CAN HOT SPOTS POLICING REDUCE CRIME IN URBAN AREAS? AN AGENT‐BASED SIMULATION*

• Author: ANTHONY A. BRAGA,ELIZABETH R. GROFF,DAVID WEISBURD,ALESE WOODITCH
• Date: 01 February 2017
• DOI: http://doi.org/10.1111/1745-9125.12131
• Published date: 01 February 2017

CAN HOT SPOTS POLICING REDUCE CRIME IN

URBAN AREAS? AN AGENT-BASED SIMULATION∗

DAVID WEISBURD,1,2 ANTHONY A. BRAGA,3

ELIZABETH R. GROFF,4and ALESE WOODITCH4

1Department of Criminology, Law and Society, George Mason University

2Institute of Criminology, Hebrew University of Jerusalem

3School of Criminology and Criminal Justice, Northeastern University

4Department of Criminal Justice, Temple University

KEYWORDS: crime hot spots, police innovation, crime prevention, agent-based mod-

eling, hot spots policing

Over the past two decades, there has been a growing consensus among researchers

that hot spots policing is an effective strategy to prevent crime. Although strong evi-

dence exists that hot spots policing will reduce crime at hot spots without immediate

spatial displacement, we know little about its possible jurisdictional or large-area im-

pacts. We cannot isolate such effects in previous experiments because they (appropri-

ately) compare treatment and control hot spots within large urban communities, thus,

confounding estimates of area-wide impacts. An agent-based model is used to estimate

area-wide impacts of hot spots policing on street robbery. We test two implementations

of hot spots policing (representing different levels of resource allocation) in a simulated

borough of a city, and we compare them with two control conditions, one model with

constant random patrol and another with no police officers. Our models estimate the

short- and long-term impacts on large-area robbery levels of these different schemes of

policing resources. These experiments reveal statistically significant effects for hot spots

policing beyond both a random patrol model and a landscape without police. These

simulations suggest that wider application of hot spots policing can have significant

impacts on overall levels of crime in urban areas.

After decades of increasing crime rates, the United States experienced a surprising

crime drop during the 1990s that continued through the 2000s. Policy makers, academics,

and journalists attempted to sort out the various explanations for the puzzling crime

decrease, such as a strong economy, improved policing, high imprisonment rates,

∗Additional supporting information can be found in the listing for this article in the Wiley Online

Library at http://onlinelibrary.wiley.com/doi/10.1111/crim.2017.55.issue-1/issuetoc.

This article was originally prepared for the National Academy of Sciences Round Table on Crime

Trends, chaired by Richard Rosenfeld. We would like to thank the panel for their helpful com-

ments and encouragement. We would also like to thank the Office of Research Computing at

George Mason University, who allowed us access to ARGO, a research computing cluster, to run

our ABM models, and Jayshree Sarma for her generous assistance on this project.

Direct correspondence to David Weisburd, Department of Criminology, Law and Society, George

Mason University, 4400 University Drive, MS 6D12, Fairfax, VA 22030 (e-mail: dweisbur@

gmu.edu).

C2017 American Society of Criminology doi: 10.1111/1745-9125.12131

CRIMINOLOGY Volume 55 Number 1 137–173 2017 137

138 WEISBURD ET AL.

stabilizing crack cocaine markets, immigration, new gun policies, and demographic shifts

(e.g., Blumstein and Wallman, 2006; Levitt, 2004; Zimring, 2007). A careful read of the

available scientific evidence suggests that no single factor can be invoked as the cause

of the 1990s crime decline; rather, the explanation seems to lie with several mutually

supportive, reinforcing factors. Although it is difficult to specify their exact contribu-

tions, innovative police strategies are commonly credited as plausible influential factors

in the crime drop (Wallman and Blumstein, 2006; Weisburd and Braga, 2006; Zimring,

2012).

The most compelling scientific evidence supporting the crime-control effectiveness of

innovative policing strategies comes from controlled evaluations designed to test inter-

ventions at a subset of places within jurisdictions (National Research Council, 2004).

Over the past two decades, the results of a series of rigorous evaluations have suggested

that police can be effective in addressing crime and disorder when they focus on small

units of geography with high rates of crime known as hot spots (see Braga, Papachris-

tos, and Hureau, 2012, 2014; Lum, Koper, and Telep, 2011; National Research Council,

2004; Telep and Weisburd, 2011; Weisburd and Eck, 2004). Policing strategies and tactics

focused on these areas are usually referred to by researchers as hot spots policing or place-

based policing (Sherman and Weisburd, 1995; Weisburd, 2008). The evidence base for the

effectiveness of hot spots policing in reducing crime and disorder at crime hot spots is es-

pecially strong. In a Campbell systematic review of hot spots policing strategies, Braga,

Papachristos, and Hureau (2012, 2014) found that 20 of 25 tests from 19 experimental and

quasi-experimental evaluations resulted in significant and meaningful reductions in crime

or disorder. Importantly, they also did not find evidence of immediate spatial displace-

ment. Crime did not simply shift from hot spots to nearby areas (Braga, Papachristos,

and Hureau, 2012, 2014; Weisburd et al., 2006); indeed, such areas were more likely to

evidence a diffusion of crime-control benefits (Clarke and Weisburd, 1994).

Although the evidence supporting the effectiveness of hot spots policing in reducing

crime and disorder at hot spots is strong, to date, we know little about whether and

to what degree such tactics will influence crime rates in the larger urban areas within

which such strategies are implemented. Because we do not have estimates of the area-

wide benefits of this approach, it is difficult to assess whether hot spots policing strate-

gies applied broadly in jurisdictions would have meaningful effects on overall crime

trends. As such, it is difficult to know whether the emergence of hot spots policing,

a key police innovation that was widely adopted by U.S. police departments over the

course of the 1990s (Weisburd and Lum, 2005), can claim any credit for the 1990s crime

drop.

In this article, we draw on agent-based modeling (Gilbert, 2008; Gilbert and Troitzsch,

2005; Grimm and Railsback, 2005; Johnson and Groff, 2014) to examine the crime-

prevention impacts of hot spots policing on street robbery patterns in an urban area in-

cluding four police beats, which we call a borough. As we detail in the next section, the

findings from current studies do not allow us to draw direct conclusions about impacts of

hot spots policing practices on overall crime trends in urban areas. At the same time, the

findings of these studies allow us to draw specific conclusions about the impacts of such

strategies in small areas or hot spots, including possible displacement and diffusion of

crime-control benefits. By applying these estimates, and more general knowledge about

geographic crime concentrations and their causes in cities, we develop models that esti-

mate expected reductions in street robberies across the simulated police beats (and the

HOT SPOTS POLICING AND URBAN AREA CRIME REDUCTION 139

borough overall) that result from hot spots policing strategies. We focus on street rob-

beries because they have been perceived as a key measure of violence in studies of the

crime drop (see Blumstein, 2006) and a major source of fear among the public (Tompson,

2012).

LIMITATIONS OF MAKING INFERENCES FROM HOT SPOTS

POLICING EXPERIMENTS ON CRIME IMPACTS IN

JURISDICTIONS/LARGE URBAN AREAS

In the 1970s, the aim of evaluations of police programs was generally focused on large

urban-area impacts. The question that police and scholars asked was whether general

strategies of policing led to crime reductions in a city or large sections of a city (e.g., bor-

oughs, precincts, districts, or beats). Overall, the answer during that period was that such

strategies were ineffective. For example, in 1974, the Police Foundation published an in-

fluential report on a large experimental field trial testing the effects of random preventive

police patrol across police beats in Kansas City (Kelling et al., 1974). The results were

unequivocal—increasing or decreasing random preventive police patrol in beats had no

impact on crime or fear of crime. In another large-scale study supported by the National

Institute of Justice, Spelman and Brown (1984) examined rapid response to citizen calls

in four American cities. By interviewing 4,000 victims, witnesses, and bystanders related

to 3,300 serious crime calls, they found that rapid response increased apprehension of

offenders in only a limited set of circumstances.

The results from studies like these led police scholars to make inferences about the

impacts of policing strategies on crime rates in cities. As a result, some scholars were

skeptical of the ability of the police to do anything about crime. David Bayley concluded

in 1994 (p. 3), for example, that:

The police do not prevent crime. This is one of the best kept secrets of modern life.

Experts know it, the police know it, but the public does not know it. Yet the police

pretend that they are society’s best defense against crime and continually argue that

if they are given more resources, especially personnel, they will be able to protect

communities against crime. This is a myth.

When reflecting on the more general acceptance of criminologists for this underlying

skepticism about the traditional mechanism of police deterrence through detection of

crime or apprehension of offenders, Michael Gottfredson and Travis Hirschi (1990: 70)

wrote that, “no evidence exists that augmentation of police forces or equipment, differ-

ential patrol strategies, or differential intensities of surveillance have an effect on crime

rates.”

The emergence of hot spots policing must be considered in this context. Sherman and

Weisburd (1995), in their introduction to the first large-scale field experiment in hot

spots policing, argued that a key reason for the failure of policing strategies to impact

crime was that they were dispersed across jurisdictions and did not focus on the specific

places where crime was concentrated. In drawing on initial findings that crime was con-

centrated at micro-geographic hot spots (e.g., Pierce, Spaar, and Briggs, 1988; Sherman,

1987; Sherman, Gartin, and Buerger, 1989), they suggested that the police should not

dilute the dosage of police interventions across entire beats but should focus it on crime