Federal Deposit Insurance Corporation
Office of Audits
Office of Inspector General
Washington, D.C. 20434

DATE: March 18, 2003

MEMORANDUM TO: Michael J. Zamorski, Director, Division of Supervision and Consumer Protection

FROM: Russell A. Rau [Electronically produced version; original signed by Stephen M. Beard for Russell A. Rau], Assistant Inspector General for Audits

SUBJECT: The Division of Supervision and Consumer Protection's Examination Assessment of Subprime Lending (Audit Report No. 03-019)

The Federal Deposit Insurance Corporation’s (FDIC) Office of Inspector General has completed an audit of the Division of Supervision and Consumer Protection's (DSC) assessment of subprime lending in the course of safety and soundness examinations. (Note: The Federal Financial Institutions Examination Council (FFIEC) 1999 Interagency Guidelines for Subprime Lending define the term "subprime lending" as extending credit to borrowers who exhibit characteristics indicating a significantly higher risk of default than traditional bank lending customers. The FFIEC was established by the Congress to promote improved and consistent examination and supervision policies and procedures among the five financial institution regulatory agencies. The FFIEC includes representatives of the Federal Deposit Insurance Corporation, Board of Governors of the Federal Reserve System (FRB), Office of the Comptroller of the Currency (OCC), Office of Thrift Supervision (OTS), and National Credit Union Administration (NCUA).) Our objective was to determine whether DSC has taken reasonable steps to ensure that institutions: (1) manage risks associated with subprime lending programs effectively and price loans based on risk, (2) establish adequate allowance levels to cover losses, and (3) maintain capital levels that reflect the additional inherent risks associated with subprime lending. We conducted this audit because of concerns stemming from recent financial institution failures involving subprime lending activities. Additional information on the objective, scope, and methodology is contained in Appendix I.

As early as 1997, regulators were cautioning financial institutions regarding the potential risk associated with subprime lending. Subprime lending provides a credit source to borrowers that may not otherwise be available due to concerns with borrowers' credit history or repayment capacity. Financial institutions generally price subprime loans commensurate with the added risk and should consider this risk in establishing all reserves for loan losses and capital levels. Financial Institution Letter (FIL) # 44-97, dated May 2, 1997, entitled Risks Associated with Subprime Lending, discusses the fact that recent examinations revealed a number of financial institutions that were engaged in subprime lending activities without properly assessing or controlling the risks associated with this type of lending. As a result, many institutions have suffered losses, which in turn has jeopardized the overall financial health of those institutions. According to the FIL, banks typically get involved with subprime lending in the following areas:

• Lending directly to subprime borrowers,
• Purchasing subprime dealer paper or loans acquired through brokers,
• Lending directly to financing companies involved with subprime lending,
• Participating in loan syndications providing credit to such financing companies, and
• Acquiring asset-backed securities issued by these financing companies. (Note: The American Bankers Association’s Banking & Finance Terminology defines dealer paper as "An arrangement made by a bank to purchase loans from a dealer who sells automobiles or other durable goods to the public. Customers purchase from the dealer on credit, the dealer sells the loan to the banks, and the dealer's customers effectively become borrowers of the banks." The American Bankers Association’s Banking & Finance Terminology defines syndications as "Loans made by a group of banks to one borrower. In most cases the dollar amount requested exceeds the amount that the individual banks are willing or able to lend. Each bank receives a pro rata share of the income based on its level of participation in the credit." According to the American Bankers Association’s Banking & Finance Terminology, asset-backed securities are "Securities that have promised interest and principal payments that are backed by cash flows from an asset or portfolio of assets that generate the cash flows.")

Although FDIC-insured institutions with significant subprime asset exposures represent less than 2 percent of all FDIC-insured institutions, subprime activities contributed to 10 of 31 failures and 90 percent of insurance losses during the past 5 years. (Note: According to the FDIC Division of Insurance and Research Briefing Notes dated September 30, 2002, "Subprime assets include residential or other consumer credits to individuals with limited or blemished credit histories, residential mortgages with high loan-to-value (LTV) ratios, payday loans, or residual interests in securitized subprime loans. Only insured institutions with subprime assets equal to at least 25 percent of Tier 1 Capital are included in this analysis." According to the FDIC Division of Insurance and Research Briefing Notes dated September 30, 2002, "Poor bank management, weak internal controls, and fraud often magnified losses to the insurance fund.") The subprime market is attractive to lenders because of the high loan yields and servicing fee income generated by this type of lending; however, the higher costs associated with servicing and collecting on subprime borrowers can erode earnings as delinquencies rise. (Note: According to the American Bankers Association’s Banking & Finance Terminology, "Servicing of mortgage loans, credit card receivables, or other financial assets includes, but is not limited to, collecting principal, interest, and escrow payments from borrowers; paying taxes and insurance from escrowed funds; monitoring delinquencies; executing foreclosure if necessary; temporarily investing funds pending distribution; remitting fees to guarantors, trustees, and others providing services; and accounting for and remitting principal and interest payments to the holders of beneficial interests in the financial assets.") According to the FDIC Division of Insurance and Research (DIR), subprime loans originated in 2000 have thus far exhibited higher and faster-rising charge-off rates than similarly-aged loans. Also, losses associated with higher default rates are associated with subprime lending. (Note: The FDIC Division of Insurance and Research (DIR) was established on June 30, 2002, as a result of the merger of the Division of Insurance (DOI) and the Division of Research and Statistics (DRS).) FIL # 44-97 also identified the controls that help mitigate these risks, including the requirement that, "Effective lenders must appropriately stratify the additional default risk and price the subprime products accordingly. This requires constant monitoring and testing of credit scoring models to ensure that projected results are in line with actual performance."

In March 1999, the FFIEC issued the Interagency Guidelines on Subprime Lending. The guidelines defined subprime lending and required regulators to evaluate the related capital levels at examinations. (Note: The Federal Deposit Insurance Act Section 10 (d) requires the appropriate federal banking agency to conduct annual full-scope on-site examinations of each insured depository institution.) In January 2001, the FFIEC issued Expanded Guidance for Evaluating Subprime Lending Programs (Expanded Guidance). This Expanded Guidance supplements the FFIEC guidelines issued in March 1999 and is specifically tailored to institutions that have subprime lending programs with an aggregate credit exposure greater than or equal to 25 percent of Tier 1 Capital. (Note: DSC’s Manual of Examination Policies provides a definition of Tier 1 Capital as "the sum of: common stockholders' equity (common stock and related surplus, undivided profits, disclosed capital reserves, foreign currency translation adjustments, less net unrealized losses on available-for-sale equity securities with readily determinable fair values); noncumulative perpetual preferred stock; minority interests in consolidated subsidiaries; minus all intangible assets (other than limited amounts of mortgage servicing rights and purchased credit card relationships and certain grandfathered supervisory goodwill); identified losses (to the extent that Tier 1 Capital would have been reduced if the appropriate accounting entries to reflect the identified losses had been recorded on the institution's books); investments in securities subsidiaries subject to section 337.4; and deferred tax assets in excess of the limit set forth in section 325.5(g)." In addition, the Expanded Guidance refines the definition of subprime lending and clarifies the agencies' expectations regarding an institution's risk management processes and provides a more detailed discussion of the supervisory expectations for examinations of subprime lending programs.

DIR's April 2002 semiannual report on economic conditions and emerging risks in banking notes that one of the four main risks to FDIC-insured institutions is the risk of subprime lending. (Note: The other emerging risks identified by DIR were: (1) accounting transparency for off-balance sheet entities and its effect on risk assessment, (2) interest rate sensitivity of specialized mortgage lenders, and (3) commercial lending in formerly fast-growing metropolitan areas.) According to the report, the single most likely source of significant insurance losses related to the failure of FDIC-insured institutions in the near-term is subprime consumer lending. According to DIR, the problems with subprime lenders have not been simply the result of higher than expected loan losses but the tendency for the credit models used by subprime lenders to underpredict actual losses. On April 1, 2002, the FDIC Financial Risk Committee issued a report stating that subprime lenders and other banks with riskier business models continue to make up a disproportionate share of problem banks. (Note: According to the FDIC Division of Insurance and Research, the function of the FDIC Financial Risk Committee (FRC) is to bring together representatives from various FDIC divisions and offices to determine and discuss the risk exposure to the deposit insurance funds and the probability of bank failures. The mission of the FRC is to ensure that the contingent loss reserve adequately reflects risk to the insurance funds.)

The FDIC estimated that as of June 30, 2002, 128 FDIC-insured institutions (1.35 percent of all insured institutions) had significant holdings of subprime assets (see Table 1 below). As of that same date, 9,464 FDIC-insured institutions had $8.037 trillion in total assets, with $66.2 billion in subprime assets at the 128 institutions with exposure in subprime lending; however, a much larger percentage of institutions may have more limited involvement in this type of lending.

Source: DSC Internal Supervisory Database Reports (Note: According to DSC, the Internal Supervisory Database report is a confidential supervisory document and thus is not available to the public.)

According to the DIR September 2002 Briefing Notes on Subprime Lending, the financial condition of the nation’s 128 insured subprime lenders deteriorated through mid-year 2002 as losses escalated and several composite examination ratings were downgraded. (Note: The FDIC DSC Manual of Examination Policies defines a composite examination rating as an overall rating given to a bank based on the 6 components of the CAMELS rating. Financial institution regulators use the Uniform Financial Institution Rating System to evaluate a bank's performance. The performance areas identified by the CAMELS acronym are Capital Adequacy, Asset Quality, Management, Earnings, Liquidity, and Sensitivity to Market Risk. A rating of 1 through 5 is given, with 1 having the least regulatory concern and 5 having the greatest concern.) According to the Briefing Notes, a worsening employment picture, relatively high consumer debt burdens, and record bankruptcy levels adversely affected subprime loan quality through 2001 and early 2002. Also, heightened corporate layoffs pushed the national unemployment rate to 5.7 percent in August 2002, from 5.5 percent in February 2002. Bankruptcy filing rates continued to climb through mid-2002, and the second quarter 2002 personal bankruptcy filing rate was at a 10-year high.

In September 2002, DIR reported that increased foreclosure activity and higher subprime residential mortgage delinquencies had hampered credit quality among subprime mortgage lenders. Twenty-one (41 percent) of the 51 FDIC-supervised insured institutions that held subprime loan exposures exceeding 25 percent of Tier 1 Capital as of mid-2002 had composite ratings of 3, 4, or 5. As of June 2002, the median 1-4 family residential loan delinquency ratio among subprime mortgage lenders was 2.43 percent, roughly 1.5 times the ratio reported by prime retail institutions. (Note: According to the DIR September 2002 Subprime Briefing Notes, prime retail institutions are defined as insured institutions with 1-4 family real estate and consumer loans exceeding 25 percent of capital that are not in the DSC Internal Supervisory Database.) The FDIC has also reported that most subprime mortgage lenders experienced decelerating mortgage loan growth during the recent recession, which could translate into higher delinquency and loss rates prospectively. (Note: According to DIR the U.S. economy officially entered a recession in March 2001. The DIR December 2001 Briefing Notes on the U.S. Economy states that: (1) real gross domestic products (GDP) fell 1.3 percent in the third quarter of 2001, for the first time since early 1993; (2) in the aftermath of the September 11 attacks, growth in consumer spending slowed considerably, and personal consumption expenditures grew by an annualized 1.0 percent in the third quarter, down from a 2.5 percent growth a quarter earlier; (3) business investment continued to fall throughout the year, and nonresidential fixed investment fell by an annualized 8.5 percent in the third quarter, following a 14.6 percent decline in the second quarter; and (4) businesses continue to cut inventories as they face weaker demand, while private inventories shrank by $61.7 billion in the third quarter.) In general, compared with mortgages underwritten between 1996 and 2002, subprime first mortgages and home equity loans originated in 2002 experienced higher, early-stage delinquencies.

According to DIR's First Quarter 2002 Regional Outlook Report, subprime lenders face an increased risk of default due to the potential inability of subprime borrowers to repay their mortgages on a timely basis. The report also states that the FDIC estimates that fewer than 1 percent of all insured institutions have significant subprime residential mortgage exposures. Nevertheless, a much larger number of institutions probably have some limited involvement in subprime mortgage lending. The future impact will be determined based on the risk in the bank’s loan portfolios, their management of this risk, and the continuance of the recessionary period. The subprime mortgage market has not been through a full recessionary cycle; therefore, the repercussions from this type of lending are still unknown.

In its September 2002 Subprime Briefing Notes, DIR reported that performance measures have deteriorated among insured institutions that have subprime consumer loans. (Note: According to the DIR September 2002 Subprime Briefing Notes, consumer loans include credit cards, automobile loans, and other loans to individuals not secured by real estate.) The median consumer loan delinquency ratio reported by all subprime lenders was 3.5 percent, over 1.5 times the level reported by prime retail institutions. Although past-due consumer loan ratios moderated among subprime lenders during the first half of 2002, consumer loan losses rose dramatically, and consumer loan growth slowed. According to DIR, slower loan growth might have contributed to loss trends and could magnify delinquency and loss ratios as loan portfolios age. As of September 30, 2002, the nation’s 20 subprime credit card lenders reported especially high consumer loan delinquencies and charge-offs. (Note: According to the DIR September 2002 Subprime Briefing Notes, subprime credit card lenders include insured institutions with subprime credit card loans-to-Tier 1 Capital ratios exceeding 25 percent.) For these institutions, the median ratio of delinquent credit cards was 11.9 percent at mid-year 2002, which is more than 3.5 times the median delinquency rate reported by all non-subprime insured institutions with credit card-to-Tier 1 Capital ratios exceeding 25 percent. Although the delinquency rate increased only slightly, the median net credit card loss ratio among subprime credit card lenders almost doubled, rising from 8.5 percent in June 2001, to 16.1 percent in June 2002.

DSC has taken reasonable steps to ensure that institutions: (1) manage risks associated with subprime lending programs effectively and price loans based on risk, (2) establish adequate allowance levels to cover losses, and (3) maintain capital levels that reflect the additional inherent risks associated with subprime lending. Specifically, the interagency policies and procedures for examinations of subprime banks provided examiners with the necessary guidance to identify and assess the condition of subprime loan programs in insured institutions and the examiners adequately implemented them. The procedures specifically address the management of risk associated with subprime programs, stress the need for banks' risk management programs to address loan pricing, and set forth the requirements for calculating and maintaining adequate allowances for loan and lease losses (ALLL) and capital levels. FDIC examiners conducted pre-examination planning that included steps to look for indications of subprime programs and generally followed the interagency subprime examination procedures involving examinations of capital levels during onsite examinations. In addition, DSC maintains a quarterly database to assist in monitoring the condition of FDIC-insured institutions with subprime programs. Further, examiners noted that institutions had implemented corrective actions as a result of DSC examination findings related to the banks' subprime lending activities, including requirements for maintaining adequate levels of capital and adequate allowances to cover losses. (See Appendix II: DSC's Assessment, Identification, and Monitoring of Banks with Subprime Lending Programs.)

During the course of our audit, we identified one issue that may warrant management's attention. Specifically, existing guidance may not be sufficient for ensuring that custom credit scoring models correctly predict the creditworthiness of borrowers. (Note: The American Bankers Association’s Banking & Finance Terminology defines a custom credit scoring model as a custom-built statistical analysis model used by banks to estimate the credit worthiness of credit applicants.) As a result, there is a potential for a lack of consistency in onsite examinations of banks with subprime lending programs, particularly with regard to allowances for losses and capital level calculations. Also, in order for lenders to appropriately stratify the additional default risk and price the subprime products accordingly, constant monitoring and testing of credit scoring models is required to ensure that projected results are in line with actual performance. (See Appendix III: Reviews of Custom Credit Scoring Models.)

On March 10, 2003, the Director, Division of Supervision and Consumer Protection (DSC) provided a written response to the draft report. The response is presented in Appendix VII of this report. In reference to our observation related to procedures for reviews of custom credit scoring models, DSC management stated that the FDIC plans to offer additional training on the topic of custom credit scoring for a select group of specialists.

In addition, during the draft report phase of the audit, DSC requested that certain sensitive information in the draft report be deleted because of references to other regulators' subprime banks. We have done so, and deleting the information did not materially affect the report.

The purpose of the audit was to review the Division of Supervision and Consumer Protection's (DSC) process for assessing subprime lending risks during safety and soundness examinations. Our specific objective was to determine whether DSC takes reasonable steps to ensure that institutions:

• manage risks associated with subprime lending programs effectively and price loans based on risk,

• establish adequate allowance levels to cover losses, and

• maintain capital levels that reflect the additional inherent risks associated with subprime lending.

In order to accomplish our objective related to how subprime lending risk is addressed, we reviewed DSC's processes for:

• identifying banks' risks associated with subprime lending and conducting examinations of banks with subprime loan programs,

• monitoring and tracking the conditions of the banks' subprime loan programs between examinations, and

• implementing corrective actions when weaknesses are identified in banks' subprime loan programs.

The audit was conducted from March 21, 2002 through December 9, 2002, in accordance with generally accepted government auditing standards. The scope of the audit included a sample of FDIC-supervised banks identified as subprime lenders as of December 2001. In addition, our audit tested adherence to the following subprime related criteria:

• Interagency Guidelines on Subprime Lending (March 1999)
• Expanded Guidance for Evaluating Subprime Lending Programs (January 2001)
• Examination Documentation (ED) Module for Subprime Lending (July 2001)

To accomplish our objective, we reviewed subprime related policies, procedures, laws, and regulations, and identified related DSC and corporate performance goals and results. We also reviewed DSC management and tracking reports and reviewed DSC headquarters and regional office files and field office workpapers for a sample of subprime banks. In addition, we interviewed DSC and DIR headquarters staff, Assistant Regional Directors (ARDs), Case Managers, and Examiners-in-Charge responsible for the supervision of the sample banks. The audit team also attended the Federal Financial Institutions Examination Council (FFIEC) Supervisory Update: Credit Scoring and Subprime Lending training seminar held in June 2002 at the Federal Reserve Bank of Dallas.

The sample selected was based on the universe of all FDIC-insured banks identified by DSC as having subprime loan programs. These institutions have subprime loan portfolios that constitute 25 percent or more of Tier 1 Capital. According to the December 2001 DSC Internal Supervisory Database, there were 128 of these institutions, and the FDIC was the supervisory agency for 60 of these institutions. We focused our review on 11 FDIC-supervised banks rated 3, 4, and 5. For the 11 banks in the sample, we conducted visits to the regional and field offices, reviewed examination workpapers, and conducted interviews with Examiners, Case Managers, and ARDs. (Note: We did not visit the field office or review examination workpapers for one bank in our sample because we decided that a visit to the site would not materially affect our audit results and, thus, would not be cost effective.)

Of the 11 institutions in our sample, their areas of concentration in subprime lending were as follows: 1 was involved in mortgage lending, 2 were involved in credit card lending, 5 were involved in auto lending, 2 were involved in single family residential (SFR) construction lending, and 1 was involved in commercial lending (asset-backed lending). The percentage of subprime loans in the banks' loan portfolios ranged from 15 percent to 90 percent. We also determined that subprime issues were not the cause of the banks' current problems in three cases.

We used the FDIC's Bank Information Tracking System (BITS) to identify examination dates and to obtain examination comments. These data were later verified by our onsite reviews of examination workpapers and reports. We also used the DSC Internal Supervisory Database to identify background information on subprime banks. The information contained in the database is sent to the DSC Risk Management and Application Section by the regional offices on a quarterly basis. We did not conduct a data reliability assessment to verify the accuracy and completeness of this database. However, because the data are regularly reviewed by DSC, their use was reasonable for accomplishing our audit objectives.

DSC's Regional Office Review Program is organized into three categories: Examination and Supervision, Management, and Administration. For the most part, the program is structured in a checklist type format. This format was developed to provide consistency among the reviews of each region while allowing flexibility in addressing region-specific issues. Each regional office is subject to internal review once every 2 years. The results of the reviews are used to: (1) provide feedback to Regional Directors; (2) inform other regions of best practices; (3) serve as tangible feedback as to the effectiveness of the Division’s policies, practices, and procedures; and (4) serve as tests of the various control objectives.

The DSC internal review manual has a section entitled Review Procedures: Subprime Lending and Loan Penetration Review Worksheet. The procedures require the review of: how the regional offices identify banks with subprime programs, preplanning documents related to the subprime lending risk assessment for banks, and how results of examinations are presented in the report of examination. The regional office internal review criteria were revised in May 2002. The revised requirements add more focus to the subprime loan penetration at both the bank and regional office level. For the period March 2001 through May 2002, six regional office reviews were conducted. Based on our review of the Internal Review Report comments, we found that no material findings or recommendations were made. During that time period, only one regional office review had been conducted under the new requirements. The review comments address four of the six required criteria: not addressed were the use of the examination documentation modules (Ed Modules) and the examination report discussion of the subprime loan sample. (Note: In January 2001, the FFIEC issued the Expanded Guidance for Evaluating Subprime Lending Programs. This expanded guidance supplements the guidelines issued in March 1999, refines the definition of subprime lending, clarifies the agencies' expectations regarding an institution's risk management processes, and provides supervisory expectations for examinations of subprime lending programs.) For the five regional office reviews conducted under the 2001 criteria, the reviews generally addressed the review points. One review did not address the use of examination documentation modules, and one did not discuss the pre-examination planning related to the subprime loan portfolio.

There are no specific Corporation goals that address subprime lending; however, DSC's general goals regarding managing bank risks and resolution of problem banks would cover banks with subprime programs. The specific goal "to initiate examinations in accordance with requirements" is important because subprime loan programs are primarily identified through onsite examinations. Also the goal of "following up on problem banks in a timely manner" is relevant to subprime lending because many of the banks involved are on the problem bank list.

The FDIC 2001 Program Performance Report discusses the performance results for the 2001 goals that the FDIC implemented. The report provides detailed analysis and discussion of how the FDIC accomplished each of its performance goals and objectives for 2001. According to the report:

• The target of 100 percent was not met as it relates to the goal of initiating examinations in accordance with requirements. During 2001, the FDIC initiated 2,575 safety and soundness examinations. However, 67 institutions (2.6 percent) did not have onsite examinations initiated in accordance with statutory requirements. Of these 67 examination delinquencies, 11 institutions were due for an examination by the FDIC and 56 were past due by the state authorities under the alternating examination program. Of the 11 institutions due for an examination by the FDIC, 4 of the examinations were postponed because the institutions were scheduled to merge, convert, or relinquish their charters. The other seven examination delinquencies were due to the following reasons:
  
  
  • Two institutions were affected by asset growth or change in capital category, which modified the institutions’ examination intervals.
  • Two institutions were in the process of satisfying the requirements of outstanding enforcement actions and requested postponement of their regular examinations.
  • One institution’s examination was postponed as an accommodation to bank management because of a recent information system conversion.
  • Two institutions were in the process of finalizing financial data (Call Report information) after the completion of a recent merger.

  All seven institutions were scheduled for examinations during 2002; six of these institutions were scheduled for the first quarter of 2002 and one was scheduled for April 2002.

• The target of mailing examination reports to FDIC-supervised banks within 90 days of the examination start date was met. Examination reports were mailed to banks within an average of 54 days of the examination start date.
  
  
• The target of completing follow-up examinations for problem institutions within 12 months of completion of the prior examination was met. No problem institutions were delinquent for an examination under statutory requirements.
  
  
• The target of processing and mailing FDIC-supervised problem institution examination reports to the institution within 45 days of receipt of the report by the Regional Office was met. On average, problem institution examination reports were mailed to the institution within 44 days of receipt of the report by the Regional Office.

Based on our reviews of federal laws and regulations, we did not identify any that exist for subprime lending as defined by the FFIEC's Interagency Guidelines for Subprime Lending. However, during discussions held with DSC management, we found that the federal banking agencies were proposing changes in the reporting requirements for Call Reports and for the Home Mortgage Disclosure Act (HMDA) that would require banks to provide information on activities related to subprime lending. The FFIEC has since withdrawn its proposal to change the Call Report requirements; however, the changes related to HMDA reporting will be implemented on January 1, 2004. Appendix V discusses the proposed changes in detail.

No instances of possible fraud or illegal acts were noted during the audit. DSC pursues or refers to the appropriate legal authority any suspicious activities identified during examinations.

We reviewed the policies and procedures for examinations of banks with subprime lending programs and found that they provide examiners with the guidance needed to assess whether institutions effectively manage the risks associated with subprime lending. Appendix IV discusses the policies and procedures in detail. The same guidance was issued to bank management through financial institution letters that spell out the federal bank regulatory agencies' expectations related to the management of subprime loan programs and the requirements for calculations of capital and allowances for losses. However, our review found that some examiners were not always satisfied with the level of detail in the policies related to capital calculations. During discussions with DSC management, we were informed that the guidance was structured in a manner that would allow it to be used by all of the federal banking regulators during their on-site examinations.

Examiners and case managers expressed concern about the lack of detail in the January 2001 FFIEC Expanded Guidance for Evaluating Subprime Lending Programs (Expanded Guidance) related to calculating capital reserves for subprime loans. (Note: In January 2001, the FFIEC issued the Expanded Guidance for Evaluating Subprime Lending Programs. This expanded guidance supplements the guidelines issued in March 1999, refines the definition of subprime lending, clarifies the agencies' expectations regarding an institution's risk management processes, and provides supervisory expectations for examinations of subprime lending programs.) The interagency policy was written in a broad framework because of the many agencies that must follow it. According to several DSC examiners, the policy is vague in that it does not provide weights or benchmarks to use when reviewing the factors to be considered in calculating appropriate capital levels related to the subprime loan portfolio. Examiners explained that as a result of the lack of this type of language in the guidance, it is difficult to get bank management's agreement on the examiners’ method for calculating the required capital levels. There is also concern that the methodology used to calculate required capital levels during examinations could vary from bank to bank even when the factors being considered are the same or similar.

According to the Expanded Guidance, each subprime lender is responsible for quantifying the amount of capital needed to offset the additional risk in subprime lending activities. The analysis should be tailored to reflect the size, concentration level, and relative risk of the institution's subprime lending activities and should consider the following elements:

• Portfolio growth rates;
• Trends in the level and volatility of expected losses;
• The level of subprime loan losses incurred over one or more economic downturns, if such data/analyses are available;
• The impact of planned underwriting or marketing changes on the credit characteristics of the portfolio, including the relative levels of risk of default, loss in the event of default, and the level of classified assets;
• Any deterioration in the average credit quality over time due to adverse selection or retention;
• The amount, quality, and liquidity of collateral securing the individual loans;
• Any asset, income, or funding source concentrations;
• The degree of concentration of subprime credits;
• The extent to which current capitalization consists of residual assets or other potentially volatile components;
• The degree of legal and/or reputation risk associated with the subprime business line(s) pursued; and
• The amount of capital necessary to support the institution's other risks and activities.

The Expanded Guidance states that because subprime lending poses more risk than standard lending, it is expected that the starting point for determining capital levels would be at a minimum one and a half to three times greater than what is appropriate for non-subprime assets of a similar type. The guidance also states that refinements should be made based on the particular circumstances of each bank. As a result, the capital ratios should be set well above the averages for their traditional peer groups or other similarly situated institutions that are not engaged in subprime lending.

We discussed the matter with DSC Washington and regional level managers who were involved in developing the guidance. They informed us that they are aware that some examiners feel the policy is vague; however, the policy was written in that manner to facilitate its use by all of the banking regulators. Empirical data on FDIC-insured banks was compared to data for unregulated financial companies in developing the policy. The unregulated companies' "set-asides" for these types of loans were two to three times higher than what insured institutions were allotting. Basically, the market was imposing a higher coverage requirement on unregulated financial companies; thus, the Expanded Guidance calls for higher coverage.

During our review, we determined that DSC has provided training to examiners related to the Expanded Guidance. The topic is covered in the DSC Commissioned Examiner Training School, and the FFIEC has published a subprime lending training CD-ROM that can be used by examiners. We also found that during each DSC Regional Conference over the past 2 years, subprime presentations were included as breakout sessions or training modules. In addition, the Expanded Guidance provides the names of subprime lending specialists for each region and Washington, and encourages examiners and case managers to avail themselves of these specialists, when necessary. We understand that because the Expanded Guidance is an interagency effort, it must be sufficiently flexible to be used by each agency involved. We are also satisfied with the efforts that DSC has made to provide examiners with supplemental training and to provide subject matter experts to assist examiners during onsite examinations. The factors used to determine capital requirements for subprime loans must be evaluated separately and then weighed to come up with a final amount. Because each set of circumstances will be different, the weights will vary based on the individual bank. Therefore, although examiners and case managers were concerned about the difficulty of obtaining management agreement on the method used to determine capital requirements, it appears DSC has taken adequate steps to address this concern. Accordingly, we have not made a recommendation regarding this matter.

According to DSC personnel, examinations of subprime lending institutions begin the same as for all safety and soundness examinations, with the examiners risk-scoping the bank’s activities prior to starting onsite examinations. For the onsite portion of the examinations, examiners are required to follow the Expanded Guidance for Evaluating Subprime Lending Programs issued on January 31, 2001. Based on our review of examination workpapers, we determined that examiners are generally complying with the risk-scoping and examination requirements for subprime lenders. As a result, DSC has the capability to identify institutions that may have subprime lending programs prior to onsite examinations and determine the conditions of the subprime loan portfolios during onsite examinations.

The initial identification of subprime banks is usually done by examiners during the pre-examination planning and onsite fieldwork for safety and soundness examinations. In preparation for an examination, examiners are required to perform pre-examination planning to assess the risk and focus the examination. When conducting pre-examination planning, examiners review offsite monitoring data, including Call Reports and Uniform Bank Performance Reports (UBPRs). (Note: According to the American Bankers Association’s Banking & Finance Terminology, an institution’s quarterly Consolidated Report of Condition and Income contains a balance sheet, income statement, and other detailed financial schedules containing information about the institution. According to the Uniform Bank Performance Report Users Guide, the UBPR is designed to be used by bank examiners and bank management evaluating the condition of banks. The report provides the bank's data, data for a peer group of banks similar in size and economic environment, and percentile rankings.) In addition, as noted in Appendix VI, there are several other offsite monitoring applications available to examiners for use when conducting offsite risk analyses.

During our review of examination workpapers for the banks in our sample, we determined that 10 of the 11 banks in our sample had pre-examination planning memorandums that documented the work that needed to be done to identify potential subprime loans. In 2 of the 11 banks in our sample, examiners were alerted by offsite monitoring programs that subprime activity was probably occurring at the banks. In seven instances, examiners identified subprime loans during onsite examinations, and in one instance the bank had been a subprime lender since the mid-1990s. We did not review the examination workpapers for the eleventh bank because we decided that a visit to the site would not materially affect our audit results and, thus, would not be cost effective.

During onsite examinations, examiners are required to follow the Interagency Guidelines on Subprime Lending, issued on March 4, 1999, and the Expanded Guidance for Evaluating Subprime Lending Programs, issued on January 31, 2001. Examiners have the option of using the Subprime Lending ED module (contained in the DSC Regional Director Memorandum #00-004) to document the results of their reviews. We found that only four examiners chose to use the module to document their reviews of subprime lending activities during onsite examinations. However, based on our reviews of examination workpapers for the banks in our sample, we determined that the examiners generally completed the required subprime examination procedures. In all 10 cases, as required by the guidelines:

• Subprime activities were scoped in the Pre-Examination Planning Memorandum.
• Internal control reviews were documented.
• Institutions' strategic/business plans were verified.
• Subprime lending policies and procedures were identified.
• Independent reviews of subprime activities were performed and documented.
• Methodologies for ALLL assessments and calculations reviewed were documented.
• Management effectiveness was assessed by examiners.
• Adequacy of communication and information systems was documented.
• Institutions' efforts to track profitability trends were reviewed and documented.

In 9 of 10 instances:

• Credit risk analysis was documented.
• Transaction testing on subprime portfolios was identified.
• Institutions' methodologies for determining capital to offset the risks taken in subprime activities were assessed.
• The adequacy of servicing and collection departments maintained by the institutions was documented.

In 8 of 10 instances:

• Offsite monitoring reports were used by examiners to risk scope the examination.
• Stress testing for capital adequacy was documented in examiner workpaper files.

Based on discussions with examiners, we found that in 7 of the 11 banks in the sample, examiners discovered the banks' subprime loan programs during onsite examinations. In addition, in six cases the banks were not aware they had subprime loan programs until examiners pointed them out.

• In one instance the bank was affected by the implementation of the Interagency Guidelines on Subprime Lending, which defines "subprime" and sets acceptable criteria for capital and ALLL. As a result of the new policy, the bank's capital levels and ALLL levels were considered inadequate, relative to the risk in its subprime portfolio.
  
  
• In four instances, a single bank employee (i.e., the bank President or a loan officer) originated all or a significant portion of the subprime loans without the knowledge of bank management.

Based on our review of DSC and DIR databases and management reports, the FDIC has an effective program to monitor the condition of FDIC-supervised subprime lenders and to track corrective actions related to subprime activities. In addition to the quarterly monitoring of all subprime lenders (described below), DSC now monitors subprime credit card banks with more intensity.

The DSC Internal Supervisory Database is DSC's internal mechanism for tracking banks engaged in lending activities that inherently pose an increased risk to the institution and thereby to the deposit insurance funds. The regional offices are required to report quarterly on institutions with subprime loan programs, institutions that participate in high loan-to-value (HLTV) lending, institutions with payday lending activities, and institutions heavily involved in securitized transactions. (Note: The DSC Internal Supervisory Database report instructions state that HLTV loans are typically junior liens on owner-occupied single-family residences, but there is limited collateral protection because the combined loan amounts often exceed the value of the home. The DSC Internal Supervisory Database report instructions state that securitized institutions are those which, as a significant part of their business, make or purchase loans for sale or securitization and have those loans serviced and held off-premises by third parties or affiliates. The DSC Capital Markets Examination Handbook defines securitization as the process where interests in loans, generally mortgages, and other receivables, including credit cards and automobile loans, are packaged, underwritten, and sold in the form of asset-backed securities. One of the benefits of the securitization process is that it converts relatively illiquid assets (loans) into readily marketable securities with reasonably predictable cash flows.) These data are included in the DSC Internal Supervisory Database as four separate reports.

Interagency procedures require examiners to identify institutions in which subprime loans constitute 25 percent or more of Tier 1 Capital. The DSC Internal Supervisory Database reports are used to track the data related to the subprime portfolios collected during examinations. Examiners are required to complete a Subprime Data Collection Form, which includes 31 different data fields. In addition, they must prepare written comments that:

• identify the risk level and trends in lending activity;
• provide an assessment of management, including loan underwriting and administration practices, and of the internal controls regarding these activities; and
• identify any notable trends in delinquencies, net losses, or examination loan classifications.

The regional offices forward the data to Washington where it is downloaded into a database and used to develop the DSC Internal Supervisory Database reports.

The Offsite Risk Evaluations of Subprime Lenders Report is another tool that DSC uses to monitor FDIC-insured subprime lenders. The report includes a combination of results from several offsite monitoring applications to track and assess the condition of subprime lenders between examinations. The information is included in the DSC Internal Supervisory Database reports. If a bank fails three or more of seven monitoring tests, it is flagged in the report. The monitoring tests and criteria used to develop the report are enumerated in Table 2 below. Appendix VI describes the offsite monitoring systems used to develop the Offsite Risk Evaluation Reports in more detail.

In an attempt to identify additional sources of information related to subprime lending programs at FDIC-insured institutions, the regulators proposed changes to Call Report and Home Mortgage Disclosure Act (HMDA) reporting requirements. On May 31, 2000, the federal banking agencies published a notice for comment regarding the inclusion of additional reporting items in the Call Reports for subprime lending activities. However, on December 9, 2002, the FFIEC voted to drop the proposal due to responses from the industry and Congress. As a result, the regulators will have to rely more heavily on the onsite reviews conducted during safety and soundness examinations. In addition, on February 15, 2002, the Federal Reserve Board issued a final rule amending the reporting requirements for HMDA that would make more lenders subject to the HMDA reporting requirements. The effective date for implementation of the changes is January 1, 2004. Appendix V discusses these issues in detail.

During our audit, we obtained copies of corrective actions related to the sample banks' subprime lending activities from the regional office correspondence files. (Note: The DSC Formal and Informal Actions Procedure Manual states that the FDIC generally initiates formal or informal corrective action against institutions with a composite safety and soundness or compliance rating of "3", "4", or "5", unless specific circumstances warrant otherwise. Formal action is generally initiated against an institution with a composite rating of "4" or "5" for safety and soundness or compliance if there is evidence of unsafe or unsound practices and/or conditions, or concern over a high volume or severity of violations at the institution. The FDIC may initiate informal action when a financial institution is found to be in a marginally unsatisfactory condition or when action is needed to address specific concerns. Informal action is generally appropriate for institutions that receive a composite rating of "3" for safety and soundness. This rating indicates the institution has weaknesses that, if left uncorrected, could cause the institution’s condition to deteriorate. The FDIC may issue formal actions pursuant to Section 8 of the Federal Deposit Insurance (FDI) Act (12 U.S.C. section 1818). These include termination of federal deposit insurance; cease-and-desist action; removal, prohibition, and suspension actions; and civil money penalties. Additionally, Section 38 of the FDI Act (12 U.S.C. section 1831o) authorizes the FDIC to issue prompt corrective action directives to undercapitalized institutions. Informal actions such as Bank Board Resolutions and Memorandums of Understanding are voluntary commitments made by the board of directors of a financial institution. They are neither publicly disclosed nor legally enforceable.) The effective dates of the corrective actions ranged from March 2000 through July 2002. In all cases, a Bank Board Resolution (BBR), Memorandum of Understanding (MOU), Cease and Desist Order (C&D), or Prompt Corrective Action Capital Plan was put in place, or an existing action was modified to address the subprime lending issues. Where applicable, the actions required bank management to take steps to improve risk management programs and to increase capital and ALLL. For the sample, there were six C&Ds, four MOUs, and one BBR in place. In addition, in all cases where corrective actions were implemented, DSC required progress reports, and/or conducted visitations to monitor the status of the bank.

In 9 of the 11 institutions in the sample, examiners noted improvements since the corrective actions were put in place. In two instances, examiners told us that the banks were not improving and were considered failure candidates; however, for one of the banks, the potential failure was not the result of the subprime loan portfolio.

We tracked the time between the start of the examination that identified the subprime-related problem and the implementation of corrective actions. We found that the time frames ranged from 2 to 10 months. Three of the 11 banks had time frames that exceeded 6 months (8, 9, and 10 months). For these three banks, we found that the FDIC and state banking regulators took steps to increase offsite monitoring or initiated onsite visits to mitigate any additional risk in the interim period between the examinations and implementation of corrective action.

In addition, 9 of the 11 institutions have since ceased subprime lending activity, either because of the corrective actions in place or because management was not aware that they were making subprime loans and was not willing or capable to service and collect on the loans. In six of the eight instances, the institutions are seeking to be either sold, acquired, or merged.

The FFIEC Expanded Guidance may not provide enough detail to guide examiners through reviews of the custom credit scoring models used by banks to predict the creditworthiness of borrowers. Although the use of custom models is not common among FDIC-supervised banks, those institutions that use them run the risk of making credit underwriting decisions based on results that may not be predictive of the borrowers' actual creditworthiness.

In May 1997 DSC developed the Credit Card Specialty Bank Examination Guidelines Manual to provide examiners with examination techniques for those institutions involved primarily in credit card operations. A credit card specialty bank is defined as any bank whose total loans exceed 50 percent of total managed assets and whose credit card loans exceed 50 percent of total loans and credit card receivables securitized and sold. The section of the manual that specifically addresses custom credit scoring systems provides definitions of terms related to credit scoring systems, explains the characteristics of the different types of scoring models, and provides a summary of examination goals. The summary of examination goals includes the following:

• Identify the types of scoring systems used by management.
  
  
• Determine how management uses the scores in making credit decisions.
  
  
• Examine credit cutoff points and odds charts to assess the level of risk that management is taking.
  
  
• Verify periodic validation and calibration procedures, review the results of the procedures, and assess the strength of the credit scoring system.
  
  
• Evaluate whether credit bureau, behavioral, and other scores enhance account management programs.
  
  
• Determine whether collection scores are used and assess their contribution to the collection process.
  
  
• Review the number of overrides. (Note: The DSC Credit Card Specialty Bank Examination Guidelines Manual defines an override as a decision made by bank management on an applicant that is contrary to the decision made by the credit scoring system.)
  
  
• Verify whether management reviews override reports and whether their performance is adequately tracked.
  
  
• Determine whether overrides have a material adverse effect on asset quality.

As noted above, these procedures were specifically designed for examinations of credit card specialty banks and were not specifically geared towards examinations of institutions with subprime loan programs.

The more current requirements for reviews of custom credit scoring models are contained in the DSC Regional Directors Memorandum (RD Memo) 00-004, dated January 24, 2000, entitled Subprime Lending Examination Procedures. The RD Memo contains the Examination Documentation Module (ED Module) developed for subprime lending that may be used at examinations of institutions that conduct subprime lending activities. (Note: The ED Modules are examination tools that focus on risk management practices and guide examiners to establish the appropriate examination scope. The modules incorporate questions and points of consideration into examination procedures to specifically address a bank's risk management strategies for each of its major business activities. The modules direct examiners to consider areas of potential risk and associated risk control practices, thereby facilitating a more effective supervisory program. The modules have been incorporated into a software package called Examination Documentation to provide an easy workpaper creation tool.) The use of ED Modules during safety and soundness examinations is optional. According to the ED Module, during onsite subprime lender examinations, examiners should:

• Determine whether models are custom or vendor-supplied and what they are being used for (e.g., approvals, pricing, loss estimates, collections).
  
  
• Determine whether the custom model development process is consistent with the bank’s risk appetite and desired level of subprime loans.
  
  
• Review the documentation supporting the institution’s scoring models and determine whether:
  
  
  • Models are empirically derived and statistically sound.
  • Scoring is based upon a developmental population that captures the behavioral and credit characteristics of the subprime population targeted for the products offered.
  • Assumptions and customer characteristics are reviewed frequently and updated to effectively predict credit performance.
  • Credit scores permit the institution to predict overall risk and the potential impact on collection activities.
  • Models consider delinquencies and losses over varying economic conditions.
  
  
  
• Evaluate the ongoing monitoring and maintenance process in view of the portfolio’s performance and determine whether:
  
  
  • Models are actively monitored, adjusted, and periodically revalidated.
  • Systems continue to reflect current underwriting standards and risk parameters.
  • Management maintains an adequate portfolio chronology log to record significant events related to the credit acquisition process for each subprime portfolio.
  • Scoring systems are supervised and maintained in accordance with vendor-provided specifications and recommendations.
  • Models are adjusted to account for unexpected events.

While each of the 12 elements of criteria listed above is important, guidance is not provided on how models should be monitored, adjusted, or validated. These steps are critical to ensuring the model is reasonably reliable and could include, for example, comparing the results of institutional credit scoring to other indices such as the Fair, Isaac and Company (FICO) credit bureau scores in order to assess the validity of the model. (Note: FICO is the Fair, Isaac and Company credit score developed to predict the creditworthiness of credit applicants using a statistical model. According to the Fair, Isaac and Company booklet entitled Understanding Your Credit Score, credit scoring estimates the repayment probability based on the information in the credit application and a credit bureau report.) Another example is that the guidance requires examiners to determine whether credit scoring is based upon a developmental population that captures the behavioral and credit characteristics of the subprime population targeted for the products offered; however, it does not provide instructions as to how to determine what the appropriate behavioral and credit characteristics should be for the borrowers targeted.

DSC management informed us that FDIC-supervised banks do not often use custom credit scoring models. According to DSC, although examiners are encountering more banks with custom models, banks usually rely on the FICO scores provided by credit bureaus or on a behavioral model that is similar to FICO. During the audit, we determined that only 1 of the 11 sample banks used a custom credit scoring model. The remaining banks in the sample relied on the FICO credit rating scores provided by the credit bureaus.

For the one bank in our sample that used a custom credit scoring model, we reviewed the workpapers and the examination report to determine the extent of examination procedures applied. (Note: We did not discuss the review of the custom credit scoring model with the Examiner-in-Charge.) While we found documentation in the workpapers that DSC reviewed the bank's custom scoring model, we found that only 2 of the 12 criterion required by the Subprime Lending ED Module were discussed. Those criterion related to determining whether the model was custom made or vendor-supplied and what it was being used for, and whether scoring was based upon a developmental population that captured the behavioral and credit characteristics of the subprime population targeted for the products offered.

According to the FDIC Third Quarter 2002 Regional Outlook, the stress in the consumer sector could particularly challenge subprime and concentrated consumer lenders. Some of these lenders have not experienced subprime lending through a full economic cycle. Without constant monitoring and testing, banks run the risk of relying on models that project results that are out of line with the actual performance of borrowers. In addition, if custom credit scoring models do not identify the additional default risk associated with subprime borrowers, banks cannot ensure that they are pricing the subprime loan products appropriately. Consequently, the custom credit scoring models used by banks should be thoroughly evaluated to ensure that the results they provide are useful in assessing the risk presented by subprime borrowers. Without adequate examination coverage of these models, the FDIC cannot ensure that those institutions that use custom credit scoring models are making sound credit underwriting decisions related to their subprime loan portfolios and properly determining loan loss allowances and capital levels.

The Expanded Guidance provides the names of subprime lending specialists for each region and Washington, and encourages examiners and case managers to avail themselves of these specialists, when necessary. To strengthen examination coverage of custom credit scoring models, DSC could advise examiners to obtain concurrence from an appropriate subprime lending specialist on the scope and results of examination procedures for the assessment of custom credit scoring models used by banks for subprime loan programs.

In response to industry and consumer queries, the federal bank regulators have defined characteristics that are attributable to subprime borrowers and lenders and have identified the concerns that center on subprime credit card banks.

"Subprime lenders" are institutions that systematically target the subprime market through lending programs that employ tailored marketing, underwriting standards, and risk selection.

The term "subprime program" refers to the process of acquiring, on a regular or targeted basis either through origination or purchase, subprime loans to be held in the institution’s portfolio or accumulated and packaged for sale. Subprime lending programs may also target borrowers with questionable repayment capacity evidenced by low credit scores or high debt-burden ratios. Subprime lending does not refer to individual subprime loans originated and managed in the ordinary course of business as exceptions to prime risk selection standards.

The federal banking agencies have defined "subprime borrowers" using the characteristics that are most often associated with these borrowers. The term "subprime" refers to borrowers that typically have weakened credit histories that include payment delinquencies, previous charge-offs, judgments, bankruptcies, foreclosures, repossessions, high default probability, and poor debt-to-service ratios. While some of these attributes are not always indicative of a subprime borrower, taken as a whole they generally represent potential troubled or problem borrowers.

The federal banking agencies have also taken measures to ensure that financial institutions address the increased risks associated with subprime lending through the issuance of guidance and policy statements. The following procedures and guidelines have been issued.

In March 1999, the FFIEC issued Interagency Guidelines on Subprime Lending that defined subprime lending as extending credit to borrowers who exhibit characteristics indicating a significantly higher risk of default than traditional bank lending customers. Regulators were required to review and evaluate the capital levels at examinations and through offsite monitoring techniques. In addition, the guidance states that measures can be implemented and enforced if the capital levels are deemed to be inadequate. Specifically, the guidance stresses the need for banks' risk management programs to address loan pricing and requires the following for risk management programs:

• Planning and strategy should be consistent with overall business strategy of the bank;
  
  
• Staff expertise requires specialized skills and knowledge;
  
  
• Lending policy should establish the framework for pricing decisions and profitability analysis; and
  
  
• Management should conduct reviews of credit scoring, pricing, and the adequacy of the ALLL models.

In January 2001, the FFIEC issued Expanded Guidance for Evaluating Subprime Lending Programs. This Expanded Guidance supplements the guidelines issued in March 1999 and is specifically tailored to institutions that have subprime lending programs with an aggregate credit exposure greater than or equal to 25 percent of Tier 1 Capital. In addition, the Expanded Guidance refines the definition of subprime lending, clarifies the agencies' expectations regarding an institution's risk management processes, and provides a more detailed discussion of the supervisory expectations for examinations of subprime lending programs. The guidance also provides additional risk factors and examination procedures to consider in the following areas:

• Allowance for Loan and Lease Losses (ALLL);
• Capital Adequacy;
• Portfolio Review and Analysis;
• Classification Guidelines;
• Cure Programs; and
• Predatory or Abusive Lending Practices.

To address risk management expectations, the guidance states that management’s ability should be judged by the quality of the risk management and control processes in place, and the extent to which management is adhering to those processes. When a primary supervisor determines that an institution’s risk management practices are materially deficient, the supervisor may instruct the institution to discontinue its subprime lending programs.

For ALLL, the guidance states that classified loans are considered loans that are not adequately protected by the sound worth and repayment capacity of the borrower or the pledged collateral. Full liquidation of the debt may be in jeopardy. The ALLL covering subprime loans that are not classified should be sufficient to absorb estimated losses on outstanding balances over the current operating cycle, which is typically 12 months. The board and management are responsible for determining the adequacy of the ALLL and documenting the methodology that determines the balance of the ALLL.

For capital, the policy states that each subprime lender is responsible for quantifying the amount of capital needed to offset the additional risk in subprime lending activities. Such lenders are also responsible for documenting the methodology and analysis supporting the specified level of capital. The methodology should be tailored to reflect the size, concentration, and risk posed to the institution by the subprime lending activities. The guidance lists several potential factors to be considered when determining the appropriate amount of capital and states that since subprime lending possesses more risk than standard lending, it is expected that the capital levels would be at a minimum one and a half to three times greater than what is appropriate for non prime assets of a similar type. The capital adequacy analysis should also include stress testing as a tool for estimating unexpected losses in subprime lending pools. Shock tests of basic assumptions will assist in determining a portfolio’s susceptibility to changes in market and business conditions.

This memorandum contains the ED module developed for subprime lending to be used at examinations of institutions that conduct subprime lending activities. Topics in the ED module include the following items:

• Preliminary review, policy considerations, internal controls, audit or independent reviews, information and communication systems;
  
  
• Portfolio analysis, allowance for loan and lease losses, servicing and collections, scoring models, profitability, and capitalization; and
  
  
• Third parties, securitizations, managerial effectiveness, transaction-level testing, portfolio quality, and stress testing of capital adequacy.

Under each section, detail is provided as to the type of work that should be performed in order to achieve a level of confidence that the bank is either administering the program effectively or not.

This guide was issued in May 1997 to aid DSC examiners in the proper evaluation techniques for institutions with credit card portfolios. Credit card specialty banks have unique characteristics. According to DSC, this guide is to be used in conjunction with the guidance pertaining to subprime lending.

On January 8, 2003, the FFIEC issued interagency guidance for account management and loss allowances for credit card lending. According to the guidelines, recent examinations of institutions engaging in credit card lending have disclosed a wide variety of account management, risk management, and loss allowance practices, a number of which were deemed inappropriate. As of March 31, 2002, there were 5 FDIC-supervised institutions that were heavily involved in subprime credit card receivables.

Federal regulators issued the guidelines to more tightly monitor specialty credit card lenders, especially those that have increased their subprime lending business. The rules require companies to more carefully monitor how much credit to extend to customers with too much debt, be more consistent in the way they declare loans worthless, and maintain stronger reserves against bad loans and customer fees. The guidelines are aimed at the large specialty credit card companies whose business is issuing credit and who can incur significant losses when customers default.

The guidance applies to all institutions under the member agencies' supervision that offer credit card programs. It describes the expectations for prudent risk management practices for credit card activities, particularly with regard to credit line management, over-limit practices, and workout and forbearance practices. The guidance also specifically addresses income recognition and loss allowance practices for credit card lending and states that recent examinations of credit card lenders have revealed a variety of income recognition and loss allowance practices. Such practices have resulted in inconsistent estimates of incurred losses and, accordingly, the inconsistent reporting of loss allowances. To address the account management, risk management, and loss allowance practices for credit card lending, the guidance provides procedures for:

• Credit Line Management;
• Over-limit Practices;
• Minimum Payment and Negative Amortization;
• Workout and Forbearance Practices;
• Income Recognition and Loss Allowance Practices; and
• Policy Expectations.

There are two ED modules that address credit card activities. The first one addresses merchant credit card lending. It provides the core and expanded analysis procedures to be conducted in the course of an examination. The impact analysis is also included in the module. The second module addresses consumer credit card activities and is to be applied to institutions that specialize in credit cards and to those institutions where credit card receivables are not the primary business focus. Both sections focus on the examination procedures outlined in the credit card specialty examination guidelines.

The analyses prescribed in the ED modules apply to merchant credit card activities and credit card activities in general. Specific details concerning all aspects of credit card lending are contained in the supplemental guidelines for credit card specialty examinations.

There are no federal laws, rules, or regulations concerning subprime lending, defined by the FFIEC Interagency Guidelines as follows: "extending credit to borrowers who exhibit characteristics indicating a significantly higher risk of default than traditional bank lending customers." However, the agencies proposed the following reporting requirements that would require banks to provide information on activities related to subprime lending.

On May 31, 2000, the federal banking agencies published a notice for comment in the Federal Register regarding the inclusion of additional reporting items in the Call Reports for subprime lending activities. (Note: The notice was published on Wednesday May 31, 2000, in Federal Register Volume 65, pages 34801 - 34819.) One of the major stumbling blocks for those who provided comments was the failure of the agencies to provide a definition for "subprime" upon which they could all agree.

On July 12, 2002, the agencies published in the Federal Register a Proposed Agency Information Collection Activities; Comment Request. The public comment deadline was September 10, 2002. Under the proposal, all banks that have subprime consumer loan programs would have to report the total amount of consumer loans in those programs. Those banks with significant programs, defined as programs that make up at least 25 percent of their Tier 1 Capital, would have to supply more detailed data on their subprime loans. Draft Call Report Schedule RC-XX, entitled Consumer Loans in Subprime Lending Programs, would require banks to report the following information related to their consumer loans in subprime lending programs:

• total amount of consumer loans;
• outstanding balances of consumer loans;
• past due and nonaccrual consumer loans; and
• charge-offs and recoveries of consumer loans.

However, the industry was concerned because there are no standard industry-wide approaches to defining subprime and subprime programs. According to the proposal, the information would enable the agencies to better plan their examinations of banking institutions and to monitor offsite the extent of, changes in, and performance of subprime lending programs of banking institutions.

On January 7, 2003, the FFIEC announced that they had dropped the proposal to require that subprime lending data be included in Call Reports. As a result, the regulators will have to rely more on the subprime lending reviews conducted during onsite examinations.

The Federal Reserve Board issued a final rule amending the reporting requirements under Regulation C, which implements HMDA. (Note: The final rule was published on Friday, February 15, 2002, in Federal Register Volume 67, pages 7222 - 7251.) The changes were intended to make more lenders subject to the HMDA reporting requirements and to obtain better data for the federal banking agencies to use to combat predatory lending. Under the current regulations, non-depository lenders are subject to HMDA based on the percentage of mortgage loans they originate. The Federal Reserve anticipates that this additional requirement will capture 98 percent of first lien subprime mortgages and 95 percent of second lien subprime mortgages.

The requirements in the final rule should provide additional assistance to the federal regulators in determining which financial institutions as well as non-depository lenders are engaged in subprime lending activities. It should also provide information pertaining to the extent of the subprime lending activities undertaken by these lenders. On May 2, 2002, the Federal Reserve released a press notice that extends the effective date for the implementation of the changes to Regulation C to January 1, 2004.

APPENDIX VI

The Offsite Risk Evaluations of Subprime Lenders Report uses a combination of several offsite monitoring programs to flag banks possibly involved in subprime lending activities. The following are descriptions of those programs:

The Growth Monitoring System (GMS) is an offsite rating tool that identifies institutions that have grown rapidly and/or have a funding structure highly dependent on non-core funding sources. GMS focuses on the relationship between loan growth and non-core funding sources. Rapid growth is an indicator of institutions experiencing significant changes in their financial statements. GMS focuses on identifying institutions that, due to a rapid change, may require increased monitoring. GMS calculates the growth percentile ranking for all institutions, including thrifts.

The Statistical CAMELS Offsite Rating (SCOR) uses Call Report data to identify institutions likely to receive a CAMELS downgrade at the next examination. SCOR uses statistical techniques to estimate the relationship between Call Report data and examination results. The system uses the correlation of financial data and examination data to rate all CAEL components (capital, assets, equity, and liquidity), including Sensitivity to Market Risk and Management. SCOR rates banks on a scale of 1 to 5 and produces a probability of downgrade percentage for each component and composite rating. The ratings are based on 13 independent ratios.

The Real Estate Stress Test (REST) is used to measure exposure to real estate risk. DIR has been conducting stress testing of subprime and other consumer lenders to evaluate their vulnerability to higher credit losses and slower loan growth. These tests are designed not so much to identify specific problem institutions but to develop an upper-bound estimate of FDIC insurance losses over the coming year. The model was developed based on an analysis of the real estate crisis in New England in the early 1990s, to determine what exposures led to the downturn. The model attempts to simulate what would happen to banks today if they encountered a real estate crisis similar to that of New England. The REST model uses Call Report data to forecast a financial institution's condition over a 3- to 5-year period and scores the institutions on the CAMELS scale of 1 to 5.

The Risk-Related Premium System (RRPS) assigns each insured institution to one of three capital groups and to one of three supervisory subgroups for the purposes of determining an assessment risk classification. Call Report data and Thrift Financial Reports are used. The model identifies institutions with atypically high-risk profiles among those in the best-rated category, to determine whether there are unresolved supervisory concerns regarding the risk-management practices of these institutions.

The Stress Analysis Model (SAM) is used to gauge risk associated with subprime lending practices. DIR conducts a stress test to flag institutions that, in the absence of abnormally high loan returns, would be unable to survive. The Mild Stress Scenario, which is used for subprime institutions, is more responsive to nonperforming loans and is likely to flag institutions with high nonperforming loans. This scenario assumes that an institution will experience a reduction in the return on loans when the rate of return on assets exceeds the prime interest rate by a specific number of percentage points.

Federal Deposit Insurance Corporation
550 17th St. NW Washington, DC 20429
Division of Supervision and Consumer Protection

DATE: March 10, 2003

MEMORANDUM TO: Stephen M. Beard, Deputy Assistant Inspector General for Audits

FROM: Michael J. Zamorski [Electronically produced version; original signed by Michael J. Zamorski], Director, Division of Supervision and Consumer Protection

SUBJECT: Draft Report Entitled The Division of Supervision and Consumer Protection's Examination Assessment of Subprime Lending (Assignment No. 2002-806)

Thank you for the opportunity to review and respond to the Office of Inspector General’s (OIG) draft report entitled The Division of Supervision and Consumer Protection's Examination Assessment of Subprime Lending. The draft report indicates that the Division of Supervision and Consumer Protection (DSC) has taken reasonable steps to ensure that institutions manage risks associated with subprime lending programs effectively, establish adequate allowance levels to cover losses, and maintain capital levels that reflect the additional inherent risks associated with subprime lending. As a result of these findings, the OIG has made no formal recommendations regarding the assessment of subprime lending.

As noted in various interviews with OIG staff, the FDIC has been a leader in analyzing subprime lending activities and in developing interagency guidelines to address risk management practices for insured depository institutions with subprime lending programs. Through the FFIEC, the FDIC continues to participate in additional substantial initiatives to address the risks associated with subprime lending activities. For example, on January 8, 2003, the Federal banking agencies issued guidance regarding Account Management and Loss Allowance Practices for Credit Card Lending Activities. The guidance addresses, in part, risk management practices for subprime credit card lenders. The FDIC also continues with various training and outreach efforts, including the subprime lending presentations for the FDIC-sponsored Commissioned Examiner Seminar. As discussed with OIG staff, the FDIC also plans to offer additional training on the topic of custom credit scoring for a select group of specialists that may require such expertise.

Thank you again for the opportunity to comment on the OIG’s draft report on subprime lending. We appreciate the OIG’s recognition of our efforts in this area and look forward to working with you again in the future.