Credit Reform: Greater Effort Needed to Overcome Persistent Cost
Estimation Problems (Letter Report, 03/30/98, GAO/AIMD-98-14).

Pursuant to a congressional request, GAO reported on the government's
measurement of subsidy costs for federal direct loans and loan
guarantees, focusing on whether: (1) agencies completed estimates and
reestimates of subsidy costs; (2) GAO could readily discern any trends
including improvements in subsidy estimates; (3) GAO could readily
identify the causes for changes in subsidy estimates; and (4) agencies
with discretionary credit programs initially underestimated credit
subsidy costs in response to the incentive created by the availability
of permanent, indefinite budget authority for credit reestimates.

GAO noted that: (1) after over 6 years of experience with credit reform,
agencies continue to have problems in estimating the subsidy cost of
credit programs; (2) the lack of timely reestimates as well as the
frequent absence of documentation and reliable information limit the
ability of agency management, the Office of Management and Budget (OMB),
and Congress to exercise intended oversight; (3) GAO found problems with
the availability and reliability of subsidy estimates, reestimates, and
supporting documentation in its cross-cutting review of 10 programs for
fiscal years (FY) 1992 through 1998; (4) in the audits of the FY 1996
financial statements, three of the five largest credit agencies received
disclaimers or qualified opinions related to their credit programs; (5)
auditors were unable to find support for agency data on such items as
delinquencies and prepayments for loans receivable and liabilities for
loan guarantees; (6) problems with the reliability and validity of the
underlying credit data raise questions about the basis for the subsidy
estimates included in the budget; (7) GAO would expect to find that, for
a given cohort, the annual changes in reestimates due to technical
factors would be smaller; (8) because component data were not available,
GAO could not determine whether this occurred; (9) however, GAO did note
that overall subsidy rates for a given cohort varied widely; (10) GAO
observed a similar pattern of fluctuations in subsidy estimates at the
program level; (11) subsidy rate estimates for any given program
continued to fluctuate widely from year to year with no pattern or
particular trend; (12) the intersection of credit reform and the Budget
Enforcement Act of 1990--that is, the fact that original subsidy
appropriations must compete under the discretionary caps while
reestimates are outside them--may offer an incentive for agencies with
discretionary programs to underestimate subsidy costs initially to
permit more loans or loan guarantees within a given appropriation level;
(13) however, available data were not sufficient to assess whether a
credit program's budgetary treatment affected its initial subsidy
estimates; (14) better information on factors underlying changes in
subsidy rates is needed to identify and understand why these estimates
change; (15) while OMB provides a user's guide and some training on the
subsidy model, it has not provided agencies with clear definitions of
each component or sufficient guidance on how to use the model; and (16)
while no single agency is successful in all aspects of credit reform
implementation, some progress is being made at each of the agencies
studied.

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

 REPORTNUM:  AIMD-98-14
     TITLE:  Credit Reform: Greater Effort Needed to Overcome Persistent 
             Cost Estimation Problems
      DATE:  03/30/98
   SUBJECT:  Credit
             Subsidies
             Financial statement audits
             Loan accounting systems
             Direct loans
             Budget administration
             Government guaranteed loans
             Future budget projections
             Accounting procedures
             Financial management systems
IDENTIFIER:  Farm Operating Loan Program
             USDA Single Family Housing Direct Loan Program
             Federal Family Education Loan Program
             William D. Ford Federal Direct Loan Program
             Mutual Mortgage Insurance Fund
             SBA 7(a) General Business Loan Program
             SBA Disaster Loan Program
             VA Guaranty and Indemnity Fund
             VA Direct Loan Guaranty Loan Program
             HUD General and Special Risk Insurance Fund
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


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


Report to the Chairman, Committee on the Budget, U.S.  Senate

March 1998

CREDIT REFORM - GREATER EFFORT
NEEDED TO OVERCOME PERSISTENT COST
ESTIMATION PROBLEMS

GAO/AIMD-98-14

Credit Subsidy Estimates

(935217)


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

  CBO - Congressional Budget Office
  CFO - Chief Financial Officer
  FASAB - Federal Accounting Standards Advisory Board
  FFELP - Federal Family Education Loan Program
  FHA - Federal Housing Administration
  FSA - Farm Service Agency
  HUD - Department of Housing and Urban Development
  MMI - Mutual Mortgage Insurance
  OIG - Office of the Inspector General
  OMB - Office of Management and Budget
  RHS - Rural Housing Service
  SBA - Small Business Administration
  SFFAS - Statement of Federal Financial Accounting Standards
  USDA - Department of Agriculture
  VA - Department of Veterans Affairs
  VBA - Veterans Benefits Administration

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


B-277070

March 30, 1998

The Honorable Pete V.  Domenici
Chairman, Committee on the Budget
United States Senate

Dear Mr.  Chairman: 

The Federal Credit Reform Act of 1990\1 was enacted to require that
the budget reflect a more accurate measurement of the government's
subsidy costs for federal direct loans and loan guarantees and to
permit better cost comparisons both among credit programs and between
credit and non-credit programs.  The credit subsidy cost is the
government's estimated net cost, in present value terms, of direct or
guaranteed loans over the entire period the loans are outstanding. 
Because the government now has over 6 years of experience with credit
reform, you asked us to determine (1) whether agencies completed
estimates and reestimates of subsidy costs, (2) whether we could
readily discern any trends including improvements in subsidy
estimates, and (3) whether we could readily identify the causes for
changes in subsidy estimates.  You also asked us whether agencies
with discretionary credit programs initially underestimated credit
subsidy costs in response to the incentive created by the
availability of permanent, indefinite budget authority for credit
reestimates.\2

We reviewed subsidy estimates and supporting documentation prepared
for two domestic credit programs at each of the five largest credit
agencies--the Departments of Agriculture (USDA), Education, Housing
and Urban Development (HUD), and Veterans Affairs (VA), and the Small
Business Administration (SBA).  Each of the agencies confirmed that
the reported data were an accurate reflection of the data they had. 
While we did not independently verify that data, we found, and agency
staff later agreed, that there were problems with these data, as
discussed later in the report.  The causes of changes in subsidy
estimates reported in appendix II are those identified by credit
agencies' staff. 


--------------------
\1 Appendix I contains a summary discussion of credit reform
requirements. 

\2 Initial estimates of subsidy costs compete under caps, while
subsidy reestimates are covered by permanent, indefinite budget
authority. 


   BACKGROUND
------------------------------------------------------------ Letter :1

There is a long history of problems with federal credit programs.  We
reported many of the problems, such as poor recordkeeping and cost
data, prior to credit reform.\3 For example, we reported in November
1989 that federal agencies' long-standing deficiencies in financial
management systems and accounting procedures had precluded accurate,
comprehensive recording and reporting of the full extent of credit
losses.\4 Agencies have had perennial problems tracking loan payments
due and loan guarantees made in federal budgets.  Also, prior to
credit reform, the cash-based budget distorted choices between direct
loans and loan guarantees.  A direct loan initially looked like a
grant since the budget included as a cost the face value of a direct
loan, ignoring that at least some part of the loan would be repaid. 
Conversely, loan guarantees looked free when they were made because
the budget ignored the fact that some would result in default costs. 
The Office of Management and Budget (OMB), GAO, the Congressional
Budget Office (CBO), and others reported on the need to change the
way credit programs were budgeted. 

In response to these reports and a growing recognition of federal
financial management problems, the Congress enacted a series of laws
designed to improve financial management and the quality and use of
cost data in decision-making.  To address the deficiencies in
recognizing the cost of credit programs, the Federal Credit Reform
Act of 1990 was enacted as part of the Omnibus Budget Reconciliation
Act of 1990.  Credit reform was intended to ensure that the full cost
of credit programs would be reflected in the budget so that the
executive branch and the Congress might consider these costs when
making budget decisions.  Accounting standards for credit programs
were developed to be consistent with the intent of this act.  To
address broader problems in federal financial management, the Chief
Financial Officers (CFO) Act of 1990 required the development and
maintenance of integrated agency accounting and financial management
systems, including financial reporting and internal controls, that
provide for development and reporting of cost information.  The
Government Management Reform Act of 1994 expanded the CFO Act to
provide for audits of the annual financial statements of the 24 CFO
agencies.  The largest credit programs are found in these agencies
and audits include a review of agencies' subsidy estimates and actual
loan performance data.  Accurate cost information also is key to
improvements in the efficiency and effectiveness of federal programs
as envisioned by the Government Performance and Results Act of 1993. 
The Debt Collection Act of 1982 provided for OMB to require agencies
to report debt information to OMB and to the Department of the
Treasury.  Finally, the Debt Collection Improvement Act of 1996
expanded collection tools and authorities available to agencies and
called for centralized servicing of some debt.  Federal financial
management, including credit program management, continues to reap
the benefits of these laws. 

While all of these laws sought to improve federal financial
management, a major change for budgeting was the Federal Credit
Reform Act included in the Omnibus Budget and Reconciliation Act of
1990.  This act changed the budgetary treatment of credit programs so
that their costs could be compared more appropriately both with each
other and with other federal spending.  Credit reform requires
agencies to estimate the net cost to the government over the full
term of the credit of new direct or guaranteed loans to be made in
the budget year and to record that cost in the budget on a
present-value basis.  Unless OMB approves an alternative proposal,
agencies are required to reestimate this cost annually as long as any
loans in the cohort\5 are outstanding.  The Balanced Budget Act of
1997 amended the Federal Credit Reform Act to simplify and clarify
subsidy estimation requirements.  OMB also has simplified guidance
for credit subsidy estimation.  Appendix I contains a more detailed
description of credit reform requirements and recent changes. 

Credit programs may be either discretionary\6 or mandatory\7 as
defined in the Budget Enforcement Act of 1990.  Appropriations for
the subsidy cost of discretionary credit programs are counted under
the discretionary spending caps and so must compete with other
discretionary programs for the funding available under these limits. 
Like other mandatory programs, mandatory credit programs receive
automatic appropriations for whatever amount of credit is needed to
meet the estimated demand for services by beneficiaries. 

All credit programs automatically receive any additional budget
authority that may be needed to fund reestimates.\8 For discretionary
programs this means there is a difference in the budget treatment of
original subsidy cost estimates and of subsidy cost reestimates.  The
original estimated subsidy cost is counted under the discretionary
caps, but any additional appropriation for upward reestimates of
subsidy cost is exempt from the caps.  This design could result in a
tendency to underestimate the initial subsidy costs of a
discretionary program.  Portraying a loan program as less costly than
it really is when competing for funds under the discretionary caps
means more or larger loans or loan guarantees could be made with a
given appropriation since the program then could rely on automatic
funding for subsequent reestimates to cover any shortfalls.  This
built-in incentive is one reason to monitor subsidy reestimates. 
Monitoring reestimates is a key control over tendencies to
underestimate costs as well as a barometer of the quality of
agencies' estimation processes. 

The development of credit reform requirements reflects in part
decisionmakers' interest in analyzing the causes of changes in
subsidy estimates.  Understanding which of the components of subsidy
expense--interest, net defaults, fees and other collections, and
other subsidy costs--are the key drivers of reestimates can both
improve the quality of estimates and yield insights into program
operations.  OMB developed and provided agencies with a computer
model to calculate the total estimated subsidy rate and the
components of subsidy expense based on agency-developed cash flow
information.  In the development of accounting standards for credit
programs, the Federal Accounting Standards Advisory Board (FASAB)
indicated that these data would be valuable for making credit policy
decisions, monitoring portfolio quality, and improving credit
performance.\9

Current accounting standards and OMB guidance require agencies to
recognize, and disclose in the financial statement footnotes, the
four components separately for the fiscal year during which direct or
guaranteed loans are disbursed.  However, for programs that disburse
over more than 1 year, the current disclosure aggregates subsidy
component data for the current year with subsidy costs from prior
years.  In addition, changes in law and program administration often
occur.  Thus, loans disbursed from programs over multiple years have
different program characteristics and the current year's financial
statement disclosures do not represent the program characteristics or
expenses of any given year of the program.  Because the requirement
in its present form does not provide the kind of useful information
that was intended, FASAB now is considering revising these standards. 

Agencies now have prepared eight budgets under credit reform
requirements and there should be 6 years of actual data available. 
Because of different program requirements, resource and expertise
levels, and levels of commitment and interest, agencies have taken
different approaches to making subsidy estimates.  Preparing subsidy
estimates is complex for a number of reasons, including projecting
cash flows many years into the future and assessing the effect of
economic changes on a particular program and its borrowers.  Further,
in some--if not all--agencies, budget office staff must integrate
information from staff and systems in both the finance and program
offices.  While the present value-based budgeting of credit reform is
a major step forward, its success depends heavily on the quality of
these complex subsidy estimates.  The independent review of agency
records and data in the annual financial audit is an important step
in monitoring the subsidy estimates and improving their reliability. 

When credit reform was enacted, it generally was recognized that
agencies did not have the capacity to implement fully the needed
changes in their accounting systems in the short run and that the
transition to budgeting and accounting on a present-value basis would
be difficult.  However, policymakers expected that once agencies
established a systematic approach to subsidy estimation based on
auditable assumptions, present value-based budgeting for credit would
provide them with significantly better information than the former
cash-based system.  Despite the difficulties with implementation,
including current data problems, present value-based reporting for
credit avoids a number of the problems of cash reporting.  Therefore,
we believe that making credit reform work is important. 


--------------------
\3 Financial Management:  Additional Actions Needed to Improve
Federal Financial Management Systems (GAO/AFMD-90-14, April 27, 1990)
and Financial Audit:  Veterans Administration's Financial Statements
for Fiscal Years 1987 and 1986 (GAO/AFMD-89-23, November 30, 1988). 

\4 Federal Credit and Insurance:  Programs May Require Increased
Federal Assistance in the Future (GAO/AFMD-90-11, November 16, 1989). 

\5 A cohort includes those direct loans or loan guarantees of a
program that are subsidized by an appropriation for a given fiscal
year even if disbursements occur in subsequent years. 

\6 Funding for discretionary spending programs is provided in
appropriations acts. 

\7 Mandatory programs generally are entitlement programs for which
the amount of funding depends on eligibility and benefits rules
contained in substantive law. 

\8 It was recognized that data were limited or unreliable in the
early years of credit reform.  This could impede the ability of
agencies to make reliable estimates.  Permanent, indefinite budget
authority for upward reestimates of subsidy costs was provided. 
Agencies with discretionary credit programs then could reestimate
subsidy costs as required without being limited by the constraints of
budgetary spending limits. 

\9 FASAB was created by OMB, Treasury, and GAO to consider and
recommend accounting principles for the federal government.  It
published Statement of Federal Financial Accounting Standards (SFFAS)
No.  2, Accounting for Direct Loans and Loan Guarantees, in July
1993. 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :2

After over 6 years of experience with credit reform, agencies
continue to have problems in estimating the subsidy cost of credit
programs.  The lack of timely reestimates as well as the frequent
absence of documentation and reliable information limit the ability
of agency management, OMB, and the Congress to exercise intended
oversight.  We found problems with the availability and reliability
of subsidy estimates, reestimates, and supporting documentation in
our cross-cutting review of 10 programs.  None of the 10 programs in
our study had for our review all of the required budget request,
budget execution,\10 and reestimate subsidy rates and supporting
documentation for fiscal years 1992 through 1998.  While
documentation generally is more available in recent years, the
availability of timely reestimates of subsidy rates did not improve
over time.  Of the 10 programs we examined, 8 either failed to do
reestimates for certain years or produced them too late to be
included in the next budget formulation or audit cycles.  The subsidy
costs of one program--the SBA Disaster Loan Program--were not
reestimated for budgets or the financial statement audits during this
period but were reestimated for the fiscal year 1999 Budget.  The
programs that did not make reestimates (HUD, USDA, and SBA) all
received waivers from OMB. 

Data used as the basis for subsidy estimates were not always
reliable.  In the audits of the fiscal year 1996 financial
statements, three of the five largest credit agencies received
disclaimers or qualified opinions related to their credit programs. 
Auditors were unable to find support for agency data on such items as
delinquencies and prepayments for loans receivable and liabilities
for loan guarantees.  This also could reflect problems with
historical data since agencies with loan guarantee programs rely on
lenders or intermediaries for loan performance data.  Moreover, we
found--and agency officials acknowledge--discrepancies between the
subsidy rates reported in the President's Budget and those provided
to us by the agencies as well as discrepancies within the data
provided to us by the agencies.  In other cases, agencies were unable
to provide supporting documentation for the numbers in the Budget. 

Problems with the reliability and validity of the underlying credit
data raise questions about the basis for the subsidy estimates
included in the Budget.  Agencies have had several years to obtain
and refine historical data and estimation methodologies.  Over time,
we would expect to find that, for a given cohort, the annual changes
in reestimates due to technical factors would be smaller.  Because
component data were not available, we could not determine whether
this had occurred.  However, we did note that overall subsidy rates
for a given cohort varied widely.  For example, estimates and
reestimates of the fiscal year 1992 cohort of Education's Federal
Family Education Loan Program (FFELP) changed direction each of the 6
years for which we had data--estimated subsidy rates ranged from
26.30 percent to 16.99 percent. 

We observed a similar pattern of fluctuations in subsidy estimates at
the program level.  Subsidy rate estimates for any given program
continued to fluctuate widely from year to year with no pattern or
particular trend.  For example, the subsidy estimates and reestimates
of the fiscal years 1992 through 1997 cohorts in VA's Loan Guaranty
Direct Loan Financing Account\11 changed direction in each of the 6
years of data and estimated subsidy rates ranged from 4.80 percent to
1.18 percent.  Subsidy rate estimates can change for many reasons,
including programmatic redesign and changes in the economy.  For
example, a change in the discount rate will cause the subsidy rate to
change, even if the cash flows are unaffected.  Financial statement
audit findings for the credit agencies we reviewed would lead us to
conclude that at least some of the fluctuations are caused by
weaknesses in agency data used to develop the cash flow estimates. 

The intersection of credit reform and the Budget Enforcement Act of
1990--that is, the fact that original subsidy appropriations must
compete under the discretionary caps while reestimates are outside
them--may offer an incentive for agencies with discretionary programs
to underestimate subsidy costs initially to permit more loans or loan
guarantees within a given appropriation level.  If this incentive did
not exist and absent any overriding economic trend or revision of
historical data, one would expect reestimates to lower the original
subsidy estimate as often as they raise it, and the patterns would be
similar for discretionary and mandatory programs.  However, available
data were not sufficient to assess whether a credit program's
budgetary treatment affected its initial subsidy estimates.  We found
somewhat similar patterns when we compared discretionary and
mandatory programs.\12 We found that the estimated subsidy rates for
8 of the 15 discretionary cohorts increased in the first reestimate
following the initial appropriation, while first reestimates for 7 of
the 12 mandatory cohorts decreased.  This result is not conclusive. 
No real conclusions can be drawn from this observation about whether
some discretionary programs may have sought to benefit from initially
underestimating subsidy costs.  Other factors such as changes in the
economy--especially in interest rates--data errors, or more
historical data may have contributed to changes in reestimates. 

Better information on factors underlying changes in subsidy rates is
needed to identify and understand why these estimates change.  In
theory, data on the four components of subsidy expense--interest, net
defaults, fees, and other subsidy costs--calculated by the OMB
subsidy model as part of the estimation process could be used to
identify possible causes of changes.  However, we could not perform
such an analysis because component data frequently were missing or
inaccurate.  Accordingly, meaningful component data were not
available to be used internally or by OMB for budget formulation or
program management. 

While OMB provides a user's guide and some training on the subsidy
model, OMB has not provided agencies with clear definitions of each
component or sufficient guidance on how to use the OMB subsidy model
to correctly calculate the components.  If OMB did so, agency staff
could improve the consistency and accuracy of component data, making
it useful for budget decision-making and oversight as envisioned by
budget and accounting requirements.  For example, we recently used
the component data to identify a large error in SBA's subsidy
estimates.\13 Although SBA's component data were flawed, it provided
a quick indication that there was an error in the fiscal year 1997
subsidy estimates for the section 7(a) General Business Loan Program. 
Correcting this error enabled SBA to guarantee approximately $2.5
billion more in section 7(a) small business loans.  OMB and SBA
officials acknowledged that better oversight and improved internal
controls at both OMB and SBA are needed to prevent similar errors in
the future. 

While no single agency yet is successful in all aspects of credit
reform implementation, some progress is being made at each of the
agencies we studied.  The increasing audit attention on credit
budgeting and accounting has focused efforts on improving subsidy
estimation.  Sustained greater commitment by agency management is
needed to continue this progress and, to succeed, it is important
that OMB continue to be a part of this effort. 


--------------------
\10 The budget execution estimate is made after an agency receives an
appropriation and when the agency obligates the government for a
direct loan or makes a loan guarantee commitment. 

\11 These estimates and reestimates were completed for the fiscal
year 1997 President's Budget. 

\12 Mandatory and discretionary programs are treated differently
under the Budget Enforcement Act.  Discretionary programs compete
under fixed-dollar caps.  Mandatory credit programs are automatically
funded for whatever amount of credit is needed for a given program
design and set of program beneficiaries.  As a result, for mandatory
programs there would not be an incentive to initially underestimate
subsidy costs. 

\13 Credit Subsidy Estimates for the Sections 7(a) and 504 Business
Loan Programs (GAO/T-RCED-97-197, July 16, 1997). 


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

The objectives of our work were to determine (1) whether agencies
completed estimates and reestimates of subsidy costs, (2) whether we
could readily identify any trends including improvements in subsidy
estimates as reported by the agencies, and (3) whether we could
readily identify the causes for changes in subsidy estimates.  You
also asked us whether agencies with discretionary credit programs
initially underestimated credit subsidy costs in response to the
incentive created by the availability of permanent, indefinite budget
authority for credit reestimates.  We selected a sample of 10
programs from the five agencies with the largest domestic federal
credit programs:  the Departments of Agriculture, Education, Housing
and Urban Development, and Veterans Affairs, and the Small Business
Administration.  We generally selected programs that have the most
credit outstanding or highest loan levels.  Both direct loan and loan
guarantee programs are represented.  Table 1 in the following
discussion of the availability of subsidy estimates and supporting
documentation contains a list of the 10 programs we examined. 

We requested that agencies provide budget data and information for
the selected programs for fiscal years 1992 through 1998.  The data
requested included (1) descriptions of the credit program and
highlights of program changes over the years, (2) spreadsheets
showing estimated or reestimated cash flows of each cohort, (3) input
to and output from OMB's credit subsidy model, and (4) documentation
of agency efforts to revise the subsidy estimation process.  For each
cohort in fiscal years 1992 through 1998, we extracted the subsidy
rate estimates used in the President's budget request, budget
execution, and all reestimates.  These data are included in appendix
III. 

Our work reports the subsidy rate data and documentation as provided
by the agencies.  We interviewed staff who prepared the subsidy
estimates and obtained written confirmation from each agency that the
data in the tables in appendix III were accurate and represented all
of the data the agency had.  However, we found problems with these
data.  While we did not independently verify the accuracy of these
data, we did compare the budget request subsidy rates confirmed by
the agencies to the rates reported in the appropriate Budget Appendix
and Budget Credit Supplement.  We found that agency-confirmed rates
differed from the Budget in nine instances although only three
differences were greater than half a percentage point.\14 In two of
these three instances, the agencies later provided documentation to
support the rates in the Budget.  In the third instance, we used the
rate produced by the OMB subsidy model because the agency's cash flow
spreadsheets best supported it.  In addition, we reviewed recent
financial statement audit reports for these credit agencies and
programs as one gauge of the reliability of these data. 

While we examined data for all 10 programs in the 5 credit agencies,
specific examples used in our work discuss only those programs or
agencies with comparable data.  For example, we had comparable data
from only seven programs to use in our analysis of the most recent
subsidy estimates for the fiscal year 1992 and fiscal year 1998
cohorts.  This is because all of VA's credit programs (including some
not examined in this report) were consolidated into two programs for
fiscal year 1998 (a direct loan program and a loan guarantee
program)\15 and Education's direct loan program only began in fiscal
year 1994. 

We used the data in appendix III to try to identify trends in subsidy
estimates.  Appendix II includes graphs of total subsidy rates by
cohort for nine programs and graphs profiling the subsidy rates of a
given program cohort over time for eight programs.  We did not
prepare either graph for SBA's Disaster Loan Program because SBA had
not included subsidy reestimates for any cohort in the President's
Budget prior to fiscal year 1999.  We also did not profile a cohort
of USDA's Farm Operating Loans because we did not have enough
comparable data for our review. 

To further understand the causes for changes in subsidy rates, we
then analyzed the four components of subsidy expense (interest, net
defaults, fees and other collections, and other subsidy costs)
required to be reported by SFFAS No.  2 and calculated by OMB's
subsidy model.  We also compared the budget execution estimate to the
first reestimate for all credit programs and analyzed whether there
was a different pattern in the direction of the reestimates for
direct loan programs and loan guarantee programs or for mandatory and
discretionary programs. 

Our work was conducted in Washington, D.C., from September 1996
through January 1998 in accordance with generally accepted government
auditing standards.  We requested comments on a draft of this report
from the following officials or their designees:  the Director of the
Office of Management and Budget, the Secretary of Agriculture, the
Secretary of Education, the Secretary of Housing and Urban
Development, the Administrator of the Small Business Administration,
and the Acting Secretary of Veterans Affairs.  All of the entities
provided written comments, which are discussed in the "Agency
Comments and Our Evaluation" section and reprinted in appendixes IV
through IX. 


--------------------
\14 After conversations with agency staff about the other six
instances where rates differed, we generally decided to use the rates
provided by the agencies because documentation was available to
support these rates. 

\15 This consolidation of programs was authorized by the 1998
Department of Housing and Urban Development, Department of Veterans
Affairs, and Related Agencies Appropriations Act. 


   PROBLEMS PERSIST WITH AGENCIES'
   ESTIMATES OF SUBSIDY COST
------------------------------------------------------------ Letter :4

In each of the agencies in our study, we found problems with the
availability of estimated subsidy rates and supporting documentation
and with the reliability of the subsidy rate estimates.  In 8 of the
10 programs we examined, agency staff either failed to do reestimates
for certain years or completed them too late to be included in the
budget formulation or audit cycles.  While some progress has been
made at some agencies, audits of financial statements continue to
show that serious problems remain.  Effective implementation of
credit reform is highly dependent on the availability of accurate
data. 


      SUBSIDY RATE ESTIMATES AND
      SUPPORTING DOCUMENTATION
      WERE NOT CONSISTENTLY
      AVAILABLE
---------------------------------------------------------- Letter :4.1

Agencies did not have for our review all of the estimated subsidy
rates and supporting documentation for the seven budgets they
prepared under credit reform for fiscal years 1992 through 1998. 
Also, the availability of subsidy rate estimates and reestimates
generally did not improve over time.  All nine programs\16 in
existence in 1992 when credit reform became effective had for our
review budget request and budget execution subsidy estimates for the
first 2 years under credit reform--fiscal years 1992 and 1993.  The
first reestimates of subsidy cost should have been done for the
fiscal year 1994 Budget.  Of the 10 programs we examined, 8 either
failed to do reestimates for certain years or produced them too late
to be included in budget formulation or audit cycles.  Only five of
the nine programs that should have completed reestimates for the
fiscal years 1994 and 1995 Budgets had them for our review.  Starting
with the fiscal year 1997 Budget, OMB Circular A-11 provided that
agencies could forgo completing reestimates under certain
circumstances.  For fiscal years 1997 and 1998, 5 of the 10 programs
made timely reestimates.  One of the programs, SBA's Disaster Loans
Program, did not have subsidy reestimates in the Budget until the
recently released fiscal year 1999 Budget.  The programs that did not
make reestimates generally reported staff constraints in completing
reestimates while preparing their budget request or difficulties
obtaining data on which to base reestimates.  All reported that they
had received waivers from OMB.  We received written copies of OMB
waivers to USDA and HUD permitting late completion of estimates. 
While USDA's waiver was effective for fiscal year 1996 and future
budgets, HUD's waiver was only for fiscal year 1997.  HUD staff told
us that OMB gave them waivers orally for other years, and OMB did not
disagree.  SBA also told us that the waivers were approved by OMB
orally, and OMB agreed. 

On the other hand, for estimates and reestimates that were made, the
availability of documentation has improved somewhat.  In fiscal years
1996 through 1998, most programs did have documentation for estimates
and reestimates that were made.  Recent credit reform guidance
explicitly requires agencies to document subsidy estimation.\17 In
the early years of credit reform, fiscal years 1992 through 1995,
most programs did not have for our review supporting documentation
for all completed budget estimates and reestimates.  An official at
Education characterized early credit program files as "woefully
lacking."\18 HUD staff said that early credit estimates were prepared
by OMB staff and that HUD and OMB do not have supporting
documentation. 

Table 1 shows, by program and fiscal year, whether agencies had for
our review all subsidy estimates for the budget request and budget
execution, made subsidy reestimates, and had supporting documentation
for these estimates that we requested for our review.  A check
indicates that the agency had for our review all estimated rates,
rates for reestimates they made, and had all supporting documentation
contained in output from the OMB subsidy model and agency cash flow
spreadsheets.  It says nothing about the quality of the data.  We
discuss in some detail our concerns about data reliability later in
this report. 



                                     Table 1
                     
                       Budget Request and Budget Execution
                       Subsidy Estimates, Reestimates, and
                             Supporting Documentation

                               President's Budget for Fiscal Year
                ----------------------------------------------------------------
Agency/
Program
(Loan
type)           1992      1993      1994      1995     1996     1997     1998
----------  --  --------  --------  --------  -------  -------  -------  -------
USDA: Farm
Service
Agency,
Farm
Operating
Loans
Program
(Direct)

USDA:                                        
Rural
Housing
Service,
Single
Family
Housing
Program
(Direct)

Education:
Federal
Family
Education
Loan
Program,
Stafford
Loans
(Guarantee
)

Education:      \a        \a
Wm. D.
Ford
Direct
Loan
Program,
Stafford
Loans
(Direct)

HUD:
Mutual
Mortgage
Insurance
Fund
(Guarantee
)

HUD:
Housing
Program's
General
and
Special
Risk
Insurance
Fund Sect.
223(f)
Refinance
(Guarantee
)

SBA: 7(a)                                                              
General
Business
Loans
Program
(Guarantee
)

SBA:
Disaster
Loans
Program
(Direct)

VA:                                                                  
Guaranty
and
Indemnity
Fund
(Guarantee
)

VA: Loan                                                              
Guaranty
Direct
Loan
(Direct)
--------------------------------------------------------------------------------
Note:   indicates agencies that had for our review (1) all estimated
budget request and budget execution subsidy rates and reestimated
subsidy rates and (2) all supporting documentation from OMB's credit
subsidy model and cash flow spreadsheets. 

\a The William D.  Ford Direct Loan Program was established in fiscal
year 1994. 


--------------------
\16 The William D.  Ford Direct Loan Program was established in
fiscal year 1994. 

\17 OMB Circular A-11 has required agencies to document how a credit
program's subsidy is calculated since the 1995 version issued as
guidance for preparation of the fiscal year 1997 Budget.  SFFAS No. 
2 discusses using a systematic methodology and creating and
maintaining a consistent database; it does not address documentation
of the estimation process.  An issue paper, Model Credit Program
Methods and Documentation for Estimating Subsidy Rates and the Model
Information Store (96-CR-7, May 1, 1996), developed by the
Government-wide Audited Financial Statements Task Force, Subgroup on
Credit Reform, discusses reasonable methods of estimating subsidy
rates and recommends using an auditable subsidy estimation procedure
that formalizes and documents loan performance assumptions.  It
discusses the advantages of a documented model--greater availability
for update, more conducive to review and comment, and more readily
transferrable between analysts. 

\18 Because we requested that agencies provide output from the OMB
subsidy model and the underlying cash flow spreadsheets as
documentation of their subsidy estimates, the two Education programs
are shown in table 1 as having incomplete documentation in each year. 
Education uses the OMB subsidy model earlier in its estimation
process than other agencies, adjusting the results to develop the
subsidy rate used in the President's Budget.  Thus, Education's OMB
model output and cash flow spreadsheets do not support the rates
included in the Budget.  Further, Education did not provide alternate
supporting documentation for those subsidy rates. 


      RELIABILITY OF SOME SUBSIDY
      ESTIMATE DATA IS
      QUESTIONABLE
---------------------------------------------------------- Letter :4.2

The reliability of credit data is questionable for a number of
reasons including (1) the poor results of financial statement audits,
(2) discrepancies we found between subsidy rates reported in the
President's Budget and the data confirmed to us by the agencies, (3)
subsidy rate estimates not always supported by documentation, (4)
acknowledgements from some staff that component data were
questionable, and (5) staff reports of difficulties with systems
support.  First, financial statement audits raised questions about
data reliability.  Three of the largest credit agencies, HUD, USDA,
and Education, received disclaimers or qualified opinions\19 on their
fiscal year 1996 financial statement audits, in part, because of
problems associated with their credit programs.  HUD received a
qualified opinion because the Federal Housing Administration's (FHA)
credit-related accounts were not reported on the present value basis
required by SFFAS No.  2.\20 Consequently, HUD's Office of the
Inspector General (OIG) was unable to audit the credit-related
account balances.  USDA's OIG gave a qualified audit opinion on the
fiscal year 1996 financial statements of the rural development
mission area because it was unable to obtain sufficient support for
credit program receivables and estimated losses on loan guarantees
and the related credit reform program subsidy and appropriated
capital used.  USDA's Farm Service Agency Farm Operating Loans were
included as part of the consolidated audit of USDA's fiscal year 1996
financial statements which received a disclaimer of opinion.  The
Farm Service Agency does not prepare separate financial statements. 
Education received a disclaimer of opinion on the department's fiscal
year 1996 financial statements because it was unable to support the
reasonableness of the amounts shown for loans receivable and
liabilities for loan guarantees.  Education's OIG also was unable to
render an audit opinion due to auditor concerns about the quality of
the data supporting subsidy estimates of the Federal Family Education
Loan Program.  This could reflect problems with historical data since
agencies with loan guarantee programs rely on lenders or
intermediaries for loan performance data.  In some cases, the
auditors recommended that agencies establish and document a process
for the development of subsidy estimates and make reestimation of
subsidy costs a priority. 

Although VA and SBA both received unqualified audit opinions on their
fiscal year 1996 financial statements, the auditors of these agencies
reported internal control weaknesses related to estimating credit
subsidies.  For example, SBA's auditors reported that the agency used
inconsistent and unreliable data to reestimate the Disaster Loan
Program.  Because of the questionable data, OMB and SBA had not
included any reestimate of this program until the fiscal year 1999
Budget or the fiscal year 1997 financial statements.  In addition,
VA's auditors reported that the agency does not efficiently and
reliably accumulate the financial information needed to comply with
credit reform accounting requirements and that significant credit
reform-related adjustments were necessary to the financial
statements. 

Second, we found discrepancies between the subsidy rates reported in
the President's Budget and the data provided and confirmed to us by
the agencies\21

in about 10 percent of our sample (7 of 68 rates).  Some agencies
agreed, after we pointed out inconsistencies, that certain data they
had provided and certified as accurate in fact were not.  Third,
agencies also had for our review subsidy rate estimates that were not
always supported by the documentation.  For example, none of
Education's documentation supported its estimated subsidy rates. 
Also, agencies, such as HUD, had difficulty identifying the fiscal
year of available subsidy rates and documentation and whether the
rates were budget execution estimates or reestimates.  Fourth, staff
from seven of the eight programs whose component data we were able to
examine acknowledged that the data were questionable.  We found that
component data were questionable because they were not consistent
with program characteristics.  For example, VA's direct loan program
showed the net default component as negative because cash flows from
loan sales were included with recoveries from defaulted loans.  As a
result, recoveries exceeded defaults.  To avoid erroneously
indicating that higher defaults would reduce the subsidy rate,
proceeds from loan sales should have been included with the "other
cash flow" component. 

Fifth, and finally, data reliability depends in part on having
adequate information systems, and effective top management commitment
is vital to ensuring that these are provided.  Today, as 4 years ago,
staff in most agencies we examined report difficulties with systems
support.  For example, staff in three of the five agencies we
reviewed--HUD, VA, and USDA--reported inadequate actual data on loan
performance and computer systems support.  These agencies have
efforts underway to refine their data and/or improve their estimation
processes.  Staff at USDA and VA have worked with their offices of
the inspector general or OMB to refine their cash flow spreadsheets
and reestimate calculations.  HUD staff reported efforts underway to
develop a new integrated accounting system and a need to re-engineer
budget and accounting processes. 

Staff in the other two agencies--Education and SBA--told us that
systems support was not an issue and reported that credit reform
implementation has become a high-level priority for their agencies. 
Specifically, SBA administrators had a commitment to meet the
requirements of the Federal Credit Reform Act and sought to improve
their capacity to make better subsidy cost estimates.  Both agencies
reported that they had developed new computer systems, significantly
refined their historical information stores, and are using contractor
support.  Although upper-management commitment is necessary, it does
not result in instant improvement.  Audits continue to report
problems at both agencies, both have difficulty obtaining historical
information, and SBA and others have identified errors in SBA subsidy
estimates. 

OMB and credit agency staff acknowledge that budget and financial
systems for credit programs could have problems with conversions to
the year 2000.\22 In a February 1998 report on agencies' progress in
addressing Year 2000 conversion, OMB reported that SBA and VA were
demonstrating sufficient progress, USDA and HUD were making some
progress but OMB still had concerns, and Education was making
insufficient progress.  In testimony in September 1997,\23 we
reported that the Veterans Benefits Administration, where its housing
credit programs are located, has developed an agencywide plan and
created a program management organization but will need to strengthen
management and oversight of Year 2000-related activities to avert
serious disruption of its ability to disseminate benefits.  Since our
testimony, VA has taken action to address some of our concerns. 


--------------------
\19 A disclaimer of opinion on financial statements is issued
whenever auditors are unable to satisfy themselves that the overall
financial statements are fairly presented.  A disclaimer may arise
because of a severe limitation in the scope of the audit, perhaps due
to a lack of documentation and/or uncertainties about the amount of
an item or outcome of a matter that materially affects financial
position.  A qualified opinion on financial statements can be used
only when the auditor believes that the overall financial statements
are fairly stated other than with regard to the noted qualification. 
The qualification may result from a limitation on the scope of the
audit, failure to follow generally accepted accounting principles,
use of different accounting principles during one of the years
included in the statements, or circumstances that prevented the
auditor from knowing that the statements were fairly presented. 

\20 Because FHA is a government corporation, it reports costs in its
financial statements in accordance with private sector accounting
standards.  However, because FHA's financial results are material to
HUD's financial statements, it should comply with SFFAS No.  2 when
it is included in HUD's consolidated financial statements.  By not
complying with credit reform's accounting standards, HUD is not
accurately reporting the costs of its programs in its financial
statements. 

\21 Although it is unclear without further investigation which rates
were accurate, we generally included in appendix III tables those
rates provided by the agencies because they were supported by some
documentation. 

\22 On January 1, 2000, computer systems worldwide could malfunction
or produce inaccurate information simply because the date has
changed.  Unless corrected, such failures could have a costly,
widespread impact.  The problem is rooted in how dates are recorded
and computed.  For the past several decades, systems have typically
used two digits to represent the year--such as "97" for 1997--to save
electronic storage space and reduce operating costs.  In such a
format, however, 2000 is indistinguishable from 1900.  This ambiguity
could cause systems to malfunction in unforeseen ways or to fail
completely. 

\23 Veterans Affairs Computer Systems:  Action Underway Yet Much Work
Remains To Resolve Year 2000 Crisis (GAO/T-AIMD-97-174, September 25,
1997). 


   SUBSIDY RATES GENERALLY
   FLUCTUATED
------------------------------------------------------------ Letter :5

Over time, some fluctuation in subsidy rates would be expected within
a given group of loans or guarantees (a cohort) and among different
cohorts of the same program.  Reasons include loans or guarantees
made at different interest rates than anticipated; programmatic
redesign; better information on technical factors such as defaults,
prepayments, and fees from more actual experience; or unanticipated
changes in the economy including interest rate changes.  Agencies
have had several years to obtain and refine historical data and
estimation methodologies.  Over time we would expect to find that,
for a given cohort, the annual changes in reestimates due to
technical factors would be smaller.  We believe that agencies can
improve their abilities to forecast such factors as defaults,
recoveries, prepayments, and fee revenue through better modeling and
more and better historical data.  As this improvement occurs, the
variability in the subsidy rate from year to year for a given cohort
caused by these factors (as opposed to economic factors such as
interest) should diminish.  Because reliable component data were not
available, we could not readily determine whether this had occurred. 
However, total subsidy estimates within a given cohort often varied
widely over time.  Moreover, estimates for different cohorts within
the same program also differed widely.  These sharp variations raise
questions about the causes for these changes and the reliability of
the underlying data.  The success of credit reform budgeting relies
on reasonable estimates informed by timely, appropriate actual data. 

We analyzed the data from several perspectives: 

  -- Comparing Estimated Subsidy Rates Over Time For a Given
     Cohort--In the programs we examined, we found that reestimates
     often were large and changed direction over time--increasing or
     decreasing estimated subsidy cost.  During the period of time
     between a given cohort's budget execution estimate and its most
     recent reestimate, we found that reestimates generally
     fluctuated both up and down.  A total of 74 percent of the 23
     cohorts we analyzed\24 had at least one reestimate increasing
     the subsidy rate and at least one reestimate that decreased the
     subsidy rate.  For example, estimates and reestimates of the
     fiscal year 1992 cohort of Education's FFELP program changed
     direction each of the 6 years of data, estimated subsidy rates
     ranged from 26.30 percent to 16.99 percent.  This could result
     from any number of factors, including changes in assumptions
     about cash flows as agencies gained experience in estimating
     subsidy rates or developed better actual loan data, and changes
     in the timing of loan activity or interest rates.  Only VA's
     Loan Guaranty Direct Loan Financing Account did not have both
     increases and decreases in its subsidy estimate, as shown in
     figure II.17.  Graphs of selected cohorts are included in
     appendix II.  Viewed another way, we compared two specific
     subsidy rate estimates--budget execution and the first
     reestimate--in eight programs for which we had appropriate data
     (27 cohorts).  We found that the first reestimates of the
     subsidy rates were higher than the budget execution rates for 13
     cohorts and were lower for 14 cohorts.  We also found that 15 of
     the 27 cohorts had changes of at least 20 percent--7 increases
     and 8 decreases.  One cohort in the SBA 7(a) program increased
     by 126 percent. 

  -- Comparing Estimated Subsidy Rates for Different Cohorts of a
     Given Program--We also analyzed the estimated subsidy rates for
     a given program by comparing the most recent estimates or
     reestimates for different cohorts.  For example, as shown in
     figure II.16, the subsidy estimates and reestimates of the
     fiscal years 1992 through 1997 cohorts in VA's Loan Guaranty
     Direct Loan Financing Account\25 changed direction in each of
     the 6 years of data, and estimated subsidy rates ranged from
     4.80 percent to 1.18 percent.  We found that the fiscal year
     1998 President's Budget showed that six of eight programs had a
     lower estimated subsidy rate for their new fiscal year 1998
     credit than they reestimated for their fiscal year 1992 credit,
     the oldest cohorts in our study.  Only HUD's Federal Housing
     Administration's Mutual Mortgage Insurance Fund and VA's Loan
     Guaranty Direct Loan Financing Account had estimated subsidy
     rates for fiscal year 1998 credit that were higher than the
     reestimates of the fiscal year 1992 cohort.  This relatively
     consistent pattern of lower estimated subsidy rates in fiscal
     year 1998 may reflect changes in economic conditions such as
     lower interest rates, data errors, and/or changes by agencies
     designed to lower the subsidy cost such as increasing fees,
     reducing the share of the loan receiving the government
     guarantee, and improving debt collection.  To determine the
     cause of specific subsidy rate differences would require
     examining the detailed assumptions used to estimate a program's
     cash flows over the full term of the credit.  Graphs of the most
     recently estimated subsidy rates for all cohorts in 9 of the 10
     programs in our sample are included in appendix II.  SBA's
     Disaster Loan Program is not included because the agency had not
     reestimated the program's subsidy costs at the time of our
     review. 

  -- Comparison of Estimated Subsidy Rates for Direct Loans to Those
     for Loan Guarantees--Loan type was not a predictor of whether
     subsidy rates increased or decreased from the budget execution
     rate to the first reestimate.\26 For example, of the three
     direct loan programs and five loan guarantee programs for which
     we had appropriate data, only one program--VA's Guaranty and
     Indemnity Fund guarantee program--did not have at least one
     cohort with an upward reestimate and at least one cohort with a
     downward reestimate. 


--------------------
\24 We examined only cohorts for which we had at least three
reestimates or budget execution estimates and two reestimates. 

\25 These estimates and reestimates were completed for the fiscal
year 1997 President's Budget. 

\26 This analysis is independent of whether the credit program is
treated as discretionary or mandatory in the budget. 


   AVAILABLE DATA NOT SUFFICIENT
   TO ASSESS WHETHER BUDGETARY
   TREATMENT AFFECTED INITIAL
   SUBSIDY ESTIMATES
------------------------------------------------------------ Letter :6

To obtain some insight on the potential effect of credit reform's
automatic appropriation (unconstrained by discretionary spending
limits) for reestimates, we compared the budget execution estimate to
the first reestimate\27 for mandatory programs and for discretionary
programs.  As explained previously, initial appropriations for
discretionary programs must compete with other programs for the
specified amount of funding available under the discretionary
spending limits set in law.  Mandatory credit programs are
automatically funded for whatever amount of credit is needed for a
given program design and set of program beneficiaries.  Both
discretionary and mandatory credit programs automatically receive
funding for the cost of reestimates without regard to Budget
Enforcement Act limits.  Thus, agencies with discretionary credit
programs could benefit from initially underestimating subsidy rates. 
If the pattern in the direction of reestimates for discretionary and
mandatory programs were the same, it would be an indication that this
provision of law was not affecting original estimates.  It may be
difficult to determine whether agencies intentionally underestimated
subsidy costs in initial estimates given data unreliability and the
number of other factors (such as changes in interest rates or other
economic conditions) that could affect subsidy estimates and
reestimates. 

We do know of one instance in which the issue was raised.  SBA, an
agency with discretionary credit programs, hired Price Waterhouse to
conduct a diagnostic review of SBA's existing internal controls. 
This September 1997 study said that "the credit subsidy process is
not viewed as a way of assessing the future risk and costs of the
program for management purposes.  Rather, the rate calculation is
perceived [by SBA] to be a tool for gaming the congressional
appropriations process." In commenting on a draft of this report, SBA
officials disagreed with this conclusion.  SBA officials stated that
the Price Waterhouse report was incorrect and, due to its special
nature, it was not corrected.  In support of their position, SBA
officials cited the quality of their data and staff and SBA's
commitment to have accurate and credible subsidy rates.  However, we
found, as discussed earlier, an error in SBA's subsidy estimation
methodology and that their component data were not always correct. 
Also, when we discussed SBA's concerns with a Price Waterhouse staff
member, he stated that its report was accurate for the period it
covered, spring 1997.  He described its report as based on interviews
and a review of documentation. 

The available data we were able to obtain were not sufficient to
assess whether a credit program's budgetary treatment affected its
initial subsidy estimates.  We found somewhat similar patterns when
we compared discretionary and mandatory programs.  We found that the
estimated subsidy rates for 8 of the 15 discretionary cohorts
increased in the first reestimate following the initial
appropriation, while first reestimates for 7 of the 12 mandatory
cohorts decreased. 

This result is not conclusive.  No real conclusions can be drawn from
this observation about whether some discretionary programs may have
sought to benefit from initially underestimating subsidy costs.  Any
firm conclusion about the reasons for reestimates would require
better data and more in-depth study.  Other factors, such as changes
in the economy--including interest rates--or more historical data,
may have contributed to these reestimates.  Further, as audits have
demonstrated, much of the data are not reliable.  Also, sensitivity
analyses and other sources showing the key variables that affect
subsidy rates were not consistently available. 


--------------------
\27 The first reestimates of subsidy rates are important because they
are the first opportunity for agencies to revise subsidy estimates
used in budget execution and receive permanent, indefinite budget
authority for any increase.  HUD's Mutual Mortgage Insurance (MMI)
program has an option to use a different budgetary treatment
specified in its authorizing legislation.  As a policy decision with
OMB, MMI uses liquidating account reserves as envisioned in its
authorizing legislation, not the permanent, indefinite budget
authority available under credit reform.  See our earlier report
(GAO/AIMD-94-58, September 26, 1994) for a more detailed discussion
of this issue. 


   LACK OF RELIABLE COMPONENT DATA
   HAMPERS ABILITY TO DETERMINE
   THE CAUSES OF CHANGES IN
   SUBSIDY ESTIMATES
------------------------------------------------------------ Letter :7

Data on the four components of subsidy expense--interest costs, net
defaults, fees and other collections, and other subsidy costs--could
be used to examine the causes for changes in subsidy rate estimates. 
Ideally, these data, calculated by the OMB subsidy model as part of
the subsidy estimation process, would provide a ready basis to
analyze such changes and thus identify possible policy responses. 
However, these component data were frequently missing or inaccurate,
and thus we were unable to use them for identifying causes of changes
in estimates. 

SFFAS No.  2 provides general definitions of these components and
requires agencies to disclose them in financial statements.  The
Federal Accounting Standards Advisory Board commented on the
importance of such data stating that "the cost component information
would be valuable for making credit policy decisions, monitoring
portfolio quality, and improving credit performance.  Information on
interest subsidies and fees would help in making decisions on setting
interest rates and fee levels.  Information on default costs would
help in evaluating credit performance." It also could be useful as a
performance measure to comply with the Government Performance and
Results Act.  With better data, decisionmakers could compare these
components across programs and agencies to see the effect of
programmatic differences. 

The potential usefulness of the component data was recently
demonstrated by our analysis of SBA component data.  Although SBA's
subsidy component data were flawed, they still provided a quick
indication to GAO that there was an error in the fiscal year 1997
subsidy estimates for the section 7(a) General Business Loan
Program.\28 Correcting this error enabled SBA to guarantee
approximately $2.5 billion more in section 7(a) small business
loans.\29 OMB and SBA officials acknowledged that better oversight
and improved internal controls at both OMB and SBA are needed to
prevent similar errors in the future. 

A model developed by OMB staff to calculate credit subsidies
aggregates detailed data on defaults, recoveries, prepayments, and
other cash flows to calculate the components of subsidy expense on a
present-value basis.  In describing the OMB credit subsidy model, OMB
guidance to agencies says "use of a common subsidy model ensures
comparability and uniformity among all Federal credit program subsidy
estimates."\30 However, VA, USDA, SBA, and HUD did not distribute
their cash flows consistently among the components.  OMB Circulars
A-11 and A-34 did not provide definitions of the subsidy components. 
The user's guide for OMB's subsidy model did not provide sufficiently
clear definitions of the components to ensure that the components
could be calculated accurately.  Agency staff said they did not have
a clear understanding of the definitions and thus were unsure about
where the OMB model allocated detailed data for each of their
program's cash flows in calculating the subsidy expense components. 
Agency staff were unclear about what the data produced by the model
represented.  In an earlier report reviewing the credit subsidy
model,\31 we recommended that OMB revise the model to specifically
identify which data were used by the model in the subsidy
calculations. 

Further, OMB did not require agencies to use the subsidy model at a
specific point in their estimation process as they are developing
subsidy estimates for their budget submissions.  Education officials
noted that, with the participation and approval of OMB, Education
uses the model as an interim step in subsidy estimation, not at the
end as do other agencies.  Education uses a methodology that makes
minor adjustments to subsidy rates produced by the model.  As a
result, Education staff did not provide component data for its
reestimated subsidy rates. 

Clear definitions of each subsidy component and specific OMB
instructions and assistance to agencies in using its subsidy model
could provide better assurance of accurate and comparable component
data.  Since our inquiries, staff at VA, USDA's Rural Housing Service
(RHS), and HUD worked with OMB staff to clarify requirements and
address problems with component data. 

Further, staff from five of eight programs whose component data we
were able to examine acknowledged that the data were questionable. 
We found that component data were inaccurate because they were not
consistent with program characteristics.  RHS staff said that, before
fiscal year 1996, the component data were incorrect because they
adjusted underlying data in using early versions of the OMB subsidy
model that did not accommodate their program characteristics.  As
discussed earlier, VA's direct loan program showed the net default
component as negative in each of the 7 years that data were
available.  This erroneously indicates that higher defaults would
reduce the subsidy rate.  Component data provided by the agencies are
shown in the tables in appendix III. 

This potentially useful information was not understood by agencies,
often was unavailable, sometimes was not accurate, and thus was not
used to inform program management or budget decision-making.  In a
letter responding to a recent GAO report on OMB's subsidy model, OMB
acknowledged that it would be useful for the revised subsidy model to
have a facility to display, at the option of the user, the
calculation of subsidy percentages and components. 


--------------------
\28 Credit Subsidy Estimates for the Sections 7(a) and 504 Business
Loan Programs (GAO/T-RCED-97-197, July 15, 1997). 

\29 Had the actual subsidy rate been known in advance, the Congress
could have chosen whether to provide the same amount of budget
authority for a higher loan amount or to hold the loan amount
constant and provide less budget authority. 

\30 OMB Circular A-34, section 12.7 (December 1995). 

\31 Credit Reform:  Review of OMB's Credit Subsidy Model
(GAO/AIMD-97-145, August 29, 1997). 


   IMPROVEMENTS UNDERWAY
------------------------------------------------------------ Letter :8

There have been a number of recent efforts to clarify and simplify
implementation of the Credit Reform Act.  The Balanced Budget Act of
1997 included some changes to the Credit Reform Act and OMB has
changed its guidance as well.  OMB is piloting a new reestimation
methodology--called the "balances approach"--at HUD that it states
will simplify the process.  However, this new approach does not
calculate the components of subsidy expense (interest, net defaults,
fees and other collections, and other subsidy costs) over the entire
term of the loans as does the current reestimate methodology.  The
data from which to calculate components would remain available with
the balances approach, but what is lost is having the component data
calculated as a part of the reestimate process.  Further, if there is
no requirement to report or review the data in that way, agencies
would have less incentive than now to make the calculations and use
the data.  OMB also is formulating a new approach to discounting cash
flows that it states will improve accuracy without adding difficulty
for agency staff. 

The Credit Task Force of the Accounting and Auditing Policy
Committee,\32 which includes OMB, Treasury, credit agency
participants, and GAO, has been studying credit reform implementation
in preparation for the first audit of the fiscal year 1997
governmentwide consolidated financial statements.  The group has
proposed guidance for agencies on methods and documentation for
estimating subsidy rates and creating a store of historical data.\33
It also developed draft guidance\34 on preparing and auditing subsidy
estimates that will be useful for budget, accounting, and auditing
staff.  The Credit Reform Committee of the Chief Financial Officers'
Council also has devised ways agencies can simplify implementation of
credit reform. 

Other resources are available to help agencies enhance their
capacities to make subsidy estimates.  For example, OMB has provided
short-term technical assistance with estimation and modeling to VA,
SBA, and HUD.  As those efforts continue, staff in agencies who
report that they lack adequate resources for research or systems
development could adapt strategies or data system formats that have
been used successfully in other agencies.  However, management-level
commitment at all of the credit agencies and OMB is critical to
continuing these efforts and to ensuring that the implementation of
credit reform is an agency priority. 


--------------------
\32 This task force formerly was known as the Subgroup on Credit
Reform of the Government-wide Audited Financial Statements Task
Force. 

\33 Government-wide Audited Financial Statements Task Force, Subgroup
on Credit Reform, Model Credit Program Methods and Documentation for
Estimating Subsidy Rates and the Model Information Store (96-CR-7,
May 1, 1996). 

\34 Government-wide Audited Financial Statements Task Force, Subgroup
on Credit Reform, Preparing and Auditing Direct Loan and Loan
Guarantee Subsidies under the Federal Credit Reform Act (96-CR-14,
October 16, 1997). 


   CONCLUSIONS
------------------------------------------------------------ Letter :9

Greater sustained commitment by management at OMB and the credit
agencies is needed to produce the useful data needed to fully
implement credit reform.  Effective implementation requires timely,
readily available, accurate estimates that are comparable among
credit programs.  Although there are indications that some agencies
have taken this seriously, problems with the availability and
reliability of subsidy estimates continue to permeate all agencies'
efforts at implementation.  While agencies are working to improve
their subsidy estimation processes, agency staff continue to report
that credit reform implementation often is not a priority of top
management.  This is indicated both by the failure to ask questions
about estimates in program and budget reviews and by the reported
lack of sufficient computer systems to support the subsidy estimation
process.  Greater commitment by OMB and the credit agencies is needed
to address pervasive problems with the availability and reliability
of subsidy estimate data and documentation.  Since agencies are most
responsive to issues in which there is demonstrated interest,
continued oversight would increase the likelihood that credit reform
would be implemented as intended.  Better and more reliable data are
needed to facilitate this oversight. 

The availability of automatic funding for reestimates of subsidy
costs creates an incentive for agencies with discretionary programs
to initially underestimate subsidy costs.  Whether or not agencies
are responding to this incentive is unclear.  Because the data
generally are not reliable and because other factors, such as
economic fluctuations (including changes in the interest rate) could
have caused changes in reestimates, more in-depth study and better
data would be needed to draw a firm conclusion. 

Accurate, consistent data on subsidy expense components could be used
effectively by program managers and executive and congressional
decisionmakers as originally intended--to monitor program
implementation, consider program changes, and compare direct loan and
loan guarantee programs designed for the same purpose.  Currently,
data are inaccurate or missing and agency staff said they do not
understand the data.  Therefore, component data have not been
available to inform decision-making. 

While no single agency yet is successful in all aspects of credit
reform implementation, some progress is being made at each of the
agencies we studied.  Over time, the scrutiny of financial statement
audits will continue to bring greater discipline to the estimation
process and greater accuracy to the reported subsidy costs. 


   RECOMMENDATIONS
----------------------------------------------------------- Letter :10

We recommend that the Secretaries of Agriculture, Education, Housing
and Urban Development, and Veterans Affairs, and the Administrator of
Small Business improve 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. 

So that agency staff can aggregate data from their cash flows into
the OMB subsidy model accurately and consistently, we recommend that
the OMB Director ensure that OMB staff (1) provide detailed guidance
and definitions of the four subsidy components (interest, net
defaults, fees and other collections, and other subsidy costs) and
(2) revise the OMB subsidy model to provide agencies with the formula
for calculating each component.  We also recommend that the OMB
Director ensure, to the extent possible, that agencies prepare
accurate subsidy estimates, use consistent definitions of subsidy
components, and have appropriate documentation. 

Finally, we recommend that the OMB Director work toward identifying
ways OMB can further assist agencies to more rapidly and accurately
implement credit reform.  These might include providing additional
direct assistance to the agencies, developing prototypes for
estimating methodologies, and prompting interagency forums for the
exchange of information on problems and best practice solutions by
working level staffs. 


   AGENCY COMMENTS AND OUR
   EVALUATION
----------------------------------------------------------- Letter :11

OMB and each of the five credit agencies we examined commented on a
range of implementation problems and progress.  OMB officials
commented that the report's focus on subsidy estimation is valuable. 
However, OMB officials stated that our analytical methodology was
questionable because the report did not distinguish between the
effect of interest rates on initial subsidy rates and the effect of
default and other technical factors.  Our methodology was designed to
isolate the effects of interest rates by using component data and
budget execution rates.  It is true that interest rates would change
each year even if the default and other technical assumptions remain
constant.  Unfortunately, we could not isolate these effects because
agencies frequently did not provide these data or provided inaccurate
data.  We used budget execution rates as the starting point for our
analyses.  This reduced the effect of interest rate changes in the
months between budget request and budget execution. 

OMB also stated that several of the report's general conclusions
about subsidy estimates were not supported by the evidence.  We
disagree.  First, OMB officials noted that unless an adjustment is
made for the effect of different discount rates, it is impossible to
draw valid conclusions about the accuracy of subsidy rates by
observing that they have fluctuated over time.  We did not draw
conclusions about the accuracy of subsidy rates.  However, we did
observe that the rates fluctuated over time and that some fluctuation
would be expected.  We said in the report that the reliability of
credit data is questionable for a number of reasons.  Our
characterization of the data was based primarily on results of the
audits of the fiscal year 1996 financial statements as well as some
discrepancies we identified between rates provided to us by agencies
and those reported in the President's Budget.  Second, OMB officials
said that we could not draw conclusions about a program from the size
or direction of subsidy reestimates unless the effect of interest
rate reestimates is removed.  Our report did not draw conclusions
about programs from the size or direction of subsidy reestimates. 
Third, OMB officials were concerned that our discussion of the timing
of reestimates could lead to the incorrect conclusion that the budget
formulation process for subsidy rates for new loans is not being
informed by the experience on existing loan cohorts.  We disagree. 
Such a conclusion regarding the effect of not performing reestimates
is, in fact, correct.  Given that we found three of the five agencies
received waivers of the reestimate requirement, it would appear that
their budget formulation is not being informed by the most recent
experience on existing loan cohorts.  (See appendix IV.)

Agriculture officials stated that the report's comments and
suggestions will improve budget formulation and accounting for
programs under credit reform.  They further stated that FSA is
working to address the concerns noted in the report.  (See appendix
V.)

Education officials strongly agreed with the report's emphasis on the
importance of the Federal Credit Reform Act of 1990.  Their comments
note that, over the past 5 years, Education has steadily increased
staff, contractor, and system resources dedicated to developing
accurate and timely credit estimates.  Education officials also
raised a question about their perception that our report implied that
"significant shifts in subsidy reestimates over time are necessarily
a bad thing." Our report did not imply this.  Rather, we said that
over time, some fluctuation in subsidy rates would be expected, some
estimates varied widely, and these sharp variations raise questions
about the causes for these changes and the reliability of the
underlying data.  Education officials also provided some
clarification of their use of the OMB subsidy model.  (See appendix
VI.)

HUD officials acknowledged that the agency has experienced some
reporting and estimation problems and stated that the agency has made
significant progress since the enactment of the Federal Credit Reform
Act of 1990.  They noted that the draft report documents the
difficulty that virtually every agency is experiencing in dealing
with credit reform requirements.  HUD officials stated that the
problem with cost estimates has less to do with the level of effort
devoted to data collection and the timeliness of reestimation than it
has to do with inherent limitations of the net present value
technique in cost estimation.  We agree that credit reform has been a
challenge to agencies.  However, cash basis reporting for credit
programs, while easier to accomplish, does not reflect their costs. 
When sufficient attention is devoted to net present value estimates
of costs as required under the Federal Credit Reform Act, these
subsidy costs will be a much better basis for budgeting.  (See
appendix VII.)

SBA officials' expressed concern that we quoted a Price Waterhouse
report prepared at SBA's request.  SBA officials stated that the
Price Waterhouse report was incorrect and, due to its special nature,
was not corrected.  In support of their position, SBA officials cited
the quality of their data and staff and SBA's commitment to have
accurate and credible subsidy rates.  However, as discussed earlier,
we found an error in SBA's subsidy estimation methodology and that
their component data were not always correct.  Also, when we
discussed SBA's concerns with a Price Waterhouse staff member, he
stated that its report is accurate for the period it covered, spring
1997, and was based on interviews and a review of documentation. 

SBA officials said their practices represent the leading edge of
compliance with the Federal Credit Reform Act and requested that we
acknowledge this if we agreed.  Our report recognized that some
progress in implementation is being made at each of the agencies we
studied, and we specifically acknowledged SBA's commitment and
efforts to improve.  (See appendix VIII.)

VA officials agreed with the report's recommendations and provided
their insights regarding the complexities of credit estimation and
its evolution over the years at VA.  They also provided some
technical comments, which we have addressed as appropriate.  (See
appendix IX.)


--------------------------------------------------------- Letter :11.1

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

This report was prepared under the direction of Christine Bonham,
Assistant Director, Budget Issues, who may be reached at (202)
512-9576.  Other major contributors to this report are listed in
appendix X.  Please contact me at (202) 512-9142 if you or your staff
have any questions concerning the report. 

Sincerely yours,

Susan J.  Irving
Associate Director, Budget Issues


BACKGROUND:  CREDIT REFORM
=========================================================== Appendix I

The federal government uses direct loans and loan guarantees as tools
to achieve numerous program objectives, such as assistance to
housing, agriculture, education, small businesses, and foreign
governments.  At the end of fiscal year 1996, the face value of the
government's direct loans and loan guarantees totaled a reported $973
billion, of which $167 billion was in direct loans and $806 billion
was in loan guarantees. 

After over 20 years of discussion about the shortcomings of using
cash budgeting for credit programs and activities, the Federal Credit
Reform Act of 1990 was enacted on November 5, 1990, as Title 13B of
the Omnibus Budget Reconciliation Act of 1990, Public Law 101-508. 
The Federal Credit Reform Act changed the budget treatment of credit
programs so that their costs can be compared more accurately with
each other and with the costs of other federal spending.  It also was
intended to ensure that the full cost of a credit program over its
entire life would be reflected in the budget when the loans were made
so that the executive branch and the Congress might consider that
cost when making budget decisions. 

In addition, it was recognized that credit programs had different
economic effects than most budget outlays, such as purchases of goods
and services, income transfers, and grants.  In the case of direct
loans, for example, the fact that the loan recipient was obligated to
repay the government over time meant that the economic effect of a
direct loan disbursement could be much less than a noncredit budget
transaction of the same dollar amount.  The change in economic
behavior resulting from loan guarantees occurred when the loan was
made, not when the government's cost was included in the federal
budget.  Thus, for both direct loans and loan guarantees the budget
did not reflect the change in economic behavior. 

CREDIT REFORM WAS DESIGNED TO
REMOVE DIFFICULTIES CAUSED BY CASH
TREATMENT

Before credit reform, it was difficult to make appropriate cost
comparisons between direct loan and loan guarantee programs and
between credit and noncredit programs.  Credit reform requirements
were formulated to address the factors that caused this problem. 

Two key principles of credit reform are (1) the definition of cost in
terms of the present value of cash flows over the life of a credit
instrument and (2) the inclusion in the budget of the costs of credit
programs in the year in which the budget authority is enacted and the
direct or guaranteed loans first may be disbursed. 

CREDIT REFORM WAS DESIGNED TO
ALLOW APPROPRIATE COST COMPARISONS

Before credit reform, credit programs--like other programs--were
reported in the budget on a cash basis.  This cash basis distorted
costs and, thus, the comparison of credit program costs with other
programs intended to achieve similar purposes, such as grants.  It
also created a bias in favor of loan guarantees over direct loans
regardless of the actual cost to the government. 

Loan guarantees appeared to be free in the short run while direct
loans initially appeared to be as expensive as grants because the
budget did not recognize that at least some of the guaranteed loans
would default and that some of the direct loans would be repaid. 

For direct loans, the budget for most discretionary accounts used
revolving funds, which showed budget authority and outlays in the
amount that loan disbursements, in the current year, exceeded
repayments received from all past loans in that budget year.  This
cash approach overstated direct loan costs in the initial years of a
program when loan disbursements were likely to be greater than
repayments.  Conversely, this treatment understated costs in later
years when loan repayments were more likely to be much larger
relative to disbursements.  In contrast, for loan guarantees, the
budget did not record any outlays when the guarantees were made
(except the negative outlay resulting from any origination fees),
even though the program was likely to entail future losses.  Budget
authority and outlays were recorded only when defaults occurred. 

Credit reform changed this treatment for direct loans and loan
guarantees made on or after October 1, 1991.  It required that budget
authority to cover the cost to the government of new loans and loan
guarantees (or modifications to existing credit instruments) be
provided before the loans, guarantees, or modifications are made. 
Credit reform requirements specified a net cost approach using
estimates for future loan repayments and defaults as elements of the
cost to be recorded in the budget.  This puts direct loans and loan
guarantees on an equal footing; it permits the costs of credit
programs to be compared with each other and with the costs of
non-credit programs when making budget decisions. 

CREDIT REFORM IDENTIFIES THE
GOVERNMENT'S COST OF CREDIT
ACTIVITIES

Credit reform requirements separate the government's cost of
extending or guaranteeing credit, called the subsidy cost, from
administrative and unsubsidized program costs.  Administrative
expenses receive separate appropriations.  They are treated on a cash
basis and reported separately in the budget.  The unsubsidized
portion of a direct loan is that which is expected to be recovered
from the borrower. 

The Federal Credit Reform Act defines the subsidy cost of direct
loans as the present value of disbursements--over the loan's life--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).  In
making these calculations, agencies must include the cost to the
federal government of borrowing the funds.  The act defines the
subsidy cost of loan guarantees as 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). 

Agencies prepare these cost estimates on a net present value basis as
a part of their budget request.  For the budget years we reviewed,
agencies then recalculated the subsidy rate when they extended credit
by updating the approved rate for any changes in interest rates and
legislation.  They make direct loans and loan guarantee commitments
as possible under this appropriation.  Later, after the end of the
fiscal year, agencies reestimate subsidy costs based on actual
experience and expected economic changes. 

CREDIT PROGRAMS NOW USE THREE
BUDGETARY ACCOUNTS

The Federal Credit Reform Act set up a special budget accounting
system to record the budget information necessary to implement credit
reform.  It provides for three types of accounts--program, financing,
and liquidating--to handle credit transactions. 

Credit obligations and commitments made on or after October 1,
1991--the effective date of credit reform--use only the program and
financing accounts.  The program account receives separate
appropriations for administrative and subsidy costs of a credit
activity and is included in budget totals.  When a direct or
guaranteed loan is disbursed, the program account pays the associated
subsidy cost for that loan to the financing account.  The financing
account, which is nonbudgetary,\1 is used to record the cash flow
associated with direct loans or loan guarantees over their lives.  It
finances loan disbursements and the payments for loan guarantee
defaults with (1) the subsidy cost payment from the program account,
(2) borrowing from the Treasury, and (3) collections received by the
government.  Figure I.1 diagrams this cash flow. 

   Figure I.1:  Credit Reform Cash
   Flow Simplified

   (See figure in printed
   edition.)


If subsidy cost calculations are accurate, the financing account will
break even over time as it uses its collections to repay its Treasury
borrowing. 

Direct loans and loan guarantees made before October 1, 1991, are
reported on a cash basis in the liquidating account.  This account
continues the cash budgetary treatment used before credit reform.  It
has permanent indefinite budget authority\2 to cover any losses. 
Excess balances are transferred periodically--at least annually--to
the Treasury. 

In addition to the three accounts specified in the Federal Credit
Reform Act, OMB has directed that discretionary credit programs or
activities with negative subsidies must have special fund receipt
accounts.  These accounts hold receipts generated when the program or
activity shows a profit or when a downward reestimate of subsidy
costs indicates that the financing account balance is too high.  OMB
guidance provides that discretionary programs cannot use these
receipts unless they are appropriated, while mandatory programs may
use the receipts without appropriation action. 

OMB AND TREASURY PROVIDE
IMPLEMENTATION GUIDANCE

OMB and the Department of the Treasury provide guidance on
implementing credit reform.  OMB's written guidance is contained
primarily in OMB Circulars A-11, A-34, and A-129.\3 OMB also has
issued memorandums to provide additional implementation guidance
addressing specific situations.  Treasury's guidance is provided in
materials such as Basic Transactions Relating to Guaranteed Loans and
Subsidies, which contains a number of illustrative cases developed by
its Financial Management Service and distributed to agencies as
examples of how to account for credit reform transactions. 
Accounting guidance, consistent with the intent of the Federal Credit
Reform Act, is found in Accounting for Direct Loans and Loan
Guarantees, Statement of Recommended Accounting Standards, Number 2. 
This guidance was developed by the Federal Accounting Standards
Advisory Board (FASAB) and approved in July 1993, by OMB, the
Department of the Treasury, and GAO. 

IMPLEMENTATION GUIDANCE HAS
CHANGED

Fiscal year 1998 is the seventh year that credit programs have been
required to comply with credit reform.  Agencies that operate credit
programs and those that provide implementation guidance--OMB and
Treasury--have had to address a variety of situations for which the
Federal Credit Reform Act does not provide explicit direction.  OMB
and Treasury have refined their guidance to agencies based on greater
experience with the processes and data requirements for implementing
credit reform and on more information on agencies' limitations and
abilities. 

The Balanced Budget Act of 1997 amended the Federal Credit Reform Act
to clarify and simplify the requirements for subsidy cost estimation. 
Among other changes, the Federal Credit Reform Act was amended to
require agencies to make loans and guarantees using the technical
assumptions such as default, recoveries, and fees included in the
President's Budget for the year in which funds are obligated.  As a
result, the dollar amount of loans approved by the Congress will not
be increased or decreased by subsequent changes to technical
assumptions.  The act also was amended to require agencies to return
to the Treasury any excess funds in accounts for pre-1992 credit, and
to change budgetary treatment of credit from the Federal Financing
Bank. 

At the same time, OMB has drafted revised guidance to credit
agencies.  The changes include eliminating the requirement to
reestimate annually the change in subsidy cost due to changes in
interest rates as disbursements are made for a cohort's loans and
loan guarantees.  Instead, agencies are required to do only one
interest rate reestimate when the cohort is 90 percent disbursed. 
Interest rate reestimates before a cohort is fully disbursed are of
questionable validity since the discount rate will continue to
change.  The reestimates thereby cause large swings in subsidy
estimates with no value added to management decision-making or the
reliability of budgetary or financial reporting.  OMB also has
developed an alternative, simplified method for agencies to calculate
reestimates--called the "balances approach." This new approach, which
is being tested at HUD, looks forward, projecting and discounting
remaining cash flows from a cohort and comparing them to the current
balance owed to Treasury.  The method used to date looks both
backward and forward, requiring agencies to revise estimates of all
cash flows for a cohort--those that already have occurred and those
in the future.  However, this new approach does not calculate the
components of subsidy expense (interest, net defaults, fees and other
collections, and other subsidy costs) for the entire term of the
loans as does the current methodology.  The data from which to
calculate components would remain available with the balances
approach, but what is lost is having the component data calculated as
a part of the reestimate process.  Further, if there is no
requirement to report or review the data in that way, agencies would
have less incentive than now to make the calculations and use the
data. 

Several interagency groups also reviewing agencies' implementation
and the current requirements of credit reform have made
recommendations that have been adopted or endorsed by OMB and
Treasury.  The Credit Reform Committee of the Chief Financial
Officers Council has recommended certain actions that would simplify
budget execution and accounting.  The Credit Task Force of the
Accounting and Auditing Policy Committee (formerly the Credit
Subgroup of the Government-wide Audited Financial Statements Task
Force) has issued three papers.  The first outlines an ideal model
for estimating and documenting subsidy rates.\4 The paper recognizes
that credit agencies are many years away from being able to implement
such a method, but discusses reasonable methods for subsidy rate
estimation and discusses the types of actual loan data that might be
maintained to support agency subsidy estimates.  The second paper
provides draft guidance for agencies' budget and accounting staff and
auditors for preparing and auditing direct loan and loan guarantee
subsidy estimates.\5 The third paper outlines recommended changes in
the accounting standards for direct loans and loan guarantees and
interpretations of the standards on the display of the components of
subsidy expense.\6


--------------------
\1 Nonbudgetary accounts may appear in the budget document for
information purposes, but are not included in the budget totals for
budget authority or budget outlays.  They do not belong in the budget
because they show only how something is financed, and do not
represent the use of resources. 

\2 Permanent budgetary authority is available as a result of
permanent legislation and does not require annual appropriation. 
Indefinite budget authority is budget authority of an unspecified
amount of money. 

\3 OMB Circular A-11, Preparation and Submission of Budget Estimates;
OMB Circular A-34, Instructions on Budget Execution; and OMB Circular
A-129, Managing Federal Credit Programs. 

\4 Model Credit Program Methods and Documentation for Estimating
Subsidy Rates and the Model Information Store, Government-wide
Audited Financial Statements Task Force, Subgroup on Credit Reform
(96-CR-7, May 1, 1996). 

\5 Preparing and Auditing Direct Loan and Loan Guarantee Subsidies
Under the Federal Credit Reform Act, Government-wide Audited
Financial Statements Task Force, Subgroup on Credit Reform (96-CR-14,
October 16, 1997).  In October 1997, this paper was submitted to the
Accounting and Auditing Policy Committee for official release as a
Technical Interpretation. 

\6 Disclosure Requirements for Credit Subsidy Expense Issue Paper,
Government-wide Audited Financial Statements Task Force, Credit
Subgroup.  In September 1997, this paper was submitted to the
Accounting and Auditing Policy Committee for review and
recommendation to FASAB. 


PROGRAM DESCRIPTIONS AND GRAPHS OF
ESTIMATED SUBSIDY RATES OF
SELECTED PROGRAMS
========================================================== Appendix II

This appendix provides a brief description of the programs we
selected, indicates whether the program is discretionary or
mandatory, illustrates changes in the programs' estimated subsidy
rates from two perspectives, and includes a brief explanation of the
more significant changes in rates as provided by agencies. 

The first graph for a program shows the most recently estimated total
subsidy rates for each year's cohort of credit.  These figures
provide a snapshot of agencies' most recent estimates of the
government's subsidy expense for these credit programs.  Differences
in the subsidy rate estimates for the different cohorts of a program
may be due to improved estimates (perhaps from greater experience) or
to changes in program characteristics or economic conditions. 

The second graph for a program profiles the estimated subsidy rates
of a given cohort over time.  We graphed the fiscal year 1992 cohorts
because they were more likely to have the most extensive subsidy
reestimate data.  We graphed fiscal year 1994 data for Education's
Ford Direct Loan Program because the program began in that year and,
therefore, it is the first year for which data were available.  These
figures depict changes over time in the agencies' knowledge about the
cost of loans funded in a given fiscal year.  Estimated subsidy rates
for some programs have greater variability, as seen in the different
vertical scales of the graphs. 

We did not prepare either the total estimated subsidy rate or the
cohort profile graphs for SBA's Disaster Loan Program.  We did not
prepare the cohort profile for the USDA/Farm Service Agency (FSA)
Farm Operating Loan Program.  Neither of the agencies provided
sufficient and/or consistent data. 

The figures in this appendix are based on data as reported and
verified by agencies.  We did not independently verify the accuracy
of these data.  The data points used to create these figures are
shown in bold on the tables in appendix III. 

FARM SERVICE AGENCY, U.S. 
DEPARTMENT OF AGRICULTURE

Farm Operating, Direct Loans, Discretionary

Loans are made to family farmers who are unable to obtain credit from
private and cooperative sources for farm operating purposes such as
purchasing livestock, poultry, and farm ranch equipment; purchasing
feed, seed, or fertilizer; meeting other farm or ranch operating
expenses; and paying family living expenses.  The use of loan funds
is intended to help provide farmers with the opportunity to conduct
successful farm operations. 

Agency budget officials attributed the relatively large drop in
subsidy rates between the fiscal year 1997 budget execution estimate
and the fiscal year 1998 budget request to the spread between the
interest rates charged to borrowers and Treasury interest rates that
represent the agency's cost of capital.  The spread in interest rates
increased nearly 100 percent for fiscal year 1998.  Further, a
program change for fiscal year 1998 reduced write-offs without
acquired property by more than 50 percent.  (See figure II.1.)

Because of changes over time in the way the FSA's Farm Operating
estimated subsidy rate data were aggregated, we were not able to
graph a profile of an individual cohort over time. 


   Figure II.1:  USDA/FSA's Farm
   Operating Direct Loan Program's
   Total Subsidy Rates, by
   Cohort--as Estimated for the FY
   1998 President's Budget

   (See figure in printed
   edition.)


RURAL HOUSING SERVICE, U.S. 
DEPARTMENT OF AGRICULTURE

Single Family Housing, Direct Loans, Discretionary

Single family loans are made to very low- and low-income families who
are without adequate housing and cannot obtain credit from other
sources.  Funds may be used to build, purchase, repair, or refinance
homes in rural areas.  Borrowers are required to "graduate" from the
direct loan program when their incomes are sufficient to afford
credit from the private sector. 

For figure II.2, agency staff attributed the relatively large drop in
subsidy rates between "Reestimated FY 1995" and "FY 1996 Execution"
primarily to the decrease in Treasury discount interest rates and the
increase in borrower interest rates.  This increase in borrower
interest rates is due to a change in the Rural Housing Service's
(RHS) regulations in fiscal year 1996 which reduced the payment
assistance to borrowers. 

Figure II.3 is influenced by P.L.  102-142, 742, which required
execution rates to be at or below the rates published in the
President's fiscal year 1992 Budget.  The agency used the rate in the
President's Budget because rates based on actual data would have been
higher.  In effect, the estimated subsidy rate was limited by law. 
According to agency staff, this resulted in a large first reestimate
of fiscal year 1992 subsidy expense. 

   Figure II.2:  USDA/RHS' Section
   502 Single Family Housing
   Direct Loan Program's Total
   Subsidy Rates, by Cohort--as
   Estimated for the FY 1998
   President's Budget

   (See figure in printed
   edition.)

   Figure II.3:  USDA/RHS' Section
   502 Single Family Housing
   Direct Loan Program's Total
   Subsidy Rates--Profile of the
   FY 1992 Cohort

   (See figure in printed
   edition.)

FEDERAL FAMILY EDUCATION LOAN
PROGRAM, DEPARTMENT OF EDUCATION

Stafford, Guaranteed Loans, Mandatory

The Federal Family Education Loan Program (FFELP) is intended to
encourage private lending to vocational, undergraduate, and graduate
students enrolled at eligible postsecondary institutions to help pay
for educational expenses.  The loans are insured by a state or
private nonprofit guaranty agency and reinsured by the federal
government.  Generally, a borrower is not required to make any
payments on the principal while still in school. 

As shown in figure II.4, a relatively large decrease in the estimated
subsidy rate occurred between the rate used for fiscal year 1997
budget execution and the rate requested in the President's fiscal
year 1998 Budget.  According to Department of Education staff, this
decrease was due primarily to (1) decreases in interest rates from
when the reestimates were calculated to when the budget request was
formulated and (2) legislative proposals included in the budget
request. 

For figure II.5, Education staff attributed the relatively large
increase from the third reestimate to the fourth reestimate of the
fiscal year 1992 cohort's subsidy rate to a new reestimate
methodology that showed higher defaults.  The fifth reestimate used a
methodology that showed defaults comparable to those in the third
reestimate. 

   Figure II.4:  Education's
   FFELP/Stafford Guaranteed Loan
   Program's Total Subsidy Rates,
   by Cohort--as Estimated for the
   FY 1998 President's Budget

   (See figure in printed
   edition.)

   Figure II.5:  Education's
   FFELP/Stafford Guaranteed Loan
   Program's Total Subsidy
   Rates--Profile of the 1992
   Cohort

   (See figure in printed
   edition.)


FORD DIRECT LOAN PROGRAM,
DEPARTMENT OF EDUCATION

Stafford, Direct Loans, Mandatory

To help defray costs of education at a participating school, loans
are made directly from the federal government to vocational,
undergraduate, and graduate students.  Generally, a borrower is not
required to make any payments on the principal while still in school. 

For figure II.6, Department of Education staff explained that the
drop in estimated subsidy rates between the latest reestimate for
fiscal year 1996 and fiscal year 1997 budget execution was due
primarily to decreases in interest rates from when the reestimates
were calculated to when the budget request was formulated.  Because
of programmatic design, small changes in interest rates result in
relatively larger changes in subsidy rates for direct loans than for
loan guarantees. 

For figure II.7, the subsidy estimate increased in the second
reestimate and decreased in the third reestimate.  Education staff
said they used a different methodology for the second reestimate that
showed higher defaults.  They attributed the decrease between the
second and third reestimates of the fiscal year 1994 cohort to using
a methodology that showed defaults comparable to those in the first
reestimate. 

   Figure II.6:  Education's
   Ford/Stafford Direct Loan
   Program's Total Subsidy Rates,
   by Cohort--as Estimated for the
   FY 1998 President's Budget

   (See figure in printed
   edition.)

   Figure II.7:  Education's
   Ford/Stafford Direct Loan
   Program's Total Subsidy
   Rates--Profile of the FY 1994
   Cohort

   (See figure in printed
   edition.)


HOUSING PROGRAMS, DEPARTMENT OF
HOUSING AND URBAN DEVELOPMENT

Federal Housing Administration (FHA) Mutual Mortgage Insurance Fund,
Section 203(b), Guaranteed Loans, Discretionary

To help people become homeowners, HUD provides insurance to lenders
against losses on mortgage loans.  These mortgage loans may be used
to finance the purchase of one-to-four family housing that is
proposed, under construction, or existing, as well as to refinance
indebtedness on existing housing. 

As shown in figure II.8, there was a relatively large increase in the
reestimated subsidy rates (in this case, a smaller negative subsidy)
between the fiscal year 1994 and 1995 cohorts.  According to HUD
officials, this increase was driven primarily by an increase in total
claim rates--from 4.95 percent for the fiscal year 1994 cohort in the
1998 Budget to 8.01 percent for the fiscal year 1995 cohort in the
1998 Budget. 


   Figure II.8:  HUD's Mutual
   Mortgage Insurance Fund's
   (203(b)) Guaranteed Loan
   Program's Total Subsidy Rates,
   by Cohort--as Estimated for the
   FY 1998 Budget Request and
   Reestimated for the FY 1997
   President's Budget

   (See figure in printed
   edition.)

   Figure II.9:  HUD's Mutual
   Mortgage Insurance Fund's
   (203(b)) Guaranteed Loan
   Program's Total Subsidy
   Rates--Profile of the FY 1992
   Cohort

   (See figure in printed
   edition.)


HOUSING PROGRAMS, DEPARTMENT OF
HOUSING AND URBAN DEVELOPMENT

FHA General and Special Risk Insurance Fund, Multifamily Refinance
223(f), Guaranteed Loans, Discretionary

Lenders are insured against loss on the purchase or refinance of
existing multifamily housing projects.  Only rental housing projects
not requiring substantial rehabilitation are eligible. 

Figure II.10 shows a sharp decline in estimated subsidy rates for
this program between the budget execution rates in fiscal years 1996
and 1997.  According to HUD officials, this decline reflects the use
of updated assumptions in fiscal year 1997.  These new assumptions
incorporated an additional 5 years of performance data as well as the
initial experience of FHA's mortgage sales. 

As shown in figure II.11, there was a sharp drop between the second
and third reestimated subsidy rates of the fiscal year 1992 cohort. 
HUD officials attributed this downward shift to a lower estimate for
defaults and a higher estimate for fees. 


   Figure II.10:  HUD/FHA's
   General and Special Risk
   Insurance Fund, Multifamily
   Refinance (223(f)) Guaranteed
   Loan Program's Total Subsidy
   Rates, by Cohort--as Estimated
   for the FY 1998 Budget Request
   and Reestimated for the FY 1997
   President's Budget

   (See figure in printed
   edition.)

   Figure II.11:  HUD/FHA's
   General and Special Risk
   Insurance Fund, Multifamily
   Refinance (223(f)) Guaranteed
   Loan Program's Total Subsidy
   Rates--Profile of the FY 1992
   Cohort

   (See figure in printed
   edition.)


SMALL BUSINESS ADMINISTRATION

7(a) General Business, Guaranteed Loans, Discretionary

Lenders are guaranteed against loss from loans to small businesses
that are unable to obtain financing in the private credit market but
can demonstrate the ability to repay loans.  Guaranteed loans are
made available to low-income business owners or businesses located in
areas of high unemployment; nonprofit sheltered workshops and other
similar organizations that produce goods or services; and to small
businesses being established, acquired, or owned by handicapped
individuals. 

For figure II.12, Small Business Administration (SBA) officials
attributed the increase in reestimated subsidy rates between fiscal
years 1994 and 1995 primarily to an increase in the assumed purchase
rate from 16.85 percent to 17.25 percent.  The purchase rate is the
percent of remaining principal and interest on defaulted guaranteed
loans that SBA expects to pay in claims from lenders.  The following
year's decrease, they explained, resulted primarily from the
imposition of new and/or modified program fees. 

For figure II.13, SBA officials explained that the changes in subsidy
rates estimated for the fiscal year 1992 cohort were due in part to
changes in the discount rate and in part to differences between
anticipated and actual purchase activity, fee collections, and
recoveries. 

   Figure II.12:  SBA's 7(a)
   Program's Total Subsidy Rates,
   by Cohort--as Estimated for the
   FY 1998 President's Budget

   (See figure in printed
   edition.)

   Figure II.13:  SBA's 7(a)
   Program's Total Subsidy
   Rates--Profile of the FY 1992
   Cohort

   (See figure in printed
   edition.)


SMALL BUSINESS ADMINISTRATION

Disaster, Direct Loans, Discretionary

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.  The loans may be used to
repair and/or replace property to predisaster conditions. 

SBA did not provide sufficient data on this program to allow us to
graph either the total subsidy rate estimates for each cohort or a
profile of an individual cohort. 



VETERANS BENEFITS ADMINISTRATION,
DEPARTMENT OF VETERANS AFFAIRS

Guaranty and Indemnity Fund, Guaranteed Loans, Mandatory

These loans assist 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.  Lenders
are guaranteed partial repayment of loans made to these individuals. 

As shown in figure II.14, a relatively sharp decline in reestimated
subsidy rates occurred from the fiscal year 1993 cohort to the fiscal
year 1995 cohort.  Department of Veterans Affairs (VA) staff
attributed this decline primarily to changes in foreclosure rates,
discount rates, and funding fees, as well as the application of
actual cohort data. 

Figure II.15 shows a decline between the third and fourth reestimates
of the fiscal year 1992 cohort.  VA officials said this decline was a
result of increased inflows from recoveries and the effect of having
more actual cohort data. 


   Figure II.14:  VA/VBA's
   Guaranty and Indemnity Fund,
   Guaranteed Loan Financing
   Account's Total Subsidy Rates,
   by Cohort--as Estimated for the
   FY 1997 President's Budget

   (See figure in printed
   edition.)

   Figure II.15:  VA/VBA's
   Guaranty and Indemnity Fund,
   Guaranteed Loan Financing
   Account's Total Subsidy
   Rates--Profile of the FY 1992
   Cohort

   (See figure in printed
   edition.)

Note:  Fiscal year 1992 budget execution estimates were calculated
quarterly.  Therefore, the data provided were not comparable to the
other reported data and were not graphed. 


VETERANS BENEFITS ADMINISTRATION,
DEPARTMENT OF VETERANS AFFAIRS

Loan Guaranty Direct Loans, Mandatory

This program makes home loans on favorable terms to members of the
general public--both veterans and nonveterans--purchasing a VA-owned
property.  These properties include homes that VA has acquired as a
result of foreclosures on VA guaranteed loans. 

Figure II.16 shows that a dramatic drop in reestimated subsidy rates
occurred between the fiscal years 1994 and 1995 cohorts.  According
to VA staff, this decrease is associated primarily with the inclusion
of actual cohort data as well as a significant increase in the
estimated proceeds from loan sales. 


   Figure II.16:  VA/VBA's Loan
   Guaranty Direct Loan Financing
   Account's Total Subsidy Rates,
   by Cohort--as Estimated for the
   FY 1997 President's Budget

   (See figure in printed
   edition.)

   Figure II.17:  VA/VBA's Loan
   Guaranty Direct Loan Financing
   Account's Total Subsidy
   Rates--Profile of the FY 1992
   Cohort

   (See figure in printed
   edition.)

Note:  Fiscal year 1992 budget execution estimates were calculated
quarterly.  Therefore, the data provided were not comparable to the
other reported data and were not graphed. 


ESTIMATED SUBSIDY RATES OF
SELECTED PROGRAMS
========================================================= Appendix III

This appendix presents a summary of the completeness of estimated
subsidy rate data and supporting documentation provided by each
agency.  The estimated subsidy rates, as reported and confirmed by
each agency, are also shown for the 10 programs.  We did not examine
the quality of the data underlying the subsidy estimates or the
agencies' estimation process. 

We determined whether the budget request estimate, the budget
execution estimate, and all reestimates for each fiscal year were
supported by output from OMB's credit subsidy model and cash flow
spreadsheets.  Table 1 summarizes the completeness of these data. 
Although we did not examine the data quality, some quality issues
jumped out.  We found 11 instances where data confirmed by agencies
did not agree with those reported in the President's Budget or Credit
Supplement.  When questioned about this, agencies provided additional
data supporting the President's Budget in four instances.  For the
remaining seven cases, we show the estimated subsidy rates that were
supported in agency documentation. 

We also did not evaluate the timeliness or frequency of the
reestimates.  Currently, agencies do not necessarily reestimate prior
year cohorts on the same schedule.  According to the June 23, 1997,
version of OMB Circular A-11, section 33.5(s), "Reestimates must be
made at the beginning of each fiscal year, as long as any loans in
the cohort are outstanding, unless a different plan is approved by
OMB." In other words, the first reestimate of the fiscal year 1996
cohort generally should be included in the fiscal year 1998 budget
request.  OMB has permitted some agencies to vary from this schedule. 
USDA, for example, received a waiver from OMB allowing its
reestimates to be prepared in the middle of the fiscal year rather
than the beginning.  Therefore, updated cost information from USDA's
first reestimate of the fiscal year 1996 cohort would not be included
in the budget submission until the fiscal year 1999 budget request. 

Although we did not independently verify the data provided by
agencies, the numerous modified audit opinions these agencies have
received on their credit programs, the discrepancies we found between
the Budget and the data provided to us by the agencies, and other
work we have done\1 raise serious concerns about the quality of the
data.  For example, USDA's Office of the Inspector General (OIG)
rendered a qualified opinion on the fiscal year 1996 financial
statements of the rural development mission area because the OIG was
not able to obtain sufficient, competent evidence to support the
agency's credit program receivables and estimated losses on loan
guarantees and the related credit program subsidy and appropriated
capital used.  For fiscal year 1996, Education's OIG also was unable
to render an audit opinion on the department's credit activities due
to auditor concerns about the integrity of the data supporting
estimates of the Federal Family Education Loan Program.  HUD received
a qualified audit opinion from its OIG in fiscal year 1996 because
FHA's credit-related accounts were not converted to a present value
basis, as required by the Statement of Federal Financial Accounting
Standards, No.  2, Accounting for Direct Loans and Loan Guarantees. 
Consequently, the OIG was unable to audit the credit-related account
balances. 

Tables III.1 through III.10 present the subsidy rate estimates
reported and confirmed by each agency.  The columns represent each
budget year--1992 through 1998--since credit reform was enacted. 
Each table is horizontally divided into five sections.  The top
section of each table shows the total subsidy rate estimated and
reestimated for each cohort, as calculated at different points in
time.  The four bottom sections represent the components of subsidy
expense--interest, net defaults, fees and other collections, and
other subsidy costs.\2 These four components add to the total subsidy
rate. 

The following is an excerpt of the top section from table III.3,
presented to illustrate how subsidy rate data are arrayed.  Because
the bottom sections organizationally mirror the top section, only the
top section is shown in this example. 


                        (Total Subsidy Rate (in percent))

                                          Cohorts
            --------------------------------------------------------------------
Time of
estimate       FY 92     FY 93     FY 94     FY 95     FY 96     FY 97     FY 98
----------  --------  --------  --------  --------  --------  --------  --------
Request        22.60     17.60     14.24     15.31     22.84     19.77     15.09
Execution      22.60     17.60     15.59     15.31     22.84     22.75
Reestimate     16.99     17.55     17.07     23.09     19.55
 1
Reestimate     18.73     18.90     22.13     20.60
 2
Reestimate     17.78     22.74     20.00
 3
Reestimate     26.30     18.96
 4
Reestimate     17.74
 5
--------------------------------------------------------------------------------
The seven columns labeled "Cohorts" contain the available budget
request, budget execution, and reestimated subsidy rates for the
credit funded by appropriations in the indicated fiscal year.  The
left-hand column contains the type of estimate prepared and gives
some indication about when the estimate was calculated.  Thus, the
first row, labeled "Request," shows the subsidy rate used to prepare
the budget year request for that fiscal year.  In the sample table
above, the fiscal year 1992 budget request used an estimated subsidy
rate of 22.60 percent.  Reading down the fiscal year 1992 column, the
budget execution rate shows the subsidy rate estimated for the fiscal
year 1992 cohort after the beginning of fiscal year 1992--after the
appropriation was received but before loans were made.  Reestimate
data are shown in the order in which they were prepared. 

The shaded cells represent estimates that will be made in the future
when future budgets are prepared (e.g., the budget execution estimate
for fiscal year 1998 and the first reestimate of fiscal year 1997
should have been done for the fiscal year 1999 budget.) The numbers
from these tables that support the graphs in appendix II are shown in
bold. 

For cells that are neither shaded nor filled in with data, the
estimated subsidy rate was not provided.  The absence of the data
might reflect incomplete agency files or indicate that the agency did
not prepare estimates.  Notes to the tables provide additional
explanatory information in some instances when data were not
provided. 

OMB has periodically released updated versions of its credit subsidy
model.  Several of these versions aggregated the component data
differently.  For example, early versions of the model combined the
two components "Fees and Other Collections" and "Other Subsidy
Costs." We showed all comparable data provided by the agencies. 



                              Table III.1
                
                  USDA/FSA: Farm Operating Direct Loan
                         Program Subsidy Rates

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Total Subsidy Rate
----------------------------------------------------------------------
Request                   \a     \a     \a     \a  13.70  13.29   6.57
Execution                \ab    \ab    \ab     \a  12.98  12.59
Reestimate 1              \a     \a  14.05  13.07
Reestimate 2              \a  10.51  14.18
Reestimate 3            8.96  10.72
Reestimate 4           14.32
Reestimate 5
----------------------------------------------------------------------
\a USDA provided subsidy rate estimates that were divided into four
sub-risk categories.  Therefore, the data were not comparable to the
other reported data and were not included in this table. 

\b USDA provided quarterly subsidy rate estimates.  Therefore, the
data were not comparable to the other reported data and were not
included in this table. 

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Interest Component of Subsidy Expense
----------------------------------------------------------------------
Request                   \a     \a     \a     \a   1.14      -   0.16
                                                           0.73
Execution                \ab    \ab    \ab     \a   1.04      -
                                                           0.39
Reestimate 1              \a     \a   1.13   0.25
Reestimate 2              \a      -   0.89
                               2.39
Reestimate 3                      -
                               1.86
Reestimate 4            3.95
Reestimate 5

Net Defaults Component of Subsidy Expense
----------------------------------------------------------------------
Request                   \a     \a     \a     \a  12.59  13.85   6.23
Execution                \ab    \ab    \ab     \a  11.96  12.84
Reestimate 1              \a     \a  12.56  12.72
Reestimate 2              \a  13.08  12.66
Reestimate 3                  12.84
Reestimate 4           12.75
Reestimate 5

Fees Component of Subsidy Expense
----------------------------------------------------------------------
Request                   \a     \a     \a     \a      -      -      -
                                                   21.43  23.75  21.76
Execution                \ab    \ab    \ab     \a      -      -
                                                   20.37  21.83
Reestimate 1              \a     \a      -      -
                                     21.72  21.56
Reestimate 2              \a      -      -
                              26.55  18.40
Reestimate 3                      -
                              24.24
Reestimate 4               -
                        8.51
Reestimate 5

All Other Component of Subsidy Expense
----------------------------------------------------------------------
Request                   \a     \a     \a     \a  21.40  23.92  21.93
Execution                \ab    \ab    \ab     \a  20.35  21.96
Reestimate 1              \a     \a  22.08  21.67
Reestimate 2              \a  26.37  19.04
Reestimate 3                  23.98
Reestimate 4            6.12
Reestimate 5
----------------------------------------------------------------------


                              Table III.2
                
                  USDA/RHS: Section 502 Single Family
                  Housing Direct Loan Program Subsidy
                                 Rates

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Total Subsidy Rate
----------------------------------------------------------------------
Request                22.64  24.35  20.02   9.20  20.99   8.30  12.81
Execution              22.64     \a     \a  24.36  14.30  14.18
Reestimate 1           27.16  19.73  23.07  23.66
Reestimate 2           23.81  20.12  22.60
Reestimate 3           24.31  18.90
Reestimate 4           24.78
Reestimate 5
----------------------------------------------------------------------
\a USDA provided quarterly subsidy rate estimates.  Therefore, the
data were not comparable to the other reported data and were not
included in this table. 

\b Comparable data for "Fees" and "Other Subsidy Cost" components
were not provided. 

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Interest Component of Subsidy Expense
----------------------------------------------------------------------
Request                       28.67  23.08   5.23  27.29   9.63  15.69
Execution                        \a     \a  27.94  17.06  17.01
Reestimate 1           30.49  22.28  24.91  27.62
Reestimate 2           28.21  23.08  25.80
Reestimate 3           29.00  22.41
Reestimate 4           27.51
Reestimate 5

Net Defaults Component of Subsidy Expense
----------------------------------------------------------------------
Request                        1.68   1.48   5.20   0.26   1.59   1.12
Execution                        \a     \a   1.13   1.08   1.20
Reestimate 1            1.05   1.60   1.34   1.19
Reestimate 2            0.91   1.52   0.95
Reestimate 3            0.73   0.57
Reestimate 4            0.93
Reestimate 5

Fees Component of Subsidy Expense
----------------------------------------------------------------------
Request                          \b      -      -      -      -      -
                                     10.10  16.46   7.98   7.74   7.10
Execution                        \a     \a      -      -      -
                                            12.59  10.63   7.12
Reestimate 1             \ab      -      -      -
                              13.46  33.63   5.75
Reestimate 2               -      -      -
                        9.63  13.50  13.31
Reestimate 3               -      -
                        9.88  13.59
Reestimate 4               -
                       10.09
Reestimate 5

All Other Component of Subsidy Expense
----------------------------------------------------------------------
Request                          \b   5.56  15.22   1.41   4.83   3.09
Execution                        \a     \a   7.89   6.79   3.09
Reestimate 1              \b   9.32  30.46   0.59
Reestimate 2            4.33   9.01   9.15
Reestimate 3            4.45   9.51
Reestimate 4            6.43
Reestimate 5
----------------------------------------------------------------------


                              Table III.3
                
                  Education: FFELP/Stafford Guaranteed
                       Loan Program Subsidy Rates

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Total Subsidy Rate
----------------------------------------------------------------------
Request                22.60  17.60  14.24  15.31  22.84  19.77  15.09
Execution              22.60  17.60  15.59  15.31  22.84  22.75
Reestimate 1           16.99  17.55  17.07  23.09  19.55
Reestimate 2           18.73  18.90  22.13  20.60
Reestimate 3           17.78  22.74  20.00
Reestimate 4           26.30  18.96
Reestimate 5           17.74
----------------------------------------------------------------------
Note:  According to officials from Education, they do not use the OMB
credit subsidy model as the final step in the subsidy estimation
process, as is done in other agencies.  Education uses it at an
earlier point in the process--before adjustments are made for
external factors.  Therefore, output from this model did not reflect
data used to prepare the budget request and budget execution
estimated subsidy rates.  Reestimates, however, are not adjusted and
thus reflect the results from OMB's model. 

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Interest Component of Subsidy Expense
----------------------------------------------------------------------
Request                                     15.58  18.14  15.85  10.35
Execution                            11.31  15.58  18.14  15.85
Reestimate 1
Reestimate 2
Reestimate 3
Reestimate 4
Reestimate 5

Net Defaults Component of Subsidy Expense
----------------------------------------------------------------------
Request                                      6.76   8.47   7.85   6.81
Execution                             8.06   6.76   8.47   9.46
Reestimate 1
Reestimate 2
Reestimate 3
Reestimate 4
Reestimate 5

Fees Component of Subsidy Expense
----------------------------------------------------------------------
Request                                         -      -      -      -
                                             3.78   3.77   4.86   3.45
Execution                                -      -      -      -
                                      4.89   3.78   3.77   3.50
Reestimate 1
Reestimate 2
Reestimate 3
Reestimate 4
Reestimate 5

All Other Component of Subsidy Expense
----------------------------------------------------------------------
Request                                         -   0.00   0.93   1.39
                                             3.25
Execution                             1.11      -   0.00   0.94
                                             3.25
Reestimate 1
Reestimate 2
Reestimate 3
Reestimate 4
Reestimate 5
----------------------------------------------------------------------


                              Table III.4
                
                  Education: Ford/Stafford Direct Loan
                         Program Subsidy Rates

                             ((In percent))

                                               Cohorts
                                --------------------------------------
Time of Estimate                  FY94    FY95    FY96    FY97    FY98
------------------------------  ------  ------  ------  ------  ------
Total Subsidy Rate
----------------------------------------------------------------------
Request                          11.39   14.45   16.54   10.17   14.65
Execution                        12.42   14.45   16.54   10.38
Reestimate 1                     18.59   17.38   14.40
Reestimate 2                     19.47   15.47
Reestimate 3                     14.29
----------------------------------------------------------------------
Notes:  (1) The Ford Direct Loan Program was established in fiscal
year 1994.  (2) According to officials from Education, they do not
use the OMB credit subsidy model as the final step in the subsidy
estimation process, as is done in other agencies.  Education uses it
at an earlier point in the process--before adjustments are made for
external factors.  Therefore, output from this model did not reflect
data used to prepare the budget request and budget execution
estimated subsidy rates.  Reestimates, however, are not adjusted and
thus reflect the results from OMB's model. 

                             ((In percent))

                                               Cohorts
                                --------------------------------------
Time of Estimate                  FY94    FY95    FY96    FY97    FY98
------------------------------  ------  ------  ------  ------  ------
Interest Component of Subsidy Expense
----------------------------------------------------------------------
Request                                   9.68   12.64    3.16    8.20
Execution                         9.39    9.68   12.64    3.16
Reestimate 1
Reestimate 2
Reestimate 3

Net Defaults Component of Subsidy Expense
----------------------------------------------------------------------
Request                                   7.59    6.88    9.56    8.13
Execution                         8.76    7.59    6.88    9.56
Reestimate 1
Reestimate 2
Reestimate 3

Fees Component of Subsidy Expense
----------------------------------------------------------------------
Request                                  -3.53   -4.00   -3.95   -2.95
Execution                        -6.45   -3.53   -4.00   -3.95
Reestimate 1
Reestimate 2
Reestimate 3

All Other Component of Subsidy Expense
----------------------------------------------------------------------
Request                                   0.72    1.03    1.41    1.28
Execution                         0.72    0.72    1.03    1.62
Reestimate 1
Reestimate 2
Reestimate 3
----------------------------------------------------------------------


                              Table III.5
                
                  HUD: Mutual Mortgage Insurance Fund
                (Section 203(b)) Guaranteed Loan Program
                             Subsidy Rates

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Total Subsidy Rate
----------------------------------------------------------------------
Request                    -      -      -      -      -      -      -
                        2.60   2.70   2.79   2.78   2.77   2.33   2.62
Execution                  -      -      -      -      -      -
                        2.60   2.57   2.79   1.95   2.77   2.88
Reestimate 1               -      -      -      -
                        2.84   3.67   3.40   1.79
Reestimate 2               -      -      -
                        2.78   3.49   3.15
Reestimate 3               -
                        3.10
Reestimate 4
Reestimate 5
----------------------------------------------------------------------
\a Comparable data for "Fees" and "Other Subsidy Cost" components
were not provided. 

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Interest Component of Subsidy Expense
----------------------------------------------------------------------
Request                 0.00          0.00          0.00   0.00   0.00
Execution               0.00          0.00   0.00   0.00
Reestimate 1            0.00   0.00   0.00   0.00
Reestimate 2            0.00   0.00   0.00
Reestimate 3            0.00
Reestimate 4
Reestimate 5

Net Defaults Component of Subsidy Expense
----------------------------------------------------------------------
Request                 2.49          2.26          1.70   1.85   1.46
Execution               2.49          2.26   1.84   1.70
Reestimate 1            2.33   1.40   1.40   2.18
Reestimate 2            1.88   1.40   1.26
Reestimate 3            1.71
Reestimate 4
Reestimate 5

Fees Component of Subsidy Expense
----------------------------------------------------------------------
Request                   \a            \a             -      -      -
                                                    5.24   5.21   4.88
Execution                 \a            \a      -      -
                                             4.69   5.24
Reestimate 1              \a      -      -      -
                               5.92   5.42   5.10
Reestimate 2               -      -      -
                        5.72   5.72   5.14
Reestimate 3               -
                        5.92
Reestimate 4
Reestimate 5

All Other Component of Subsidy Expense
----------------------------------------------------------------------
Request                   \a            \a          0.78   1.03   0.81
Execution                 \a            \a   0.90   0.78
Reestimate 1              \a   0.84   0.62   1.13
Reestimate 2            1.06   0.82   0.72
Reestimate 3            1.11
Reestimate 4
Reestimate 5
----------------------------------------------------------------------


                              Table III.6
                
                   HUD/FHA: General and Special Risk
                 Insurance Fund, Multifamily Refinance
                223(f), Guaranteed Loan Program Subsidy
                                 Rates

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Total Subsidy Rate
----------------------------------------------------------------------
Request                 1.51   2.98   3.14   3.20   2.30      -      -
                                                           0.79   0.75
Execution               1.51   2.98   3.34   3.40   2.51      -
                                                           0.79
Reestimate 1            1.91   3.04   1.49   1.67
Reestimate 2            1.91   0.71
Reestimate 3               -
                        0.21
Reestimate 4
Reestimate 5
----------------------------------------------------------------------
\a Comparable data for "Fees" and "Other Subsidy Cost" components
were not provided. 

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Interest Component of Subsidy Expense
----------------------------------------------------------------------
Request                        0.00   0.00   0.00   0.00   0.00   0.00
Execution                      0.00                        0.00
Reestimate 1                   0.00   0.00   0.00
Reestimate 2            0.00   0.00
Reestimate 3            0.00
Reestimate 4
Reestimate 5

Net Defaults Component of Subsidy Expense
----------------------------------------------------------------------
Request                        7.39   7.64   7.82   7.24   5.21   5.02
Execution                      7.39                        5.21
Reestimate 1                   7.55   6.85   7.27
Reestimate 2            6.78   5.96
Reestimate 3            5.52
Reestimate 4
Reestimate 5

Fees Component of Subsidy Expense
----------------------------------------------------------------------
Request                          \a     \a      -      -      -      -
                                             4.62   4.94   6.00   5.77
Execution                        \a                           -
                                                           6.00
Reestimate 1                      -      -      -
                               4.51   4.44   4.54
Reestimate 2               -      -
                        4.88   5.25
Reestimate 3               -
                        5.73
Reestimate 4
Reestimate 5

All Other Component of Subsidy Expense
----------------------------------------------------------------------
Request                          \a     \a   0.00   0.00   0.00   0.00
Execution                        \a                        0.00
Reestimate 1                   0.00      -      -
                                      0.92   1.06
Reestimate 2            0.00   0.00
Reestimate 3            0.00
Reestimate 4
Reestimate 5
----------------------------------------------------------------------


                              Table III.7
                
                   SBA: 7(a) Guaranteed Loan Program
                             Subsidy Rates

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Total Subsidy Rate
----------------------------------------------------------------------
Request                 0.60   1.35   2.35   2.73   2.01   2.68   1.80
Execution               4.85   5.21   2.15   2.74   1.06   2.54
Reestimate 1            4.94   4.97   2.91   2.97   2.40
Reestimate 2            3.83   3.67   2.70   3.61
Reestimate 3
Reestimate 4
Reestimate 5
----------------------------------------------------------------------
                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Interest Component of Subsidy Expense
----------------------------------------------------------------------
Request                                      0.00   0.00   0.00   0.00
Execution                             0.00   0.00   0.00   0.00
Reestimate 1            0.00   0.00   0.00   0.00   0.00
Reestimate 2            0.00   0.00   0.00   0.00
Reestimate 3
Reestimate 4
Reestimate 5

Net Defaults Component of Subsidy Expense
----------------------------------------------------------------------
Request                                      5.13   5.08   8.25   5.33
Execution                             4.88   5.07   5.51   8.05
Reestimate 1            6.79   6.95   4.82   5.51   5.88
Reestimate 2            5.73   5.68   4.68   5.61
Reestimate 3
Reestimate 4
Reestimate 5

Fees Component of Subsidy Expense
----------------------------------------------------------------------
Request                                         -      -      -      -
                                             2.52   3.22   5.57   3.53
Execution                                -      -      -      -
                                      2.87   2.47   4.57   5.50
Reestimate 1               -      -      -      -      -
                        1.85   1.97   2.18   2.81   3.48
Reestimate 2               -      -      -      -
                        1.85   2.01   1.98   2.00
Reestimate 3
Reestimate 4
Reestimate 5

All Other Component of Subsidy Expense
----------------------------------------------------------------------
Request                                      0.11   0.15   0.00   0.00
Execution                             0.13   0.15   0.12   0.00
Reestimate 1            0.00   0.00   0.27   0.27   0.00
Reestimate 2            0.00   0.00   0.00   0.00
Reestimate 3
Reestimate 4
Reestimate 5
----------------------------------------------------------------------


                              Table III.8
                
                   SBA: Disaster Direct Loan Program
                             Subsidy Rates

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Total Subsidy Rate
----------------------------------------------------------------------
Request                22.50   8.21  12.79  12.67   8.46   7.90  11.44
Execution              33.93  20.58  22.99  31.54  28.08  20.02
Reestimate 1
Reestimate 2
Reestimate 3
Reestimate 4
Reestimate 5
----------------------------------------------------------------------
\a Comparable data for "Fees" and "Other Subsidy Cost" components
were not provided. 

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Interest Component of Subsidy Expense
----------------------------------------------------------------------
Request                                                -      -   0.89
                                                    3.61   1.57
Execution                             9.70  19.61  15.46  11.91
Reestimate 1
Reestimate 2
Reestimate 3
Reestimate 4
Reestimate 5

Net Defaults Component of Subsidy Expense
----------------------------------------------------------------------
Request                                            11.00   9.46  11.15
Execution              19.00  11.93  12.31  10.85  11.58   8.10
Reestimate 1
Reestimate 2
Reestimate 3
Reestimate 4
Reestimate 5

Fees Component of Subsidy Expense
----------------------------------------------------------------------
Request                                                -   0.00   0.00
                                                    1.46
Execution                 \a     \a      -   0.00      -   0.00
                                      1.74          1.58
Reestimate 1
Reestimate 2
Reestimate 3
Reestimate 4
Reestimate 5

All Other Component of Subsidy Expense
----------------------------------------------------------------------
Request                                             2.54   0.00      -
                                                                  0.59
Execution                 \a     \a   2.72   1.08   2.62   0.00
Reestimate 1
Reestimate 2
Reestimate 3
Reestimate 4
Reestimate 5
----------------------------------------------------------------------


                              Table III.9
                
                  VA/VBA: Guaranty and Indemnity Fund,
                   Guaranteed Loan Financing Account
                             Subsidy Rates

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Total Subsidy Rate
----------------------------------------------------------------------
Request                 1.15   1.82   1.36   1.18   1.56   1.47  0.49\
                                                                     a
Execution                 \b     \b   1.36   1.18   1.56  0.49\
                                                             ac
Reestimate 1            2.26   1.11   0.98   0.69  0.42\
                                                       a
Reestimate 2            2.30   1.44   1.16  0.78\
                                                a
Reestimate 3            2.35   1.77  0.98\
                                         a
Reestimate 4            1.01  1.14\
                                  a
Reestimate 5           1.15\
                           a
----------------------------------------------------------------------
\a VA has consolidated its loan guarantees into one account and its
direct loans into another (Public Law 105-65).  The fiscal year 1998
Budget was prepared on this basis.  Thus, data for the fiscal year
1998 request, the fiscal year 1997 budget execution estimate, and
associated reestimates of prior years included in this table are not
comparable with other data shown. 

\b Subsidy rate estimates were calculated quarterly; therefore, the
data provided were not comparable to the other reported data and were
not included. 

\c VA consolidated its programs midway through fiscal year 1997.  The
rate of 1.47 percent was effective from October 1996 through March
1997.  From April through September 1997, the consolidated fiscal
year 1998 rate of 0.49 percent was used. 

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Interest Component o
----------------------------------------------------------------------
Request                               0.00   0.00   0.00   0.00  0.00\
                                                                     a
Execution                             0.00   0.00   0.00  0.00\
                                                             ac
Reestimate 1                          0.00   0.00  0.00\
                                                       a
Reestimate 2                   0.00   0.00  0.00\
                                                a
Reestimate 3            0.00   0.00  0.00\
                                         a
Reestimate 4            0.00  0.00\
                                  a
Reestimate 5           0.00\
                           a

Net Defaults Compone
----------------------------------------------------------------------
Request                               4.38   3.31   3.86   3.95  2.60\
                                                                     a
Execution                             4.38   3.31   3.86  2.60\
                                                             ac
Reestimate 1                          2.69   2.99  2.45\
                                                       a
Reestimate 2                   3.06   2.77  2.64\
                                                a
Reestimate 3            4.33   3.04  2.41\
                                         a
Reestimate 4            2.66  2.03\
                                  a
Reestimate 5           2.54\
                           a

Fees Component of Su
----------------------------------------------------------------------
Request                                  -      -      -      -      -
                                      2.11   1.64   1.87   1.95  1.91\
                                                                     a
Execution                                -      -      -      -
                                      2.11   1.64   1.87  1.91\
                                                             ac
Reestimate 1                             -      -      -
                                      1.45   1.95  1.84\
                                                       a
Reestimate 2                      -      -      -
                               1.19   1.38  1.78\
                                                a
Reestimate 3               -      -      -
                        1.09   0.87  1.40\
                                         a
Reestimate 4               -      -
                        1.08  0.87\
                                  a
Reestimate 5               -
                       1.08\
                           a

All Other Component
----------------------------------------------------------------------
Request                                  -      -      -      -      -
                                      0.92   0.50   0.43   0.52  0.20\
                                                                     a
Execution                                -      -      -      -
                                      0.92   0.50   0.43  0.20\
                                                             ac
Reestimate 1                             -      -      -
                                      0.27   0.35  0.19\
                                                       a
Reestimate 2                      -      -      -
                               0.43   0.24  0.08\
                                                a
Reestimate 3               -      -      -
                        0.90   0.39  0.03\
                                         a
Reestimate 4               -      -
                        0.57  0.03\
                                  a
Reestimate 5               -
                       0.32\
                           a
----------------------------------------------------------------------


                              Table III.10
                
                   VA/VBA: Loan Guaranty Direct Loan
                    Financing Account Subsidy Rates

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Total Subsidy Rate
----------------------------------------------------------------------
Request                11.28   8.35   2.34   2.34   3.11   1.56  2.36\
                                                                     a
Execution                 \b     \b   2.34   2.34   3.11  2.61\
                                                             ac
Reestimate 1            3.90   1.09   3.17   1.18  1.76\
                                                       a
Reestimate 2            3.90   1.27   4.80  1.38\
                                                a
Reestimate 3            4.11   3.52  2.05\
                                         a
Reestimate 4            4.38  0.41\
                                  a
Reestimate 5           0.21\
                           a
----------------------------------------------------------------------
\a VA has consolidated its loan guarantees into one account and its
direct loans into another (Public Law 105-65).  The fiscal year 1998
Budget was prepared on this basis.  Thus, data for the fiscal year
1998 request, fiscal year 1997 budget execution estimate, and
associated reestimates of prior years included in this table are not
comparable with other data shown. 

\b Subsidy rate estimates were calculated quarterly; therefore, the
data provided were not comparable to the other reported data and were
not included. 

\c VA consolidated its programs midway through fiscal year 1997.  The
rate of 1.56 percent was effective from October 1996 through March
1997.  From April through September 1997, the consolidated fiscal
year 1998 rate of 2.61 percent was used. 

                             ((In percent))

                                           Cohorts
                       -----------------------------------------------
Time of Estimate        FY92   FY93   FY94   FY95   FY96   FY97   FY98
---------------------  -----  -----  -----  -----  -----  -----  -----
Interest Component of Subsidy Expense
----------------------------------------------------------------------
Request                                  -      -      -      -      -
                                     17.90   9.77  14.76  18.21  17.32
                                                                    \a
Execution                                -      -      -
                                     17.90   9.77  14.76
Reestimate 1                             -      -      -
                                     12.39  11.35  11.11
                                                      \a
Reestimate 2                      -      -      -
                               9.33  12.39  11.35
                                               \a
Reestimate 3               -      -      -
                       13.20   9.33  12.39
                                        \a
Reestimate 4               -      -
                       11.04  9.34\
                                  a
Reestimate 5               -
                       10.99
                          \a

Net Defaults Component of Subsidy Expense
----------------------------------------------------------------------
Request                                  -      -      -      -      -
                                     71.22  71.77  66.50  73.08  69.54
                                                                    \a
Execution                                -      -      -
                                     71.22  71.77  66.50
Reestimate 1                             -      -      -
                                     69.41  71.82  74.86
                                                      \a
Reestimate 2                      -      -      -
                              77.86  67.81  86.92
                                               \a
Reestimate 3               -      -      -
                       77.51  75.55  83.47
                                        \a
Reestimate 4               -      -
                       77.33  90.97
                                 \a
Reestimate 5               -
                       81.19
                          \a

Fees Component of Subsidy Expense
----------------------------------------------------------------------
Request                                  -      -      -      -      -
                                      3.87   3.83   3.68   3.92  3.80\
                                                                     a
Execution                                -      -      -
                                      3.87   3.83   3.68
Reestimate 1                             -      -      -
                                      3.79   3.80  0.88\
                                                       a
Reestimate 2                      -      -      -
                               4.06   3.76  0.90\
                                                a
Reestimate 3               -      -      -
                        1.92   4.01  0.88\
                                         a
Reestimate 4               -      -
                        1.92  1.05\
                                  a
Reestimate 5               -
                       1.11\
                           a

All Other Component of Subsidy Expense
----------------------------------------------------------------------
Request                              95.33  87.72  88.05  96.77  93.02
                                                                    \a
Execution                            95.33  87.72  88.05
Reestimate 1                         88.76  88.16  88.61
                                                      \a
Reestimate 2                  92.53  88.76  100.5
                                              5\a
Reestimate 3           96.74  92.41  98.78
                                        \a
Reestimate 4           94.66  101.7
                                7\a
Reestimate 5           93.50
                          \a
----------------------------------------------------------------------


(See figure in printed edition.)Appendix IV

--------------------
\1 For example, see HUD Management:  FHA's Multifamily Loan Loss
Reserves and Default Prevention Efforts (GAO/RCED/AIMD-95-100, June
5, 1995); Small Business Administration:  Credit Subsidy Estimates
for the Sections 7(a) and 504 Business Loan Programs
(GAO/T-RCED-97-197, July 16, 1997); and Homeownership: 
Appropriations Made to Finance VA's Housing Program May Be
Overestimated (GAO/RCED-93-173, September 8, 1993). 

\2 The fourth component category, other subsidy costs, primarily
reflects estimated prepayments. 


COMMENTS FROM THE OFFICE OF
MANAGEMENT AND BUDGET
========================================================= Appendix III

See comment 1. 



(See figure in printed edition.)

See comment 2. 

See comment 3. 

See comment 4. 

See comment 5. 

See comment 6. 

Now on p.  17. 



(See figure in printed edition.)

See comment 7. 

See comment 8. 

See comment 9. 



(See figure in printed edition.)

See comment 10. 


The following are GAO's comments on the Office of Management and
Budget's March 11, 1998, letter. 

GAO COMMENTS

1.  See "Agency Comments and Our Evaluation" section of the report. 

2.  We did not draw conclusions about the accuracy of subsidy rates
from our observation that subsidy rates have fluctuated over time. 
Our report stated that the reliability of credit data was
questionable for a number of reasons.  Our characterization of the
data was based primarily on results of the audits of the fiscal year
1996 financial statements as well as some discrepancies we identified
between rates provided to us by agencies and those reported in the
President's Budget. 

3.  The report does not propose a causal relationship between
reestimates and subsidy rate fluctuations.  Further, changes in
assumptions cause the subsidy rate for a cohort of loans to fluctuate
from year to year, not reestimates. 

4.  We did not assess whether any assumptions used in subsidy rate
estimation were accurate.  Further, we did not draw conclusions about
a program from the size or direction of subsidy reestimates.  We
acknowledged that the fluctuations in subsidy rates could be due to
credit extended at different interest rates than anticipated or to
unanticipated changes in the economy, as well as other factors.  We
clarified the report language to state that interest rate changes are
an example of such unanticipated economic changes. 

5.  Report text revised to clarify that the patterns are similar. 

6.  We compared the subsidy rate estimated for budget execution
against the first subsidy reestimate for the following reasons. 
First, we used budget execution (rather than budget request) because
it eliminated the effect of interest rate changes in the months
between when the budget request rate was formulated and the time the
government is obligated or committed for the loans.  Second, credit
reform guidance requires agencies to have appropriations of budget
authority to cover the full estimated net present value cost of
outstanding credit.  Third, we used the first reestimate because if
an agency sought to benefit from initially underestimating subsidy
costs and there was oversight by OMB to ensure that agencies have
sufficient appropriations of budget authority, the agency would have
to obtain budget authority to cover the shortfall as soon as
possible--that is, in the first reestimate. 

7.  Given that three of the five agencies received waivers of the
reestimate requirement, it would appear that their budget formulation
is not being informed by the most recent experience on existing loan
cohorts.  The Credit Task Force of the Accounting and Auditing Policy
Committee, which includes OMB and GAO representation, proposed that
agencies that have OMB approval be allowed to use a combination of
actual and projected data as the basis for reestimates.  By allowing
agencies to begin the reestimation process earlier, this approach
should reduce the need for waivers--permitting agencies to use more
recent actual data to inform budget formulation and to include them
in the President's Budget and financial statement audits. 

8.  We agree that this would be useful information.  However, we
could not obtain sufficient component data from agencies to evaluate
whether subsidy estimation has improved over time, as indicated by
smaller annual changes in reestimates due to technical factors.  Our
report does include a section discussing a number of recent efforts
to clarify and simplify implementation of the Credit Reform Act.  In
other ongoing work, we are evaluating data reliability, barriers to
credit reform implementation, and agency plans to overcome those
barriers.  We will also report on notable best practices in credit
agencies as appropriate. 

9.  We do not dispute that OMB staff devote a substantial amount of
time to credit issues.  However, our conclusion that greater
sustained commitment is needed reflects our concerns about the
availability and reliability of credit data despite the fact that
agencies now have prepared eight budgets since credit reform became
effective for fiscal year 1992.  Further, component data were not
used to inform program management or budget decision-making. 

10.  OMB's attachment containing specific comments has not been
included, but our report has been modified as appropriate to reflect
the comments contained in the attachment. 




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



(See figure in printed edition.)



(See figure in printed edition.)

See comment 1. 

Now on p.  46. 


The following are GAO's comments on the Department of Agriculture's
March 11, 1998, letter. 

GAO COMMENTS

1.  Report text was revised to reflect the agency's comment. 




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

See comment 1. 



(See figure in printed edition.)

See comment 2. 

See comment 3. 


The following are GAO's comments on the Department of Education's
March 5, 1998, letter. 

GAO COMMENTS

1.  See "Agency Comments and Our Evaluation" section of the report. 

2.  The report text was clarified to reflect Education's comment. 

3.  Education provided all total subsidy rate estimates we requested
but did not provide subsidy components for each of these rates. 
Moreover, although Education explained why the requested
documentation--OMB model output--would not be helpful, it did not
provide alternative supporting documentation. 




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

See comment 1. 



(See figure in printed edition.)

See comment 2. 

See comment 3. 



(See figure in printed edition.)

Now on p.  4. 

See comment 4. 

See comment 5. 



(See figure in printed edition.)

See comment 3. 

See comment 6. 


The following are GAO's comments on the Department of Housing and
Urban Development's March 4, 1998, letter. 

GAO COMMENTS

1.  An explanation of agencies' reported need for waivers has been
added to the report. 

2.  The report was clarified to specify that agencies can improve
their ability to forecast defaults, recoveries, prepayments, and fee
revenue through better modeling and more and better historical data. 

3.  Our report acknowledges the implementation difficulties involved
with credit reform.  However, once agencies establish a systematic
approach to subsidy estimation based on auditable assumptions,
present-value-based budgeting for credit will provide significantly
better information than the former cash-based system. 

4.  Our statement that causes for changes in estimates cannot readily
be identified was based on the fact that much of the component data
that would point to causes for such changes was missing, inaccurate,
or inconsistent with other data reported by agencies. 

5.  We clarified the report to indicate that the patterns were
similar and that the data were inconclusive.  As the report states,
any firm conclusion about the reasons for changes in reestimates
would require better data and more in-depth study. 

6.  HUD's attachment containing technical comments/corrections has
not been included, but our report has been modified as appropriate to
reflect the comments contained in the attachment. 




(See figure in printed edition.)Appendix VIII
COMMENTS FROM THE SMALL BUSINESS
ADMINISTRATION
========================================================= Appendix III

See comment 1. 

Now on pp.  19-20. 

See comment 2. 

See comment 3. 

See comment 4. 



(See figure in printed edition.)

See comment 5. 

See comment 6. 

See comment 7. 

See comment 8. 

Now on p.  2. 

See comment 9. 

Now on p.  3. 

See comment 10. 

See comment 11. 

Now on p.  6. 

See comment 12. 

Now on p.  8. 

See comment 13. 

See comment 14. 

Now on p.  9. 



(See figure in printed edition.)

See comment 15. 

See comment 16. 

Now on p.  10. 

See comment 17. 

Now on p.  10. 


The following are GAO's comments on the Small Business
Administration's March 5, 1998, letter. 

GAO COMMENTS

1.  See "Agency Comments and Our Evaluation" section of the report. 
The report text also was clarified to reflect SBA's position and our
evaluation. 

2.  See "Agency Comments and Our Evaluation" section of the report. 

3.  Data specific to individual agencies are shown in tables
contained in appendix III.  In other ongoing work, we are evaluating
data reliability, barriers to credit reform implementation, and
agency plans to overcome those barriers.  We also will report on
notable best practices in credit agencies as appropriate. 

4.  Throughout our report, we cite changes in economic conditions
and/or interest rates as the first item in a list of possible causes
of changes in subsidy rates, thus emphasizing their importance. 

5.  See "Agency Comments and Our Evaluation" section of the report. 

6.  See comment 3.  The results of the February 1998 meetings will be
considered in a future report on this work. 

7.  While a comparison of modeling methodologies used by the five
credit agencies would be interesting, it is outside the scope of this
report. 

8.  A review of the adequacy of accounting standards and budget
guidance was outside the scope of this report.  However, in other
ongoing work, we are evaluating barriers to credit reform
implementation and agency plans to overcome those barriers.  The
adequacy of accounting standards and budget guidance, if identified
as barriers, would be considered in that work.  Further, as we noted
in our report, changes to accounting standards and budget guidance
for credit programs are being considered. 

9.  A brief discussion of these changes is contained in appendix I. 

10.  Recognition that SBA reestimated the Disaster Loan Program for
the fiscal year 1999 budget was added to the "Results in Brief"
section. 

11.  The report text was revised as suggested. 

12.  Appendix III includes tables by agency and program that
illustrate where component data were unavailable. 

13.  Our report does address oversight by OMB.  The Subgroup on
Credit Reform of the Government-wide Audited Financial Statements
Task Force (now the Credit Task Force of the Accounting and Auditing
Policy Committee) has been working on data collection, analysis, and
documentation issues for several years and has proposed guidance. 

14.  Recent financial statement audit reports cited in our report
text have raised questions about data reliability.  Our report also
describes discrepancies between data in the President's Budget and
data provided and confirmed by agencies. 

15.  We agree that poor document retention is one cause of
differences between agency-confirmed rates and those reported in the
President's Budget. 

16.  The data provided by SBA does not permit us to prepare graphs
similar to other programs.  We graphed estimated subsidy rates from
two perspectives.  The first graph (for example, see figure II.12)
showed the most recently estimated total subsidy rate for each year's
cohort of credit--the rates for all of the outstanding cohorts were
calculated and based on historical and economic data updated as of
the same point in time.  Using the budget execution rates as
suggested by SBA would not be a similar approach because the rates
were calculated at different points in time.  The second graph (for
example, see figure II.13) showed the estimated and reestimated
subsidy rates for a given cohort over time beginning with budget
execution.  It would not be possible to do this for the Disaster Loan
Program since SBA provided only the initial rate graphed in this
analysis. 

17.  Specific support exists on pp.  10 through 16. 




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



(See figure in printed edition.)



(See figure in printed edition.)

Now on p.  13. 

See comment 1. 



(See figure in printed edition.)



(See figure in printed edition.)

See comment 2. 

Now on p.  7. 

See comment 2. 

Now on p.  16. 

See comment 2. 

Now on p.  25. 

See comment 3. 

Now on p.  61. 

See comment 2. 

Now on pp.  84 and 86. 


The following are GAO's comments on the Department of Veterans
Affairs' March 11, 1998, letter. 

GAO COMMENTS

1.  We revised table 1 to show that VA had provided all requested
rates and documentation for fiscal years 1992 and 1993.  We made the
change for fiscal year 1992 because VA stated that the cash flows
showing only 5 of the 15 years of the credit maturity represented all
of the output created by the OMB/VA model for that year.  For fiscal
year 1993, VA provided only one of the four quarterly execution cash
flows requested.  However, we discussed VA's comments with VA staff
who told us that they did not revise the cash flows for each of the
four quarters of fiscal year 1993. 

2.  Report text was revised. 

3.  Account title was revised. 


MAJOR CONTRIBUTORS TO THIS REPORT
=========================================================== Appendix X

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

Christine Bonham, Assistant Director
Carolyn Litsinger, Senior Evaluator-in-Charge
Carol Henn, Senior Evaluator

OFFICE OF THE GENERAL COUNSEL,
WASHINGTON, D.C. 

Charles Roney, Assistant General Counsel
Edda Emmanuelli-Perez, Attorney Advisor

RELATED GAO PRODUCTS

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

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). 

Debt Collection:  Improved Reporting Needed on Billions of Dollars in
Delinquent Debt and Agency Collection Performance (GAO/AIMD-97-48,
June 2, 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). 

*** End of document. ***