Is Airport Security Killing 500 People a Year?

So asserts cryptographer and security critic Bruce Schneier in a recent smack at the costly security theater inflicted on us at airports. "Security theater" is actually too kind a term; it makes light of a heavy burden that includes not only enormous cost and trouble with little benefit, but the erosion of dignity, civility, personal liberty, and a healthy balance between government power and individual rights. (In disclosure I must say I feel this more keenly every time I fly.) Schneiier recently debated former Transportation Safety Authority administrator Kip Hawley on the issue, and finds Hawley's argument unconvincing:

[Hawley] wants us to trust that a 400-ml bottle of liquid is dangerous, but transferring it to four 100-ml bottles magically makes it safe. He wants us to trust that the butter knives given to first-class passengers are nevertheless too dangerous to be taken through a security checkpoint. He wants us to trust the no-fly list: 21,000 people so dangerous they’re not allowed to fly, yet so innocent they can’t be arrested. He wants us to trust that the deployment of expensive full-body scanners has nothing to do with the fact that the former secretary of homeland security, Michael Chertoff, lobbiesfor one of the companies that makes them. He wants us to trust that there’s a reason to confiscate a cupcake (Las Vegas), a 3-inch plastic toy gun (London Gatwick), a purse with an embroidered gun on it (Norfolk, VA), a T-shirt with a picture of a gun on it (London Heathrow) and a plastic lightsaber that’s really a flashlight with a long cone on top (Dallas/Fort Worth).

Additionally, there’s actual physical harm: the radiation from full-body scanners still not publicly tested for safety; and the mental harm suffered by both abuse survivors and children: the things screeners tell them as they touch their bodies are uncomfortably similar to what child molesters say.

In 2004, the average extra waiting time due to TSA procedures was 19.5 minutes per person. That’s a total economic loss—in –America—of $10 billion per year, more than the TSA’s entire budget.

So it's ridiculous, insulting, stupid, dehumanizing, a sop to friends of those in power, an erosion of civil liberties, and a huge waste of time and money. But is it actually killing us?

Schneier says it is, by virtue of traffic deaths; he citeswork by security analysts John Mueller and Mark Stewart in their book Terror, Security, and Money. Here's the passage:

The increased automobile deaths due to people deciding to drive instead of fly is 500 per year. Both of these numbers are for America only, and by themselves demonstrate that post-9/11 airport security has done more harm than good.

He doesn't say how they got that figure. I'm assuming it's from taking (hopefully accurate) estimates of how many miles are being driven by those who shirk the security lines for their cars, then deriving 500 deaths from the average number of deaths per 100,000 miles.*

If they're right, though, then the inefficiencies of post-911 airport security, by putting more people on the road, have already killed more people right here in the U.S. than the 9/11 attacks did. In any case, if these measures do put people on the road, a sane, rational policy (I know that's asking a lot) would take that into account.

via Schneier on Security: Harms of Post-9/11 Airline Security.

Note added 4/6/12: *The National Highway Transportation Safety Administration says there are roughly 1.14 deaths per 100,000,000 traffic miles; thus it would require about 43,900,000 traffic miles to produce 500 deaths. That sounds like a lot of miles; it would be about 14% of the roughly 305 billion miles Americans drive each year. The NHTSA data I found does show an increase in miles driven after 2001, but that increase seems roughly in line with prior annual increases in traffic. Then again, I believe those post-2001 surges took place amid massive drops in miles flown and despite hikes in fuel prices, which usually decrease driving. Doubtless other variables come into play as well, and I'm not sure how Mueller and Stewart parsed them. I've written Mueller and Schneier asking how those figures were derived, and will post that here if and when I get it.

__4/6/12, 1:23pm EDT: __John Mueller, author of the book from which Schneier cited the 500 deaths per year figure, kindly and quickly responded to an email request for the source of that information. It comes from a paper by Cornell University researchers Garrick Blalock, Vrinda Kadiyali, and Daniel H. Simon, "The Impact of Post-9/11 Airport Security Measures on the Demand for Air Travel," published in The Journal of Law and Economics in November 2007. PDF here. Here's the relevant passage:

If the inconvenience of security deters travelers from flying, some of them might choose to travel by automobile instead. As is consistent with our finding that the negative effect of baggage screening is greatest for trips of less than 500 miles, we expect this substitution would be especially likely on short trips, for which driving is most feasible. Because air transportation is safer than road transportation, the increase in driving could lead to more traveler fatalities. In fact, we have shown that the substitution of road for air transportation following 9/11 led to an increase in driving-related fatalities (Blalock, Kadiyali, and Simon forthcoming). As part of that analysis, we estimated the reduced-form relationship between air passenger volume and driving-related deaths. Using fatalities in commercial vehicles to control for time trends, weather patterns, economic conditions, and unobserved highway conditions, we found that a decrease of 1 million enplanements leads to an increase of 15 driving-related fatalities. Applying that relationship to the estimated reduction in originating passenger volume due to baggage screening, we estimate that in the fourth quarter of 2002 approximately 129 individuals died in automobile accidents that resulted from travelers substituting driving for flying in response to the inconvenience associated with baggage screening. ⁿ¹⁶

Although both our revenue and fatalities estimates are very rough approximations, the numbers are of an order of magnitude that warrant attention. These costs must be weighed against the difficult-to-measure benefits of better security.

And here a related footnote:

n16 To calculate the number of fatalities, we multiply the estimated number of trips not taken because of baggage screening, 3.2 million, by the number of enplanements associated with each trip (2.7). This calculation gives us the total reduction in enplanements associated with baggage screening. Finally, we multiply this figure by the number of additional driving-related fatalities resulting from a decrease of 1 million enplanements, 15, to arrive at our figure of 129 additional driving-related fatalities.

The Mueller figure (500/year) that Schneier uses is extrapolated from the 129/quarter-year figure calculated for 2002. As Blalock et alia note, that figure is rough, but "of an order of magnitude that warrants attention."

I'm certainly convinced of that.