oversight

Small Business Administration: A Review of SBA's Estimate of Impact of Legislative Proposals for the 7(a) Loan Guarantee Program

Published by the Government Accountability Office on 1999-12-15.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

             aG-A                           O
        Accountability * Integrity * Reliahbility
United States General Accounting Office                                                             Resources, Community, and
Washington, DC 20548                                                                            Economic Development Division


          B-284140

          December 15, 1999

          The Honorable Donald A. Manzullo
          Chairman, Subcommittee on Tax,
           Finance, and Exports
          Committee on Small Business
          House of Representatives

          Subject: Small Business Administration: A Review of SBA's Estimate of Impact of
                   Legislative Proposals for the 7(a) Loan Guarantee Program

          Dear Mr. Chairman:

         On July 29, 1999, the House Small Business Committee reported out H.R. 2615, which
         included provisions to modify the Small Business Administration's (SBA) 7(a) loan program.'
         These provisions include (1) increasing the maximum portion of a loan that SBA may
         guarantee from $750,000 to $1 million, (2) adding incentives to encourage lenders to make
         small loans of $150,000 or less,2 (3) placing a cap of $2 million on the total size of a loan that
         SBA may guarantee, and (4) imposing a penalty for the early prepayment of some loans. 3
         SBA prepared two estimates of the impact that these legislative proposals would have on the
         average size and number of loans that it guarantees. First, in a June 29, 1999, response to
         you, SBA estimated the impact of an increase in the maximum loan guarantee on the
         availability of small loans. SBA concluded that increasing the guarantee would result in its
         guaranteeing fewer loans. Second, SBA estimated the collective impact-on the average size
         of loans and on the volume of loans guaranteed-of increasing the maximum guarantee and
         including the small loan incentives and the $2 million cap. According to SBA officials, the
         second response was provided to your staff on July 29, 1999, and to the Committee staff on
         an earlier date. SBA concluded that implementing these proposals together would not affect
         the average size and number of loans that it guarantees. According to SBA, both estimates
         were prepared within a short time of being requested. SBA officials also stated that they
         intended these estimates to represent their best judgment within the time available.

         SBA worked with the Congressional Budget Office (CBO) to estimate the proposals' impact
         on the credit subsidy rate. The credit subsidy rate is the net present value of the federal costs



             The House passed H.R. 2615 on Aug. 2, 1999.

         2
          The three incentives are (1) increasing the loan guarantee percentage from 75 percent to 80 percent for loans of more than
         $100,000 but less than or equal to $150,000; (2) reducing the guarantee fee from 3 percent to 2 percent for loans of more than
         $100,000 but less than, or equal to, $150,000; and (3) allowing lenders to retain one-quarter of the 2-percent guarantee fee on
         loans of $150,000 or less.
         3
          The penalty applies to loans with terms of 15 years or more that are prepaid within the first 3 years.




                                                                    GAO/RCED-00-49R Changes to SBA's 7(a) Loan Program
 B-284140

expected per dollar loaned through the program. CBO and SBA determined that the
proposals would not result in a significant change in the credit subsidy rate.

You asked us to describe how SBA prepared these estimates of the legislation's impact on the
average size of loans and the number of loans SBA makes and, if appropriate, recommend
ways that SBA could have improved its estimates. You further asked us to describe how SBA
analyzed the impact of the legislative proposals on the credit subsidy rate for the 7(a) loan
program and, if appropriate, recommend ways to improve this analysis.

In summary, we found:

* SBA's estimates of the legislative proposals' impact on the average size of loans and the
    volume of loans represent the professional judgment of SBA's loan program officials.
    SBA's analysis could have been enhanced had it considered, within the available time,
    factors such as (1) the relative size of the program change, as compared with prior
    program changes; (2) the upward trend in the average loan size; and (3) the economic
    context surrounding historical changes to the program.

*   With respect to its estimate of the impact of the legislative proposal on the credit subsidy
    rate, SBA used a cash-flow model that incorporated the legislative changes in H.R. 2615.
    SBA's conclusion that the legislative proposal would not significantly change the credit
    subsidy rate is based on assumptions regarding cash flows that appear optimistic:. That
    is, the prepayment assumptions were more optimistic than the cash flows used to prepare
    the fiscal year 1998 financial statements, which received an unqualified audit opinion.

SBA's Assessments of the Impact of Legislative Proposals on Average Loan Size and
Loan Volume Were Based on Professional Judgment

 For both estimates, the assessments of SBA officials were based on their professional
judgment and a consideration of historical program data. SBA officials did not document
their analysis and cannot explain how the various factors they considered were weighed in
their assessment. SBA officials believed that their approach to estimating loan volume
provided results similar to those that could have been obtained through more rigorous
methods. Furthermore, SBA officials noted that they did not have the econometric modeling
experience, the tools, or the time to prepare more rigorous estimates. Finally, SBA officials
stated that they intended these estimates to represent their best judgment only. Nonetheless,
we believe that within the time available, SBA could have considered other relevant issues.

In its June 29, 1999, estimate of the effect of increasing the loan guarantee amount from
$750,000 to $1 million, SBA relied on its experience with a $250,000 increase in the loan
guarantee amount that occurred in 1988. Specifically, in fiscal year 1988, the maximum
guarantee for 7(a) loans was increased from $500,000 to $750,000. From fiscal year :L988
through fiscal year 1990, the average size of a 7(a) loan grew from $213,300 to $247,c900-a 16-
percent increase. SBA concluded that a similar increase of $250,000 in the maximum
guarantee amount (in this case from $750,000 to $1 million) would also result in a 16-percent
increase in average loan size-from $230,000 to $267,000. According to SBA, such an increase
would result in the agency's funding 6,400 fewer loans (assuming total program funding of
$10.5 billion).




2                                          GAO/RCED-00-49R Changes to SBA's 7(a) Loan Program
B-284140

The credibility of the estimates could have been enhanced if other factors, such as the
following had been considered:

*    First, the $250,000 increase in the loan guarantee amount made in 1988 represented a 50-
     percent increase from the $500,000 base. The current legislative proposal represents a 33-
     percent increase in the loan guarantee amount ($250,000 from a $750,000 base). A 33-
     percent increase in the loan guarantee amount would likely lead to a smaller increase in
     the average loan size than the 16-percent increase that occurred when the maximum loan
     guarantee amount was increased by 50 percent.

*    Second, SBA officials did not incorporate the upward trend in the average loan size into
     their analysis. The average loan size increased by about 13 percent from 1986 through
     1988. Given this upward trend in average loan size, one would expect that the impact of
     the 1988 program change was less than the observed 16-percent increase in average loan
     size, although SBA officials attributed the entire 16-percent increase to the change.

*    Third, general economic conditions in 1999 differ from those in 1988, when the first
     program change occurred. For example, the prime interest rate 4 is currently 1.75
     percentage points lower today than it was during 1988. Such a difference in the state of
     the economy will influence the demand for commercial loans and thus possibly the
     impact of a program change.

SBA officials believe that these factors are appropriate to consider but disagree that
including these factors would have been any more appropriate than SBA's chosen approach.
We believe that including all appropriate factors would have helped SBA improve the
credibility of its estimates, even if the estimates themselves would have been the same.

According to SBA officials, on July 29, 1999, SBA provided your staff its second estimate of
the effect on loan size and volume of legislative changes. This estimate considered not only
increasing the loan guarantee limit from $750,000 to $1 million but also adding the small loan
incentives and implementing a $2 million cap. While SBA officials believe that raising the
guarantee amount to $1 million would increase the average size of loans, adding the small
loan incentives would both increase the demand for smaller loans and encourage lenders to
offer them. Thus, SBA officials concluded that the proposals would offset each other and
that the number and size of loans SBA guarantees would be unchanged.

Like the earlier estimate, this estimate was based on professional judgment, and SBA did not
document its analysis. SBA did not prepare a precise estimate of the impact the incentives
for small loans might have on loan size and volume. To determine the impact of the three
incentives for small loans, SBA considered other agency programs that had undergone
similar changes. SBA's loan volume data showed that three of its loan programs-the Export
Working Capital program, the SBAExpress program, and the LowDoc program-had an
in-crease in loan volume when changes were made in the programs. 6 However, SBA could not
document how it came to this conclusion. In addition, SBA's experience with changes to the
LowDoc and SBAExpress programs may not be comparable to the impact that the three

'Interest rates for SBA loans are negotiated between the borrower and the lender but are subject to SBA madmums, which are
pegged to the prime rate. The prime rate is a key indicator of interest rates charged to commercial borrowers by banks.

'Under the Export Working Capital program, SBA guarantees up to 90 percent of a secured loan or $750,000, whichever is less.
The LowDoc and SBAExpress programs streamline the lending process. The LowDoc program uses a shortened, one-page
application, to which SBA responds within 36 hours. The SBAExpress program allows lenders to use their own application
forms for loans up to $150,000 with a 50-percent guarantee rate.




3                                                       GAO/RCED-00-49R Changes to SBA's 7(a) Loan Program
 B-284140

 currently proposed incentives could have on the size of average loans. That is, while both
 sets of changes encourage lenders to make small loans, they do so in different ways, and
 therefore the magnitude of the impact could be different.

SBA's Analysis of the Impact of Legislative Proposals on Credit Subsidy Rate
Appears Optimistic

According to a director in SBA's Office of the Chief Financial Officer, in order to estimate the
impact on the credit subsidy rate, SBA used a cash-flow model that incorporated the changes
in H.R. 2615. SBA modeled the proposed legislation, assuming there would be no change in
the volume and mix of loans. 6 According to SBA, any increase in the subsidy rate from the
small loan incentives is more than offset by the decrease in the subsidy rate resulting from
the provision of the prepayment penalty. Overall, SBA estimated that the credit subsidy rate
would decline from 1.16 to 1.12 percent.

SBA's estimates of the credit subsidy rate resulting from the proposed legislation may be
optimistic. While including the legislative provision for a penalty imposed on the early
prepayment of loans, SBA's model assumed an increase in the frequency with which
borrowers prepay their loans within the 3-year period in which the penalty applies. However,
according to CBO, one should expect the frequency of prepayments to decline when a
prepayment penalty is introduced. If the prepayment rate were reduced, the prepayment
penalties might not offset the increase in the subsidy rate from the small incentives, as SBA
estimated. However, CBO found that the overall impact of the legislative proposals on the
credit subsidy rate was negligible. We did not perform a detailed analysis of SBA's credit
subsidy model or determine its reasonableness.

Scope and Methodology

To gather information on SBA's analysis for determining the impact of the legislative
proposals under H.R. 2615, we met with officials from SBA's Office of Financial Assistance,
Office of Congressional and Legislative Affairs, and Office of the Chief Financial Officer. We
also reviewed program and budget information from SBA.

We conducted our work from October 1999 through November 1999 in accordance with
generally accepted government auditing standards.

Agency Comments

We provided a draft copy of this report to the SBA for its review and comment. We spoke
with the Associate Administrator for SBA's Financial Assistance office and other SBA
officials, who said that the draft of this report implied that their estimates were intended as a
scientific modeling methodology and that we did not recognize the time constraints under
which they prepared the estimates. Furthermore, they believe that their methodology was
sound and the consideration of additional conditions would not result in any significant
difference in the final estimates. In response, we revised the report to make clear that we do
not suggest that, within the time that SBA had to prepare the estimates, it should have used
sophisticated modeling technique, such as economic forecasting. In addition, we revised the
report to emphasize the amount of time SBA had to prepare the estimates, and we added
information to better describe how SBA intended to portray these estimates.


'According to an official of SBA's Office of the Chief Financial Officer, the size of loans has little influence on credit subsidy rate.


4                                                           GAO/RCED-00-49R Changes to SBA's 7(a) Loam Program
B-284140

With regard to the time SBA had to prepare these estimates, SBA officials said that the time
they had to prepare their estimates was imposed by the requesters. The first estimate was
provided to your staff within 3 days of being requested. The second estimate was provided to
Committee staff the same day it was requested by them and to your staff 6 days later,
according to SBA officials. However, SBA did not have a record of how much time it had to
prepare these estimates. Nonetheless, we believe that SBA, within the time it says it had,
could have improved the credibility of its estimates had it considered other relevant factors,
which could include the three that we discuss in the report. SBA officials said that they
would take our concerns into consideration when preparing estimates in the future. The SBA
officials also provided technical changes, which we have made, as appropriate.



We are sending copies of this report to interested congressional committees and the
Honorable Aida Alvarez, Administrator of SBA. We will also make copies available upon
request.

Please call me at (202) 512-7631 if you or your staff have any questions. Major contributors to
this report were Michael Clements, LaSonya Roberts, and Mathew Scire.

Sincerely yours,




Stwnleyins
Associate Director, Housing and
 Community Development Issues




(385834)


5                                         GAO/RCED-00-49R Changes to SBA's 7(a) Loan Program