Study: US mortgage algorithms perpetuate racial bias in lending

US mortgage algorithms perpetuate racial bias, costing Black and Latino borrowers up to half a billion dollars more in interest annually compared to white borrowers with similar credit scores, according to researchers. What happened Researchers at UC Berkeley found that both online and face-to-face mortgage lenders charge higher interest rates to Black and Latino borrowers, resulting in significant financial disparities. The study focused on 30-year, fixed-rate, single-family residential loans issued from 2008 to 2015 and guaranteed by Fannie Mae and Freddie Mac. Black and Latino borrowers pay 5.6 to 8.6 basis points higher interest on purchase loans and 3 basis points more on refinance loans compared to White and Asian borrowers, r esulting in USD 250 million to USD 500 million in additional interest payments annually. Why it happened The shift from human bias to algorithmic bias in lending discrimination is attributed to the increased use of machine learning algorithms in mortgage lending decisions. Intended to create fair systems, these algorithms have a disparate impact on minority borrowers. The pricing disparities result from algorithms that use machine learning to target applicants who might shop around less with higher-priced loans. This "algorithmic strategic pricing" may be based on geography or applicant characteristics, potentially exploiting areas with fewer financial services or borrowers with limited access to a range of products. For lenders, th e practice amounts to 11 to 17 percent higher profits on purchase loans to minorities. What it means For Black and Latino borrowers, US mortgage algorithms perpetuate and exacerbate existing wealth gaps between racial groups. The findings raise legal questions about potential violations of US fair lending laws and the rise of statistical discrimination, and highlight the need for more robust oversight and regulation of AI-driven financial decision-making processes. System 🤖 Developer: Quicken Loans Country: USA Sector: Banking/financial services Purpose: Reduce credit risk Technology: Machine learning Issue: Bias/discrimination

Date it happened

2018-11-01

This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.