LSE - Small Logo
LSE - Small Logo

G. Agustin Markarian

February 3rd, 2023

Living close to a police killing depresses voter turnout, especially among Black voters.

0 comments | 1 shares

Estimated reading time: 14 minutes

G. Agustin Markarian

February 3rd, 2023

Living close to a police killing depresses voter turnout, especially among Black voters.

0 comments | 1 shares

Estimated reading time: 14 minutes

Every year police officers in the United States kill over 1,000 people. In the past decade, high-profile police killings of unarmed Black Americans in the US have triggered national and international protests, but do police killings also affect how people participate in elections? In new research, G. Agustin Markarian finds that police killings that occurred before elections in California, Florida, and Ohio reduced voter turnout by three percentage points in the 2016 presidential election among registered voters who lived nearby. In addition, in these three states, Black voter turnout decreased by nearly six percentage points among registered voters living near a police killing.

Living close to where the police have killed someone is linked to many negative outcomes: nearby killings by police impact the mental health of people of the same ethnicity, reduce the number of 911 calls members of the community make, and hamper students’ academic performances. Living near policing killings also affects people’s political sentiments: living close to where one has occurred reduces people’s beliefs that they can influence politics (also known as political efficacy), and their trust in government. These effects are particularly pronounced among Black Americans.

People are more likely to engage in conventional forms of political participation, like voting or donating to campaigns, when they believe they can influence government and have trust in it. While political efficacy and trust in government have independent effects on political participation, their combination is a good predictor of voter turnout. In my research, I adapt a previously developed model to examine how this relationship works. Those with high levels of trust in government and political efficacy are more likely to participate in politics through conventional means. A reduction in either dimension reduces the likelihood of someone frequently participating in conventional ways. Since police killings reduce some nearby residents’ trust in government and/or their political efficacy, we should expect that some people will vote less often, substitute voting with protests, or withdraw from politics altogether.

Comparing Pre- and Post-Election Police Killings 

To investigate the effect of nearby police killings on voter turnout, I collected voter data from California, Florida, and Ohio using publicly available voter files. Police killing data come from Mapping Police Violence, a crowd-sourced dataset tracking national police killings since 2015. I geocoded all voters and police killings and compared how close they were to one another. I focused on voters living within a mile of police killings occurring in the six months before and after the 2016 general election.

Testing the impact of nearby police killings on voter turnout is not easy. The location of police killings is not randomly distributed. In other words, because of deeply rooted racial segregation and racialized policing practices, police are more likely to kill people in high-crime, high-surveillance, and low-income communities of color – multiple factors associated with lower voter turnout and, therefore, confounding any simple observational study. However, the timing of police killings is unlikely to be linked with the timing of elections, meaning that there is a unique opportunity to observe a naturally occurring experiment.

Suppose the timing of police killings is randomly distributed, an assumption largely supported by the data. In that case, places that had a police killing before the 2016 general presidential election should be similar to places that had a police killing after it. I test that assumption in Figure 1 by comparing registered voters living near a pre-election police killing to those living near a post-election police killing. We see that voters are similar across demographic characteristics, and they are likely similar across unobservable characteristics as well. The only meaningful difference across the two groups is that one set of voters experienced a proximal police killing before the 2016 general election while the others did not. Additionally, we can take advantage of the longitudinal (panel) nature of the data by comparing changes in voter turnout rates between 2012 and 2016 across the two groups.

Figure 1 – Voter turnout by demographic

Nearby Police Killings Reduce Voter Turnout, Particularly Among Black Voters 

In Figure 2, I compare voter turnout rates over time across the two groups of voters (those that experienced a nearby pre-election police killing before the 2016 election and those that experienced one after the 2016 election). The results are then disaggregated by voters’ race and ethnicity. Voter turnout rates are relatively consistent among the two groups between 2008 and 2012, suggesting that without a pre-election police killing, participation trends are consistent in the treatment and control groups. However, we observe drastically different voter turnout trends between 2012 and 2016, particularly among Black voters (Figure 2.3). Between 2012 and 2016, voter turnout rates increase by three percentage points more in the post-election killing group than in the pre-election killing group (Figure 2.1). These results suggest that nearby pre-election police killings reduce voter turnout. My analysis indicates that these differences are statistically significant when observing all voters and only Black voters but not when only studying White or Latino voters.

Figure 2 – Voter Turnout Trends by Group by Race

Further analysis shows that the distance between a police killing and a voter matters, as voters living closer to a police killing are more demobilized than those living further away. We can see these effects in Figure 3, which compares average voter turnout for each group at 0.1-mile intervals from their associated killing by voters’ race.

Figure 3 – Average Turnout by Distance From Police Killing by Group by Race

Because people living further away from police killings are less likely to live in high-crime, high-surveillance, high-poverty neighborhoods – factors negatively associated with voter turnout – we should expect a positive slope in both lines. However, I hypothesize that living near a pre-election police killing has an independent negative effect on voter turnout. If that hypothesis is correct, the line for the pre-election killing group should have a steeper positive slope than the post-election killing group line.

Photo by Koshu Kunii on Unsplash

This hypothesis is supported by Figure 3 when we look at all voters (Figure 3.1), Black voters (Figure 3.3), and, to a lesser degree, Latino voters (Figure 3.4). Black voters living within 0.1 miles of a pre-election police killing are about eight percentage points less likely to vote than similar voters who did not experience a pre-election police killing. However, Black voters living 1.5 miles away from a pre-election police killing are just as likely to vote as their comparison group. I do not find any statistically significant effects when formally analyzing White and Latino voters.

My research suggests that police killings are not only a tragedy in their own right; they disrupt communities’ political structures. Pre-election police killings demobilized nearby registered voters, reducing their turnout by about three percentage points. About 35,000 registered voters living within a mile of a pre-election police killing were demobilized in these three states. Black voters were particularly demobilized, meaning police killings may increase gaps in political participation. My findings help shed light on the role that policing practices play in shaping political behavior by focusing on an increasingly important topic in political science: police violence.


About the author

G. Agustin Markarian

G. Agustin Markarian is an assistant professor at Loyola University Chicago. His research lies at the intersections of political representation and public policy, focusing on racial, ethnic, and class-based inequalities. 

Posted In: Justice and Domestic Affairs

Leave a Reply

Your email address will not be published. Required fields are marked *

LSE Review of Books Visit our sister blog: British Politics and Policy at LSE

RSS Latest LSE Events podcasts

This work by LSE USAPP blog is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported.