Crime and (Electoral) Punishment: Violence and Voting in the United States

Abstract

What are the relationships between incidences of crime, perceptions of crime, and vote choice? In this paper, I study the 1988 American presidential election (a race marked by its focus on crime) and compare it to the 1984 and 1992 elections (races with their sights set on other issues), ultimately showing an inverse, statistically significant, and unique relationship between local reported violent crime rates and constituents’ perceptions of crime as an important national problem in 1988 alone. Furthermore, across all three elections, I show limited interactive effects on perceptions of crime from crime rates and constituents’ attention to elite cues about crime, and I find no significant relationships between incidences or perceptions of crime and constituents’ presidential vote choices. To make this argument, I estimate a series of linear probability models on data from the American National Election Study and Uniform Crime Reporting Program. Altogether, my results suggest crime does not align with traditionally-conceived theories of public opinion and retrospective voting, and I suggest this results from the issue’s raced nature. I conclude by discussing directions for future research.