On November 24, Republicans on the House Committee on Financial Services issued a report regarding the CFPB’s approach for determining discrimination in the auto lending industry. The report questions the CFPB’s proxy methodology and its authority to bring claims against banks involved in indirect auto lending under the Equal Credit Opportunity Act’s (ECOA) disparate impact theory. According to the report, disparate impact “is a controversial legal theory of liability in discrimination cases.” The report further states that, even if it assumes that the ECOA permits disparate impact claims, the CFPB is nonetheless required to identify the following to establish a prima facie case: (i) a specific policy or practice adopted by the creditor; (ii) disparate impact on a prohibited basis; and (iii) a causal relationship between the challenged practice and the alleged disparate impact. The report states, “[d]ocuments obtained by the Committee show that the Bureau will likely have difficulty proving any one of these requirements, much less all three.” Notably, the report criticizes the CFPB’s adoption of the Bayesian Improved Surname Geocoding proxy method, which “combines surname- and geography-based information into a single proxy probability for race and ethnicity,” labeling it as “faulty and unreliable.” The report further suggests that the CFPB observed the method to be “less accurate . . . than some proprietary proxy methods that use nonpublic data.” In closing, the report comments on the CFPB’s “ambition to eliminate dealer markup” by summarizing (i) a December 2013 settlement in which the CFPB used its leverage over a bank holding company to negotiate the settlement terms; (ii) the agency’s plans to increase the number of individual enforcement actions on dealer markup and compensation policies; and (iii) potential ECOA rulemaking to “promulgate a regulation prohibiting lenders from compensating dealers based on the terms of a loan.”