A study of indirect auto financing commissioned by the American Financial Services Association found that the CFPB’s proxy methodology for measuring disparities in auto dealer reserve is “conceptually flawed in its application and subject to significant bias and estimation error.”  Among the study’s other key findings was that the CFPB’s preferred alternative dealer “compensation” methods, namely the use of a fixed fee, fixed percentage of the amount financed, or hybrid of the two, may increase the cost of credit for consumers.

Conducted by Charles River Associates, the study used an array of industry data and a data base consisting of approximately 8.2 million new and used motor vehicle retail installment contracts originated during 2012 and 2013.  It measured disparities in dealer reserve using the Bayesian Improved Surname Geocoding (BISG) methodology used by the CFPB in supervisory examinations to proxy for race and ethnicity.  (The CFPB explained its use of the BISG proxy method in a white paper that accompanied the release of the CFPB’s proposed larger participant rule for the auto financing market and a special edition of Supervisory Highlights describing the CFPB’s fair credit supervisory activity in “the indirect automobile lending market.”)

The study’s other key findings were the following:

  • When appropriately considering the relevant market complexities and adjusting for proxy bias and error, the observed variations in dealer reserve are largely explained.  The researchers found little evidence that dealers systematically charge different reserves on a prohibited basis and instead found that reserve variations could “largely be explained by objective factors other than race and ethnicity.”
  • The use of biased race and ethnicity proxies creates significant measurement error, which likely results in overstated disparities.  The researchers found that the BISG proxies overestimated minority population counts and that the use of such proxies “can contribute to inflated estimates of alleged consumer harm.”
  • In consent orders involving indirect financing, the Department of Justice has recognized that dealer reserves depend on objective, observable business factors.  The researchers concluded that failure to consider legitimate business factors for observed disparities increases the potential for reaching erroneous conclusions.  The researchers also noted that these factors are generally unknown to the financial institution and regulators.
  • A portfolio-level analysis of retail installment contracts acquired from different dealerships with different operating models, cost structures, pricing policies, competitive landscapes and the like may create the appearance of differential pricing on a prohibited basis when none exists.  The researchers stated that “[g]iven the highly competitive nature of automotive finance, each financial institution observes the pricing of only a subset of a dealer’s contract portfolio, rather than the entire portfolio.”  The researchers also noted that the assignment of contracts is not random and suggested that “conclusions about dealer compensation patterns cannot be ascertained from the analysis of the contracts assigned to a given individual financial institution.”

The study concludes that the “[f]ailure to consider either competition or pricing complexities allows for the application of an overly-simplistic and biased analytical framework, which leads regulators to pursue overly onerous civil-money penalties from financial institutions.”  In this regard, the study notes that “[g]iven the asymmetric nature of information between dealers and financial institutions, financial institutions and their regulators are in a less than ideal position to evaluate the pricing dynamics of transactions at dealers.”  Notwithstanding these limitations, the researchers conclude, based upon their more refined analysis, that “these pricing dynamics are largely explained by several objective factors, rather than by race and ethnicity.”