CFPB prioritizes fair lending, machine learning and privacy in digital engagement

      Comments Off on CFPB prioritizes fair lending, machine learning and privacy in digital engagement

Members of the auto finance industry continue to have a strong appetite for the development of their digital origination and service platforms. Much of the industry also wants to use data in new and creative ways to make it easier for customers to access credit.

However, the desire for growth and innovation can be tempered by regulatory oversight. Here, we discuss the resurgence of regulatory attention on disparate impact, this time applied to machine learning and agency data that reveals the broad use of digital platforms across most borrower segments, associated updated expectations for data security and identity verification.

Protected classes and disparate impact: a reminder

the Consumer Financial Protection Bureau (CFPB) has always stated that fair lending and credit discrimination will be a priority for the agency. Federal laws such as the Equal Credit Opportunity Act (ECOA) prohibit a creditor from discriminating against any applicant because of race, color, religion, national origin, sex , marital status or age. These attributes are often referred to as “protected status”. This is called “disparate treatment”.

It goes without saying that denying someone a car loan because of their gender, race, or other protected status is clearly illegal. There has been much more ambiguity about whether an underwriting system that is ostensibly neutral – meaning it does not consider any of these protected statuses – also violates credit discrimination laws. depending on the result. This is often referred to as a “disparate impact” policy, a policy that negatively impacts a certain group based on their characteristics.

The CFPB believes that creditors can violate fair lending laws based on this disparate impact theory. Consider the agency’s 2013 bulletin which stated that indirect auto finance companies are subject to the ECOA and can be held liable under the disparate impact theory if discriminatory price disparities are discovered – a position controversy that provoked a strong reaction from many players in the automotive finance sector. In 2018, Congress used the Congressional Review Act to repeal the 2013 CFPB bulletin. Since then, the CFPB has been prohibited from enacting a similar rule or issuing guidelines without the authority of Congress and, therefore, there has been no disparate impact seismic development in recent years.

Disparate impact, treatment and outcomes

Fast forward to 2022: the CFPB recently published a blog post which indicates that it intends to leverage the Consumer Financial Protection Act to target discriminatory practices. While the official name of this law may not be familiar to everyone, most of us in the auto finance industry are familiar with what this law prohibits – Unfair, Deceptive, or Abusive Acts or Practices (UDAAP). The CFPB stressed that it is concerned about both disparate treatment and disparate impact, and that both fall within its mandate to eliminate unfair practices.

Readers will recall that an act or practice is considered unfair if:

  • It causes or is likely to cause substantial harm to consumers;
  • The harm is not reasonably avoidable by consumers; and
  • The harm is not outweighed by compensating benefits to consumers or competition.
© Can Stock Photo / murrstock

Risk assessment: show your work

The CFPB does not mention the auto finance industry by name, but has a different goal: the use of machine learning and artificial intelligence to market and advertise credit. This message coincided with March 16 from the Office announcement that it updated the unfair portion of the UDAAP exam manual to include credit discrimination. The Bureau’s message is that examiners will require supervised companies to show their risk assessment process and discriminatory results in all consumer credit markets.

In practice, the focus on discriminatory outcomes effectively amounts to focusing on disparate impact. In sectors other than auto finance, the CFPB has already taken steps to define options for how computer models that help value collateral are accurate and fair and effectively prevent algorithmic bias. The CFPB has made it clear that it will closely examine how companies rely on automated decision-making models and any potential discriminatory results.

Consumer Behavior Trends

As part of its obligation to provide an annual report to Congress on the consumer credit card market, the CFPB shared its market results on digital engagement. Many points apparently apply to all types of credit, such as auto financing. For example, the CFPB found that data from 2020 shows that more consumers than ever are paying their credit card bills online or using a mobile app. In fact, around 60% of consumers have made at least one online payment by credit card during the year. Paper payments have declined over the same period. While this trend is consistent across all age groups, the data revealed that the use of digital payment by older consumers accelerated over this period as these consumers became more familiar with the technology during the COVID-19 pandemic.

Along with the increased use of digital creative and service platforms, there is an increased risk of fraud. The financial companies that dominate the digital services market, and even those that are catching up, have invested huge resources in their cybersecurity protections and customer authentication practices. Financial institutions active in the digital market – probably almost all readers right now – may want to review the Federal Financial Institutions Review BoardAuthentication and access to the services and system of financial institutions orientation, which was updated in 2021.

There are many valuable elements to this advice, but companies expanding their digital lending programs will want to pay close attention to strategies to mitigate risk in email systems, internet access, customer call centers, and other areas. other entry points to a financial institution’s information system. . Single-factor authentication, for example, is often inadequate, and a layered security structure that changes based on transactional risk is recommended.

As expected, the current leadership of the CFPB has redoubled its attention to equitable access to funding, as evidenced by its publications and actions. As digital advancements and consumer acceptance increase, lenders will want to regularly review their compliance platforms and marketing programs.

Kelly Lipinsky is a Member (Partner) of McGlincheyof Consumer Financial Services Regulation and Compliance Group. She is also an executive member of the firm’s Cleveland office.

The 23rd edition 2022 Big Wheels Auto Finance Data Report is now available, providing exclusive statistics on the auto loan and lease industry and a ranking of the top 200 automotive financiers in the country. Order your report.