February 17, 2017
CFPB Evaluates Use of Alternative Data and Modeling Techniques in the Consumer Lending Credit Process
By Obrea O. Poindexter, Oliver I. Ireland, and Crystal N. Kaldjob
On February 16, 2017, the Consumer Financial Protection Bureau (CFPB or Bureau) issued a Request for Information (RFI) regarding use of alternative data and modeling techniques in the consumer lending "credit process." The CFPB defines the "credit process" broadly as all processes and decisions made by a creditor during the full life cycle of the credit product, including marketing, pre-screening, fraud prevention, application, underwriting, account management, credit authorization, setting of pricing and terms, and servicing and collection of debts. According to the Bureau, it is seeking information about the benefits and risks of using alternative data and modeling techniques in consumer lending as the agency begins to consider future activity to encourage responsible use of these techniques and reduce regulatory burdens that could impede their use. The Bureau is also interested in information as it relates to small business lending.
The RFI comes at a time of heightened regulatory interest in and scrutiny of alternative data and modeling techniques and their potential impact on consumers. In November 2016, the Bureau published a RFI seeking public comment on various aspects of data aggregation services, including market practices and potential market developments with respect to companies that offer products and services that use financial account data with the permission of the consumer.
INFORMATION ON TYPES OF ALTERNATIVE DATA AND MODELING TECHNIQUES The Bureau is seeking information on the types of alternative data and modeling techniques, together with information on their stages of development, the extent to which such alternative data and modeling techniques have been adopted, and their uses and impact. The RFI cites several examples of alternative data and modeling techniques, which the Bureau acknowledges do not represent the universe of such data and modeling techniques that are available in the market. The alternative data identified in the RFI include data showing trends or patterns in traditional loan repayment data; data related to payments on non-loan products that require regular payments, such as telecommunications, rent, utility, or insurance payments; checking account transaction and cashflow data; data related to a consumer's stability, such as frequency of changes in residence or employment; educational and employment data (for example, degrees obtained or job positions held); behavioral data and social media connections. The RFI also enumerates several "alternative modeling techniques," which the Bureau defines as non-traditional modeling techniques used in automated credit processes.
INFORMATION ON POTENTIAL BENEFITS AND RISKS TO CONSUMERS Through the RFI, the Bureau seeks information about the potential benefits and risks to both consumers and market participants of the use of alternative data and modeling techniques in the credit process. The Bureau says that alternative data and modeling techniques have the potential to provide consumers with benefits such as
1 2017 Morrison & Foerster LLP | mofo.com Attorney Advertising
greater access to credit, enhanced creditworthiness predictions and lower costs of credit. According to the Bureau, alternative data and modeling techniques also present potential risks to consumers and market participants, such as data accuracy issues, loss of transparency about sources and types of data used, loss of ability to access data or to correct incorrect data, reduced ability to improve credit rating, discrimination and "unintended and undesirable side effects". The Bureau noted that certain alternative data variables could serve as proxies for groups that are protected by federal or state fair lending laws. In addition to information about potential risks, the Bureau requests information about how such risks can be mitigated.
REQUEST FOR COMMENT In general, the Bureau requests information concerning past, current and potential uses of alternative data and modeling techniques, and is interested in the types of data and techniques used, policies and procedures in place to ensure responsible use, comparisons of alternatives with traditional data and modeling techniques and an empirical understanding of current and potential consumer benefits and risks. The categories of information about which the Bureau is requesting information include: (1) alternative data; (2) alternative modeling techniques; (3) potential benefits and risks to consumers and market participants; and (4) specific statutes and regulations. The specific information sought by the Bureau includes the following.
Alternative Data: Information concerning the types of alternative data used in decisions in the credit process (both categories of data and specific data elements or variables used), including, for each, (1) a description of specific decisions for which the data is used, the source of the data, the breadth or coverage of the data, whether the data contain both positive and negative observations, and the quality of the data; (2) the process for deciding whether to use that type of data; (3) whether the data are used alongside other traditional or alternative data; and (4) whether there are types of alternative data that have been evaluated, but are not being used.
Alternative Modeling Techniques: Information concerning (1) the types of alternative modeling techniques used in decisions in the credit process, including a detailed explanation of the modeling technique and how it transforms inputs into outputs; and (2) for each type of technique used, a description of the model development and governance process (including, the specific tests or validations performed to assess compliance with fair lending or other regulatory requirements), a description of any judgmental or subjective or discretionary decisions made in the development phase, and a description of how, if at all, the process handles sample selection, measurement error, and correlations with characteristics prohibited under fair lending laws.
Potential Benefits and Risks to Consumers and Market Participants: Information concerning what the evidence suggests about the potential benefits for consumers in the use of alternative data and modeling techniques to improve risk assessments (such as for more accurate pairings with appropriate credit products), increases in access to affordable credit, lower prices or more convenient decisioning processes and information concerning what the evidence suggests about the potential benefits for market participants of use of alternative data and modeling techniques, such as increased predictability, risk assessments, lower operational costs and competitive advantages. The RFI also requests information concerning what the evidence suggests about the potential risks to consumers and market participants, including information on impacts to consumer privacy, decreased transparency, and lack of historical performance data related to certain alternative data and modeling techniques, and steps to mitigate such risks.
2 2017 Morrison & Foerster LLP | mofo.com Attorney Advertising
Specific Statutes and Regulations: Information concerning (1) the specific challenges and uncertainties that market participants face in complying with the Equal Credit Opportunity Act and Regulation B in using alternative data and modeling techniques; (2) the prevalence of proxying for certain characteristics to test for potential fair lending risks; (3) the use of demographically conscious model development techniques to ensure that models or modeling techniques do not result in unlawful discrimination; and (4) certain other information for those that evaluate models for fair lending risks, including for example, the methods used in evaluating the fair lending risks of alternative data and modeling techniques, and steps used to determine whether alternative data serves as a proxy for a prohibited basis and steps to determine whether alternative data has a disproportionately negative impact on a prohibited basis. Similarly, the Bureau seeks information on the challenges and uncertainties that market participants face in complying with the Fair Credit Reporting Act and Regulation V, including challenges to companies generating, selling or brokering alternative data, the challenges to consumer reporting agencies assembling or evaluating alternative data in implementing accuracy and dispute procedures and disclosing file information and the challenges to lenders when using alternative data. Finally, the Bureau requests information on the challenges and uncertainties market participants face in complying with the prohibition on unfair, deceptive and abusive acts or practice.
The Bureau's agenda with respect to FinTech and emerging financial services is as yet unclear. The November 2016 RFI on consumer access to financial account and account-related data was directly related to Section 1033 of the Dodd-Frank Act, and an area of potential controversy between data aggregators and providers of financial accounts. The concerns related to alternative data and credit modeling techniques are primarily credit underwriting issues and only secondarily a compliance issue. While the RFI highlights the potential benefits, such as greater credit access and lower costs, it also highlights the potential for discriminatory effects. Market participants should closely monitor developments in this space, including the possibility of Bureau action. Comments are due by May 19, 2017.
Obrea O. Poindexter (202) 887-8741 firstname.lastname@example.org
Oliver I. Ireland (202) 778-1614 email@example.com
Crystal N. Kaldjob (202) 887-1687 firstname.lastname@example.org
3 2017 Morrison & Foerster LLP | mofo.com Attorney Advertising
Financial Services Team
Michael J. Agoglia Alexis A. Amezcua Elizabeth Balassone Roland E. Brandel Sarah Nicole Davis Henry M. Fields Joseph Gabai Angela E. Kleine Jim McCabe James R. McGuire Mark David McPherson Ben Patterson Sylvia Rivera Nicholas Alan Roethlisberger Grant C. Schrader William L. Stern Nancy R. Thomas Lauren Lynn Wroblewski
(415) 268-6057 (415) 268-6557 (415) 268-7585 (415) 268-7093 (415) 268-7478 (213) 892-5275 (213) 892-5284 (415) 268-6214 (415) 268-7011 (415) 268-7013 (212) 468-8263 (415) 268-6818 (213) 892-5734 (415) 268-7534 (415) 268-6635 (415) 268-7637 (213) 892-5561 (415) 268-6458
Leonard N. Chanin Rick Fischer Adam J. Fleisher Natalie A. Fleming Nolen Calvin D. Funk Julian E. Hammar Oliver I. Ireland Crystal N. Kaldjob Steven M. Kaufmann
(202) 887-8790 (202) 887-1566 (202) 887-8781 (202) 887-1551 (202) 887-6930 (202) 887-1679 (202) 778-1614 (202) 887-1687 (202) 887-8794
James M. Bergin Meghan E. Dwyer Tiffani B. Figueroa David J. Fioccola Marc-Alain Galeazzi Adam J. Hunt Jessica Kaufman Mark P. Ladner Jiang Liu David H. Medlar Barbara R. Mendelson Michael B. Miller Judy Man Ni Mok Jeffrey K. Rosenberg Mark R. Sobin Joan P. Warrington
(212) 468-8033 (212) 336-4067 (212) 336-4360 (212) 336-4069 (212) 336-4153 (212) 336-4341 (212) 336-4257 (212) 468-8035 (212) 468-8008 (212) 336-4302 (212) 468-8118 (212) 468-8009 (212) 336-4073 (212) 336-4130 (212) 336-4222 (212) 506-7307
Washington, D.C. (continued)
Donald C. Lampe Jeremy R. Mandell Amanda J. Mollo Obrea O. Poindexter Ryan J. Richardson Sean Ruff Trevor R. Salter Nathan D. Taylor
(202) 887-1524 (202) 887-1505 (202) 778-1609 (202) 887-8741 (202) 887-8761 (202) 887-1530 (202) 887-1527 (202) 778-1644
4 2017 Morrison & Foerster LLP | mofo.com Attorney Advertising
About Morrison & Foerster: We are Morrison & Foerster--a global firm of exceptional credentials. Our clients include some of the largest financial institutions, investment banks, Fortune 100, technology and life science companies. We've been included on The American Lawyer's A-List for 13 straight years, and Fortune named us one of the "100 Best Companies to Work For." Our lawyers are committed to achieving innovative and business-minded results for our clients, while preserving the differences that make us stronger. This is MoFo. Visit us at www.mofo.com. Because of the generality of this update, the information provided herein may not be applicable in all situations and should not be acted upon without specific legal advice based on particular situations. Prior results do not guarantee a similar outcome.
5 2017 Morrison & Foerster LLP | mofo.com Attorney Advertising