The historic $98 million auto lending discrimination settlement between Ally Financial and Ally Bank (collectively, “Ally”), and the United States Department of Justice (“DOJ”) and Consumer Financial Protection Bureau (“CFPB”) launched a warning shot at the indirect auto lending industry and signaled that the DOJ and CFPB mean business when it comes to fair lending enforcement against non-mortgage creditors. The settlement is important for several reasons, including that:
It is the federal government’s largest ever auto loan discrimination settlement;
It is the first nationwide auto lending settlement
It resolves the first joint DOJ/CFPB fair lending investigation;
In contrast to a large majority of other federal fair lending settlements, which focused solely on African-American and Hispanic borrowers, the Ally settlement also resolves claims that Asian/Pacific Islander borrowers paid higher auto loan prices than non-Hispanic white borrowers; and
In contrast to fair lending settlements involving residential mortgage loans for which there is data identifying the borrowers’ actual race and ethnicity, the government imputed Ally’s borrowers’ race and ethnicity using a controversial proxying methodology; this means that borrowers who are compensated for being allegedly harmed based on their minority status could, in fact, be Caucasian.
This client alert summarizes the Ally settlement and highlights the key compliance takeaways for indirect auto lenders and other creditors.
The Ally settlement stems from a CFPB examination conducted between September 2012 and November 2012. The CFPB’s examination included an evaluation of Ally’s lending policies, procedures, internal controls, and fair lending compliance practices, and an analysis of loan data on the auto loans Ally funded between April 1, 2011 and March 31, 2013.
In accordance with Section 706(g) of the Equal Credit Opportunity Act (“ECOA”) and a December 6, 2012 Memorandum of Understanding between DOJ and the CFPB, on November 7, 2013, the CFPB referred the matter to the DOJ because it had concluded that Ally had engaged in a pattern or practice of lending discrimination. DOJ initiated its investigation, which it coordinated with the CFPB, on November 13, 2013, and the parties announced their settlement a little more than a month later, on December 20, 2013. The settlement terms are set forth in a Consent Order with the CFPB dated December 19, 2014 and a Consent Order with DOJ dated December 23, 2013.
Ally’s Auto Pricing Practices
According to the settlement documents, Ally is one of the largest indirect auto finance companies in the United States. Ally has relationships with more than 12,000 automobile dealers across the United States, and since April 2011, it funded nearly three million auto loans.
Auto dealers often arrange third-party financing for the consumers to whom they sell cars. Although the process varies, typically a dealer selling a car to a consumer will submit a loan application on the consumer’s behalf to one or more auto lending companies. If the lender approves the consumer’s application, it will determine the minimum interest rate - the “buy rate” - at which it will agree to purchase the loan. The lender bases its buy rate on its cost of funds, as adjusted to account for the consumer’s credit history, and other risk factors associated with the transaction.
Like many indirect auto lenders, Ally allows the dealers from which it buys loans to mark up the interest rate charged to a consumer above the risk-based buy rate. According to the DOJ’s complaint (“DOJ Complaint”), Ally retains a portion of the profits earned from the mark ups and pays the remainder to dealers. The DOJ Complaint also states that Ally caps the amount of dealer mark up at 200 or 250 basis points, depending on the contract term and the consumer’s credit tier.
CFPB and DOJ Claims
The CFPB and DOJ allege that between April 1, 2011 and March 31, 2012, Ally’s African-American borrowers paid an average of 29 basis points more in dealer mark up than similarly situated non-Hispanic white borrowers, their Hispanic borrowers paid 20 basis points more, and their Asian/Pacific Islander borrowers paid 22 basis points more. According to the settlement documents, these differences amounted to $200-$300 in interest over the life of the borrowers’ retail installment contracts.
The CFPB and DOJ claim that Ally did not perform any fair lending monitoring of dealer mark ups until March 2013. Further, the DOJ Complaint states that Ally’s fair lending monitoring since March 2013 “has been entirely inadequate.” Both the DOJ and CFPB claim that the dealer mark up disparities against African-American, Hispanic and Asian/Pacific Islander borrowers resulted from Ally’s dealer mark up policy and lack of compliance monitoring.
Additionally, the DOJ Complaint indicates that the government’s analyses also showed that the African-American, Hispanic and Asian/Pacific Islander borrowers with the strongest credit had the largest interest rate disparities, meaning that “Ally discriminates most severely against [the most creditworthy minority borrowers.]”
Summary of Settlement Terms
The Consent Orders contain monetary provisions and detailed compliance obligations. The monetary provisions require Ally to deposit $80 million into a settlement fund and to hire a settlement administrator to distribute the $80 million to borrowers identified by the CFPB and DOJ. Additionally, the CFPB Consent Order requires Ally to pay an $18 million penalty to the CFPB.
The Consent Orders also require that Ally implement a compliance plan that includes, among other things:
A dealer compensation policy that limits dealer mark up to no more than Ally’s current caps;
Regular notices to all dealers explaining ECOA, Ally’s expectations with respect to ECOA compliance, and the dealer’s obligations to price retail installment contracts in a non-discriminatory manner, including exercising discretion to set a consumer’s contract rate;
Quarterly analysis of dealer-specific retail installment contract pricing data for disparities on a prohibited basis;
Quarterly and annual analysis of portfolio-wide retail installment contract pricing data for disparities on a prohibited basis that reflects the same methods and controls the CFPB and DOJ applied in their analyses;
Appropriate corrective action with respect to dealers who are identified in the quarterly analysis of dealer-specific retail installment contract pricing data for disparities on a prohibited basis, culminating in the restriction or elimination of dealers’ ability to exercise discretion in setting a consumer’s contract or exclusion of dealers from future transactions with Ally;
Remuneration of affected consumers within 60 days of an analysis that identifies statistically significant disparities on dealer mark up on a prohibited basis within an individual dealer’s transactions with Ally;
Remuneration of affected consumers by March 1 of each calendar year until the termination of the Consent Orders, if the portfolio-wide analysis for the preceding calendar year identifies statistically significant dealer mark up disparities for that group at or above the agreed upon target.
Interestingly, Ally may submit a non-discretionary dealer compensation plan to the government during the pendency of the Consent Orders for review. If the government issues a determination of non-objection and Ally implements the non-discretionary dealer compensation plan, it will no longer be required to conduct analyses and future corrective action that are otherwise required under the Consent Order.
As noted above, the CFPB and DOJ focused their fair lending analysis on differences across race and ethnicity in the amount of dealer mark up. The DOJ Complaint and both agencies’ Consent Orders contain information about the government’s statistical analyses that may be useful for other institutions in developing their own fair lending compliance monitoring.
Proxying for Race and Ethnicity
Unlike residential mortgage lenders, which are required by statute to gather information about loan applicants’ race and ethnicity, auto lenders do not obtain - and in fact are prohibited from obtaining - race and ethnicity information about their borrowers. The CFPB and DOJ therefore had to use a proxy approach to impute borrower race and ethnicity for their analyses.
The use of proxies in fair lending cases involving auto lenders is not new, and auto lenders, credit card issuers, and other non-mortgage consumer lenders have used geographic and surname proxying for years to evaluate their fair lending risk. In the Ally case, the CFPB and DOJ used a proxying approach, the Bayesian Improved Surname Geocoding (“BISG”) method, that combines geography and surname approaches to estimate a joint probability that a borrower is a member of a particular race or ethnicity. Although the CFPB and DOJ appear convinced that the BISG method is appropriate to assess disparities in auto lending, it has not been subject to rigorous testing or validation within the consumer lending environment and we have observed that the BISG method has been shown to have a high error rate when tested against Home Mortgage Disclosure Act (“HMDA”) data.
Analysis of Mark Up
According to the settlement documents, the CFPB and DOJ focused their fair lending analysis on dealer mark up, i.e., the difference between each borrower’s contract rate and Ally’s buy rate, on non-subvented loans. The DOJ Complaint states that the dealer mark up disparities were statistically significant at the 95 percent confidence rate, meaning that the probability that such disparities occurred by random chance is less than 5 percent.
The DOJ Complaint states that because dealer mark ups are not determined by creditworthiness or other objective criteria related to borrower risk, “it is not proper to include factors measuring creditworthiness and other objective criteria related to borrower risk in the statistical analysis of interest rate mark up disparities.” In any case, it appears that the agencies tested whether controlling for such objective risk factors made a difference and concluded that it did not.
Although the Ally Consent Orders reflect a negotiated settlement and are not binding on other institutions, the settlement documents provide several insights about how the CFPB and DOJ view ECOA compliance and enforcement.
The CFPB Consent Order indicates that the CFPB examined not only whether Ally had violated ECOA, but also whether it had sufficient “processes for managing compliance with federal consumer financial laws.” This reflects the CFPB’s continued focus on whether financial institutions have adequate compliance systems.
Similarly, both the CFPB and DOJ emphasized Ally’s failure to implement an effective compliance program in their settlement documents and press releases. This is consistent with guidance set out in the CFPB’s March 2013 Bulletin 2013-02, “Indirect Auto Lending and Compliance with the Equal Credit Opportunity Act” (the “CFPB Bulletin” or “Bulletin”) advising indirect auto lenders subject to CFPB jurisdiction to develop a robust fair lending compliance management system. According to the CFPB Bulletin, a strong fair lending compliance program for an indirect auto lender would include the following:
An up-to-date fair lending policy statement;
Regular fair lending training for all employees involved with any aspect of the institution’s credit transactions, as well as all officers and Board members;
Ongoing monitoring for compliance with fair lending policies and procedures;
Ongoing monitoring for compliance with other policies and procedures that are intended to reduce fair lending risk (such as controls on dealer discretion);
Review of lending policies for potential fair lending violations, including potential disparate impact;
Depending on the size and complexity of the financial institution, regular analysis of loan data in all product areas for potential disparities on a prohibited basis in pricing, underwriting, or other aspects of the credit transaction;
Regular assessment of the marketing of loan products; and
Meaningful oversight of fair lending compliance by management and, where appropriate, the financial institution’s board of directors.
CFPB-regulated institutions should consider evaluating their compliance infrastructure before they are examined by the CFPB to determine whether enhancements may be appropriate, and all lenders should consider whether they have adequate controls in place to manage fair lending risk.
Discretion Remains a Major Focal Point
Like a majority of the fair lending investigations settled during the past several years, the Ally case focused on discretionary pricing. Although price negotiation has been a long-standing practice in mortgage, auto, and other types of lending, and price discretion allows lenders to respond to market pressures and competition, the government appears to be on a crusade to eliminate it. The fact that the Consent Orders incentivize Ally to adopt a non-discretionary dealer compensation plan is further evidence that the government disfavors discretionary pricing in consumer lending.
The CFPB Bulletin similarly focuses on discretionary pricing and encourages indirect auto lenders to consider eliminating mark ups. According to the Bulletin, lenders allowing dealer mark ups should consider:
Imposing controls on dealer mark up and compensation policies, or otherwise revising dealer mark up and compensation policies, and also monitoring and addressing the effects of those policies in the manner described below, so as to address unexplained pricing disparities on prohibited bases; or
Eliminating dealer discretion to mark up buy rates and fairly compensating dealers using another mechanism, such as a flat fee per transaction, that does not result in discrimination.
The Bulletin goes on to state that indirect auto lenders that retain dealer mark up may want to consider adopting a monitoring and corrective action system with the following features:
Sending communications to all participating dealers explaining the ECOA, stating the lender’s expectations with respect to ECOA compliance, and articulating the dealer’s obligation to mark up interest rates in a non-discriminatory manner in instances where such mark ups are permitted;
Conducting regular analyses of both dealer-specific and portfolio-wide loan pricing data for potential disparities on a prohibited basis resulting from dealer mark up and compensation policies;
Commencing prompt corrective action against dealers, including restricting or eliminating their use of dealer mark up and compensation policies or excluding dealers from future transactions, when analysis identifies unexplained disparities on a prohibited basis; and
Promptly remunerating affected consumers when unexplained disparities on a prohibited basis are identified either within an individual dealer’s transactions or across the lender’s portfolio.
Given the high degree of government scrutiny, lending institutions that allow for discretionary pricing should consider evaluating whether they have sufficient controls in place to manage the related fair lending risk.
Responsibility for Third-Party Originators
Consistent with recent fair lending settlements with wholesale mortgage lenders, and the Department of Housing and Urban Development’s disparate impact rule, the CFPB and DOJ claim that Ally is legally liable for the pricing disparities resulting from the activities of independent, third-party auto dealers. The CFPB Bulletin takes the same position, stating that indirect auto lenders should not operate under the incorrect assumption that they are not responsible for third-party dealer mark ups. Although there are legal defenses to this view, lending institutions should recognize that the CFPB and DOJ are likely to continue taking action against lenders for the acts of third-party originators
The Ally settlement documents include very specific fair lending monitoring requirements that, while not binding on any other lender, provide insights as to compliance features the CFPB and DOJ may favor.
First, the settlement documents indicate that because dealers do not use borrower credit and other risk factors in determining mark up, the CFPB and DOJ do not believe it is appropriate to control for those factors when testing for mark up disparities. Although this is useful information, it raises a number of questions. For example, did the agencies use any controls in their analyses, such as loan amount or market indicators, or were the disparities they cited based on raw differences? Further, if a lender could show that credit and risk factors influenced mark up, would it be appropriate to control for them?
Second, the settlement documents provide for both within dealer and portfolio-wide monitoring. This is consistent with the guidance in the Bulletin, and suggests that the CFPB and DOJ believe that indirect auto lenders are responsible for both dealer-level price disparities, even though they typically fund only a tiny percentage of any given dealer’s loans, and cross-dealer disparities, even though they may arise regardless of whether any given dealer his disparities within its own operations.
Third, the settlement documents require Ally to pay borrower remuneration for both dealer-level and portfolio-wide disparities, but do not provide any detail on how such remuneration should be calculated. As noted above, the CFPB Bulletin also addresses remuneration, stating that indirect auto lenders that retain dealer mark up may wish to include remuneration for both dealer-level and across-portfolio disparities. This raises a number of compliance questions, including whether monitoring programs should include remuneration, and, if so, under what circumstances.
Although the compliance monitoring requirements in the Ally settlement provide useful insights into the CFPB and DOJ’s views, they should not be viewed as a blueprint for every lender. Indirect auto lenders should consider their size, complexity and particular areas of risk when designing a fair lending compliance program. Further, they should carefully evaluate the pros and cons of including a remuneration component.
As noted above, race and ethnicity proxying, including the BISG method, has been shown to have high error rates. Among other matters, this means that borrowers receiving remuneration for alleged discriminatory pricing may not even be members of the racial or ethnic group imputed to them by the proxy process. Nevertheless, because the CFPB and DOJ appear to favor the BISG approach, auto and other non-mortgage lenders may want to consider using it in their fair lending testing programs.
* * * * *
The Ally settlement is likely the first in a new wave of fair lending enforcement against non-mortgage consumer lenders. Institutions should consider evaluating their fair lending compliance processes in light of the settlement to ensure that they are adequately managing their risk.
 In the second-largest DOJ auto lending settlement, Compass Bank paid $1.75 million in 2007 to resolve allegations that its auto lending policies and procedures discriminated against applicants on the basis of marital status. United States v. Compass Bank, Case No. CV 07-B-0102-S (N.D. Ala. Feb. 21, 2007), http://www.justice.gov/crt/about/hce/documents/compasssettle.pdf.
 Although the claims against Ally focus predominantly on differences in dealer mark up, this particular allegation refers to differences in “interest rate.”
 12 U.S.C. § 2801 et seq.
 See 12 C.F.R. § 1002.5(b).
 See, e.g., United States v. Nara Bank et al., Case No. CV-09-7124-RGK (C.D. Cal. Nov. 18, 2002)(partial consent decree).
 Geographic proxying involves assigning a race or ethnicity to a borrower based on the racial or ethnic composition of the census tract in which the borrower resides. Surname proxying involves using Census Bureau data to assign ethnicity (and in some limited contexts, race) based on the borrower’s last name.
 The DOJ Complaint states that the “BISG method is recognized by social scientists, statisticians, and economists as a tested and accurate way to determine differences in experiences based on race or ethnicity for large groups of individuals for whom self-identified race and ethnicity data is not available.” The DOJ Complaint does not, however, acknowledge that while the BISG method has been used for research purposes, primarily in the health care field, it has not been previously used to impose legal liability on a financial services provider.
 Subvented loans are loans that are subsidized by the auto manufacturer to promote sales.
 Complaint at 9, United States v. Ally Financial Inc. and Ally Bank, Case No. CV 13-15180, (E.D. Mich. Dec. 20, 2013), http://www.justice.gov/crt/about/hce/documents/allycomp.pdf.
 See generally Loan Originator Compensation Requirements Under the Truth in Lending Act (Regulation Z), Bureau of Consumer Fin. Protection, 78 Fed. Reg. 11,280 (Feb. 15, 2013); CFPB Bulletin 2013-02, Indirect Auto Lending and Compliance with the Equal Credit Opportunity Act, available at http://files.consumerfinance.gov/f/201303_cfpb_march_-Auto-Finance-Bulletin.pdf. See also United States v. Countrywide Financial Corporation et al., Case No. CV11-10540 (C.D. Cal. Dec. 28, 2011)(consent order); Consumer Financial Protection Bureau and United States v. National City Bank, Case No. CV-01817-CB (W.D. Penn. Dec. 23, 2013)(consent order); United States v. Wells Fargo Bank, N.A., Case No. CV-12-1150 (D.D.C. Sep. 21, 2012)(consent order).
 See, e.g., United States v. Countrywide Financial Corporation et al., Case No. CV11-10540 (C.D. Cal. Dec. 28, 2011)(consent order); Conciliation Agreement between U.S. Department of Housing and Urban Development and MortgageIT, Inc., FHEO Case No. 00-09-0001-8 (Sept. 30, 2013); Consumer Financial Protection Bureau and United States v. National City Bank, Case No. CV-01817-CB (W.D. Penn. Dec. 23, 2013)(consent order); United States v. Wells Fargo Bank, N.A., Case No. CV-12-1150 (D.D.C. Sept. 21, 2012)(consent order).
 See generally Meyer v. Holly, 537 U.S. 280 (2003); Comments Concerning Proposed Rulemaking on the “Discriminatory Effects Standard of the Fair Housing Act,” available at http://www.aba.com/aba/documents/news/DisparateImpactLetter11712.pdf.