 

In a March 1998 report\15 on credit reform estimation problems, we
indicated that the five key credit agencies had problems estimating
the subsidy cost of credit programs.  During this prior review, we
examined data for the same 10 programs (listed in the "Objectives,
Scope, and Methodology" section) discussed in this report to identify
trends and causes for the changes in subsidy estimates.  The
resulting report noted that the lack of timely reestimates, as well
as the frequent absence of documentation and reliable information,
limited the ability of agency management, OMB, and the Congress to
exercise intended oversight.  The report contained broad
recommendations for improving oversight of credit reform
implementation including ensuring that (1) estimates are prepared
accurately and (2) documentation supporting subsidy estimates
included in the budget and financial statements is prepared and
retained.  Appendix I provides additional background information on
estimating the cost of credit programs. 


--------------------
\8 In accordance with the Federal Credit Reform Act of 1990, the
subsidy cost of direct and guaranteed loans does not include
administrative costs of the program. 

\9 Present value is the value today of a stream of payments in the
future discounted at a certain interest rate.  For the period we
reviewed, when calculating the present value of loan subsidy costs,
agencies must use as the discount rate the average annual interest
rate for marketable U.S.  Treasury securities with similar maturities
to the loan or guarantee. 

\10 The act requires OMB to coordinate with the Congressional Budget
Office in developing estimation guidelines. 

\11 FASAB was created by OMB, Treasury, and GAO to develop and
recommend accounting principles for the federal government.  These
three agencies approved Statement of Federal Financial Accounting
Standards (SFFAS) No.  2, Accounting for Direct Loans and Loan
Guarantees, in July 1993. 

\12 See footnote 4. 

\13 The Technical Release contains other practices of a technical
nature that we did not include in this discussion. 

\14 OMB Circular A-11 requires budgetary reestimates greater than $1
million or 5 percent of the pre-reestimate cost estimate to be
included in the budget submission.  Lesser reestimates may be
reported cumulatively with future reestimates.  However, for
financial statement purposes, agencies are required to include
reestimates that are significant to the financial statements. 

\15 Credit Reform:  Greater Effort Needed to Overcome Persistent Cost
Estimation Problems (GAO/AIMD-98-14, March 30, 1998). 


   OBJECTIVES, SCOPE, AND
   METHODOLOGY
------------------------------------------------------------ Letter :3

Our objectives were to assess (1) the ability of agencies' to
reasonably estimate the cost of their loan programs, including
whether they used practices identified in the Credit Reform Task
Force's Technical Release as being effective in making these
estimates and (2) the status of agencies' efforts to ensure that
computer systems used to estimate the cost of credit programs are
Year 2000 compliant.  We selected a sample of 10 programs--5 direct
loan programs totaling $52.1 billion and 5 guaranteed loan programs
totaling $558.1 billion--from the five agencies with the largest
domestic federal credit programs:  the Small Business Administration
and the Departments of Education, Housing and Urban Development,
Veterans Affairs, and Agriculture.  We generally selected programs
that had the most credit outstanding or highest loan levels at each
agency.  Specifically, these programs were the following: 

  -- 7(a) General Business Loans Program and Disaster Loan Program,
     which totaled 72 percent of SBA's loan guarantees and 73 percent
     of its direct loans, respectively;

  -- Federal Family Education Loan Program and William D.  Ford
     Direct Loan Program, which totaled 100 percent of Education's
     loan guarantees and 50 percent of its total loans receivable,
     respectively;

  -- Mutual Mortgage Insurance Fund and General and Special Risk
     Insurance Fund Section 223(f) Refinance, which totaled 81
     percent of HUD's loan guarantees;

  -- Guaranty and Indemnity Fund and the Loan Guaranty Direct Loan
     Program, which totaled 100 percent of VA's post credit reform
     loan guarantees and 69 percent of its total loans receivable,
     respectively; and

  -- Farm Service Agency Farm Operating Loans Program and Rural
     Housing Service Single Family Housing Program, which totaled 20
     percent\16 of USDA's direct loans. 

Generally, to accomplish these objectives, we evaluated the process
the agencies used to estimate the cost of their loan programs during
fiscal year 1997, including whether the agencies used practices
outlined in the Credit Reform Task Force's Technical Release that
enhanced their ability to reasonably estimate loan program costs.  In
addition, we determined whether agencies had a reliable basis for the
underlying assumptions for their estimates of loan program
performance by assessing the support for key cash flow assumptions. 
We used the financial statement audit work of the respective agency
auditors as a starting point for our analyses.  Further, we obtained
information on the status of agencies' efforts to ensure that
computer systems used to estimate the cost of credit programs are
Year 2000 compliant. 

Our work was conducted in Washington, D.C., and St.  Louis, Missouri,
from September 1997 to November 1998 in accordance with generally
accepted government auditing standards.  We requested written
comments on a draft of this report from the following officials or
their designees:  the Administrator of Small Business and the
Secretaries of Education, Housing and Urban Development, Veterans
Affairs, and Agriculture.  All of the entities provided written
comments, which are discussed in the respective "Agency Comments and
Our Evaluation" sections of this report and are reprinted in
appendixes III through VII.  Further details of our objectives,
scope, and methodology are in appendix II. 


--------------------
\16 At the end of fiscal year 1997, USDA reported that $79.3 billion,
nearly 78 percent of the direct loan portfolio, was either disbursed
prior to the implementation of the Federal Credit Reform Act in
fiscal year 1992 or consisted of price support loans that were
excluded by Section 502(1) of the Federal Credit Reform Act.  Thus,
we focused on these two programs because they were the domestic loan
programs subject to credit reform that had disbursed the largest
volume of loans in less than 4 years. 


   NOT ALL KEY CREDIT AGENCIES
   COULD MAKE REASONABLE ESTIMATES
   OF LOAN PROGRAM COSTS
------------------------------------------------------------ Letter :4

For the 10 credit programs at the five key credit agencies we
reviewed, only SBA and Education were able to reasonably estimate the
cost of their credit programs for financial reporting purposes and
received unqualified opinions on their fiscal year 1997 financial
statements.  However, the data that Education used to prepare its
budget estimates, which were different from the data used to prepare
its financial statements, had not been validated.  Further, SBA made
errors in the reestimate it submitted for budget purposes.  HUD, VA,
and USDA were not able to prepare reasonable estimates, which
contributed to their qualified opinions or disclaimers\17 of opinion
on their fiscal year 1997 financial statements.  These problems also
call into question the reliability of the loan program data these
agencies submitted to the Congress for future budget decisions.  HUD,
VA, and USDA have prepared action plans to correct some of their loan
cost estimation problems.  In addition, during 1998, HUD focused
considerable effort on making reasonable cost estimates of its loan
programs.  Further, while not required to prepare formal action
plans, both SBA and Education have planned or acted to correct
deficiencies in their loan estimation process. 


--------------------
\17 A qualified opinion on financial statements is issued when the
auditor believes that a particular aspect of the statements may be
misstated.  The qualification may result from a nonpervasive
departure from federal accounting standards or the inability of an
auditor to obtain audit satisfaction on a component of the financial
statements.  A disclaimer may arise because of a severe limitation in
the scope of the audit, perhaps due to the lack of documentation
and/or uncertainties about the amount of an item, or the outcome of a
matter that significantly affects financial position. 


      SBA PREPARED REASONABLE COST
      ESTIMATES BUT HAD WEAKNESSES
      IN ITS REESTIMATE PROCESS
---------------------------------------------------------- Letter :4.1

SBA based its fiscal year 1997 estimates of loan program costs on
reliable records of historical loan performance data and was
therefore able to make a reasonable estimate of the cost of these
programs on its fiscal year 1997 financial statements.  However, for
the two programs we reviewed, SBA initially made errors in the
reestimates of its loan program's costs, which its independent public
accountant uncovered, including using incorrect discount rates, which
required large adjustments to the draft financial statements.  As a
result of making these adjustments, SBA received an unqualified
opinion on its fiscal year 1997 financial statements.  However,
because the fiscal year 1997 budgetary reestimate was included with
the fiscal year 1999 President's budget prior to the audit
adjustments, the budgetary reestimate contained erroneous data. 

Since the inception of credit reform in 1992, SBA has placed
significant emphasis on gathering reliable key cash flow data and,
with OMB's assistance, developed sophisticated cash flow models to
estimate future loan performance and cost.  Beginning in 1992, SBA
devoted considerable resources to evaluating its existing financial
management systems and determining what modifications would be
necessary to allow it to reasonably estimate loan program costs under
credit reform.  Since these initial efforts, SBA has continued to
further refine its estimates of loan program costs and the related
underlying assumptions.  For example, during 1997, SBA hired
consultants to study and develop refined loss and recovery estimates
for the Disaster Loan Program. 

SBA followed a number of practices that enhanced its ability to make
reasonable financial statement and budgetary estimates of loan
program costs for the two programs we reviewed.  For example, SBA
developed an extensive database of historical cash flow information,
which provided a reliable basis for its estimates of credit program
costs.  Further, SBA, with assistance from OMB, established
sophisticated, well organized cash flow models that, when compared
with actual historical data, reasonably estimated future loan
performance.  Because of this database and SBA's sophisticated cash
flow model, SBA was able to calculate reasonable estimates of loan
program costs without significant manual intervention or requests for
data from outside entities.  SBA also routinely compared estimated
loan performance to actual costs recorded in the accounting system to
assess the reasonableness of its estimates of future loan performance
and costs.  Finally, in preparing their estimates for the two
programs we reviewed, individuals from SBA's program, budget, and
accounting offices coordinated their work. 

However, during the audit of SBA's fiscal year 1997 financial
statements, the independent public accountants identified material
internal control weaknesses\18

related to estimating the cost of credit programs.  Specifically, the
independent public accountant reported that incorrect data, including
discount rates, were used in the 1997 reestimate, and errors existed
in some of the reestimated cash flow models.  In aggregate, these
errors resulted in SBA recording over $221 million in adjustments to
its financial statements, which enabled the independent public
accountant to render an unqualified opinion.  However, these
adjustments were not identified until after SBA submitted its fiscal
year 1999 budget to OMB and, as a result, this budget submission and
the President's budget misstated the cost of these loan programs. 
Further, the independent public accountant reported that SBA lacked
adequate internal controls over the estimation process.  For example,
SBA did not retain the cash flow models for one of the programs we
reviewed for fiscal years 1992 through 1997.  Although these models
are normally a part of performing the reestimates, and should be
retained as a matter of routine record-keeping, SBA was able to
calculate a reasonable financial statement reestimate by using a more
recent cash flow model. 

Because SBA received an unqualified audit opinion on its fiscal year
1997 financial statements, OMB did not require SBA to prepare a
formal action plan to address the weaknesses identified in the
estimation process.  However, in the audit report on the fiscal year
1997 financial statements, the independent public accountant made
recommendations, with which we concur, which addressed the material
internal control weaknesses described above including developing
formal policies and procedures for estimating the cost of credit
programs and implementing a formal supervisory review process to
identify and correct potential errors.  In response to these
recommendations, SBA developed and implemented formal policies and
procedures including a formal supervisory review process.  In
addition, SBA adopted other practices, such as performing sensitivity
analyses and calculating its fiscal year 1998 reestimate earlier than
the fiscal year 1997 reestimates, to allow sufficient time to include
any necessary adjustments resulting from the audit in the data
presented in the President's budget.  These actions should help SBA
ensure that future errors are detected and corrected promptly and
that budgetary and financial estimates of loan program costs are
reasonable. 


--------------------
\18 Internal controls provide the framework for the accomplishment of
management objectives, accurate financial reporting, and compliance
with laws and regulations.  When estimating the cost of loans,
effective internal controls serve as checks and balances to help
ensure the reasonableness of the estimate.  Material weaknesses in
internal controls are significant deficiencies in the agency's
internal controls, which could significantly affect the financial
statements or a performance measure and not be detected promptly by
the agency. 


         AGENCY COMMENTS AND OUR
         EVALUATION
-------------------------------------------------------- Letter :4.1.1

SBA agreed with the findings in this report.  (SBA's comments are
reprinted in appendix III.)


      EDUCATION'S FINANCIAL
      STATEMENT ESTIMATES WERE
      REASONABLE, HOWEVER BUDGET
      ESTIMATES WERE QUESTIONABLE
---------------------------------------------------------- Letter :4.2

The Department of Education was able to prepare reasonable credit
program estimates for its fiscal year 1997 financial statements,
based on information obtained through a significant data gathering
effort from its guaranty agencies.\19

However, the audited estimates differed materially from the credit
subsidy estimates based on Education's own database, which raises
questions about the validity of Education's database.  Further,
Education's credit program estimates for its budget submission for
these two programs were based on the questionable data from its own
database.  Until the information in the existing database is
determined to be reliable, Education will continue to expend
considerable time and resources in order to make reasonable loan
program cost estimates. 

While the IG concluded that Education's fiscal year 1997 financial
statement estimates were reasonable, several internal control
weaknesses were reported.  For example, the IG reported that
Education needed to establish the validity of its principal database,
the National Student Loan Data System (NSLDS),\20 to provide a basis
for preparing reliable loan estimates and to establish sufficient
controls to detect material errors in its loan estimates.  Data from
NSLDS were used to prepare the fiscal year 1997 budgetary estimates
because Education's staff believed that the data were reliable. 
However, the weaknesses found in Education's internal controls raise
questions about the quality of this database. 

Based on Education's continuing validation efforts, it believed that
estimates based on data from NSLDS would be similar to estimates
based on data received from the guaranty agencies.  Education
therefore planned to use the guaranty agencies' data to validate the
estimates that were based on NSLDS data.  As part of this plan,
Education prepared two estimates for financial statement
purposes--one based on data from NSLDS and the other based on data
from the guaranty agencies.  The IG audited the estimates based on
the data provided by the guaranty agencies\21 and concluded that
these estimates differed materially from the estimates based on
Education's own database.  Further, the IG reported that "Education's
own procedures for preparing its loan estimates were not sufficiently
rigorous to detect this material misstatement" and that "Education's
ability to continue to prepare auditable loan estimates for its
financial statements depends on establishing a reliable store of
up-to-date historical loan data."

The IG audited the guaranty agency-based estimates and found them to
be reasonable.  However, obtaining data from the guaranty agencies
was a time-consuming way for Education to develop its estimates of
loan program costs and should be viewed as a short-term solution
only.  It is not feasible for Education to carry out this process
year after year instead of relying on the principal system that was
designed to provide this information. 

In addition to these data problems, the auditors identified some
instances where Education's estimation practices could be improved. 
For example, Education did not have documented policies and
procedures for calculating the cost of credit programs, including a
formal review process, to routinely identify and correct potential
errors.  These important internal controls could help ensure the
reasonableness of Education's complex estimates in the future.  While
Education did some limited sensitivity analyses, it did not document
the analyses.  Such documented analyses would have allowed its
auditors to focus their audit procedures on the key cash flow
assumptions.  For fiscal year 1997, the IG audited more than 20 cash
flow assumptions to validate the assumptions used by the model to
estimate the cost of the loan programs.  In the future, conducting a
complete sensitivity analyses and documenting the results could save
significant time and effort by helping focus management and the
auditors' attention on the reasonableness of the assumptions that
have the greatest impact on the program's cost. 

Using guaranty agency data, Education was able to calculate
reasonable financial statement estimates of credit program costs,
partly due to the sophistication of its credit subsidy model. 
Education developed its own method for estimating and reestimating
credit program costs, called the Budget Loan Model (BLM) System. 
This model uses a series of assumptions to estimate cash flows over
the life of the loans.  BLM is used in concert with the OMB credit
subsidy model because Education believes it captures the unique
requirements of its program.  During the fiscal year 1997 audit, BLM
was reviewed and the auditors determined that BLM included and
modeled all key elements of Education's various loan program
requirements, such as loan term and repayment grace periods, and used
historically valid data obtained from the guaranty agencies. 

In addition to developing its sophisticated cash flow model,
Education followed other effective practices for cost estimation,
such as documenting its cash flow model, by developing a Technical
Manual and User's Guide.  Also, for financial statement purposes,
agency staff compared estimates of future loan performance, based on
the data obtained from the guaranty agencies, to actual costs
recorded in the accounting system and determined that the financial
statement estimates reasonably predicted future loan performance. 
Further, calculating the estimates of loan program costs at Education
was a coordinated effort between the accounting, budget, and program
staff.  For example, representatives from these three offices met on
a regular basis and jointly developed the cash flow assumptions used
in the financial statement estimates.  In addition, Education
compared the cash flow models to program requirements and determined
that its models accurately captured all material aspects of the
credit programs.  Finally, Education also calculated timely
reestimates for both budgetary and financial statement purposes.\22
However, as discussed previously, the NSLDS data used for the budget
reestimates were questionable. 

Since Education received an unqualified audit opinion on its fiscal
year 1997 financial statements, the agency was not required by OMB to
prepare a formal action plan to address any financial management
issues.  However, according to Education officials, Education has
efforts underway to address the challenges it faces in preparing
reasonable estimates of its loan program costs.  For example,
Education is continuing its efforts to review and correct inaccurate
or incomplete data in NSLDS, including providing detailed technical
instructions to data originators and providers (schools, lenders, and
guaranty agencies).  These efforts should help Education address its
major loan cost estimation challenges.  However, until these efforts
are successfully completed and Education can rely on the data in
NSLDS, it will be forced to repeatedly undergo an onerous process in
order to make reasonable estimates of loan program costs. 

In its June 1998 audit report on Education's fiscal year 1997
financial statements, the IG made several recommendations, which we
concur with, related to improving the agency's loan cost estimation
process.  Specifically, the IG recommended that Education (1)
maintain documentation of the source of the data used in developing
assumptions for its cash flow models and the models themselves, (2)
validate the data used in the models, (3) update data annually to
reflect the current activity, (4) perform and document sensitivity
analyses to identify factors that significantly impact the loan
estimates or may vary in the future as well as factors that rely on
assumptions not based on current data, (5) establish clearly defined
roles and responsibilities for staff and groups responsible for
developing estimates of Education's loan programs, (6) develop
formalized policies and procedures for estimating the cost of credit
programs, and (7) perform quality assurance reviews of loan estimates
and document the results of these reviews.  Education agreed with
these recommendations and as discussed above has acted or plans to
act to address these recommendations. 


--------------------
\19 These agencies serve as intermediaries between the government and
the lender.  They are responsible for reviewing student applications
and approving loans, reviewing and paying claims to lenders when
defaults occur, and collecting on defaulted loans. 

\20 NSLDS prescreens student financial aid program applications,
reports on student status confirmation, and tracks borrowers.  This
system contains information regarding loans made, insured, or
guaranteed for Education's direct and guaranteed loan programs.  Its
purposes are to (1) ensure that accurate and complete data on student
loan indebtedness and institutional lending practices are available,
(2) screen applications to identify prior loan defaults, (3) provide
a database to research and identify trends and patterns, (4) support
audits and program reviews, and (5) calculate default rates. 

\21 The data received from the guaranty agencies, which was used to
calculate the financial statement estimates, was subjected to
specific audit procedures by the guaranty agencies' auditors and was
determined to be materially correct by the auditors. 

\22 Education did not issue its fiscal year 1997 audited financial
statements until June 15, 1998, 3 months after the March 1, 1998,
deadline for submission to OMB, due to the difficulties with the
conversion to a new general ledger system, delays in completing
reconciliations with Treasury, and preparation of loan subsidy
estimates. 


         AGENCY COMMENTS AND OUR
         EVALUATION
-------------------------------------------------------- Letter :4.2.1

Education stated that it does not believe our report provides a basis
to conclude that data from NSLDS are of questionable validity.  They
further stated that NSLDS data were highly comparable to data
received from the guaranty agencies and that adjustments made to the
loan cost estimates as a result of the fiscal year 1997 financial
statement audit process were also reflected in the agency's budget
forecasts.  Education therefore concluded that the budget estimates
were highly reliable."

We disagree.  Our conclusion that the data from NSLDS are of
questionable validity is based on our review of the IG's fiscal year
1997 financial statement audit report and supporting work papers.  We
also held numerous discussions with IG staff responsible for the
audit.  Based on this work, we determined that (1) the data in NSLDS
have never been validated by the agency, despite the fact that the
IG, beginning with the fiscal year 1995 audit, recommended this be
done in order to provide a basis for preparing reasonable loan cost
estimates, (2) material differences were noted by IG staff between
the data in NSLDS and that provided by the guaranty agencies, and (3)
the adjustments made to the loan cost estimates as a result of the
fiscal year 1997 audit were made during the summer of 1998--several
months after the fiscal year 1997 budget estimates were submitted to
OMB as part of the President's fiscal year 1999 budget--and,
therefore, these adjustments could not have been reflected in those
budget estimates. 

Further, we reviewed the December 1998 letter the IG submitted to the
House Majority Leader and the Chairman, House Committee on Government
Reform and Oversight, in response to their request that the IG update
its assessment of the most significant challenges facing the
Department of Education.  The IG's letter identified improving the
data integrity of Education's information management systems,
including NSLDS, as one of Education's most significant management
challenges.  The report stated that the Student Financial Assistance
loan programs contain inaccurate and incomplete data. Specifically,
the IG reported that the September 1998 audit of NSLDS found that
about 3.7 million loan records totaling $10.7 billion had not been
updated with lender-provided loan status and principal and interest
balance data. 

Until Education corrects its inaccurate loan data and successfully
completes a validation of NSLDS, any loan cost estimates prepared
based on NSLDS will continue to be questionable.  (Education's
comments are reprinted in appendix IV.)


      HUD DID NOT PREPARE
      REASONABLE FISCAL YEAR 1997
      ESTIMATES, BUT RECENT
      IMPROVEMENTS HAVE BEEN MADE
---------------------------------------------------------- Letter :4.3

At the end of fiscal year 1997, HUD was unable to provide adequate
supporting data for its financial statement credit subsidy estimates. 
This lack of supporting data also calls into question the quality of
HUD's budget submission related to its credit subsidy estimates. 
Since then, HUD, with the assistance of independent contractors, has
focused significant effort on this area and has made considerable
progress towards developing the supporting data necessary to
reasonably estimate loan program costs, including those for the two
programs we reviewed.  These revised data are currently undergoing an
audit, which, once completed, will help determine the reliability of
the data and, thus, the reasonableness of HUD's loan cost estimates. 

Most of HUD's loan guarantees are made by the Federal Housing
Administration (FHA) which, as a government corporation, follows
private sector generally accepted accounting principles (GAAP).  FHA
received an unqualified audit opinion on its fiscal year 1997
financial statements prepared in accordance with GAAP.  However, in
order to consolidate FHA's financial results into HUD, credit program
cost information must be converted to federal accounting standards. 
HUD has had difficulty making this conversion due to the differences
in the two accounting approaches for credit programs.  Under SFFAS
No.  2, when estimating the liability and related expense for future
defaults on guaranteed loans, FHA must estimate, for the life of the
loans, all cash disbursements related to the loan guarantee and the
associated collateral (for example, payment of default claims and
costs to dispose of foreclosed property) as well as all cash receipts
(for example, loan guarantee premiums and proceeds from the sale of
foreclosed properties).  GAAP considers most of the same receipts and
disbursements but does not include loan guarantee premiums (a
significant cash receipt for FHA) when calculating the same liability
and related expense.  Under GAAP, the loan guarantee premiums are
generally reported as revenues.  Further, calculating the present
value of receipts and disbursements is not required under GAAP. 
Because of the different methods of calculating this liability and
related expense, the GAAP-based amount would be significantly
different from what would be calculated under SFFAS No.  2. 

For fiscal year 1997, FHA was unable to prepare financial statements
that complied with the requirements of SFFAS No.  2 in time to be
audited and included in HUD's consolidated financial statements.  As
a result, the auditors issued a qualified opinion on HUD's fiscal
year 1997 financial statements.  However, an independent public
accounting firm is currently auditing FHA's fiscal year 1997 balance
sheet prepared in accordance with SFFAS No.  2 as part of its audit
of the opening balances for the fiscal year 1998 financial statement
audit. 

While HUD was not recording the cost of its loan guarantee programs
on its financial statements in accordance with the requirements of
SFFAS No.  2, it was estimating the future cash flows of its loan
guarantee programs for budget purposes.  In the spring of 1998, we
evaluated the cash flow models used to develop the fiscal year 1997
budget for the two programs we reviewed and identified numerous
problems, such as formula errors and inconsistent calculations of
cash flow assumptions.  We also determined that these cash flow
models were not documented and the Mutual Mortgage Insurance (MMI)
Fund model required extensive manual data entry, which increased the
likelihood of errors.  Additionally, HUD was unable to provide
supporting data for many of the cash flow assumptions in the models. 
These problems with the cash flow models and the supporting data
raise concerns over the reliability of HUD's fiscal year 1997 budget
submission for its credit subsidy estimates. 

In the summer of 1998, HUD, with the assistance of independent
contractors, focused significant effort on correcting the errors in
these models.  These contractors assisted in gathering and developing
sufficient, relevant, and readily available supporting data as a
basis for the estimates of loan program cost estimates; performed
extensive detailed analyses of these cash flow models; identified
additional errors; and revised the models.  The contractors also
helped HUD implement other effective cost estimation practices.  For
example, the contractors compared cash flow models to program
requirements and determined that these revised models accurately
captured all significant aspects of the program, such as loan
origination fees and rebates of premiums when loans are repaid early. 
In addition, the contractors assisted HUD in documenting these cash
flow models, including sources of data and the mechanics of the
model. 

HUD, with the assistance of its contractors, also followed other
effective cost estimation practices.  For example, HUD recently
performed sensitivity analyses for its credit programs, including the
two we reviewed, to identify key cash flow assumptions.  And, by
working together, accounting, budget, and program staff focused their
efforts on gathering and documenting the basis for the assumptions
that had the greatest impact on HUD's credit subsidy estimates.  HUD
determined that an independent actuarial review provided the basis
for three of the six key cash flow assumptions for the MMI model--the
primary single family guaranteed loan program.  For the remaining key
cash flow assumptions for this program, HUD determined and documented
that the basis for estimating future loan performance was historical
experience from the accounting system. 

While HUD has generally improved its estimation process for the two
programs we reviewed, other improvements could be made.  For example,
comparing its estimates of future loan performance to actual cash
flows recorded in the accounting system would enable HUD to determine
whether these estimates reasonably predicted future loan performance. 
In making this comparison, we found that the average claim amount\23
used in the 1997 budget submission was consistent with historical
experience for the MMI Fund.  However, when the contractors were
updating HUD's fiscal year 1997 cash flow models for financial
reporting purposes, they misinterpreted a report and used it to
calculate an estimated average claim amount for the MMI Fund that was
significantly less than the actual amount recorded in the accounting
system. 

When we informed HUD of the error in the revised cash flow model, HUD
changed the average claim amount to be consistent with actual costs
recorded in the accounting system.  As a result, the estimated
program cost recorded in the draft financial statements increased
$1.3 billion.  If HUD had compared the estimated future loan
performance used in the models to actual costs recorded in the
accounting system, it would have detected this error in the average
claim amount. 

Also, for the two programs we reviewed, HUD did not prepare timely
credit subsidy reestimates for budgetary and financial statement
purposes.  HUD obtained permission from OMB to routinely prepare
budget reestimates in the summer following the reporting year. 
However, SFFAS No.  2 requires annual reestimates each year as of the
date of the financial statements if the reestimate would
significantly affect the amounts presented.  According to the
Director of HUD's Housing Budget Office, actual data from the
accounting system were not available in time to prepare reestimates
in the fall--the same time that the staff were formulating the annual
budget.  HUD management has refined its reestimate approach which
should have allowed for timely reestimates to be included in the
current year's budget and financial statements.\24

Because HUD received a qualified opinion on its fiscal year 1997
financial statements, it was required by OMB to prepare an action
plan to address identified financial management issues related to the
loan program cost estimation process.  This plan included
accumulating supporting data for estimating the cost of its loan
programs and reviewing its cash flow models to identify additional
improvements that could reduce the chance of error.  Further, the
plan included routinely reestimating the cost of its loan programs
timely and including the reestimates in both the current budget cycle
and the current year's financial statements.  Additionally, the plan
included establishing formal policies and procedures that include a
formal supervisory review process.  The plan also provides for
performing comparisons of estimated to actual loan performance.  This
plan, if fully implemented, should help HUD prepare reasonable
estimates of loan program costs. 


--------------------
\23 The average claim amount is the average amount paid to a lender
when borrowers default on their insured mortgage and is also a key
cash flow assumption for the Mutual Mortgage Insurance Fund program. 

\24 However, HUD did not complete all components of its credit
subsidy reestimates in time to be fully included as part of the
fiscal year 2000 President's Budget. 


         RECOMMENDATIONS
-------------------------------------------------------- Letter :4.3.1

In its audit report on the fiscal year 1997 financial statements, the
IG included a recommendation, with which we concur, that HUD develop
and implement a plan to prepare the FHA data needed to meet SFFAS No. 
2 requirements.  As previously discussed, HUD has taken steps to
address most of the problems related to reasonably estimating the
cost of its loan programs.  To help ensure that HUD is able to
reasonably estimate the cost of its loan programs, we recommend that
the Secretary of Housing and Urban Development or his designee take
the following actions: 

  -- Complete efforts to work with independent contractors to
     accumulate sufficient, relevant, and reliable data to estimate
     the cost of credit programs. 

  -- Implement plans to compare estimated cash flows to actual cash
     flow experience to validate the quality of the estimates as part
     of the annual reestimation process. 

  -- Implement its revised reestimate approach that will result in
     timely credit subsidy reestimates for both financial statements
     and budget submissions. 

  -- Implement existing plans to develop written policies and
     procedures including a formal supervisory review process for
     estimating the cost of credit programs. 


         AGENCY COMMENTS AND OUR
         EVALUATION
-------------------------------------------------------- Letter :4.3.2

HUD did not take exception to the findings discussed in this report
and agreed with and stated it plans to implement our recommendations. 
(HUD's comments are reprinted in appendix V.)


      MAJOR DEFICIENCIES IN BASIC
      LOAN ACCOUNTING SYSTEM
      PRECLUDED VA FROM MAKING
      REASONABLE ESTIMATES OF LOAN
      PROGRAM COSTS
---------------------------------------------------------- Letter :4.4

During fiscal year 1997, VA had serious problems\25 performing basic
accounting for its loan programs and, therefore, did not have a
reliable basis for the loan program cost estimates included in its
financial statements.  These problems contributed to the qualified
audit opinion on VA's fiscal year 1997 financial statements.  They
also raise doubts about the reliability of loan program cost
information submitted to OMB for budgetary purposes.  Further, we
found that VA did not record its guarantee obligations on the loans
it sold.  These weaknesses not only affect VA's ability to make
reasonable cost estimates, but also call into question its ability to
effectively manage and monitor its vendee loan program. 

During fiscal year 1997, VA transferred the management of its direct
loan portfolio to an outside servicer.\26 VA hoped that the transfer
would reduce the number of staff resources needed and resolve
existing internal control weaknesses and obsolescence issues related
to its computer system.  However, the data transferred to the
servicer, which up to that point had been maintained by more than 40
VA regional offices, were incomplete and inconsistent and immediately
created loan servicing problems.  Further, VA closed down its own
loan servicing system without putting in place procedures designed to
ensure that it maintained accountability over the loan portfolio. 
These procedures should have included maintaining an inventory of the
loans in the portfolio and a loan origination database to be used in
conjunction with the servicer's system.  It also included other
procedures for monitoring the amount and timing of cash due from
borrowers. 

As a result, VA management did not know the number or amount of
direct loans outstanding at year-end, which VA estimated to be at
least $2.1 billion, or whether the amount of cash received from the
servicer during the year was correct.  Further, because the servicer
did not have an accurate inventory, the servicer was, according to
VA, unable to allocate over $3 million in payments received after the
transfer and, therefore, did not have correct payment histories for
the affected loans.  Without this basic information, VA was unable to
reliably track the performance of its existing loans or reasonably
estimate the future performance of its loans or the cost of its
credit programs. 

VA's loan accounting problems were further exacerbated by its
improper treatment of loans sold to investors with a guarantee of
prompt payment of future principal and interest.  During fiscal year
1997, VA sold about $1 billion in loans and, since 1992, has sold
approximately $9 billion in loans.  Because VA guaranteed future
principal and interest payments on the sold loans, it is responsible
for future losses resulting from such occurrences as delinquencies
and defaults of the underlying loans.  According to SFFAS No.  2,
future losses should have been estimated and a subsidy expense and
related liability should have been established for future defaults or
delinquent payments when the loans were sold. 

Prior to fiscal year 1997, VA did not record the subsidy expense or
the liability for potential future defaults on the loans it sold. 
Once we identified this error, VA, in consultation with OMB,
estimated an additional expense as part of an aggregate adjustment
for future losses and related liability for the loan sales not
recorded between 1992 and 1997.\27 Because the adjustment was
aggregated in the financial statements with other adjustments related
to direct loans, we were unable to determine what portion of this
$376 million estimate was related directly to the loan sales
activity, therefore, we did not attempt to determine the
reasonableness of the adjustment.  However, because of the lack of
critical financial data, VA's ability to reasonably estimate the cost
of its guarantee obligations related to loans sold is severely
hampered.  In order to further refine this estimate, VA recently
hired an outside contractor to reconstruct the historical data on
prior loan sales and develop a model to estimate the cost of loans
sold with a guarantee.  In addition, the contractor plans to assess
VA's current cash flow models for direct and guaranteed loan programs
to determine whether the assumptions are appropriate. 

The problems VA had in accounting for its loans also hindered the
agency's ability to implement effective cost estimation practices. 
For example, because VA lacked complete data about the inventory of
loans in its portfolio, it did not have a reasonable basis for
estimates of future loan performance and could not compare estimated
loan performance to actual costs recorded in the accounting system. 

VA did not use certain other estimation practices that would have
improved its cost estimation process.  For example, VA did not
perform sensitivity analyses, but instead it relied on program
managers' opinions in order to identify those assumptions that had
the greatest impact on the programs' cost.  As part of our
assignment, we performed sensitivity analyses and verified that
program managers' opinions correctly identified the key cash flow
assumptions.  However, for new programs and changes in current
program design or delivery, sensitivity analyses would help ensure
that key cash flow assumptions are appropriately identified.  Also,
VA did not have either written policies and procedures for estimating
loan program costs or a formal review process that included
representatives from the program, budget, and accounting offices. 
While VA did implement a number of effective cost estimation
practices--including calculating timely reestimates, comparing cash
flow models to program requirements, and having organized, documented
cash flow models--the combined impact of VA's serious basic
accounting weaknesses hindered its ability to make reasonable cost
estimates. 

As required by OMB, VA has prepared an action plan to address its
financial management problems.  The action plan focuses on resolving
the fundamental problems in accounting for the loan sales program and
the basic data integrity issues related to the incomplete inventory
of loans currently maintained by the servicer.  Until these basic
accounting deficiencies are resolved, VA will continue to have
difficulty making reasonable estimates of its loan program costs. 
Additionally, once VA's basic accounting problems are resolved,
management can turn its attention to implementing practices that will
further improve its ability to make reasonable loan program cost
estimates. 


--------------------
\25 We currently have additional work underway that specifically
focuses on VA's serious basic accounting problems for its loan
programs and that will include recommendations to address those
problems. 

\26 Generally, a servicer manages the loan portfolio by accounting
for the individual loans, collecting loan payments, maintaining
escrow accounts, and foreclosing on delinquent borrowers. 

\27 This expense was estimated as part of the direct loan program. 
However, OMB Circular A-11 requires that the related subsidy cost for
loans sold with a guarantee be included as part of the loan guarantee
program.  In the future, VA plans to include the subsidy cost for the
loans sold in the appropriate program. 


         RECOMMENDATIONS
-------------------------------------------------------- Letter :4.4.1

In its audit report on the fiscal year 1997 financial statements, the
IG included a recommendation, with which we concur, that VA complete
actions underway to ensure that all direct loan records are complete
and accurate.  Once VA's basic accounting issues are resolved, in
order to correct the deficiencies that we identified in VA's direct
and guaranteed loan program cost estimation processes, we recommend
that the Secretary of Veterans Affairs or his designee implement the
following cost estimation practices: 

  -- Compare estimated cash flows to actual cash flow experience to
     validate the quality of the estimates as part of the annual
     reestimation process. 

  -- Develop and implement written policies and procedures that
     include a formal supervisory review process and a coordinated
     approach between program, budget, and accounting staff for
     estimating the cost of credit programs. 

  -- Use sensitivity analysis as a tool to identify key cash flow
     assumptions. 

  -- Continue efforts, with the assistance of contractors, to create
     and use a model and develop the necessary data to calculate the
     liability for the guarantee on sold loans and record the related
     liability in the financial statements. 


         AGENCY COMMENTS AND OUR
         EVALUATION
-------------------------------------------------------- Letter :4.4.2

In commenting on our draft report, VA concurred with our
recommendations and agreed to implement them as part of its current
efforts to correct direct loan records.  However, VA did not agree
with how we characterized the magnitude of the problems it
encountered when the agency outsourced its loan portfolio. 
Generally, VA asserted that the problems discussed in this report
were limited to a small portion of its overall credit programs and
should not be considered material. 

We disagree.  The problems with VA's loans receivable and the
liability for loan guarantees were so pervasive that the IG qualified
its audit opinion on VA's fiscal year 1997 financial statements.  As
stated in its report, the IG was unable to attest to the accuracy of
the loans receivable balance "because of incomplete records and the
poor quality of the direct loan portfolio records." The IG further
reported that VA's reported $2.1 billion net credit program
receivables balance was inaccurate because VA's accounting procedures
were not being consistently followed and/or internal controls were
not operating effectively.  The IG's report also stated that because
of VA's "inadequate records, there were

  -- numerous errors in direct loan and associated escrow account
     balances and payment of taxes and insurance,

  -- significant delays in establishing new loans in the accounting
     records and processing borrowers' loan payments, and

  -- inconclusive general ledger account balances."

Specifically, with regard to the materiality of the transferred
loans, these loans were $1.2 billion, or 57 percent of the reported
$2.1 billion of net credit program receivables.  Further, we found
that VA sold $9 billion of direct loans between 1992 and 1997 without
initially recording the cost of its guarantee obligations.  We
consider each of these amounts to be significant.  (VA's comments are
reprinted in appendix VI.)


      USDA LACKS ADEQUATE SYSTEMS
      AND HISTORICAL DATA TO
      REASONABLY ESTIMATE THE COST
      OF ITS CREDIT PROGRAMS
---------------------------------------------------------- Letter :4.5

For fiscal year 1997, USDA was unable to make reasonable cost
estimates for its loan programs because it did not maintain the
historical data needed to predict future loan performance and used
computer systems that were not appropriately configured to capture
the data necessary to make such estimates.  These long-standing
problems contributed to the auditor's inability to give an opinion on
USDA's fiscal year 1997 consolidated financial statements and raised
questions about the quality of the budget data related to USDA's loan
programs. 

For the two programs we reviewed, USDA performed sensitivity analyses
and identified which assumptions had the greatest impact on the
credit subsidy.  However, because it lacked adequate historical data,
USDA based its prediction of key assumptions, such as the amount and
timing of defaults and prepayments, primarily on the opinion of
program managers.  These managers estimated future loan performance
based on their programmatic knowledge and experience without the
assistance of extensive historical loan performance data and
sophisticated computer modeling.  Program management opinion may be
an acceptable source of support for estimates when a new, unique
program is established or when significant changes have been made to
existing programs.  However, program management opinion should be
used only as an interim method and does not provide a reliable basis
for established programs.\28

Additionally, when program manager opinion is used, it should
subsequently be compared to actual cash flow data from the accounting
system to corroborate the reasonableness of management's judgment. 

The lack of historical data for the two programs we reviewed was
largely the result of system inadequacies.  For example, prior to the
implementation of FCRA, USDA's systems did not track certain key cash
flow data.  In addition, although USDA's current systems were capable
of capturing some key cash flow data at the detail level, these
systems could not summarize the data so that they were readily usable
for calculating credit subsidy estimates.  For example, USDA's
current systems were incapable of accumulating summary level
prepayment information because the systems could not distinguish
between borrowers that were completely paying off loans and borrowers
that were paying an extra amount each month.  USDA's accounting
system also did not contain the loan origination date for loans that
were modified when borrowers experienced financial hardship and were
unable to meet scheduled payments.  As a result, the number and
amount of delinquent loans in the accounting system could not be
broken out by loan origination year, and USDA was unable to track
individual loans through their entire history without extensive
manual intervention. 

USDA also lacked adequate historical data to estimate the amount of
interest subsidy borrowers would receive in the future.  The Single
Family Housing Loan Program makes low interest rate loans to low
income families who lack adequate housing and cannot obtain credit
from other sources.  For this program, the amount of interest that
USDA subsidizes is based on borrowers' income.  As the borrowers'
income increases or decreases, USDA pays more or less interest
subsidy on the loans.  Ultimately, when the borrowers are capable of
paying market interest rates for a housing loan, they "graduate" from
the program because they no longer qualify for the program.  However,
USDA did not maintain adequate records of the number of borrowers who
moved to higher income levels and when borrowers were eligible to
graduate from the program. 

Another factor affecting the reasonableness of USDA's estimates of
loan program costs is the timing of reestimates that should be made
to incorporate actual loan performance and other new information.  In
order to reasonably estimate and report the cost of loan programs,
USDA must reestimate its credit subsidies on time and include these
reestimates in the current year's financial statements and budget
submission.  However, USDA management told us that the agency lacked
sufficient staff to make prompt reestimates for the programs we
reviewed because these estimates needed to be calculated in the fall
at the same time the budget was prepared.  USDA received OMB's
permission to calculate budgetary reestimates in the summer following
the financial statement reporting year.  However, this authorization
to delay budgetary reestimates did not allow USDA to delay the
financial statement reestimates.  Further, USDA did not calculate the
fiscal year 1997 reestimate for the Single Family Direct Loan Program
until the fall of 1998--after the date agreed upon with OMB.  As a
result, the agency did not update its fiscal year 1997 estimate of
loan program costs until nearly 3 years after the original estimate
was prepared in 1995.\29

Until USDA calculates timely reestimates and includes them in its
financial statements or clearly demonstrates that the reestimates
would not be material to the financial statements, the amount of
loans, liability for loan guarantees, and cost of the credit programs
may be materially misstated on the financial statements.  Delaying
the reestimates also affects the quality of loan performance and cost
data that are provided to the Congress for budgetary considerations. 

Also, for the two programs we reviewed, USDA did not use other
practices that would enhance its ability to reasonably estimate the
cost of loan programs.  For example, USDA did not routinely compare
estimated loan performance to actual costs recorded in the accounting
system to assess how closely the estimate compared with subsequent
actual costs.  It also did not routinely compare cash flow models to
program requirements.  These comparisons would have enabled USDA to
identify and research significant differences and determine whether
assumptions related to expected future loan performance needed to be
revised.  In addition, during fiscal year 1997, USDA did not have
formal policies and procedures for calculating estimates of loan
program costs or for a formal review process that included
representatives from the program, budget, and accounting offices to
help ensure continuity and accuracy during this complicated
estimation process. 

USDA has developed an action plan to address deficiencies in
estimating the cost of its loan programs.  This plan includes
aggressive time frames and directs budget and accounting staff to
prepare reasonable and timely estimates of loan program costs and to
assemble the most accurate and reliable data available for each
credit program.  The plan also includes implementation of a number of
practices that will improve USDA's estimation process, including
revising cash flow models and comparing estimated loan performance to
actual costs recorded in the accounting system.  Additionally, the
plan calls for a task force comprised of representatives from budget,
program, accounting, and the IG offices to ensure that preparing the
loan program cost estimates is a coordinated effort.  Further, the
plan calls for, and USDA is developing, formal policies and
procedures for calculating estimates of loan program costs, including
a formal review process by representatives from the program, budget,
and accounting offices.  Finally, the plan includes documenting the
basis for the assumptions used to estimate program costs and
identifying additional sources of data that may be used to reasonably
estimate future loan performance. 

The USDA IG told us that outside contractors may be needed to
successfully implement the agency's action plan.  Other agencies have
successfully used outside contractors to assist with gathering and
developing a reliable basis for their estimates of loan program costs
and improving their cash flow models. 

While implementation of the current plan will improve USDA's ability
to prepare reasonable estimates of loan program costs, the plan does
not currently address how USDA will implement the necessary computer
system enhancements to address such problems as providing complete
and accurate prepayment and delinquent loan information.  Until these
computer system enhancements are made, USDA will continue to have
great difficulty making reasonable estimates of loan program costs
based on reliable historical data.  Further, until written policies
and procedures that include a formal supervisory review process are
developed and fully implemented, USDA will lack important controls to
help ensure that errors are detected and corrected promptly. 


--------------------
\28 Statement on Auditing Standards No.  57, Auditing Accounting
Estimates, states that agency management is responsible for
"accumulating relevant, sufficient, and reliable data on which to
base the estimate." USDA lacked corroborating evidence demonstrating
that agency staff judgment reasonably predicted actual loan
performance for an extended period, and the USDA IG has concluded
that staff judgment alone is not sufficient support for these
estimates.  We concur with the IG's conclusion. 

\29 USDA management told us that beginning with fiscal year 1998,
USDA planned to revise its reestimate process to calculate timely
reestimates and include them in its fiscal year 1998 financial
statements and budget cycle.  However, implementation of this plan
was delayed by OMB because additional testing of this revised process
was needed. 


         RECOMMENDATIONS
-------------------------------------------------------- Letter :4.5.1

In its May 1998 audit report, the USDA IG recommended, and we concur,
that the agency develop sufficient, relevant, and reliable data to
support its estimates of loan program costs.  USDA has recognized the
need to develop and better document its basis for the credit subsidy
estimates and, as described above, has developed an action plan to
address this problem.  We also recommend that the Secretary of
Agriculture or his designee take the following actions: 

  -- Implement the action plan to address deficiencies in estimating
     the cost of loan programs in a timely manner, including

comparing estimated cash flows to actual cash flow experience to
validate the quality of the estimates as part of the annual
reestimation process,

reestimating loan program costs timely and including them in the
current year's financial statements and budget submissions, and

developing and implementing written policies and procedures that
include a formal supervisory review process and a coordinated
approach between program, budget, and accounting staff for estimating
the cost of credit programs. 

  -- Ensure that the key cash flow assumptions in existing cash flow
     models are documented, including comparisons to program
     requirements. 

  -- Ensure that once all mission-critical systems are Year 2000
     compliant, computer systems are updated to capture the data
     necessary to reasonably estimate loan program costs. 

  -- Consider hiring outside contractors to assist in gathering
     sufficient, relevant, and reliable data as a basis for credit
     program estimates. 


         AGENCY COMMENTS AND OUR
         EVALUATION
-------------------------------------------------------- Letter :4.5.2

We received comments from both the Rural Development (RD) and Farm
Service Agency (FSA) components of USDA because these two components
operate the programs included in our review.  In general, RD and FSA
did not take exception to either the findings or most of the
recommendations presented in this report.  However, FSA stated that
it would not be in its best interest to use outside contractors to
assist in gathering sufficient, relevant, and reliable data as a
basis for credit program estimates as we recommended.  According to
FSA, it has had success working with the National Agricultural
Statistical Service (NASS) and believes that NASS can accomplish the
same tasks we recommended, potentially at significantly lower costs. 
We agree that NASS has assisted FSA in gathering cash flow data. 
However, based on the amount of progress that other agencies have
experienced in a short period of time with the assistance of
independent contractors, we believe that FSA may also benefit from
this type of contractor support and should explore this option as
well. 

Other comments received from RD and FSA focused primarily on the
levels of historical data needed to make reasonable credit subsidy
estimates and the steps these components have taken to address some
of the challenges they face when making these estimates.  RD and FSA
also provided clarification on various points in our report, which we
have incorporated as appropriate.  (USDA's comments are reprinted in
appendix VII.)


   YEAR 2000 COMPUTING ISSUES
   COULD AFFECT THE ABILITY OF
   AGENCIES TO ESTIMATE LOAN
   PROGRAM COSTS
------------------------------------------------------------ Letter :5

Another factor that could significantly affect the five key credit
agencies' ability to make reasonable credit subsidy estimates in the
future is the Year 2000 problem.  The Year 2000 problem is rooted in
the way dates are recorded and computed in many computer systems. 
For the past several decades, systems have typically used two digits
to represent the yearsuch as "98" for 1998to save electronic data
storage space and reduce operating costs.  With this two-digit
format, however, the Year 2000 is indistinguishable from 1900, 2001
from 1901, and so on.  As a result of this ambiguity, system or
application programs that use dates to perform calculations may
generate incorrect results when working with years after 1999.  As an
example of the potential impact, a veteran born in 1925 and therefore
turning 75 in 2000 could be incorrectly computed as being negative 25
years old (if "now" is 1900)not even born yetand hence ineligible
for benefits that the veteran had been receiving, such as a mortgage
guarantee. 

Addressing the Year 2000 problem is a major challenge for the five
key credit agencies, all of which rely on computers to process and
update records.  Unless the systems that compile loan program
information are Year 2000 compliant, the five key credit agencies may
face serious problems at the turn of the century.  Systems used to
track loans could (1) produce erroneous information on loan status,
such as indicating that an unpaid loan had been satisfied or (2)
incorrectly calculate interest and amortization schedules.  Loan
origination, default, repayment schedule, prepayment, and premium
receipts are all linked to dates.  To assist in the credit subsidy
estimation process, this date-related information must be retained
for extended periods and used to project future cash flows for the
credit agencies' loan programs.  Therefore, computer systems that
support the five key credit agencies' various loan programs are
susceptible to the Year 2000 problem. 

To avoid widespread system failures, the five key credit agencies
have been fixing, replacing, or eliminating Year 2000 noncompliant
systems.  All of the systems that provide key cash flow data for the
10 loan programs we reviewed have been identified as mission
critical.\30 According to the agencies, these mission-critical
systems that support the loan cost estimation process are either
currently Year 2000 compliant or are scheduled to meet the OMB goal
to be compliant by March 31, 1999.  However, in its Quarterly Report: 
Progress on Year 2000 Conversion, as of mid-November 1998, OMB
expressed concerns related to progress on Year 2000 conversion at
Education and USDA.  OMB noted that HUD and VA appear to be making
satisfactory progress towards being Year 2000 compliant and pointed
out that SBA was the first agency to report that all of its
mission-critical systems were Year 2000 compliant.  In previous
reports and testimonies we have raised issues over the status of Year
2000 conversion efforts at the five key credit agencies, except for
SBA.  (See the list of GAO products related to Year 2000 efforts at
the end of this report.)

To fully address Year 2000 risks that the five key credit agencies
face, data exchange environment\31 problems must also be addressed--a
monumental issue.  As computers play an ever-increasing role in our
society, exchanging data electronically has become a common method of
transferring information between federal agencies and private sector
organizations.  For example, Education's student financial aid data
exchange environment is massive and complex.  It includes about 7,500
schools, 6,500 lenders, and 36 guaranty agencies, as well as other
federal agencies.  All five key credit agencies depend on electronic
data exchanges with external business partners to execute their
lending programs.  As computer systems are converted to process Year
2000 dates, the associated data exchange environment must also be
made Year 2000 compliant.  If the data exchange environment is not
Year 2000 compliant, data exchanges may fail or invalid data could
cause the receiving computer systems to malfunction or produce
inaccurate computations.  All five key credit agencies are working on
plans to address data exchange issues with external business
partners. 

Because of these risks, the five key credit agencies must have
business continuity and contingency plans to reduce the risk of Year
2000 business failures.\32

Specifically, the five key credit agencies must ensure the continuity
of their core business processes and lending operations by
identifying, assessing, managing, and mitigating their Year 2000
risks.  These efforts should not be limited to the risks posed by
Year 2000-induced failures of internal information systems but must
include the potential Year 2000 failures of others, including
external business partners. 

The business continuity planning process focuses on reducing the risk
of Year 2000-induced business failures.  It safeguards an agency's
ability to produce a minimum acceptable level of outputs and services
in the event of failures of internal or external mission-critical
systems.  It also helps identify alternate resources and processes
needed to operate the agency core business processes.  While it does
not offer a long-term solution to Year 2000-induced failures, it will
help an agency to prepare for potential problems and may facilitate
the restoration of normal service at the earliest possible time in
the most cost-effective manner.  All of the five key credit agencies
have begun business continuity and contingency planning. 


--------------------
\30 Mission-critical systems are those that support a core business
activity or process. 

\31 The data exchange environment includes the electronic transfer
(sending or receiving) of a data set using electronic media. 
Electronic data exchanges can be made using various methods,
including direct computer-to-computer exchanges over a dedicated
network, direct exchanges over commercially available networks or the
Internet, or exchanges of magnetic media such as computer tapes or
disks.  The information transferred in a data set often includes at
least one date. 

\32 Year 2000 Computing Crisis:  Business Continuity and Contingency
Planning (GAO/AIMD-10.1.19, August 1998). 


---------------------------------------------------------- Letter :5.1

We are sending copies of this report to the Ranking Minority Member
of the House Committee on the Budget.  We are also sending copies to
the Director, Office of Management and Budget; the Secretaries of
Agriculture, Education, Housing and Urban Development, and Veterans
Affairs; the Administrator of Small Business; and interested
congressional committees.  Copies also will be made available to
others upon request. 

Please contact me at (202) 512-9508 if you or your staffs have any
questions concerning this report.  Major contributors to this report
are listed in appendix VIII. 

Linda M.  Calbom
Director, Resources, Community, and Economic Development
 Accounting and Financial Management Issues


ESTIMATING CREDIT PROGRAM COSTS
=========================================================== Appendix I

The Federal Credit Reform Act of 1990 (FCRA) was enacted to require
agencies to more accurately measure the government's cost of federal
loan programs and to permit better cost comparisons both among credit
programs and between credit and noncredit programs.  FCRA assigned to
OMB the responsibility to coordinate the cost estimates required by
the act.  OMB is authorized to delegate to lending agencies the
authority to estimate costs, based on written guidelines issued by
OMB.  These guidelines are contained in sections 33.1 through 33.12
of OMB Circular No.  A-11, and supporting exhibits.\1

The Federal Accounting Standards Advisory Board (FASAB)\2 developed
the accounting standard for credit programs, SFFAS No.  2, Accounting
for Direct Loans and Loan Guarantees, which became effective with
fiscal year 1994.  This standard, which generally mirrors FCRA,
established guidance for estimating the cost of direct and guaranteed
loan programs, as well as recording direct loans and the liability
for loan guarantees for financial reporting purposes. 

The actual and expected costs of federal credit programs should be
fully recognized in both budgetary and financial reporting.  To
determine the expected cost of a credit program, agencies are
required to predict or estimate the future performance of the
program.  This cost, known as the subsidy cost, is the present
value\3 of disbursements--over the life of the loan--by the
government (loan disbursements and other payments) minus estimated
payments to the government (repayments of principal, payments of
interest, other recoveries, and other payments).  For loan
guarantees, the subsidy cost is the present value of cash flows from
estimated payments by the government (for defaults and delinquencies,
interest rate subsidies, and other payments) minus estimated payments
to the government (for loan origination and other fees, penalties,
and recoveries). 

To estimate the cost of loan programs, agencies first estimate the
future performance of direct and guaranteed loans when preparing
their annual budgets.  The data used for these budgetary estimates
should be reestimated to reflect any changes in loan performance
since the budget was prepared.  This reestimated data is then used in
financial reporting when calculating the allowance for subsidy (the
cost of direct loans), the liability for loan guarantees, and the
cost of the program.  In the financial statements, the actual and
expected cost of loans disbursed as part of a credit program is
recorded as a "Program Cost" on the agencies' Statement of Net Costs
for loans disbursed. 

In addition to recording the cost of a credit program, SFFAS No.  2
requires agencies to record direct loans on the balance sheet as
assets at the present value of their estimated net cash inflows.  The
difference between the outstanding principal balance of the loans and
the present value of their net cash inflows is recognized as a
subsidy cost allowance--generally the cost of the direct loan
program.  For guaranteed loans, the present value of the estimated
net cash outflows, such as defaults and recoveries, is recognized as
a liability and generally equals the cost of the loan guarantee
program. 

In preparing SFFAS No.  2, FASAB indicated that the subsidy cost
components--interest, defaults, fees, and other cash flows--would be
valuable for making credit policy decisions, monitoring portfolio
quality, and improving credit performance.  Thus, agencies are
required to recognize, and disclose in the financial statement
footnotes, the four components of the credit subsidy--interest, net
defaults, fees and other collections, and other subsidy
costs--separately for the fiscal year during which direct or
guaranteed loans are disbursed.  FASAB is currently considering
revising these standards. 

In addition, nonauthoritative guidance is contained in the previously
discussed Technical Release of the Credit Reform Task Force of the
Accounting and Auditing Policy Committee, entitled Preparing and
Auditing Direct Loan and Loan Guarantee Subsidies Under the Federal
Credit Reform Act.  This Technical Release provides detailed
implementation guidance for agency staff on how to prepare reasonable
credit subsidies.  Further, the Technical Release provides suggested
procedures for auditing credit subsidy estimates. 

DEVELOPING CASH FLOW ASSUMPTIONS
AND MODELS

Agency management is responsible for accumulating relevant,
sufficient, and reliable data on which to base the estimates. 
Further, SFFAS No.  2 states that each credit program should use a
systematic methodology to project expected cash flows into the
future.  To accomplish this task, agencies should develop cash flow
models.  A cash flow model is a computer-based spreadsheet that
generally uses historical information and various assumptions
including defaults, prepayments, recoveries, and the timing of these
events to estimate future loan performance.  These cash flow models,
which should be based on sound economic, financial, and statistical
theory, identify key factors that affect loan repayment performance. 
Agencies use this information to make more informed predictions of
future credit performance.  The August 1994 User's Guide To Version
r.8 of the OMB Credit Subsidy Model provides general guidance on
creating cash flow models to estimate future delinquencies, defaults,
recoveries, etc.  This user's guide states that "In every case, the
agency or budget examiner must maintain current and complete
documentation and justification for the estimation methods and
assumptions used in determining the cash flow figures used for the
OMB Subsidy Model" to calculate the credit subsidy. 

According to SFFAS No.  2, to estimate the cost of loan programs and
predict the future performance of credit programs, agencies should
establish and use reliable records of historical credit performance. 
Since actual historical experience is a primary factor upon which
estimates of credit performance are based, agencies should maintain a
database, also known as an information store, at the individual loan
level, of historical information on all key cash flow assumptions,
such as defaults or recoveries, used in calculating the credit
subsidy cost.  Additional nonauthoritative guidance on cash flow
models may be found in the Model Credit Program Methods and
Documentation for Estimating Subsidy Rates and the Model Information
Store issue paper prepared by the Credit Reform Task Force of the
Accounting and Auditing Policy Committee.  The draft "Information
Store" Task Force paper provides guidance on the type of historical
information agencies need to reasonably estimate the cost of credit
programs.  The information store should provide three types of
information.  First, the information store should maintain key loan
characteristics at the individual loan level, such as the loan terms
and conditions.  Second, it should track economic data that influence
loan performance, such as property values for housing loans.  Third,
an information store should track historical cash flows on a
loan-by-loan basis.  The data elements in an information store should
be selected to allow for more in-depth analyses of the most
significant subsidy estimate assumptions. 

In addition to using historical databases and the cash flow models,
other relevant factors must be considered by agencies to estimate
future loan performance.  These relevant factors include

  -- economic conditions that may affect the performance of the
     loans,

  -- financial and other relevant characteristics of borrowers,

  -- the value of the collateral to loan balance,

  -- changes in recoverable value of collateral, and

  -- newly developed events that would affect loan performance. 

REESTIMATING CREDIT SUBSIDIES

Agencies prepare estimates of loan program costs as a part of their
budget requests.  Later, after the end of the fiscal year, agencies
are required to update or "reestimate" loan costs for differences
among estimated loan performance and related cost, the actual program
costs recorded in the accounting records, and expected changes in
future economic performance.  The reestimate should include all
aspects of the original cost estimate including prepayments,
defaults, delinquencies, recoveries, and interest.  Reestimates of
the credit subsidy allow agency management to compare the original
budget estimates with actual program results to identify variances
from the original estimate, assess the quality of the original
estimate, and adjust future program estimates as appropriate.  Any
increase or decrease in the estimated cost of the loan program is
recognized as a subsidy expense or a reduction in subsidy expense for
both budgetary and financial statement purposes. 

The reestimate requirements for interest rate and technical
assumptions (defaults, recoveries, prepayments, fees, and other cash
flows) differ.  For budget purposes, OMB Circular A-11 states that
agencies must reestimate the interest portion of the estimate when 90
percent of the direct or guaranteed loans are disbursed.  The
technical reestimate, for budgetary purposes, generally must be done
annually, at the beginning of every year as long as the loans are
outstanding, unless a different plan is approved by OMB,\4 regardless
of financial statement significance.  For financial statement
reporting purposes, both technical and interest rate reestimates are
required annually, at the end of the fiscal year, whenever the
reestimated amount is significant to the financial statements.  If
there is no significant change in the interest portion of the
estimate prior to the loans being 90 percent disbursed, then the
interest reestimate may be done at least once when the loans are 90
percent disbursed. 


--------------------
\1 The act requires OMB to coordinate with the Congressional Budget
Office in developing estimation guidelines. 

\2 FASAB was created by OMB, Treasury, and GAO to consider and
recommend accounting principles for the federal government.  These
three agencies approved Statement of Federal Financial Accounting
Standards (SFFAS) No.  2, Accounting for Direct Loans and Loan
Guarantees, in July 1993. 

\3 Present value is the value today of a stream of payments in the
future discounted at a certain interest rate.  For the period we
reviewed, when calculating the present value of loan subsidy costs,
agencies must use as the discount rate the average annual interest
rate for marketable U.S.  Treasury securities with similar maturities
to the loan or guarantee. 

\4 The OMB representative with primary budget authority may authorize
agencies to calculate technical reestimates for budgetary purposes
less frequently than every year when any one of four conditions are
met.  These four conditions are (1) based on periodic schedules
established in coordination with OMB, (2) when a major change in
actual versus projected activity is detected, (3) when a significant
difference is detected through monitoring "triggers" developed in
coordination with OMB, and (4) when a group of loans are being closed
out. 


OBJECTIVES, SCOPE, AND METHODOLOGY
========================================================== Appendix II

Our objectives were to assess (1) the ability of agencies' to
reasonably estimate the cost of their loan programs, including
whether they used practices identified by the Credit Reform Task
Force\1 as being effective in making these estimates and (2) the
status of agencies' efforts to ensure that computer systems used to
estimate the cost of credit programs are Year 2000 compliant. 

We selected a sample of 10 programs--5 direct loan programs totaling
$52.1 billion and 5 guaranteed loan programs totaling $558.1
billion--from the five agencies with the largest domestic federal
credit programs:  the Small Business Administration, and the
Departments of Education, Housing and Urban Development, Veterans
Affairs, and Agriculture.  We generally selected programs that had
the most credit outstanding or highest loan levels at each agency. 
Specifically, these programs were: 

  -- 7(a) General Business Loans Program and Disaster Loan Program,
     which totaled 72 percent of SBA's loan guarantees and 73 percent
     of its direct loans, respectively.  7(a) General Business Loans
     Program guarantees loans made to small businesses that are
     unable to obtain financing in the private credit market but can
     demonstrate the ability to repay the loan.  Disaster loans are
     made to homeowners, renters, businesses of all sizes, and
     nonprofit organizations that have suffered uninsured physical
     property loss as a result of a disaster in an area declared
     eligible for assistance by the President or SBA. 

  -- Federal Family Education Loan Program and William D.  Ford
     Direct Loan Program, which totaled 100 percent of Education's
     loan guarantees and 50 percent of its total loans receivable,
     respectively.  These two programs help pay for educational
     expenses incurred by vocational, undergraduate, and graduate
     students enrolled at eligible postsecondary institutions.  The
     guaranteed loans are made by private lenders, insured by a state
     or private nonprofit guaranty agency, and reinsured by the
     federal government, whereas the direct loans are made directly
     from the federal government to the students. 

  -- Mutual Mortgage Insurance Fund and the General and Special Risk
     Insurance Fund Section 223(f) Refinance, which totaled 81
     percent of HUD's loan guarantees.  The Mutual Mortgage Insurance
     Fund helps people become homeowners by providing insurance to
     lenders that finance the purchase of one-to-four family housing
     that is proposed, under construction, or existing, or lenders
     that refinance indebtedness on existing housing.  The Special
     Risk Insurance Fund Section 223 (f) Refinance insures lenders
     against loss on the purchase or refinance of existing
     multifamily housing projects. 

  -- Guaranty and Indemnity Fund and the Loan Guaranty Direct Loan
     Program, which totaled 100 percent of VA's post credit reform
     loan guarantees and 69 percent of its total loans receivable,
     respectively.  The Guaranty and Indemnity Fund assists veterans
     and certain others in obtaining credit for the purchase,
     construction, or improvement of homes on more favorable terms
     than are generally available to nonveterans.  The Loan Guaranty
     Direct Loan Program makes home loans on favorable terms to
     members of the general public--both veterans and nonveterans--
     purchasing a VA-owned property. 

  -- Farm Service Agency Farm Operating Loans Program and the Rural
     Housing Service Single Family Housing Program, which totaled 20
     percent of USDA's direct loans.\2

Farm Service Agency, Farm Operating Loans are made to family farmers
who are unable to obtain credit from private and cooperative sources
and are intended to help provide farmers with the opportunity to
conduct successful farm operations.  The Rural Housing Service,
Single Family Housing Loans are made to very low- and low-income
families who are without adequate housing and cannot obtain credit
from other sources and may be used to build, purchase, repair, or
refinance homes in rural areas. 

To gain an understanding of the credit programs and the agencies'
credit subsidy estimation process, we obtained and reviewed the
fiscal years 1996 and where available 1997 financial statement audit
work papers.  During this review, we focused primarily on the
auditor's review of the loans receivable and liability for loan
guarantees line items on the balance sheet as well as the audit of
the credit subsidy cost on the statement of operations.  In addition,
at VA, SBA, and HUD, we directly participated in the fiscal year 1997
financial statement audits as part of the federal government's first
consolidated financial statement audit. 

To assess the reasonableness of agencies' credit subsidy estimation
processes, we first performed sensitivity analyses of the SBA, HUD,
and VA cash flow models to identify the key cash flow assumptions,
which are those assumptions having the greatest impact on the credit
subsidy.  We used the sensitivity analyses performed by USDA and
Education.  To perform the sensitivity analyses, we obtained copies
of the agencies' cash flow models and performed an extensive search
to identify each root cash flow assumption\3 in the agencies' cash
flow model. 

Once identified, each root cash flow assumption was adjusted, both up
and down, by a fixed proportion.  We followed the guidance in the
Credit Reform Task Force's Technical Release Preparing and Auditing
Direct Loan and Loan Guarantee Subsidies Under the Federal Credit
Reform Act and adjusted each root cash flow assumption by 10 percent. 
To determine which root assumption had the greatest impact on the
credit subsidy, we used the adjusted cash flows as input into OMB's
credit subsidy model to recalculate the subsidy.  For the recovery
assumptions--generally the estimated amount agencies receive from
selling collateral net of cash outflows for managing, maintaining,
and selling foreclosed properties--we adjusted the recovery
assumption along with the default timing assumption to ensure that
recoveries occurred after the defaults. 

Once we identified the key cash flow assumptions, we used the
guidance in Statement on Auditing Standard No.  57, Auditing
Accounting Estimates, as well as the Technical Release to determine
whether agencies had a reliable basis--whether the agencies had
gathered sufficient, relevant, and reliable supporting data--for the
estimates of loan program cost and for their estimates of loan
program performance.  Because of VA's serious problems performing
basic accounting for its loan programs, we determined that it would
not be meaningful to further assess whether VA had sufficient,
relevant, and reliable supporting data for its estimates of loan
program costs.  When possible, we used the work of the agencies'
fiscal year 1997 financial statement auditors to determine whether
agencies had a reliable basis for their estimates of loan program
costs.  We also compared program descriptions with agencies' cash
flow models to determine whether all characteristics of the program
were appropriately modeled.  Further, we compared estimated loan
program performance to actual loan program performance when
appropriate, to determine whether material variances between the
estimates and actual performance existed.  For two agencies, USDA and
VA, this comparison was not meaningful, and therefore not performed,
because of serious data quality concerns. 

To determine whether agencies had implemented the practices
identified in the Technical Release, we interviewed agencies'
accounting, program, and budget staff and assessed the process
agencies used to estimate the cost of their loan programs.  We also
compared the process agencies used to the practices identified in the
Technical Release.  Finally, we obtained and reviewed agencies' most
recent action plans to address financial management weaknesses,
including those related to estimating credit program costs. 

We also reviewed the status of agency efforts to ensure that computer
systems that provide cash flow data used to estimate the cost of
credit programs were Year 2000 compliant.  To do this, we met with
cognizant agency officials and identified the systems that provide
data supporting key cash flow assumptions and determined whether the
agency had assured that these systems were currently, or were
scheduled to be on time for Year 2000 compliance.  We also reviewed
the agencies' Year 2000 compliance plans and status reports to OMB. 
We did not independently test the systems that provide data
supporting key cash flow assumptions to determine whether the systems
were Year 2000 compliant as reported to OMB.  Further, we discussed
the agencies' efforts to develop contingency plans designed to ensure
the continued operation of critical business processes despite system
failures.  Our work was conducted in Washington, D.C., and St. 
Louis, Missouri, from September 1997 to November 1998 in accordance
with generally accepted government auditing standards. 



(See figure in printed edition.)Appendix III

--------------------
\1 The Credit Reform Task Force, formerly known as the Subgroup on
Credit Reform of the Governmentwide Audited Financial Statements Task
Force (established to study accounting and auditing issues related to
credit reform implementation in preparation for the first audit of
the governmentwide consolidated financial statements and to provide
guidance to agencies to resolve these issues), is now a task force of
the Accounting and Auditing Policy Committee sponsored by the Federal
Accounting Standards Advisory Board (FASAB).  This task force
developed a Technical Release, Preparing and Auditing Direct Loan and
Loan Guarantee Subsidies Under the Credit Reform Act, which has been
approved by FASAB and is expected to be issued by the Office of
Management and Budget as authoritative guidance during fiscal year
1999. 

\2 At the end of fiscal year 1997, USDA reported that $79.3 billion,
nearly 78 percent of the direct loan portfolio, was either disbursed
prior to the implementation of the Federal Credit Reform Act in
fiscal year 1992 or consisted of price support loans that were
excluded by Section 502(1) of the Federal Credit Reform Act.  Thus,
we focused on these two programs because they were the domestic loan
programs subject to credit reform that had disbursed the largest
volume of loans in less than 4 years. 

\3 The root cash flow assumption is the starting point for the
assumption, which means that there are no preceding formulas or
related inputs that would affect the assumption. 


COMMENTS FROM THE SMALL BUSINESS
ADMINISTRATION
========================================================== Appendix II




(See figure in printed edition.)Appendix IV
COMMENTS FROM THE DEPARTMENT OF
EDUCATION
========================================================== Appendix II

supplementing those in the report text appear at the end of this
appendix. 



(See figure in printed edition.)


The following are GAO's comments on the Department of Education's
January 12, 1999, letter. 

GAO COMMENTS

1.  See "Agency Comments and Our Evaluation" section for Education. 

2.  We agree with Education's statement that the agency uses its
budget model in concert with rather than in lieu of the OMB credit
subsidy model.  The report was revised accordingly. 

3.  When calculating the portion of Education's direct loans
receivable that was included in the scope of our review, we included
the defaulted loan guarantees and facilities loans as part of the
total direct loans receivable universe.  However, Education does not
consider these loans to be direct loans, but does consider them to be
part of total credit program receivables.  While our review did cover
100 percent of what Education considers to be direct loans, we
believe it is appropriate to continue to reflect our scope of loans
reviewed as a percentage of total credit program receivables. 
Further, to be consistent with the information available from other
key credit agencies, we revised the direct loan scope percentage to
exclude the allowance for subsidy.  As a result, the scope
calculation for this review was revised to 50 percent of total credit
program receivables. 

4.  We agree with Education's statement that Stafford Loans should
not be included in the program titles and revised the report
accordingly. 




(See figure in printed edition.)Appendix V
COMMENTS FROM THE DEPARTMENT OF
HOUSING AND URBAN DEVELOPMENT
========================================================== Appendix II




(See figure in printed edition.)Appendix VI
COMMENTS FROM THE DEPARTMENT OF
VETERANS AFFAIRS
========================================================== Appendix II

supplementing those in the report text appear at the end of this
appendix. 



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)


The following are GAO's comments on the Department of Veterans
Affairs' January 15, 1999, letter. 

GAO COMMENTS

1.  See the "Agency Comments and Our Evaluation" section for VA. 

2.  We do not agree that VA has already corrected the accounting
issues discussed in this report.  In September 1998, we visited VA's
servicer and concluded that the problems described in this
report--including the servicer's inability to monitor the amount and
timing of cash due from borrowers, its inaccurate and incomplete loan
histories, and its inability to allocate payments received from
borrowers to the appropriate borrower's account--continued to exist. 

3.  When reviewing VA's action plan, we focused only on those
initiatives designed to address the basic accounting weaknesses VA
had with its credit programs because this was the scope of our
review.  Thus, we did not focus on other actions described by VA
because they were beyond the scope of this review. 

4.  While the specific scope of this review was the Guaranty and
Indemnity Fund and the Loan Guaranty Direct Loan Program, VA provided
us with one model for its guaranteed loan program and one model for
its direct loan program that included both the "vendee" and acquired
loans.  However, only one subsidy cost is produced per model.  Since
the direct loan model included both types of loans, we could not
determine the amount of subsidy attributable to each loan type.  As a
result, we reviewed 100 percent of VA's post credit reform loan
guarantees and 69 percent of its total loans receivable.  To further
clarify this, the report was revised to better describe our scope
determination. 

5.  This report does not focus on VA's "non-established loans."
However, this is being covered in a review that is now ongoing. 

6.  The $3 million referred to in our report relates to monthly loan
payments or loan payoff amounts that VA received which it could not
match to a borrower.  The existence of this condition calls into
question the completeness of VA's loan records and seriously
undermines its ability to monitor the performance of existing loans
and reasonably estimate the future performance of its credit
programs. 




(See figure in printed edition.)Appendix VII
COMMENTS FROM THE DEPARTMENT OF
AGRICULTURE
========================================================== Appendix II

supplementing those in the report text appear at the end of this
appendix. 



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)


The following are GAO's comments on the Department of Agriculture's
January 11, 1999, letter. 

GAO COMMENTS

1.  We agree that USDA needs to determine the appropriate amount of
detailed history needed to make reasonable predictions of future loan
performance.  The amount of history needed for each loan program
would likely vary by program type and complexity and be closely
linked to the quality and type of history agencies had available. 
However, we do not agree that loan history is of limited value where
program or economic changes have occurred.  Historical experience
should be used as the baseline for an agency's credit subsidy
estimate.  Once this baseline is established, the incremental changes
in cash flows due to expected changes in the current and forecasted
economic conditions as well as changes in program design or delivery
should be adjusted for.  In addition, USDA should compare the
estimates of loan performance for the changed credit program to the
most recent historical experience to ensure that current estimates
are reasonably predicting actual loan program performance.  However,
as discussed in our report, USDA was not routinely comparing its
estimates of loan program costs with actual historical experience. 

Further, we agree that extensive pre-credit reform detailed loan
history may not be required in all cases and, in some cases, reliable
summary level information may be acceptable.  However, because nearly
73 percent of USDA's reported loan portfolio is comprised of
pre-credit reform loans, we believe that this experience is relevant
to USDA's current estimates of loan performance and, therefore,
should be considered at some level. 

2.  While we agree that USDA has made system changes to help control
funds, track cohorts, and respond to financial and budgetary
reporting needs, further work is needed.  USDA officials told us that
the current systems configuration does not allow the systems to
summarize the data so that it is readily usable for calculating
credit subsidy estimates.  Until these systems are able to readily
provide reliable key cash flow data in a format that can be easily
used in the subsidy estimation process, USDA's ability to calculate
reasonable credit subsidy estimates will continue to be impaired. 

3.  We agree that some progress has been made by USDA in the past 2
years in addressing the challenges it faces preparing reasonable
estimates of its loan program costs; however, the benefits of these
and planned future actions have not yet been fully realized. 
Further, as explained in comment 2, until USDA's systems readily
provide reliable supporting data for the credit subsidy estimates,
the ultimate success of these actions may be jeopardized. 

4.  Although we acknowledge that USDA has a large number of loan
programs, the amount of work needed to prepare reasonable credit
subsidy estimates can be reduced by optimizing its computer systems'
abilities and appropriately configuring these systems to readily
provide reliable data for the loan cost estimation process. 

5.  We did not intend to imply, and did not state in our draft
report, that the note interest rate varied based on the borrower's
income.  To avoid further confusion on this point, the report was
revised to clarify that the amount of interest subsidy paid by USDA
changes when a borrower's income changes.  Further, the lack of
support for USDA's interest subsidy assumption for this loan program
could affect the estimates of the amount of interest subsidy that
would be recaptured (the amount of interest subsidy a borrower may be
required to repay upon sale of the property). 

6.  The Year 2000 section of this report focused on the status of
systems managed at the USDA agency level and did not address systems
managed by components such as Rural Development.  We did not verify
the status of Rural Development's Year 2000 compliance efforts. 

7.  The report was revised to reflect the agency's comment. 

8.  See "Agency Comments and Our Evaluation" section for USDA. 

9.  We do not agree that because of ever-evolving credit reform
standards that agencies appear to be in a no-win situation and
oversight agencies continue to raise the bar on their expectations. 
The requirements and standards have changed little since the Federal
Credit Reform Act became effective in 1992 and the related accounting
standards became effective in 1994.  Since this time, OMB, Treasury,
and the Accounting and Auditing Policy Committee's credit reform task
force (which included representatives from USDA) have been working to
help provide agencies with detailed guidance on how to implement
credit reform requirements.  Further, as demonstrated by the Small
Business Administration and the Department of Education, some
agencies are able to prepare reasonable credit subsidy estimates. 
Finally, when forecasting loan repayments for a credit program whose
performance is directly linked to economic conditions, econometric
models are an appropriate tool because they would consider the impact
of economic conditions on estimated future loan repayments. 
Econometric modeling techniques are not new and have been
successfully used by at least one of the five key credit agencies in
their estimation processes. 


MAJOR CONTRIBUTORS TO THIS REPORT
======================================================== Appendix VIII

ACCOUNTING AND INFORMATION
MANAGEMENT DIVISION, WASHINGTON,
D.C. 

Dan Blair, Assistant Director
Shirley Abel, Assistant Director
Julia Duquette, Audit Manager
Marcia Carlsen, Senior Auditor
Meg Mills, Communications Analyst
McCoy Williams, Assistant Director
Mary Papadopulos, Senior Auditor
Alana Stanfield, Assistant Director
Elizabeth Kreitzman, Audit Manager
Christina Quattrociocchi, Senior Auditor
Suzanne Lightman, Auditor
Chris Bonham, Assistant Director
Carolyn Litsinger, Senior Evaluator

RESOURCES, COMMUNITY, AND ECONOMIC
DEVELOPMENT DIVISION, WASHINGTON,
D.C. 

Austin Kelly, Senior Economist

ST.  LOUIS FIELD OFFICE

Christie Arends, Senior Auditor
Arthur Brouk, Audit Manager
Gary Brown, Auditor




RELATED GAO PRODUCTS
============================================================ Chapter 0

CREDIT REFORM

Credit Reform:  Greater Effort Needed to Overcome Persistent Cost
Estimation Problems (GAO/AIMD-98-14, March 30, 1998). 

Credit Reform:  Review of OMB's Credit Subsidy Model
(GAO/AIMD-97-145, August 29, 1997). 

Credit Subsidy Estimates for the Sections 7(a) and 504 Business Loan
Programs (GAO/T-RCED-97-197, July 15, 1997). 

Credit Reform:  Case-by-Case Assessment Advisable in Evaluating
Coverage, Compliance (GAO/AIMD-94-57, July 28, 1994). 

Federal Credit Programs:  Agencies Had Serious Problems Meeting
Credit Reform Accounting Requirements (GAO/AFMD-93-17, January 6,
1993). 

YEAR 2000 COMPUTING CRISIS

Year 2000 Computing Crisis:  Significant Risks Remain to Department
of Education's Student Financial Aid Systems (GAO/T-AIMD-98-302,
September 17, 1998). 

Year 2000 Computing Crisis:  Strong Leadership and Effective
Partnerships Needed to Reduce Likelihood of Adverse Impact
(GAO/T-AIMD-98-277, September 2, 1998). 

Year 2000 Computing Crisis:  Progress Made in Compliance of VA
Systems, But Concerns Remain (GAO/AIMD-98-237, August 21, 1998). 

Year 2000 Computing Crisis:  Business Continuity and Contingency
Planning (GAO/AIMD-10.1.19, August 1998). 

Year 2000 Computing Crisis:  Actions Needed on Electronic Data
Exchanges (GAO/AIMD-98-124, July 1, 1998). 

Year 2000 Computing Crisis:  USDA Faces Tremendous Challenges in
Ensuring That Vital Public Services Are Not Disrupted
(GAO/T-AIMD-98-167, May 14, 1998). 

Year 2000 Computing Crisis:  Strong Leadership Needed to Avoid
Disruption of Essential Services (GAO/T-AIMD-98-117, March 24, 1998). 

Veterans Affairs Computers Systems:  Actions Underway Yet Much Work
Remains to Resolve Year 2000 Crisis (GAO/T-AIMD-97-174, September 25,
1997). 

Year 2000 Computing Crisis:  An Assessment Guide (GAO/AIMD-10.1.14,
September 1997). 

Veterans Benefits Computer Systems:  Uninterrupted Delivery of
Benefits Depends on Timely Correction of Year 2000 Problems
(GAO/T-AIMD-97-114, June 26, 1997). 

Veterans Benefits Computers Systems:  Risk of VBA's Year 2000 Efforts
(GAO/AIMD-97-79, May 30, 1997). 


*** End of document. ***