353Do police-involved deaths increase homicide rates in the post-Ferguson Era? Evidence from 44 US cities

Abstract

Abstract Background Starting in 2014, homicide rates increased in several US cities, reversing a two-decade downward trend. A number of commentators blamed the events of Ferguson, Missouri, where police killed Michael Brown, leading to protests and greater scrutiny of similar events in other cities. It is still unclear whether effect is causal and what the mechanism would be. Methods Using the FBI’s Uniform Crime Reporting database, I derived 2011-2019 crime data from 44 US cities where protests followed a local police-involved death. The main outcome was homicide. Assaults are potentially similar as a conflict-resolution strategy, though reporting is more discretionary than homicides. Differences in effect between homicide and assault were used as a proxy for underreporting and an indicator legal cynicism. Effects were tested using interrupted time series and combined with meta-analyses. Meta-regressions were used to test the moderating effects of external investigations and sociodemographic factors. Results Homicides increased 26% (99% CI: 15%, 36%). Assaults also increased, though the effect was 15% lower than the homicides (-27%, -4%). No tested factor significantly effects. Conclusions Where police-involved deaths lead to protests, homicide rates increase. The findings support de-legitimisation of police as a causal mechanism, but not de-policing. Key messages Police-involved deaths can lead to increased homicides. Cities where police-community tensions are already high may be particularly vulnerable to homicide rate increases following police-involved deaths.