De-Policing and What to Do About It

The past several years have seen much discussion over the “Ferguson Effect” and “de-policing.” This effect first came to national attention in 2014. A police shooting in Ferguson, Missouri, touched off protests and riots, and crime increased as police grew more hesitant to enforce the law. Subsequently, in the wake of “viral” use-of-force incidents, typically captured on cell-phone or body-camera video, cities often see police retreat from discretionary law-enforcement activity, with stops and arrests declining—and murders and shootings rising.

There has been a voluminous debate over the existence of this pattern and what police reforms might stop controversial use-of-force incidents from occurring. Some critics who want less policing have celebrated cops’ withdrawal.[1] There has also been criticism of the officers pulling back, as well as speculation as to their motives. In 2017, Washington Post columnist Radley Balko called the de-policing theory a “cynical, dangerous argument” that “not only aims to shame critics into silence” but “besmirches the integrity of conscientious police officers;”[2] this year on social media, he parodied it as “murders are surging because an entire profession would rather let people die than hold their colleagues accountable when they needlessly hurt and kill people.” [3]*

*The last sentence in this paragraph has been changed. The original version attributed a comment to Radley Balko in a Washington Post column that he had quoted from another source, and a comment on social media as a statement of his own views.

Less attention, however, has been paid to the question of what to do about de-policing when it happens. [4] Even as states pass new policing laws, viral incidents—ranging from truly horrifying abuses to justifiable uses of force that look excessive to the untrained eye—show no sign of stopping soon in a country with 330 million people, high and racially disparate rates of violent crime, and widespread cell phones and body cameras. Communities must be able to weather such controversies and protests and to dispense justice as appropriate, without descending into murder and chaos, just as they routinely handle even the most severe crimes not involving police.

By reviewing the relevant academic literature and interviewing several sources—including current and former police officers, law-enforcement officials, and criminal-justice experts—I compiled insights and suggestions as to why de-policing happens and what we might do to spur officers to keep working in an unwelcoming environment. [5] Several major themes emerged from these efforts.

Police today find themselves under intense scrutiny, both from the public and from government officials. Contrary to some of the more aggressive police critics, officers are fully aware that real abuses exist and must be dealt with. But they also know that legitimate uses of force sometimes “look bad on video,” which can lead to a public outcry and, in extreme cases, even civil unrest, putting pressure on mayors, police chiefs, and other government officials to “throw them under the bus.” One way for a cop to reduce the risk of this happening is simply to police less, keeping to nondiscretionary acts such as responding to calls and pulling over to help when flagged down.

To continue being proactive in such an environment, police want assurances that their departments and government officials, at least, will treat them fairly if they have to use force. Officials can credibly make such assurances even as they hold bad cops accountable—and they can encourage proactive police work in other ways as well, within certain limits. The big question is whether they want to and have the will and the necessary political support.

Time and again, American cities have experienced the same pattern, often dubbed the “Ferguson Effect”: there is a policing incident; the incident goes “viral” and is followed by a public outcry; officers pull back from discretionary activity, including stops and arrests; and shootings and homicides go up. Such patterns have occurred in cities such as Chicago, St. Louis, Baltimore, Cincinnati, Riverside, California, [6] and, most recently, Minneapolis and Seattle. [7] In a much broader sense, de-policing has occurred across the country, with officers reporting in surveys that these nationally publicized incidents have made police less proactive. [8] In some ways, it’s a return to an older approach to policing from before the era of CompStat and data-driven proactivity—where officers were expected to answer calls but not to do much else to prevent crime.

One recent study, by Tanaya Devi and Roland Fryer, analyzed investigations of police departments accused of having a “pattern or practice” of civil rights violations. It found that such investigations can improve departments and reduce crime—but when the investigations were spurred by viral incidents, crime skyrocketed instead. The explanation that best fit the data, Devi and Fryer contended, was an “abrupt change in the quantity of policing activity,” including declines in police-civilian interactions that reached nearly 90% in Chicago and about half in Riverside and St. Louis. [9] Similarly, Cheng Cheng and Wei Long found that after a wave of highly publicized incidents beginning in late 2014, big cities with large black populations saw decreases in arrests and increases in homicides.[10] Another study, by Travis Campbell, finds that Black Lives Matter protests led to a reduction in local police shootings—but also a drop in clearance rates for property crime, reduced reporting of property crime to the police, and an increase in murders that outnumbered the decrease in police shootings, all findings basically consistent with the de-policing hypothesis. [11]

In theory, though, there is a way we could see such patterns even if de-policing didn’t cause crime: viral incidents and protests could cause de-policing directly as well as increase crime directly, with no separate process through which de-policing causes crime. For example, cops might back off under harsh criticism—after which crime might go up not because the police have backed off but because of the tension, stress, and loss of respectfor the rule of law that these incidents created. In that scenario, efforts to stop de-policing would have no effect on crime.

But this is unlikely to be a full explanation. Viral incidents may increase crime through multiple channels, but de-policing is almost certainly one of them.

Efforts to tie post-viral-incident crime increases to other effects of viral shootings—such as a decline in policing’s “legitimacy” with the public rather than a decline in concrete police activity—have generally produced weak results, as my Manhattan Institute colleague Charles Fain Lehman has explained in detail elsewhere. [12] Moreover, Devi and Fryer found that “geographic divisions within Chicago and St. Louis that witnessed the largest reductions in police activity, after the investigation, also witnessed the largest increases in crime after the investigation,” drawing an even closer correlation between changes to policing and crime. They also pointed out that both crime and policing seem less affected when a viral shooting is not coupled with an investigation of the department, perhaps because cops feel less under fire.

Speaking more broadly, the research is convincing that police presence reduces crime, both in general and when it comes to many types of proactive efforts. A 2020 study, for example, estimated that one homicide is averted for every 10 officers hired. [13] Other studies of policing levels—leveraging changes such as federal grants to support hiring police, [14] 9/11 terror alerts, [15] and the randomized assignment of foot patrols [16]—have also found that more cops mean less crime. In a recent survey of criminal-justice experts, about 60% agreed or strongly agreed that increasing police budgets would decrease crime, with only about 20% disagreeing or strongly disagreeing. [17] And regarding proactive policing, a 2018 review of the literature by the National Academies found strong evidence to support several strategies, especially “hot spots” policing and “focused deterrence” programs, which concentrate efforts on particularly high-risk places and individuals, respectively, though the evidence was mixed on many other strategies. [18]

The point is not that every single stop or arrest that any police officer could possibly make would yield benefits. There are diminishing returns as activity expands to cover increasingly marginal situations. Some of the most successful policing practices involve highly targeted strategies, [19] and one recent study suggests that more experienced officers reduce crime while making fewer arrests than their less experienced counterparts. [20] What is clear is that intense drops in overall police activity after viral incidents have the potential to increase crime and that, indeed, we have seen increases in shootings and murders to go along with cops’ withdrawal.

De-policing is the result of numerous pressures; what follows is a description of these pressures as officers see them. 

Officers interviewed cited fear of becoming “the next YouTube video” or a target of negative coverage in the media, a sentiment that Devi and Fryer also report hearing frequently in focus groups with police officers. This concern is beyond the reach of government action. But interviewees further reported pressures from government actors and even within the police department. In general, officers are keenly aware of the political forces that shape their jobs.

Mayors and governors are elected, giving them incentives to bow to public opinion, even if public opinion has been shaped by an unflattering video that does not show the whole situation. Police chiefs, meanwhile, serve at the pleasure of these elected officials. District attorneys are typically elected (or appointed by elected officials) as well. Officers have a strong sense that public outcry can translate to an officer being publicly denounced, fired, or even prosecuted—even if the officer follows his training—and that government higher-ups can’t be trusted to step up in such situations. 

Numerous interviewees also cited official policy decisions that encouraged officers to be less proactive. New York Mayor Bill de Blasio ran on a platform of scaling back the city’s then aggressive “stop, question, and frisk” regime, for instance, while a new Washington State law increases the burden required for police to involuntarily stop someone for questioning (from “reasonable suspicion” to “probable cause”). [21] A consent decree has limited proactive policing in Baltimore, [22] and American Civil Liberties Union lawsuits have had similar effects on proactive-stop-based policing efforts in Chicago [23] and Milwaukee. [24] Chicago has also banned many foot pursuits, depending on the severity of the suspected crime. [25]

Scaling back other aspects of the justice system, such as lenient bail policies or otherwise releasing individuals whom police see as dangerous, can lead to a sense of futility as well. In the words of one officer, “What’s the point of locking someone up when they’re back on the street an hour later?” Additionally, while cities have lately rushed to increase funding for police, with the president’s blessing, several cut budgets in 2020 [26] amid the “defund the police” movement and an incredibly violent summer, sending an explicit message that what police do is not valued.

Some sources further lamented officials’ unwillingness to push back against the overall narrative of racially biased policing or to frankly discuss racial gaps in crime rates. This narrative, of course, lies at the heart of the current debate—in bars and salons across the country, as well as academic journals—and the repeated cycles of protests and violence since Ferguson.

These forces, when combined, add up to low morale, resulting in abrupt de-policing after viral incidents and contributing to police departments’ mounting personnel struggles. There is a long-running recruitment crisis, and many cops hired in the 1990s will be heading into retirement soon. [27] Since the summer of 2020, many officers have fled big-city departments. A study of one large department found a nearly fourfold increase in voluntary resignations following the George Floyd protests; [28] NYPD departures, including resignations and retirements, were up 75% in 2020. [29] Baltimore is hundreds of officers short of its staffing needs, which is making it hard for the department to comply with the terms of its consent decree. [30] Notably, though, many officers who quit city departments end up working in quieter suburbs, and the limits of currently available data make it unclear how staffing levels have changed nationwide in the past year. [31]

Low morale is seen in growing tensions among rank-and-file officers, their superiors, and public officials. Police in New York [32] and Chicago [33] have symbolically turned their backs on their respective mayors; there have been reports of “blue flu,” in which police protest what they see as unfair treatment of their fellow officers by calling in sick. [34]

The roots of these tensions run deep, and obviously, it would be unreasonable (and unrealistic, at any rate) to expect every government actor to base all his public statements and policy positions on how police would respond to them. What is clear is that police see themselves as increasingly embattled and under fire from the public and government officials and that this has consequences for public safety.

Fairly or not, to put it bluntly, the price we pay for alienating the police and reducing their morale is steep. Such consequences deserve consideration as officials debate broader issues of policing and criminal justice, especially when it comes to policies that directly encourage de-policing—and particularly after a viral incident. That’s when officials need to navigate a public outcry, respond appropriately to the actual case in front of them, respect the public’s need for information and transparency, and avoid another wave of violence.

Since the spread of cell-phone video and body cameras, viral incidents have run the gamut. Some involve shocking abuses, such as the killing of George Floyd in Minneapolis, which has resulted in a murder conviction for former police officer Derek Chauvin, as well as the filing of state and federal charges against the other officers involved. Other incidents involve a public outcry following incidents where a police officer used force appropriately. A case in point occurred in Columbus, Ohio, in April—where an officer shot a teenage girl who was in the process of attacking another girl with a knife. Still other viral incidents were ambiguous or debatable, including cases where an officer had to make fast decisions and performed less than ideally, where an officer mistakenly but genuinely perceived a serious threat, or where a lack of video evidence makes it unclear what even happened.

Citizens and media outlets have every right to debate police officers’ conduct; government officials have no control over that, nor should they. What they can control is their own behavior in the wake of a viral policing incident and the way they manage police officers amid such turmoil. The message that cops get from their leaders during this process is a key factor in how they subsequently go about their jobs.

Police officers realize that bad cops exist and need to be dealt with. What they fear is a rush to judgment—being thrown under the bus by public officials—before they have a chance to defend themselves. Even as police officers rarely face outright criminal prosecution for using force, these fears are sometimes justified.

In the Columbus case, those jumping to express outrage included Ohio Senator Sherrod Brown, [35] who said on social media that, “while the verdict was being read in the Derek Chauvin trial, Columbus police shot and killed a sixteen-year-old girl,” adding that “she should be alive right now.” In a less prominent Washington, DC, incident, three officers were grappling with a man illegally carrying a loaded “ghost gun,” and, in trying to bring the man under control, one of the officers struck the man numerous times. The incident—which, again, involved an armed suspect resisting arrest—was resolved without lethal force, but the video spread online, the police chief quickly said that he was “embarrassed, disturbed, disheartened, and ashamed” by the video, and the officers were placed on leave, pending disciplinary and criminal investigations. [36]

Police officers are more willing to be proactive and engage with the community when they feel confident that they’ll be treated fairly if the situation escalates and they need to use force. This is clear from my qualitative interviews as well as from academic research.

In a series of studies in the wake of Ferguson, Justin Nix and Scott E. Wolfe surveyed law-enforcement actors, including sheriffs’ deputies and management-level officers. [37] These surveys asked numerous questions, such as whether negative publicity had affected motivation and proactivity, whether officers were willing to partner with the community to solve problems, and how the officers perceived the fairness of their departments and their own legitimacy. Nix and Wolfe ran statistical models to see how these variables related to one another. The overarching theme is that when police have a sense of fair treatment and legitimacy, the Ferguson Effect on their motivation is less pronounced.

In a 2016 study, [38] Nix and Wolfe surveyed more than 500 deputies at a southeastern sheriff ’s department, with an impressively high response rate (85%). The key finding was that deputies who thought that their organizations and command staff treated them fairly “were significantly less likely to report experiencing reduced motivation due to negative publicity surrounding law enforcement in the six months following Michael Brown’s death in Ferguson.” They were also less likely to say that their colleagues had become less motivated. Simply put, officers who think that their superiors are fair are less sensitive to the pressures laid out above.

In another study, [39] Wolfe and Nix concluded: “Deputies who reported being less motivated as a result of negative publicity surrounding law enforcement . . . indicated less willingness to partner with the community.” However, “officers who perceived fair treatment from their organization were more likely to engage in community partnerships,” and “confidence in one’s authority as a police officer appears to protect against the negativeeffects of media coverage of high-profile incidents like Ferguson.” Yet another study by Wolfe and several coauthors, relying on a survey of border patrol officers, found that a sense of fairness was especially important to job satisfaction when officers faced uncertainty, including uncertainty from negative publicity. [40]

These studies are part of a much broader literature on “organizational justice,” which boils down to whether employees believe that their organizations treat them fairly—including beliefs that promotions and other perks are distributed fairly, that superiors adhere to legitimate processes and allow employees a voice, and that superiors are generally respectful of their subordinates. This research tends to find that employees perform better when they hold such beliefs, including workers in criminal-justice fields, specifically, as Wolfe showed in a meta-analysis (coauthored with Spencer Lawson) of more than 100 studies. [41]

A closely related concept is “procedural justice,” which applies not only within police departments but also to the ways that law enforcement interacts with the public. This refers to the principles of fair processes, transparency, voice, and impartiality. As the Obama administration’s 2015 policing task force noted, “Research on internal procedural justice tells us that officers who feel respected by their supervisors and peers are more likely to accept departmental policies, understand decisions, and comply with them voluntarily.” [42] This was evident, for instance, in a survey of Buenos Aires police officers. [43] A 2016 study based on a survey of a large urban force in the U.S. similarly found that “when officers were in a procedurally fair department, they were more likely to trust and feel obligated to obey their supervisors, less likely to be psychologically and emotionally distressed, and less likely to be cynical and mistrustful about the world in general and the communities they police in particular,” as well as more likely to endorse “democratic,” community-focused policing. [44]

In a 2017 Pew survey, however, only 45% of officers said that they thought their departments’ disciplinary processes were fair, and only 27% believed that low-performing officers were held to account. [45] Police discipline is often seen as arbitrary, [46] a “jackpot,” [47] with good cops hamstrung while bad cops escape, thanks to extensive appeal options. [48]

Deep fixes to police discipline could take up an entire article on their own, [49] as could the questions of how officers should be trained and when they should be prosecuted. [50] But in the more immediate context of viral incidents, organizational and procedural justice means a fair, careful application of the procedures and laws that exist. This is largely common sense; but in my interviews, I had sources outline more specifically how officials should respond when faced with this phenomenon. What follows is my synthesis of their ideas, mixed with some insights from other sources.

Clear communication between police officials and line-level officers is a must, both before and after these incidents occur. Officers should obviously know what is expected of them and be properly trained to perform their duties. They should know what they are not allowed to do and what consequences they face otherwise. They should further understand how the process works for investigating and resolving complaints and allegations, so that they do not feel blind-sided if they or another officer face one.

After an incident occurs, it is crucial for higher-level police officials to be transparent about what is happening with the accused officer and why, to convey that officers who do their jobs properly will have the support of the agency, and to impress on officers that this is a time when the city needs a strong, professional police presence the most. Officials should also make themselves available to hear the concerns of their officers. During these times, officers feel under attack by the public and media, and a sense that their employer, at least, will “have their back” is crucial.

In cases where the accused officer did nothing wrong, officials should defend the officer as soon as it is appropriate to do so and explain to the public why police are trained to use force the way the officer did. Police officers and civilians often see the same video very differently for a host of reasons, including officers’ training. For example, officers are more attuned to the split-second nature of many of these decisions and are also aware of what police are generally trained not to do, such as shoot a suspect in the leg.

The timely release of body-camera footage can be especially helpful. Since use-of-force situations are often hectic and confusing on video, some departments have taken to releasing not only raw footage but also detailed breakdowns of these incidents where key moments are annotated, slowed down, or paused. While there is always some concern about such efforts, particularly when PR firms are involved, [51] they can be key to getting the public to understand what happened and why.

In cases where the officer was wrong, officials should emphasize to other officers and the public that such misbehavior is not tolerated—and does not taint the vast majority of officers who do their jobs in good faith. Where possible, officials should spell out to the public how officers are trained so that it is clear that the officer was acting outside what he or she was taught. (Proactively educating the public about normal training, the reasons for it, and the limits of police force would be welcome, as well.)

It’s worth pointing out that, in dealing with a specific officer at the center of a controversial incident, hasty disciplinary measures can be ineffective, given that public employment comes with a variety of statutory, [52] constitutional, [53] and often union-won contractual protections. For example, the Atlanta Police Department fired the officer who killed Rayshard Brooks last year the day after the incident, but the officer was reinstated in May of this year because the department had not followed the process laid out in the law, which included written notice 10 working days in advance. [54] (The officer’s criminal prosecution for the shooting is ongoing.)

These protections are policy variables in themselves, of course, and reforming them in certain ways could be a worthy endeavor, so long as accused officers retain the ability to mount a vigorous defense. [55] Nonetheless, carefully adhering to due process not only assures good cops that they’ll be treated fairly but also can be a prerequisite to punishing bad cops.

We know about de-policing because departments across the country track their officers’ activity. In principle, it would be easy to use those very same data to fight the problem, simply by imposing consequences on officers whose activity dropped in a clearly deliberate way.

Yet such efforts are difficult to undertake in practice and can backfire if poorly implemented. Hardly any management practice is less popular among cops themselves, the public, and politicians than “quotas.” Going back to the 1970s, numerous states have passed laws banning quotas, amid complaints that such practices force cops to make dubious or even illegal stops to meet their numbers. [56] On the other side of the spectrum, several officers told me that the quotas they’ve faced have always been trivial to meet, and thus ineffective for their weakness rather than troubling for their strength. Any official seeking to increase discretionary activity must consider the potential for such pushback, unintended consequences, and inefficacy

Given that proactive policing is a big part of a police officer’s job, it is almost inevitable that objective measures of these activities will play some role in police management. In one Pew survey, about 40% of officers reported that they were subject to some sort of quota, though these were almost always informal targets rather than official requirements. [57] Further, laws banning quotas typically apply to explicit numerical requirements and do not prohibit the more general use of objective activity measures to manage police. [58]

So when police activity falls in the wake of a viral incident, what can officers’ superiors do?

Officers, leaders, and experts I spoke with outlined numerous ways that superiors can prod officers to keep their activity up. Such activity can play a role in promotions, performance reviews, and assignments to special units. More generally, police officials and mayors can communicate that they don’t want to see a decline in activity at a time when the city needs protection the most, especially if these leaders have taken the steps outlined above to assure officers that they will be treated fairly if they need to use force.

Street-level police leaders, such as sergeants, can talk to individual officers whose work has dropped off. Even simply letting it be known that superiors are keeping close tabs on the quantity and quality of the work that is being done can have an effect. Further, departments can publicly highlight acts of good policing, as, for example, the NYPD frequently does on social media; this both praises the best officers and illustrates the benefits of policing for the public.

A somewhat out-of-the-box idea is to train officers in tactics that, among other advantages, look less violent on video. Some departments, for instance, have offered training in Brazilian jujitsu, a martial art focused on grappling rather than strikes. Early data from the Marietta, Georgia, police department suggest that officers trained this way are less likely to resort to Tasers and less likely to be injured themselves. [59] Further study of this technique in other departments, ideally with random assignment, is worth pursuing.

Meanwhile, a seemingly obvious solution to de-policing—disciplining officers who pull back—is rare outside of extenuating circumstances. It’s simply hard to punish a police officer for not doing the discretionary aspects of his job, so long as the officer has the “plausible deniability” to claim that he didn’t see anything that needed addressing. The general sense within law enforcement is that if a police officer doesn’t want to be proactive and isn’t motivated by carrots such as special-unit assignments, there’s no way to truly force the issue.

One can certainly imagine some sort of punishment for officers who deliberately cut back their activity, but the pitfalls are many. Like quotas, this punishment would face objections from police, unions, and much of the public because of its potential for encouraging dubious stops, especially when enforced in a city on edge and distrustful of the police, thanks to a viral video. It could also prove hard to enforce in a fair way, given the broader issues with police disciplinary processes noted above, not to mention the difficulty of proving a deliberate cutback by a specific officer.

Police are public employees with superiors, and it is not unreasonable to ask those superiors to step in when officers stop doing a major part of their jobs, with disastrous consequences for public safety. But the realistic options are limited.

The process of de-policing is a series of cost-benefit scenarios managed by various actors. Elected officials facing a public outcry weigh the political advantages of joining in, as well as their own convictions regarding the incident at hand and police use of force in general, against the risks of alienating police and increasing crime. Police chiefs weigh the advantages of supporting their officers against the risks of angering their elected bosses. Down the chain of command, officers weigh the benefits of being proactive and stopping crime—a reason many became cops, to begin with—against the risk that they could be the next viral video.

There is no easy solution to the problem of de-policing, but there are ways to minimize it, at least when the will to minimize it exists. Officials can avoid turning de-policing into official policy by, for example, making it harder for police to effect stops. They can work to cultivate a sense of organizational justice, in which police feel that they will be treated fairly if they need to use force. And to a limited extent, higher-ups can encourage police to keep their activity up, even when tensions run high and the temptation to de-police grows.

Please see Endnotes in PDF

Robert VerBruggen is a fellow at the Manhattan Institute where he provides policy research, writes for City Journal, and contributes to special projects and initiatives in the President’s office. Having held roles as Deputy Managing Editor of National Review, Managing Editor of The American Conservative, Editor at RealClearPolicy, and Assistant Book Editor at The Washington Times, VerBruggen writes on a wide array of issues including economic policy, public finance, health care, education, family policy, cancel culture, and public safety. VerBruggen was a Phillips Foundation Journalism Fellow in 2009 and a 2005 winner of the Chicago Headline Club Peter Lisagor Award. He holds a B.A. in journalism and political science from Northwestern University.