Shutterstock; Paul Spella / The Atlantic

For a generation, American cops have aggressively policed many urban neighborhoods based on the premise that cracking down on minor street disorder would avert spikes in more serious crime. While reviled by some, the so-called broken-windows theory is still defended by many in law enforcement. Maybe it’s time they applied it to themselves.

According to the broken-windows theory, just as a building with one broken window is vulnerable to additional vandalism, a neighborhood with visible signs of minor disorder, such as graffiti and littering, is vulnerable to criminal invasion. “It is more likely that here, rather than in places where people are confident they can regulate public behavior by informal controls, drugs will change hands, prostitutes will solicit, and cars will be stripped,” George Kelling and James Q. Wilson wrote in The Atlantic in 1982. “That the drunks will be robbed by boys who do it as a lark, and the prostitutes’ customers will be robbed by men who do it purposefully and perhaps violently.”

Kelling and Wilson advised police to focus on maintaining order in neighborhoods that hadn’t quite tipped from disorder to violent crime, and emphasized that maintaining order requires more than arresting lawbreakers. They also argued that it requires enforcing the community standards desired by residents of a given neighborhood in a way not easily reconciled with legalistic notions of due process.

“A strong, commendable desire to see that people are treated fairly makes us worry about allowing the police to rout persons who are undesirable by some vague or parochial standard,” they wrote. Many are thus reluctant to give police the discretion to perform “a function that every neighborhood desperately wants them to perform.” For example, “arresting a single drunk or a single vagrant who has harmed no identifiable person seems unjust, and in a sense it is. But failing to do anything about a score of drunks or a hundred vagrants may destroy an entire community.”

The case for aggressively enforcing even low-level laws as well as community standards, derived from Kelling and Wilson, influenced policing across the United States, most notably in New York City and Los Angeles under police chief William Bratton, as well as in cities where his deputies were hired as police chiefs. Admirers say the theory contributed to the sharp, sustained declines in violent crime in those cities and beyond, even as critics blame it for unduly onerous policing and mass incarceration.

The attorney Ken White is one of the few people to suggest applying the logic of broken windows to police officers and departments themselves. “If tolerating broken windows leads to more broken windows and escalating crime, what impact does tolerating police misconduct have?” he asked. “Under the Broken Windows Theory, what impact could it have but to signal to all police that scorn for rights, unjustified violence, and discrimination are acceptable norms? Under Broken Windows Theory, what could be the result but more scorn, more violence, and more discrimination?”

Significant evidence substantiates the premise that police misconduct is widespread, far beyond the countless examples that are captured on cellphone cameras and posted to YouTube.

Last year, USA Today published a major database of police misconduct. “Obtained from thousands of state agencies, prosecutors, police departments and sheriffs, the records detail at least 200,000 incidents of alleged misconduct, much of it previously unreported,” the newspaper stated. The records included “more than 110,000 internal affairs investigations by hundreds of individual departments and more than 30,000 officers who were decertified by 44 state oversight agencies,” as well as “22,924 investigations of officers using excessive force, 3,145 allegations of rape, child molestation and other sexual misconduct and 2,307 cases of domestic violence.” Independent Department of Justice probes into individual police departments, such as those in Ferguson, Missouri, and Baltimore, revealed agencies that routinely and brutally violated the civil rights of residents.

Similarly strong evidence suggests that police tolerate misconduct in their ranks. In major surveys of police officers, the Pew Research Center and the National Institute of Justice found that 72 percent disagree that cops in their department who consistently do a poor job are held accountable; 52 percent believe that “it is not unusual for a police officer to turn a blind eye to improper conduct by other officers” and that most cops in their department would not report a colleague they caught driving drunk; and 61 percent think that cops “do not always report even serious criminal violations that involve the abuse of authority by fellow officers.”

Even as police officers appear to let one another off the hook, they often crack down on the residents they’re supposed to protect. In 2013 alone, for example, Ferguson’s municipal court “issued over 9,000 warrants on cases stemming in large part from minor violations such as parking infractions, traffic tickets, or housing code violations,” the Department of Justice found. “Jail time would be considered far too harsh a penalty for the great majority of these code violations, yet Ferguson’s court routinely issues warrants for people to be arrested and incarcerated for failing to timely pay related fines and fees. Under state law, a failure to appear in municipal court on a traffic charge involving a moving violation results in a license suspension.”

No community should be policed so aggressively. But if Ferguson is over-policed, the police themselves seem to be under-policed. And if police believe that aggressive policing of communities works, then on what basis could they object to a dose of their own medicine?

A good place to start would be requiring police officers to police one another on the job. Pew’s survey of police officers found that 84 percent say “officers should be required to intervene when they believe another officer is about to use unnecessary force,” while just 15 percent say they should not be required to intervene. Apparently, a lot of police officers would find it reasonable if their department imposed a duty to intervene. But many cities enforce no such duty. According to the Police Use of Force Project, they include Anchorage, Atlanta, Birmingham, Boston, Buffalo, Charlotte, Chesapeake, Columbus, Dallas, Denver, Detroit, Durham, El Paso, Fort Wayne, Garland, Glendale, Greensboro, Honolulu, Indianapolis, Irving, Jacksonville, Jersey City, Kansas City, Laredo, Lexington, Lincoln, Long Beach, Louisville, Lubbock, Memphis, Mesa, Nashville, North Las Vegas, Omaha, Pittsburgh, Plano, Reno, Rochester, San Diego, San Jose, Scottsdale, St. Louis, St. Petersburg, Tampa, Toledo, Tulsa, Wichita, and Winston-Salem.

A duty to intervene would of course include preventing a colleague from needlessly firing a weapon. But it could be interpreted expansively to include, as well, needless use of a baton or pepper spray, needless shoving, or even a lower-level transgression such as needless yelling or needlessly detaining a motorist for an excessive period of time during a routine traffic stop.

More broadly, cities could crack down on cops who refrain from giving fellow cops traffic tickets, get caught fudging a minor detail in a police report, or park their car illegally. Perhaps such a policy would ultimately reduce more egregious examples of special treatment or lawbreaking on the job.

A true broken-windows approach to tackling police misconduct would also go beyond enforcing even minor laws and policies. It would insist on police adherence to neighborhood norms too.

In the law-review article “The Good Cop: Knowing the Difference Between Lawful or Effective Policing and Rightful Policing—And Why It Matters,” the Yale Law School scholar Tracey L. Meares draws a distinction between whether cops are behaving lawfully and whether they are behaving in a way that accords with community views about how they ought to police to retain the public’s support. The law has little capacity “to tell police how to arrest or stop someone in a way that will tend to support police legitimacy,” she observes. “Rookie officers spend literally hours and hours reading law to learn when they are legally allowed to stop, arrest, and search. They are not correspondingly trained about how to conduct themselves so as to create and maintain their legitimacy in the community.”

Just as the broken-windows theory posits that graffiti can lead to drug dealing, one could argue that cops who transgress against a given community’s norms––say, by lawfully but disrespectfully berating someone on a street corner––are contributing to disorder and its consequences.

Broken windows is not my preferred approach to policing. In theory, police officers who enforce order on the streets could do so without resorting to unduly punitive fines and onerous probation requirements. In theory, broken windows need not manifest as racial inequity or mass incarceration. In practice, that’s exactly what has happened, whether due to flaws in the theory itself or flawed implementations of it. Those injustices cannot be ignored even if one grants that many of the police chiefs inspired by broken windows presided over falling violent-crime rates.

Still, those falling crime rates suggest that going after little problems to deter bigger ones may work. And applying that insight to law enforcement is less problematic than applying it to civilians. Constraints on the people whom society vests with a monopoly on violence are more necessary and justifiable than constraints on civilians indulging nonviolent behavior that some see as disorderly.

Insofar as broken windows already influences policing in a given city, constituting its official response to disorder, fairness demands that it be applied to police themselves. Police unions will resist, of course, not wanting their members to be policed as their members police the public. That’s hardly a reason for politicians to back down.

Kelling and Wilson wrote that “outside observers should not assume that they know how much of the anxiety now endemic in many big-city neighborhoods stems from a fear of ‘real’ crime and how much from a sense that the street is disorderly, a source of distasteful, worrisome encounters.” The protests roiling American cities right now can be understood as conveying a similar message: The anxiety about the police endemic in big cities, especially among black people, is obviously rooted in police shootings and other egregious abuses, but may also stem from lower-level misbehavior that creates the sense that cops don’t respect the neighborhoods or people they’re supposed to protect. Police forces compelled to police themselves aggressively might well find better community relations on the other side, and that suggests that their job will get easier, not harder.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.