InfoBytes Blog
Filter
Subscribe to our InfoBytes Blog weekly newsletter and other publications for news affecting the financial services industry.
SEC charges alternative data provider with securities fraud
On September 14, the SEC announced a settlement with an alternative data provider and one of the company’s co-founders (collectively, "respondents") resolving allegations that the company violated antifraud provisions by engaging in deceptive practices and making material misrepresentations regarding alternative data. According to the order, the respondents understood that companies would share their confidential app performance data if they promised not to disclose it to third parties. As a result, the respondents assured companies that their data would be aggregated and anonymized before being used by a statistical model to generate estimates of app performance. However, the respondents, between 2014 and mid-2018, utilized non-aggregated and non-anonymized data to alter its model-generated estimates to make them more valuable to sell to trading firms. The SEC alleged that the respondents violated provisions of the Exchange Act, such as Section 10(b) and Rule 10b-5 thereunder, because their misrepresentations and other deceptive practices misled subscribers regarding how the company’s intelligence estimates were calculated. The order, to which the respondents consented, imposes civil money penalties of $300,000 and $10 million. The order also provides that the company must cease and desist from committing or causing any future violations of the Exchange Act, and prohibits the co-founder from serving as an officer or director of a public company for three years.
CFPB seeking innovation in adverse action notices when using artificial intelligence
On July 7, the CFPB released a blog post discussing the use of artificial intelligence (AI) and machine learning (ML), addressing the regulatory uncertainty that accompanies their use, and encouraging stakeholders to use the Bureau’s innovation programs to address these issues. The blog post notes that “AI has the potential to expand credit access by enabling lenders to evaluate the creditworthiness of some of the millions of consumers who are unscorable using traditional underwriting techniques,” but using AI may create or amplify risks, including unlawful discrimination, lack of transparency, privacy concerns, and inaccurate predictions.
The blog post discusses how using AI/ML models in credit underwriting may raise compliance concerns with ECOA and FCRA provisions that require creditors to issue adverse action notices detailing the main reasons for the denial, particularly because AI/ML decisions can be “based on complex interrelationships.” Recognizing this, the Bureau explains that there is flexibility in the current regulatory framework “that can be compatible with AI algorithms.” As an example, citing to the Official Interpretation to Regulation B, the blog post notes that “a creditor may disclose a reason for a denial even if the relationship of that disclosed factor to predicting creditworthiness may be unclear to the applicant,” which would allow for a creditor to use AI/ML models where the variables and key reasons are known, but the relationship between them is not intuitive. Additionally, neither ECOA nor Regulation B require the use of a specific list of reasons, allowing creditors flexibility when providing reasons that reflect alternative data sources.
In order to address the continued regulatory uncertainty, the blog post encourages stakeholders to use the Trial Disclosure, No-Action Letter, and Compliance Assistance Sandbox programs offered by the Bureau (covered by InfoBytes here) to take advantage of AI/ML’s potential benefits. The blog post mentions three specific areas in which the Bureau is particularly interested in exploring: (i) “the methodologies for determining the principal reasons for an adverse action”; (iii) “the accuracy of explainability methods, particularly as applied to deep learning and other complex ensemble models”; and (iii) the conveyance of principal reasons “in a manner that accurately reflects the factors used in the model and is understandable to consumers.”
CFPB issues 2019 fair lending report to Congress
On April 30, the CFPB issued its annual fair lending report to Congress, which outlines the Bureau’s efforts in 2019 to fulfill its fair lending mandate. According to the report, in 2019 the Bureau continued to focus on promoting fair, equitable, and nondiscriminatory access to credit, highlighting several fair lending priorities that continued from years past such as mortgage lending, student loans, and small business lending. The Bureau also highlighted three policies released over the last year to promote innovation and to facilitate compliance: the No-Action Letter Policy, the Trial Disclosure Program Policy, and the Compliance Assistance Sandbox Policy (covered by InfoBytes here). Additionally, the report discussed the Bureau’s efforts in encouraging consumer-friendly innovation to expand access to unbanked and underbanked consumers and communities. These include: (i) using alternative data in credit underwriting to expand credit access responsibly; (ii) issuing a request for information on the use of “Tech Sprints” (covered by InfoBytes here) to encourage regulatory innovation and stakeholder collaboration; (iii) continuing to enforce fair lending laws such as ECOA and HMDA, including reaching a settlement with one of the largest HDMA reporters nationwide to resolve HMDA reporting allegations; and (iv) engaging with stakeholders to discuss fair lending compliance, issues related to credit access, and policy decisions. The report also provides information related to supervision, enforcement, rulemaking, and education efforts.
Democratic Congressmen ask GAO to look at alternative data in mortgage lending
On January 16, Democratic members of the House Financial Services Committee sent a letter to the Government Accountability Office (GAO) inquiring about the benefits and drawbacks of using alternative data in mortgage lending, as well as the federal government’s role in overseeing the use of alternative data credit reporting agencies (CRAs) and lenders. The letter notes that while alternative data can be useful in helping lenders identify creditworthy potential borrowers who cannot be scored by CRAs through traditional measures, questions remain about how the use of alternative data may affect compliance with fair lending laws, including the Equal Credit Opportunity Act and Fair Housing Act. “While some alternative data, such as rental payment history, may provide an objective measure of creditworthiness, others might enable discrimination on the basis of a protected class, or infringe upon consumer privacy,” the letter cautions. The letter asks GAO to study the use of alternative data in expanding access to credit, with a particular focus on mortgage credit, and poses the following questions:
- How have different entities used alternative data to expand access to mortgage credit? Specifically, can alternative data determine consumer creditworthiness and whether a consumer is able to repay a mortgage? Additionally, are there certain alternative data sources that are better at predicting creditworthiness or some that are more likely to raise concerns about correlations with discriminatory factors? Furthermore, what federal activity has there been in this space?
- What are the potential benefits and risks associated with using alternative data and financial technology for access to mortgage credit, and are there variations in these benefits and risks across different groups, including minorities and younger borrowers?
- What potential risks does alternative data pose to fair lending compliance, and are the regulatory and enforcement agencies that govern the credit-granting system equipped to manage and prepare for an increased use of alternative data in mortgage lending?
- How do the benefits and trade-offs of other options for expanding access to mortgage credit compare to the use of alternative data in credit scoring?
Regulators issue joint statement on using alternative data in underwriting
On December 3, the Federal Reserve, the CFPB, the FDIC, the NCUA, and the OCC (agencies) issued an Interagency Statement on alternative data use in credit underwriting, highlighting applicable consumer protection laws and noting risks and benefits. (See press release here). According to the statement, alternative data use in underwriting may “lower the cost of credit” and expand credit access, a point previously raised by the CFPB and covered in InfoBytes. Specifically, the potential benefits include: (i) increased “speed and accuracy of credit decisions”; (ii) lender ability to “evaluate the creditworthiness of consumers who currently may not obtain credit in the mainstream credit system”; and (iii) consumer ability “to obtain additional products and/or more favorable pricing/terms based on enhanced assessments of repayment capacity.” “Alternative data” refers to information not usually found in traditional credit reports or typically provided by customers, including for example, automated “cash flow evaluation” which evaluates a borrower’s capacity to meet payment obligations and is derived from a consumer’s bank account records. The statement indicates that this approach can improve the “measurement of income and expenses” of consumers with steady income over time from multiple sources, rather than a single job. The statement also recognizes that the way in which entities use alternative data—for example, implementing a “Second Look” program, where alternative data is only used for applicants that would otherwise be denied credit—can increase credit access. The statement points out that use of alternative data may increase potential risks, and that those practices must comply with applicable consumer protection laws, including “fair lending laws, prohibitions against unfair, deceptive, or abusive acts or practices, and the Fair Credit Reporting Act.” Therefore, the agencies encourage entities to incorporate appropriate “robust compliance management” when using alternative data in order to protect consumer information.
Alternative data model boosts credit access, says CFPB NAL recipient
On August 6, the CFPB published a blog providing an update on credit access and the Bureau’s first-issued No-Action Letter (NAL), and reporting that use of alternative data in underwriting may expand access to credit. In 2017, the CFPB announced its first NAL to a company that uses alternative data and machine learning to make credit underwriting and pricing decisions. One condition for receiving the NAL required the company to agree to a model risk management and compliance plan, which analyzed and addressed risks to consumers and the real-world impact of its service. Through specific testing, the company worked to answer two key questions: (i) “whether the tested model’s use of alternative data and machine learning expands access to credit, including lower-priced credit, overall and for various applicant segments, compared to the traditional model”; and (ii) “whether the tested model’s underwriting or pricing outcomes result in greater disparities than the traditional model with respect to race, ethnicity, sex, or age, and if so, whether applicants in different protected class groups with similar model-predicted default risk actually default at the same rate.”
According to the Bureau, the company reported that in the access to credit comparisons, the alternative data model approved 27 percent more applicants as compared to a traditional underwriting model, and yielded 16 percent lower average APRs for approved loans, with the expansion in access to credit “occur[ing] across all tested race, ethnicity, and sex segments.” For the fair lending testing, the company reported that no disparities were found in the approval rate and APR analysis results provided for minority, female, and older applicants. Additionally, the company reported significant expansion of access to credit for certain consumer segments under the tested model, including that (i) “consumers with FICO scores from 620 to 660 are approved approximately twice as frequently”; (ii) “[a]pplicants under 25 years of age are 32 [percent] more likely to be approved”; and (iii) “[c]onsumers with incomes under $50,000 are 13 [percent] more likely to be approved.” The Bureau noted that the testing results were provided by the company, and the simulations and analyses were not separately replicated by the Bureau.
House Fintech Task Force holds hearing on alternative data
On July 25, the House Financial Services Committee’s Task Force on Financial Technology held a hearing, entitled “Examining the Use of Alternative Data in Underwriting and Credit Scoring to Expand Access to Credit.” As noted by the hearing committee memorandum, credit reporting agencies (CRAs) have started using alternative data to make lending decisions and determine credit scores, in order to expand consumer access to credit. The memorandum points to some commonly used alternative data factors, including (i) utility bill payments; (ii) online behavioral data, such as shopping habits; (iii) educational or occupational attainment; and (iv) social network connections. The memorandum notes that while there are potential benefits to using this data, “its use in financial services can also pose risks to protected classes and consumer data privacy.” The committee also presented two draft bills from its members that address relevant issues, including a draft bill from Representative Green (D-TX) that would establish a process for providing additional credit rating information in mortgage lending through a five-year pilot program with the FHA, and a draft bill from Representative Gottheimer (D-N.J.) that would amend the FCRA to authorize telecom, utility, or residential lease companies to furnish payment information to CRAs.
During the hearing, a range of witnesses commented on financial institutions’ concerns with using alternative data in credit decisions without clear, coordinated guidance from federal financial regulators. Additionally, witnesses discussed the concerns that using alternative data could produce outcomes that result in disparate impacts or violations of fair lending laws, noting that there should be high standards for validation of credit models in order to prevent discrimination resulting from neutral algorithms. One witness argued that while the concern of whether using alternative data and “algorithmic decisioning” can replicate human bias is well founded, the artificial intelligence model their company created “doesn’t result in unlawful disparate impact against protected classes of consumers” and noted that the traditional use of a consumer’s FICO score is “extremely limited in its ability to predict credit performance because its narrow in scope and inherently backward looking.” The key to controlling algorithmic decision making is transparency, another witness argued, stating that if the machine is deciding what credit factors are more important or not, the lender has “got to be able to put it on a piece of paper and explain to the consumer what was more important,” as legally required for “transparency in lending.”
CFPB issues latest fair lending report to Congress
On June 28, the CFPB issued its seventh fair lending report to Congress, which outlines the Bureau’s efforts in 2018 to fulfill its fair lending mandate. According to the report, in 2018, the Bureau continued to focus on promoting fair, equitable, and nondiscriminatory access to credit, highlighting several fair lending priorities that continued from years past such as mortgage origination, mortgage servicing, and small business lending. The Bureau also noted two new focus areas for fair lending examinations or investigations: (i) student loan origination, specifically, whether there is discrimination in underwriting and pricing; and (ii) debt collection and model use, specifically, whether there is discrimination in governing auto servicing and credit card collections, including the use of models that predict recovery outcomes. Additionally, the report highlighted several other Bureau activities from 2018, including, among other things (i) issuing guidance to facilitate the implementation of the August 2018 HMDA final rule (covered by InfoBytes here); and (ii) recommending supervisory reviews of third-party credit scoring models, noting that the “use of alternative data and modeling techniques may expand access to credit or lower credit cost and, at the same time, present fair lending risks.”
U.S. government watchdog studies fintech lending trends, recommends need for clarity on use of alternative data
In December, the Government Accountability Office (GAO) issued a report entitled “Financial Technology: Agencies Should Provide Clarification on Lenders’ Use of Alternative Data,” which addresses emerging issues in fintech lending due to rapid growth in loan volume and increasing partnerships between banks and fintech lenders. The report also addresses fintech lenders’ use of alternative data to supplement traditional data used in making credit decisions or to detect fraud. The report notes that many banks and fintech lenders would benefit from additional guidance to ease the regulatory uncertainty surrounding the use of alternative data, including compliance with fair lending and consumer protection laws. The report’s findings cover the following topics:
- Growth of fintech lending. GAO’s analysis discusses the growth of fintech lending and several possible driving factors, such as financial innovation; consumer and business demand; lower interest rates on outstanding debt; increased investor base; and competitive advantages resulting from differences in regulatory requirements when compared to traditional state- or federally chartered banks.
- Partnerships with federally regulated banks. The report addresses two broad categories of business models: bank partnership and direct lending. GAO reports that the most common structure is the bank partnership model, where fintech lenders evaluate loan applicants through technology-based credit models, which incorporate partner banks’ underwriting criteria and are originated using the bank’s charter as opposed to state lending licenses. The fintech lender may then purchase the loans from the banks and either hold the loan in portfolio, or sell in the secondary market.
- Regulatory concerns. GAO reports that the most significant regulatory challenges facing fintech lenders relate to (i) compliance with varying state regulations; (ii) litigation-related concerns including the “valid when made” doctrine and “true lender” issues; (iii) ability to obtain industrial loan company charters; and (iv) emerging federal initiatives such as the Office of the Comptroller of the Currency’s (OCC) special-purpose national bank charter, fragmented coordination among federal regulators, and the Consumer Financial Protection Bureau's (CFPB) “no-action letter” policy.
- Consumer protection issues. The report identifies several consumer protection concerns related to fintech lending, including issues related to transparency in small business lending; data accuracy and privacy, particularly with respect to the use of alternative data in underwriting; and the potential for high-cost loans due to lack of competitive pressure.
- Use of alternative data. The report discusses fintech lenders’ practice of using alternative data, such as on-time rent payments or a borrower’s alma mater and degree, to supplement traditional data when making credit decisions. GAO notes that while there are potential benefits to using alternative data—including expansion of credit access, improved pricing of products, faster credit decisions, and fraud prevention—there are also a number of identified risks, such as fair lending issues, transparency, data reliability, performance during economic downturns, and cybersecurity concerns.
The GAO concludes by recommending that U.S. federal financial regulators, including the CFPB, Federal Reserve Board of Governors, Federal Deposit Insurance Corporation, and the OCC communicate in writing with fintech lenders and their bank partners about the appropriate use of alternative data in the underwriting process. According to the report, all four agencies indicated their intent to take action to address the recommendations and outlined efforts to monitor the use of alternative data.
CFPB’s Project Catalyst Issues First “No-Action” Letter to Consumer Lending Firm
On September 14, the CFPB’s Project Catalyst initiative issued its first “no-action” letter to a consumer lending firm that provides an online lending platform that uses alternative data when making lending decisions. As previously discussed in InfoBytes, Project Catalyst explores innovation in the consumer financial services sector and examines the potential challenges facing consumers, entrepreneurs, and investors. With the issuance of the no-action letter—at the lender’s request—the CFPB indicated that it does not, at the present, intend to take enforcement action against the lender under the Equal Credit Opportunity Act. However, the letter does not waive the Bureau’s right to choose to “conduct supervisory activities or engage in an enforcement investigation” should the lender fail to comply with the outlined terms. Further, the letter stipulates that the Bureau has the right to evaluate other matters concerning the lender. According to a press release issued by the Bureau, the lender has agreed to “share certain information with the CFPB regarding the loan applications it receives, how it decides which loans to approve, and how it will mitigate risk to consumers, as well as information on how its model expands access to credit for traditionally underserved populations.”
Earlier this year the CFPB issued a request for information seeking input about the use of alternative data, and it believes the information it will receive under the terms of the no-action letter will help to “further its understanding of how these types of practices impact access to credit generally and for traditionally underserved populations, as well as the application of compliance management systems for these emerging practices.” (See previous InfoBytes summary here.)