In 1969, Philip Zimbardo, a psychologist from Stanford University, ran an interesting field study. He abandoned two cars in two very different places: one in a mostly poor, crime-ridden section of New York City, and the other in a fairly affluent neighborhood of Palo Alto, Calif. Both cars were left without license plates and parked with their hoods up.
After just 10 minutes, passersby in New York City began vandalizing the car. First they stripped it for parts. Then the random destruction began. Windows were smashed. The car was destroyed. But in Palo Alto, the other car remained untouched for more than a week.
Finally, Zimbardo did something unusual: He took a sledgehammer and gave the California car a smash. After that, passersby quickly ripped it apart, just as they’d done in New York.
This field study was a simple demonstration of how something that is clearly neglected can quickly become a target for vandals. But it eventually morphed into something far more than that. It became the basis for one of the most influential theories of crime and policing in America: “broken windows.”
Thirteen years after the Zimbardo study, criminologists George L. Kelling and James Q. Wilson wrote an article for The Atlantic. They were fascinated by what had happened to Zimbardo’s abandoned cars and thought the findings could be applied on a larger scale, to entire communities.
“The idea [is] that once disorder begins, it doesn’t matter what the neighborhood is, things can begin to get out of control,” Kelling tells Hidden Brain.
In the article, Kelling and Wilson suggested that a broken window or other visible signs of disorder or decay — think loitering, graffiti, prostitution or drug use — can send the signal that a neighborhood is uncared for. So, they thought, if police departments addressed those problems, maybe the bigger crimes wouldn’t happen.
“Once you begin to deal with the small problems in neighborhoods, you begin to empower those neighborhoods,” says Kelling. “People claim their public spaces, and the store owners extend their concerns to what happened on the streets. Communities get strengthened once order is restored or maintained, and it is that dynamic that helps to prevent crime.”
Kelling and Wilson proposed that police departments change their focus. Instead of channeling most resources into solving major crimes, they should instead try to clean up the streets and maintain order — such as keeping people from smoking pot in public and cracking down on subway fare beaters.
The argument came at an opportune time, says Columbia University law professor Bernard Harcourt.
“This was a period of high crime, and high incarceration, and it seemed there was no way out of that dynamic. It seemed as if there was no way out of just filling prisons to address the crime problem.”
An Idea Moves From The Ivory Tower To The Streets
As policymakers were scrambling for answers, a new mayor in New York City came to power offering a solution.
Rudy Giuliani won election in 1993, promising to reduce crime and clean up the streets. Very quickly, he adopted broken windows as his mantra.
It was one of those rare ideas that appealed to both sides of the aisle.
Conservatives liked the policy because it meant restoring order. Liberals liked it, Harcourt says, because it seemed like an enlightened way to prevent crime: “It seemed like a magical solution. It allowed everybody to find a way in their own mind to get rid of the panhandler, the guy sleeping on the street, the prostitute, the drugs, the litter, and it allowed liberals to do that while still feeling self-righteous and good about themselves.”
Giuliani and his new police commissioner, William Bratton, focused first on cleaning up the subway system, where 250,000 people a day weren’t paying their fare. They sent hundreds of police officers into the subways to crack down on turnstile jumpers and vandals.
Very quickly, they found confirmation for their theory. Going after petty crime led the police to violent criminals, says Kelling: “Not all fare beaters were criminals, but a lot of criminals were fare beaters. It turns out serious criminals are pretty busy. They commit minor offenses as well as major offenses.”
The policy was quickly scaled up from the subway to the entire city of New York.
Police ramped up misdemeanor arrests for things like smoking marijuana in public, spraying graffiti and selling loose cigarettes. And almost instantly, they were able to trumpet their success. Crime was falling. The murder rate plummeted. It seemed like a miracle.
The media loved the story, and Giuliani cruised to re-election in 1997.
George Kelling and a colleague did follow-up research on broken windows policing and found what they believed was clear evidence of its success. In neighborhoods where there was a sharp increase in misdemeanor arrests — suggesting broken windows policing was in force — there was also a sharp decline in crime.
By 2001, broken windows had become one of Giuliani’s greatest accomplishments. In his farewell address, he emphasized the beautiful and simple idea behind the success.
“The broken windows theory replaced the idea that we were too busy to pay attention to street-level prostitution, too busy to pay attention to panhandling, too busy to pay attention to graffiti,” he said. “Well, you can’t be too busy to pay attention to those things, because those are the things that underlie the problems of crime that you have in your society.”
Questions Begin To Emerge About Broken Windows
Right from the start, there were signs something was wrong with the beautiful narrative.
“Crime was starting to go down in New York prior to the Giuliani election and prior to the implementation of broken windows policing,” says Harcourt, the Columbia law professor. “And of course what we witnessed from that period, basically from about 1991, was that the crime in the country starts going down, and it’s a remarkable drop in violent crime in this country. Now, what’s so remarkable about it is how widespread it was.”
Harcourt points out that crime dropped not only in New York, but in many other cities where nothing like broken windows policing was in place. In fact, crime even fell in parts of the country where police departments were mired in corruption scandals and largely viewed as dysfunctional, such as Los Angeles.
“Los Angeles is really interesting because Los Angeles was wracked with terrible policing problems during the whole time, and crime drops as much in Los Angeles as it does in New York,” says Harcourt.
There were lots of theories to explain the nationwide decline in crime. Some said it was the growing economy or the end of the crack cocaine epidemic. Some criminologists credited harsher sentencing guidelines.
In 2006, Harcourt found the evidence supporting the broken windows theory might be flawed. He reviewed the study Kelling had conducted in 2001, and found the areas that saw the largest number of misdemeanor arrests also had the biggest drops in violent crime.
Harcourt says the earlier study failed to consider what’s called a “reversion to the mean.”
“It’s something that a lot of investment bankers and investors know about because it’s well-known and in the stock market,” says Harcourt. “Basically, the idea is if something goes up a lot, it tends to go down a lot.”
A graph in Kelling’s 2001 paper is revealing. It shows the crime rate falling dramatically in the early 1990s. But this small view gives us a selective picture. Right before this decline came a spike in crime. And if you go further back, you see a series of spikes and declines. And each time, the bigger a spike, the bigger the decline that follows, as crime reverts to the mean.
Kelling acknowledges that broken windows may not have had a dramatic effect on crime. But he thinks it still has value.
“Even if broken windows did not have a substantial impact on crime, order is an end in itself in a cosmopolitan, diverse world,” he says. “Strangers have to feel comfortable moving through communities for those communities to thrive. Order is an end in itself, and it doesn’t need the justification of serious crime.”
Order might be an end in itself, but it’s worth noting that this was not the premise on which the broken windows theory was sold. It was advertised as an innovative way to control violent crime, not just a way to get panhandlers and prostitutes off the streets.
‘Broken Windows’ Morphs Into ‘Stop And Frisk’
Harcourt says there was another big problem with broken windows.
“We immediately saw a sharp increase in complaints of police misconduct. Starting in 1993, what you’re going to see is 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.”
The problem intensified with a new practice that grew out of broken windows. It was called “stop and frisk,” and was embraced in New York City after Mayor Michael Bloomberg won election in 2001.
If broken windows meant arresting people for misdemeanors in hopes of preventing more serious crimes, “stop and frisk” said, why even wait for the misdemeanor? Why not go ahead and stop, question and search anyone who looked suspicious?
There were high-profile cases where misdemeanor arrests or stopping and questioning did lead to information that helped solve much more serious crimes, even homicides. But there were many more cases where police stops turned up nothing. In 2008, police made nearly 250,000 stops in New York for what they called furtive movements. Only one-fifteenth of 1 percent of those turned up a gun.
Even more problematic, in order to be able to go after disorder, you have to be able to define it. Is it a trash bag covering a broken window? Teenagers on a street corner playing music too loudly?
In Chicago, the researchers Robert Sampson and Stephen Raudenbush analyzed what makes people perceive social disorder. They found that if two neighborhoods had exactly the same amount of graffiti and litter and loitering, people saw more disorder, more broken windows, in neighborhoods with more African-Americans.
George Kelling is not an advocate of stop and frisk. In fact, all the way back in 1982, he foresaw the possibility that giving police wide discretion could lead to abuse. In his article, he and James Q. Wilson write: “How do we ensure … that the police do not become the agents of neighborhood bigotry? We can offer no wholly satisfactory answer to this important question.”
In August of 2013, a federal district court found that New York City’s stop and frisk policy was unconstitutional because of the way it singled out young black and Hispanic men. Later that year, New York elected its first liberal mayor in 20 years. Bill DeBlasio celebrated the end of stop and frisk. But he did not do away with broken windows. In fact, he re-appointed Rudy Giuliani’s police commissioner, Bill Bratton.
And just seven months after taking over again as the head of the New York Police Department, Bratton’s broken windows policy came under fresh scrutiny. The reason: the death of Eric Garner.
In July 2014, a bystander caught on cellphone video the deadly clash between New York City police officers and Garner, an African-American. After a verbal confrontation, officers tackled Garner, while restraining him with a chokehold, a practice that is banned in New York City.
Garner died not long after he was brought down to the ground. His death sparked massive protests, and his name is now synonymous with the distrust between police and African-American communities.
For George Kelling, this was not the end that he had hoped for. As a researcher, he’s one of the few whose ideas have left the academy and spread like wildfire.
But once politicians and the media fell in love with his idea, they took it to places that he never intended and could not control.
“When, during the 1990s, I would occasionally read in a newspaper something like a new chief comes in and says, ‘I’m going to implement broken windows tomorrow,’ I would listen to that with dismay because [it’s] a highly discretionary activity by police that needs extensive training, formal guidelines, constant monitoring and oversight. So do I worry about the implementation about broken windows? A whole lot … because it can be done very badly.”
In fact, Kelling says, it might be time to move away from the idea.
“It’s to the point now where I wonder if we should back away from the metaphor of broken windows. We didn’t know how powerful it was going to be. It simplified, it was easy to communicate, a lot of people got it as a result of the metaphor. It was attractive for a long time. But as you know, metaphors can wear out and become stale.”
These days, the consensus among social scientists is that broken windows likely did have modest effects on crime. But few believe it caused the 60 or 70 percent decline in violent crime for which it was once credited.
And yet despite all the evidence, the idea continues to be popular.
Bernard Harcourt says there is a reason for that:
“It’s a simple story that people can latch onto and that is a lot more pleasant to live with than the complexities of life. The fact is that crime dropped in America dramatically from the 1990s, and that there aren’t really good, clean nationwide explanations for it.”
The story of broken windows is a story of our fascination with easy fixes and seductive theories. Once an idea like that takes hold, it’s nearly impossible to get the genie back in the bottle.
The Hidden Brain Podcast is hosted by Shankar Vedantam and produced by Maggie Penman, Jennifer Schmidt, Renee Klahr and Rhaina Cohen. Our supervising producer is Tara Boyle. You can also follow us on Twitter @hiddenbrain, and listen for Hidden Brain stories each week on your local public radio station.