Measuring Bias in Consumer Credit (joint with Will Dobb ie, Daniel Paravisini, and Vikram Pathania)
Abstract:
This paper tests for bias in consumer lending decisions using administrative data from a high-cost lender in the United Kingdom. We motivate our analysis using a simple model of lending, w hich predicts that profits should be identical for different groups at the margin if loan examiners are unbiased. We identify the profitability of mar ginal applicants exploiting variation from the quasi-random assignment of l oan examiners. We find significant bias against non-native and older loan a pplicants when using the firm's preferred measure of long-run profits. In c ontrast, there is no evidence of bias when using a short-run measure used t o evaluate examiner performance, suggesting that our results are due to the misalignment of firm and examiner incentives. We conclude by showing that a decision rule based on machine learning predictions of long-run profitabi lity can simultaneously increase profits and eliminate bias.
DTSTAMP:20260406T000042Z DTSTART:20180718T130000Z DTEND:20180718T140000Z SEQUENCE:0 TRANSP:OPAQUE END:VEVENT END:VCALENDAR