Arnold Kling  

Surveillance Costs and Benefits

Usually Look on the Bri... The Common Sense of Bayesianis...

In this essay, I argue that in the wake of the London bombings that screening passengers is still not cost effective.

I am pretty sure that any cost-benefit analysis of "equal-opportunity" screening would reach an adverse determination. Crude racial profiling would be just as bad, because terrorists would simply work around the profiles. The approach that makes the most sense to me is to search primarily on the basis of the risk characteristics of the individual, with proximity to a "potential target" only a secondary factor. Today, we do it the other way around, which is extremely cost-inefficient.

UPDATE: In my essay, I argued that bomb-detecting sensors seemed like a better idea than video cameras. I had no technical basis for this argument. But Peter Huber backs me up

Comments and Sharing

CATEGORIES: Cost-benefit Analysis

TRACKBACKS (1 to date)
TrackBack URL:
The author at Anomaly UK in a related article titled Policing Terrorism writes:
    Possibly we in Britain see counter-terrorism differently -- Schneier, like Arnold Kling, is thinking in terms of preventing a one-off attack like 9/11, while we naturally think in terms of winning an extended campaign, in which we take hits but use int... [Tracked on July 29, 2005 11:07 AM]
COMMENTS (3 to date)
James writes:

Arnold writes, "I think that the top security priority should be to set up a system to monitor the Department of Homeland Security. I am not kidding." And this sounds great. I can sympathize. But would it work? Historically, the overwhelming trend has been for governments to increase, not reduce their powers over their citizens. Why would this change now?

Todd FLemming writes:

I come from a slightly different place. I am a security practitioner with only an undergraduate degree in Economics. Please pardon my lack of experience in your discipline. I do however enjoy your blog. Here is my take; You are correct. Homegrown terrorism is a big threat and a big opportunity. Keep in mind the second worst terrorist event on US soil was completely homegrown and had nothing to do with Muslim extremism (that we know of). The covert nature of these activities necessitate that the general public does not understand the ongoing operations only the ultimate results be they success or failure. Our intelligence community must get effective faster than the emergence of the threat. This unfortunately is not happening.

Security cameras can no more be classified as inferior technology than say a mechanical lock on a door. All of these security devices have there appropriate applications and inappropriate applications. Cameras are highly efficient for assessment and verification but as you correctly indicate very costly and inefficient for surveillance. Unfortunately explosive sensor technology is not yet to the point where it can be deployed as you suggest. Even if it was it would be highly prone to nuisance alarms which would in turn require some authentication. The most expensive (or ineffective) technology is that which is deployed and not used due to the desensitization caused by constant false alarms.

Screening makes us all feel better but in actuality the marginal reduction in risk is very small. The current process of checking ID’s is notable here. All you need is what appears to be a state or federal government issued idea with your picture on it. The idea says nothing about you or if the credential is actually genuine. I agree random screening if it is truly random is also ineffective. Yet screening based on specific characteristics and profiles could be effective. Keep in mind, I don’t think they need to be racial for the reasons you indicate. For example someone with bulging clothes wearing a heavy coat in the summer and appearing to be nervous would be a good candidate. (I know what your thinking civil libertarians) That is for further screening not shooting. This would not only be costly due to the labor costs of the actual screening but also because of the training required.

As far as auditing Homeland Defense, that probably goes for all government agencies. All government spending should have some independent mechanism to test for “reasonableness” and economic efficiency. The question is how you do this without just creating another inneficent bureaucratic instrument. Unfortunately one tenet of quality is that you can't inspect quality into the system. You must have a process that is quality oriented. I would agree that some “kaizen like” or continuous improvement mechanism must be built into our security process.

I mean no disrespect to those who have lost their lives to terrorists but you as an economist should understand that we need to put the costs of saving human life in perspective. We lose substantially more people annual to cancer, heart disease and automobile accidents. Does it make sense for society to spend more resources protecting us from terrorism or should we allocate more resources to other areas in which far more lives are lost?

Wild Pegasus writes:

Crude racial profiling would be just as bad, because terrorists would simply work around the profiles.

I don't think the terrorists are recruiting a lot of white people for suicide bombings.

- Josh

Comments for this entry have been closed
Return to top