Contingent Urban Geographies of Violent Crime: Racial Segregation and the Impact of Public Housing in Atlanta

Abstract

Neighborhood variations in crime incidence rates are most commonly interpreted through the lens of social disorganization theory, or a "communities and crime" perspective. This approach typically articulates explanation for crime by focusing on the characteristics of communities—a unitary scale most commonly equated with neighborhoods. We argue that this perspective fails to recognize the importance of broader urban geographic contexts, and offer an extension that sees geographically contingent processes functioning at multiple scales simultaneously. We develop this perspective applied to the "spread effects" of public housing on violent crime in surrounding neighborhoods: these spread effects are conditioned by the nature of the urban contexts through which they operate. Specifically, deeply divided and racialized patterns of residential segregation at least partially define the contexts that condition public housing's effect on crime. We examine our perspective using early 1990s block group data for the City of Atlanta and find substantial evidence in support of our perspective. In particular, we find that Techwood Homes, the nation's first federally constructed public housing project, exerted different geographic spread effects in predominantly White than in predominantly Black portions of the city. By failing to recognize the complexity and contingency of public housing's geographic effect on crime in surrounding neighborhoods, previous approaches substantially overestimate crime in White areas, and underestimate crime in Black areas.