To Ensure Fair and Equal Treatment

Transcription

1 Self-Testing To Ensure Fair and Equal Treatment Self-testing is a voluntary undertaking designed to ensure compliance and manage legal and business risk. Self-testing offers the compliance manager a window into the marketing and sales practices and service encountered by potential and established customers of a financial institution. It produces information that is not available or produced naturally by the bank s business processes. Information acquired through self-testing can help your bank ensure compliance with regulatory laws and guidelines. It can help the financial institution detect and limit legal, business, and reputation risk and address issues that can improve the bank s ability to succeed in an increasingly diverse and competitive marketplace. by Paul C. Lubin and Bob Homeyer Monitoring the marketing and sales practices and quality of assistance encountered by consumers is becoming even more important as banks expand their product offerings into varied product lines and leverage multiple sales channels, personnel, and third parties to market their products and services. The expansion into new and different product lines and delivery mechanisms can foster increased risk in an environment without adequate oversight. Self-testing, which limits the risk associated with these new activities, can help a financial institution safely expand into new product lines and sales channels. The two most popular and accepted methods of self-testing are preapplication inquiries or mystery shops and post-application customer interviews. Self-testing programs can either proactively seek to uncover (and correct) deficiencies before they result in complaints or allegations of violations of the law (e.g., discrimination or unfair sales prac- tices) or respond to an order or decree after an investigation as a follow-up mechanism. The most effective approach is a preemptive or a proactive strategy to uncover (and correct) any deficiencies in marketing, sales, and service processes before they result in customer complaints, allegations of discrimination, and referrals to regulatory authorities. Advantages Offered Monitoring sales and service practices provides a financial institution with ABA Bank Compliance March/April

2 valuable information. It can help the institution ensure that it is in compliance with the law and regulatory guidelines. A self-test designed specifically to measure adherence to the law (e.g., the Fair Housing Act or Equal Credit Opportunity Act) will also indirectly provide valuable information about the sales and service process and suggest areas where the financial institution can improve. By regularly monitoring the experiences encountered by consumers, the financial institution can detect and then resolve issues that may represent violations of the law before they result in complaints and allegations that can have a negative impact on reputation and sales. A plan that systematically tests a bank s sales and service practices will be viewed positively by third parties, regulators, and enforcement agencies. It is a proactive step to help prevent violations of the law and regulatory guidelines and to ensure that customers are treated fairly and honestly. Hence it has been viewed as a mitigating factor by regulators and enforcement agencies when reviewing issues uncovered during an examination or investigation. The information generated through selftesting designed specifically to measure adherence to ECOA and the FHA is privileged under ECOA and the FHA. To encourage self-testing, Congress in 1996 created a legal privilege for data gathered on a voluntary basis to assess compliance with ECOA and the FHA. The regulators defined self-testing as voluntary activities carried out by a third party that collect information assessing compliance that is not readily available or collected in loan files, in applicant records, or through normal, everyday business practices. The purpose was to encourage financial institutions to use more creative types of activities such as mystery shopping, post-applications, and customer feedback to help detect discriminatory practices and compliance with ECOA and the FHA that cannot be readily identified through file reviews and on-site inspections by field auditors. Civil rights groups, community activists, government regulators, and enforcement agencies regularly use mystery shopping and post-application surveys to help detect violations of the law and acts against public policy. An example of this is the recent HUD paired testing program assessing the treatment of minorities and nonminorities in the preapplication stage of the loan process. In addition, HUD provides Fair Housing Initiatives Program funds to community groups to test for discrimination in lending. And let s not forget about the activist organizations, news organizations, and class-action attorneys who use mystery shopping and other methods to test the sales practices of financial institutions. Defining the Focus of the Self-Testing Program Self-testing programs have a specific goal. This is especially important for fair lending testing, where ECOA and the FHA provide a legal privilege that protects the information collected from self-tests that were specifically designed to measure compliance with the two acts. A broad diagnostic self-testing program will test for adherence to the law and regulatory guidelines across the organization. It will identify patterns and practices that may represent violations of the law, such as discrimination or viola- tions of Regulation B and ECOA and the FHA. Such a program will test various product lines and delivery channels, such as mortgage loans originated through the bank, mortgage company, and broker network or consumer and home equity loans sold through bank branches, the telephone unit, or the World Wide Web. These tests should be conducted yearly to adequately measure new product lines and delivery channels and to identify changes in marketing, sales, and compensation practices, policies, and personnel. A self-test can also be designed to specifically detect patterns indicative of misleading sales and service practices and violations of Section 5 of the Federal Trade Commission Act. It can help detect patterns indicative of predatory lending. Because of the in-depth nature of the self-testing program, the detailed information it produces can be leveraged for business purposes that is, improving sales and service to foster account growth and customer retention. Another approach can be referred to as temperature check or evaluation. This approach is much smaller in scale and cannot be used to determine whether patterns or practices exist across the organization. Instead, a smaller but more concentrated battery of tests are conducted in selected major markets and delivery channels. For example a multistate bank can examine its preapplication loan process in three to five major markets deemed higher risk. In a self-test designed to measure adherence to ECOA and the FHA, a financial institution can select major markets and markets with high proportions of protected classes. Testing would be conducted under the premise that if a problem or major area of risk 4 March/April 2003 ABA Bank Compliance

3 exists, the issue will manifest itself in these markets. The information gathered in these markets can then be applied to other markets. Still another approach is to target markets and products where HMDA data, account growth, withdrawn applications, and statistical modeling indicate a greater source of risk, e.g., substantially higher minority denial rates, disproportionately fewer minority loan applications, higher minority application withdrawal rates, or a disproportionate number of minorities and older customers carrying credit insurance. Such an approach seeks to determine whether the cause is due to disparate treatment. Disparate treatment can include prescreening or discouraging minorities from applying, steering protected classes toward certain products without regard to their suitability, or providing minorities with a different level of assistance and coaching. Regardless of the approach the financial institution takes, the self-test will detect risk at the branch and employee levels. Such information should be used to determine whether violations of fair lending laws and disparate treatment occurred. How to Test for Fair Lending Preapplication Testing Preapplication inquiries involve using mystery shoppers (or, as we call them in self-testing programs, testers ) posing as potential customers or loan applicants and inquiring about the available products. Testers enact scenarios designed to mirror the approach of the average customer so as not to arouse Examples of Findings in Self-Testing Programs Preapplication Testing few or no overt differences; more subtle differences: different information, different questioning, or different products discussed; more favorable treatment of minorities: emphasizing market share in emerging customer segments, and Increased sensitivity to minorities. the use of credit scoring in the sales process at times resulting in differential treatment; gender differences, e.g., Hispanic female versus Hispanic male, African American female versus African American male, etc.; insufficient staffing to handle Spanish-language calls; telephone based service differences, minority versus nonminority; differences in: approval time quoted, closing costs quoted, and/or products discussed; poor or no explanation of points and APR; problematic markets; and inconsistency in small business loan information based on dedicated loan officers versus bankers. Post-application Testing at times a greater willingness to help nonminorities; nonminorities more likely to say the loan officer offered a positive opinion on likely approval; older customers report receiving less information and help than younger customers; confusion as to the role of the mortgage broker and his or her relationship to the lender between minorities and nonminorities; at times minorities more likely to say delays cost us money ; minorities more likely to report being contacted by the lender to ascertain interest in the lender s products; minorities more likely to say the lender did not return my calls ; and nonminorities more likely to report the loan officer offered a recommendation. ABA Bank Compliance March/April

4 Self-Testing for Community Banks Why self-test: to protect reputation; to keep community focus; to ensure third-party adherence to protocols and laws and regulations; laws/regulations apply; ECOA, the FHA, the FTC Act, and TILA; the Gramm-Leach-Bliley Act; interagency guidelines; and nondeposit investments. What to examine: sales and service practices; treatment of protected and nonprotected classes; the degree to which customers feel welcome and invited to do business with the financial institution; needs discovery and suitability of products offered; access to service and information; courtesy and comfort in dealing with protected and nonprotected classes of consumers; office and employee performance. Self-testing methods: matched pair testing; mystery shopping; post-application assistance audits; focus groups; complaint monitoring; and file reviews. What to do with the results: keep the results privileged; work with legal counsel; communicate results with senior management that can act on the results ensure accountability by developing and then monitoring the results of an action plan. Remedies that work: Proactively use self-testing to detect and remedy problems on an ongoing basis. Include fair and equal treatment in the company s mission statement. Establish tools and protocols to ensure consistency in sales and service process. Train to build awareness of the law and instill knowledge that fair treatment builds customer loyalty and sales. Integrate fair treatment and fair lending into ongoing employee training. suspicion, and to create an environment in which the representatives are providing their normal day-to-day service. Depending on your primary area of interest and business, you can test for discrimination or violation of the law (ECOA, the FHA, or the Truth in Lending Act), regulatory guidelines (e.g., nondeposit products), as well as misleading sales practices (Federal Trade Commision Act). The inquiries can cover mortgage products, home equity products, refinance loans, unsecured or consumer loans, small business loans, auto loans, nondeposit investment products, insurance, deposit products, or a combination of these. Test Design and Development Differential treatment based on race is the most common factor that is tested. However, many financial institutions are also conducting programs designed to uncover bias by age (e.g., an older applicant looking for a first-time homebuyers loan) or gender (male versus female applicant seeking a small business loan or a married versus single potential loan applicant). To test for racial discrimination, matched pair, triads, sandwich, or even quad test designs are set up, where the testers are matched in every way except one either race, gender, or age, depending on the focus of the testing. theory behind this approach is that the salient differences present will be due to this sole test factor. In the case of testing for discrimination, you should match the testers by age (within five years), price of home and/or amount of loan desired, stage in the lending process (just starting to look or making an offer) debt-to-income ratios, income (within $5,000), available 6 March/April 2003 ABA Bank Compliance

5 down payment, etc. In a matched pair a nonminority is matched against a minority. In triads and quads a number of specific minority testers can be matched against the nonminority, such as African-Americans, Hispanics, or Asians. In a sandwich test one member of the test group (e.g., the minority) is sandwiched between two members of the control group (e.g., the nonminority). You should review the demographics of your current and potential customer bases to determine the depth of testing necessary and areas where you may be at risk. If the sole focus is to test for misleading sales practices or issues associated with fair treatment and consumer disclosure, monadic tests, or one-off tests are appropriate and will cover more offices and employees. In monadic test designs an office or employee is tested once because the financial institution is not specifically testing for discrimination. The demographic makeup of the testers should match the customer base, and the profiles and scenarios should approximate customer inquiries. In a self-testing program you must also decide the selection process for the specific representatives or branches to be evaluated. While it is of course preferable to have some testing done at every branch, this often is not possible given budgetary constraints. In this case, you may elect to test half or a smaller but significant proportion of your branches in person, and conduct telephone tests at the others, if branch calls are a reasonable entry channel for potential loan customers. Call centers should also be evaluated; the number of tests is dependent on the number of employees and average call load per day. Implementing a Preapplication Self-Testing Program Testers should undergo a comprehensive training (four to six hours) and fill out a very detailed questionnaire (10 to 15 pages, with a two-page narrative for qualitative evaluation). The testers are not told the true objective of the study is compliance related (e.g., to detect instances of disparate treatment, adherence to nondeposit investment sales guidelines, or misleading sales practices) because this may bias them to discover or amplify treatment issues. The testers are told the study is being conducted as a quality control audit; it is in fact a double-blind test because the representatives also don t know they are being tested. Informing employees about the ways you intend to measure performance and then providing feedback is a good practice. It serves to motivate employee adherence to company protocols and sets the tone for the program in your employees minds. If you are conducting a self-testing program for the first time, it is a good idea to conduct the tests before informing your sales and service staff. However, you should inform employees that your position is to test and provide feedback to them on a regular basis. They should know the protocols you expect them to adhere to, how they will be measured, and the reasons for the selftesting. However, you should not coach as to field dates, approach mechanisms, and locations, as this could affect the success of the testing. During the field phase of the mystery shopping, it is important to set a schedule that prevents detection. Criteria related to staffing, customer traffic, and employee administrative duties should be factored into the schedule, as well as cross-branch communication between the loan officers. Post-application Interviews If the objective of preapplication testing is to uncover quality of assistance offered to potential customers right up to, but stopping short of, the submission of an application, post-application interviews evaluate the process from beginning to end. A comprehensive fair lending evaluation program, in fact, uses both methods so the entire process is evaluated from start to finish. A telephone-based interview works best because it produces the highest level of cooperation and ensures feedback from the appropriate customers. While a self-administered mail survey is more cost effective, it will yield a lower cooperation rate, and the responses from the customers that respond may be far different from those who did not return the survey. A random sampling of customers or loan applicants should be utilized. In a fair lending test, four groups of customers (nonminority approved, nonminority denied, minority approved, and minority denied) are contacted and interviewed about the quality of assistance they received and their satisfaction with the process, the representative(s) who ABA Bank Compliance March/April

6 serviced their applications and their experiences during the loan closings. Both the interviewers and the respondents are not informed of the true purpose of the study; it should be presented as a quality of service audit, just like the preapplication test. To generate the sample for a postapplication study, you should go back to applicants no more than three to six months after the application date (because respondents typically don t remember the interaction much beyond that time frame). A sample of at least 100 completed interviews in each quota group is necessary for adequate significance testing. The applicant questionnaire can be largely similar to the tester questionnaire in the areas of measurement specifically, access, courtesies exhibited, sales practices, product exposure, demonstration and recommendation, terms quoted, closing actions, and satisfaction. You should, however, be sure to measure the assistance the customer received during the sales encounter and for loans offered and discussed and the underwriting and closing process. Analyzing the Results There are a number of problems that can be brought out or identified by the results of a self-testing program. For example, in matched-pair fair lending tests, results should be examined at the unit, office, or employee level. The most obvious issues are blatant examples of disparate treatment, such as when a loan officer schedules an appointment with one member of the matched pair and actively avoids another. Other examples include one member getting short shrift (e.g., basically Because of the sensitivity of the information collected, you should take steps to keep the information privileged and protected. Discuss the matter with legal counsel. Most financial institutions either work with their legal departments or retain outside counsel. Whether you use the company s legal department or outside counsel, be sure that program communications and the availability of the data generated by the program are limited to counsel and to a controlled number of need to know individuals who can take action on the results. It is particularly important that the individuals exposed to the data are in a position to correct the legal and business issues that self-testing programs uncover. This will help ensure the success of the program, as well as its usefulness and lebeing handed a brochure and told to call with any questions) while the other gets a 45-minute appointment during which a number of loan options are discussed and an application is encouraged. At the corporate level, when you examine the total results by race you should statistically test the data at the 80 percent (directional) and 95 percent (significant) confidence levels to uncover current or potential differences in quality of service between protected and nonprotected classes of consumers. In order for the statistical testing to have the most value, you should have 100 tests, or, at a minimum, 30 tests per group. The main areas of review: access; courtesy; identifying needs; product exposure and demonstration of appropriate products (features and benefits discussed); terms quoted and voluntary rate and fees disclosures; creativity or making suggestions for the best product to fit the customer s situation; in lending, how an applicant can improve his or her chances for loan approval; closing actions (including encouragement to apply) and tester satisfaction; product terms and pricing met customer expectations; and customer perceptions and what they like and dislike. It s important to recognize that while there may be no instances of blatant disparate or unfair treatment, subtler issues might exist and need to be addressed. The subtler issues can include a directional difference in mean approval time that favors the nonminority applicant or quickly covering certain elements, such as the existence of a prepayment penalty or that credit insurance is optional. After the results of your self-testing program have been tabulated, the next step is a data review with compliance management, in which a consensus of findings and how to disseminate the information is reached. Following this, a formal presentation of the critical findings should be conducted with management. Securing and Using the Results 8 March/April 2003 ABA Bank Compliance

7