More police presence makes communities safer

The claim

Increased police staffing and patrol presence deters crime by raising the probability of detection and arrest. Communities experiencing high crime rates need more officers on the street, faster response times, and stronger enforcement. Reducing police budgets or staffing, as reformers propose, will produce predictable rises in violent crime and harm the communities that most need protection.

The mechanism

The core deterrence argument is grounded in rational choice theory: when the expected cost of crime — the probability of apprehension multiplied by the severity of punishment — rises, criminal behavior should decline. Police presence raises both visible deterrence (would-be offenders see officers) and actual probability of arrest. This is the theoretical foundation for patrol staffing models and is not without empirical support. The mechanism breaks down, however, when it confronts several complications: the geographic concentration of deterrent effects, the role of community trust in reporting rates, the distinction between presence and clearance, and the counterfactual efficiency of alternative public safety investments.

The claim conflates several distinct propositions: (1) more police reduce crime through deterrence; (2) more police improve crime clearance rates; (3) patrol presence is more cost-effective than other safety investments; and (4) the reform proposition — “fewer police” — reduces safety. These propositions do not rise or fall together, and the evidence is substantially different for each.

The evidence

Police staffing elasticity: Chalfin and McCrary

The most rigorous econometric estimate of the police-crime relationship comes from Aaron Chalfin and Justin McCrary’s 2018 Journal of Economic Literature review and their earlier 2018 American Economic Review paper. Using measurement-error correction to address the attenuation bias that plagues earlier OLS studies, they estimate that a 10% increase in police size reduces violent crime by approximately 3–10%, depending on specification and crime type. The cost-benefit calculation is, by their own accounting, favorable — an additional officer generates crime cost savings that exceed the officer’s salary. However, the confidence interval is wide, the effect is an average across heterogeneous cities, and it says nothing about the deployment model within departments.

Mello (2019) in the Journal of Public Economics used variation in federal COPS (Community Oriented Policing Services) grant funding as a natural experiment. Cities that received COPS grants hired more officers; comparing grant recipients to similarly situated non-recipients, Mello estimates approximately 0.9% reduction in violent crime per 1% increase in police. This near-unit elasticity is among the stronger estimates in the literature but applies specifically to the marginal officer hired under a community policing framework — not general patrol staffing.

Hot-spots policing: the strongest evidence

The cleanest evidence that police activity reduces crime concerns not staffing levels but deployment concentration. Lawrence Sherman and colleagues pioneered hot-spots policing research in Kansas City in the early 1990s, finding that concentrating patrol on small, high-crime micro-locations (blocks, intersections) produced substantial crime reductions with no evidence of geographic displacement to adjacent areas. David Weisburd’s 2016 Justice Quarterly meta-analysis of 65 hot-spots randomized experiments found consistent, significant crime reductions with an average effect size of approximately 0.2 standard deviations — stronger than almost any other policing intervention in the experimental literature.

The critical finding from hot-spots research is that it does not primarily require more officers; it requires different deployment of existing officers. The “Koper curve” (1995) suggests that patrol doses of 15 minutes produce maximum deterrence before officers should rotate to the next hot spot. This is a deployment optimization problem, not a headcount problem.

Camden, NJ: the restructuring case

In 2013, Camden, New Jersey dissolved its city police department — not to reduce police but to escape an unworkable union contract and restructure to a county-wide community policing model. Officer headcount actually increased, but the deployment philosophy changed fundamentally: officers were assigned to walk beats, build community relationships, and respond to social service needs. Use-of-force incidents fell dramatically. Violent crime declined approximately 42% over the five subsequent years by NJ State Police Uniform Crime Report data. Camden is routinely misrepresented in the “defund police” debate: it did not reduce police, it rebuilt them. But the case demonstrates that deployment model and community legitimacy can produce crime reductions that headcount alone does not.

The trust and reporting problem

A dimension systematically underweighted in the staffing debate is the relationship between police-community trust and crime reporting. Desmond, Papachristos, and Kirk (2016) in the American Sociological Review documented a “Ferguson effect” in Milwaukee after a high-profile police killing: 911 calls in Black neighborhoods fell significantly for months afterward. If victims and witnesses do not report crimes because they distrust police, the deterrence mechanism partially breaks down — police cannot clear crimes they do not know about, and victims have no pathway to resolution. Generalized patrol increases in low-trust communities can depress reporting without producing compensating deterrence gains.

Clearance rates for violent crime have declined across most US cities, with homicide clearance falling from approximately 90% in 1965 to roughly 50% in 2020 (BJS). Adding officers has not reversed this trend. The clearance problem is at least partly a function of community cooperation, which depends on legitimacy, not presence.

Alternative models: community violence interruption

Jeffrey Butts and colleagues’ 2015 Annual Review of Public Health meta-analysis of Cure Violence and similar community violence interruption (CVI) programs found 19–24% reductions in shootings in targeted areas. CVI deploys credible messengers — people with street credibility who can interrupt retaliatory violence cycles — at substantially lower cost per unit of crime reduction than uniformed patrol. A 2021 RAND analysis of CVI programs in multiple cities found cost-per-crime-prevented figures competitive with or lower than incremental patrol staffing. These programs do not replace police; they address the violence dynamics that respond poorly to deterrence-only models.

Mental health co-responders

Approximately 20% of police calls nationally involve mental health crises (NAMI estimates). Sending armed officers as first responders to psychiatric emergencies produces worse outcomes on average — higher use-of-force rates, criminalization of mental illness — and consumes patrol time without safety benefit. Eugene, Oregon’s CAHOOTS program, operational since 1989, sends mental health crisis teams without officers to 17–24% of 911 calls; less than 1% of CAHOOTS calls require police backup. Denver’s STAR program showed zero arrests in its first six months across approximately 750 crisis calls. These co-responder models free officer time for law enforcement functions while producing better outcomes for crisis callers.

Cross-national context

Japan (2.3 officers per 1,000 population) and Finland (1.5 per 1,000) have substantially fewer police per capita than the US (approximately 2.3 per 1,000, though this is an undercount given the US definitional inconsistencies) and dramatically lower violent crime rates. Germany (3.0) and the Netherlands (2.8) are closer to the US in staffing but have far lower homicide rates. The UK (2.3) has a homicide rate roughly one-fifth of the US. These cross-national comparisons are confounded by inequality, gun access, social safety net, and incarceration history — but they do not support the claim that police staffing is the primary driver of cross-national violence differences.

Who benefits

Police unions — particularly the Fraternal Order of Police (FOP), the International Brotherhood of Teamsters’ corrections and police affiliates, and large municipal unions like the NYPD’s PBA — have structural interests in headcount maximization. The FOP spent approximately $1 million on federal lobbying in 2022 and endorses candidates primarily on staffing and qualified immunity grounds. Benefit accrues directly to dues-paying officers through overtime, pension formulas calculated on headcount norms, and political leverage. The FOP’s political interest in framing any policing debate as a binary headcount question — rather than a deployment optimization or accountability question — is substantial.

Law enforcement equipment vendors (Axon, formerly TASER International; Motorola Solutions; ShotSpotter, now SoundThinking) sell hardware and software that scales per agency and per officer. The ShotSpotter contract, for example, charges cities per square mile monitored; larger patrol areas require more coverage. These companies have lobbied against budget reductions and fund research centers that produce favorable analysis.

Politicians running on public safety platforms in high-crime jurisdictions benefit from the simple, legible signal of officer headcount. Hiring announcements are easy to stage; restructuring deployment models and investing in community violence interruption programs are harder to communicate and slower to show results.

The counter

The individual-mechanism account of policing has genuine force within a defined scope. Hot-spots policing is among the best-evidenced crime reduction interventions in social science — the effect is real, replicable, and not easily explained away. Chalfin and McCrary’s cost-benefit math is defensible as an average across American cities; in cities with severe officer shortages relative to their crime concentrations, hiring more officers probably does reduce crime. The 2020–2022 period of significant violent crime increases in many US cities, coinciding with retirements, resignations, and morale declines in police departments following the 2020 protests, is consistent with the staffing hypothesis — though it is also consistent with pandemic disruption, court backlogs, and firearm purchase surges that make causal attribution genuinely uncertain.

The strongest version of the individual claim is not “patrol presence deters crime at scale” but “deterrence requires a functional police department, and departments that fall below a staffing floor experience measurable crime increases.” That claim is better supported than the broad version and is worth taking seriously in resource allocation debates. The contested verdict here reflects genuine empirical heterogeneity, not an absence of evidence for police effectiveness.

References

Butts, J. A., Roman, C. G., Bostwick, L., & Porter, J. R. (2015). Cure violence: A public health model to reduce gun violence. Annual Review of Public Health, 36, 39–53. https://doi.org/10.1146/annurev-publhealth-031914-122509

Chalfin, A., & McCrary, J. (2018). Are US cities underpoliced? Theory and evidence. Review of Economics and Statistics, 100(1), 167–186. https://doi.org/10.1162/REST_a_00694

Desmond, M., Papachristos, A. V., & Kirk, D. S. (2016). Police violence and citizen crime reporting in the Black community. American Sociological Review, 81(5), 857–876. https://doi.org/10.1177/0003122416663494

Koper, C. S. (1995). Just enough police presence: Reducing crime and disorderly behavior by optimizing patrol time in crime hot spots. Justice Quarterly, 12(4), 649–672. https://doi.org/10.1080/07418829500096281

MacDonald, J., Klick, J., & Grunwald, B. (2016). The effect of privately provided police presence on crime: Evidence from a business improvement district. Journal of the Royal Statistical Society: Series A, 179(3), 831–846. https://doi.org/10.1111/rssa.12133

Mello, S. (2019). More COPS, less crime. Journal of Public Economics, 172, 174–200. https://doi.org/10.1016/j.jpubeco.2018.12.003

Sherman, L. W., & Rogan, D. P. (1995). Effects of gun seizures on gun violence: ‘Hot patrols’ in Kansas City. Justice Quarterly, 12(4), 673–693. https://doi.org/10.1080/07418829500096281

Weisburd, D., Telep, C. W., Hinkle, J. C., & Eck, J. E. (2010). Is problem-oriented policing effective in reducing crime and disorder? Criminology & Public Policy, 9(1), 139–172. https://doi.org/10.1111/j.1745-9133.2010.00617.x

Weisburd, D. (2016). Does hot spots policing inevitably lead to unfair and abusive police practices, or can we maximize both fairness and effectiveness? Justice Quarterly, 33(3), 1–24. https://doi.org/10.1080/07418825.2015.1103379

Roman, J. K. (2021). Can community violence intervention programs reduce US homicides? RAND Corporation. https://www.rand.org/pubs/perspectives/PEA1259-1.html