About . . . . . . Classes . . . . . . Books . . . . . . Vita . . . . . . . Links. . . . . . Blog

by Peter Moskos

June 22, 2016

Attacking Broken Windows Again

There's a report out by the newfangled NYC Department of Investigation Office of the Inspector General for the NYPD (you know, OIG-NYPD, for short): "An Analysis of Quality of Life Summonses, Quality of Life Misdemeanor Arrests, and Felony Crime in New York City, 2010-2015."

The report is surprisingly good, in terms of data analysis and presentation. (I love, for instance, how somebody cared and took the time to explain how the data in the charts should be read.) Though it seems strangely political that one of the first things the office does is produce a report to be spun as "Broken Windows Doesn't Work" (Despite evidence to the contrary). From the OIG report:
Issuing summonses and making misdemeanor arrests are not cost free. The cost is paid in police time, in an increase in the number of people brought into the criminal justice system and, at times, in a fraying of the relationship between the police and the communities they serve.
The report limits itself (sort of) to:
what, if any, data-driven evidence links quality-of-life enforcement--defined narrowly for purposes of this Report as quality-of-life criminal summonses and quality-of-life misdemeanor arrests--to a reduction in felony crimes.
From 2010 to 2015, it doesn't find any. Both misdemeanor enforcement and crime went down. Ergo, Broken Windows must be broken.

But no Broken Windows advocate thinks there's a one-to-one correlation between misdemeanor summonses and lower crime! Both can go up. Or (in the ideal Broken Windows world) both can go down.

In the six years before 2010 (the starting year for the report) misdemeanor arrest in NYC went up substantially (190,346 to 245,400) and murders went down (570 to 471). Of course when arrests are up and crime is down, the anti-Broken Windows klatch says correlation doesn't mean causation (even though sometimes it actually does). But of course when the data works for them, correlation "proves" Broken Windows doesn't work.

But perhaps, to give police a bit too much benefit of the doubt, the NYPD simply reassessed what needed to be done. Some would call this problem-solving policing. And the NYPD actually has a pretty good track record of this over the past 25 years. Tactics change. Times change. Reassessment is a key to problem solving. As old problems go away and new problem appear, police don't need to keep making the same quality-of-life arrests.

But I mentioned "too much benefit of the doubt" because police were and are wedded to the idea that all arrests are good, and more arrests are better. This is wrong. And the recent reduction in small-scale enforcement happened not because police under Bloomberg and Kelly wanted to reassess their strategies but because the department was dragged kicking-and-screaming by lawsuits into the political reality of a lower-crime New York City.

This OIG report does a great job in linking police enforcement to violent crime.
Higher quality-of-life enforcement rates in precincts with higher proportions of residents who are Hispanic or living in [public housing] may be related to violent crime rates in those precincts. (p. 44)
You think? We need minor arrests and citations, especially when they're given to major criminals.

What's interesting is that when one takes violent crime into account:
White residents receive higher[!] rates of quality-of-life enforcement, and precincts with higher proportions of residents who are black or males aged 15-20 receive lower[!] rates of quality-of-life enforcement than would be anticipated given these precincts’ violent crime rates.

This goes against type. It could mean (à la Ghettoside), that given the crime rate, communities with high-crime are actually under-policed. Or it could mean there is no connection at all between violence and police enforcement, and police just happen to be harassing blacks in high-crime areas. (And these position are not necessarily mutually exclusive.)

Anyway, I applaud the report for at least considering violent crime as a relevant factor. Because it is. Such politically-incorrect honesty is shockingly rare. But more importantly -- though I think there is a link between good Broken Windows policing and a reduction in serious crime -- quality-of-life issues deserve police attention for their own sake. Even if the Broken Windows theory (unattended disorder leads to more disorders and serious crime) can't be proved, we still need Broken Windows policing because order maintenance and quality-of-life issues matter for their own sake.

The major problem with this report is that it doesn't take 911 and 311 calls for service into account. This is a serious omission. Police get called to deal with "minor" issue because neighbors don't think they're so minor. Police have little control over whom they interact with. As Bratton put it in 2014:
“The idea that we can engage in policing that’s racially proportionate is absurd,” he told reporters after a panel discussion in Manhattan about Broken Windows.

Quality-of-life enforcement, he said, is driven primarily by complaints made to the city’s 311 hotline, meaning police action is in response to citizen complaints.

“We go where the calls come from, we go where the help is needed, we go where the victims are, and that’s the reality," Bratton said. "If those numbers are racially disparate, or disproportionate, well, that’s the reality."
Assuming calls for service are concentrated in high-crime minority areas (because they are), what are police supposed to do? Wait for some white people to walk by engaging the public?

The other half of the story, the bad half, is that quality-of-life crackdowns come from nervous and insecure precinct commanders. Very few people call 911 to ask cops to stop and arrest people for nickel-bags of weed. But it happened 83,000 times in 2010. That's non-intelligence-driven policing. Too often commanders face Compstat pressure and need to "do something" to keep the brass off their back. Quality-of-life policing can and must be part of real policing, and not just a way to generate numbers or revenue.

The 71 in Crown Heights, for instance, went crazy giving tickets to bicyclists. People were not complaining about bikes without bells. But a commander wanted "numbers," and "numbers" he got. This wasn't real policing, much less quality-of-life enforcement or Broken Windows policing. But the data in the report can't distinguish between good misdemeanor enforcement (Broken Windows) and bullshit misdemeanor enforcement (Zero Tolerance).

Or take this example I wrote about in 2009:
Police know the difference between “good” and “bullshit” stats. One ranking NYPD officer told me he neither asks for nor approves of bullshit citations from those under him. He gave an example of a public park closed at night: “If the park were used by people to party—smoking and drinking--we would encourage citations. But if people were just using the park as a shortcut coming home from work, I wouldn’t want officers citing those people. That’s an excellent use of discretion.” He’s right, and an officer under him acknowledged his superior’s ideals. But he added, “I’d love it if I always had enough good C’s [criminal citations], but I need numbers. And if I don’t have enough stats and CompStat is coming up, I don’t care if they’re bullshit. I’ll take whatever the f*ck I can get!” In a world where “better stats” and “more stats” are synonymous, the tail has long since started to wag the dog.
Broken Windows is not Zero-Tolerance enforcement. Key to Broken Windows is proactive order-maintenance policing that targets quality-of-life issues and public fear. Neighborhoods with more violence fear should be targeted more heavily for selective misdemeanor enforcement.

With less policing, crime and violence rise. (The former especially in neighborhoods with more criminals and the latter especially in neighborhoods with public drug dealing.) Shamefully, many in the police-are-the-problem camp refuse to accept any cause and effect between less policing and more crime, particularly in cities such as Baltimore and Chicago beset with passionate but unfocused calls for "police reform."

Hopefully we'll see police as part of the solution and demand proactive, smart, quality-of-life policing responding to citizens' fear in high-crime areas. But then police will focus disproportionately on blacks and hispanics in high-crime neighborhoods.

Or we can continue down the wrong road and see police as part of the problem. As murders increase nationally, we shouldn't be debating the crime rise and quibbling semantics over the "Ferguson Effect" and "Broken Windows."

If we restrict policing and police discretion in order to return to the failed call-and-response police model of the 1980s and early 1990s, police will still focus disproportionately on blacks and hispanics in high-crime neighborhoods. But less with the citation book and more with the crime-scene tape.


Andy D said...

I guess the question, as in all public policy, is: from an administrative/civilian control point of view, how do you ensure that the stats you are getting are good, rather than bullshit? I think Compstat etc are credited with improving crime in NYC, but at a certain point you end up with, as you point out, the tail wagging the dog. This seems to be why when NYC-style Compstat-driven policing was exported to placed like Bmore and Chicago, it failed: you get the "letter of the Compstat" (more stats) rather than the "Spirit of Compstat" (more GOOD discretionary stats.)

Is this because bosses don't buy in (i.e. they just want "MORE NUMBERS!" to look good at the Compstat meeting rather than "Good Numbers?") Is it because the beat cops don't buy in? (i.e. "This is bullshit. Where can I grab some quick easy stats?") Or something else?

Even in my (smaller, more rural) department I have seen the progression through my career: When I first started 14 years ago, the guys on the street were very reactive: they didn't do shit until they got a call. They were mainly older and nearing retirement. Over the first few years a lot of younger, more aggressive guys came to the street. They (we) started aggressively pursuing Quality of Life and pro-active enforcement, and the crime rates went down. But we became a victim of our own success. The bosses kept requiring MORE STATS despite dropping crime numbers. So as more and more successive waves of new guys hit the street, we got more numbers but the quality of the numbers went down. People seem to be getting hassled who don't deserve it. Instead of using discretion to issue warnings or taking no action for people who don't need a summons, a summons gets issued to everyone it seems. The community can tell. They get irritated and cooperate less with us (even the "good" people.)

To me, in my department, this is because the bosses didn't (don't) "get it." They just want MORE all the time. And a lot of the youngest guys on the street don't get it either; they equate more enforcement with good enforcement and look at me strangely when I say someone doesn't deserve a summons or an arrest.

If we can't seem to get it right in a small department, how do you get it right in Chicago or Baltimore?

Jay said...

1. Do you have any idea who actually did the research and wrote the report? It’s exceptionally good for something coming out of a government bureaucracy.

2. Could the data on QOL enforcement, violent crime, and race be summarized this way: “Precincts with more violent crime get more QOL policing– especially when those precincts are mostly White.”

3. Could the title of your post be “Attacking Broken Windows Again – This Time with Good Evidence”?

4. The JRCD meta-analysis you linked to concludes that there is “an overall statistically significant, modest crime reduction effect.” It also says that the modest crime-reduction that QOL policing does produce does not come from making misdemeanor arrests or in other ways focusing on disorderly individuals. It comes from making the area appear more orderly.

The term Broken Windows is itself misleading. It implies focusing on making the neighborhood look more orederly – literally fixing the broken windows. In practice, what it came to mean was – to use a phrase from the original article – “kicking ass,” which is precisely what the JRCD article says does not work.

5. I haven’t kept up with the research, but when I was looking at this stuff 20 years ago, it seemed clear that you have three important variables – predatory crime, disorder, and fear. The strongest link was between disorder and fear. People feel safer when the neighborhood looks safer. Rates of victimization have less of an impact on fear. It also seemed that cops (and government generally) could do much more about disorder than about predatory crime. So QOL policing is important. It can make people feel better about their neighborhood. But only if the focus is on the actual quality of life, not on kicking ass for the sake of numbers.

David Madden said...

Hello everyone. Long time listener, first time caller...
Andy, its like we work for the same department. Stats for the sake of stats. We work in a culture where the "performers" are rewarded, even though they screw up anything more complicated than a traffic stop. How do we change the system to reward hard work that cant be quantified?
Btw, I just finished Cop in the Hood. Great book, lots of great ideas and insights. Too bad that only stat crazy dipshits seem to become chiefs and would dimiss your ideas out of hand.

ICP said...

Jay--the (relatively) new inspector of the NYPD IG's office is Phil Eure, a guy people in the oversight world hold in high esteem for his integrity and professionalism. I would wager it's his report.

Peter--well said--and in the spirit of James Q: "Even if the Broken Windows theory (unattended disorder leads to more disorders and serious crime) can't be proved, we still need Broken Windows policing because order maintenance and quality-of-life issues matter for their own sake."