Mortgage loans to nonoccupants as an indicator of racial redlining

Where racial redlining prevents potential residents of a neighborhood from obtaining mortgage loans, a greater number of houses will he sold to investors and a greter number of residents will rent homes owned by such investors. It may be possible, therefore, to measure the extent of redlining by using HMDA data on loans made to nonoccupants. This study models the flow of mortgage credit to nonoccupants in nine MSAs, using traditional economic and demographic variables and variables describing the racial composition of the neighborhood. The percentage of the census tract population that is black has a small but statistically significant coefficient in Los Angeles, Chicago, and Nashville, and the Hispanic population variable is statistically significant in Los Angeles, Chicago, Boston, and Albuquerque. The model explains a high percentage of the variation in mortgage lending to nonoccupants across census tracts and is robust with respect to alternative formulations of the dependent variable, and the independent variables have the hypothesized signs.