The Consumer Financial Protection Bureau made it clear that it will continue to target auto finance lenders as one of its top supervisory and enforcement priorities in the Fair Lending Report of the Bureau of Consumer Financial Protection , which was released in June 2019.  In addition to adding student loan origination to its watchdog list the CFPB will target model-use practices in auto servicing debt collection in an effort to more closely monitor discriminatory policies and practices based upon consumer data. The report specifically referenced the use of models that predict recovery outcomes.

In a world of increasingly available consumer data, lenders continue to augment the scope of information they may choose to evaluate to underwrite and service auto loans. Examples of alternative data include:

• Data showing trends or patterns in traditional loan repayment data.
• Payment data relating to non-loan products requiring regular (typically monthly) payments, such as telecommunications, rent, insurance, or utilities.
• Checking account transaction and cash flow data and information about a consumer’s assets, which could include the regularity of a consumer’s cash inflows and outflows, or information about prior income or expense shocks.
• Data that some consider to be related to a consumer’s stability, which might include information about the frequency of changes in residences, employment, phone numbers or email addresses.
• Data about a consumer’s educational or occupational attainment, including information about schools attended, degrees obtained, and job positions held.
• Behavioral data about consumers, such as how consumers interact with a web interface or answer specific questions, or data about how they shop, browse, use devices, or move about their daily lives.
• Data about consumers’ friends and associates, including data about connections on social media.

In the report, the bureau devoted an entire subsection to a modeling discussion in its summary of steps taken to improve access to credit. In fact, the bureau directly acknowledged that “[t]he use of alternative data and modeling techniques may expand access to credit or lower credit cost and, at the same time, present fair lending risks.” The bureau also seemed to acknowledge that part of the purpose for its supervisory activities is to educate the bureau regarding modeling techniques, to “keep pace” with technological advances, and to “learn about the models and compliance systems” available via third-party vendors. In taking a hands-on approach to learning, the bureau can, at the same time, assess fair lending risks to consumers. It seems that education is leading the bureau beyond monitoring data use in credit applications to monitoring data use in all facets of auto finance servicing.

The move is particularly interesting, given the CFPB’s no-action letter to Upstart Network, Inc., which was issued in September 14, 2017, actively monitored in 2018, and referenced in the June 2019 Fair Lending Report. There, the scope of the no-action letter was “limited to Upstart’s automated model for underwriting applicants for unsecured non-revolving credit.” Upstart mixes both traditional underwriting factors, such as credit score and income, with non-traditional data points, such as education and employment history.

The CFPB made clear that its issuance of a no-action letter would not serve as an official endorsement of or expression of the bureau’s views on the use of any particular modeling techniques. While no-action letters are not binding on the bureau, the Upstart no-action letter, in conjunction with the June 2019 Fair Lending Report, seems to indicate that modeling techniques in general will receive heightened scrutiny from the bureau going forward.

Ultimately, it remains to be seen whether the bureau’s exploration into the impact of alternative data on credit access will result in an enforcement action involving model use within auto servicing. Given the bureau’s announcement that the issue is now squarely on its radar, in the least, CFPB investigations seem inevitable.

Notably, the bureau seems to recognize the potential benefits to using alternative data beyond traditional credit file data to provide access or better pricing for those consumers who face barriers to accessing credit or those that traditionally pay more for credit. However, the bureau seems to embrace the idea with caution, ever vigilant to protect nondiscriminatory access to credit, lest the techniques or the data itself present fair lending threats.

Accordingly, to the extent that auto finance companies are using modeling techniques via a third-party vendor or their own proprietary formula in their vehicle recovery processes, they would do well to proactively examine the methods and data used for any potentially discriminatory impact on consumers.