Ferguson Effect: De-policing Hypothesis

A disputed proposal that scrutiny of police conduct may reduce proactive policing and thereby contribute to violent crime.

Version

v1 · 2026-10-31 · History

Domain-specific #

14496

Domain group

Social Sciences

Origin domain

Criminology & Forensic Studies

Subdomains

Policing, Criminal Justice → Criminology & Forensic Studies

Core Idea

The de-policing version of the Ferguson effect proposes that publicity and scrutiny after controversial police use of force may reduce discretionary proactive activity, which may in turn weaken deterrence or enforcement contact and contribute to violent crime. Both causal links are disputed empirical questions. A separate community-legitimacy account focuses on community responses and cooperation rather than requiring officer pullback.^[1]

How would you explain it like I'm…

Less Checking, More Crime?

Some people have an idea: after police get a lot of criticism for hurting someone, maybe officers do less stopping and checking on their own. And maybe that could lead to more serious crime. Scientists treat each step as a question to check — nobody has proven it's true.

Does Criticism Make Police Pull Back?

The Ferguson effect is a name for some ideas about police and crime. One version, called de-policing, suggests a chain of steps. First, when police use of force becomes controversial and gets lots of public attention, officers might do less of the extra, self-started police work. Second, with less of that work, there might be fewer chances to stop crime. Third, that might lead to more violent crime. Each step is a guess that researchers need to test, not a proven fact. Researchers also separate this from a different idea about communities losing trust in police.

De-Policing Chain Hypothesis

The de-policing version of the Ferguson effect proposes a causal chain. Publicity and scrutiny after a controversial police use of force might lead officers to cut back on discretionary, proactive activity — work they choose to do rather than respond to calls. Less proactive activity might weaken deterrence or reduce enforcement contact, contributing to a rise in violent crime. Each link in this chain is a hypothesis to investigate, not an established result of the step before it. In his 2016 research assessment, Rosenfeld separated this de-policing pathway from a different explanation focused on communities becoming alienated and seeing police as less legitimate.

 

The de-policing hypothesis is one version of the so-called Ferguson effect. It proposes a three-link causal chain: public attention and scrutiny following controversial police use of force may lead officers to reduce discretionary proactive activity; that reduction may weaken deterrence or decrease enforcement contact; and the weakened deterrence or contact may contribute to increased violent crime. Each arrow in the chain is an empirical claim requiring its own investigation, not a demonstrated consequence of the event before it. Evidence for one link does not establish the others. Rosenfeld's 2016 research assessment distinguishes this de-policing pathway from a separate account that centers on community alienation and the perceived legitimacy of police, rather than on changes in officer behavior.

Structural Signature

The hypothesis relates a publicity and scrutiny event, a possible officer response, a proposed deterrence link and a separately measured crime outcome. Stops and arrests can indicate activity but do not directly measure motive or deterrence.

Clarity

Distinguish a proposed cause, a recorded change in police activity and a measured crime outcome. A stop or arrest count does not reveal officer motivation by itself.

Manages Complexity

Examine each proposed link separately and consider policy changes and other explanations. This keeps a sequence of events from being mistaken for proof that one caused the next.

Abstract Reasoning

If the proposed chain held, evidence would need to support both the activity response and its contribution to crime. Fewer arrests without an unexpected crime rise challenge an automatic version; fewer stops alongside increased violence can have multiple explanations.

Knowledge Transfer

The same question can be examined in different cities, but their incidents, activity measures and local policies differ. Findings from one context do not automatically establish results elsewhere.

Example

Before Freddie Gray's arrest, Morgan and Pally found Baltimore arrests declined after Ferguson publicity while seasonally adjusted crime showed very weak evidence of an unexpected increase. Later post-Gray changes occurred amid local unrest and competing explanations. This is a partial test and contrast, not proof of a causal chain.^[2]

In Chicago, investigatory stops declined around late 2015 and gun violence rose in 2016. The Crime Lab identified overlapping video, investigation, agreement and legal changes, and compared New York's large stop decline without a comparable homicide increase. The observations leave both the reason for Chicago's decline and its effect on violence unresolved.^[3]

Neighborhood in Abstraction Space

Ferguson Effect: De-policing Hypothesis sits in a sparse region of the domain-specific corpus (100^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (2835 abstractions)

Nearest neighbors

• Null Ritual — 0.72
• Internal validity — 0.70
• Regression Discontinuity Design — 0.70
• Confirmation Fatigue — 0.70
• Shooting bias — 0.69

Computed from structural-signature embeddings · 2026-10-31

Not to Be Confused With

Deterrence is one proposed intermediate link. Legitimacy can form part of an alternative account. Neither is equivalent to the whole de-policing hypothesis. Arrests and stops are imperfect indicators, and the community-legitimacy interpretation need not involve officer withdrawal. The two city reports are historical investigations, not a universal verdict.^[1]

References

[1] Richard Rosenfeld, Documenting and Explaining the 2015 Homicide Rise Research Directions, National Institute of Justice, June 2016, pp. 2–4, 18–23. The report is the author's historical research assessment. registry ↩a ↩b

[2] Stephen L. Morgan and Joel A. Pally, Ferguson, Gray, and Davis An Analysis of Recorded Crime Incidents and Arrests in Baltimore City, March 2010 through December 2015, Johns Hopkins University, March 15, 2016, executive summary pp. 2–4 and joint conclusions p. 68. registry ↩

[3] University of Chicago Crime Lab, Gun Violence in Chicago, 2016, January 2017, pp. 3–4 and 24–25, figures 37–38. registry ↩