Small Business Administration: Observations on the Disaster Loan Program

Published by the Government Accountability Office on 2003-05-01.

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

                            United States General Accounting Office

GAO                         Testimony
                            Before the Committee on Small Business
                            and Entrepreneurship, U.S. Senate

For Release on Delivery
Expected at 9:30 a.m. EDT
Thursday, May 1, 2003       SMALL BUSINESS
                            Observations on the
                            Disaster Loan Program
                            Statement of Davi M. D’Agostino
                            Director, Financial Markets and Community Investment

This is a work of the U.S. government and is not subject to copyright protection in the
United States. It may be reproduced and distributed in its entirety without further
permission from GAO. However, because this work may contain copyrighted images or
other material, permission from the copyright holder may be necessary if you wish to
reproduce this material separately.
Madam Chair and Members of the Committee:

I am pleased to be with you today at this roundtable to discuss the role of
the Small Business Administration’s (SBA) Disaster Loan Program in
responding to the September 11, 2001, terrorist attacks, general
performance measures for the program, and the effects of SBA’s program
to sell loans to private investors on disaster loans and their borrowers. As
you know, the effects of the September 11 attacks were felt not only in
New York but also around our country, with the economic damage
occurring in states as far west as California. The unique nature of the
attacks and the government’s response required SBA to make
unprecedented efforts to expand its disaster lending coverage and to be
flexible in its efforts to serve those needing assistance. Notwithstanding
SBA’s extraordinary performance in responding to the September 11
attacks, our work showed that the Disaster Loan Program’s performance
measures do not fully or adequately reflect SBA’s actual performance. In
reviewing SBA’s loan sales program, which includes disaster loans, we
identified three areas needing improvement: tracking borrower inquiries
and complaints; sales budgeting and accounting, which affect the
reliability of SBA financial statements and budget information; and
reporting on the operational benefits of the loan sales.

My remarks today will focus on SBA’s (1) response to the September 11
terrorist attacks; (2) performance plans and measures for its Disaster Loan
Program; and (3) loan asset sales program, which involves selling disaster
and other loans.1 My comments are based on our recent reports on SBA’s
Disaster Loan Program (Small Business Administration: Response to
September 11 Victims and Performance Measure for Disaster Lending,
GAO-03-385, Jan. 29, 2003) and loan asset sales program (Small Business
Administration: Accounting Anomalies and Limited Operational Data

 For information on assistance provided to small businesses in the Lower Manhattan area
after September 11 by SBA and other government agencies, please see U.S. General
Accounting Office, September 11: Small Business Assistance Provided in Lower
Manhattan in Response to the Terrorist Attacks, GAO-03-88 (Washington, D.C.: Nov. 1,

Page 1                                                                     GAO-03-721T
          Make Results of Loan Sales Uncertain, GAO-03-87, Jan. 3, 2003).2 Both are
          available on our Web site: www.gao.gov.

          The nature of the September 11 attacks and subsequent government
Summary   actions presented SBA’s Disaster Loan Program with new and difficult
          challenges. Specifically, small businesses in both the declared disaster
          areas and around the nation suffered economic injury. SBA sought to
          respond to the concerns of small businesses in the months following
          September 11 by extending eligibility for economic injury loans
          nationwide—a marked change from earlier disasters that affected
          primarily businesses in one geographic location. In addition, SBA modified
          both the terms and lending practices of its Disaster Loan Program—for
          example, by reducing the amount of documentation some borrowers
          needed to provide. Congress supported these efforts with supplemental
          appropriations that allowed SBA to offer larger loans to a relatively broad
          population of victims. By the end of fiscal year 2002, the agency had
          worked with individuals and businesses in all 50 states, the District of
          Columbia, and the U.S. territories, approving 9,700 loans totaling
          $966 million.

          We found that SBA had adapted its Disaster Loan Program to respond to
          the needs of September 11 victims but that SBA’s performance measures
          did not provide congressional decision makers with an accurate
          description of the program’s performance. For example, two of SBA’s six
          performance measures assessed only one discrete step in the loan
          application and disbursement processes—the application process. In
          addition, some output measures3 had not kept up with SBA’s actual
          progress in assisting disaster victims. Further, we identified features in
          SBA’s description of its Disaster Loan Program in the 2002 and 2003
          performance plans that made assessing the agency’s progress in attaining

           Also see April 29, 2003, testimony before the Subcommittee on Government Efficiency and
          Financial Management, Committee on Government Reform, U.S. House of Representatives.
          U.S. General Accounting Office, Small Business Administration: Loan Accounting and
          Other Financial Management Issues Impair Accountability, GAO-03-676T (Washington,
          D.C.: Apr. 29, 2003).
           According to Office of Management and Budget (OMB) guidance, outputs are the level of
          activity that can be produced or provided over a given period of time or by a specific date.
          Outcomes are the intended results, effects, or consequences that occur from carrying out
          program activities. OMB, Preparation and Submission of Strategic Plans, Annual
          Performance Plans, and Program Performance Reports, Circular No. A-11, Part 6.
          (Washington, D.C: June 2002).

          Page 2                                                                        GAO-03-721T
             its strategic goals difficult. For example, although SBA guidance
             recommended that program goals be outcome oriented, SBA’s 2003
             performance goal was output oriented.

             Our review of SBA’s five loan sales from August 1999 to January 2002
             revealed that 85 percent of the $4.4 billion in loans sold were disaster
             assistance home and business loans. SBA established some policies to
             protect borrowers whose loans were sold. For example, disaster loans less
             than 2 years old were not sold because they typically required more
             servicing and sometimes had to be increased to cover exigencies, such as
             revised physical damage estimates. In trying to determine how borrowers
             reacted to having their loans sold, we found that SBA relied on borrower
             inquiries and complaints to determine whether purchasers of the loans
             were using prudent loan servicing practices. However, information on
             borrowers’ reactions was incomplete because SBA did not have a
             comprehensive process to capture the inquiries and complaints it receives.
             Moreover, we found serious issues in SBA’s budgeting and accounting for
             the loans sold, as well as the remainder of the portfolio. For example, SBA
             incorrectly calculated the accounting losses on the loan sales and lacked
             reliable financial data to determine the overall financial impact of the
             sales. In addition, there were significant unexplained declines in the
             subsidy allowance for the disaster program. We discussed these issues
             with SBA’s auditor who subsequently withdrew its “clean” financial
             statement audit opinions for fiscal years 2000 and 2001 and disclaimed an
             opinion for 2002. SBA is continuing to work on resolving its accounting
             and financial reporting problems. Finally, our analysis of the operational
             benefits from loan sales suggested that some benefits that SBA reported,
             such as reductions in servicing and workload volume, either had not yet
             materialized or were overstated.

             When disasters such as floods, tornadoes, or earthquakes strike, federal,
Background   state, and local government agencies coordinate to provide assistance to
             disaster victims. SBA, through its Disaster Loan Program, is part of this
             effort. SBA provides loans to households and businesses without credit
             available elsewhere at a maximum rate of 4 percent and up to a 30-year
             term. For households or businesses with credit available elsewhere, SBA
             provides loans at a maximum rate of 8 percent and, for businesses, up to a
             3-year term. Business loans are available up to $1.5 million,4 loans for

              Even if a business receives a loan to cover both physical damage and economic injury, the
             total loan amount generally cannot exceed $1.5 million.

             Page 3                                                                      GAO-03-721T
physical damage to homes are available up to $200,000, and loans for the
repair or replacement of personal property are available up to $40,000.

Like other federal programs, SBA’s Disaster Loan Program follows
performance measurement guidelines under the Government Performance
and Results Act (GPRA) of 1993.5 GPRA requires agencies to set multiyear
strategic goals in their strategic plans and corresponding annual goals in
their performance plans, measure performance toward the achievement of
those goals, and report on their progress in their annual performance
reports.6 Annual performance plans are sent to Congress soon after the
transmittal of the President’s budget and provide a direct linkage between
an agency’s long-term goals and mission and day-to-day activities. Related
annual performance reports describe the degree to which performance
goals have been met. Guidance from the Office of Management and Budget
(OMB) indicates that performance plans should include measures of
outcomes—intended results—when the outcomes can be achieved during
the fiscal year covered by the plan. Otherwise, the guidance recognizes
that the performance plans will predominantly include measures of
outputs (program activities) rather than outcomes.

In 1999, SBA began a loan asset sales program, at the direction of OMB, to
reduce the amount of debt the agency owned and serviced. OMB is
interested in increasing loan asset sales in order to improve the
management of loan assets and to transfer loan servicing responsibilities
to the private sector. Our review focused on SBA’s first five loan sales
through January 2002 in which 110,000 loans with an outstanding balance
of $4.4 billion were sold. Approximately 85 percent of the dollar volume of
loans SBA sold were disaster assistance loans made directly by SBA, most
of which have below-market borrower interest rates. The remaining 15
percent were mostly defaulted 7(a) loans, made by SBA’s lending partners
(primarily banks).

P.L. 103-62, GPRA 1993.
 OMB provides guidance on developing these plans in “Preparation and Submission of
Strategic Plans, Annual Performance Plans, and Annual Program Performance Reports,”
Circular No. A-11, Part 6 (Washington, D.C: June 2002).

Page 4                                                                  GAO-03-721T
                         In the weeks and months following the terrorist attacks, SBA and
SBA Expanded and         Congress faced the challenge of responding to the lingering effects of the
Changed the Terms of     attacks and subsequent federal actions on small businesses throughout the
                         country. SBA responded first in Lower Manhattan, then expanded its
Its Disaster Loan        response as additional parts of the New York City and Pentagon areas
Program in Response      were designated disaster areas. Ultimately, SBA helped small businesses
                         around the country with disaster lending. In response to the concerns
to the September 11      expressed by small businesses, SBA and Congress modified the program,
Attacks                  expanding eligibility for economic injury loans to small businesses around
                         the country, providing translators for applicants, modifying the size
                         standards for small businesses, expediting the loan approval and
                         disbursement processes, and providing larger loans.

SBA’s Response Covered   SBA’s response to the terrorist attacks began on September 11, when SBA
Small Businesses         officials arrived in Lower Manhattan to begin coordinating the agency’s
Nationwide               efforts. The initial disaster area in New York City and New Jersey
                         eventually expanded to include additional counties in Connecticut,
                         Massachusetts, New Jersey, New York, and Pennsylvania. Maryland,
                         Virginia, and parts of the District of Columbia were also declared disaster
                         areas for SBA purposes. As the United States began to deploy military
                         personnel in response to the terrorist attacks, small businesses nationwide
                         affected by the loss of employees called up as military reservists were
                         eligible to apply for a disaster loan under the Military Reservist Economic
                         Injury Disaster Loan (EIDL) program.7 Small businesses across the nation
                         that were adversely affected by the lingering effects of the attacks and
                         subsequent government action, such as airport closings and the
                         precipitous drop in tourism, were also eligible to receive disaster loans
                         under SBA’s Expanded EIDL program. In essence, the entire country was
                         deemed a disaster area.

                         More than half the loans went to small businesses outside the area of the
                         attack sites in New York City and at the Pentagon, with businesses in
                         Florida and California receiving the second and third largest share of loans
                         (see fig. 1). Loans ranged from $300 to $1.5 million, with $50,000 as the
                         most frequently disbursed amount (11 percent of all loans). Businesses
                         outside the immediate sites of the attacks generally received slightly more
                         than those close by, in part because they did not have access to the

                          The Military Reservist EIDL program is available to small businesses whenever the
                         government calls military reservists to duty, not just during federally declared disasters.

                         Page 5                                                                          GAO-03-721T
                                       resources available in New York City. The loans were spread among
                                       industries, with no single type of business accounting for most of the
                                       funds (see fig. 2). The manufacturing sector received the most funds,
                                       followed by professional, scientific, and technical services; transportation
                                       and warehousing; wholesale trade; and accommodation and food services.

Figure 1: Geographic Distribution of SBA September 11 Loan Disbursements





                                                                3%                                 Fla.

                                       Dollars in millions

                                                $100 to $400

                                                $50 to $100

                                                $20 to $50

                                                $10 to $20

                                                Less than $10

Source: GAO analysis of SBA data.

                                       Page 6                                                              GAO-03-721T
                             Figure 2: SBA September 11 Business Loan Disbursements, By Industry

SBA and Congress             In the months after the terrorist attacks, small business owners affected by
Modified the Disaster Loan   the terrorist attacks presented a number of concerns to Congress about
Program in Response to       SBA’s Disaster Loan Program. SBA officials regarded these comments as
                             valuable feedback and worked with Congress to make several
Complaints from Small        modifications to the program for September 11 victims:
                             •   First, in October 2001, SBA issued regulations to make economic injury
                                 disaster loans available to small businesses nationwide, an
                                 unprecedented change to the Disaster Loan Program, according to SBA
                                 officials. SBA’s Expanded EIDL program enabled businesses outside
                                 the declared disaster areas to apply for loans to cover “ordinary and
                                 necessary” operating expenses that could not be met because of the
                                 attacks or related actions of the federal government between
                                 September 11 and October 22, 2001.

                             •   Second, SBA printed informational packets in languages such as
                                 Spanish and Chinese; provided multilingual staff at its offices who
                                 could speak Mandarin Chinese, Croatian, Arabic, and Spanish; and was
                                 prepared to send employees with additional language capabilities to
                                 New York City.

                             Page 7                                                           GAO-03-721T
•   Third, in February 2002, SBA modified the size standards for all
    September 11 loan applicants, allowing borrowers to take advantage of
    recent inflation-based adjustments.8 In addition, in March 2002, SBA
    increased the size threshold for travel agencies adversely affected by
    the attacks from $1 million in annual revenues to $3 million.

•   Fourth, to expedite loan processing, loan officers streamlined their
    needs analysis, calculating economic injury loans using the applicant’s
    annual sales and gross margin. By the end of fiscal year 2002, SBA was
    processing September 11 business loans, on average, in 13 days
    compared with 16 days for disaster assistance business loans
    processed in fiscal year 2001. To further expedite disbursement to
    those in the World Trade Center and Pentagon disaster areas, SBA
    decreased the amount of documentation needed to disburse up to

•   Fifth, in January 2002, Congress approved supplemental appropriations
    for SBA of $150 million, raised the maximum loan amount from $1.5
    million to $10 million, and deferred payments and interest for 2 years.9
    Congress also created the Supplemental Terrorist Activity Relief
    (STAR) program to provide assistance to small businesses affected by
    the terrorist attacks through SBA’s 7(a) loan guaranty program, which
    is not part of the Disaster Loan Program. Under the STAR program,
    SBA reduced the fee charged to lenders on new 7(a) loans from 0.50
    percent of the outstanding balance of the guaranteed portion of the
    loan to 0.25 percent. As of the end of fiscal year 2002, SBA had
    guaranteed about 4,700 STAR loans for $1.8 billion.

Some small businesses affected by the terrorist attacks maintained that
SBA’s underwriting criteria—for example, collateral requirements—were
too restrictive. They testified that SBA had withdrawn their applications
because they would not use their homes as collateral. They argued that it
was too risky to use their homes as collateral, especially since the survival
of their businesses was uncertain. SBA, however, did not change its
underwriting criteria for September 11 victims. SBA officials said that the

 In January 2002, SBA increased the revenue-based thresholds for determining the size of
businesses by the rate of inflation. In February 2002, SBA retroactively applied the
inflation-adjusted size standards to all businesses applying for September 11 loans,
allowing more businesses to seek assistance.
Emergency Supplemental Appropriations for Recovery and Response to Terrorist
Attacks on the United States Act, 2002 P.L. 107-117 (Emergency Supplemental Act of 2002).

Page 8                                                                      GAO-03-721T
                       agency makes every effort to approve each application by applying more
                       lenient credit standards than private lenders. However, the officials said
                       that they adhered to their credit standards to minimize losses and program

                       SBA data indicate that the 52 percent rate for withdrawing and declining
                       September 11-related loan applications was not out of line when compared
                       with other disasters or with private lenders. The primary reasons SBA
                       identified for withdrawing September 11 loan applications was a lack of
                       Internal Revenue Service (IRS) records to corroborate applicants’ income,
                       and applicants’ failure to provide additional information SBA had
                       requested. SBA officials said that the most common reasons for declining
                       September 11 loan applications were inability to repay the loan and
                       unsatisfactory credit. According to SBA, these were also the primary
                       reasons for withdrawing or declining nearly two-thirds of all SBA disaster
                       loan applications in fiscal year 2001.

                       SBA officials believed that many of the complaints about the disaster
                       program resulted from the mismatch between victims’ expectations of
                       SBA’s disaster program and the nature of the program. SBA officials told
                       us that they tried to minimize public confusion about the nature of the
                       assistance available from SBA by working closely with the media and
                       public officials to provide accurate information about the Disaster Loan

                       The six performance indicators SBA currently uses to measure the
SBA’s Disaster         Disaster Loan Program are
Program Performance
                       •    field presence within 3 days of a declaration,10
Measures Do Not
Capture the Scope of   •    loans processed within 21 days,
the Agency’s Efforts   •    customer satisfaction rate,

                       •    homes restored to predisaster condition,

                         Federal assistance, including all types of SBA disaster loans, is available once the
                       President declares that a major disaster or emergency situation exists. Governors may
                       request a disaster declaration from SBA if damage is minor or moderate and a declaration
                       from the Department of Agriculture if losses are confined to agricultural production. SBA
                       offers only economic injury loans in these last two situations.

                       Page 9                                                                      GAO-03-721T
                           •    businesses restored to predisaster condition, and

                           •    initial loan disbursement within 5 days of receiving closing documents.

                           We identified several problems with these measures. For example, several
                           are output measures that did not reflect the actual progress being made.
                           Some are proxies that did not accurately represent what was being
                           measured. There is a lack of measures for intermediate or end outcomes,
                           and features in SBA’s description of the Disaster Loan Program in its
                           performance plans made assessing the program difficult. Several of the
                           limitations we found had been identified in previous GAO or SBA
                           Inspector General reports and had not been corrected.11

Three Output Measures Do   Officials from SBA’s Disaster Area Offices (DAO) questioned whether the
Not Capture Progress       three output measures—establishing a field presence within 3 days of a
                           disaster declaration, processing loan applications within 21 days, and
                           disbursing initial loan amounts within 5 days of receiving the closing
                           documents—were appropriate indicators of timely service to disaster
                           victims since they did not, for example, capture recent program
                           improvements. SBA has had a 98 percent success rate in meeting the
                           target for establishing a field presence each fiscal year since 1998. Officials
                           from the area offices said that improvements in planning, interagency
                           coordination, and technology enabled them to have staff on site within 1
                           day of a disaster declaration. According to DAO staff, delays in
                           establishing a field presence generally occurred because SBA was waiting
                           for decisions from state officials.

                           SBA data and comments from DAO officials suggested that the second
                           output measure—processing loan applications within 21 days of receipt—
                           did not reflect improvements in past performance. For example, SBA
                           aimed for an 80 percent success rate for fiscal year 2001, but the actual
                           time required for processing averaged 13 days in fiscal year 2001 and fell
                           to 12 days in fiscal year 2002. The average time required to process the

                            See U.S. General Accounting Office, Managing for Results: Opportunities for Continued
                           Improvement in Agencies Performance Plans, GAO/GGD-99-215 (Washington, D.C.: July
                           20, 1999); Small Business Administration: Status of Achieving Key Outcomes and
                           Addressing Major Management Challenges, GAO-01-792 (Washington, D.C.: June 22, 2001);
                           and Final Audit Report—Results Act Performance Measurement for the Disaster
                           Assistance Program, Small Business Administration, Office of the Inspector General, Audit
                           Report 1-06 (Feb. 15, 2001).

                           Page 10                                                                    GAO-03-721T
                            September 11 business loans was also about 13 days. DAO officials
                            attributed their faster processing times to several agencywide

                            DAO staff also suggested that another measure—the 5-day target for
                            making initial disbursements once closing documents are received—did
                            not reflect past performance and was a low threshold. Before 2002, SBA
                            had an internal goal of ordering disbursements within 3 days of receiving
                            closing documents. When SBA included this measure in the performance
                            plan, the disbursement target was increased to 5 days to accommodate
                            weekends and holidays, because SBA’s system for tracking disaster loan
                            processing could not distinguish between workdays and other days.
                            Accustomed to the stricter 3-day standard, staff were able to meet the 5-
                            day standard with ease.

                            In commenting on a draft of our report, SBA indicated that the output
                            measures were established based on what was determined to be a
                            reasonable level of service in an average year, taking into account the
                            amount of resources required. Because disasters cannot be predicted,
                            officials did not think it would be feasible to adjust production levels
                            based on a single year’s performance. Even with some program
                            improvements, they believed it would be very difficult and costly to
                            maintain such levels during periods of multiple major disasters. Although
                            SBA acknowledged that a basis for modifying some output measures might
                            exist, the officials believed that the modifications should be based on an
                            average level of projected activity that takes into consideration some
                            permanent improvements that have been made to the program.

Two “Outcome” Measures      SBA officials indicated that three measures—number of homes restored to
Actually Assessed Outputs   predisaster condition, number of businesses restored to predisaster
                            condition, and customer satisfaction—were used to assess the effect, or
                            outcomes of lending to disaster victims. But these “outcome” measures
                            also had limitations. First, while the restoration of homes and businesses
                            was a stated outcome in SBA’s strategic and performance plans, SBA did
                            not actually measure the number of homes and businesses restored.
                            Instead, SBA reported on the number of home loans approved as a proxy
                            measure for the number of homes restored to predisaster condition.
                            However, these measures assessed what are actually program outputs
                            (loans approved) rather than stated outcomes (homes and businesses
                            restored). Such proxy measures, then, were likely to have overestimated
                            the number of homes and businesses restored because borrowers might
                            cancel the loan. According to SBA, about 10 percent of the loans approved

                            Page 11                                                        GAO-03-721T
                         for September 11 victims were cancelled by borrowers. Third, these
                         indicators used annual figures that were affected by factors outside of
                         SBA’s control, such as the number of disasters that occurred during a
                         given fiscal year. A more useful indicator would be the percentage of
                         homes and businesses receiving loans that were restored each year to pre-
                         disaster conditions.

                         To measure customer satisfaction, SBA used the results of its survey of
                         successful loan applicants. (SBA also used this survey to evaluate the
                         impact of the program.) But the survey methodology had significant
                         limitations. For example, it measured the satisfaction of only a portion of
                         the customers that the disaster loan program serves. Every DAO director
                         we interviewed indicated that all disaster victims were SBA customers and
                         that a broader population should be surveyed. In 2001, we and the SBA
                         Inspector General made the same suggestion to SBA. As we indicated
                         then, the survey method SBA had been using was likely to produce
                         positively skewed responses. SBA headquarters officials indicated that
                         they were resistant to surveying those who were denied loans because
                         they presumed that the applicants’ responses would be negative.

Some Measures Did Not    Recommendations from SBA’s Inspector General, and guidance from us
Assess Intermediate or   and within SBA, have encouraged the use of outcome measures for this
End Outcomes             program. But we found that only one of the performance measures SBA
                         was using—customer satisfaction—had the potential to assess a stated
                         outcome of the Disaster Loan Program. The other intended outcomes,
                         which could have been measured annually or biannually, such as jobs
                         retained or housing restored, were not measured.

                         In addition, SBA had stopped using intermediate outcome measures it had
                         used in the past—loan currency and delinquency rates—to assess the
                         quality of disaster loans. It also had not measured another potential
                         intermediate outcome from the underwriting process—having appropriate
                         insurance. As one DAO official suggested, having coverage such as flood
                         insurance potentially reduces the number of loans required in some
                         disaster-prone areas. As we have reported previously, such insurance can
                         reduce disaster assistance costs and could reduce the effect of a disaster
                         on its victims.12

                          U.S. General Accounting Office, Disaster Assistance: Information on Federal Costs and
                         Approaches for Reducing Them, GAO/T-RECD-98-139 (Washington, D.C.: Mar. 26, 1998).

                         Page 12                                                                  GAO-03-721T
                          SBA headquarters staff said that while they recognized some of these
                          shortcomings, they had limited ability to develop and use better outcome
                          measures. The staff indicated that the very nature of disaster lending was
                          unpredictable, making it difficult to set performance targets for
                          intermediate or end outcomes. One SBA official said that the agency is
                          reluctant to measure and report intermediate or end outcomes that are
                          outside its control. Other DAO officials indicated that conducting some
                          end outcome measurement methodologies would be expensive—for
                          instance, on-site inspections of a sample of homes and businesses to
                          assess restoration.

                          We made two recommendations designed to help SBA improve its
                          performance measures for disaster lending. First, we recommended that
                          SBA revise the performance measures to include more outcome measures;
                          assess more significant outputs, such as service to applicants or loan
                          underwriting; report achievements that can be compared over several
                          years, such as percentages; and include performance targets that
                          encourage process improvement rather than maintaining past levels of
                          performance. Second, we recommended that SBA revise and expand its
                          current research to improve its measures and evaluate program impact. To
                          improve its current measures, we suggested that SBA conduct research,
                          such as surveying DAO staff and reviewing relevant literature to identify
                          new outcome measures that could be tested. To evaluate its program
                          impact, SBA needs to ensure that its survey covers all disaster loan
                          applicants and to employ other methods, such as periodic analyses of
                          regional statistics, to assess the economic impact of the program on local
                          communities. SBA generally agreed with our recommendations and said it
                          is addressing our concerns. As of this month, SBA had distributed a
                          customer service survey to help evaluate the Disaster Loan Program’s
                          impact and was developing a broader survey. We will follow up with SBA
                          regarding the status of their efforts.

SBA’s Performance Plans   We identified several features of the description of the Disaster Loan
Had Limitations           Program in the 2002 and 2003 performance plans that make it difficult to
                          assess whether SBA is making progress in attaining its strategic goal. First,
                          between 2002 and 2003, the program’s performance goal changed from an
                          outcome-oriented goal (helping families recover from disasters) to an
                          output-oriented goal (streamlining disaster lending) without the required
                          explanation. GPRA requires agencies to explain why they change
                          performance goals, and OMB generally recommends that agencies use
                          goals that are outcome-oriented.

                          Page 13                                                          GAO-03-721T
                       Second, the 2002 and 2003 performance plans do not define the linkages
                       between each program output and each intermediate or end outcome. The
                       plans do not explain how the outputs (disaster loans) are related to the
                       performance indicators (field presence, customer satisfaction, and
                       application processing time frames). Third, the plans do not explain how
                       the performance measures or indicators are related to either program
                       outcomes or outputs. Fourth, performance indicators are added to or
                       dropped from the plans without explanation, making it difficult to
                       understand how and if SBA expects to improve or sustain its loan
                       processing performance.

                       The performance plans also contain incomplete or inaccurate information
                       on some performance indicators. For example, despite OMB and SBA
                       guidance, validation and verification information on field presence and
                       loan processing measures is omitted, making it difficult to assess the
                       quality of performance data. In addition, the 2003 performance plan
                       indicates that data on the number of homes restored to predisaster
                       condition are based on on-site inspections of homes. However, SBA
                       officials indicated that they use a proxy measure—the number of original
                       home loans approved—as the actual source of data for homes restored to
                       predisaster condition.

                       We recommended revising the section of the performance plan that covers
                       the Disaster Loan Program to establish direct linkages between each
                       output and outcome and the associated performance measure; accurately
                       describe proxy measures as either outcome or output measures;
                       accurately describe the validation and verification of performance
                       measures; and explain additions, deletions, or changes from the previous
                       year’s goals and measures. SBA also agreed with this recommendation.
                       SBA informed us this month that it has undertaken a long-term review of
                       the strategic plan with the aim of revising the performance goals and
                       measures and linking performance to the new plans and goals. We will
                       monitor SBA’s progress in implementing this initiative.

                       A large portion—85 percent in the first 5 sales—of the loans sold are
Loan Assets Sales      disaster loans previously serviced by SBA. SBA’s program to sell disaster
Affect Disaster Loan   loans that it makes directly to borrowers and subsequently services results
                       in private investors owning and servicing the loans over their remaining
Borrowers and the      terms. It was difficult for us to determine the reaction of borrowers whose
Loan Program           loans were sold because of incomplete records at SBA. We identified
                       numerous errors in SBA’s accounting for the loan sales, including
                       unexplained declines in SBA’s loss allowance account for disaster loans.

                       Page 14                                                        GAO-03-721T
                         Until corrected, these errors mean that SBA’s subsidy estimates and
                         reestimates for the disaster loan program cannot be relied upon. The
                         operational benefits from selling loans that SBA has claimed may be

Information on How       SBA built in some safeguards to protect borrowers when their loans are
Selling Disaster Loans   sold. But, because SBA’s process for documenting and tracking borrower
Affects Borrowers Is     inquiries and complaints has weaknesses, we could not determine how
                         many borrowers had actually contacted SBA with complaints or concerns
Incomplete               about the loan sales.

                         Borrowers have little control over what happens to their loans if SBA
                         decides to sell them. However, SBA has some policies intended to protect
                         the integrity of the programs that provided the loans. SBA’s programs,
                         including servicing disaster loans after they are made, are designed to help
                         the borrower recover from a disaster. To protect this public policy goal,
                         SBA’s loan sales agreements with purchasers require certification that the
                         investors are qualified to purchase and service the loans and will follow
                         prudent loan servicing practices. The loan sales agreement also prevents
                         purchasers from unilaterally changing the terms and conditions of the
                         loans. In addition, SBA does not sell some disaster loans, including those
                         issued to borrowers currently residing in a federally declared disaster area
                         and those that are less than 2 years old. According to SBA, more servicing
                         is typically required in the first 2 years of a disaster loan—such as changes
                         due to revised physical damage estimates.

                         Nevertheless, we were not able to validate the way in which borrowers
                         reacted to the loan sales because SBA could not provide a reliable
                         estimate or information on the number of borrowers who had contacted
                         them about their sold loans. Complete and reliable information on
                         borrower complaints is important because SBA officials told us that when
                         a borrower complained about a servicing action they contacted purchasers
                         to collect additional information and determine whether a purchaser was
                         breaching the borrower protections. One reason why SBA’s tracking
                         system is ineffective is that borrowers with questions or complaints can
                         call or write to several different SBA offices, or to a representative of
                         Congress. Some SBA field office officials told us that SBA does not
                         provide them with clear guidance on how to respond to or document such
                         complaints. Officials from seven district offices, three servicing centers,
                         and two disaster area offices told us that they had received calls and
                         letters from borrowers who had concerns about loans that had been sold.
                         But the methods for documenting inquiries and complaints varied across

                         Page 15                                                          GAO-03-721T
offices, except for congressional letters, which were consistently
forwarded to SBA headquarters. In August 2001, SBA began providing a
toll-free number for borrowers to call with questions or complaints about
loan sales. Borrowers were informed about the toll-free number in a letter
telling them how to contact the new owner of their loan. However, field
office staff did not receive any guidance regarding the purpose and use of
the toll-free number.

Though we were unable to determine how many borrowers have
contacted SBA about their sold loans, we reviewed 133 of the 155 written
inquiries and complaints documented at headquarters, along with SBA’s
written responses, to identify the types of questions and problems
borrowers may have when their loans are sold. Our analysis showed that
almost half (65) were inquiries and concerns about their loans being sold,
requests to buy their own loans, or pleas not to have their loans sold.
However, 47 of the borrowers complained about a purchaser’s servicing
action. For example, some letters involved disagreements or frustration
with servicing decisions, such as refusing to subordinate or release
collateral,13 or imposing a fee to complete a servicing action such as
subordination. Another 18 letters were from borrowers who wanted to
defer payments or change the amount of their monthly payments because
of financial problems, and felt they were not getting appropriate treatment
from the purchasers of their loans.

To address these weaknesses in the loan sales program, we recommended
that SBA develop procedures for documenting and processing inquiries
and complaints from borrowers, and then provide guidance to the field
offices about implementing them. SBA reported to Congress in March 2003
that it would soon issue a procedural notice to its field offices providing a
uniform process for handling borrower inquiries and complaints. SBA
stated that it also intends to establish an e-mail account for use by all
employees to record and forward borrower comments to the asset sales
team at headquarters, establish a database to track borrower comments,
and enhance a tracking system used for residential borrower inquiries at a
servicing center. We will follow up with SBA to monitor its
implementation of our recommendations.

  “Subordination” occurs when a lender allows a new or existing loan to take a superior lien
to another loan. For example, a borrower with an SBA disaster home loan may want SBA
or a lender to subordinate the disaster loan to a new or refinanced home mortgage.

Page 16                                                                      GAO-03-721T
SBA’s Accounting for Loan   During our review, we found errors that we believe could have
Sales and the Remaining     significantly affected the reported results in the budget and financial
Portfolio Was Flawed        statements for fiscal years 2000 and 2001. Because of errors we identified,
                            SBA’s auditor withdrew its clean audit opinions for those years and issued
                            disclaimers of opinion. Moreover, because of these and other financial
                            management issues, the auditor has disclaimed an opinion on SBA’s
                            financial statements for 2002. Although this roundtable is not intended to
                            explore the intricacies of accounting, I will briefly comment on our
                            findings, which are fully discussed in the report and testimony cited

                            SBA incorrectly calculated the accounting losses on the loan sales and
                            lacked reliable financial data to determine the overall financial impact of
                            the sales. Further, because SBA did not analyze the effect of loan sales on
                            its remaining portfolio, its reestimates of loan program costs for the
                            budget and financial statements cannot be relied upon. In addition, SBA
                            could not explain significant declines in its loss allowance account for
                            disaster loans. Until SBA corrects these errors and determines the cause of
                            the precipitous decline in the loss allowance account, the subsidy
                            estimates and reestimates for the disaster loan program cannot be relied
                            on. These errors and the lack of key analyses also mean that congressional
                            decision makers are not receiving accurate financial data to make
                            informed decisions about SBA’s budget and the level of appropriations the
                            agency should receive.

                            We recommended that, before doing more loan asset sales, SBA correct
                            the accounting and budgeting errors and misstatements. And that SBA’s
                            Inspector General, with SBA’s independent auditors, should assess the
                            impact of the identified errors and determine if the prior audit opinions
                            need to be revised. SBA is working to respond to these recommendations
                            and, as we noted above, the auditor has withdrawn the previously issued
                            clean audit opinions because they could not be relied upon. We will be
                            monitoring SBA’s continuing efforts to resolve these issues.

                             GAO-03-87 and GAO-03-676T.

                            Page 17                                                        GAO-03-721T
Loan Sales Have Reduced    SBA reported that loan asset sales had benefited the agency’s operations
SBA’s Loan Servicing       by reducing loan servicing, and that this reduction in loan servicing
Volume, but Other          volume should help allocate resources to other areas necessary to
                           achieving SBA’s mission and help the agency to manage its loan portfolio
Operational Benefits May   more effectively. Though we found that loan servicing volume had
Be Overstated              declined for SBA disaster home loan centers, the effect on regular
                           business loans was less clear. Furthermore, despite these reductions in
                           loan servicing volumes, SBA had not yet redeployed staff to more mission-
                           critical activities, such as lender oversight and business outreach. We
                           found that loan sales have mostly reduced the servicing workloads for
                           disaster assistance loans. They have had less impact on servicing
                           workloads for 7(a) business loans, because lenders did not always consent
                           to sell these loans. Because the reduction in loan servicing has involved
                           disaster assistance loans, it was unclear to what extent loan sales would
                           help the agency realign its workforce in the district offices that primarily
                           serve small businesses.

                           SBA has also reported that the loan sales have prompted borrowers to pay
                           their loans in full, revealed inconsistencies in the application of the
                           agency’s servicing procedures, and highlighted weaknesses in its
                           information system. We found some support to show that the loan sales
                           had produced portfolio management efficiencies. But we also found that
                           some of the benefits SBA had reported began before the loan sales
                           program, or could have been caused by other factors. For example,
                           borrowers of disaster loans who refinanced their homes while lower
                           interest rates were available often paid off their disaster loans, even
                           though their disaster loans had low interest rates.

                           To provide Congress and SBA with a better understanding of the impact of
                           loan sales on SBA’s operations, we recommended that SBA conduct a
                           more comprehensive evaluation of the loan sales’ impact on the agency
                           and the cost savings from the sales. SBA recently stated that it will
                           conduct such an evaluation.15 We will follow up with SBA as it addresses
                           our recommendation.

                           Madam Chair, Members of the Committee, this concludes my prepared
                           statement. I would be happy to answer any questions at this time.

                            Hector V. Barreto, Administrator, Small Business Administration, Letter to The Honorable
                           Susan Collins, Chair, Committee on Government Affairs, U.S. House of Representatives,
                           March 7, 2003.

                           Page 18                                                                    GAO-03-721T
                  For information on this statement, please contact Davi D’Agostino,
Contacts and      Director, Financial Markets and Community Investment, at (202) 512-8678
Acknowledgments   or Katie Harris, Assistant Director, at (202) 512-8415. You may also reach
                  them by e-mail at dagostinod@gao.gov or harrism@gao.gov. Other
                  individuals who made key contributions to this testimony or related work
                  include Dan Blair, Kristy Brown, Linda Calbom, Marcia Carlsen, Emily
                  Chalmers, Patricia Donahue, Julia Duquette, David Eisenstadt, and Kay

                  Page 19                                                        GAO-03-721T