Tuesday, March 15, 2005

Why surveillance cameras don't reduce crime

Dennis Bailey at the Open Society Paradox calls me out on yesterday's Grits post about surveillance cameras, and offers up examples of surveillance "sucess stories." To me, the wisdom of proliferating security cameras should be evaluated based on the empirical results of a three-way tradeoff between security, liberty, and measurable, real-world effectiveness. So I'm happy to see his set of success stories. Taken together, the two posts supply readers with a pretty good range of cases documenting both risks and benefits, pros and cons of camera surveillance.

Some of his examples are private surveillance, not government, which is a different matter, though when enacted on a large scale, it too deserves scrutiny. I don't know anybody, though, who thinks there shouldn't be a camera behind the counter at the convenience store. Still, even in his other examples, serious crimes were committed in the cameras' presence, though perhaps it was easier to prosecute crooks (unless, I suppose, they operated at night or wore a hat).


Earlier I'd cited longitudinal research showing widespread use surveillance cameras doesn't reduce crime. They've been used long enough in Britain now to allow extensive analysis of the resulting crime data.

So Dennis' post got me wondering why, despite its internal logical consistency, the idea of camera surveillance reducing crime never played out that way in the real world in Britain where it's much more widespread (except in a few specific places like parking facilities).

A hypothesis:


1. Human beings are creative and criminals actively try to get around or disable surveillance.


2. Police have limited resources.


3. If cameras are being monitored, police resources are being directed wherever the cameras are being focused. I'm not just talking about time spent by the video-monitor watchers, but where patrol units are sent to respond. By definition, that excludes every other portion of the world where they aren't looking, which is inevitably SOMEWHERE (or if cameras are everywhere, then you get lapses like in the Atlanta case), so because of (1) above, criminals are inevitably able to adjust to the new environment.


For example, take a look at this typical quote from
one of Dennis' examples from a Washington state police chief: "I can have one person watch 11 monitors, whereas I would have to have 11 officers patrolling to see the same things." But that's not true, because what that officer is doing is DIRECTING the other officers, who must respond if the first one sees anything. More to the point, over time officers will be more likely to be directed there, so cameras represent an unplanned overallocation of the state's limited policing resources.

If that's true, then the reason camera surveillance doesn't reduce crime might be that the theory of cameras as a crime preventive fundamentally misunderstands both crime, and the way cameras usurp human police management decisions by overallocating scarce officer resources. If I'm right, in the end, at best, cameras target only the most stupid criminals, for a while.

16 comments:

Dennis Bailey said...

Hey, If I have been able to get you to agree that surveillance cameras work at least against "stupid" criminals, I feel like I'm making headway. However, it does make me wonder whether I'd rather be accosted by a smart or intellectually inferior miscreant.

Gritsforbreakfast said...

For a while, and also under very particular, now well-proven circumstances -- again, we now have empirical data as to where they work and where they don't As a general principle, though, and as a long-term crime reduction strategy, I think camera systems are expensive and misguided.

BTW, you probably don't care whether you're assaulted by a smart or dumb criminal - that's the point! If the total number of assaults don't go down over time, what have you gained?

Peter said...

Another thing to factor into the debate are the estimates made within the criminal justice system as to who they actually catch.

Years ago, I worked for the Prison Service in the UK. At that time, it was estimated that we catch about 5% of sex criminals. Not prosecute, not convict, but catch.

Thus, pretty much all of our intelligence on sex crime is based on the least sophisticated and competent 5% of sex criminals.

I do not know even the numbers back then for other kinds of crime, but I would guess that it doesn't get much better. I suspect that we're getting better and better at catching the same people over and over - that's certainly how it went in the UK, since recidivism (and reconviction) was pervasive.

Interestingly, in my 3 years there, I don't recall a single instance of video evidence being used to help convict someone. My memory may be flawed though.

Final thought - one man watching 11 cameras is actually a serial process IMO, not a parallel one. I do not believe a person's attention can be meaningfully directed at 11 screens simultaneously, even if you're watching 11 TV shows and your brain is helping out by filling in the spaces.

bhurt said...

Actually, widespread camera use, especially if it allows for reductions in police force, make crime easier for smart crooks.

Let's perform a thought experiment. I want to go mug someone. Consider the case without surveillance, where we have a police force with 11 officers roaming around more or less at random. The nearest officer might be 10 blocks away, or the nearest officer might be just around the corner. The risk to me, as a mugger, is that I don't know where the nearest officer is.

Now, consider the case with surveillance. The police force is now 2 guys- one watching the cameras, and one waiting on call to go deal with trouble. Now I can mug people with impunity- because I know where the cops are. I know that even if I mug someone right in front of a camera, it'll be a couple of minutes at least before the police arrive. I can get their wallet and be long gone before the cops get here. Now all I have to do is defeat the cameras, which requires high tech equipment like ski masks, wide brimmed hats, and overcoats.

Even if the second cop is out wandering around, so there's a chance the lone cop is around the corner, you've still made it much safer for me to be a mugger- as the odds of the cop being just around the corner have dropped by over an order of magnitude (11 cops that could be just around the corner vr.s 1 cop that could be just around the corner).

There is no replacement for boots on the ground- none.

bhurt said...

Actually, widespread camera use, especially if it allows for reductions in police force, make crime easier for smart crooks.

Let's perform a thought experiment. I want to go mug someone. Consider the case without surveillance, where we have a police force with 11 officers roaming around more or less at random. The nearest officer might be 10 blocks away, or the nearest officer might be just around the corner. The risk to me, as a mugger, is that I don't know where the nearest officer is.

Now, consider the case with surveillance. The police force is now 2 guys- one watching the cameras, and one waiting on call to go deal with trouble. Now I can mug people with impunity- because I know where the cops are. I know that even if I mug someone right in front of a camera, it'll be a couple of minutes at least before the police arrive. I can get their wallet and be long gone before the cops get here. Now all I have to do is defeat the cameras, which requires high tech equipment like ski masks, wide brimmed hats, and overcoats.

Even if the second cop is out wandering around, so there's a chance the lone cop is around the corner, you've still made it much safer for me to be a mugger- as the odds of the cop being just around the corner have dropped by over an order of magnitude (11 cops that could be just around the corner vr.s 1 cop that could be just around the corner).

There is no replacement for boots on the ground- none.

Peter Kaufmann said...

Here is a personal experience.

I was waiting for a train in a subway station when two men started fighting a few meters from one of may dome surveillance cameras. It was kind of brutal; they even crashed their heads against steel pillars.

Nobody engaged, everyone was waiting for the sentry which is pretty well present at stations that have no cameras.

Guess what? Nobody came. Not after 5 minutes and not after 20 minutes. (I was still there because my train had a severe delay.)

-- Peter

Anonymous said...

And of course everyone there, including you, just stood around and watched the bloody show unfold, eh? We expect cameras and more cops to make up for the fearful apathetic society we live in. Then, you get to go home, tell all your friends about the beating you observed and feel you're getting your very own ten minutes of fame. If you'd all get in there together and STOP criminals, crime would actually take a downturn.

Let me know when you get up the nerve to be real citizens. I want to watch. Anonymously, of course!

Anonymous said...

i dont get it
hi

Anonymous said...

And if you do get involved trying to stop a fight, then you too can be arrested. You 'appear' to be involved in the fight and it would take some evidence to prove otherwise... hopefully the cameras might help here... but who knows?

It still isn't a substitute for properly trained officers on the beat.

Digital Camera said...

A Camera cant' reduce crime until human being are conscious about crimes

Fairings said...

There is always the human factor that can allow people to commit crimes just because of the failures of the security guards.

Mexican Breakfast Food said...

They don't work well here because people destroy them so government never replaces those cameras and security cannot be accomplished.

Hidden Cameras said...

Well the key is to use hidden cameras. For example a bird house outdoors or something similar. They can't disable what they don't see or suspect.

Security Cameras said...

Yes, in an ideal world, we need human resources like the cops to police all areas but that's not very practical. It requires a lot of manpower and financial resources. So the only thing left is using security cameras...also security cameras might not help much in intervening a fight but they will definitely help if someone is gravely injured from the fight or god forbid even killed! The cameras are definitely useful to find the perpetrator and figure our the sequence of events.

security cameras said...

You are right security cameras do not reduce crime. We need human resources to do that but it can catch crime.

Maria Jose Tobar said...

I've always been the type of person to laugh at commercial security, even in Albuquerque. I always look at the security cameras with a smile on my face. I've heard nobody even watches though.