Internet Explorer 11 is not supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

The Clouded Legacy of ‘Broken Windows’ Policing

It's been with us for nearly four decades, but we still can't definitively answer the question of whether it prevents crime in our cities.

Mayor Rudy Giuliani and Commissioner William Bratton conduct a press conference on New York City's crime statistics. (Photo courtesy New York Daily News)
We are coming up on the 40th anniversary of the most influential piece of journalism in the history of law enforcement: “Broken Windows,” by George L. Kelling and James Q. Wilson, which appeared in The Atlantic in March of 1982. Modestly presented but compelling in its arguments, their article generated a revolution in policing that is very much with us today.

Kelling and Wilson argued that what residents of cities feared most was not violent crime — though this obviously concerned them — but disorder: tangible signs of neighborhood neglect that grow out of minor offenses and lead to a chaotic experience on the streets and a gradual increase in crimes of all kinds.

Their icon for this decline in community life was a beguilingly simple one: a window. “One broken window,” they wrote, “is a signal that no one cares, and so breaking more windows costs nothing. … A piece of property is abandoned, weeds grow up, a window is smashed. Adults stop scolding rowdy children; the children, emboldened, become more rowdy. … Families move out, unattached adults move in. … Fights occur. Litter accumulates. People start drinking in front of the grocery; in time, an inebriate slumps on the sidewalk and is allowed to sleep it off. Pedestrians are approached by panhandlers.”
Worst of all, the unpunished perpetrators of disrespectful acts graduate to truly frightening offenses: burglary, armed robbery, violent attacks on those they dislike. The law-abiding residents of the neighborhood become dejected, sullen and solitary.

Fortunately, the authors wrote, there is a remedy for all this: Police must target the most routine of offenses — not just vandalism but subway fare evasion: graffiti on building walls; illegal dice games that end in bitter quarrels; open containers of alcohol carried in full view on the street; urinating in public.

Crack down on all of these, Kelling and Wilson insisted, and public order would return. The community would come to life again. Equally important, remember that the fare-jumpers of today turn out to be the violent criminals of tomorrow. Stop them and deal with them right at the beginning, and there will be dramatically fewer felonies in the future. “Disorder and crime,” the article concluded, “are usually inextricably linked in a kind of developmental sequence.”

IN THE YEARS THAT FOLLOWED PUBLICATION of “Broken Windows,” police departments all over the country began to apply its proposed remedies. Those were associated most conspicuously with New York City Mayor Rudolph Giuliani and Police Commissioner William Bratton, but they were a genuinely national phenomenon.

“Broken windows policing,” or “quality of life law enforcement,” as it was known in much of the country, was the spark for a whole host of innovations that seemed to follow from its premises. Most important was the doctrine that the job of police was not simply to respond to crime but to prevent it — an idea that goes back to Sir Robert Peel in London in the 19th century but which found a new home in American police departments in the 1980s and 1990s. Another was CompStat, a technological effort pioneered in New York City to document the incidence of serious crime in a community and hold district police commanders accountable for stopping it. This was linked to “predictive policing,” the identification of violent-crime hot spots and the flooding of those enclaves with a heavy police presence. And most notably, there was “stop and frisk,” the practice of allowing police officers the discretion to detain, question and examine individuals on urban streets based on “reasonable suspicion” that they might have committed a crime, rather than the older, more restrictive requirement that they find “probable cause” of an offense.

The influence of that 1982 article spread far beyond the immediate confines of municipal police work. The U.S. Justice Department employed it in its “Weed and Seed” anti-crime initiative. Downtown chambers of commerce created business improvement districts, in which commercial businesses hired civilian monitors to detect and report littering, loitering and non-criminal activity that seemed to erode the sense of order in a central city. Public schools started using it, suspending pupils for minor instances of class disruption that were perceived to bring disorder to an entire classroom.

ALL OF THESE STRATEGIES are, to one extent or another, the children of “broken windows.” The question for us now is what they have accomplished—or, more generally, how valid was the whole “broken windows” idea in the first place. Was the precipitous decline of crime in the 1990s and later a clear consequence of this approach to policing? Or was there simply a correlation that its proponents mistakenly labeled as a cause?
Given four decades of experience with ”broken windows,” one might think it possible to reach some definitive conclusion about what it has accomplished. But it isn’t. There are a multitude of variables in play in any community that implements a new form of policing; tracing out the impact of any one of them is an extremely difficult exercise.

More disturbingly, “broken windows” has, like so many other interventions in present-day public life, become highly vulnerable to politics and ideology. Conservatives generally believe it has worked and continues to work; progressives tend to denounce it as counterproductive and often racist.

Bernard Harcourt, a Columbia law professor, sees “a tremendous amount of disorder that erupts as a result of broken windows policing, with complaints skyrocketing, with settlements of police misconduct cases skyrocketing, and of course with incidents, brutal incidents, all of a sudden happening at a faster and faster clip."

A study conducted at Northeastern University in 2018 declared flatly that “the body of evidence for the broken windows theory does not stand,” as the researchers wrote in the Annual Review of Criminology. “There is no consistent evidence that disorder induces higher levels of aggression or makes residents feel more negative toward the neighborhood.”

There are equally forceful arguments for the opposite point of view. Matt DeLisi, coordinator of criminal justice studies at Iowa State University, wrote of the progressive polemicists that “they’re ideologically opposed to the proactive policing that Broken Windows fosters, because it draws sharp moral lines and is unafraid to make judgments about environments and behaviors. … In the real world, the impact of Broken Windows has been profound.”

The most rigorous research I have come across, conducted by the George Mason Center for Evidence-Based Crime Policy, doesn’t come down on either side. “There is no clear answer,” the researchers wrote, “as to the link between crime and disorder and whether existing research supports or refutes broken windows theory.”

THERE DOES SEEM TO BE SOME CONSENSUS around the idea that “broken windows,” in some cities in the first couple of decades of the new century, fell into overuse. Kelling himself thought so toward the end of his life. “So do I worry about the implementation of broken windows? A whole lot … because it can be done very badly,” he confessed.

More specifically, “broken windows” policing came to be confused in many places with zero tolerance, a notion that most believers in “broken windows” reject. Zero tolerance means virtually certain arrest and arraignment for the most minor offenses, with little flexibility on the arresting officer’s part. “Broken windows,” in its original formulation, calls for officer discretion and heavy involvement by the community in restoring order to its streets.

Zero tolerance led to what is now conceded to have been an excessive use of stop and frisk. Police officers in some cities were judged on the number of stop-and-frisk actions they conducted. In 2008 in New York City, police made nearly 250,000 stops for what was labeled “furtive movement.” Fewer than 1 percent of these stops turned up an illegal weapon. In 2016, the U.S Justice Department branded the Baltimore stop-and-frisk program as racist due to what it called unwarranted use of the method in Black neighborhoods.

If there’s one alternative explanation for fluctuating urban crime rates that has gained serious traction, it’s ”collective efficacy,” developed by the Harvard sociologist Robert J. Sampson. Collective efficacy is essentially the degree of cohesion that exists in a neighborhood, the mutual trust through which neighbors help each other out of difficult situations, or return a lost wallet without rifling through it. In Sampson’s view, it is this form of efficacy, not quality of life law enforcement, that has the most effect on the local crime rate.

It’s a plausible theory, but it leaves a crucial question unanswered. Perhaps it’s the level of order maintenance by police that sustains collective efficacy in the first place. Indeed, a recent study in Lowell, Massachusetts, concluded that the maintenance of order in a community, whether by police or residents, makes more difference in preventing felonies than a hard-line arrest policy.

So there we are. “Broken windows” policing may have given us safer streets, or it may not have. There is no disputing that crime declined in the years after it was widely implemented, but as the statisticians like to say, that could simply be a matter of correlation, not causation. Perhaps it could. Personally, though, I lean toward the conclusion that some correlations are simply too strong to ignore. The link between “broken windows” and crime looks to me like one of them.
Alan Ehrenhalt is a contributing editor for Governing. He served for 19 years as executive editor of Governing Magazine. He can be reached at
From Our Partners