On September 14, 2017, the CFPB issued a no-action letter – the first one ever issued by the agency – to a marketplace lender, stating that the agency had no present intention to take enforcement or supervisory action against the lender under the Equal Credit Opportunity Act (ECOA) relating to the lender’s underwriting model, and especially its use of certain alternative data fields.  The letter expires after three years, by its own terms.

As detailed in the lender’s request for a no-action letter, it uses a variety of common credit-bureau-based data fields in underwriting, but also uses other, “alternative data” that seemed to the be focus of the CFPB’s issuance of the letter.  We saw several significant aspects to the underwriting model detailed in the no-action letter request:

  • The requester excludes all residents of West Virginia from its underwriting process. This exclusion apparently did not raise any redlining concerns for the CFPB.
  • The underwriting process also requires that the applicant have a valid email account. We sometimes have been concerned that such a requirement could have a disparate impact (possibly on the basis of age) because it makes the credit product unavailable to those customers who are not digitally engaged, but the CFPB made no comment on this requirement, apparently not finding fault with it.
  • The underwriting process also uses the identity of the college attended by the applicant. This strikes us as very strange and in tension with the CFPB’s repeated attacks on the use of cohort default rate and other school-specific variables in student loan underwriting.
  • The lender also stated that it uses public records (liens and judgments) as part of its underwriting process, which is notable given the removal of such information from credit reports offered by the three large credit bureaus (although this data continues to be separately available). We blogged about this change in the credit bureaus’ inclusion of public records here.

The letter (and the CFPB’s press release) makes the point that the use of this information is being used to expand access to credit to individuals, particularly younger ones, without sufficient credit experience to have a credit score.  We have known for some time that the CFPB has indicated that it will be more understanding about the use of alternative data when it is used to expand access to credit for consumers who do not have sufficient credit bureau history, but it’s unclear how that inclination will be weighed against particular data elements that have been criticized by the CFPB (like, for example, school-specific variables).

Nevertheless, we view the no-action letter as confirmation of our view that the CFPB wishes to encourage the use of alternative data to expand access to credit to “thin file” or “no file” consumers.  The problem is that there are so many limitations in the letter – and so many factual questions that are not addressed by the CFPB at all – that it serves to provide essentially no reliable guidance to the industry.  Moreover, as we detailed on this blog when the NAL process was finalized, a no-action letter does not bind the CFPB, any other federal or state governmental agency, or private litigants, and is not entitled to any deference from courts.  Our criticism of the Bureau’s no-action letter policy was followed by similar comments in the Treasury Report released earlier this year, which noted the “stringent standards that must be met before the agency will even consider a regulated party’s request” and criticized the inaccessibility of the process to industry participants.

For these reasons, the no-action letter still leaves a great deal of uncertainty surrounding other types of alternative data that may be used in underwriting models.  The CFPB could promote access to credit in a much more significant way if it provided guidance on the use of a more expansive set of alternative data than the handful of attributes noted in the no-action letter.  Unless and until that happens, lenders will need to make a judgment about the use of alternative data and equip themselves with empirical data to show the predictive power of the data, and that it is being used to expand access to credit.  That is the focus we have been concentrating on with our lender clients who make extensive use of alternative data in their underwriting models.

An informative new American Banker podcast discusses recent and possible future changes to traditional credit scores, what they mean for industry, and possible industry responses.

The podcast begins with a discussion of changes that will take effect on July 1, 2017 to the public record data standards used by the “Big 3” consumer reporting agencies (CRAs) for the collection and updating of civil judgments and tax liens.  The new standards, which will apply to new and existing public record data on the CRAs’ credit reporting databases, create new verification requirements for data about civil judgments and tax liens, such as certain minimum consumer personal identifying information and a minimum frequency of courthouse visits to obtain new and updated data of at least every 90 days.

The experts participating in the podcast suggested that under current credit score models, the change could result in small credit score increases for impacted consumers averaging about 10 to 11 points.  However, they indicated that credit scores generated by newer credit score models under development that consider other data are unlikely to be impacted.

Among the topics discussed was the potential benefits and challenges in using alternative data in credit score models.  This past February, the CFPB issued a request for information seeking information about the use of alternative data and modeling techniques in the credit process.




The CFPB has issued a request for information (RFI) that seeks information about the use of alternative data and modeling techniques in the credit process.  On March 21, 2017 from 12:00 to 1:00 p.m. ET, Ballard Spahr attorneys will hold a webinar: The New Frontier of Alternative Credit Models: Opportunities, Risks and the CFPB’s Request for Information.  A link to register is available here.

According to the CFPB, the RFI stems from the Bureau’s desire “to encourage responsible innovations that could be implemented in a consumer-friendly way to help serve populations currently underserved by the mainstream credit system.”  The CFPB had signaled the likelihood of future action relating to alternative credit data in a May 2015 report, “Data Point: Credit Invisibles,” that reported the results of a research project undertaken by the CFPB to better understand the demographic characteristics of consumers without traditional credit reports or credit scores.  The report, which the RFI cites, concluded that the current credit reporting system is precluding certain populations from accessing credit and taking advantage of other economic opportunities.

In conjunction with the RFI’s issuance, the CFPB held a field hearing on alternative credit data in Charleston, West Virginia at which Director Cordray gave remarks.  (In a break from its prior practice, the CFPB did not publish advance notice of the field hearing on its website.)

In the RFI’s Supplementary Information, the CFPB states that it not only seeks information relating to consumer credit but, “because some of the Bureau’s authorities relate to small business lending,” it “welcomes information about alternative data and modeling techniques in business lending markets as well.”  To that end, for many of the specific questions asked in the RFI on which the CFPB seeks comments, the CFPB asks commenters to describe “any differences in your answers as they pertain to lending to businesses (especially small businesses) rather than consumers.”  (The CFPB notes the ECOA’s coverage of consumer and business credit and that it has begun the process of writing regulations to implement Dodd-Frank Section 1071, which requires data collection and reporting for lending to women-owned, minority-owned, and small businesses.)  Comments on the RFI must be received on or before May 19, 2017.

The Supplementary Information includes a discussion of alternative data and modeling techniques in which the CFPB provides examples of the types of data and modeling techniques that have been labeled “alternative.”  It also discusses prior research by other federal regulators, such as the FTC’s report on big data.  (The CFPB notes that the non-traditional data that might be used to assess borrower creditworthiness could include “big data.”  To address the growing interest in the use of “big data” and “machine learning” by a wide range of businesses, we recently held a webinar, “Big Data and Computer Learning – Lots of Opportunity and Lots of Legal Risk.”)

In the Supplementary Information, the CFPB lists potential consumer benefits and risks it has identified and states that it intends to use the information gleaned from the RFI’s questions “to help maximize the benefits and minimize the risks” from the use of alternative data and modeling techniques.  The RFI contains 20 specific questions (most of which have numerous subsidiary questions) that are divided into four sections: alternative data, alternative modeling techniques, potential benefits and risks to consumers and market participants, and specific statutes and regulations as they pertain to alternative data and modeling techniques.  The CFPB notes that although each question speaks generally about all decisions in the credit process, “answers can differentiate, as appropriate, between uses in marketing, fraud detection and prevention, underwriting, setting or changes in terms (including pricing), servicing, collections, or other relevant aspects of the credit process.”

The CFPB states in the RFI that it not only seeks to understand the benefits and risks stemming from the use of alternative data and modeling techniques, but “also to begin to consider future activity to encourage their responsible use and lower unnecessary barriers, including any unnecessary regulatory burden or uncertainty that impedes such use.”  We hope the CFPB’s issuance of the RFI reflects its recognition of the complexity of the issues involved in the use of alternative data and modeling techniques and the need for it to carefully consider the interests of all stakeholders.