About | HeinOnline Law Journal Library | HeinOnline Law Journal Library | HeinOnline

LOG IN
LOG IN


18 J. Legal Analysis 1 (2026)

handle is hein.journals/jlegan18 and id is 1 raw text is: Journal of Legal Analysis, 2026, 18, 1-18 https://doi.org/10.1093/jIa/laaf0l3 Article Differential validity in fair lending Spencer Carol, Talia Gillis *, Scott Nelson3 'Independent, United States of America 2Law, Columbia University, United States of America Finance, The University of Chicago Booth School of Business, United States of America *Corresponding author. Talia Gillis, Law, Columbia University, 435 west 116th Street, New York, NY 10027. E-mail: tbg2117@columbia.edu Abstract Fair lending's disparate impact doctrine aims to address lending disparities. But which disparities? Traditional fair lending has narrowly focused on equal outcomes-examining differences in loan approval rates or interest rates. However, this singular focus overlooks other dimensions of disparities that are essential for fair credit access. This article challenges the conventional emphasis on equal outcomes, demonstrating how it has failed to address deep-rooted inequalities in traditional credit allocation while also stifling innovation in machine-learning and alternative data. We argue that disparities in the validity of creditworthiness predictions-the accuracy with which a model identifies creditworthy applicants-importantly impact equal access to credit and, in particular, the extension of credit to the creditworthy. Despite mounting empirical evidence of the harm of validity disparities, traditional fair lending enforcement inadequately recognizes this disparity dimension, a gap that may become increasingly harmful as lending decisions rely on advanced statistical methods. Future regulatory guidance, enforcement, and supervision should explicitly recognize validity inequalities across protected groups while addressing the accompanying challenges of this more comprehensive perspective on disparities, which is essential for equitable credit allocation. 1. INTRODUCTION There is emerging consensus that antidiscrimination law is in urgent need of reform (e.g. Barocas and Selbst 2016; Kim 2017; Huq 2019; Ho and Xiang 2020; Yang and Dobbie 2020; Gillis 2022; Kim 2022; Starr 2023). Whether due to the increasing use of alternative data (Kaul 2021) or machine-learning tools (FinRegLab 2021), fair lending's disparate impact framework is ill-equipped to achieve fair outcomes in modern underwriting (Gillis and Spiess 2019; Aggarwal 2021; Hurlin et al. 2024; Sargeant 2023). At the same time, the algorithmic fairness literature, originating in computer science and statistics, has highlighted the many ways statistical predictions can result in and perpetuate disparities across groups (Dwork et al. 2012; Hardt, Price and Srebro 2016; Kusner et al. 2017; Chen et al. 2019), as well as the inherent tensions between competing fairness definitions (Chouldechova 2017; Kleinberg et al. 2017; Friedler et al. 2021). However, much of this literature remains disconnected from the practical realities of specific decision-making contexts, particularly fair lending, and rarely addresses how law © The Author(s) 2026. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (https://creativecommons.org/licenses/by-nc/4.0), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact reprints@oup.com for reprints and translation rights for reprints. All other permissions can be obtained through our RightsLink service via the Permissions link on the article page on our site-for further information please contact journals.permissions@oup.com.