Current Regulatory Challenges in Consumer Credit Scoring Using Alternative Data-Driven Methodologies

Credit is a crucial determinant of financial success for most US consumers, but not all consumers can access it. This financial exclusion is partially due to traditional credit-risk scoring and approval processes that cannot assess the creditworthiness of “credit invisible” or “thin file” consumers––that is, consumers who do not have enough traditional data depicting their financial payment history. Consequently, some consumer-reporting agencies and lenders turn to alternative data credit-scoring systems as a way to increase financial inclusion. The enormous complexity of these alternative consumer credit-scoring systems, however, raises significant accuracy and transparency issues—most of which stem from their secret, legally protected status—as well as heightened concerns over the use of discriminatory and biased scoring practices using nontraditional behavioral data. If these issues are not addressed, alternative data-driven credit-scoring systems can potentially amplify transparency and discrimination issues, preventing consumers from understanding the factors that impact their credit scores. At the same time, they can position underprivileged groups to face increased discrimination in terms of both accessing credit and receiving favorable interest rates. This Note proposes four regulatory solutions and suggests enhancements to the Model Fairness and Transparency in Credit Scoring Act developed by legal and technology scholars Hurley and Adebayo. The current regulatory framework can better address discrimination by requiring lenders to disclose how they define “creditworthiness” so that consumers can gain a better understanding of the standards to which they are being held. It can also push lenders to foster more appropriate credit standards. Moreover, federal legislation is needed to curtail or prohibit the use of nontraditional behavioral data, especially data derived from a consumer’s social networks, which can unfairly penalize consumers for their social or cultural associations. If regulatory agencies should regulate these firms under the presumption that behavioral data is inherently discriminatory until proven otherwise. Finally, regulators should seek to incentivize firms using alternative credit scoring methodologies to seek no-action letters