Issue Ownership and Local Context: How Crime Shapes Law Enforcement Elections

• In Event: Poster Session: Urban Politics & International Order, Environmental Issues, and IPE
  In Poster Session: Poster Session: Urban Politics

Abstract

Since the 1960s, the Republican Party in the United States has embraced the tagline of “law and order”. The literature on issue ownership (Budge and Farlie, 1983; Petrocik, 1996) would trace this link between the Republican Party and “law and order” politics to voters’ belief that Republican candidates are best equipped to address the issue of crime. The theory of retrospective voting (Fiorina, 1977, 1981) would posit that violent crime rates are one metric voters can use to assess a candidate’s expected performance in office (Arnold and Carnes 2012; Hopkins and Pettingill 2018). Yet one fundamental gap exists in this literature of issue ownership and retrospective voting as it pertains to “law and order” politics: do voters make distinctions among different levels of government and elected offices when assessing a candidate’s expected performance in dealing with crime? Arceneaux (2005, 2006) finds that citizens do make meaningful distinctions among different levels of government in assessing the performance of incumbents but that these attitudes only influence vote choice when it comes to “easily accessible issues.”

To our knowledge, previous research has not put this question to the test when it comes to the issue of law and order, and limited work tests the strength of retrospective voting in local politics (see Berry and Howell 2007; Hopkins and Pettingill 2018). In other words, we currently know little about whether voters hold politicians accountable for local conditions to different degrees across levels of government. More specifically, we do not know the extent to which voters hold two key elected officials responsible for prevailing crime rates: sheriffs and prosecutors.

The literature on issue ownership and the history of law and order politics would lead us to expect that increases in local crime rates are associated with greater support for Republican candidates. Furthermore, given that sheriffs and prosecutors are often the most visible embodiment of law and order policies (Soss and Weaver 2017), a logical expectation is that evidence of retrospective voting based on change in violent crime rates would be strongest in elections for sheriff and prosecutor as opposed to elections for higher offices such as the president.

To put our theory to the test, we leverage a new dataset on local elections (Warshaw, de Benedictis-Kessner and Velez 2022) providing us with the election returns and candidate information for over 2,000 county-level elections for sheriffs and prosecutors over the last 30 years in the United States. We combine these data with county-level demographic and socioeconomic metrics from the U.S. Census Bureau and the FBI’s Uniform Crime Reporting (UCR) Program Data. Using a series of two-way fixed effects and multilevel Bayesian linear regression models, we find that an increase in local violent crime is associated with increased support for Republican prosecutor candidates, though not candidates for sheriff or president. This relationship between crime and vote choice is not driven by year-specific effects of local violent crime rates on support for Republican candidates for prosecutor. Moreover, we also show that changes in national crime rates, contrary to changes in local crime rates, do not significantly affect vote choice across all election types.

Overall, our paper makes the following contributions: (1) we provide initial evidence that voters use local violent crime rates to inform vote choice in local elections, (2) our finding that local violent crime is associated with support for Republican prosecutor candidates lends evidence to the theory that the Republican Party owns “law and order” issues in the United States, and (3) we show that voters associate crime differently between those officials responsible for law (prosecutors) and order (sheriffs). This work highlights the importance of studying the effect of local context on voting behavior, has several important implications for the study of issue ownership and retrospective voting in local politics, and provides numerous avenues for future research.

Authors

• Joseph Ryan Loffredo, Massachusetts Institute of Technology

• Gabrielle Peloquin-Skulski, Massachusetts Institute of Technology