The “broken windows” theory gained currency as crime rates started going down in the 90s. So proponents could say “look it’s working.“ Of course crime went down everywhere including the many places that didn’t implement proactive policing. Just like crime rates went up everywhere after the pandemic— whether the areas were red or blue governed.
Just because two things occur together doesn’t mean one causes the other.
What it does indicate is that as you noted police have a questionable impact on crime and thats largely because of misallocation of time and resources.
If you look at statistics, police (other than those on TV) just aren’t very effective at solving crimes and historically never have been. (San Francisco PD’s poor closure rate is not uncommon.)
FBI data show only about half of all murders are solved. Over the past decade, “consistently less than half of all violent crime and less than twenty-five percent of all property crime were cleared,” according to William Laufer and Robert Hughes in a 2021 law review article.
Add to those figures the assumption that only about half of crimes actually get reported. So it’s seems likely police have a limited impact on crime rates.
Yes I think a part of the problem is what police actually do.
Officers spend only about 4 percent of their time on violent crimes. The rest is responding to noise complaints, traffic accidents and calls from Walmart about customers trying to misuse coupons. And paper work.
So if we want police to do a better job of solving actual crimes then we need to make sure they’re focusing their limited time and resources on the most serious violations.
Which means reforming not just police procedures and training, but also limiting the number of laws they are expected to enforce.
You and Matt Yglesias should debate this one. Here's his case for proactive policing-- specifically, for using things like fare evasions and traffic violations as pretext to search for illegally carried guns:
I fall more on your side because I think pretextual searches are a priori bad for liberty and corrupting to the rule of law: it seems intuitively obvious that in an even-handed justice system designed for a free people, the enforcement effort applied to any offense X should be proportionate to the actual social harms of X, not to the level of excuse it gives you to search people who might statistically be more likely to be carrying contraband. But a more purely first-order consequentialist person would be more sympathetic to Matt Y's take, so it may be worth specifically digging into the empirical rights and wrongs of the "NYC has low violent crime because they pretextually search lots of people for guns" argument that he's leaning on.
EDIT: never mind, now that I read the piece again I see you do give some evidence against the pretextual searches -> low crime thing. Thanks and apologies for morning skimming brain.
As a sociologist, I have to disagree. Proactive policing seems to have worked pretty well in NYC. What doesn't work is the idea of police waiting for service calls long after a crime has taken place. Broken windows is a good theory, and crime in NYC dropped precipitously after it was implemented. But sociologists and criminologists like to "churn" theories, so soon we were on to new, more dysfunctional theories that didn't work as well. One note: proactive policing doesn't work without crime mapping. You must station resources near where crime occurs.
I think it's been speculatively debunked-- there's a meta-analysis that purports to debunk it out of Northeastern University, but we'd need to know what the variables are in order to assess quality. The sheer number of studies involved doesn't inspire confidence. 200+?Not that many instances of the program happened, and no fair tapping more than one instance of the same site. Some researchers seem to have decided they just didn't like it, then proposed spurious reasons. Several admitted that the effect was real, but they didn't agree with the theory. (Never underestimate the lengths we academics- particularly those of us with no street smarts- will go to score points on other academics.) So that was it for them. Some pro studies were Kellog and Wilson 1982 and Kellog and Scuse 2001. Admittedly, the "broken windows" idea has to be bolstered with improved community ties, political economic supports, resistance to bad gentrification, good (probably dismounted) policing, and other good things.
Totally hear you on academics, lol. But yeah, that's my understanding of the debunking of broken windows. It's not that broken windows aren't associated with crime. Broken windows are associated with many of the things that are also associated with crime -- broken community ties, poor political representation, economic isolation, displacement, bad policing, etc. But in the case of broken windows themselves, the relationship does not seem to be causal. Or at least we can say with a decent amount of confidence that trying to fix broken windows through stop-and-frisk type policing doesn't seem to fix the broken windows or deter serious crime.
The show "We Own This City" I think was making the case for how destructive it is to incentivize easy arrests and let bad metrics guide behavior.
Such a great show!
The “broken windows” theory gained currency as crime rates started going down in the 90s. So proponents could say “look it’s working.“ Of course crime went down everywhere including the many places that didn’t implement proactive policing. Just like crime rates went up everywhere after the pandemic— whether the areas were red or blue governed.
Just because two things occur together doesn’t mean one causes the other.
What it does indicate is that as you noted police have a questionable impact on crime and thats largely because of misallocation of time and resources.
If you look at statistics, police (other than those on TV) just aren’t very effective at solving crimes and historically never have been. (San Francisco PD’s poor closure rate is not uncommon.)
FBI data show only about half of all murders are solved. Over the past decade, “consistently less than half of all violent crime and less than twenty-five percent of all property crime were cleared,” according to William Laufer and Robert Hughes in a 2021 law review article.
Add to those figures the assumption that only about half of crimes actually get reported. So it’s seems likely police have a limited impact on crime rates.
Yes I think a part of the problem is what police actually do.
Officers spend only about 4 percent of their time on violent crimes. The rest is responding to noise complaints, traffic accidents and calls from Walmart about customers trying to misuse coupons. And paper work.
So if we want police to do a better job of solving actual crimes then we need to make sure they’re focusing their limited time and resources on the most serious violations.
Which means reforming not just police procedures and training, but also limiting the number of laws they are expected to enforce.
Amen, amen, amen.
You and Matt Yglesias should debate this one. Here's his case for proactive policing-- specifically, for using things like fare evasions and traffic violations as pretext to search for illegally carried guns:
https://www.slowboring.com/p/why-most-gun-arrests-in-dc-dont-lead
I fall more on your side because I think pretextual searches are a priori bad for liberty and corrupting to the rule of law: it seems intuitively obvious that in an even-handed justice system designed for a free people, the enforcement effort applied to any offense X should be proportionate to the actual social harms of X, not to the level of excuse it gives you to search people who might statistically be more likely to be carrying contraband. But a more purely first-order consequentialist person would be more sympathetic to Matt Y's take, so it may be worth specifically digging into the empirical rights and wrongs of the "NYC has low violent crime because they pretextually search lots of people for guns" argument that he's leaning on.
EDIT: never mind, now that I read the piece again I see you do give some evidence against the pretextual searches -> low crime thing. Thanks and apologies for morning skimming brain.
As a sociologist, I have to disagree. Proactive policing seems to have worked pretty well in NYC. What doesn't work is the idea of police waiting for service calls long after a crime has taken place. Broken windows is a good theory, and crime in NYC dropped precipitously after it was implemented. But sociologists and criminologists like to "churn" theories, so soon we were on to new, more dysfunctional theories that didn't work as well. One note: proactive policing doesn't work without crime mapping. You must station resources near where crime occurs.
It's my understanding that broken windows has been pretty thoroughly debunked. So I'd be interested to see the studies I've missed.
I think it's been speculatively debunked-- there's a meta-analysis that purports to debunk it out of Northeastern University, but we'd need to know what the variables are in order to assess quality. The sheer number of studies involved doesn't inspire confidence. 200+?Not that many instances of the program happened, and no fair tapping more than one instance of the same site. Some researchers seem to have decided they just didn't like it, then proposed spurious reasons. Several admitted that the effect was real, but they didn't agree with the theory. (Never underestimate the lengths we academics- particularly those of us with no street smarts- will go to score points on other academics.) So that was it for them. Some pro studies were Kellog and Wilson 1982 and Kellog and Scuse 2001. Admittedly, the "broken windows" idea has to be bolstered with improved community ties, political economic supports, resistance to bad gentrification, good (probably dismounted) policing, and other good things.
Totally hear you on academics, lol. But yeah, that's my understanding of the debunking of broken windows. It's not that broken windows aren't associated with crime. Broken windows are associated with many of the things that are also associated with crime -- broken community ties, poor political representation, economic isolation, displacement, bad policing, etc. But in the case of broken windows themselves, the relationship does not seem to be causal. Or at least we can say with a decent amount of confidence that trying to fix broken windows through stop-and-frisk type policing doesn't seem to fix the broken windows or deter serious crime.