Did de‐policing cause the increase in homicide rates?

• Author: Richard Rosenfeld,Joel Wallman
• Date: 01 February 2019
• Published date: 01 February 2019
• DOI: http://doi.org/10.1111/1745-9133.12414

DOI: 10.1111/1745-9133.12414

RESEARCH ARTICLE

DE-POLICING AND THE HOMICIDE RAISE

Did de-policing cause the increase in homicide rates?

Richard Rosenfeld1Joel Wallman2

1University of Missouri—St. Louis

2Harry Frank Guggenheim Foundation

Correspondence

RichardRosenfeld, Department of

Criminology and Criminal Justice,

Universityof Missouri—St. Louis, St. Louis,

MO63121.

Email:r ichard_rosenfeld@umsl.edu

Wethank the special editor and reviewers for

helpfulcomments on an earlier draft.

Research Summary: Widespread protests and demands

for accountability in the wake of broadly publicized police

killings of unarmed civilians coincided with the marked

upturn in homicide levels, especially in large U.S. cities, in

2015. Many observers, including prominent political fig-

ures, claimed that de-policing caused the homicide rate

to rise: Fearing increased legal liability and publicity,

the police employed less proactive enforcement and made

fewer arrests, producing an increase in homicide levels. We

use structural equation modeling to estimate the simultane-

ous relationship between arrest and homicide rates between

2010 and 2015 in 53 large cities. We find no evidenceof an

effect of arrest rates on city homicide rates for any offense

category for any yearin t his period, including 2015, the year

of the spike in homicide levels.

Policy Implications: The results of our analysis revealthat

declining rates of arrest did not produce the rise in homi-

cide levels. But arrests, the chief formal reaction to crime,

can play an important supporting role in crime preven-

tion. Ample research findings have demonstrated the effec-

tiveness of proactive policing practices, such as targeted

patrol, in reducing crime rates. The success of such tactics

depends, in turn, on citizens’ belief that criminal offenders

face an appreciable risk of arrest. It is equally important,

however, that residents of communities that have experi-

enced unwarranted police violence believethat law enforce-

ment is fairly applied. Lack of trust and confidence in the

police militates against crime prevention as a co-production

Criminology & Public Policy. 2019;18:51–75. wileyonlinelibrary.com/journal/capp © 2019 American Society of Criminology 51

52 ROSENFELD AND WALL MA N

of police and community and may promote violence as a

form of self-help.

KEYWORDS

arrests, de-policing, homicide

1INTRODUCTION

After falling almost continuously for more than two decades, U.S. homicide rates rose abruptly

in 2015. A dominant narrative attributed the increase to de-policing. The media catchphrase “the

Ferguson Effect” conveyed the idea that the police had disengaged from proactive enforcement out

of fear of heightened legal liability or of having their identities exposed on social or traditional media,

and that less enforcement led to more crime.

Other than a few case studies and anecdotal media reports, we are awareof no systematic nationwide

studies of the relationship between police disengagement and the rise in homicide levels. To estimate

a possible relationship between police activity and the 2015 spike in homicide rates, we examined the

association between homicide rates and arrest rates in 53 large U.S. cities between 2010 and 2015.

Our analysis is based on structural equation panel models in which the simultaneous relationship1

between homicide and arrest rates is incorporated. We found no association between reduced rates of

arrest and elevated rates of homicide after controlling for their simultaneous relationship. Whatever

sparked the 2015 homicide rise, it was not a reduction in arrest rates in large U.S. city police agencies.

2BACKGROUND

As reports began to emerge in 2015 that homicide rates were increasing in several big cities, some

observers countered that claims of a new crime wave were overblown and that the longstanding U.S.

crime decline had not ended (Bialik, 2015; Coates, 2015; Davey & Smith, 2015; Pyrooz, Decker,

Wolfe, & Shjarback, 2016). The critics were correct that the crime drop had not ended, but they were

mistaken to downplay the significance of the homicide rise. It was not a media creation. It was real,

large, abrupt, and in need of explanation (Rosenfeld, Gaston, Spivak, & Irazola, 2017).

The nationwide homicide rate increased by 11.4%, to 4.9 homicides per 100,000 population in 2015

from 4.4 in 2014 (ucr.fbi.gov). That was the largest percentage change in homicide levels in a single

year since 1968 (Rosenfeld et al., 2017). During the same period, homicide rates in large cities grew

by 17.2%.2Not all big cities experienced an increase in homicide rates in 2015. In some, they were

flat or fell. Others had small increases. Rosenfeld et al. (2017) found that homicide rates increased

by more than 25% between 2014 and 2015 in 14 of the 46 large cities examined, and in 9 cities, the

increases exceeded 50%. The heterogeneity in the rise in homicide levels across the big cities requires

explanation. If the de-policing explanation is correct, other things being equal, we should observe

greater homicide rate increases in cities where arrest rates were declining.

2.1 Police violence and community unrest

In August 2014, Michael Brown wasshot and killed by a police officer in Ferguson, Missouri, a suburb

of St. Louis. Michael Brown was Black, and the officer is White. Brownwas unarmed, although some