Lies, Damned Lies, and Statistics

It’s the beginning of the year, and that means it’s time for police departments, police unions, and police critics to argue over the meaning of last year’s crime statistics. Here in New York, the debate has fallen along familiar lines—the police and the administration touting greater safety (and claiming credit for it), while scaremongers howl that the numbers are overstated, and critics complain that the arrest patterns show racial and social disparities. There is a great overview of the typical retrenchment, played out with this year’s numbers, by Al Baker and David Goodman in the Times. My thoughts, as usual, are that everyone is a little bit right and a whole lot wrong.

Statistics played a huge role in the Bloomberg-Kelly era – every year Commissioner Kelly touted a drop in crime, and critics pointed out the geometric increase in stops and frisks. From 2002 through 2006 the number of stops increased from 97,296 to 506,491. That wasn’t public at the time, of course, because the NYPD, in violation of a legal settlement, was refusing to disclose them. At the CCRB, we knew that stops were way up, because complaints of stops were way up. In the impeccable bureaucratese that I favored at the time, I wrote in our 2005 report that “complaints involving abuse of authority allegations such as ‘question and/or stop’ have risen at rates higher than complaints in which force, discourtesy, or offensive language allegations are lodged.”

One statistic we always tracked internally, once we got the stop-and-frisk numbers, was how many stop-and-frisk complaints there were in comparison to the number of stops. It was a remarkably stable number—from 2006 through 2012, the CCRB would receive, on average, one complaint for every 300 stops conducted by officers.

Criticizing stop-and-frisk (which by that time had been found unconstitutional and was already in decline) was a central prong of de Blasio’s campaign, and led to the early narrative once he took office. Without the enormously aggressive stop-and-frisk program, so the critics said, crime would immediately bounce back to 1970s levels. During de Blasio’s first week, there were more than a few articles suggesting that a crime wave was upon us. After all, there were eight murders in the first five days of 2014, two more than in the first five days of 2013—the city was “on pace” to see over five hundred murders once again.  (in the end, 2014 turned out to be a record low year for murders).

This year brought the welcome spectacle of two police commissioners behaving like six-year olds in the sandbox, calling each other names and questioning each others’ success. The personal animus between Kelly and Bratton appears untethered to any policy choices. The first time they went at it, when Bratton replaced Kelly in 1994, Kelly was the advocate for community policing and Bratton wanted a tougher hand. Now they have traded roles, and you would think from their public statements that each has simply reached into the other’s press kit from twenty years ago.

For the record, the crime statistics this year are pretty straightforward, as they have been for about ten years: nothing has changed. Murders apparently went up from 333 to 350, but in a city of 8.4 million people, that means your chances of getting murdered last year increased from 0.0039643% to 0.00416667%. Higher than your chances of winning the Powerball, but still. Minor crimes are reported as going down. And here’s the thing. Of course the numbers are fake. Or at least a little fake. If you walk into a precinct and say that someone punched you in the face and took your wallet, the desk officer can say that’s a felony robbery. Or she can say that it’s a misdemeanor assault and a misdemeanor larceny. One of those helps her get a promotion, and one of them doesn’t, and you will almost certainly never be the wiser. Crime statistics are so pervasively and uniformly underreported that when the LA Times analyzed five years of data—while Bratton was commissioner—and discovered that 14,000 crimes were misclassified, the result in the New York papers was a collective yawn. 

But you can only fake the statistics so much. The old saw that you can’t fake murder statistics because you can’t hide the bodies isn’t true, because whether or not a death is a “murder” as opposed to some act of negligence can’t always be determined when the person dies. (Three guesses as to whether the 333 murders in 2014 included Eric Garner, for example). Rape statistics are totally unreliable because rape is so dramatically underreported. But there are only so many tricks. Whatever you were doing five years ago, you are probably doing it now. So the statistics are probably a pretty good measure of relative crime from one year to the next, even if the top-line numbers are fudged. And over the past decade, crime decreases (or occasional increases) have been in the single-digit percentages. Crime came down twenty years ago, and stayed down. That’s really about it. That’s why when the NYPD reports its yearly statistics in historical terms, the most recent year it uses for comparison is 2001. 

There is only one real story from the crime statistics last year. Ray Kelly’s stop-and-frisk program has been wholly, totally, and completely discredited as a crime fighting tool. In 2012, with the program already starting to decline, the NYPD stopped 532,911 people. By 2014 that had dropped to 45,787, and while the final 2015 numbers are not yet released, as of October they were on track to be about 25,000. That’s a 95% decrease in three years. If stop-and-frisk had had any impact on crime whatsoever, it would somehow have been reflected in the numbers (and even if you think there is some lag, we are now two years past the close of the stop-and-frisk era). Statistics really only tell stories when you see big moves. And we have seen a big move downward in stops, without any big move in crime. The case is closed: stop-and-frisk did not lower crime.

That’s the police side. But as most of you know, I am really as interested (or more) in my old shop, the CCRB. The 2015 CCRB report isn’t out yet, and I will write about it when it is. But the most appalling misstatements about criminal justice statistics this past year did not come from either police commissioner. They came from Richard Emery, the new CCRB board chair. In 2015, Emery, in an effort to show how much complaints have dropped, allowed the following chart to be printed in the CCRB’s mid-year report:

Wow what an enormous drop in complaints! - Except that all of those numbers are year-end figures, except for the last one, which covers only a six-month period. This is the sort of chart that an eighth-grade math teacher would use as an example of dishonest statistics. Emery compounded the mendacity by saying that the drop in complaints represents a “shift in the NYPD culture towards civilians." This is utter nonsense. As I said above, in most years, the CCRB would get one complaint for every 300 stops. In 2014, with stops way down and complaints kind-of-down, it received one stop and frisk complaint for every 43 stops. In relation to actual police contact with civilians, there was actually an increase in complaints to the CCRB—not surprising, given the massive attention to police conduct that has been in the media in the past few years.

From every perspective, the lesson of crime statistics this past year has been the lesson of stop and frisk. Stop and frisk has been nearly abolished while crime is steady – almost a controlled experiment showing that the program didn’t lower crime. And complaints are down almost entirely because stop and frisk is down—there is no evidence that the stops actually conducted are any different than they used to be. If you were to look at overtime abuse, PD promotions to specialized squads, and likely dozens of factors I haven’t thought of, stop and frisk would likely play a role. All of the rest of it is just noise.

Small-bore Solutions to Big-time Problems

As someone who spent close to a decade investigating police misconduct, researching and drafting policy recommendations, and working with stakeholders on issues of policing, I support the goals of the current movement, but am sometimes frustrated by the proposed resolutions. The current wave of activists wants to draw attention to the problem of overpolicing in this country—either as embodied by police shootings, primarily of young black men, or as the militarization of police departments, or as zero-tolerance policing masquerading as “Broken Windows” policing. So far so good. There is a crisis of overpolicing in this country, and for too long, nobody paid attention to it. I should know—a decade ago, much of my job involved trying to get newspapers to run stories about police misconduct more prominently.

But beyond publicizing the problem, what demonstrators are asking for most is increased criminal prosecution of police. But I don’t think that throwing cops in jail will deter police misconduct any more than throwing street-level offenders in jail deters muggings. And it kind of saddens me when people who criticize criminal over-enforcement in one context cry for punishment in the other. The criminal law is erratic in its application and filled with procedural complexities that can scuttle a case. Police defendants routinely get great lawyers, usually waive their right to a jury trial, and present technical defenses that are effective with judges. That’s how our criminal justice system, which is predicated upon the principle that it is better to let ten guilty people go free than to convict one innocent person, is designed. I understand that in most contexts it doesn’t fulfill that design. But I don’t feel comfortable demanding aggressive prosecution and stiff criminal penalties as soon as the accused is a police officer.

Sure, when Walter Scott or Laquan McDonald is shot while running or walking away from the police, murder charges should be brought. In both of these cases, murder charges have been brought. Just as charges have long been brought against police officers when the evidence is strong enough. It is true that it is hard to meet the beyond-a-reasonable-doubt standard to convict a police officer. But it’s hard to meet that standard for anyone; and it certainly isn't impossible. After all, the Manhattan District Attorney won a homicide conviction against Bryan Conroy for shooting Ousamane Zongo in 2005, and the Eastern District of New York convicted three cops in the Louima case. Cops have long been routinely indicted and convicted on perjury charges based on video and audio recordings.

What’s more interesting to me is what we can do to limit police misconduct in the first place, rather than how we punish it once it happens. Demilitarizing police departments and ending policies like stop-and-frisk can improve police-community relations, but most civilian encounters that go bad come from police who are responding to a call. And what I saw, over and over again, is that police officers are more likely to make mistakes when they respond to a call with insufficient or inaccurate information. And those mistakes can be lethal.

Most often, officers are given inadequate information because they speed to a scene after hearing a radio code—and little else—from their dispatcher. The woman who made the call that led to Henry Louis Gates getting detained in his own living room said she was calling because her neighbor urged her to, and that it looked to her like the guy had lost his keys and was trying to get into his own house. When Khiel Coppin’s mother called 911, Coppin could be heard in the background saying he had a gun, but his mother told the operator that she thought he didn't. And when a bystander called 911 to say that Tamir Rice was pointing a gun at people in an Ohio park, she said that the gun was “probably fake.” But none of the officers were given all the information that they could have been. They were only given a numbered code (say, 10-52-K, which means in New York that there is a dispute and there is a knife) and if they were lucky, a brief description.

It should be technologically feasible for any department to play the entirety of a 911 call to a responding officer on their way to the scene. I am not aware of any department that does so. While most departments get lots of federal funding for sophisticated equipment, the day-to-day technology of law enforcement is archaic. At the CCRB, we put in funding requests for digital recorders and were denied every year--as late as 2008 we were using manual recorders and cassette tapes for interviews. In many police precincts, they still have carbon copy forms to fill out, which means they still use a typewriter.

 New York allows officers to play back descriptions that were given to them by dispatch to quell civilians who believe they were improperly stopped.  This is a small and extremely powerful tool—when we studied thousands of stop and frisk complaints in 2000, we found that only one had been filed after an officer had played a description back over the radio. But this power still falls short of the easy ask I’m making—let dispatch play for the officers, in full, the call that they are responding to. Radio codes make a lot of sense when there are dozens of officers listening over a scanner. Each car cannot be troubled to hear the long-winded report of a civilian. But once a unit confirms that it is responding to a call, the officers deserve all the information the department can give them. They can hear the whole call, and learn from the caller’s tone, inflection, and language whether the caller fears for her life or is merely annoyed.

We can’t know for certain whether the officers responding to the Gates call, the Coppin call, or the Rice call would have behaved differently if they had been able to hear the full details and the tone of voice of these callers. But they might have. And I could recite dozens of other cases—none of them in the news—where officers might have taken a different stance if they had known everything that had been said to dispatch. And it certainly couldn’t do any harm. It may seem counterintuitive or unbearably small-bore to say that we should give more information to line officers responding to a call. But small-bore solutions are often the foundation of addressing big problems. Playing full calls to responding officers can give them the information they need to assess a situation properly and accurately. And most importantly, it could save lives.

Pat Lynch, Quentin Tarantino, and the Priorities of Public Unions

My article on Pat Lynch and Quentin Tarantino appeared in the November 11, 2105 edition of Newsweek.  The text is reprinted below for non-Newsweek subscribers. 

Patrick Lynch has a problem.

When first elected head of the Patrolman’s Benevolent Association union in 1999, Lynch was, at 35, the youngest PBA president ever, and he was seen as a force of reform in an ossified union that had gone years without pay raises.

But he immediately took on the role he has played ever since: chief defender of police officers who kill civilians. It served him well for over a decade, but recent events suggest that the shtick is getting old.

In 2004, when Police Commissioner Ray Kelly said the shooting of Timothy Stansbury in a Brooklyn housing project stairwell “appeared unjustified,” Lynch and the PBA issued a vote of no-confidence in Kelly, demanding that the “anti-cop” commissioner resign.  

After the detectives in the shooting that killed Sean Bell were acquitted, Lynch took the lead in defending the verdict—never mind the fact that his union doesn’t represent detectives. Ousmane Zongo, Gidone Busch, Khiel Coppin—whenever there was a cop standing over a corpse, there was Lynch in his pompadour, tie and windbreaker telling us all to move along, nothing to see here.

And for years everyone believed him. I should know. From 2001 through 2008, I served at the Civilian Complaint Review Board, the city agency charged with investigating complaints of officer misconduct. We investigated thousands of complaints a year. We issued recommendations regarding the use of pepper spray, policing demonstrations and executing search warrants.

And every time we tried to draw attention to the issue of police misconduct, there was Lynch standing in our way, reminding everyone that crime was down. And every time he seemed to win.

I went to forums all around the city after Bell was killed and spoke at the community council meetings after Stansbury was shot. There were always a vocal few who complained about racial profiling, over-policing and violence, but most people shrugged and moved on.

When we gave our budget testimony before the New York City Council every year, we were conveniently scheduled just after the police commissioner’s morning testimony and just before the afternoon statements of the district attorneys, so the council members could go get lunch while we spoke to an empty chamber.

But in the last two years, all that has changed. After the shooting of Michael Brown and the death of Eric Garner, just about any violent arrest can land a cop in the paper. And there have been dozens of rallies, marches, protests and seminars calling for police reform. Lynch can’t keep the lid on everyone.

But that hasn’t stopped him from trying. At a rally two weeks ago—one that wasn’t very large and would likely have been quickly forgotten—filmmakerQuentin Tarantino said that if “there’s murder going on, then you need to rise up and stand up against it. I’m here to say I’m on the side of the murdered.”  

Unable to resist, Lynch frothed up his membership and called for a police boycott of Tarantino’s latest film. Wrapped up in his latest media battle, Lynch has been alarmingly silent on the news that came out a week later and is much more important to his membership: On October 25, a neutral arbitrator charged with approving a new contract for the PBA presented Lynch with a draft that limits raises to 1 percent a year. The PBA has been without a contract since 2010.  

While other unions held talks with the city, winning concessions for their members, Lynch has held fast, defending cops accused of wrongdoing rather than entering the negotiating room.

Right now, Lynch resembles no one so much as Randi Weingarten, the teachers union chief who spent much of her time rallying against charter schools and teacher accountability rather than negotiating for member benefits.

Her successor, Michael Mulgrew, agreed to accountability measures that allow the city to remove problem teachers and reaped tremendous reward, including 18 percent raises, new leadership posts and fellowship tracks for outstanding teachers.

It’s true to a point that Tarantino stuck his foot in his mouth. The legally precise definition of “murder” includes a degree of intent to kill that doesn't apply to most of the police-involved shootings of the past year. Officer Michael Slager was charged with murder, however, in the shooting of Walter Scott.  

But a drunk driver who runs over a 3-year-old on a tricycle hasn’t committed murder in the legal sense either, and we still want to see him punished. If Lynch is rallying the troops because Tarantino didn’t say that he stands with “the victims of criminally negligent homicide,” then he is resting his legitimacy on a thin reed.

And now that his membership is starting to see where all this posturing has got them in terms of a contract, Lynch’s post may be far from secure. He survived a weak challenge to his leadership this summer—three of the insurgent trustees were under indictment for allegedly fixing traffic tickets—but that was before the arbitrator released his findings.

In a department that is now 46 percent minority, where most officers are justifiably proud of the fact that they don’t engage in misconduct, a union boss who tells you what movies to see rather than getting you a raise may find that his time is limited.

The Mulgrew model offers a blueprint for what a legitimate challenger to Lynch may look like. Increase accountability and allow the department to fire problematic officers, and in exchange provide for real wage increases and better career paths for the rank and file.

New York Police Department officers read the papers, and they know which way the wind is blowing. If Lynch keeps singing the only tune he seems to know, his days appear numbered.

Everybody Loves a Firefighter

One highlight of my job as the spokesman and policy director of New York’s Civilian Complaint Review Board was the NYPD’s Advancing Community Trust program. Twice a year, the entire graduating class of the police academy would attend a day-long session at the Apollo Theatre. It started with group of high-profile panelists (Al Sharpton, Herbert Daughtry, Calvin Butts–one year Wycleaf Jean came) talking about policing in minority neighborhoods. The rest of the day was a series of role-play exercises and lectures. I gave an hour-long presentation on the CCRB and accountability.

But giving my presentation was not what made the day so enlightening. Instead, it was the keynote speech at the end by Wilbur Chapman. After a career in the NYPD and a controversial stint running the Bridgeport PD, Chapman has been brought back as the Deputy Commissioner for Training. I don’t know that he had any responsibilities other than to give a speech at ACT twice a year. But what a speech it was.

Chapman’s speech, “Everyone Loves a Firefighter,” stoked the bunker mentality, comradery, and paranoia already common in of most patrol officers. Chapman pointed out that the outside world would never give them credit for what they did to protect it, and that if they had been looking for easy accolades, it wasn’t too late to withdraw and join the Fire Department. Firefighters, after all, sit around making dinner for each other for eighty percent of their lives—and when they get caught stealing jewelry from a burning home, or drivinh a ladder while high on cocaine, you don’t hear about it in the tabloids and you don’t see protests at City Hall. Sanitation workers, Corrections officers, teachers, EMTs, building inspectors – all of them, Chapman said, live in a bubble, free from the constant second-guessing and irrational criticism faced by cops. The only people who will ever support you, he made clear, are each other.

Chapman was boosting morale, putting a spring in the step of officers who had felt beaten down all day. You could see the officers who had slumped in their chairs sitting up and paying attention. Still, it was almost incendiary—it encouraged and promoted the blinders that stand in the way of reform, the belief that no one else can really understand, so you may as well turn your back on anyone asking questions. But from Chapman’s perspective, it had one additional benefit. It was—and remains—pretty much true.

The current crisis on policing has been pitched consistently in terms of race—white officers against black civilians. You wouldn’t guess, from reading most columnists, that in a very short period of time, the NYPD has made enormous gains in diversity, and as older (predominantly white) officers age out of the force, it is on track to match the city’s demographics. Meanwhile, the FDNY has long been a bastion of white privilege—while the NYPD, on its own initiative, has increased minority representation on its force to 46%, the FDNY remains, even after a seven-year lawsuit, nearly 90% white. Corrections officers are notoriously abusive—a series of articles in the New York Times last year received national attention, but has not coalesced into any noticeable activism. The transit workers union fought the De Blasio administration tooth and nail to keep its members from being subject to criminal penalties when they kill pedestrians by running stoplights, and the progressive Working Families Party took the TWU’s side. A beat cop would be right to wonder, where is the “Inmates Lives Matter” movement? Where are the “Jail Killer Bus Drivers” signs? What makes us such special and inviting targets?

It has been well-documented that violence against police or on-the-job-murders are not on the rise, but what most police officers are talking about when they say there is a “war on cops” is not an increase in physical attacks on police, but an increase in how quickly and how unfavorably police are judged. It takes three years for a Corrections officer is videotaped and admits that he "stomped on an inmate's head" to go through the disciplinary process and be fired. But editorials call for cops to be immediately fired, or even arrested days after an incident. No one is demanding the resignations of the teacher and the administrator in South Carolina—who after all called the officer to the classroom to remove a student, watched him do so, and stated afterwards that he had done precisely what they had asked.

Those of us who work for or seek police reform, I believe, have an obligation to acknowledge that the problems are complex, the people involved in creating them are numerous (and often include ourselves), and that sometimes, it is our statistics or narratives that need correction. When the claim that a black person is killed by the police every 28 hours is given four pinnochios by the Washington Post, and the creator of the statistic writes back to say that so long as people join the movement, truth is “not the point,” she is playing the same game as Donald Trump. When a dashcam video supports the police version of an incident, we should acknowledge it. Otherwise we risk being the boy who cried wolf, and giving the police unions and officers good reason to ignore us when we point out real abuses, unjust policies, and the need for reform.

Police silence is a problem. The bunker mentality is a real obstacle to truth. We will not break it down in a day. But we will never break it down if we fail to acknowledge its legitimate roots, and to hold ourselves to an even higher standard than those that we critique.

Broken Windows Policing Doesn't Mean What You Think It Means

If you have followed police issues at all over the past decade, you have probably heard the term “Broken Windows Policing.” You might think that it means something along the lines of what the New York Times said it means this January, a policing strategy premised on the belief that “summonses issued and arrests made for minor offenses preclude the eruption of major crimes.” If so, you are completely wrong. But don’t worry, you are not alone.

George Kelling and James Wilson unveiled Broken Windows in the Atlantic in 1982. Their article synthesized two entirely unrelated lines of research into a single theory on police, community involvement, and civic pride, a theory that essentially remains untested because it has been so misunderstood that it has not really been implemented anywhere. 

First, the “broken windows” part of the theory, which postulates that people who would ordinarily not engage in vandalism will do so if they see that their neighborhood permits it. The key study here involved leaving a car on the streets of Palo Alto – no one touched it until the researcher broke one window, at which point, everyone took their turn and reduced it to scraps. This conclusion makes basic sense to all of us – you would never think to mark up the bathroom wall at Le Bernadin, but if you’re in CBGBs, what’s the difference. 

Second, the “policing” side. For this prong, Kelling and Wilson looked to a study regarding increased use of foot patrols in Newark. Patrol officers would enforce a set of informal rules: drunks could sit on stoops, but could not lie down. Strangers to the neighborhood were “sent on their way” if they loitered. The foot patrols did not reduce crime—they only reduced people’s perception of crime. Residents surveyed reported that they felt safer, though evidence showed crime had not decreased.

Kelling and Wilson hypothesized that this feeling of greater security was the stirring of an actual greater security. The idea is that once the community feels safe, people will take control of their neighborhood. Combined with the lesson of the vandalized car, the Broken Windows Theory of Policing guesses that by enforcing a community’s standards of order, patrol officers will empower that community to drive away more serious crime. Reading the article, you get the sense of the police as a bunch of low-level scolds, riding the buses and subways and kicking off people who drink or smoke, but not actually arresting anyone. Kelling has recently stated that to this day he is “not long on arrests as an outcome.”

This is a very different image than one has of the NYPD during (with apologies to Howard Safir and Bernie Kerik) the Bratton-Kelly-Bratton era.  During that era, the NYPD has engaged in two often-conflated policing strategies that in fact are quite distinct. The first is to enforce traditionally “minor” crimes vigorously, with summonses and arrests. This is not Broken Windows (where there are community-specific standards for what is enforced and arrest is rarely part of the equation); it is “zero tolerance,” a strategy traditionally associated with summons and arrest quotas.

The second strategy is stop-and-frisk.  Stop-and-frisk as a policing strategy, rather than a patrol tactic, is now associated with Ray Kelly, whose theory was and remains that if violent criminals know that there is a high chance they will be frisked for guns while walking down the street, they will not walk down the street armed. The only way to make violent criminals think there is a high chance they will be frisked is to stop and frisk an enormous number of people. The consequences of this strategy have been dealt with at length elsewhere, including federal court, and I won’t get into them more here.

The origins of this theory, however, are perhaps less well understood. The stop-and-frisk theory’s origin was a simple change in policy that Bill Bratton instituted as head of the transit police in 1990: he aggressively pursued fare-beating.  The theory then was that anyone that commits a major crime in the subway probably got on the subway by jumping a turnstile.  Stop the turnstile beaters, and you cut down crime. The theory worked spectacularly, and Bratton deserves enormous credit for this simple change. The drop in crime on the subways had big implications for the city--safe subway commutes were a precondition for major changes seen in Brooklyn and Queens in the past twenty years. 

When Bratton first became commissioner of the NYPD, he applied what he had learned in the transit system to the streets.  If muggers can be stopped at the turnstile, then maybe they can be stopped at the door of their apartment as well. 

But subways are a closed system, and going in without paying is at least actually breaking the law. The streets of New York are an open system, and the only way to stop the gun-packers was to go and find them. Stops and frisks increased initially under Bratton in an acknowledged search for guns.  They increased much more under Kelly, and in Bratton’s second stint the practice was shut down. But it was Bratton’s brainchild, an outgrowth of his successful transit strategy.

And while Bratton has largely curtailed stop-and-frisk, he remains loyal to what he calls Broken Windows, but which is actually zero tolerance.  It is not Broken Windows because not only are the enforcement standards the same across the city, and not only does it rely on arrests rather than a nudge, but it is not Broken Windows because the windows aren’t getting fixed. For the theory to work, the physical disorder must also be cleared away.  But a kid who lives in the Walt Whitman houses, seeing that his apartment has mold and the buzzers on his doors don’t work, who knows he can get a summons for riding his bike on the sidewalk or arrested for a joint, does not think that the police are enforcing community standards.

There is a movement now to “end Broken Windows policing.”  It’s hard to know what to make of such a plea, in light of the facts on the ground in New York. Because despite the fact that police commissioners have been citing to the article for decades, in New York at least, Broken Windows policing can’t be ended—it was never actually begun.