Persistent Racial Disparities in U.S. Home Mortgage Lending: Evidence of Algorithmic and Systemic Bias from HMDA Data (2007-2016)
- Publicada
- Servidor
- Preprints.org
- DOI
- 10.20944/preprints202508.0113.v1
This study analyzes racial disparities in U.S. home mortgage lending using comprehensive Home Mortgage Disclosure Act (HMDA) data spanning 2007-2016, encompassing nearly one million loan applications. Our analysis reveals persistent and signicant racial disparities in loan approval rates, with Black and American Indian/Alaska Native applicants experiencing substantially higher rejection rates compared to White and Asian applicants. Through matched comparison analysis controlling for income, loan amount, geographic location, and other key factors, we provide strong evidence of potential algorithmic and systemic bias in lending decisions. Black applicants face approval rates 21.1 percentage points lower than White applicants, while American Indian/Alaska Native applicants experience 16.6 percentage point gaps. Most critically, our bias detection analysis demonstrates that even when controlling for identical nancial proles, minority applicants are denied at signicantly higher rates, suggesting discriminatory decision-making processes that cannot be explained by traditional risk factors alone. These ndings have profound implications for housing equity, wealth accumulation, and fair lending policy.