Section 1473(q) of the Dodd-Frank Act (now codified at 12 U.S.C. § 3354(q)) amended the Financial Institutions Reform, Recovery, and Enforcement Act of 1989 (“FIRREA”) to instruct the CFPB, Fed, OCC, FDIC, NCUA, and FHFA (collectively, the “agencies”) to jointly develop regulations for quality control standards for automated valuation models (“AVMs”), defined as “any computerized model used by mortgage originators and secondary market issuers to determine the collateral worth of a mortgage secured by a consumer’s principal dwelling.” As part of the rulemaking process, the Small Business Regulatory Enforcement Fairness Act of 1996 (“SBREFA”) requires the CFPB to convene a Small Business Review Panel to consider whether the rule could have a significant economic impact on a substantial number of small entities. Accordingly, on February 23, 2022, the CFPB released an outline of proposals and alternatives under consideration by the agencies to seek informed feedback and recommendations from small businesses likely to be subject to the rule.
As amended, subparts (1) – (4) of FIRREA Section 1125(a) mandate that the agencies establish four specific quality control standards for AVMs. FIRREA Section 1125(a)(5) also affords the agencies discretion to adopt standards designed to “account for any other such factor that the agencies…determine to be appropriate.” As such, the CFPB’s SBREFA outline proposes creating a fifth such discretionary quality control standard “designed to protect against unlawful discrimination.”
In support of its proposal, the CFPB asserts that algorithmic systems such as AVMs are subject to Federal nondiscrimination laws, including the Equal Credit Opportunity Act (“ECOA”), because a lender evaluating an applicant’s collateral could use an AVM “in a way that would treat an applicant differently on a prohibited basis or result in unlawful discrimination against an applicant on a prohibited basis.” The CFPB then notes that it recognizes three different methods of proving discrimination under the ECOA and its implementing regulation (“Regulation B”): (1) overt discrimination; (2) disparate treatment; and (3) disparate impact. It is worth mentioning that overt discrimination has been viewed by federal regulators such as DOJ and the FDIC as a blatant type of disparate treatment that does not require an inference or presumption based on circumstantial evidence. However, it appears that the CFPB considers these theories to be distinct from one another.
The third method of proving discrimination articulated by the CFPB, disparate impact, has been a controversial theory of liability because it imposes liability on a creditor even where the creditor had no intent to discriminate against an applicant. Rather, the theory presumes that the creditor has treated applicants fairly and consistently in accordance with some facially neutral policy or procedure of the creditor. Of course, the disparate impact theory gained traction in the subprime lending cases post-2008 and then loomed large in the CFPB’s enforcement actions against indirect auto lenders, the latter of which were scrutinized by Congress in its decision to rescind the CFPB’s indirect auto lending guidance using the Congressional Review Act. In fact, it remains a legal question whether disparate impact claims are cognizable under the ECOA since the United States Supreme Court (“Supreme Court”) has never considered the issue, though civil rights advocates point to the Supreme Court’s willingness in the 2015 Inclusive Communities case to recognize the theory for discrimination claims brought under the Fair Housing Act (“FHA”).
Thus, should the agencies adopt a final rule that relies upon disparate impact under the ECOA as a legal basis to justify imposing a quality control standard on AVMs (or muddies the waters by relying upon both the ECOA and the FHA without distinction), it is possible that the rule could be challenged under the Administrative Procedures Act as not in accordance with the law. Alternatively, if the CFPB were to bring an enforcement action against a creditor for allegedly violating either the final rule’s quality control standard or ECOA itself on the basis of disparate impact, the creditor could defend itself by arguing among other things that disparate impact claims are not cognizable under the ECOA. Indeed, the ECOA lacks any “results-oriented” language like the “otherwise make available” language of the FHA or the “otherwise adversely affect” language of the Age Discrimination in Employment Act, which the Supreme Court, in decisions issued a decade apart, relied on in recognizing disparate impact liability.
Even if the plain language of the ECOA could not support a disparate impact claim, the CFPB might argue that the statute’s anti-discrimination provision is ambiguous (by asserting, for instance, that the word “discriminate” could be interpreted to encompass both intent-based and effects-based actions), in which case the CFPB may expect the reviewing court to grant its interpretation Chevron deference. See Chevron, U.S.A. v. Natural Resources Defense Council, Inc. 476 U.S. 837 (1984). But this argument also might prove difficult because Chevron deference is appropriate only when it appears that Congress has “delegated authority to the agency generally to make rules carrying the force of law, and … the agency interpretation claiming deference was promulgated in the exercise of such authority.” See Public Citizen, Inc. v. U.S. Dept. of Health and Human Services, 332 F.3d 654, 659 (D.C. Cir. 2003) (quoting U.S. v. Mead Corp., 533 U.S. 218 (2001)). In examining Regulation B, which was originally issued by the Fed and subsequently readopted by the CFPB, the only references to the concept of disparate impact appear in 12 C.F.R. § 1002.6(a) and Official Interpretation 6(a)-2. However, these provisions merely summarize the ECOA’s legislative history and Supreme Court precedent under Title VII of the Civil Rights Act, and even then, acknowledge only that the ECOA “may” prohibit acts that are discriminatory in effect. The CFPB has articulated its belief that disparate impact is cognizable under the ECOA elsewhere, including in a compliance bulletin and its examination manual, but those materials carry no force of law under the CFPB’s own recently-adopted rule. Thus, a reviewing court could conclude that the mere recitation of legislative history and of a judicial doctrine developed under an unrelated statute was not an actual exercise of rulemaking authority under the ECOA, and therefore that the agencies’ interpretation of the ECOA as expressed in Regulation B is not entitled to Chevron deference. In that circumstance, the reviewing court would be free to resolve any purported ambiguity in the ECOA according to its own construction, affording respect to the agencies’ position only to the extent it is persuasive. And should either of these issues – disparate impact under the ECOA or the availability of Chevron deference – ultimately be appealed to the Supreme Court, there may well be four justices willing to grant certiorari to consider them.
The uncertain outcome of any challenge to the CFPB’s use of disparate impact in a rulemaking or in enforcing the ECOA, given the stakes involved, suggests that the CFPB may seek to resolve matters via settlement rather than risking litigation in federal court. However, only time will tell whether the CFPB is spoiling for a fight.