Mortgage Performance and Housing Market Discrimination,” Cityscape

This analysis1 examines claims that observed patterns of mortgage default rates by race can be used to make inferences about racial discrimination in household access to mortgage credit. The analysis concludes that observed higher default rates for black households provide no evidence at all about discriminatory treatment in mort-gage lending. In a series of well-crafted empirical papers, Berkovec, Canner, Gabriel, and Hannan (BCGH, 1993, 1994) have provided new evidence about determinants of default on resi-dential mortgages and about loan losses arising from defaults. These studies were based upon previously unexploited data sources maintained by the U.S. Department of Housing and Urban Development (HUD) on individual loans insured by the Federal Housing Ad-ministration (FHA). A large sample of FHA loans was matched to borrower characteris-tics by HUD and to neighborhood (census-tract) characteristics by the researchers. Empirical analyses based on these data document the importance of loan-to-value ratios, borrower income and assets, demographics, and neighborhood housing market character