| Abstract: | Lenders are assumed to use formal credit scoring schemes in order to evaluate borrower credit worthiness. Variables used in these schemes may be measured with error resulting in credit scores which include the effects of biased parameter estimates, and in lending decisions that appear to be discriminatory although lenders are not prejudiced. Regulations which restrict the information used in credit scoring schemes may produce undesirable credit supply results. Theoretical models are supplemented with illustrative empirical analysis of mortgage lending in which use of information on property location is prohibited. The empirical results indicate that the quantitative impact of such regulations is modest. |
