Analysis of Discrimination in Prime and Subprime Mortgage Lending and the Impact of Predatory Lending Legislation

Abstract

The recent development of the subprime mortgage lending market allows us to examine implications of taste-based and information-based theories of discrimination in an important asset class, namely, residential real estate. The existing studies on race-based discrimination in mortgage lending use single equation regressions. At the individual loan level, we show that race is correlated with both observable and unobservable risk variables and is therefore correlated with the disturbance term. To rectify this problem we specify a system of equations and use a novel econometric technique (Full Information Maximum Likelihood) that does not need to identify instrumental variables for system identification. Using a unique data set we find that African Americans and Minorities are no more likely to be rejected than Whites in both prime and subprime markets. This evidence is in contrast to the predictions of either taste-based or information-based discrimination theories. The individual loan level results are robust when we use neighborhood data. We also find that predatory lending legislation, such as the New Jersey Home Ownership Security Act (NJHOSA), which intends to curtail lending abuses in subprime market, results in tighter subprime credit, but generally does not have a significant impact on lending discrimination. These results are supportive of the private interest theory of regulation.