About . . . . Classes . . . . Books . . . . Vita . . . . Blog. . . . Podcast

by Peter Moskos

April 2, 2014

Bratton says morale was low in the NYPD under Kelly...

...And Bratton is right. Should be the end of the story, despite what you'll read in the Post and the Daily News.

More interesting is what is buried in the Daily News story:
Police made 12,495 stops between October and December — down a staggering 86% from 89,620 during the same time period in 2012. And of the stops during the last quarter of 2013, 16% resulted in an arrest. That’s up from 6% over the same period in 2012.
Raising the hit-rate of stops is a great indicator that more stops are based on actual real articulatable reasonable suspicion (you know, what is legally required) and not just quota pressure (er, productivity goals). 2,000 arrests from 12,500 stops is better than 5,400 arrests from 90,000 stops. Of course if these data simply mean more stops are unrecorded, this "improvement" could mean nothing...


Anonymous said...


What percentage of these stops do you think never happened? If senior police personnel can juke crime stats, why can't patrol officers juke their stop and frisk stats to keep management off of their backs. What would be the harm in an officer just making stuff up about a stop that resulted in nothing. No one is going to jail or being ticketed. It would be different if an officer faced pressure to write traffic tickets and lied about someone speeding because you need real driver and a car with a plate. Workers have always found creative ways to get around rules they feel to be stupid.


PCM said...

Good points. If I had to guess, I'd guess that something like 15% to 20% of stops were fictitious.... For all the reasons you say. (Of course even then not all stops were recorded.)

Matt said...

While not disagreeing with the sentiment of what you've said, I'm not sure that "2,000 arrests from 12,500 stops is better than 5,400 arrests from 90,000 stops" is necessarily true, at least not unreservedly true.

Fewer stops leads to fewer arrests, and that will inevitably mean that some offenders will escape when they would otherwise have been caught. In the present case (assuming, for the sake of argument, that all the previous stops were genuine and that the year-on-year change in offending population is small), roughly 3,400 offenders have not been identified by police who would have been identified last year. Many of these will be drug offenders, but some will be people in possession of illegal weapons and some of them will have gone on to use those weapons to commit more-serious crimes.

The overall number of such offences will, of course, be fairly small (but impossible to determine, particularly in the context of the overall crime drop). Nevertheless fewer stops almost inevitably mean that, for example, robberies will occur that otherwise would not have done.

This may be a price that people are prepared to accept in order to reduce the number of stops occurring. I only make this point because it is easy to forget that stop-and-search laws have to balance the rights of potential victims as well as those of people being searched.

PCM said...

You logic is impeccable. But so far it hasn't happened.

But I'm glad you raise the trade-off on rights/crime issue. Some say it doesn't matter (ie: their interpretation of constitutional rights trumps all... crime increase be damned).

I don't agree with that, but at least it's an honest opinion.

It bother me more when people refuse to discuss the possible link between stops and crime.

Each shooting does x amount of harm. Each stop does y amount of harm (a small fraction of x) but might also contribute to preventing x. So what's an acceptable ratio? Are 1 million stops worth it to provent 1 killing? No. Would 1,000 stops be worth it to prevent 100 killings? Yes. Somewhere between those two is a number I'm willing to accept.

The problem is -- for defenders of stops -- right now stops are way down and so are killings and robberies.... So what does that mean?

Based on the data we have right now, I'm questioning whether these stat-pressure driven stops have *any* crime reduction benefit (despite what seems to be the impeccable logic).