How Much Does Racial Bias Affect Mortgage Lending? Evidence from Human and Algorithmic Credit Decisions
利用抵押贷款申请的机密数据，研究发现少数族裔申请人信用评分较低、杠杆较高，且更难获得算法批准；可观察风险因素解释了大部分拒绝差异，剩余1-2个百分点的差异可能由未观察到的风险因素解释，表明差别对待的作用比以往研究认为的更有限。
ABSTRACT We assess racial discrimination in mortgage approvals using confidential data on mortgage applications. Minority applicants tend to have lower credit scores and higher leverage, and are less likely to receive algorithmic approval from race‐blind automated underwriting systems (AUS). Observable applicant‐risk factors explain most of the racial disparities in lender denials. Further, exploiting the AUS data, we show there are risk factors we do not observe, and these factors at least partially explain the residual 1 to 2 percentage point denial gaps. We conclude that differential treatment plays a more limited role in generating denial disparities than previous research suggests.