Job Market Paper:
Abstract: The welfare costs of crime are disproportionately borne by individuals living in predominately African-American or Hispanic neighborhoods. This paper uses two regression discontinuity designs to show that Federal housing policies established in the wake of the Great Depression make present-day contributions both to this inequity in the distribution of crime within cities and to the overall volume of crime in a city. To stabilize housing markets in the 1930’s, a newly formed Federal agency constructed maps of 239 US cities; these maps purported to grade neighborhoods in terms of lending risk, the riskiest neighborhoods being labeled in red and colloquially said to have been “redlined”. Redlined neighborhoods faced decades of reduced credit access relative to neighborhoods assigned higher grades. First, I use a spatial regression discontinuity design to show that these neighborhood color assignments made in the late 1930s causally influence the present day distribution of crime across neighborhoods in Los Angeles, California. I use a second regression discontinuity design that relies on between-city variation in which cities were mapped to show that these housing policies did not merely redistribute present day crime within cities, but increased the overall volume of city-level crime. An extension of the latter analysis suggests that (a) redlining decreased crime in neighborhoods that were not redlined and that (b) increases in city-level racial segregation were one channel through which redlining increased city-level crime.