The 'demand gap' between police numbers and violence

People often question the relationship between police numbers and activity, and crime.

Wilson and Boland (1) were among the first to examine this, concluding that "Police resources and police activity independently affect the robbery rate after controlling for various socioeconomic factors". When one Australian state brought in legislation that gave police the power to stop, search and detain a person without warrant under particular circumstances or tell them to move on, they found increases in the use of these powers reduced burglary and vehicle theft (2). In an innovative natural experiment, Eric Piza and colleagues found that "police layoffs were associated with significant increases of overall crime, violent crime, and property crime" (3). And when Hunter Boehme and Scott Mourtgos explored when Los Angeles prevented officers from using minor infractions as a pretextual reason to further investigate drivers, it didn't have the desired effect of reducing racial differences in stops. It did however link to an increase in both violent and property crimes post-intervention (4). But what about overall police numbers, regardless of activity?

Police numbers and violence over time

For another writing project (a third edition of my book 'Intelligence-Led Policing') I am updating a couple of charts showing police numbers and crime. Previously, I looked at all recorded crime; however, crime recording practices have changed so much over the decades in both the US and the UK, that it seemed more prudent to examine violence. There is less variation in recording practices and violence has been less vulnerable to changing public reporting practices. This blog post doesn't have an editorial comment about the charts - it just shows the data and the sources. Make up your own mind about what it all means.

Changes in the US since 1960

First the US. Using 1960 as an index year, this graphic charts changes in fulltime sworn personnel per capita as a percentage change since 1960. The percentage change is not by raw numbers, but rather by officers per capita, because the US population has grown considerably over the time of the chart. The demand gap between available personnel and reported violence increases dramatically in the 1980s and 1990s, and never really closes.

The sworn officer data are drawn from paper records reported in UCR reports sourced from Temple University's library until 1984, and thereafter from the FBI's Crime Data Explorer. Population data used to weigh the sworn officer numbers are sourced from Macrotrends population data. Crime data in the first two editions of my book were compiled from a variety of sources. Fortunately for the third edition, I could draw on a recent blog post by Jeff Asher, where he generously provides a repository of US national crime rates going back to 1960. Cheers Jeff! If you are interested specifically in murder, then I highly recommend subscribing to Jeff Asher's substack page where he has historical murder data for the same period. And if you want to hear Jeff talking about crime data in the US, the Reducing Crime podcast has an episode featuring Jeff.

Changes in the US since 1960

In the UK, the violence/police demand gap is even more pronounced during the 1980s and 1990s. Note the scale changed compared to the US chart.

For reference, the sources for this chart are more of a data hodge-podge. Officer numbers back in the early 1960s are grabbed from Hansard a couple of times, and then interpolated. Hansard is useful to about 1985 where there is then some overlap with an obscure and now buried House of Commons research paper from 2001. I say buried because the link has disappeared. Fortunately, I grabbed a copy before it evaporated. A Home Office statistical bulletin covers 1998 to about 2005, and then from there Home Office statistical reports are all generally more reliable.

UK population data are sourced from Macrotrends until 2002 whereupon I used the population numbers reported by the Home Office in their police strength report as a denomiator for police officer numbers, so the chart is comparable to the US chart. Violence numbers are drawn from data going back to the late 1800s, and I use column AD (Total violence against the person). From 1980 I flip to the Crime Survey of England and Wales, accessible from Worksheet A1 here.

I'm still researching data sources, so I cannot guarantee these data are reliable nor that the charts will stay as is for the book. But in the meantime, I hope you find this interesting.

References

(1) Wilson, J. Q., & Boland, B. (1978). The effect of the police on crime. Law and Society Review, 12(3), 367-390.

(2) Wang, J. J. J., & Weatherburn, D. (2021). The effect of police searches and move-on directions on property and violent crime in New South Wales. Journal of Criminology, 54(2), 383–401.

(3) Piza, E. L., & Chillar, V. F. (2021). The effect of police layoffs on crime: A natural experiment involving New Jersey’s two largest cities. Justice Evaluation, 4(2), 176-196.

(4) Boehme, H. M., & Mourtgos, S. M. (2024). The effect of formal de-policing on police traffic stop behavior and crime: Early evidence from LAPD's policy to restrict discretionary traffic stops. Criminology and Public Policy.