Ally Financial's recent settlement with the government over auto dealer markups on indirect auto loans is an enforcement case based on leverage, not law. The Justice Department and the Consumer Financial Protection Bureau isolated a large bank creditor to make an example of dealer compensation practices the CFPB condemned last March.
Prosecutors alleged discriminatory loan charges based on a portfolio analysis of hundreds of thousands of loans made by car dealers nationwide, without any allegation of discriminatory pricing by the people who actually set the loan rates—the dealers themselves, who the Dodd-Frank Act largely shields from CFPB enforcement.
Whether the indirect lender who settled intended to discriminate was also a moot point to the government based on a theory of disparate impact built around statistical analysis measuring the difference of average markups among borrower groupings using flawed proxies.
"Whether or not Ally consciously intended to discriminate makes no practical difference to consumers," said CFPB Director Richard Cordray. "In fact, we do not allege that Ally did so. Yet the outcome, and the harm to consumers, is the very same here."
But it does make a difference. The challenge of demonstrating illegal discrimination using disparate impact is to observe the Equal Credit Opportunity Act's goal of not stereotyping individuals while conducting analyses that rely on statistical differences ascribed to groups labeled by their prohibited basis characteristics. This challenge is compounded by using proxies as substitutes for true characteristic information.
The CFPB and Justice Department assert proxy analysis for enforcement, but the examination procedures the CFPB cites as its authority applies proxy analysis for the purpose of examination scoping for disparate treatment. That is, proxies are considered appropriate for identifying red flags for further investigation, not as sufficient for proving a violation of ECOA. As the Supreme Court found in Watson v. Fort Worth Bank & Trust, it is inappropriate "to hold a defendant liable for unintentional discrimination on the basis of less evidence than is required to prove intentional discrimination."
Yet this is precisely what results from the government's use of error-ridden proxies and statistical tests to impose liability in the absence of discriminatory intent. The law requires the government to prove two things—that plaintiffs are members of a prohibited basis group and that they individually suffered markups not applied to similarly-situated non-minority borrowers. The government fails this burden on both counts. First, the government has no proof that particular individuals are actually prohibited basis group members — they only have proxy-based probabilities — and the law is not satisfied by the probability of being a minority.
Second, the government does not — and cannot — prove that any prohibited basis group member paid a dealer markup that was not paid by a similarly situated non-minority group member. In fact, for every markup paid by an alleged minority group member there was one or more non-minority group members within the portfolio of borrowers who paid the same markup or more. In the absence of discriminatory intent, this does not prove a denial of equal credit opportunity under the law.
CFPB's allegations amount to little more than non-random samples resulting in non-random differences. Such a departure from the laws of chance, as the Watson Court notes, provides an unrealistic basis to assert unlawful discrimination: "It would be equally unrealistic to suppose that [lenders] can eliminate, or discover and explain, the myriad of innocent causes that may lead to statistical imbalances in the composition of their [loan portfolio's pricing]."
The greatest irony is that the government's remedial order compensates African-Americans, Asians and Hispanics for paying the exact same loan markups that non-minorities paid, but for whom compensation will not be granted. In other words, similarly situated consumers with exactly the same markup amounts will be treated differently by the government solely on the basis of their race or ethnic origins.
The government has overwhelming power and resources to pursue enforcement. Leveraging a settlement on threat of litigation with less evidence than is required to prove intentional discrimination is an abuse of power. Compensating presumptively-minority borrowers who pay the same markup as individual non-minority borrowers undermines equal credit opportunity policy.
Richard Riese is senior vice president of the Center for Regulatory Compliance at the American Bankers Association and was a co- author of the 1999 Interagency Fair Lending Examination Procedures.