Summary of 28th Annual Conference on Fair Lending and Consumer Financial Protection

Skadden, Arps, Slate, Meagher & Flom LLP

Anand S. Raman Austin K. Brown Darren M. Welch Nicole M. Cleminshaw Andrew Hanson

Year In Review

Anand Raman, the head of Skadden’s Consumer Financial Services (CFS) practice, began the conference by providing a summary of notable events and trends over the past year relating to consumer financial services compliance and enforcement, including enforcement actions by the Consumer Financial Protection Bureau (CFPB) and prudential regulators, statutory and regulatory changes, and activity by state regulators.

Key issues discussed in this session included:

CFPB Staffing Changes and Enforcement Trends. The CFPB hired new Associate Director of Supervision, Enforcement and Fair Lending Bryan Schneider at the end of 2019 and hired new Enforcement Director Thomas Ward early in 2020. The impact that these new hires will have on the CFPB’s enforcement strategy is not yet clear, but the CFPB continues to actively investigate consumer compliance issues and initiate enforcement actions.

Notably, enforcement actions over the past year have disproportionately come in the form of lawsuits, rather than consent orders, compared with earlier years. For example, of the 20 actions that were filed between April 1, 2019, and April 1, 2020, 12 were lawsuits and eight were consent orders. In contrast, between April 1, 2015, and April 1, 2016, the CFPB filed 43 enforcement actions, of which 18 were lawsuits and 25 were consent orders.

On March 3, 2020, the Supreme Court heard oral argument in the matter of Seila Law v. Consumer Financial Protection Bureau to consider two questions: (i) whether the vesting of substantial executive authority in the CFPB, an independent agency led by a single director, violates the separation of powers; and (ii) whether, if the CFPB is found unconstitutional on the basis of the separation of powers, 12 U.S.C. §5491(c)(3) can be severed from the Dodd-Frank Act. The CFPB argues that the single-director structure is unconstitutional, and a decision by the Court invalidating the single-director structure could affect how the Bureau pursues enforcement in the future. The Court’s decision is expected in the coming months.

Prudential Regulators. Anand noted that the prudential regulators also continue to actively examine consumer compliance issues — including in the areas of fair lending and unfair or deceptive acts or practices (UDAP) — with a particular emphasis on nonpublic resolutions of consumer compliance matters. Key issues that the prudential regulators have reviewed over the past year include overdraft fee assessment practices, commercial lending disclosures and broker compensation.

Fair Lending. The CFPB and the federal prudential regulators did not enter into any public fair lending enforcement actions over the past 12 months, although the Department of Justice and the Department of Housing and Urban Development entered into settlements relating to redlining and automobile loan pricing.

Statutes and Regulations. The CFPB proposed revisions to Regulation F, which implements the Fair Debt Collection Practices Act. The proposed revisions would prohibit a debt collector from calling a consumer about a particular debt more than seven times within a seven-day period and from engaging in more than one telephone conversation with a consumer about a particular debt within a seven-day period.

The CFPB also issued a final rule delaying the implementation of the underwriting requirements of its 2017 payday lending rule, which had originally been proposed under the Bureau’s former Director Richard Cordray. The underwriting requirements would have made it an unfair and abusive practice to make certain payday and vehicle title loans without determining that the borrower had an ability to repay the loan, but the CFPB subsequently determined that these “ability-to-repay” provisions would unduly restrict access to credit.

With respect to small business data collection under the Dodd-Frank Act and the Equal Credit Opportunity Act (ECOA), the CFPB agreed, as part of a settlement of litigation against the agency relating to its delay in issuing a rule, to publish an “outline of proposals under consideration and alternatives considered” by September 15, 2020. The CFPB also agreed to convene a Small Business Advocacy Review panel, which would, among other things, issue a report on the subject.

On September 10, 2019, the CFPB issued new or revised policies relating to no-action letters, Compliance Assistance Sandbox approvals and trial disclosure approvals, all of which are administered by the CFPB’s Office of Innovation. The purpose of these policies is to encourage institutions to provide innovative products and services to consumers by addressing or resolving regulatory uncertainty that may be preventing the institutions from implementing a product or service.

State Update. States have aggressively enforced consumer financial protection laws, particularly in light of a perceived reduction in activity by the CFPB, with the New York Department of Financial Services, Massachusetts Attorney General, and California Department of Business Oversight as notable examples. State agencies have taken action across a variety of industries, and have been particularly active in the auto lending space.

Audience questions submitted during the presentation included:

  • What are the fair lending implications of enhanced consideration of deposits under the Community Reinvestment Act (CRA) regulatory proposals from the Office of the Comptroller of the Currency (OCC) and the Federal Deposit Insurance Corporation (FDIC)? Regulators and plaintiffs have at times sought to leverage CRA-related concepts in redlining enforcement actions and litigation, including analysis of how banks have defined their assessment areas. Under the OCC and FDIC proposal, deposits would be used to identify areas where some banks must designate additional assessment areas, which could result in changes in the geographic areas where regulators (and banks) assess redlining risk.
  • Have you seen scrutiny regarding closure of loan production offices in an indirect product line? Regulators have indicated that loan production office locations may be relevant to redlining analyses in some circumstances, but there is little clear guidance in this area, and even less with respect to fair lending analysis of indirect product lines. With respect to indirect product lines in particular, an equally (or more) relevant consideration may be the location of third-party originators, such as loan brokers that do business with the lender.

Consumer Compliance Hot Topics

Skadden counsel Austin Brown and associates Nicole Cleminshaw and Andrew Hanson led a discussion of hot topics in consumer compliance over the past year, including emerging issues regarding fair lending and unfair, deceptive or abusive acts or practices (UDAAP) related to the COVID-19 pandemic. Much of the consumer compliance enforcement activity over the past year has related to UDAAP, with several enforcement actions also relating to the Fair Debt Collection Practices Act and the Fair Credit Reporting Act.

Although there were no public fair lending enforcement actions by the CFPB over the past year, the CFPB entered into a consent order against mortgage lender Freedom Mortgage Corp. for submitting erroneous Home Mortgage Disclosure Act data under Regulation C in June 2019. The presenters pointed out that the Freedom Mortgage Corp. order related primarily to the reporting of race, ethnicity, and sex information, which the CFPB alleged was reported incorrectly on an intentional basis in many cases.

Key issues discussed in this session included:

Compliance With Foreclosure, Forbearance and Other Loss Mitigation Guidance and Regulations. The presenters discussed efforts by federal and state governments to address the hardships created by the COVID-19 pandemic, including requirements and guidelines prohibiting foreclosures and encouraging or requiring loan servicers to offer forbearance and other loss mitigation options to borrowers. Under the CARES Act, for example, a borrower with a federally-backed mortgage loan experiencing a hardship due directly or indirectly to the COVID-19 emergency may request forbearance from the borrower’s servicer, and such forbearance shall be granted where the borrower provides a certification of hardship. Likewise, servicers may not pursue foreclosure processes for federally-backed mortgage loans over the 60-day period that began on March 18, 2020. These and other similar requirements at the state level for nonmortgage loans raise fair lending risk, inasmuch as institutions may not be treating similarly situated customers consistently.

The presenters outlined best practices to potentially mitigate fair lending risk relating to foreclosure, forbearance and other loss mitigation programs, including:

  • making loss mitigation decisions centrally, using objective guidelines;
  • notifying relevant staff, including loan officers or others who might not typically be involved in servicing decisions, regarding the changes, and establishing an internal hotline for questions regarding eligible products and procedures;
  • updating automated loss mitigation models with the new guidelines to test how well they work and to monitor outcomes periodically; and
  • testing the consistency of loss mitigation outcomes.

Risks Related to Implementation of the Paycheck Protection Program (PPP). Implementation of the Paycheck Protection Program has led to a number of compliance issues, including fair lending risks under the ECOA relating to the prioritization of existing customers for PPP loans. In particular, it has been widely reported that some institutions have issued guidelines prohibiting from eligibility loan applicants who do not already have a loan account with the institution. Several reasons have been offered for such a policy, including minimizing the burden on staff, who are already stretched thin by COVID-19 issues and concerns, to compliance with know-your-customer rules.

The presenters highlighted that, depending on the composition of the existing customer base, a customer-only policy could lead to a higher denial rate for minority-owned businesses. In addition, any other underwriting overlays could create denial rate disparities and expose institutions to further fair lending risk. The presenters recommended that institutions carefully monitor underwriting decisions relating to the PPP program to determine whether minority-owned businesses are being adversely affected by policies and procedures. They noted that fair lending enforcement relating to the PPP program could take months or even years to develop, necessitating careful documentation of decisions and attention to treating similarly situated customers equally.

Emerging Fair Lending Issues. Loan pricing continues to be an area of fair lending scrutiny by regulators, particularly with respect to relationship and competitive discounts. In particular, regulators continue to investigate whether institutions are offering competitive price matching and relationship discounts equally to minority and nonminority borrowers. The presenters discussed best practices in this area, including:

  • documenting relationship pricing and competitive offer standards and ensuring personnel are trained on the standards;
  • requiring personnel to document instances where discounts are provided and the specific reason for the discount;
  • periodically reviewing instances of discounts to assess consistency with policy; and
  • periodically reviewing instances where discounts are not provided to confirm accuracy.

A second fair lending hot topic is the enhanced regulatory scrutiny resulting from the banking industry’s shift into digital banking. As banks reduce their physical footprint, fair lending risk may arise inasmuch as branches are closed disproportionately in majority-minority communities, or an institution leaves majority-minority communities altogether. Steps to mitigate fair lending risk relating to the closure of branches potentially include documenting the specific reasons for any particular branch closure and evaluating whether branch closures will disproportionately affect majority-minority communities.

Emerging UDAAP Issues. The presenters discussed several emerging UDAP and UDAAP issues, focusing on the application of the UDAP prohibition to small business lending practices, recent rulemaking relating to the prohibition against abusive acts or practices, and the development of debt collection rules tied to UDAAP. Although the application of UDAP to small business lending practices is not new, regulators have recently focused their enforcement activities on protecting small businesses from unfair or deceptive acts or practices, including with respect to the disclosure of material terms and conditions of contracts. The presenters made the point that enforcement actions in the consumer space may foreshadow small business enforcement in the future, and encouraged participants to consider issues such as payment processing, credit reporting and overdraft practices affecting small businesses through a UDAP lens.

The presenters also described the CFPB’s recent application of UDAAP principles to debt collection practices, including a recent enforcement action alleging that the number of debt collection calls to a borrower constituted an unfair practice.

Audience questions submitted during the presentation included:

  • What are your thoughts on conducting fair lending analysis of Small Business Administration (SBA) PPP loans, and the data points that could be areas of focus? PPP underwriting and pricing rules are standardized, so discretion is already quite limited. However, potential areas of focus include whether the institution is marketing the PPP program throughout its trade area, regardless of the minority composition of any community; whether any overlays are being applied differently to similarly-situated small businesses; and whether any rule prioritizing existing customers is disproportionately affecting minority-owned businesses. Appropriate documentation of the reasons for underwriting and pricing decisions will be a key factor in future fair lending enforcement actions relating to the PPP, and institutions may wish to consider reviewing fair lending compliance through statistical analyses and “second looks” at denials. 
  • Are there other opportunities beyond special purpose credit programs to help disadvantaged customers and communities that comply with ECOA? ECOA prohibits discrimination in any aspect of a credit decision. However, the Special Purpose Credit Program provisions of Regulation B allow institutions to benefit certain classes of persons, subject to specific rules. Any guidelines issued by the SBA regarding the prioritization of loans to minority-owned businesses may be relevant to an analysis of Regulation B compliance. However, the SBA has not yet issued specific guidelines regarding this issue.

Machine Learning and Digital Marketing

Skadden counsel Darren Welch led a discussion regarding machine learning, digital marketing and other emerging technology-based issues in consumer financial services. Machine learning — an advanced computing methodology which helps identify predictive patterns from large data sets without human involvement — has the potential to result in more predictive models, increased access to credit, and better terms and conditions, thereby benefiting the industry and consumers alike. As with other advances in technology, however, machine learning also presents a number of compliance issues — including unique challenges as well as some issues present in traditional models — in the areas of fair lending, fairness and technical compliance. Darren discussed these compliance issues and a number of recommended best practices.

Key issues discussed in this session included:

Fair Lending Testing. With the prevalence and complexity of machine learning models used for underwriting and other purposes, it is important for companies to assess the adequacy of fair lending testing methodologies. Two key types of fair lending testing are (i) to identify potential less discriminatory alternatives and (ii) to assess whether variables may serve as a close proxy for a prohibited factor. While regulators have not clearly articulated expectations regarding fair lending testing methodologies for machine learning models, important questions include how to determine whether an alternative is materially better than the challenged practice, whether an alternative must be considered if it results in some drop in model performance, and how to assess different and conflicting impacts of an alternative on different borrower groups. Skadden has worked with clients to develop metrics and protocols to address these questions.

Explainability. It is important that lenders understand how complex models work, and that they are able to explain to consumers the reasons why a model resulted in an adverse outcome. In addition, to mitigate fair lending risk, lenders may wish to ensure that they can provide intuitive reasons as to why nontraditional data elements are predictive of risk.

Nontraditional Data. The use of nontraditional data elements — i.e., data not found in traditional credit bureau reports or reported by the consumer on the application — in lending models has the potential to expand access to credit in some circumstances, as regulators have indicated. However, these data elements can present elevated compliance risk, and it is important to consider appropriate fair lending risk management, including carefully reviewing data elements, documenting the rationale as to why nontraditional data elements are predictive, fair lending testing for alternatives and proxies, and considering other options to mitigate fair lending risk resulting from third-party models.

Digital Marketing. Recommendations in recent regulatory guidance regarding internet marketing include monitoring audiences reached by marketing, understanding third-party algorithms, carefully reviewing geographic filters and offering consumers the best products for which they are eligible. Regarding marketing through social platforms, including Facebook, while certain targeting options have been eliminated for credit models, other algorithms used by social media platforms that cannot be controlled by advertisers may consider prohibited basis variables, and Darren recommended that lenders consider whether and how they use such platforms for marketing.

Audience questions submitted during the presentation included:

  • Describe model “differential validity” concerns. Fair lending testing often focuses on the outcomes for different borrower groups, depending on how the model predicts risk for those groups. A model may predict risk better or worse for certain borrower groups, e.g., it may overpredict or underpredict risk for minority applicants relative to how well it predicts risk for nonminority applicants. In that situation, the model could be said to have a “differential validity” for different borrower groups, which can present general fairness and fair lending issues. If a model significantly overpredicts risk for a borrower group, this could potentially present fair lending concerns even if there is no disparity in outcomes.
  • What are the limitations of the CFPB’s no-action letter process? The CFPB’s prior no-action letter policy involved several limitations, including the length of time to obtain the no-action letter, an expiration period on no-action letters, and a number of conditions such as continued fair lending testing. The CFPB has attempted to address some of those concerns in its revised no-action letter policy. Even if a lender is able to obtain a no-action letter from the CFPB for a practice, the company could still be subject to criticism or challenge, including by community groups and private litigants, and thus lenders may wish to maintain elements of their compliance risk management programs for covered practices.
  • Do regulators expect fair lending monitoring and analysis of machine learning decision models, including marketing models? Regulators have indicated that they expect lenders to engage in testing of models, as appropriate, and this applies to machine learning models. With respect to so-called “black box” models, where a lender may not have visibility into the model, fair lending risk management can be more complex and may include different steps, including fair lending certifications, third-party review and outcomes testing in coordination with the model developer. Testing of marketing models, in particular, generally focuses on penetration in different geographic areas or to borrowers of different actual or estimated race and ethnicity groups, for example. Recent regulatory fair lending guidance recommends that lenders monitor the audience reached by marketing efforts.

This memorandum is provided by Skadden, Arps, Slate, Meagher & Flom LLP and its affiliates for educational and informational purposes only and is not intended and should not be construed as legal advice. This memorandum is considered advertising under applicable state laws.

BACK TO TOP