The Franchise, Policing, and Race: Evidence from Arrests Data and the Voting Rights Act

Abstract

This paper investigates the relationship between the franchise and law enforcement practices using evidence from the Voting Rights Act (VRA) of 1965. We find that, following the VRA, black arrest rates fell in counties that were both covered by the legislation and had a large number of newly enfranchised black voters. We uncover no corresponding patterns for white arrest rates. The reduction in black arrest rates is driven by less serious offenses, for which police might have more enforcement discretion. Importantly, our results are driven by arrests carried out by sheriffs-who are always elected. While there are no corresponding changes for municipal police chiefs in aggregate, we do find similar patterns in covered counties with elected rather than appointed chiefs. We also show that our findings cannot be rationalized by alternative explanations, such as differences in collective bargaining, changes in the underlying propensity to commit crimes, responses to changes in policing practices, and changes in the suppression of civil right protests. Taken together, these results document that voting rights, when combined with elected, rather than appointed, chief law enforcement officers, can lead to improved treatment of minority groups by police.

Enfranchisement and Incarceration after the 1965 Voting Rights Act

“…We further contribute a test of how perceptions of threat can be activated via key political junctures like enfranchisement. At first glance, our results appear inconsistent with more recent work that finds that Black arrest rates declined in response to the VRA (Facchini, Knight, and Testa 2020). Yet, the apparent difference is largely a function of emphasis.…”

“…In addition, we study an important outcome near the end of the carceral process. Further work, in the vein of Facchini, Knight, and Testa (2020), could illuminate which places within this process contribute most to the end result.…”

“…In Equation 2, the parameter of interest is β-the average post-1965 difference in incarceration trends between states covered by Section 5 and those not covered. Finally, following (among others) Cascio and Washington (2014) and Facchini, Knight, and Testa (2020), we estimate a long-difference specification in which we again estimate Equation 2, but exclude data from 1965 to 1975. By focusing only on "long-term" differential changes, a long-difference specification like this is well-suited to situations where agents require some time to adapt to changes (here, passage of Section 5) by excluding the adjustment period from estimation.…”

The Economics of Policing and Public Safety

“…The Civil Rights Act increased the ability of Black Americans to vote in specific counties, making the preferences of Black civilians about policing more important in local elections. Facchini, Knight, and Testa (2020) finds that (elected) sheriff's deputies reduced their level of engagement in response to this increased electoral power. Specifically, in the affected counties, fewer Black people were arrested for low-level offenses-the kinds of incidents for which the benefit of crime reduction is likely lower relative to the direct legitimacy cost of arrests.…”

Voting rights and the resilience of Black turnout

“…Using a similar research design, Facchini et al. (2020) find that preclearance decreased the rate at which Black individuals were arrested for non‐felony offenses in covered counties with elected sheriffs and a high concentration of Black voters. To identify the effects of enfranchisement on labor market disparities, Avenancio‐León and Aneja (2019) employ a triple‐difference design that compares racial differences in labor market outcomes within covered counties to those within neighboring uncovered counties.…”