Read on Twitter
New working paper joint with @benconomics, @emilyweisburst and @MWillJr that considers the race-specific effects of a larger police force in the U.S. https://www.nber.org/system/files/working_papers/w28202/w28202.pdf
We find that, in most cities, more police manpower leads to reductions in homicide, with every 10-17 officers hired abating one homicide. In per capita terms, the effects are twice as large for Black versus white victims.
We also consider the extent to which investments in police manpower expand civilian interactions with the criminal justice system, or create "net widening" effects, focusing on differences by race in the burdens and benefits of enforcement activity.
Investments in police manpower lead to more arrests for lower level "quality-of-life crimes", which typically do not have victims and carry small misdemeanor penalties. The burden of these arrests falls disproportionately on Black civilians.
At the same time, we find that arrests for the most serious offenses fall with police manpower, especially for Black Americans. While low level arrests can accumulate over time, this finding suggests that police hiring has the potential to reduce both crime and incarceration.
Critically though, the average effects described above mask important variation in the quality of policing across cities. In cities with relatively large Black populations, the returns to investments in police manpower are smaller and perhaps non-existent for Black civilians.
Likewise, in these cities, larger police forces lead to a greater number of arrests for "quality of life" offenses, in particular for Black civilians, without the reduction index crime arrests that we observe elsewhere.
The pattern of findings provides empirical support for two important propositions: First, given cross-city heterogeneity, it is the presence of police officers more so than the number of arrests they make which drives the public safety returns to investments in law enforcement.
Second, by showing that the cities with the largest Black populations don't share equally in the benefits of policing while sustaining the greatest burdens, we provide novel empirical support for the narrative that Black communities are both over- and under-policed.
By focusing on the size of a city’s police force, we provide
historical evidence on a critical policy estimand that is implicated by the Defund movement and which, for many years, has been the primary means by which municipal policymakers have invested in public safety.
historical evidence on a critical policy estimand that is implicated by the Defund movement and which, for many years, has been the primary means by which municipal policymakers have invested in public safety.
This study provides an estimate of the historical trade-offs of investments in police and, critically, the resulting implications for communities of color. Looking backward, our results suggest that reducing the size of a city's police force has led to reduced public safety.
This is a cautionary tale for at least some incarnations of the Defund movement. At the same time, there are many examples of alternatives to police which have been shown to be effective in reducing crime.
These include investments in physical infrastructure (e.g., green space, fixing abandoned properties & street lighting), CBT, youth summer jobs and social supports more generally. A challenge though is that social programs are notoriously difficult to scale.
Given that we have never seen a large shift in the policy regime from one that is intensive in policing to one that is intensive in the use of alternatives to policing, research cannot tell us very much about what a shift like this will mean for public safety in the future.
What this research suggests though is that while Black communities might benefit in a number of ways from less police presence, Black communities also disproportionately bear the risks of reduced investments in policing.
There is much more detail in the paper, especially in the supplementary appendix which considers heterogeneity with respect to age and gender as well as some other outcomes such as officer safety.
