Everyone knows about New York’s spectacular crime turnaround, with murder down 68 percent and overall felonies down 50 percent in the five years since 1993. But how we accomplished that turnaround—and what we’ve learned about how to police New York—isn’t widely known or fully understood. It should be, because the New York story adds up to a textbook on how to police any big city.

Our success rested on two major changes. First, we had to remake the NYPD into an effective, focused organization. Second, we had to use this instrument actually to police the city by developing strategies and tactics that would prevent and uproot crime rather than just react to it. In practice, of course, these were parallel, often overlapping efforts. The re-engineering was more challenging than the strategizing; and if we hadn’t changed the way the department did business, it wouldn’t have mattered what field strategies we chose, because we wouldn’t have been able to translate them into practice.

Professors of business administration and organizational management would have caught on to what we were doing right away, since it was no different from the restructuring and re-engineering that had transformed American business in the late 1980s and early 1990s. Like the corporate CEOs of that era, we began with a large, unfocused, inward-looking, bureaucratic organization, poor at internal communication or cooperation and chronically unresponsive to intelligence from the outer world. We reduced layers of management, drove responsibility down to the operating units, improved communication and data processing, tightened accountability, and rewarded results. In short order, we had the NYPD’s bureaus and divisions competing with criminals, not with one another.

Police work is by nature decentralized and discretionary. The cop in the field, the front-line supervisor, the precinct commander—these are the real decision makers in day-to-day police work. The only way you can control a police department from headquarters is if your aim is to prevent police from doing anything, rather than to have them function effectively—and for many years that was precisely the aim of the NYPD. The organization didn’t want high performance; it wanted to stay out of trouble, to avoid corruption scandals and conflicts in the community. For years, therefore, the key to career success in the NYPD, as in many bureaucratic leviathans, was to shun risk and avoid failure. Accordingly, cops became more cautious as they rose in rank, right up to the highest levels.

One anecdote that a deputy chief likes to tell perfectly captures how risk-averse the department had become. One weekend in the late eighties, when he was serving as duty captain in upper Manhattan, he arrived at a crime scene where cops had arrested two drug dealers, one of whom had tried to flee to an apartment. The deputy chief helped the cops secure search warrants for the apartment and a safe they found inside it—which contained drugs, cash, and weapons. The next day, the borough commander—“apoplectic with rage,” the deputy chief recalls—called him in to yell at him for seeking a warrant. “It’s people like you who cause problems in the department,” the commander roared. If something like that happens again, the commander ordered, “you will walk away.” Sums up the deputy chief: “I, as a captain, was not supposed to encourage these officers to do police work. My job was to stop them from doing police work.” After all, something could go wrong.

As for its management philosophy, the NYPD combined the worst of both worlds: it was a micromanaged organization that was strategically adrift. “Cops felt,” as one prosecutor remembers it, “as if the brass were checking up on them, not backing them up.” Management consisted of sending sergeants around to make sure that patrolmen were at their assigned posts. Woe to the cop who wasn’t there—even for a good reason—when the sergeant came at regular intervals to sign his memo book. The message: just sit there and get your ticket punched.

We had to change the department’s methods and mind-set. The first step, when author Bratton became commissioner in 1994, was to draw as many people as possible into the planning process, especially the precinct and unit commanders, who corresponded to a corporation’s middle managers, and the sergeants and lieutenants, who were, in effect, the front-line supervisors. “Bratton ran an organization that was open to talent,” says Jack Maple, a former NYPD deputy commissioner for crime-control strategies. “It was an organization of inclusion, where people weren’t afraid to come up with the wildest ideas.” Maple himself was a prime case in point. When only a Transit Police lieutenant, he caught Bratton’s attention with an idea-crammed memo on how to control robberies in the subway, and he rose rapidly thereafter.

We brought almost 500 people into the planning process, serving on 12 re-engineering teams that questioned everything in heated debate that sometimes escalated to a free-for-all. “An organization as big and as venerable as the NYPD accumulates a lot of bad habits,” points out John Timoney, who was chief of department in 1994 and first deputy commissioner in 1995 and is now Philadelphia’s police commissioner. “We did things a certain way because we had always done them that way. We had to banish the phrase,’We have always’ from our vocabularies. We had to start asking, ‘How should we do it?’ and ‘How can we do it better?’”

Ideas and innovations bubbled up. The commissioner didn’t originate most of the ideas that ended up transforming the way we did business; instead—and crucially—he created the atmosphere in which paradigm-breaking ideas could flourish. A December 1994 plan of action listed more than 600 recommendations, of which we adopted more than 400. We raised recruiting standards and improved training for the real world of police patrol. We redesigned the uniform with input from the officers who wear it, and, with the advice of the police unions, we revamped the archaic discipline system to make it swifter. We retooled our job-performance evaluation system. We even changed our super-secret internal-affairs process, bringing precinct commanders into internal investigations, which strengthened our capacity to prevent and detect police corruption and abuse. Formerly, only a half-dozen internal-affairs officials attempted to control misconduct in a 38,000-person department.

But the most important reform we made was decentralizing the department, devolving power to the precinct commanders and creating a career path for them to ascend. In the old NYPD, precinct commanders had little genuine authority. They couldn’t conduct an anti-prostitution operation or use plainclothes officers in anti-drug operations; they couldn’t secure search warrants and conduct searches. These constraints reflected a deep mistrust of the precincts, a fear that something could go terribly wrong out there that would embarrass the command staff. But as Robert Johnson, head of the private First Security Services and a member of Bratton’s kitchen cabinet, puts it: “A management team that concentrates on preventing failure usually forecloses success. When you don’t trust your basic resource, it’s hard to perform your basic business.”

The precincts are the primary unit of policing, and the precinct commanders are policing’s equivalent of corporate line managers. It was just plain crazy to limit their options. Worse, the precinct commander’s job was all downside risk in 1993, just before we arrived; there was no real way to succeed and a dozen ways to fail. An outstanding precinct commander’s performance against crime would have gone largely unnoticed—nobody was monitoring that kind of success—but a corruption scandal or a mishandled community incident could set a career back years. No wonder that captains tried to hurry through their precinct commands and move on to less career-threatening assignments in the bureaucracy at 1 Police Plaza.

We gave precinct commanders—typically people with 15 years’ experience, a college education, and a sophisticated knowledge of the city and the department—the authority to run what amounted to miniature police departments. John Timoney devised a new career ladder for these commanders, ascending from one of the 35 “C” precinct houses with moderate workloads to one of 31 higher-pressure “B” houses and ultimately to one of the 10 extremely busy “A” houses. As a commander rose through these steps, he could advance from captain to deputy inspector to inspector without ever leaving the precincts. A few battlefield promotions made it clear that precinct command was the place to shine in the NYPD.

The good commanders reveled in their new authority, and their precincts became natural arenas for team building. Their cops were energized. Cops like to do police work: stinging a drug dealer with a buy-and-bust operation, executing a search warrant at the apartment of a gun dealer, even catching someone blasting a car radio with a decibel meter—these are interesting jobs compared with regular patrol. They are also team activities that give the workers a shared sense of purpose and a renewed sense of energy.

Accountability goes hand in hand with decentralization; you can’t give all that power away without a means of maintaining strategic oversight. The NYPD does that through its now-famous Compstat process, which uses computerized crime statistics, electronic crime maps, and intensive crime-control management meetings to guide and monitor the department’s anti-crime strategies.

From the start, Jack Maple had insisted on timely weekly data about crime in the precincts, an approach completely foreign to the NYPD, where crime data typically lagged events by up to three months. Because we couldn’t live with flying blind, not knowing what was actually happening precinct by precinct and week by week, Maple, Chief Louis Anemone, and Anemone’s staff developed the basic format for the weekly report that the department has been using ever since. It showed weekly felony crime and arrest data for every precinct, comparing it with the totals for the previous week and the month- and year-to-date totals. Compstat was the computer-file name of this report, a contraction of “comparison statistics.” Maple and Anemone began going over the new data at meetings with precinct personnel, quizzing commanders about crime in their precincts and what they were doing about it. Soon they were using pin maps and acetate overlays to display the patterns of criminal activity. They began calling in special unit commanders and narcotics commanders. The Compstat process was off and running.

As Compstat grew more sophisticated, we began computer mapping the crime patterns and displaying the maps on large overhead screens. We could identify local crime increases almost immediately and respond to them rapidly with effective measures before they could add up to a big, citywide crime spike. You could see the clusters of shootings, robberies, burglaries, and car thefts. We mapped arrest and patrol activity and compared crime incidents with police response. If the two didn’t match up, you knew you were doing something wrong. We compared our deployment patterns with time-of-day graphs that showed when crime spikes were occurring. Compstat’s maps helped make sure that we were putting our resources where the problems were, and when they were happening. We could quickly assess whether new strategies and tactics worked or failed.

Compstat enforced cooperation among the department’s many bureaus and units. Every week, a different group of commanders from a particular part of the city stands up at the podium to get grilled about crime and prevention in their precincts, but every other relevant special unit and task force commander for that area of the city is also present and must be prepared to respond. Just having them all in one room instantly cut the bureaucratic Gordian knot. Chief Anemone and Deputy Commissioner Maple could broker solutions to current problems on the spot. Was drug dealing causing an increase in shootings? Narcotics enforcement could be called in swiftly. Were burglaries on the rise in a community? The warrant division could search for wanted burglars in the area. Were auto thieves following a certain route to a nearby highway? Highway units could set up a checkpoint.

Above all, as John Timoney puts it, “Compstat is the greatest accountability tool ever.” It’s an instrument for holding precinct commanders responsible for crime in their areas, rewarding them if they push crime down and removing them if they don’t come up with plans to do so. In addition, it gets the whole department, top brass included, involved in thinking about how to push back crime, and it lets precinct commanders know, on a weekly basis, that their bosses support their efforts. Eventually, we used Compstat to manage everything from civilian complaints to overtime to police auto accidents.

It’s not too strong a statement to say that we reinvented police strategy in 1994. Before then, the prevailing criminological wisdom held that the police couldn’t do much about crime and that police strategies and tactics didn’t really matter. In that spirit, the NYPD and most other police departments spent almost no time thinking about anti-crime strategies. Police brass lurched from emergency to emergency, with no one looking at the overall picture.

Strategy is a way of seeing things whole, of focusing on the entire system of crime and how it operates. Since 1994, the NYPD hasn’t just been solving crimes; it has been dismantling criminal enterprises and support systems. It has been taking away the things that criminals need to function: their guns, their fences, their chop shops and auto exporters, their drug-buying and prostitution customers, their buildings and apartments, their cars, and the unpoliced sectors of the city where crime used to thrive.

In early 1994, the department developed strategies on guns, drugs, youth crime, domestic violence, and quality of life. Though much has been made of quality-of-life enforcement as key to the overall New York strategy, it was hardly the whole story. To say that “zero tolerance” policing turned New York around, as if driving away squeegee men and panhandlers could by itself cut the robbery and burglary rates, is a gross oversimplification. To succeed, we had to employ the quality-of-life strategy in concert with a range of strategies targeting felony crime.

But quality-of-life enforcement is important for three reasons. First, most neighborhoods are usually more concerned about prostitution, low-level drug dealing, excessive noise, underage drinking, and other minor offenses than major crimes. Citizens want the police to do something about these highly visible disturbances. Second, as George Kelling has persuasively argued in the pages of City Journal, disorderly environments breed both crime and fear. Third, criminals who commit serious crimes frequently commit minor violations as well; quality-of-life enforcement lets cops intervene with this population and sometimes prevent serious crimes before they happen.

In 1994, we took quality-of-life enforcement to the streets. Author Bratton had pioneered this style of policing in Boston in the 1970s. He imported it to New York when he became chief of the Transit Police in 1990, and it had transformed the subway from a place where young thugs thought they could get away with anything into a place where they felt they could get away with nothing, with a steep drop in crime as a result. So, too, with the city as a whole: as Kelling has observed, New York City had “depoliced” its streets in the quarter-century prior to 1994. Police officers were walking by disorderly conditions and letting them fester. They were openly giving freedom of the streets to the drug dealers, the gangs, the prostitutes, the drinkers, and the radio blasters. A sense of fear and anarchy pervaded many neighborhoods. The traditional order-keeping forces, the responsible adults in these communities, played less of a role as their own fear and uncertainty grew. They—along with the wrongdoers—had gotten the message that even the cops didn’t care, and they were understandably hesitant to put themselves on the line.

The newly empowered precinct commanders targeted prostitution, public drinking and underage drinking, street drug dealing, and excessive noise. They instructed patrol officers to intervene in street disorder and to try to restore a sense of civility and safety to neighborhoods. “We had always been telling the cops what not to do in street situations,” says Michael Julian, formerly chief of personnel and now a private security executive. “They needed training in what they could do.”

As part of the quality-of-life initiative, the police checked identification of people stopped for minor offenses. When the checks turned up a wanted person, a parole violator, or a repeat offender, cops arrested and searched him, instead of letting him off with a relatively toothless desk-appearance ticket, as is usual with minor offenders.

This intensified police presence had an almost immediate impact on illegal guns. We flooded shooting hot spots, identified through the Compstat maps, with both uniformed and plainclothes enforcement teams. People carrying illegal guns quickly realized that they risked gun charges after being arrested for minor crimes. After rising briefly in 1994, gun apprehensions then began to fall; the gunslingers were leaving their weapons at home. With far fewer guns on the street, far fewer people were shot and far fewer were killed. In month-to-month comparisons from 1993 to 1994, homicide was down 32 percent in September 1994, 46 percent in October, 28 percent in November, and 34 percent in December.

At the same time, we brought the department’s detectives more forcefully into play. We told them to make shootings a priority and to investigate every one as if it were a murder. “There was no good reason why the detective response to a killing was always so much greater than to a shooting,” says John Timoney. “An aggravated-assault victim is just a lucky homicide victim whose assailant missed.” To keep on top of the issue, Jack Maple had the operations center beep him every time a shooting occurred. In 1994, he got beeped nearly 4,500 times, day and night.

We enlisted detectives in an anti-gun offensive, instructing them to grill anyone arrested in a shooting or on gun charges about how and where he had acquired his weapon. The murderer who fired on Jewish students in a van on the approach ramp to the Brooklyn Bridge, killing one boy, was in custody within a few days; within a week, so was the gun dealer from whom he had purchased his weapon. By the end of 1994, the NYPD had arrested more than 200 gun dealers and confiscated their supply of weapons. Between 1993 and the end of 1995, handgun homicides declined by 40 percent, and shootings declined by more than 2,000 cases.

We fed information about gun trafficking from the precinct interrogations to the NYPD’s joint task force with the federal Bureau of Alcohol, Tobacco, and Firearms, which in turn used this intelligence to make arrests for illegal gun trafficking in the main states that supply weapons to New York City. The task force also began closely scrutinizing applicants for federal firearms-dealing licenses, in the belief that some dealers would use the right to buy guns in quantity to sell them illegally to unlicensed New Yorkers. In the first years of the program, 92 percent of applicants for new licenses and for license renewals either failed to win approval or—once they knew the police were watching them—withdrew their applications.

Precinct detective squads adopted a new intelligence-gathering strategy. They began to question all arrestees not only about crimes that they might have committed themselves but also about any other crimes that they knew anything about. Detectives asked arrestees about open homicides, about robbery and burglary patterns, and about where drugs or firearms were for sale. Some talked in return for a promise of more lenient treatment, but many were willing to talk gratis, swelling the NYPD’s supply of criminal intelligence and case leads. A car thief turned in a fence, who turned in a father-and-son gun-dealing team. A gun-crime debriefing in the 46th Precinct in the Bronx led to several arrests in a year-old murder case and to arrests for a recent stabbing and a carjacking in the neighboring 43rd Precinct.

Detectives began using the department’s improved intelligence to identify robbery patterns more quickly and to apprehend the robbers on the second or third crime of the pattern rather than the ninth or tenth. Identifying a pattern early, staking out the locale, and catching the guy can prevent 20, 30, or even 50 robberies a year. We also told detectives to pursue all accomplices in any robbery, instead of playing the old detective game of building clearance rates by closing cases with just one arrest. Our goal was preventing crime, not closing cases. Robbery fell by more than 26,000 incidents, or 31 percent, in 1994 and 1995. By 1998, it was down 55 percent from its 1993 level.

The Detective Bureau also placed a new emphasis on warrants and on finding wanted fugitives, since warrant absconders who fail to show up for their court appearances are frequently recidivists who will continue to commit crimes for as long as they are free. We expanded the warrant division and doubled warrant apprehensions between 1993 and 1995. Detective Bureau arrests of all kinds, including warrant apprehensions, are up about 138 percent since 1993.

To control burglaries and thefts of car radios, we went after fences who traffic in stolen goods, another thing that the NYPD had never bothered to do. Early on, police used sting operations to nail two major fences in Brooklyn, recovering enough stolen merchandise to fill a warehouse. Eventually, the fences helped identify a who’s who of Brooklyn burglars; one fence even had security videotape showing many of the burglars’ faces. Even when it doesn’t lead directly to mass burglary arrests, as this operation did, shutting down local fences can have a dramatic impact on neighborhood burglary rates. It may take burglars a while to find another outlet for their stolen goods, and they can’t go back to work until they can move the hot merchandise. Burglary was down 25,000 incidents in 1994 and 1995, or 25 percent. By 1998, it was down 53 percent.

When we arrived, New York City, and especially the borough of Queens, looked like the car-theft capital of the world, with more than 112,000 cars, worth about $400 million, stolen in 1993. Organized-crime groups stole an estimated 70 percent of them, for parts and for export. Using the same logic that we had applied to burglary, we went after chop shops and auto exporters. In 1995, we set up more than 30 phony fencing operations to catch car thieves, and we conducted more than 60 operations in which undercover officers offered supposedly stolen parts to dealers. We also used stings to make cases against exporters of stolen cars to Russia, the Dominican Republic, Saudi Arabia, Nigeria, and Colombia. We requested district attorneys in Queens and Brooklyn to prosecute auto thieves routinely and send them to jail. Car theft plummeted by more than 40,000 cases in 1994 and 1995, or 36 percent. By 1998, it was down 61 percent.

We knew, as Michael Julian puts it, that “drugs were causing probably half of the crime and a lot of the fights and gunplay.” There was no way to reduce crime without going after narcotics. But the department had long been combating drugs in the wrong way, arresting street dealers and seizing drugs—commodities that were expendable, and almost infinitely renewable, as far the drug gangs were concerned. We believed that we had to concentrate on the local drug gangs and dismantle them. These are the critical middlemen between the street dealers and the international drug importers.

Narcotics Division commander Pat Harnett developed a new strategy in 1995 and 1996: he established turf-based drug units that would concentrate on reclaiming individual neighborhoods, investigating and dismantling the drug gangs operating in an area,however long it took. They made drug buys, infiltrated organizations, cultivated informants, planted wiretaps, investigated bank accounts, and built solid cases against all the participants in a gang or group of gangs. When a case was ready, they would roll up 20 or 30 members of a gang in a single day. The department supplemented these investigations with civil actions and eviction proceedings against drug dealers’ buildings and apartments. It also sent out uniformed patrolmen in force, to hold the gains and discourage drug buyers from reentering the neighborhood. Everywhere the drug initiatives have gone, they have driven down murders, shootings, and street crime.

All these strategies—and a host of other ones targeting youth crime or family violence, for example, or using civil lawsuits to hamstring criminal enterprises—send one clear message to the criminal population of New York: “It’s not your street,” as one cop put it. “This street belongs to the people of New York. We are going to take it back.”

Many criminologists, as you’d expect, are sticking to their guns and insisting that nothing we have done had the slightest effect on New York’s crime drop. Crime fell not just in New York but nationwide, they say, proving that some impersonal force much bigger than the NYPD has come to bear on all of urban America. In truth, however, in 1994 and 1995, the drop in New York crime accounted for more than half of the crime decline in the entire country: we weren’t part of the national trend; we drove the trend.

But what reasons do the experts proffer to explain the drop in crime? The crime-prone youth population shrank, they argue, and employment expanded. In New York, however, the teenage population was stable in the early 1990s, not declining, and the minority teenage population was actually on the rise. Between 1990 and 1994, New York City public high school enrollments were up 12 percent. Throughout the period of steep crime decline, moreover, New York had an unemployment rate of between 8 and 10 percent, just about double the national rate. According to the demographic and economic theories, crime should have been raging out of control instead of falling off the charts.

“The experts will never forgive Bratton and Giuliani for proving them wrong,” says Tom Repetto of the Citizens’ Crime Commission. “They want to believe that crime can only be reduced by sweeping social change. But they do have a fallback position: if the police did reduce crime, they did it by illegitimate means.”

The claim of “illegitimate means” is growing more insistent, accompanied by accusations that the NYPD routinely uses unconstitutional searches and “Gestapo tactics.” Policing in New York, critics charge, has been “militarized,” as if SWAT teams were roving around the city on search-and-destroy missions. The tragic shooting of Amadou Diallo in February and the outrageous abuse of Abner Louima in 1997 have fanned this talk significantly. Many who claim to speak for minority communities say that they now fear the police more than they fear the criminals.

The fear may be understandable, but the problem has been overdrawn. The NYPD is not a racist or brutal department. It won its gains against crime not by abusing citizens but by strategically managing and focusing its resources. Its problem is a problem of attitude: cops need to be more respectful in their encounters with the public. With crime under control, we have the breathing space to provide much better training in the human side of policing and to build trusting relationships in every neighborhood in the city. The NYPD rose to the challenge of crime in the mid-1990s; it certainly can rise to this challenge now.

Though the department can do much more to win the confidence of minority communities, it has already done the most important thing: today, those communities receive the level of police protection and service they deserve. They aren’t written off as unimportant and unpoliceable. The Compstat revolution had the effect of allocating resources to the neighborhoods that needed the help most. “All the dots representing crime on Compstat maps are the same size, whether the victim is Leonardo DiCaprio or Fred Mertz,” says Jack Maple. “The resources go where the problems are.” As a result, minority neighborhoods like East New York and Washington Heights have seen huge crime drops. Kids can play on the street again; they don’t go to bed to the sound of gunfire anymore.

What we learned above all from the New York experience is that police can control and manage virtually every type of crime in virtually every neighborhood. No place is unpoliceable; no crime is immune to better enforcement efforts. Though underachieving in the past, American police departments can take the lead in restoring safety and order to communities all over the country. After a generation of police executives who were convinced that cops couldn’t cut crime, a new group of leaders is following the New York example and sending the message that police can make a difference. These leaders are junking the old reactive model, in which police responded to crimes and filed reports, in favor of a new strategic policing that gets criminals on the run and keeps them running. American police departments are beginning to live up to the boast posted by an anonymous officer in the NYPD Command Center in 1995: “We’re not report takers,” it read. “We’re the police.”