The Consumer Financial Protection Bureau (CFPB)’s assault on dealer participation began with a simple question. Fueled by a now-discredited study and the series of class-action suits settled by several captives in the mid-2000s — settlements that resulted in the industry’s unofficial 250 basis-point ceiling on dealer rate markups — the bureau wondered if the cap had eliminated discriminatory credit pricing. Not only did finance sources not have the answer, they didn’t have adequate monitoring and testing programs in place to determine whether they did, according to internal CFPB memos.
In September 2012, four options were on the table: (1) Write an unfair, deceptive, or abusive acts or practices rule addressing its fair lending concerns regarding dealer markups, potentially banning the practice altogether; (2) develop a rule that clarified the liability of finance sources operating under the Equal Credit Opportunity Act regarding disparities caused by markups; (3) write a Truth in Lending Act rule requiring that markups be disclosed to consumers; or (4) advance a notice of proposed rulemaking to signal the bureau’s focus on dealer markups and highlight the potential UDAAP, ECOA, and TILA issues, “without deciding upon an approach.”
Due to the National Automobile Dealers Association (NADA)’s hard-fought exemption for new-car dealers, the bureau opted for an ECOA-based supervision and enforcement approach laid out in its controversial guidance on dealer participation. An internal memo shared between bureau officials days after the regulator published its compliance bulletin on March 21, 2013, would reveal that rulemaking aimed at banning markups or forcing dealers to disclose them to consumers was back on the table.
“There are several concerns with a rulemaking approach. First, the legal authority for all of the potential rulemakings is unclear given our lack of authority over dealers,” the April 3, 2013, memo reads, in part. “Second, the bureau would face considerable pressure from external groups if it sought to regulate or ban the practice itself — pressure that should not be underestimated.”
The bureau realized just how intense that pressure can be when Congress sent to President Trump’s desk an approved resolution of disapproval of the CFPB’s dealer participation guidance under the Congressional Review Act. On May 21, Trump’s signature marked the end of the industry’s more than five-year effort to protect a key source of dealer income. It also marked the first time the CRA, which prohibits a similar rule from being introduced unless authorized by Congress, was used to repeal a rule that had been in effect for several years.
Regulate by Enforcement
The CFPB alleged in its five-page fair lending guidance that allowing auto dealers to mark up interest rates on retail installment sale transactions as compensation for services rendered create a significant risk of unintentional, disparate impact discrimination — and that the provision in the ECOA’s Regulation B that lets creditors off the hook for another creditor’s violation doesn’t apply. The reason, according to the bureau’s guidance, is the provision doesn’t limit a creditor’s liability for its own violation, and a policy that allows dealers to violate the ECOA would mean the creditor is liable.
The bulletin offers two options to finance sources operating in the indirect finance channel: Eliminate dealer markups and replace them with another compensation model (such as flat fees) or impose controls, monitor compliance, and address the bureau’s fair lending concerns. The same memo reveals the bureau’s directors believed they already had the facts they needed to back their claims.
The April 3, 2013, memo lists the CFPB’s former point man to the auto finance industry, Rick Hackett, and 15 other bureau officials as recipients, including Hackett’s colleague in the regulator’s Office of Fair Lending, Patrice Ficklin. It documents such outreach efforts as speaking engagements at conferences staged by the Consumer Bankers Association (CBA) and the American Financial Services Association (AFSA) prior to the release of its guidance. Officials appeared to hope supervision and enforcement alone could move the industry away from markups. But the memo lists another option: a “market-tipping consent order.”
“… Our outreach strategy, including the compliance bulletin, is aimed at spurring similar marketwide changes,” the memo states, in part. “These actions may ultimately prove successful in generating a coattail effect, but another enforcement-based option is to attempt to enter into a consent order with several auto lenders, enough to tip the market away from discriminatory practices in particular, or markup more generally.”
The memo called for the bureau to put several top indirect auto finance sources “under examination or investigation” and makes clear directors were aware of the risks associated with their strategy, particularly potential challenges to its data integrity. Another bulletin, issued a year earlier, mentions neither “rate markups” or “auto finance” but does state that the bureau agreed with a 1994 policy statement on discrimination in lending by 10 federal agencies, including the U.S. Department of Justice (DOJ). It recognized three methods of proving lending discrimination under the ECOA, including evidence of disparate impact — a legal theory that says a policy can be deemed discriminatory if it has an adverse impact on a protected group, even if unintentional.
Because the ECOA prohibits dealers and auto finance sources from collecting race information, the CFPB needed a method for identifying the ethnicity or race of the borrower. So in 2012, the bureau hired Dr. Bernard Siskin and his statistical analysis firm, Philadelphia-based BLDS LLC, to adapt a statistical proxy method originally developed for health outcome research.
The approach, known as the “Bayesian Improved Surname Geocoding” proxy method, combines surnames with geographical information into a single proxy probability for race and ethnicity. The bureau, however, kept its mathematical formulas, computer code, and other information related to its methodology under wraps until Sept. 17, 2014, when, in response to congressional pressure, it issued a white paper that explained some aspects of its methodology.
“Finally, publicizing our methodology in the short term opens our methodology up to attack and further questions,” Ficklin wrote in a June 2013 draft memorandum. “News reports are already labeling it as racial profiling and junk science, and these aspersions may increase if we reveal greater specificity (although those attacks may continue regardless).”
Questioning the Methodology
To construct its proxy, the bureau first calculated an applicant’s probability of belonging to a specific race and ethnicity by comparing the individual’s surname and address with U.S. Census Bureau data to determine the likelihood the borrower belongs to one of six racial and ethnic categories. Analysts then used Bayes’ theorem to determine each applicant’s “composite probabilities.”
Internal CFPB documents reveal the reason the bureau refused to release details of its methodology early on.
“[We] have reason to believe that our proxy [methodology] is less accurate in identifying the race/ethnicity of particular individuals than some proprietary proxy methods that use nonpublic data,” Ficklin wrote in a May 2013 memo. “These proprietary methods are likely to achieve a greater level of accuracy … but we have chosen not to use them because our use of nonpublic data compromises our ability to encourage improvement of compliance management in the nonmortgage lending industry.”
A month earlier, another memo detailed a meeting with an unnamed finance source in December 2012. Its directors were told that, based on disparate impact statistics, they had violated the ECOA. But the finance source, which was later allowed to quietly enter into a Memorandum of Understanding with the bureau, responded by hiring Siskin to represent it before the CFPB. When he reviewed the data set analyzed by the bureau, he added explanatory variables — e.g. metropolitan statistical area and whether the customer financed a new or used car — and found that the disparities almost completely disappeared “after controlling for those variables alone.”
The internal memo also shows that the bureau’s Office of Research (OR) analyzed both approaches and found both were valid and reasonable, “but under different assumptions about the underlying cause of the disparities.” In short, the CFPB’s method assumed that members of different classes experience different outcomes because dealers treat them differently. Siskin’s method underestimated the same racial disparities the bureau’s method seemed intent on overestimating.
“In light of the above, we recommend applying OR’s method rather than adopting the alternative proposed by [Siskin] in proceeding with the upcoming [Potential Action and Request for Response] and [Action Review Committee] determinations, at least for now,” the CFPB memo, penned by Ficklin, read. “… We would add two important caveats. First, the alternative method proposed by [Siskin] is not invalid or unreasonable, and thus could potentially suggest a lower bound in disparities that we should bear in mind as we make decisions on how to proceed in the current auto lending matters.
“Second, OR will continue to evaluate ways of enhancing its method, and additional PARR responses and discussions with other regulators and academics may help identify adjustments or alternative methods for consideration.”
On Sept. 17, 2013, the bureau formally advised Ally that it was preparing to initiate an enforcement action based on a review of more than one million loans the finance source booked between April 2011 and March 2012. The regulator said Ally violated the ECOA by charging African-American and Hispanic borrowers an average of 28.9 and 19.6 basis points more for non-subvented loans and 22 and 14.3 basis points higher for subvented loans than similarly situated non-Hispanic white borrowers, respectively, while Asians paid an average of 21.5 basis points higher for non-subvented loans.
On average, these disparities were expected to cost more than 213,000 consumers $192.32 each over the full life of the loan, totaling more than $41 million in possible direct damages, according to an October 2013 memo in which bureau officials sought former CFPB Director Richard Cordray’s authorization to seek settlement with the finance source.
In that same memo, the bureau acknowledged its proxy methodology and reliance on the disparate impact doctrine could be challenged in court. The document also reveals why the bureau ultimately decided to proceed.
“Ally may have a powerful incentive to settle the entire matter quickly without engaging in protracted litigation,” the memo states, in part. That incentive was Ally’s pending application before the Federal Reserve to change its status from that of a bank holding company to a financial holding company. Maintaining that status, which was set to expire in December 2013, would be difficult if the bureau found that Ally violated the ECOA. And if the Fed didn’t approve its application, Ally would have been forced to divest its insurance and used-car remarketing operations.
But Ally didn’t just roll over. It argued the bureau’s methodology overestimated disparities or produced findings that may not be evidence of discrimination. The bureau was unmoved, citing lack of evidence. So on Dec. 20, 2013, five days before Ally’s deadline to achieve financial holding company status, the DOJ and the CFPB announced a $98 million settlement with Ally Financial and Ally Bank — the largest auto loan discrimination settlement in history. It called for $80 million in compensation for the approximately 235,000 African-American, Hispanic and Asian/Pacific Islander borrowers who paid higher interest rates than non-Hispanic white borrowers between April 2011 and December 2013. Ally was also ordered to pay $18 million to the bureau’s civil penalty fund.
The settlement didn’t have the market-tipping effect the regulator had hoped for. Ally refused to eliminate dealer participation despite the enticement of more favorable settlement terms. As one Ally exec told the bureau, eliminating dealer participation would be “corporate suicide.”
The CFPB knew other finance sources would make similar claims and had anecdotal evidence that eliminating markups would lead to market-share losses. So in December 2014, the CFPB and DOJ made an offer to a group of finance sources: If they collectively imposed stricter markup caps or eliminated markups altogether, the bureau would not impose fines against them. That strategy also failed.
Andrew Koblenz, the NADA’s executive vice president of legal and regulatory affairs and general counsel, would later describe Ally’s decision to retain its dealer participation model as a major turning point. “That spelled the end of the realistic effort to get the world to move to flat fees as a government-mandated policy matter,” he said.
The Industry Strikes Back
Even before the Ally settlement was announced, Republican lawmakers were already questioning the bureau’s lack of transparency regarding its proxy methodology. Their complaints boiled over at a Nov. 13, 2013, Senate Banking Committee hearing, where Sen. Jerry Moran (R-Kan.) grilled Cordray about why the bureau hadn’t divulged its methodology for determining disparate impact discrimination. The director acknowledged the CFPB’s lack of transparency but said they were grappling with jurisdictional issues, including the dealer exemption.
The following day, the bureau hosted an industry forum at its Washington, D.C., headquarters. It consisted of three panels, one of which pitted a single dealer representative against three consumer advocacy groups. Also present were representatives of other federal agencies who confirmed they were also using proxy methodologies to determine whether discrimination exists in lender portfolios.
Rounding out the speaker roster were representatives from the NADA, CBA, AFSA, and the National Association of Minority Automobile Dealers (NAMAD). The representative for the latter, Damon Lester, questioned the bureau’s use of proxies for determining the existence of discrimination, saying, “They’re not particularly useful in proving violations.”
Paul Metrey, the NADA’s vice president of regulatory affairs and chief regulatory counsel, took aim at the bureau’s preference toward flat fees, arguing that the model wouldn’t address the “whatever fair lending risk may have existed for the customer” because dealers would “have the discretion to select the finance source to which to sell the contract.”
Participating on the forum’s indirect auto finance panel was Bill Himpler, AFSA’s executive vice president. He shared details of a meeting the association had with Cordray two months prior to the Nov. 14, 2013, event, noting that the bureau had no concrete idea as to whether or not “the [flat fee] ‘cure’ will kill the patient.” He added that the association discovered that one one-hundredth of 1% of all auto loan sales make it to the CFPB’s consumer complaint database. Himpler said his association would commission an independent study to challenge the bureau’s approach.
“Without this important analysis, it’s like putting the cart before the horse,” Himpler said.
The Next Chapter
The industry was in need of research to counter consumer-advocate studies like the Center for Responsible Lending (CRL)’s 2011 study, “Under the Hood: Auto Loan Interest Rate Hikes Inflate Consumer Costs and Loan Losses.” Among its many claims, which have since been discredited by an analysis conducted by The Washington Post, was that U.S. consumers pay an estimated $25.8 billion in interest rate markups over the lives of their loans.
The study was cited at the CFPB’s forum by a representative of the National Association for the Advancement of Colored People (NAACP), including its claim that dealer markups increased the odds of a loan default or repossession by 12.4% and 33%, respectively. The study was even cited during a debate preceding last month’s House vote on the resolution of disapproval of the bureau’s guidance — nearly three years after an executive with the CRL acknowledged the study “is not a perfect data set.”
AFSA delivered the industry’s response almost a year after Himpler made his promise at the CFPB forum — too late for BMO Harris, which became the first major auto finance source to switch to flats in April 2014. Conducted by Charles River Associates, AFSA’s study concluded that the bureau’s proxy methodology overestimated the African-American population by 41%.
“After analyzing the CFPB’s methodology in detail and reviewing 8.2 million finance contracts, it concluded that the CFPB’s method was ‘conceptually flawed in its application and subject to significant bias and estimation error,” the NADA’s Koblenz told F&I and Showroom in August 2015, a month after BB&T became the second finance source to move to flats.
For its part, the NADA released its Fair Compliance & Policy Program at its annual convention in January 2014. Based on a consent order developed by the DOJ, the voluntary program was designed to address the bureau’s fair lending concerns by having dealers document reasons for variations in rate markdowns. To this day, however, the bureau has never endorsed the program.
During an April 2014 interview with F&I and Showroom, former CFPB official Hackett said the program was a “good step forward for the industry,” but added, “In the wrong hands, it could be a totally useless pretext for what you want to do.”
Hackett had already departed the bureau when Koblenz addressed the F&I industry at the magazine’s September 2015 conference, where he broke down the association’s three-year effort to defend dealer participation into three chapters. The NADA official noted that the industry’s odyssey began in the wake of the 2008–’09 financial crisis and continued through the 2010 discussions preceding the creation and passage of the CFPB-creating Dodd-Frank Act.
In the summer of 2011, the Federal Trade Commission staged a series of roundtables designed to identify consumer protection issues pertaining to auto financing. The debates centered on the notion that dealers shouldn’t be compensated for helping consumers secure vehicle financing, with consumer advocates calling dealer markups nothing more than “surcharges” and “dealer kickbacks.”
The Senate Banking Committee also released a report in 2010, which claimed that dealers “routinely mark up loan rates” above what borrowers would need to pay to qualify for credit. “What the [committee] was telling this new agency is that what dealers get paid is excess,” Koblenz said during his address. “So that’s what was at stake during the Dodd-Frank debates, and that’s what led to the aggressive lobbying and aggressive education that the NADA and the dealers around the country did. And that’s what led to the exclusion of dealers from Dodd-Frank [under the Brownback amendment].”
The CFPB had changed its tune by the time a bureau official addressed finance executives at the 2013 Consumer Bankers Association conference, the representative telling attendees that dealers deserve to be compensated. Koblenz said Chapter One ended at that moment, and Chapter Two ended with the Ally settlement and its refusal to embrace flats. “That brings us to Chapter Three. … Now we’re moving into a situation where they’re trying to improve the model,” he said.
Despite mounting criticism, the bureau and DOJ entered into a consent order in July 2015 with American Honda Finance Corp. to resolve the regulators’ claims that the captive violated the ECOA by charging African-American, Hispanic, and Asian/Pacific Islanders higher dealer markups than non-Hispanic white borrowers over a four-year period beginning in January 2011.
Honda agreed to pay $24 million in restitution and reduce its markup caps from 2.25% and 2% to 1.25% and 1%, respectively. Fifth Third Bank and Toyota Financial agreed to similar deals with the CFPB and the DOJ in September 2015 and February 2016, respectively.
Insiders said at the time that the bureau favored the reduced caps after finding “a considerable reduction or effective elimination” of disparities when there was only a 1% availability of markup across the board. They note that the bureau would take action against a finance source if its methodology uncovered a 10-basis-point markup disparity when comparing similar deals with identical terms.
The Tide Turns
Koblenz arrived at Industry Summit 2015 about 40 days after the House Financial Service Committee approved H.R. 1737 by a 47-10 vote, with 13 out of the 23 Democrats on the committee voting in favor of the legislation that aimed to repeal the bureau’s dealer participation bulletin and add a few steps to its guidance-writing activities.
The bill, the “Reforming CFPB Indirect Auto Financing Guidance Act,” was approved that November by a 332-96 House vote, with 88 Democrats approving of the legislation. The bill, however, was not acted upon by the Senate before the end of the 114th Congress, but emerging from the effort was a key figure in the fight ahead: Rep. Jeb Hensarling (R-Texas), chairman of the powerful House Financial Service Committee
Five days after the House vote on H.R. 1737, Hensarling and his fellow Republican committee members released a report critical of the bureau’s use of the disparate impact doctrine. Based on findings from a slew of internal CFPB memos, “Unsafe at Any Bureaucracy: CFPB Junk Science and Indirect Auto Lending” revealed that the bureau was aware its methodology was flawed and prone to overestimation, yet pushed forward with its discrimination claims.
Hensarling’s committee released a second report in January 2016. It was critical of the bureau’s handling of the $98 million Ally settlement, noting that some checks had gone to white borrowers. It also charged that the bureau’s theory of liability when it came to finance sources could not meet the legal test of disparate impact.
The reports set the stage for the three-hour grilling Cordray endured in March 2016, during his semi-annual report to Congress. He was asked why the bureau hadn’t complied with Hensarling’s three-month-old request for documents that would reveal how much more Ally’s African-American customers had paid in interest rate markups.
And it wasn’t just Republican asking the tough questions. “When you discriminate [against dealers who adjust interest rates to meet a payment a customer asks for], that is discrimination against African-Americans,” said Rep. David Scott (D-Ga.). “When your rule and your actions deny them access to that car, how are they going to get a job? The unintended consequence is that you’re strangling the poor dealer and denying the very customers that you’re supposedly trying … to help.”
By July, nine industry trade groups rang in the bureau’s five-year anniversary by throwing their support behind the Senate version of H.R. 1737. The bill, S.B. 2663, was introduced by Sen. Moran. That September, Hensarling introduced H.R. 5983, the Financial CHOICE Act of 2016. It contained language found in 73 other CFPB-altering bills introduced during the 114th Congress, including S. 2663 and H.R. 1737.
Critics called the bill an “anthology of the committee’s efforts over the past six years” to defang the CFPB and deregulate the financial services industry. Undeterred, Hensarling’s committee passed the measure six days after he introduced it in the House, but it also ended up sitting in House for the rest of that year. Supporters, however, viewed the legislation as the perfect replacement for the Dodd-Frank Act once Donald Trump won the presidency in November 2016.
The Trump Effect
The CFPB was facing challenges from all sides once Trump took office. On Inauguration Day, Jan. 20, 2017, he issued an executive order calling for a regulatory freeze across all federal agencies. Hensarling’s committee delivered its third critical report on the CFPB the next day. It alleged that Cordray violated federal law governing the agency’s rulemaking activities when the bureau published its “larger participant rule” in 2015 — a move that significantly expanded the regulator’s supervisory authority to encompass nonbank entities that make, acquire or refinance at least 10,000 loans or leases in a year.
The bureau’s single-director structure was also under assault. A federal appeals court hearing a case involving a mortgage company ruled it unconstitutional in October 2016. By March 2017, there was another fight brewing regarding the bureau’s proposed regulation limiting the use of binding arbitration. The CFPB published its final rule banning mandatory arbitration clauses in finance contracts on July 19. Congressional Republicans responded the next day by filing a joint resolution of disapproval under the CRA.
Thanks to the eleventh-hour vote by Vice President Mike Pence to break the U.S. Senate’s 50-50 split vote — the House having approved the resolution on July 25, 2017 — Congress sent the resolution to the White House in October 2017 for Trump’s signature. The effort would prove critical a month later, when the Congressional Accountability Office (GOA) announced that the bureau’s March 2013 guidance falls under the CRA and could be repealed using the same process used to scrap the bureau’s arbitration rule.
Prompted by Sen. Pat Toomey’s March 2017 request to review whether the bureau’s guidance fell under the CRA, the GAO issued its response 16 days after Cordray informed bureau officials via email that he was stepping down amid intensifying calls for Trump to fire the embattled director. But even that decision wasn’t free of controversy.
Cordray formally resigned on Nov. 24 and elevated his former chief of staff, Leandra English, to deputy director to put her in the line of succession. Hours later, the White House named its budget director, Mick Mulvaney, acting director under the Federal Vacancies Reform Act.
English filed suit on Nov. 28, but her request for a restraining order to block Mulvaney’s appointment was rejected three days later. This past January, her request for a preliminary injunction to remove Trump’s appointee was also rejected by a Trump-appointed U.S. District Judge. But the challenge set the stage for an appeal by English, whose case is still ongoing.
Legal insiders said the GAO’s findings gave Mulvaney all the authority he needed to rescind the guidance. Instead, he defanged the bureau’s Office of Fair Lending, which led its examinations of dealer participation, ordered a review of the CFPB’s finalized payday lending rules, called off a four-year investigation into a subprime lender, dropped a lawsuit against a group of four online payday lenders, and initiated an examination of the bureau’s policies and practices.
This past March, a familiar CFPB foe, Sen. Moran, introduced S.J. Res. 57, the resolution of disapproval of the bureau’s guidance under the CRA. It was approved by the Senate by a 51-47 on April 18, setting the stage for the House vote 20 days later and Trump’s eventual approval two weeks after that.
Sensing the end was near, NADA officials held a press briefing ahead of the House vote. David Regan, executive vice president of legislative affairs for the association, called the bureau’s compliance bulletin a rule masquerading as guidance, adding that it was designed to pressure lenders to restructure the $1.1 trillion auto finance market by eliminating “auto loan discounts to dealerships.”
“And this was done without notice and comment and without jurisdiction over dealers,” he said, adding that the bureau’s own data reveals a 20% error rate in its proxy methodology.
“It was only after a very aggressive multi-year bipartisan congressional oversight that we were able to see exactly how flawed the approach was,” Regan added. “The bottom line is there’s no real surprise here. Congress has been on this case for many years. However, I would say this is not a categorical prohibition of future trends to regulate in the area. I think Congress clearly intended when they passed the CRA that this wouldn’t [preclude the bureau from regulating in this area], that it would be an opportunity to send the agency back to the drawing board.”
Join F&I and Showroom's Gregory Arroyo and Hudson Cook LLP's Eric Johnson on Wednesday, June 6 (11 a.m. PT/2 p.m. ET), for Part One of the magazine's compliance webinar series, "Who Will Take Up the CFPB's Torch." To register, click here.