Author Description

This study was undertaken to address concerns raised about concentrations of FHA defaults in neighborhoods and among lenders in a 1997 study by the National Training Information Center (NTIC). This report was completed in two phases between December 1997 and September 2000. The first phase culminated with a March 1998 report which examined whether FHA defaults were concentrated among a group of high-default neighborhoods and high-default lenders. The report used statistical analysis to distinguish between patterns caused by chance from those attributable to specific factors such as the loan-to-value characteristics of the loan. After controlling for these factors, the statistical analysis found evidence of non-random default concentrations, at a substantially lower scale than that found in the more limited NTIC analysis. Moreover, the particular neighborhoods and lenders identified as high default changed from year to year, suggesting transitory causes which might not be amenable to policy changes. Finally, it was thought that control for differences in applicant credit histories, which was not possible for the first phase of the analysis, might explain the remaining non-random concentrations.