A United States Fair Lending Perspective on Machine Learning.

Patrick Hall,Arjun Ravi Kannan,Steven Dickerson,Nicholas Schmidt,Raghu Kulkarni,Benjamin Cox

doi:10.3389/frai.2021.695301

Patrick Hall, Arjun Ravi Kannan + Show 4 more

Open Access

https://doi.org/10.3389/frai.2021.695301

Copy DOI

Abstract

The use of machine learning (ML) has become more widespread in many areas of consumer financial services, including credit underwriting and pricing of loans. ML’s ability to automatically learn nonlinearities and interactions in training data is perceived to facilitate faster and more accurate credit decisions, and ML is now a viable challenger to traditional credit modeling methodologies. In this mini review, we further the discussion of ML in consumer finance by proposing uniform definitions of key ML and legal concepts related to discrimination and interpretability. We use the United States legal and regulatory environment as a foundation to add critical context to the broader discussion of relevant, substantial, and novel ML methodologies in credit underwriting, and we review numerous strategies to mitigate the many potential adverse implications of ML in consumer finance.

Highlights

Within the financial services industry, lenders’ use of machine learning (ML) to measure and identify risk in the provision of credit can benefit both financial institutions (FIs) and the consumers and businesses that obtain credit from lenders
Transparency into the intricacies of ML systems is achieved today by two primary technical mechanisms: directly interpretable ML model architectures and the post hoc explanation of ML model decisions. These mechanisms are important in lending, because under Equal Credit Opportunity Act (ECOA)’s implementing regulation, Regulation B, and the Fair Credit Reporting Act (FCRA), the principal reasons for many credit decisions that are adverse to the applicant must be summarized to consumers through a set of short written explanations known as “adverse action notices.”
While questions remain as to which methods will be most useful for ensuring compliance with regulatory requirements, variants of constrained models, Shapley values, and counterfactual explanations appear to be gaining some momentum in the broader lending community (Bracke et al, 2019; Bussman et al, 2019)

Summary

A United States Fair Lending Perspective on Machine Learning

Patrick Hall 1,2*, Benjamin Cox 3, Steven Dickerson 4, Arjun Ravi Kannan 4, Raghu Kulkarni 4 and Nicholas Schmidt 5,6. ML’s ability to automatically learn nonlinearities and interactions in training data is perceived to facilitate faster and more accurate credit decisions, and ML is a viable challenger to traditional credit modeling methodologies. In this mini review, we further the discussion of ML in consumer finance by proposing uniform definitions of key ML and legal concepts related to discrimination and interpretability. We use the United States legal and regulatory environment as a foundation to add critical context to the broader discussion of relevant, substantial, and novel ML methodologies in credit underwriting, and we review numerous strategies to mitigate the many potential adverse implications of ML in consumer finance

INTRODUCTION

Traditional Methods for Identifying Discrimination

CONCLUSION