Fintech Development from Washington
Federal agencies issue interagency statement on using alternative data in credit underwriting. 
On December 3, 2019, the Federal Reserve Board, the Consumer Financial Protection Bureau, the Federal Deposit Insurance Corporation, the Office of the Comptroller of the Currency, and the National Credit Union Administration (agencies) published a joint statement addressing potential risks and benefits from the use of alternative data in credit underwriting. In the statement, the agencies acknowledge bank and nonbank financial firms’ use or contemplated use of alternative data, which the agencies define as information “not typically found in the consumer’s credit files of the nationwide consumer reporting agencies or customarily provided by consumers as part of applications for credit.”
The statement notes that alternative data may expand access to credit for certain consumers and enable them to obtain additional loan products or more favorable pricing or terms. The statement also explains that a well-designed compliance management program provides for a thorough analysis of relevant consumer protection laws and regulations to ensure that firms understand the opportunities, risks, and compliance requirements before using alternative data.
For additional information, Consumer Compliance Outlook published an article on compliance risks in using alternative data in 2017: Carol A. Evans, “Keeping Fintech Fair: Thinking About Fair Lending and UDAP Risks,” Consumer Compliance Outlook, Issue 2 2017.
The Consumer Financial Protection Bureau (Bureau) issued three policies in September 2019 to promote innovation and facilitate compliance.
The three policies, discussed below, are: (1) the No-Action Letter policy, (2) the Trial Disclosure Program policy, and (3) the Compliance Assistance Sandbox policy. The policies were issued after the Bureau proposed them in 2018 and solicited and reviewed public comments on each from a diverse array of stakeholders. The revised polices are applicable as of September 10, 2019.
The Bureau’s updates its No-Action Letter policy(policy).
In 2016, the Bureau issued its first No-Action Letter policy, under which Bureau staff would issue a statement that they have no present intention to recommend the initiation of a supervisory or enforcement action against a company for providing a product or service under certain facts and circumstances. The Bureau’s 2019 policy streamlines the application review process and makes other changes to the Bureau’s 2016 policy. In reviewing applications, the revised policy states it will consider the potential consumer benefits and risks, how the applicant will mitigate the risks, the statutory and/or regulatory provisions for which the applicant seeks a No-Action Letter, and why a No-Action Letter is needed.
On the same day it issued the new policy, the Bureau issued its first No-Action Letter under the new policy. The Bureau responded to a request by the Department of Housing and Urban Development (HUD) on behalf of more than 1,600 housing counseling agencies (HCAs) that participate in HUD’s housing counseling program. In general, under the No-Action Letter, the Bureau will not make supervisory findings or bring a supervisory or enforcement action under RESPA Section 8 or its authority to prevent unfair, deceptive, or abusive acts or practices under the Dodd–Frank Wall Street Reform and Consumer Protection Act against HUD-certified HCAs that have entered into certain fee-for-service arrangements with lenders for prepurchase housing counseling services.
The Bureau’s new Trial Disclosure policy allows entities seeking to improve consumer disclosures to conductin-market testing of alternative disclosures for a limited time with the Bureau’s permission.
The Dodd–Frank Act gives the Bureau the authority to provide legal protections for entities to conduct certain trial disclosure programs. The Bureau’s Trial Disclosure policy outlines how trial disclosure programs will be approved and some of their terms and conditions. The Bureau has stated that the new policy streamlines the application and review process as compared with its 2013 policy, in addition to other differences.
The Bureau’s Compliance Assistance Sandbox policy allows a provider of a financial product or service facing regulatory uncertainty involving certain consumer protection laws to apply for a safe harbor approval.
The policy provides for the issuance of official staff “approvals” under Truth in Lending Act (TILA), the Electronic Fund Transfer Act (EFTA), and the Equal Credit Opportunity Act (ECOA). Those laws provide protection from liability if a regulated entity acts in good faith conformity with an interpretation or approval issued by authorized Bureau staff. Under the policy, after evaluating a product or service for compliance with the relevant law and other application criteria, Bureau staff can issue an official approval. Compliance in good faith with the terms of the approval will have a “safe harbor” from liability for specified conduct during the testing period.
HUD issued a rulemaking proposal to amend its disparate impact rule.
In August 2019, HUD published a notice in the Federal Register seeking comment on a proposed disparate impact framework for establishing legal liability for facially neutral practices that have unintended discriminatory effects on classes of persons protected under the Fair Housing Act.
In 2015, the U.S. Supreme Court upheld the use of “disparate impact” theory to establish liability under the Fair Housing Act for facially neutral practices that disproportionately affect a protected class without a legally sufficient justification. Texas Department of Housing & Community Affairs v. Inclusive Communities Project, Inc. 135 S. Ct. 2507 (2015). HUD’s proposed rule follows the prior June 2018 advance notice of proposed rulemaking, in which HUD solicited comments on amending the disparate impact standard set forth in the HUD’s 2013 final rule. The rule proposes to replace HUD’s current discriminatory effects standard, (24 C.F.R. §100.500), with a new standard and to provide parties with three methods of defending their algorithmic models to assess factors, such as risk or creditworthiness, where they can show their models achieve legitimate objectives. The deadline for submitting comments was October 18, 2019.
The Bureau published a blog post in August 2019 to provide an update on the Bureau’s first No-Action Letter.
In 2017, under the Bureau’s original No-Action Letter, the Bureau issued a No-Action Letter to Upstart Network, Inc., a company that uses alternative data and machine learning in making credit underwriting and pricing decisions. This No-Action Letter was issued under the 2016 No-Letter policy. Upstart’s underwriting model uses both traditional underwriting data and alternative data. The No-Action Letter applies to enforcing the Equal Credit Opportunity Act (ECOA) and Regulation B for Upstart’s use of alternative data and machine learning in credit decisions and is specific to that entity only.
The recent blog post provides an overview of the outcomes of Upstart’s testing of its underwriting and pricing models with respect to both increasing access to credit and fair lending. The company reported higher rates of loan approval as well as higher rates of approval of minority and other borrowers under its tested model as compared with under a traditional model. For example, the results provided showed that the tested model approved 27 percent more applicants than the traditional model and yielded 16 percent lower average annual percentage rates (APRs) for approved loans.
In addition, Upstart’s analysis of the tested model and the traditional model indicated that the approval rate and APR results for minority, female, and 62-and-older applicants did not yield disparities that would require further fair lending analysis. Upstart will continue to report on its outcomes on the period covered by the No-Action Letter.
NOTABLE RESEARCH
The nonprofit research organization FinRegLab released two studies on the use of certain alternative data in credit underwriting decisions.
The July 2019 report, titled The Use of Cash-Flow Data in Underwriting Credit; Empirical Research Findings , summarizes the research conducted on data provided by nonbank financial services providers that rely on information about applicants’ cash flow to originate consumer and small businesses loans. FinRegLab analyzed the predictiveness of cash-flow variables and credit scores based on loan performance and compared those outcomes with those of traditional scores and variables. Outcomes were also compared with those of models that combine both cash flow and traditional metrics. Where information was available, the report also evaluates the extent to which the companies participating in the study increased access to credit for underserved populations and whether the use of cash-flow metrics increases fair lending risk in credit decisions. The report finds that the cash-flow variables and scores tested were predictive of credit risk and that study participants are serving borrowers who may have historically faced constraints on their ability to access credit.
A second report, titled The Use of Cash-Flow Data in Underwriting Credit: Small Business Spotlight , was released in September 2019. This report focuses on the increased use of cash-flow data in small business lending and on the policy and market issues that affect further future use. The report finds that the use of cash-flow data is being adopted in small business lending more widely than in consumer lending by a variety of credit providers. For example, the report notes that some banks have begun analyzing their existing customers’ deposit data to permit faster underwriting. FinRegLab expects a future release of a third companion report, providing a broader market context and policy analysis for both consumer and small business lending.