TABLE OF CONTENTS

EXECUTIVE SUMMARY ii

BACKGROUND 1

OBJECTIVES, SCOPE, AND METHODOLOGY 2

LOSS RATES 3

TABLES

1. 7(a) General Business Loan and Banking Industry Loss Rates 6

A Table 1: Ending Year Portfolio Balance and Rate of Growth, FY 1986 to

1995 10

Table 2: 7(a) Combined Commercial Loss Rates (Pre- and Post-Credit

Reform Loans), FY 1992 to 1995 10

Table 3: 7(a) General Business Loan Purchase Rates, FY 1986 to 1995 11

Table 4: 7(a) General Business Loan Combined Actual Loss Rates,

FY 1992 to 1995 11

B Glossary 12

C Office of the Chief Financial Officer Comments 14

D Office of Financial Assistance Comments 18

E Contributors to This Report 19

ABBREVIATIONS

ABA American Bankers Association

C&I Commercial and Industrial Loans

Fed Federal Reserve Bank

FDIC Federal Deposit Insurance Corporation

GAO General Accounting Office

OCC Office of the Comptroller of the Currency

OCFO Office of the Chief Financial Officer

OFA Office of Financial Assistance

OIG Office of the Inspector General

SBA Small Business Administration

SIC Standard Industrial Classification

SOP Standard Operating Procedure

EXECUTIVE SUMMARY

This inspection was initiated at the request of the Administrator of the Small Business Administration (SBA). It focuses on the methodology used to compute loss rates for the 7(a) general business loan program and on the use of these rates by SBA officials. Each year the Office of the Chief Financial Officer (OCFO) publishes an Annual Loss Study which includes the "commercial," "actual," and "ultimate" loss rates for each of SBA’s loan programs. The commercial, or one-year, loss rate was developed to compare program performance with that of private sector lenders. The actual loss rate is a cumulative rate, while the ultimate loss rate is a projection of future losses. Because loss and default rates are sometimes used interchangeably, we have included a discussion of default rates in this inspection.

The methodology used by SBA to calculate the commercial, or one-year, loss rate is fundamentally valid. While there are slight differences between the SBA and the banking industry regarding the specific elements composing the numerator and denominator, the added components in SBA’s calculation have a minimal effect on the loss rate.

SBA’s commercial loss rate is not comparable with banking industry loss rates because the loan objectives and portfolios differ significantly. SBA and banking objectives vary substantially because the latter are for-profit businesses charged with providing dividends to their shareholders. SBA’s loan guarantee program is designed to promote small business formation and growth. Further, SBA’s loan portfolio is significantly different from that of the banking industry. First, SBA’s loss rate reflects a small business loan portfolio, while banking industry loss rates include loans to all businesses, large or small. Second, banking industry loans cut across all industries, while SBA’s are concentrated in two Standard Industrial Classification (SIC) codes--retail trade and services. Third, the banking industry commercial and industrial (C&I) loss rates exclude real estate loans, while real estate loans are not separated from SBA’s loan portfolio. Differing charge-off policies also affect the timing of charge-offs and, therefore, the loss rates.

The usefulness of the commercial loss rate is limited by two factors: (1) a rapidly expanding portfolio and (2) the three-to-four year lag between loan disbursals and charge-offs. SBA’s portfolio has increased by 118 percent in the past ten years. During the same period, the banking industry portfolio has increased only five percent. Although the commercial loss rate shows trends in SBA’s loan portfolio over time and may indicate improved performance for loans made five to ten years ago, it provides little information about the behavior of loans disbursed during the last three-to-four years. Thus, SBA’s rapid portfolio growth can create a misleading picture with regard to loss rates because most losses on unseasoned loans are unlikely to have occurred.

SBA program officials no longer use the commercial loss rate as a management tool. They indicated that they now focus on cohort performance and other indicators such as purchase, currency, and subsidy rates. Since the FY 1992 implementation of credit reform, the OCFO has tracked all loans on a yearly cohort basis. This information provides the best measure of loan performance because cohorts at the same stages of the maturation process can be compared.

For these reasons, we recommend that SBA discontinue the use of the commercial loss rate for the 7(a) general business loan program.

Loss and default rate data are not always properly identified, and may inadvertently misrepresent 7(a) loan program performance. We found that the presentation of SBA’s loss rates in the Annual Loss Studies and in some submissions to Congress needs additional clarification. Moreover, because there is no standard definition for the term "default" either in SBA or in the banking industry, there can be confusion over what SBA default rates represent--a loan payment delinquency rate or a purchase rate. According to program officials, it is not unusual for people to confuse the loss and default rates. We recommend that the OCFO and OFA ensure that information on loss and default rates is more clearly defined and labeled.

While the actual and ultimate loss rates are used primarily for internal purposes, the Annual Loss Studies in which they are published contain useful program performance information. Credit reform, however, has significantly changed the way loan information is tracked. This has necessitated two different methods of presenting loan information in the Annual Loss Study--one for pre- and one for post-credit reform data. We recommend that the OCFO determine, in consultation with appropriate program officials, whether revisions can be made in the format of the Annual Loss Study to clarify loan loss data.

In response to our other two recommendations, the OCFO plans to condense the Annual Loss Rate Study and present the information in a clearer and more consistent manner.

BACKGROUND

This OIG inspection was initiated at the request of the Administrator of the SBA to validate the methodology used to compute loss rates for the 7(a) business loan program. SBA’s OCFO, which is responsible for calculating the loss rates for the 7(a) program, publishes an Annual Loss Study which includes the "commercial," "actual," and "ultimate" loss rates for each of SBA’s loan programs.

Prior to 1989, SBA calculated only the actual and ultimate loss rates for loan programs. The actual loss rate represents total cumulative losses, while the ultimate loss rate is an estimate of total future losses on loans presently outstanding. Neither rate has an equivalent in the private banking sector. According to SBA officials, the actual loss rate does not provide a fair comparison with the yearly loss rate used by the banking industry because it is calculated on a cumulative basis. Therefore, the Comptroller’s Office developed a commercial, or one-year, loss rate for each loan program to permit SBA program officials to compare program performance with that of private sector lenders. Although the commercial loss rate was also used in the past to evaluate 7(a) loan trends, SBA program management currently uses better measures to assess loan portfolio trends, including subsidy, purchase, and delinquency rates.

Early in the inspection, the OIG team found that there is confusion surrounding the use of 7(a) loss and default rates, which are sometimes used interchangeably. Moreover, neither SBA nor the banking industry has a standard definition of "default." We therefore included in this inspection a discussion of the default rate.

OBJECTIVES, SCOPE, AND METHODOLOGY

This inspection focuses on the methodology used by SBA’s OCFO to calculate 7(a) general business loan program loss rates and on the use of these rates by SBA program officials. The financial data employed to calculate both loss and default rates is entered into the loan accounting system by SBA field offices and loan centers. It was not, however, within the scope of this review to audit those offices to determine the accuracy of the data being input. It should be noted that the OIG’s Auditing Division is currently conducting an audit of 7(a) liquidations which will test the accuracy of loan charge-offs and recoveries in a statistical sample of loans.

To conduct this inspection, we interviewed OCFO and OFA officials and reviewed the OCFO’s Annual Loss Studies; the relevant standard operating procedures (SOP); the annual financial statement audits performed by Cotton & Company; and reports by the OIG, the General Accounting Office (GAO) and others that address the 7(a) business loan program. To determine the methods used in the banking industry to calculate loss rates, we interviewed representatives of the ABA and the three major Federal bank regulatory agencies--the Office of the Comptroller of the Currency (OCC), the Federal Reserve (Fed), and the Federal Deposit Insurance Corporation (FDIC). We conducted our work between October 1995 and March 1996 in accordance with the Quality Standards for Inspections issued in March 1993 by the President’s Council on Integrity and Efficiency.

Commercial Loss Rate. SBA computes and uses a "commercial loss rate" to compare the performance of 7(a) loans with those of private sector lenders. The commercial loss rate is a one-year measurement which is calculated by dividing net charge-offs for the year by the average dollar value of loans outstanding.

Following the purchase of a defaulted guaranty loan from a lender, SBA charges off the portion of the loan that is determined to be uncollectible or of such little value that its continuance as a bankable asset is unwarranted. In determining gross charge-offs, SBA includes (1) the amount charged off less any reversals of previously charged-off items, (2) recoverable expenses, and (3) losses that occur on judgments.

Validity of the Methodology. We found the methodology SBA uses to calculate its commercial loss rate to be consistent with that used by Federal regulators in reporting banking industry loss rates. All FDIC insured banks are required to complete quarterly reports on their financial status. From these reports, the FDIC calculates industry loss rates by dividing net charge-offs by the average outstanding loan portfolio balance. There are, however, slight differences between the SBA and the banking industry with regard to the specific elements composing the numerator and denominator. For example, banks do not add expenses to the loan balance, while SBA adds recoverable costs; SBA’s commercial loss rate denominator includes judgments and acquired collateral that the banking industry excludes; and SBA averages the ending outstanding loan portfolio for the current and prior fiscal years, while the FDIC uses several different methods to calculate the outstanding loan balance. The added components in SBA’s calculation, however, are minor and have a minimal effect on the loss rate. Overall, SBA uses methodology similar to that of the banking industry, and we believe it to be fundamentally sound.

Comparability with the Banking Industry. SBA and bank objectives vary substantially. Banks are for-profit businesses charged with providing dividends to their shareholders. SBA is a "public-private partnership that uses tax payer dollars to leverage private capital that would not otherwise be made available to the small business sector." More importantly, while the method used by SBA to calculate a one-year loss rate is comparable with that used by the banking industry, the objects of comparison, i.e., the loan portfolios, are not. SBA uses commercial and industrial (C&I) loans in making comparisons. Even when only C&I loans are taken into account, SBA’s loan portfolio is significantly different from that of the banking industry. First, the industry loss rates include loans to both large and small business. According to SBA, small business loans make up only 39 percent of all C&I loans. Second, banking industry loans cut across all industries, while SBA’s portfolio is concentrated in only a few areas. SBA reports that 63 percent of its outstanding loans at the end of FY 1995 had been awarded to businesses in two SIC Code categories--retail trade and services. Third, whereas SBA includes real estate loans in its small business loan portfolio, banks track real estate loans separately.

In addition, SBA and banks have substantially different charge-off policies. Because banks must promptly charge-off the amount of any loan loss that becomes confirmed, they typically charge loans off their books prior to pursuing recoveries. In contrast, SBA’s policy is to make all reasonable efforts to collect the entire debt prior to charge-off. This can last several years, delaying the charge-off and its effect on the loss rate. Because of these significant differences, we believe that the 7(a) commercial loss rate should not be used to make comparisons between SBA loan losses and those of the private banks.

Limitations of the Commercial Loss Rate. The usefulness of the commercial loss rate is also constrained by two factors: (1) a rapidly expanding portfolio and (2) the three-to-four year lag between loan disbursals and charge-offs. SBA’s portfolio has increased by 118 percent in the past ten years, while the banking industry portfolio has grown only five percent. (See Table 1 in Appendix A).

During this period, SBA’s average outstanding loan balance (the loss rate denominator) increased significantly, while loan losses (the numerator) gradually declined. As a result, SBA’s commercial loss rate for guaranty loans has fallen from 3.87 percent in 1986 to .81 percent in 1995. (See Table 1.) Although the rate provides a useful view of SBA’s loan portfolio over time and may indicate improved program performance for loans made five to ten years ago, it provides little information about the performance of loans disbursed during the last three to four years, i.e., those loans that may not be sufficiently "seasoned" to reflect losses. Thus, SBA’s rapid portfolio growth can create a misleading picture with regard to loss rates because most losses on unseasoned loans are unlikely to have occurred.

The banking industry’s loss rates on commercial and industrial loans have steadily declined from 1.54 percent to .25 percent between 1986 and 1995 and are slightly lower than SBA’s commercial loss rates. (See Table 1.) Regulatory officials attribute the decline primarily to (1) more conservative underwriting by banks, (2) improvements in the national economy, (3) greater diversification in lending portfolios, in terms of loan type and geography, and (4) more sophisticated portfolio management and loan review systems.

Default Rates. The terms "loss rate" and "default rate" are sometimes used interchangeably in the media and in Congressional hearings. According to program officials, the loss and default rates are often confused due to the lack of a standard definition of the term default--within either SBA or the private banking industry. In the banking industry a loan is in technical default if the borrower is in violation of some specific provision of the loan agreement--even though the borrower may be current in his/her loan payments. Usually, however, banks do not place a loan in liquidation based on a technical default. According to Federal regulators, banks do not report a default rate, so no aggregate data on industry defaults is available. Banks do, however, track and report delinquent loans, and the FDIC compiles data on past due loans and computes delinquent loan ratios.

SBA’s Loan Servicing SOP defines default as a delinquency in a payment on a loan. OFA officials suggest, however, it is not entirely clear whether a default occurs when a payment on the loan is 30 days past due (and therefore no longer current), when the payment is 60 days past due, or when the loan is in liquidation. The OCFO treats SBA’s purchase rate (also called the "repurchase" rate) as the default rate. Under SBA’s guaranty agreement with lenders, a lender can request that SBA purchase the guaranteed portion of the outstanding loan if an installment remains unpaid for more than 60 days. The purchase rate is calculated by dividing total purchases for one year by total disbursements for the year. Because the outstanding portfolio (the denominator) has grown significantly while purchases (the numerator) have remained relatively constant, the combined pre- and post-credit reform purchase rates can be misleading. (See Table 3 in Appendix A.) Nevertheless, because purchases are the earliest indication of potential loan losses, there is a benefit to using purchase rates to monitor loan performance.

Even when the purchase rate is used as the default rate, there may still be some confusion. For example, a table labeled "Default Rates," which was submitted by SBA to a March 1995 hearing of the House Committee on Small Business, did not represent either delinquency rates or actual purchase rates, but forecasted purchase rates for annual cohorts without labeling them as such. In short, SBA needs to make a stronger effort to clarify the meaning of the default rates it cites and to apply them in a consistent manner.

Actual and Ultimate Loss Rates. In addition to the commercial loss rate, the OCFO also publishes actual and ultimate loss rates in the Annual Loss Studies. (See Table 4 in Appendix A for historical data.) There are no aggregate banking industry rates comparable to SBA’s actual and ultimate loss rates. Moreover, these loss rates are used for internal SBA purposes.

SBA has used the actual loss rate to track losses from the inception of the 7(a) loan program. It is composed of cumulative net charge-offs divided by cumulative disbursements. Because the actual loss rate is not limited to one year of data, it is significantly higher than SBA’s commercial loss rate and banking industry loss rates.

The ultimate loss rate is an estimate of total future losses on loans presently outstanding. The OCFO computes it by adding an estimate of losses on loans that have not yet been purchased (projected losses) and an estimate of losses on loans that have already been purchased (allowance for losses) to the cumulative net losses. The projected losses and allowance for losses figures are calculated using historical loan data and projections about future loan behavior. These estimates provide the basis for establishing a "reserve" for future losses on the existing pre-1992 portfolio.

It is assumed that most of the losses on loans in the pre-credit reform cohort will occur within the next ten years, and when all charge-offs and recoveries have been made, the actual and ultimate loss rates will converge. For loans approved after FY 1991, the cost of Federal credit is estimated based on the subsidy rate when the loan is disbursed, as opposed to a forecast of future losses. Thus the use of the subsidy rate to estimate loan costs will eliminate the need for an ultimate loss rate calculation.

The ratio of cumulative net losses to total cumulative disbursements.

An estimate of future losses on loans that have already been purchased.

Synonymous with the term write-off. It is the process by which SBA recognizes a loss on its records, removes uncollectible accounts from its active receivables, and authorizes cessation of loan office collection activity.

A group composed of loans approved in the same fiscal year. Cohorts may be further classified according to loan type and loan program. Loans approved prior to October 1, 1991, are tracked as a single cohort.

The ratio of net charge-offs to the average outstanding loan balance. It is calculated annually and is designed to produce a loss rate that is comparable to those published by the private sector.

Cumulative total of the principal amount loaned under a particular program. Depending on the loan type, e.g., direct or guaranteed, the disbursement of monies is made by SBA or the participating lender.

Provides a measure of Federal agencies’ costs for financial assistance programs. All Federal direct and guaranteed credit is required to be treated consistently across the Federal Government. This process is based on the calculation of a subsidy rate for each distinct loan program.

A classification used when it is necessary to enforce collection or when the Agency’s interest in collateral may be in jeopardy as a result of any type of insolvency proceedings.

A judicial determination in court that a borrower owes SBA money.

Gross charge-offs less recoveries.

Loans approved prior to October 1, 1991.

An estimate of future losses on loans that have not yet been purchased.

The ratio of total purchases to total disbursements.

Expenses incurred by SBA on behalf of the borrower that are charged to the borrower’s loan account, increasing the amount owed to SBA.

An estimate of the cost of Federal credit approved since October 1, 1991. It is based on a present value analysis of expected cash flows for each credit program for each approval year (cohort). By using the subsidy rate, the cost of Federal credit is recognized at the time the decision is made to approve program levels.

An estimate of total future losses on loans presently outstanding. It is computed by adding the projected losses and the allowance for losses to cumulative net losses.