Police-involved deaths and the impact on homicide rates in the post-Ferguson era: a study of 45 US cities

This study investigated whether in the post-Ferguson era, homicide rates increased in cities where there was a protested police-involved death. It also tests for evidence of two potential mechanisms. To test for evidence of legal cynicism, effects between homicide and aggravated assault rates are compared; a gap would suggest reduced reporting and community disengagement from police. The moderating influence of state or federal investigation was examined as a potential indicator of de-policing. Using an interrupted time series design, I analysed trends in 45 US cities with a protested police-involved death. Results were combined using a meta-analysis, and meta-regressions were used to test for moderating effects. A funnel plot and Egger’s regression were used to test for bias in event selection. Averaged across all cities, there was an acute and largely sustained increase of 31.6% in the homicide rate, which was significantly larger than the effect in aggravated assaults. Effects were not significantly moderated by state or federal investigations. There was no evidence of bias in event selection. The findings suggest that police-involved deaths can have wider-reaching increases in violence in the communities they are meant to protect.