Recruitment seminar.

Abstract: In the post World War II period, most U.S. cities experienced large movements of population from the city centers to the suburbs. In this paper we provide causal evidence that this process of suburbanization can be explained by the rise of violent crime in city centers. We do so by proposing a new instrument to exogenously predict violent crime. This instrument uses as time variation the U.S. national levels of lead poisoning. Cross-sectional variation comes from a proxy for soil quality, which explains the fate of lead in soil and its subsequent bioavailability. Using data for more than 300 U.S. cities, results show that the increase in violent crime from the level in 1960 to its maximum in 1991 decreased the proportion of people living in city centers by 15 percentage points. This increase in crime moved almost 25 million people to the suburbs. As a result of suburbanization, we find that people remaining in the city center are more likely to be black people, consistent with the “white flight" phenomenon. We then demonstrate that this suburbanization process had aggregate effects on the city. Exploiting a spatial equilibrium model, we determine that violent crime had externalities on productivity and amenities.