Agreement Marks New Standard for Fair Lending Oversight in Financial Technology
December 1, 2020 | WASHINGTON, D.C. — Today the NAACP Legal Defense and Educational Fund, Inc. (LDF) and the Student Borrower Protection Center (SBPC) announced that they have entered into an agreement with Upstart Network, Inc. under which the parties will collaborate on a review of Upstart’s fair lending outcomes and assess best practices in the use and testing of alternative data in fintech credit models.
In recent years, regulators, lawmakers, civil rights organizations, and consumer advocates have expressed concern that the use of educational criteria in underwriting can lead to discriminatory outcomes. These risks particularly can affect communities of color, including students who attend minority-serving institutions such as Historically Black Colleges and Universities (HBCUs) and Hispanic Serving Institutions (HSIs). LDF and SBPC have raised these specific concerns with Upstart, which utilizes education data and other alternative variables in underwriting.
Under the agreement announced today the civil rights law firm of Relman Colfax will monitor and conduct regular public reporting over a two-year period to assess fair lending implications from Upstart’s underwriting model and to educate the public on policy issues surrounding AI underwriting models and how alternative data can be best used to promote equitable access to credit. As more companies experiment with and deploy alternative data and machine learning algorithms in underwriting, the work accomplished under this agreement will serve as a guidepost to align industry underwriting and fair lending testing practices with the goals of equity and financial inclusion.
“As our work has shown, lenders’ use of alternative data—including data related to borrowers’ higher education—can reinforce systemic barriers and result in discrimination,” said SBPC Executive Director Seth Frotman. “This agreement offers a model for companies to protect borrowers from the discriminatory effects of these practices and ensure lenders cannot repackage age-old discrimination that has locked Black borrowers out of consumer credit markets. As more companies experiment with and deploy machine learning algorithms to facilitate lending determinations, we believe the work accomplished under this agreement will serve as a guidepost to align industry practices with the goals of equity and fair credit access.”
“This novel agreement represents one of the first instances where a private algorithm used in the consumer lending context will be monitored in an ongoing, systematic way for bias,” said Ashok Chandran, Assistant Counsel at LDF. “We are hopeful that this agreement will serve to eliminate any racial bias in Upstart’s underwriting algorithm, as we have raised concerns about in previous communications with the company. More broadly, we hope this agreement will have broader implications for the lending industry. This includes the establishment of clearer standards, approaches, and best practices for fintech lenders that rely on algorithmic decision-making processes or alternative data that affect access to credit for Black communities.”
“This monitoring agreement is reflective of LDF’s historic commitment to promoting educational equity for Black students and other students of color, particularly those attending HBCUs or HSIs,” said Liliana Zaragoza, Assistant Counsel at LDF. “Through this agreement, we look forward to creating a roadmap to ensure that new technologies do not undermine Black people’s equal access to education by ensuring Black borrowers are not penalized for choosing HBCUs.”
“The consumer credit system leaves far too many Americans out. We founded Upstart to solve that problem, and so we’re excited to partner with thought leaders on fairness in credit access. Their perspective has already helped us make several improvements earlier this year. We welcome all ideas to help us achieve our mission, and hope the transparency of this process can lay out a path for others interested in making credit more fair and accessible for all,” said Paul Gu, Co-founder, Senior Vice President of Product and Data Science at Upstart.
###