Schneier on Security
A blog covering security and security technology.
« Should the Government Stop Outsourcing Code Development? |
| Fifth Annual Movie-Plot Threat Contest »
March 31, 2010
Security Cameras in the New York City Subways
The New York Times has an article about cameras in the subways. The article is all about how horrible it is that the cameras don't work:
Moreover, nearly half of the subway system’s 4,313 security cameras that have been installed — in stations and tunnels throughout the system — do not work, because of either shoddy software or construction problems, say officials with the Metropolitan Transportation Authority, which operates the city’s bus, subway and train system.
I certainly agree that taxpayers should be upset when something they've purchased doesn't function as expected. But way down at the bottom of the article, we find:
Even without the cameras, officials said crime in the transit system had dropped to a record low. In 1990, the system averaged 47.8 crimes a day, compared with 5.3 so far this year. “The subway system is safer than it’s ever been,” said Kevin Ortiz, an authority spokesman.
No data on how many crimes were solved by cameras, but we know from other studies that their effect on crime is minimal.
Posted on March 31, 2010 at 1:24 PM
• 28 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
From the original article: "Part of the problem is that the authority is trying to retrofit a century-old transit system with the latest technology. Recorders are often housed in communications rooms, which have to be carved out of creaky, cramped subway stations. The requisite wiring and electrical equipment, built for the air-conditioned sterile space of a server room, do not always withstand conditions underground."
Unless they're trying to do something far more complicated than simply recording, playing back, and forwarding video, this is not rocket science.
You could take a SheevaPlug embedded computer running Linux and a 1 TB hard drive, stick them in a waterproof case with power and video feed coming in and Ethernet or encrypted WiFi coming out, and that would be a box the size of large paperback book that would last for years in a subway tunnel. And if, for some bizarre reason, it didn't, it would be cheap enough to replace frequently.
But then, that assumes people actually want to accomplish the goal and know what they're doing.
Cameras are a great example of post-crime "security", great for determining possible paths and even for identifying criminals (though it did not work in the Dubai assassination), but, "prevention" uses are a pipe dream, simply "forcing" a criminal to find other ways to do what they want.
and, I'm not even an "expert".
“Even without the cameras, officials said crime in the transit system had dropped to a record low. In 1990, the system averaged 47.8 crimes a day, compared with 5.3 so far this year. ‘The subway system is safer than it’s ever been…’
During the Monday morning commute, signs of beefed up security could be seen throughout the subways in response to the attacks in Russia. Officers on the midnight patrol, who usually get off at 8 a.m., were kept on through the rush hour, joining day patrols. And heavily armed special units were also assigned to the transit system on Monday.”
Crime throughout NYC has dropped a similar amount since 1990. Some wags attribute the change to implementation of the "broken windows" theory – by aggressively enforcing public disorder crimes their perpetrators a disrupted before they commit serious offenses. Others credit Roe v. Wade – fewer unwanted births, fewer miscreants, less crime. Otherwise, while flooding the transit system with submachinegun toting coppers may have some localized effect, like “Bruce sez,” public CCTV doesn’t prevent crime and only rarely helps solve them.
"Even without the cameras..."
Puts me in mind of former NYC Mayor Giulianni claiming responsibility for the nationwide drop in the crime rate.
In many of these circumstances, the installation work can be sub standard, and is mass installation with no quality control or post inspection.
Let me recommend that they replace those faulty cameras with these:
They perform as per their specifications, are a fraction of the cost, and they eliminate that troublesome "integration" problem.
This is also an example of organizations having the money to make large capital purchases without allocating the operating budget for maintenance and repair.
@Bill Bremer: "This is also an example of organizations having the money to make large capital purchases without allocating the operating budget for maintenance and repair."
A few years back, I read in SharkTank about an employee who was commended at work for negotiating such a fine bargain on a system. He got it at such a discount because he opted out of the operational manuals.
The first year in production, the maintenance costs were astronomical. The following year, the same employee who was commended for such a great deal was once again commended by management, this time for significantly reducing maintenance costs. How did he do it? He purchased the manuals.
Improper priorities and rewards are problematic. Entities need to anticipate maintenance needs. If they think that is expensive, they should compare it to the alternative.
Most cameras are only as valuable as how much they deter. I don't think they are ever completely useless, as even one that doesn't work will deter some criminals. But, as demonstrated, not many, especially not the real problems.
I thought it was here that I had read about a study that showed putting up security cameras caused the crime to move to where the cameras couldn't see, but I haven't been able to find the article so far.
How funny that security camera footage plays such a large part in so many television and film dramas these days.
"Even without the cameras, officials said crime in the transit system had dropped..."
They should say even without the working cameras. Perhaps the cameras, working or not is providing a deterrence effect; criminals don't know which ones are working on any given day.
"Unless they're trying to do something far more complicated than simply recording, playing back, and forwarding video, this is not rocket science."
Doing one is easy. Doing thousands is hard. Doing anything through a government procurement system, especially in New York, is even more difficult. It is easy for you or me to spec off-the-shelf components; it is far harder to sell a procurement manager on why all of them have to be water-resistant when a competitor's product is $3 cheaper per unit and _some_ of them don't have to be water-resistant. Also, the waterproof case means heat dissipation problems.
Wi-Fi . . . in an underground tunnel? In many cases the problem _is_ the wiring. When they built the subway system, no one ever thought of needing more than a little copper for phone and switching. Additional wiring must survive cold, heat, damp, roaches, and last but not least, determined vandals.
Ha ha. Other than being obvious, vulnerable to vandalism, and creating massive premise liability for the client ("I felt safe . . . there were cameras . . . I was mugged . . . I'm suing the MTA for damages, they installed cameras they KNEW DIDN'T WORK.")
Bingo. The techs (and their bosses) often get paid whether or not the installation 'works.'
How many Teamsters does it take to change a light bulb? Twenty you gotta problem with dat?
I assumed the goal was not detering crime on the subway, but helping to track the movements of people in a transportation system in which they were otherwise almost impossible to follow. No idea how well the cameras serve this purpose, although presumably not well if they do not work. Of course, if you want to encourage criminals, terrorists, and other undesirables to use the subway then TELL them the cameras don't work....
The crime rate "even without the cameras" is an unknown, because (at least until now) most people probably were unaware the cameras didn't function.
"Additional wiring must survive cold, heat, damp, roaches, and last but not least, determined vandals."
Don't forget rats. When the phones in my (NYC) building went dead a couple years ago, the repairman explained it was because rats had chewed through the cables under the street.
Can't speak as to why crime is down. Interesting since today's youth is generally considered to be more violent because of exposure to violence and sex on TV and video games. Wasn't it Stalin who ranked #1 on Grand Theft Auto?
The funny thing is that they indicate that crime has fallen, when in fact any such study can be skewed. If anyone here has watched the Brazilian movie Elite Troop, the commanders on that movie played a game of "not my corpse". If anyone found a corpse on the territory they are assigned to, they would move the corpse so it would be "found" on some other commander's turf and so it becomes someone else's statistical problem. Can't say this is what has happened but I wouldn't be surprised if that had something to do with it.
People, as Bruce and many others have mentioned, prefer much more the feeling of safety, that something can be done, rather than tangible security improvement that they cannot feel or see but actually have some effect.
It's always difficult to implement things when money is tight. Cutting corners is common and most people don't like to look at the overall cost of the system besides the immediate money. Remember that the Iphone in fact costs around $1500 for a year. Granted, this is including service but it does show that the small entry fee means more will be required later. Same with the cameras. Some guy gets the idea that installing them would be a good idea, but makes not budgeting or planning for their actual use. Any why would they? DVR can record it and they can use it later. They don't worry about telling the victim that someone observing the cameras could have stopped the crime as it was happenning.
People love technology and how it makes things easier, but they don't realize that it can never replace human intuition and presence.
"The funny thing is that they indicate that crime has fallen."
Crime is far lower in NYC than, say, 15-20 years ago. It's obvious just walking around for people who live here.
It's not right to just dismiss info that's fairly well-established by just saying " any such study can be skewed"
Especially for someone who uses these sorts generalizations: " today's youth is generally considered to be more violent because of exposure to violence and sex on TV and video games"
Cameras working or not is one thing and comes with a different set of discussion all together...
The main point (which I believe) leading to drop in the crime rate is the presence of cameras itself (irrespective of them working or not). It is the sense of being "watched" is what leads to drop in crime rate. It is something which you may call as a pseudo-security - but psychologically it definitely works!
Well, now that it is well known the cameras don't work, we should see an increase in crime in the affected areas.
I didn't catch Bruce's 25th February article (which he cites above) when it came out. But boy, what a doozy.
He states "Pervasive security cameras don't substantially reduce crime. This fact has been demonstrated repeatedly: in San Francisco public housing, in a New York apartment complex, in Philadelphia, in Washington, DC, in study after study in both the U.S. and the U.K. Nor are they instrumental in solving many crimes after the fact."
In support of this claim he links to many articles -- most of which actually contradict the claim! The few that don't are so weak they can certainly be called the cherry-picking -- ironically as Bruce accuses camera supporters of being cherry-pickers.
Examining them one by one:
Far from "demonstrating" anything, the only pejorative statements are anecdotal claims, by just two people, that they personally have no recollection of this small camera installation solving a murder -- however it has definitely solved other crimes. The main thrust of the article is that the system has poor technical performance because its installation and maintenance budget is so stingy (averages just $228 per camera per year, including purchase, installation _and_ on-going maintenance!)
Behind a paywall, but I managed to obtain a copy elsewhere. It DOES NOT demonstrate that CCTV was ineffective. Quite the opposite; it found reduced crime rates in both sites studied. In the smaller site, the quantity of data collected was small and the reduced crime rate was considered not statistically significant; in the other site, the reduction of crime WAS statistically significant.
I can't understand why Bruce even included this one. It clearly states at the outset that its conclusion is exactly the opposite of his thesis: this CCTV installation WAS EFFECTIVE at reducing crimes rates, the results were statistically significant. They do conclude that there is a big difference in effectiveness between a carefully planned installation vs. just blindly installing cameras.
Far from "demonstrating" anything, this news article provides no data on effectiveness one way or the other. It quotes a representative of some sort of activist group who doesn't like them, and quotes objections by a retired police chief whose former policy was implicitly criticised by the incumbent. Apart from that, all we can say from this article is that Washington DC has very few CCTV (astonishingly few for a national capital.)
Bruce has several times in the past misrepresented this study as concluding that CCTV is ineffective. It actually:
* criticises the fact that most security measures (not just CCTV) are launched with too little thought to measuring effectiveness, in many cases making it impossible to say whether or not they have worked
* finds huge difference in effectiveness between carefully thought out installations vs. installing cameras willy-nilly
* far from demonstrating that CCTV is ineffective, this report identified a number of areas in which it was very effective
Once again, I can't understand why Bruce even cited this paper. The main conclusion is that there was a statistically significant in rate of crime!! Another conclusion is that most residents interviewed felt that the system had been installed without invading their privacy.
This is not a study. It is a policy guide for police officers. However in two appendices it does include some potted summaries of other studies. Twenty studies are cited in summary form; 14 of these showed an unequivocal benefit to CCTV. In some cases the benefit was small, but in carefully planned installations in high crime areas it was dramatic. The 6 other cases included cases where some reduction was seen but it may have simply been displacement, a case where violent crime and serious theft were reduced but petty theft increased, a case where the researchers were unable to obtain sufficient data for analysis, and two cases where the installation seems not to have worked.
This guide also cites the Gill & Spriggs Home Office study already discussed above.
A news report which in typical journalistic style cherry picks a government report to put the most controversial possible spin on it. Nevertheless the journalist must have had some integrity, because the article does conclude that CCTV reduces crime. Interestingly this article harps on quite a bit about the supposed superiority of street lighting as a crime fighting tool -- something which has been much discussed on this blog before, because many studies have shown street lighting to be so ineffective as a crime fighting tool, that there are examples of towns reducing their crime rates by disconnecting the street lights.
Another news report which puts the most sensationalistic spin on a statement. The article's headline "CCTV boom has failed to slash crime" is totally unsupported within the article itself. The statement actually does not discuss the effectiveness of CCTV for crime _prevention_; it claims that so much attention has been put on it for crime prevention, and so little for prosecution, that in many cases it is difficult to use for prosecution -- and gives suggestions on how to remedy this. The figures on clear-up rates wildly contradict those that this blog discussed back in August.
The large city near me just had a shooting of 5 people due to a "flash mob" in a park. The park has 20 cameras that saw almost nothing of the shooting. What they did see was dark and grainy as they don't have the nighttime upgrade.
"Over the years, a variety of improvements have been made to the cameras, including better focusing and long-distance capabilities. But the cameras never were intended for the kind of conditions experienced Saturday night, she said."
She means darkness as the kind of conditions that were experienced.
Oh, and the city took great pains to opt out of a recent state law change that let carry permit holders carry in parks. Apparently, the local ban had little impact on, you know, criminals.
For the NYC subways, in addition to the water/humidity, the rats, the vagrants, the cockroaches, the temperature extremes (-20 to +60 C), you've also got a really serious RFI environment. Multiple 600V arc sources.
Instead of using CCTV, the better choice (and as cost effective) is IP Video. While the unit cost of the cameras is higher, when you remove the cost of the coax (and allow for PowerOverEthernet), the cost comes out the same. Add to that the digital quality photos (and some of the cameras go down to 0.1lux (yes that is 0.1), and you get a low light solution with no moving parts.
High quality pics, low cost rugged storage, what more could you ask for?
And in related news, the U.S. Forest Service recently admitted to putting
motion-activated surveillance cameras into remote locations in National Parks.
Searching on "surveillance camera National Parks" will return links to the
"when all you have is a hammer, everything looks like a nail"
"The large city near me just had a shooting of 5 people due to a "flash mob" in a park. The park has 20 cameras that saw almost nothing of the shooting. What they did see was dark and grainy as they don't have the nighttime upgrade."
According to the on-line news reports I have just been reading, there were two investigations under way. One concerns the drive-by shooting, the other concerns the rioting and violence by the flash mob. While initial statements suggested that the shooting was due to the flash mob and was not gang related, this now appears to be untrue and available evidence indicates that the investigations are unrelated. It is believed the shooting was in revenge for a separate gang-related drive-by shooting the previous day, and it just happened that the gunmen's target was at the park.
According to these reports, the cameras didn't film the drive-by shooting because it happened to occur in an area not covered by the cameras (probably because the shots were fired from a car that was outside the park), however it hadn't even been requested because the shooters were quickly identified and were in custody 11 hours later. What the CCTV footage was actually wanted for was identifying the instigators of violence amongst the "flash mob" -- and for this purpose it may be a bit grainy but it is adequate, as it has also been stated that it is going to be used as evidence in the ensuing prosecutions.
Apparently this is the second Chattanooga "flash mob" to turn violent, the first being a student rave less than a year ago. However the problem of violent flash mobs is much worse in Philadelphia.
@Anon-y-mouse 03APR10 1219 -
Not to be nitpicky, but it was a National Forest, not a National Park. Parks come under the umbrella of the National Park Service and many already have surveillance, since many are historical parks in urban areas/are very popular (Yellowstone, Grand Canyon) and so on. I've only worked in the urban parks, so I can't say how prevalent they are in the natural parks. Natural Forests, on the other hand, tend to see less traffic since they are seen as natural resource reserves rather than recreation areas, although they do have campsites, trails, etc. run by the Forest Service.
However, both the Forest Service and the Park Service have had trouble with pot farmers setting up huge plantations on federal land, sometimes quite close to campsites. These setups are often very complex and always well-defended, so it may be that the rationale for the cameras in the middle of nowhere was to keep an eye out for them. Not saying it's necessarily a great idea, just that there may be a reason other than spying on blissfully ignorant campers.
I'm going to trust CTU and Jack Bauer on this one.
I would think security cameras are best for after the fact review ... this is what makes great footage for the "crime stoppers" section of the evening news ... "does anyone this idiot?". But how long do you keep all the recordings? Image how much storage is required for all these cameras.
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.