The CFPB’s annual fair lending report covering its 2019 activities is scheduled to be published in tomorrow’s Federal Register. While most of the report recycles information about which we have previously blogged, it does contain the following noteworthy information:
- The section of the report on “Innovations in access to credit” includes a subsection about “providing adverse action notices when using artificial intelligence and machine learning models.” The Bureau states that an area of innovation that it is monitoring for fair lending and credit access issues is “artificial intelligence (AI), and more specifically, machine learning (ML), a subset of AI.” The Bureau observes that “there may be questions about how institutions can comply with ECOA and FCRA adverse action notice requirements “if the reasons driving an AI decision are based on complex interrelationships.” It comments that “the existing regulatory framework has built-in flexibility that can be compatible with AI algorithms.” The Bureau gives the following examples of such flexibility:
- The Bureau states that the Regulation B (ECOA) official commentary provides that in giving specific reasons for adverse action, “a creditor need not describe how or why a disclosed factor adversely affected an application, or, for credit scoring systems, how the factor relates to creditworthiness. Thus, the Official Interpretation provides an example that a creditor may disclose a reason for a denial, even if the relationship of that disclosed factor to predicting creditworthiness may be unclear to the applicant.” The Bureau comments that this flexibility could be useful to creditors when issuing adverse action notices “based on AI models where the variables and key reasons are known, but which may rely upon non-intuitive relationships.”
- The Bureau states that “neither ECOA nor Regulation B mandate the use of any particular list of reasons. Indeed, the regulation provides that creditors must accurately describe the factors actually considered and scored by the creditor, even if those reasons are not reflected on the current sample forms.” The Bureau comments that this latitude could be useful to creditors “when providing reasons that reflect alternative data sources and more complex models.”
The Bureau also notes that new tools continue to be developed to explain AI decisions and “hold great promise…to facilitate use of AI for credit underwriting compatible with adverse action notice requirements.” It also comments that despite the flexibility of the existing regulatory framework “there still may be some regulatory uncertainty about how aspects of the adverse action notice requirements apply in the context of AI/ML.” The Bureau encourages entities to consider using the Bureau’s new innovation policies (e.g. No-Action Letter Policy) to address potential compliance issues.
- In discussing its annual risk-based prioritization process for 2019, the Bureau states that it focused its fair lending supervision efforts on mortgage origination, small business lending, student loan origination, and debt collection and model use and provides the following details:
- Mortgage origination. The Bureau continued to focus on (1) redlining and whether lenders intentionally discouraged applications from individuals living or seeking credit in minority neighborhoods, (2) assessing whether there is discrimination in underwriting and pricing processes including steering, and (3) HMDA data integrity and validation (which supports ECOA exams) and HMDA diagnostic work (monitoring and assessing new rule compliance).
- Small business lending. The Bureau focused on assessing whether (1) there is discrimination in the application, underwriting, and pricing processes, (2) creditors are redlining, and (3) there are weaknesses in fair lending related compliance management systems.
- Student loan origination. The Bureau focused on whether there is discrimination in policies and practices governing underwriting and pricing.
- Debt collection and model use. The Bureau focused on whether there is discrimination in policies and practices governing auto loan servicing and credit card collections, including the use of models that predict recovery outcomes.
- In 2019, the Bureau initiated 26 supervisory events (Matters Requiring Attention and Supervisory Recommendations) relating to fair lending.
- In 2019, the Bureau referred three ECOA matters to the DOJ, with such matters involving (1) a pattern or practice of redlining in mortgage origination based on race, (2) discrimination based on public assistance in mortgage origination, and (3) discrimination based on race and national origin in auto origination.
- In 2019, neither the CFPB nor any of the other eleven federal agencies with ECOA enforcement authority brought a public enforcement action for ECOA violations.
- In 2019, in addition to the Bureau’s three referrals to the DOJ, the FDIC referred two matters and the Federal Reserve and the NCUA each referred one matter. The matters referred by the FDIC involved discrimination in auto origination based on the applicant’s receipt of income derived from a public assistance program and discrimination in the underwriting of commercial loans based on religion. The Fed’s referral involved pricing discrimination based on national origin, race, and sex, and the NCUA referral involved discrimination on the basis of age.
- The new report does not discuss the Bureau’s plans to reexamine the disparate impact doctrine in light of the U.S. Supreme Court’s Inclusive Communities decision and to hold a symposium on disparate impact and the ECOA.
On May 20, 2020, from 12 p.m. to 1 p.m. ET, Ballard Spahr will hold a webinar, “Fair Lending and UDAAP Considerations During the COVID-19 Era.” For more information and to register, click here.