Abstract: Paper No. 167

Measuring the Incidence of Discrimination

Jan Ondrich, Stephen L. Ross, and John Yinger

December 1994 

Fair housing audits are a key tool for measuring the incidence of racial and ethnic discrimination in housing markets. Traditional nonparametric measures of incidence count the share of cases in which the majority auditor is favored (simple gross measure) or the difference between the shares of majority- and minority-favored audits (simple net measure). The problem is that these simple measures are far apart, and are often criticized as being an extreme overestimate (gross measure) or underestimate (net measure) of the true incidence of discrimination. This paper presents a bivariate probit model of agent behavior with random effects and then employs the estimated parameters to calculate a parametric measure of the incidence of discrimination. These estimates demonstrate that the simple net measure is typically an underestimate of discrimination but that the simple gross measure may be reasonable under certain circumstances.

For more information on ordering this paper, please contact the Publications Officer, Center for Policy Research, 426 Eggers Hall, Syracuse University, Syracuse, New York 13244-1020 or e-mail our Publications Officer at puboff@maxwell.syr.edu.