A Profit-Based Measure of Lending Discrimination

TL;DR

This paper introduces a profit-based measure to detect lending discrimination, revealing disparities in loan profits for different groups due to model miscalibration, and discusses fairness trade-offs.

Contribution

It proposes a novel profit-based discrimination measure and demonstrates its application to real-world data, highlighting calibration issues and fairness tensions.

Findings

Loans to men and Black borrowers yielded lower profits.

Model miscalibration caused disparities in credit risk estimation.

Including protected attributes can correct disparities but raises fairness concerns.

Abstract

Algorithmic lending has transformed the consumer credit landscape, with complex machine learning models now commonly used to make or assist underwriting decisions. To comply with fair lending laws, these algorithms typically exclude legally protected characteristics, such as race and gender. Yet algorithmic underwriting can still inadvertently favor certain groups, prompting new questions about how to audit lending algorithms for potentially discriminatory behavior. Building on prior theoretical work, we introduce a profit-based measure of lending discrimination in loan pricing. Applying our approach to approximately 80,000 personal loans from a major U.S. fintech platform, we find that loans made to men and Black borrowers yielded lower profits than loans to other groups, indicating that men and Black applicants benefited from relatively favorable lending decisions. We trace these disparities to miscalibration in the platform's underwriting model, which underestimates credit risk for Black borrowers and overestimates risk for women. We show that one could correct this miscalibration -- and the corresponding lending disparities -- by explicitly including race and gender in underwriting models, illustrating a tension between competing notions of fairness.