m_topn picture
Atlantic Monthly Sidebar

February 1989

Making Neighborhoods Safe

Sometimes "fixing broken windows" does more to reduce crime than conventional "incident-oriented" policing

by James Q. Wilson and George L. Kelling

NEW Briarfield Apartments is an old, run-down collection of wooden buildings constructed in 1942 as temporary housing for shipyard workers in Newport News, Virginia. By the mid-1980s it was widely regarded as the worst housing project in the city. Many of its vacant units provided hiding places for drug users. It had the highest burglary rate in Newport News; nearly a quarter of its apartments were broken into at least once a year.

For decades the police had wearily answered calls for assistance and had investigated crimes in New Briarfield. Not much came of this police attentiveness--the buildings went on deteriorating, the burglaries went on occurring, the residents went on living in terror. Then, in 1984, Detective Tony Duke, assigned to a newly created police task force, decided to interview the residents of New Briarfield about their problems. Not surprisingly, he found that they were worried about the burglaries- but they were just as concerned about the physical deterioration of the project. Rather than investigating only the burglaries, Duke spent some of his time investigating the buildings. Soon he learned that many city agencies--the fire department, the public-works department, the housing department--regarded New Briarfield as a major headache. He also discovered that its owners were in default on a federal loan and that foreclosure was imminent.

The report he wrote to Darrel Stephens, then the police chief, led Stephens to recommend to the city manager that New Briarfield be demolished and its tenants relocated. The city manager agreed. Meanwhile, Barry Haddix, the patrol officer assigned to the area, began working with members of other city agencies to fix up the project, pending its eventual replacement. Trash was carted away, abandoned cars were removed, potholes were filled in, the streets were swept. According to a study recently done by John E. Eck and William Spelman, of the Police Executive Research Forum (PERF), the burglary rate dropped by 35 percent after Duke and Haddix began their work.

Stephens, now the executive director of PERF tells the story of the New Briarfield project as an example of "problem-oriented policing," a concept developed by Professor Herman Goldstein, of the University of Wisconsin Law School, and sometimes also called community-oriented policing. The conventional police strategy is "incident-oriented"--a citizen calls to report an incident, such as a burglary, and the police respond by recording information relevant to the crime and then trying to solve it. Obviously, when a crime occurs, the victim is entitled to a rapid, effective police response. But if responding to incidents is all that the police do, the community problems that cause or explain many of these incidents will never be addressed, and so the incidents will continue and their number will perhaps increase.

This will happen for two reasons. One is that a lot of serious crime is adventitious, not the result of inexorable social forces or personal failings. A rash of burglaries may occur because drug users have found a back alley or an abandoned building in which to hang out. In their spare time, and in order to get money to buy drugs, they steal from their neighbors. If the back alleys are cleaned up and the abandoned buildings torn down, the drug users will go away. They may even use fewer drugs, because they will have difficulty finding convenient dealers and soft burglary targets. By the same token, a neglected neighborhood may become the turf of a youth gang, whose members commit more crimes together in a group than they would if they were acting alone. If the gang is broken up, former members will still commit some crimes but probably not as many as before.

Most crime in most neighborhoods is local: the offenders live near their victims. Because of this, one should not assume that changing the environmental conditions conducive to crime in one area will displace the crime to other areas. For example, when the New York City police commissioner, Ben Ward, ordered Operation Pressure Point, a crackdown on drug dealing on the Lower East Side, dealing and the criminality associated with it were reduced in that neighborhood and apparently did not immediately reappear in other, contiguous neighborhoods. Suburban customers of the local drug dealers were frightened away by the sight of dozens of police officers on the streets where these customers had once shopped openly for drugs. They could not--at least not right away--find another neighborhood in which to buy drugs as easily as they once had on the Lower East Side. At the same time, the local population included some people who were willing to aid and abet the drug dealers. When the police presence made drug dealing unattractive, the dealers could not--again, at least not for the time being--find another neighborhood that provided an equivalent social infrastructure.

The second reason that incident-oriented police work fails to discourage neighborhood crime is that law-abiding citizens who are afraid to go out onto streets filled with graffiti, winos, and loitering youths yield control of these streets to people who are not frightened by these signs of urban decay. Those not frightened turn out to be the same people who created the problem in the first place. Law-abiding citizens, already fearful, see things occurring that make them even more fearful. A vicious cycle begins of fear-induced behavior increasing the sources of that fear.

A Los Angeles police sergeant put it this way: "When people in this district see that a gang has spray-painted its initials on all the stop signs, they decide that the gang, not the people or the police, controls the streets. When they discover that the Department of Transportation needs three months to replace the stop signs, they decide that the city isn't as powerful as the gang. These people want us to help them take back the streets." Painting gang symbols on a stop sign or a storefront is not, by itself, a serious crime. As an incident, it is trivial. But as the symptom of a problem, it is very serious.



IN an earlier article in The Atlantic (March, 1982) we called this the problem of "broken windows": If the first broken window in a building is not repaired, then people who like breaking windows will assume that no one cares about the building and more windows will be broken. Soon the building will have no windows. Likewise, when disorderly behavior--say, rude remarks by loitering youths--is left unchallenged, the signal given is that no one cares. The disorder escalates, possibly to serious crime.

The sort of police work practiced in Newport News is an effort to fix the broken windows. Similar projects are under way in cities all over America. This pattern constitutes the beginnings of the most significant redefinition of police work in the past half century. For example:

- When a gunfight occurred at Garden Village, a low-income housing project near Baltimore, the Baltimore County police responded by investigating both the shooting and the housing project. Chief Cornelius Behan directed the officers in his Community Oriented Police Enforcement (COPE) unit to find out what could be done to alleviate the fears of the project residents and the gang tensions that led to the shooting. COPE officers worked with members of other agencies to upgrade street lighting in the area, trim shrubbery, install door locks, repair the roads and alleys, and get money to build a playground. With police guidance, the tenants organized. At the same time, high-visibility patrols were started and gang members were questioned. When both a suspect in the shooting and a particularly troublesome parole violator were arrested, gang tensions eased. Crime rates dropped. In bringing about this change, the police dealt with eleven different public agencies.

- When local merchants in a New York City neighborhood complained to the police about homeless persons who created a mess on the streets and whose presence frightened away customers, the officer who responded did not roust the vagrants but instead suggested that the merchants hire them to clean the streets in front of their stores every morning. The merchants agreed, and now the streets are clean all day and the customers find the stores more attractive.

- When people in a Los Angeles neighborhood complained to the police about graffiti on walls and gang symbols on stop signs, officers assigned to the Community Mobilization Project in the Wilshire station did more than just try to catch the gang youths who were wielding the spray cans; they also organized citizens' groups and Boy Scouts to paint over the graffiti as fast as they were put up.

- When residents of a Houston neighborhood became fearful about crime in their area, the police not only redoubled their efforts to solve the burglaries and thefts but also assigned some officers to talk with the citizens in their homes. During a nine-month period the officers visited more than a third of all the dwelling units in the area, introduced themselves, asked about any neighborhood problems, and left their business cards. When Antony Pate and Mary Ann Wycoff, researchers at the Police Foundation, evaluated the project, they found that the people in this area, unlike others living in a similar area where no citizen-contact project occurred, felt that social disorder had decreased and that the neighborhood had become a better place to live. Moreover, and quite unexpectedly, the amount of property crime was noticeably reduced.

These are all examples of community-oriented policing, whose current popularity among police chiefs is as great as the ambiguity of the idea. In a sense, the police have always been community-oriented. Every police officer knows that most crimes don't get solved if victims and witnesses do not cooperate. One way to encourage that cooperation is to cultivate the good will of both victims and witnesses. Similarly, police-citizen tensions, over racial incidents or allegations of brutality or hostility, can often be allayed, and sometimes prevented, if police officers stay in close touch with community groups. Accordingly, most departments have at least one community-relations officer, who arranges meetings between officers and citizens' groups in church basements and other neutral locales.

But these commonplace features of police work are add-ons, and rarely alter the traditional work of most patrol officers and detectives: responding to radio calls about specific incidents. The focus on incidents works against a focus on problems. If Detective Tony Duke had focused only on incidents in New Briarfield, he would still be investigating burglaries in that housing project; meanwhile, the community-relations officer would be telling outraged residents that the police were doing all they could and urging people to call in any useful leads. If a tenant at one of those meetings had complained about stopped-up drains, rotting floorboards, and abandoned refrigerators, the community-relations officer would have patiently explained that these were not "police matters."

And of course, they are not. They are the responsibility of the landlord, the tenants themselves, and city agencies other than the police. But landlords are sometimes indifferent, tenants rarely have the resources to make needed repairs, and other city agencies do not have a twenty-four-hour emergency service. Like it or not, the police are about the only city agency that makes house calls around the clock. And like it or not, the public defines broadly what it thinks of as public order, and holds the police responsible for maintaining order.

Community-oriented policing means changing the daily work of the police to include investigating problems as well as incidents. It means defining as a problem whatever a significant body of public opinion regards as a threat to community order. It means working with the good guys, and not just against the bad guys.

The link between incidents and problems can sometimes be measured. The police know from experience what research by Glenn Pierce, in Boston, and Lawrence Sherman, in Minneapolis, has established: fewer than 10 percent of the addresses from which the police receive calls account for more than 60 percent of those calls. Many of the calls involve domestic disputes. If each call is treated as a separate incident with neither a history nor a future, then each dispute will be handled by police officers anxious to pacify the complainants and get back on patrol as quickly as possible. All too often, however, the disputants move beyond shouting insults or throwing crockery at each other. A knife or a gun may be produced, and somebody may die.

A very large proportion of all killings occur in these domestic settings. A study of domestic homicides in Kansas City showed that in eight out of ten cases the police had been called to the incident address at least once before; in half the cases they had been called five times or more. The police are familiar with this pattern, and they have learned how best to respond to it. An experiment in Minneapolis, conducted by the Police Foundation, showed that men who were arrested after assaulting their spouses were much less likely to commit new assaults than those who were merely pacified or asked to leave the house for a few hours. Research is now under way in other cities to test this finding. Arrest may prove always to be the best disposition, or we may learn that some kind of intervention by a social agency also helps. What is indisputable is that a domestic fight--like many other events to which the police respond--is less an "incident" than a problem likely to have serious, long-term consequences.

Another such problem, familiar to New Yorkers, is graffiti on subway cars. What to some aesthetes is folk art is to most people a sign that an important public place is no longer under public control. If graffiti painters can attack cars with impunity then muggers may feel they can attack the people in those cars with equal impunity. When we first wrote in these pages about the problem of broken windows, we dwelt on the graffiti problem as an example of a minor crime creating a major crisis.

The police seemed powerless to do much about it. They could arrest youths with cans of spray paint, but for every one arrested ten more went undetected, and of those arrested, few were punished. The New York Transit Authority, led by its chairman, Robert Kiley, and its president, David Gunn, decided that graffiti-free cars were a major management goal. New, easier-to-clean cars were bought. More important, key people in the Authority were held accountable for cleaning the cars and keeping them clean. Whereas in the early 1980s two out of every three cars were covered with graffiti, today fewer than one in six is. The Transit Police have played their part by arresting those who paint the cars, but they have been more successful at keeping cars from being defaced in the first place than they were at chasing people who were spraying already defaced ones.



WHILE the phrase "community-oriented policing" comes easily to the lips of police administrators, redefining the police mission is more difficult. To help the police become accustomed to fixing broken windows as well as arresting window-breakers requires doing things that are very hard for many administrators to do.

Authority over at least some patrol officers must be decentralized, so that they have a good deal of freedom to manage their time (including their paid overtime). This implies freeing them at least partly from the tyranny of the radio call. It means giving them a broad range of responsibilities: to find and understand the problems that create disorder and crime, and to deal with other public and private agencies that can help cope with these problems, It means assigning them to a neighborhood and leaving them there for an extended period of time. It means backing them up with department support and resources.

The reason these are not easy things for police chiefs to do is not simply that chiefs are slaves to tradition, though some impatient advocates of community-oriented policing like to say so. Consider for a moment how all these changes might sound to an experienced and intelligent police executive who must defend his department against media criticisms of officer misconduct, political pressure to cut budgets, and interest-group demands for more police protection everywhere. With decentralized authority no one will know precisely how patrol officers spend their time. Moreover, decentralized authority means that patrol officers will spend time on things like schmoozing with citizens, instead of on quantifiable tasks like issuing tickets, making arrests, and clearing cases.

Making the community-oriented officers generalists means letting them deal with other city agencies, a responsibility for which few officers are well trained and which cuts across sensitive questions of turf and public expectations.

If officers are left in a neighborhood, some of them may start taking money from the dope dealers and after-hours joints. To prevent that, officers are frequently moved around. Moreover, the best people are usually kept in the detective squad that handles the really big cases. Few police executives want their best people settling into a neighborhood, walking around the bus stops and shopping malls.

The enthusiasts for community-oriented policing have answers for all these concerns, but sometimes in their zeal they forget that they are contending with more than mere bureaucratic foot-dragging--that the problems are real and require thoughtful solutions. Many police executives get in trouble not because the crime rate goes up but because cops are accused of graft, brutality, laziness, incivility, or indifference.

In short, police management is driven more by the constraints on the job than by the goals of the job. You cannot cope with those constraints without understanding them. This may be why some of the biggest changes toward community-oriented policing have occurred in cities where a new chief has come in from the outside with a mandate to shake up a moribund department. Lee Brown brought a community orientation to the Houston Police Department under precisely those circumstances--the reputation of the department was so bad that almost any change would have been regarded as an improvement.

What can we say to the worried police chief who is already running a pretty good department? Start with corruption: For decades police executives and reformers have believed that in order to prevent corruption, you have to centralize control over personnel and discourage intimacy between police officers and citizens. Maybe. But the price one pays for this is very high. For example, many neighborhoods are being destroyed by drug dealers, who hang out on every street corner. The best way to sweep them off the streets is to have patrol officers arrest them for selling drugs and intimidate their customers by parking police cars right next to suspected drug outlets. But some police chiefs forbid their patrol officers to work drug cases, for fear they will be corrupted. When the citizens in these cities see police cars drive past scenes of open drug dealing, they assume the police have been paid off. Efforts to prevent corruption have produced the appearance of corruption.

Police Commissioner Ben Ward, in New York, decided that the price of this kind of anti-corruption strategy was too high. His Operation Pressure Point put scores of police officers on the streets to break up the drug dealing bazaar. Police corruption is no laughing matter, especially in New York, but some chiefs now believe that it will have to be fought in ways that do not require police officers to avoid contact with people.

Consider the problem of getting police resources and managing political pressures: resources can be justified with statistics, but statistics often become ends in themselves. One police captain we interviewed said that his department was preoccupied with "stacking widgets and counting beans." He asked his superior for permission to take officers out of radio cars and have them work on community problems. The superior agreed but warned that he would be watching to see what happened to "the stats." In the short run the stats--for example, calls answered, average response time--were likely to get worse, but if community problems were solved, they would get better as citizens had fewer incidents to report. The captain worried, however, that he would not be given enough time to achieve this and that the bean counters would cut off his program.

A better way to justify getting resources from the city is to stimulate popular demand for resources devoted to problem-solving. Properly handled, community-oriented policing does generate support for the department. When Newark police officers, under orders from Hubert Williams, then the police director, began stopping city buses and boarding them to enforce city ordinances against smoking, drinking, gambling, and playing loud music, the bus patrons often applauded. When Los Angeles police officers supervised the hauling away of abandoned cars, onlookers applauded. Later, when some of the officers had their time available for problem-solving work cut back, several hundred citizens attended a meeting to complain.

In Flint, Michigan, patrol officers were taken out of their cars and assigned to foot beats. Robert Trojanowicz, a professor at Michigan State University, analyzed the results and found big increases in citizen satisfaction and officer morale, and even a significant drop in crime (an earlier foot-patrol project in Newark had produced equivalent reductions in fear but no reductions in crime). Citizen support was not confined to statements made to pollsters, however. Voters in referenda twice approved tax increases to maintain the foot-patrol system, the second time by a two-to-one margin. New Briarfield tenants unquestionably found satisfaction in the role the police played in getting temporary improvements made on their housing project and getting a commitment for its ultimate replacement. Indeed, when a department experiments with a community-oriented project in one precinct, people in other precincts usually want one too.



POLITICIANS, like police chiefs, hear these views and respond. But they hear other views as well. One widespread political mandate is to keep the tax rate down. Many police departments are already stretched thin by sharp reductions in spending that occurred in the lean years of the 1970s. Putting one additional patrol car on the streets around the clock can cost a quarter of a million dollars or more a year.

Change may seem easier when resources are abundant. Ben Ward could start Operation Pressure Point because he had at his disposal a large number of new officers who could be thrown into a crackdown on street level drug dealing. Things look a bit different in Los Angeles, where no big increases in personnel are on the horizon. As a result, only eight officers are assigned to the problem-solving Community Mobilization Project in the Wilshire district--an economically and ethnically diverse area of nearly 300,000 residents.

But change does not necessarily require more resources, and the availability of new resources is no guarantee that change will be attempted. One temptation is to try to sell the public on the need for more policemen and decide later how to use them. Usually when that script is followed, either the public turns down the spending increase or the extra personnel are dumped into what one LAPD captain calls the "black hole" of existing commitments, leaving no trace and producing no effects.

What may have an effect is how the police are deployed and managed. An experiment jointly conducted by the Washington, D.C., Police Department and the Police Foundation showed that if a few experienced officers concentrate on known repeat offenders, the number of serious offenders taken off the streets grows substantially. The Flint and Newark experiences suggest that foot patrols in certain kinds of communities (but not all) can reduce fear. In Houston problem-oriented tactics seem clearly to have heightened a sense of citizen security.

The problem of interagency cooperation may, in the long run, be the most difficult of all. The police can bring problems to the attention of other city agencies, but the system is not always organized to respond. In his book Neighborhood Services, John Mudd calls it the "rat problem": "If a rat is found in an apartment, it is a housing inspection responsibility; if it runs into a restaurant, the health department has jurisdiction; if it goes outside and dies in an alley, public works takes over." A police officer who takes public complaints about rats seriously will go crazy trying to figure out what agency in the city has responsibility for rat control and then inducing it to kill the rats.

Matters are almost as bad if the public is complaining about abandoned houses or school-age children who are not in school. The housing department may prefer to concentrate on enforcing the housing code rather than go through the costly and time-consuming process of getting an abandoned house torn down. The school department may have expelled the truant children for making life miserable for the teachers and the other students; the last thing it wants is for the police to tell the school to take the kids back.

All city and county agencies have their own priorities and face their own pressures. Forcing them to cooperate by knocking heads together at the top rarely works; what department heads promise the mayor they will do may bear little relationship to what their rank-and-file employees actually do. From his experiences in New York City government Mudd discovered that if you want agencies to cooperate in solving neighborhood problems, you have to get the neighborhood-level supervisors from each agency together in a "district cabinet" that meets regularly and addresses common concerns. This is not an easy task (for one thing, police district lines often do not match the district boundaries of the school, housing, traffic, and public works departments), but where it has been tried it has made solving the "rat problem" a lot easier. For example, Mudd reports, such interagency issues as park safety and refuse-laden vacant lots got handled more effectively when the field supervisors met to talk about them than when memos went up the chain of command of one agency and then down the chain of command of another.



COMMUNITY organizations along the lines of Neighborhood Watch programs may help reduce crime, but we cannot be certain. In particular, we do not know what kinds of communities are most likely to benefit from such programs. A Police Foundation study in Minneapolis found that getting effective community organizations started in the most troubled neighborhoods was very difficult. The costs and benefits of having patrol officers and sergeants influence the delivery of services from other city agencies has never been fully assessed. No way of wresting control of a neighborhood from a street gang has yet been proved effective.

And even if these questions are answered, a police department may still have difficulty accommodating two very different working cultures: the patrol officers and detectives who handle major crimes (murders, rapes, and robberies) and the cops who work on community problems and the seemingly minor incidents they generate. In every department we visited, some of the incident-oriented officers spoke disparagingly of the problem-oriented officers as "social workers," and some of the latter responded by calling the former "ghetto blasters." If a community-service officer seems to get too close to the community, he or she may be accused of "going native." The tension between the two cultures is heightened by the fact that in many departments becoming a detective is regarded as a major promotion, and detectives are often selected from among those officers who have the best record in making major arrests--in other words, from the ranks of the incident-oriented. But this pattern need not be permanent. Promotion tracks can be changed so that a patrol officer, especially one working on community problems, is no longer regarded as somebody who "hasn't made detective." Moreover, some police executives now believe that splitting the patrol force into two units--one oriented to incidents, the other to problems--is unwise. They are searching for ways to give all patrol officers the time and resources for problem-solving activities.

Because of the gaps in our knowledge about both the results and the difficulties of community-oriented policing, no chief should be urged to accept, uncritically, the community-oriented model. But the traditional model of police professionalism--devoting resources to quick radio-car response to calls about specific crime incidents--makes little sense at a time when the principal threats to public order and safety come from collective, not individual, sources, and from problems, not incidents: from well-organized gangs and drug traffickers, from uncared-for legions of the homeless, from boisterous teenagers taking advantage of their newfound freedom and affluence in congested urban settings.

Even if community-oriented policing does not produce the dramatic gains that some of its more ardent advocates expect, it has indisputably produced one that the officers who have been involved in it immediately acknowledge: it has changed their perceptions of the community. Officer Robin Kirk, of the Houston Police Department, had to be talked into becoming part of a neighborhood fear-reduction project. Once in it, he was converted. In his words, "Traditionally, police officers after about three years get to thinking that everybody's a loser. That's the only people you're dealing with. In community policing you're dealing with the good citizens, helping them solve problems."


Copyright 1989 by James Q. Wilson and George L. Kelling. All rights reserved.
The Atlantic Monthly; February 1989; Making Neighborhoods Safe; Volume 263, Number 2; pages 46-52.

m_nv_cv picture m_nv_un picture m_nv_am picture m_nv_pr picture m_nv_as picture m_nv_se picture