Fintech and Gender Discrimination

Abstract

Using data from a lending platform that switched from a human-based to a machine learning-based system, we find that fintech may increase gender discrimination. The rationale is that machine learning algorithms allow the platform to better decipher differences in borrower preferences between female and male borrowers. Specifically, after the switch, the platform assigned higher interest rates and better credit ratings to less price-sensitive female borrowers. These results are not driven by changes in borrower credit risk or lender preferences. Instead, the behavior is consistent with the platform’s attempt to maximize its revenue by applying price discrimination to female borrowers.

Keywords: Fintech, Gender Discrimination, Peer-To-Peer Lending, Machine Learning, Credit Rating

JEL Classification: G41, G51, J16