Integrating ecological, demographic, and criminological theory, this article examines the role of violent crime and socioeconomic disadvantage in triggering population decline in Chicago neighborhoods from 1970 to 1990. The results show that high initial levels of homicide and increases over time in the spatial proximity to homicide were associated with large losses in total population across 826 census tracts. However, we also observe sharp group differences in patterns for blacks and whites. Although both black and white populations declined in response to high initial levels of homicide and socioeconomic disadvantage, increases in neighborhood homicide, spatial proximity to homicide, and socioeconomic disadvantage were associated with black population gain and white population loss. In discussing these findings, we argue that taking violent crime and spatial processes into account resolves the apparent contradiction between Wilson's depopulation hypothesis and Massey's segregation hypothesis on the increasing concentration of urban poverty.