July 24, 2005
Deterrence and Subway Searches
Posted by Dave Hoffman

[Warning: The following post contains unrealistic economic models of the behavior of very evil people.]

Dan Solove at Prawfsblawg posted a few days ago about the new NYC subway search policy.  Key graf:

I'm dashing this post off quickly, so I won't explore the legality of this.  But regardless of the legal issues, this seems to me to be a very silly policy.  It is another big waste of money and time, as well as a needless invasion of civil liberties -- all for a cosmetic security benefit.  There are 4.5 million passengers each day on the NYC subways.  What good could a few random checks do?  The odds of the police finding the terrorist with a bomb this way are about as good as the odds of being hit by lightning. I doubt it will have much of a deterrent effect either.

Unlike Dan, I'm not an expert on privacy law.  But I do think some about deterrence, and I think he's underestimating the effect of random searches on public safety. 

Assume the police are executing a search policy (nicknamed "Ring of Steel") that results in 1000 random searches a day; this results (by my very rough calculation) in a 7% chance of being searched over the course of a year of weekdays.  We have to assume that are many intermediate steps necessary in conspiring to harm others; if the terrorists have to purchase, for example, washing machine timers, ordinary subway travel becomes hazardous to the plan.  (A cheap point: these are shorter odds than the 1 in 3000 chance of being hit by lightning over the course of your lifetime, although of course the odds of being searched on any given day are 1 in 4500).

Under these assumptions, Dan is probably wrong that the search policy will have little deterrent effect.  Thinking about the issue from a behavioral perspective, some keys are:

  1. Risk Aversion:  Terrorists are notoriously risk averse - they obey the law punctiliously until they attack. Thus, even a relatively minor risk of being caught will act as a very large deterrent, forcing terrorists to find other paths.
  2. Decision-making under uncertainty: Subway searches are going to be perceived (true or not) as targeted at certain ethnic minorities; the police, if they are smart, will also vary the number of searches per day.  These two effects together should make it hard for terrorists to intelligently evaluate the likelihood of being stopped on any given trip.
  3. Availability Cascades:  Again, smart police should perform searches in highly visible places, and trumpet "sweep" days (think of the IRS's press machine in March.)  The goal will be to convince folks that searches are much more common than they actually are, and to equate "going in the subway" with "possibly getting searched: better not bring illegal goods down there." [Maybe this sounds too big brother like for some.  Arguably, it would be better for society to pay the privacy cost of having lots of actual searches, so that folks could evaluate whether they think them to be worth it, rather than allowing the government to mind game the populace.]

Regardless, I think there is a good argument that terrorists, subject to human behavioral tics, are likely to vastly overestimate the likelihood of being caught and therefore be more deterred than rational terrorists (what a contradiction in terms that is!) would be.  So I have to disagree with Dan.  Subway searches can, if intelligently carried out, make it much more costly to plan and execute mass transit attacks.  Will terrorists then move on to other targets of opportunity?   Probably.  But forcing them to do so would be a victory.

Dan notes we should weigh the gains against the costs.  Fair enough. 

  1. Law enforcement costs:  Two kinds: (1) direct (for more police officers); (2) indirect (lost opportunities to catch crime elsewhere).  But there are also law enforcement benefits, primarily consisting of increasing the expected sanction for the transport of drugs and guns from one part of the city to another.  This should help the City to attack pan-neighborhood gangs.
  2. Civil Liberties: Again, several kinds: (1) constitutionally protected privacy (which Dan says he isn't going to talk about);  (2) constitutionally unprotected privacy (i.e.: the "private public space" so essential to living in a big city); and (3) equal protection.  The last of these really concerns me.  The privacy losses created by these searches will fall most heavily on poor and working class new yorkers, who, unlike my former law firm colleagues, will be unable to opt out of the system by regularly taking cabs/town cars instead of the subway.  This is a serious problem.  But, if it is true that searches deter attacks - and there is at least some reason to believe they will - maybe working class and poor New Yorkers will be willing to make that tradeoff? (One interesting data point is that the plan has taken flak from democrats running to oust Mayor Bloomberg for not going far enough!)

Current Affairs | Bookmark

TrackBacks (0)

TrackBack URL for this entry:

Links to weblogs that reference Deterrence and Subway Searches:

Recent Comments
Popular Threads
Search The Glom
The Glom on Twitter
Archives by Topic
Archives by Date
September 2016
Sun Mon Tue Wed Thu Fri Sat
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30  
Miscellaneous Links