Mortgage Lending Discrimination and Racial Differences in Loan Default

This article analyzes the use of racial differences in loan default to test for mortgage lending discrimination—an approach whose underlying theory has not been carefully explored. The loan approval process creates a selected sample of loans. If mortgage lenders hold minorities to a higher standard, the minority sample should have greater selection bias than the majority sample and so should exhibit lower default rates. This article examines the default approach in this framework and observes that the approach suffers from heteroskedasticity bias caused by loan defaults as a discrete dependent variable. In addition, the correlations between race and the unobservable variables in the loan approval and default equations bias the default approach away from a finding of discrimination. The article concludes that tests based on the loan approval approach should be accepted as the principal technique for determining whether mortgage lending discrimination exists.