Enclaves of Isolation: Violence and Political Participation in U.S. Cities

Does spatial proximity to violence mobilize or depress individuals for political action? Whilemany ethnographic studies have shed light on the various forms of social isolation that characterizehigh-violence American neighborhoods, the democratic consequences of proximate exposureto violence have not been well understood. Merging voter files in U.S. cities and geocoded crimedata, I test whether living in close proximity to sites of homicides affects one’s likelihood of votingin federal elections. Employing a regression discontinuity in time (RDiT) design, I providecausal evidence that close residential proximity to homicide depresses turnout by roughly fourto six percentage points, with the strongest effects observed in plurality-Black block groups andthose involving a Black victim. In further mechanism tests, I analyze 1) foot-traffic data aggregatedby census block groups to examine how violence exposure affects population movement and 2)crime-linked survey data. Together, the analyses provide significant evidence that the unequalpsychological burden of fear, shaped by an individual’s perception of their risk of victimization,may drive social isolation and the observed negative effects. Generally, I consider how persistentlyhigh and spatially concentrated rates of violence in the United States shape patterns of politicalparticipation in race-class subjugated communities and affect democratic health more broadly.