On February 16, 2017, the Consumer Financial Protection Bureau (CFPB or the Bureau) held a field hearing exploring the benefits and risks of using "alternative data" to calculate individual credit scores.
Individuals who have never had a mortgage, credit card, or other type of loan product may not have a credit score. These "credit invisible" consumers may be unable to prove creditworthiness to lenders, and thus cannot begin to build credit – a negative cycle that keeps them from being able to make necessary purchases in time of need. The CFPB estimates that 26 million Americans are credit invisible, while another 19 million have a credit history that is too old to be used by credit bureaus.
The Bureau believes that tapping into alternative sources of data may help such consumers build payment histories and gain access to credit. Such "alternative data" sources could include bills for mobile phones, rent payments, utilities, and electronic money transmissions. These types of data come from a wider variety of sources than traditional credit data and may provide consumers with additional opportunities to develop credit history.
However, the use of alternative data also raises potential concerns related to consumer privacy and discrimination. Questions such as whether the collection and use of such data will be subject to requirements under the Fair Credit Reporting Act (FCRA) and the Equal Credit Opportunity Act (ECOA) were discussed at length during the event.
The CFPB's February 16 field hearing featured remarks from CFPB director Richard Cordray, a discussion panel with participants from industries and consumer advocate groups, and a short public comment period. The Director also announced that the CFPB was releasing a formal Request for Information (RFI) regarding the use of alternative data and modeling techniques to solicit additional views on these issues.
Remarks of CFPB Director Richard Cordray
In his prepared remarks, CFPB Director Richard Cordray emphasized that the three goals of the Bureau's request for public commentary on alternative data were to gain more information on (1) the risks involved, (2) the added complexity of using alternative data, and (3) the effects on privacy and transparency.
Director Cordray asserted that the aim of incorporating alternative data into credit scores would be to promote equal access to credit without regard to race, sex, ethnic background, or any other personal characteristics. While alternative data could open more credit opportunities for consumers of underserved demographics, the Director cautioned that alternative data sources could harm the people they are intended to help.
The Director also emphasized that the Bureau will continue its "evenhanded oversight" to ensure that all lenders are playing by the same rules, by applying its oversight to both big banks and startups.
The panel participants included representatives from financial industries and consumer advocate groups. The discussion was meant to build an understanding of the information and modeling typically involved in alternative data use, and to raise ideas and concerns from affected industries and consumers.
Major themes that arose out of the discussion include the following:
- "Mainstream" vs. Social Media Data: The panelists discussed different types of alternative data, separating them into two categories: "mainstream" alternative data, such as rent payments and cell phone bill payments, and "social media" data. Both industry and consumer advocate participants agreed that mainstream alternative data represented the more realistic and reliable data source. For example, consumer advocate participants stated that because rental payments and cell phone payments are generally made with consistency, data on these payments could provide greater access to credit for currently credit-invisible consumers. By contrast, several participants were negative about the ability for social media to provide useful insights into creditworthiness. In particular, panelists were concerned about violations of fair lending laws where credit decisions were made on the basis of factors like a person's activities or friends.
- Regulatory Coverage: When asked what standards should apply to alternative data, a participant representing one of the three major credit bureaus stated that his company considered any data used for credit decisions, traditional or alternative, to be subject to FCRA's requirements. That said, he also stressed the need for careful decision-making prior to commingling alternative and traditional FCRA data. His company applies an extensive internal review prior to commingling data sets in a way that could cause alternative data not previously subject to FCRA to become subject by combining it with data used to produce credit reports.In terms of the coverage of FCRA and ECOA, financial industry representatives also made a particular point of separating data used for credit decisions, which is covered by both laws, with data used in identity theft and fraud prevention products. Even where data is derived from the same sources, industry participants argued that such data is not subject to FCRA and ECOA where it is not used to make a credit decision, but is instead used prior to underwriting in products designed to confirm the identity of a credit applicant.
- Cautious Use: Several participants emphasized the need to exercise caution in incorporating alternative data sources into traditional underwriting models. The credit bureau participant stated that his company has been monitoring the impact of alternative data sets in various test models for over a decade and has only recently begun incorporating such data into products. Both industry and consumer advocate participants stressed careful monitoring to determine whether the use of such data could create underlying improper discrimination in models or cause models to lose accuracy and predictiveness.
- Uses in Small Business Lending: One participant, from a major online small business lender, emphasized the use of alternative data in enabling fintech lenders to make loans to underserved small businesses. Where traditional bank lending has often focused on the personal credit of the small business owner, this company's model focuses on direct access to the small business borrower's bank statements, sales records, accounts receivable, payment processing records, and other financial indicators to form a better picture of the creditworthiness of the business. This participant emphasized transparency and customer engagement as crucial in the use of alternative data.
- No Prohibitions: A participant from a financial industry trade group stressed the need for regulators to avoid stifling innovation through wholesale prohibitions on the use of certain data sources. In contrast to consumer advocate participants, this individual stressed the need for both positive and negative information about consumers to maximize accuracy.
- Accuracy vs. Access: Although there were few disputes between participants, there was some tension among priorities for consumer and industry participants. Both sides agreed that use of alternative data should not create discrimination or otherwise violate applicable law, but consumer advocates were most interested in alternative data as a means of expanding access to credit. Although industry participants also supported alternative data for its ability to increase access, they were equally focused on its ability to improve the accuracy of credit scoring.By contrast, consumer groups did not appear to support the use of alternative data where it increased accuracy in a way that would negatively affect access. In other words, if alternative data accurately (and in a non-discriminatory way) would improve accuracy by clarifying that a consumer was actually a bad credit risk, consumer groups did not appear to support use of that data.
The program's public commentary was brief, but offered an opportunity for consumers to express their thoughts on the use of alternative sources of data to impact their credit scores. One group of consumers announced that according to a study, credit scores increase over time when alternative data is used to increase credit scores. Another individual requested that data accuracy should be taken into consideration when underwriting occurs. Others stated that regulators should be vigilant against the use of alternative data to pull consumers into predatory credit products.