Violent crime in many American cities began rising in the second half of 2014, after two decades of decline. The Major Cities Chiefs Association convened an emergency session in August 2015 to discuss the double-digit surge in violence besetting its member police departments. Homicides at that point were up 76% in Milwaukee, 60% in St. Louis, and 56% in Baltimore, compared to the same period in 2014; the average homicide increase among 35 cities surveyed by the Association was 19%. “Crime is the worst I’ve ever seen it,” said St. Louis Alderman Joe Vacarro in May. July 2015 was the bloodiest month in Baltimore since 1972, with 45 people killed in 30 days. Arrests, summons, and pedestrian stops had dropped in many cities, where data on such police activity were available.
I first noted the rising violence in a WSJ op-ed in May 2015... I posited a reason for it: officers were backing off of proactive policing in reaction to the hostility they were encountering in urban areas. Officers had told me about being surrounded by angry, jeering crowds who cursed and threw water bottles and rocks at them when they tried to make an arrest.
The violence surge continued into fall. Homicides in Baltimore reached their highest per capita rate in the city’s history. In October, Attorney General Loretta Lynch brought together over one hundred police chiefs, mayors, and federal prosecutors in another emergency meeting to strategize over the rising homicide rates. FBI Director James Comey noted in an October speech that “Most of America’s 50 largest cities have seen an increase in homicides and shootings this year, and many of them have seen a huge increase.”
The media confirmed the experience of law enforcement officials. In September, the data blog FiveThirtyEight found a 16% increase in homicide in the 60 largest cities so far that year. The Washington Post found a nearly 17% increase in homicides in 2015 in the 50 top cities, the largest one-year increase since 1993. (Had homicides dropped 17% in one year, mayors and police chiefs across the country would have been popping champagne corks.) The Brennan Center for Justice estimated a nearly 15% increase in homicides in 25 of the 30 largest cities in 2015. This year, many cities are still struggling with crime increases: the number of homicides in Chicago through March 11, for example, had nearly doubled from the same period in 2015, notwithstanding that homicides had already increased nearly 13% in all of 2015. Homicides in Los Angeles were up nearly 28% through March 9 and violent crime up 13%.
I first noted the rising violence in a Wall Street Journal op-ed in May 2015. And having spoken with police officers across the country, I posited a reason for it: officers were backing off of proactive policing in reaction to the hostility they were encountering in urban areas. Officers had told me about being surrounded by angry, jeering crowds who cursed and threw water bottles and rocks at them when they tried to make an arrest. Suspects and bystanders stuck cell phones in officers’ faces and refused to comply with lawful orders. Officers were continuing to answer 911 calls with alacrity, but in that large area of discretionary policing—getting out of a squad car at 1 a.m., for example, to question someone who appears to have a gun or may be casing a target—many officers were deciding to simply drive on by rather than risk a volatile, potentially career-ending confrontation that they were under no obligation to instigate.
I dubbed this latest outbreak of depolicing and the resulting emboldening of criminals the “Ferguson effect,” picking up on a term first used by St. Louis police chief Sam Dotson. The police shooting of Michael Brown in Ferguson, Mo., in August 2014 had triggered riots, die-ins, and cop assassinations. It gave rise to the angry Black Lives Matter protest movement, which asserts that racist (insert “white” whenever circumstances allow) cops are engaged in a killing spree against unarmed black men. Activists and academics denounced pedestrian stops and public order policing (otherwise known as Broken Windows policing) as racially biased and oppressive. As a result, officers were doing a lot less of such discretionary enforcement. Arrests in St. Louis city and county, for example, dropped a third after the Brown shooting; misdemeanor drug arrests in Baltimore dropped a third through November 2015.
The relationship between depolicing and crime was hardly a novel discovery; a 2005 University of Washington study of depolicing in Cincinnati following the anti-cop riots of 2001 had found a drop in arrests and a surge in crime in the city’s black areas. And I was hardly the only person to hear from police officers about their reluctance to engage in proactive enforcement. FBI director Comey reported that cops in one big city precinct “described being surrounded by young people with mobile phone cameras held high, taunting them the moment they get out of their cars.” The cops told Comey: “’We feel like we’re under siege and we don’t feel much like getting out of our cars.’” In November 2015, the acting head of the Drug Enforcement Administration, Chuck Rosenberg, said his own conversations with police officials had persuaded him that cops were worried about becoming the “next viral video,” because even if they did everything right, they could “still end up on the evening news.”
Despite the ample evidence of officers pulling back from discretionary enforcement, my May op-ed unleashed considerable opposition. The American Society of Criminologists sent out an unprecedented alert to its members in June 2015, asking them to try to disprove the Ferguson effect. The ASC helpfully provided sample rebuttals of my op-ed prepared by The Sentencing Project, an anti-incarceration advocacy group. The ASC worried that my article would somehow impede the push for federal “sentencing reform.”
It is in high-crime black neighborhoods where the police are backing off the most under the relentless charge that they are racist. And it is there where a fall-off in proactive policing is going to produce the biggest negative impact.
Criminologists David Pyrooz of the University of Colorado Boulder, Scott Wolfe of the University of South Carolina, and Scott Decker of Arizona State University were among those who responded to the op-ed, publishing a complex econometric analysis of the Ferguson effect. They later defended their analysis on this site against criticism by me and others. Pyrooz and his co-authors modelled monthly rates of change in crime rates in 81 of the 105 largest cities in the country in the twelve months before and after Ferguson. Though the rate of change in violent crime increased ten times after Ferguson, that tenfold increase was not enough to be deemed statistically significant, they say. The authors concluded that “there is no systematic evidence of a Ferguson Effect on aggregate crime rates throughout the large U.S. cities . . . in this study.”
But that 81-city average masked important changes in the nation’s crime picture. Before Ferguson, individual cities’ crime rates were largely moving downwards together; after Ferguson, crime trajectories were all over the map. Crime in some cities was still down; in others, it was way up. Variance in homicide rates increased nearly six times. And the cities with the highest homicide surges were exactly what the Ferguson effect would predict: cities with large black populations, smaller white populations, and already high rates of violent crime.
Recall that the Black Lives Matter movement has sent out a relentless message that cops are the biggest threat facing young black men today. It directs a non-stop flow of racially-tinged animus at the police profession. A typical march that I observed this Fall on Fifth Ave. in New York City featured “Fuck the Police” and “Racism Is the Disease, Revolution Is the Cure” T-shirts, “Stop Police Terror” signs, and “Hey Hey, Ho Ho, Racist Cops Have Got to Go” chants. Such rhetoric has influenced street behavior. During Black Lives Matter demonstrations in Ferguson, Mo., in August 2015, 18-year-old Tyrone Harris opened fire at police officers, and was shot in response. A crowd pelted the cops with frozen water bottles and rocks, wounding three officers, while destroying three police cars and damaging businesses. “We’re ready for what? We’re ready for war,” some protesters chanted. In Cincinnati, a small riot broke out in late July 2015 when the police arrived at a drive-by shooting scene, where a 4-year-old girl had been shot in the head and critically injured. Bystanders loudly cursed at officers who had started arresting suspects at the scene on outstanding warrants.
A recent poll of New York City officers found that active resistance to arrest had increased over the last two years. “There’s a total lack of respect out there for the police,” a female sergeant in New York told me last year, echoing cops across the country. It is in predominantly black neighborhoods where police worry that a videoed use of force will land them on TV in the role of racist cop of the week and where officers tell each other: “‘If you get out of your car, you’re crazy, unless there’s a radio call,’” in the words of a cop in the Los Angeles Police Department’s Newton Division.
In short, it is in high-crime black neighborhoods where the police are backing off the most under the relentless charge that they are racist. And it is in high-crime neighborhoods where a fall-off in proactive policing is going to produce the biggest negative impact. It is in those neighborhoods where informal social controls — above all families — have most broken down and where policing most critically takes up the slack. The per capita rate of shootings, for example, is 81 times higher in predominantly black Brownsville, Brooklyn, than in nearby Bay Ridge. Not surprisingly, the per capita rate of pedestrian stops is also higher in Brownsville than in Bay Ridge—15 times higher—because every shooting will call forth a police response geared to interrupting retaliatory gunfire. There is less public order policing to begin with in low-crime areas, because there is less disorder, but even if police have become equally gun-shy about discretionary enforcement in low-crime areas, the consequences would be less.
The Pyrooz study confirms the relationship between depolicing and an increase in violent crime. Not only have homicides spiked in predominantly black cities, but robberies — the quintessential urban street crime — registered what even the authors deem a statistically significant increase in all 81 cities. Pyrooz and his colleagues struggle mightily against their own findings, however. They attribute the sharp post-Ferguson rise in homicides in predominantly black cities not to depolicing but to the fact that such cities were somehow “primed” for a homicide increase. This is a circular, pseudo-explanation. How do we know that those cities were “primed” for a post-Ferguson crime increase? Because they had such a post-Ferguson crime increase. And why did they have a crime increase? Because they were “primed” for it. They offer a strange analogy to explain their ad hoc “priming” hypothesis: It’s like a “stock portfolio,” they say, “where some holdings increase even in a down market.” But the post-Ferguson crime increases were not random and unpredictable fluctuations among a diversified market basket of data points; they were predictable effects of a racially-driven depolicing phenomenon. If being a high-crime city “primes” that city for further crime increases, the authors need to explain why crime dropped in those same high-crime cities over the previous two decades. (Answer: because of the data-driven proactive policing revolution that started in New York City and spread nationwide.) And if it was obvious that the crime drop in black cities was about to reverse itself—by coincidence at exactly the moment when the Black Lives Matter movement kicked into high gear — the authors might have alerted us to that reversal ahead of time.
The authors claim that their study refutes my Ferguson effect hypothesis because the title of my original Wall Street Journal op-ed was: “The New Nationwide Crime Wave.” Since the post-Ferguson crime increases were not uniform across all 81 of their modelled cities, the title of the op-ed was wrong, they say, and therefore the Ferguson effect hypothesis was also wrong. I did not write the title of my op-ed, consistent with the usual practice, nor did I see it before publication. But in any case, nothing in that first op-ed or in its several follow-ups implied that crime needed to go up uniformly in every American city for there to be a Black Lives Matter-generated pull-back from proactive enforcement and a resulting effect on crime.
Pyrooz, Wolfe, and Decker repeatedly put themselves on the side of “science” against some presumed group of science-deniers. But their analysis of the motivations of people concerned about depolicing is anything but “scientific,” consisting merely of a dizzying set of non sequiturs and ungrounded speculation.
The trio alleges that their critics “want a Ferguson effect to exist” (emphasis in original) and that those critics somehow want a Ferguson effect to exist because they “believe that police are not professional enough, not trained well enough, and too hesitant under pressure to withstand the new reality that their actions can be caught on camera.” Such a belief, they say, is the true “anti-cop” position, compared to the “pro-cop” belief that “a vast majority of officers are well-trained professionals who can withstand pressure from public scrutiny.” The trio then asserts: “If one accepts this premise, we would certainly not expect large groups of officers to de-police and cause higher crime rates in our communities.”
As an initial matter, it is impossible to tell here and elsewhere whether they are arguing that depolicing is not going on, or that it should have no effect on crime rates, or that depolicing would cause higher crime rates but officers are too professional to cut back on discretionary enforcement. In any case, it is an empirical matter whether it is going on or not, regardless of whether acknowledging it is “pro-cop” or “anti-cop.”
The outrage among Black Lives Matter allies at the mere suggestion that the police may be backing off of proactive enforcement is the strangest aspect of this whole episode... Isn’t that how political pressure is supposed to work?
Pyrooz, Wolfe, and Decker need to get out there and talk to some officers. “Public scrutiny,” as they call it, is not the problem impeding urban policing today; the problem is hatred, aggression, and sometimes violent resistance to arrest. A police officer in Los Angeles tells me: “Several years ago I could use a reasonable and justified amount of force and not be cursed and jeered at. Now our officers are getting surrounded every time they put handcuffs on someone. The spirit and the rhetoric of this flawed movement is causing more confrontations with police and closing the door on the gains in communication we had made before it began.” Cops are human. It is wholly unrealistic to think that the relentless propaganda campaign against them, accompanied by a volatile, hostile street environment, is not going to lead many to hesitate before initiating encounters that politicians and the press have labelled as racist and that they are under no mandate to undertake.
Moreover, policing is political. The outrage among Black Lives Matter allies at the mere suggestion that the police may be backing off of proactive enforcement is the strangest aspect of this whole episode. A decline in pedestrian stops and Broken Windows policing is exactly what the activists have been demanding. Now they’re getting it. Isn’t that how political pressure is supposed to work?
The authors suggest that “we should spend more time worrying about the legitimacy crisis rather than a Ferguson effect on crime.” But it is the falsehoods about a police reign of terror spread by the Black Lives Matter movement that has brought on that “legitimacy crisis.” Since Wolfe, Decker, and Pyrooz pride themselves on their fealty to “science,” perhaps they could “scientifically” analyse police shootings in the context of crime. They would discover that police shootings of blacks are lower than what black violent crime rates would predict. Police shootings constitute a much lower share of black homicide deaths than of white and Hispanic homicide deaths. And police officers are two and a half times more likely to be fatally shot by a black man than a black man is to be fatally shot by a cop.
The three criminologists’ final argument against the Ferguson effect is the least “scientific” of all. They accuse their critics of “threatening people with the prospect that violent crime will increase if they protest police behaviours,” and of thereby implying that “the police should not be accountable for their behaviour.” No one is “threatening people with the prospect” of a violent crime increase; the violent crime increase was well underway before anyone noticed it and hypothesized a reason. It is a question of fact whether violent crime is rising in urban areas, regardless of any untoward implications the authors think such a fact would have for the Black Lives Matter movement. And no one is saying that “the police should not be accountable for their behaviour.” What analysts such as myself are saying is that the current frenzy of cop hatred is affecting proactive policing. If that effect somehow delegitimates the Black Lives Matter movement in the eyes of Pyrooz and his colleagues, so be it. “Science” is silent on such a matter.
This piece originally appearaed Quillette