Sometime around the end of 2009, in the wake of Abdulmutallab’s failed attempt to bring down Northwest flight 253, I was having a conversation about the incident. My fellow converser expressed hesitation to declare the incident a security breach. Her hesitation was rooted in the fact that the attempt was unsuccessful. To put her hesitation into my own words, how can we conclude that a terrorist attempt that failed nonetheless had a high probability of success? And without fully understanding Abdulmutallab’s probability of success, how can we declare that a security breach occurred?
On December 25th, 2009, Umar Farouk Abdulmutallab boarded Northwest flight 253 from Amsterdam, Netherlands to Detroit, Michigan. Abdulmutallab had a six inch packet of white power sewn into his underwear. ((http://af.reuters.com/article/worldNews/idAFLDE5BP03M20091226)) A criminal complaint later filed against Abdulmutallab would identify the substance as PETN. ((http://big.assets.huffingtonpost.com/AbdumutallabCharges.pdf)) Specifically, he carried 80 grams of PETN. ((http://www.washingtonpost.com/wp-dyn/content/article/2009/12/28/AR2009122800582.html?hpid=topnews)) There is every indication that his intent was to detonate an explosive device on the airplane. However, what he actually succeeded in doing was described by passengers as something closer to setting off firecrackers.
The event has given rise to a variety of news stories and discussions exploring potential security, communication and intelligence failures. ((For the sake of simplicity I will refer to all measures designed to protect airplanes from attack, including airport security, communication and intelligence, collectively as “security”)) ((Intelligence failures have largely been cited as responsible for the Flight 253 incident)) In discussing the incident, President Obama said “[t]he system has failed in a potentially disastrous way.” ((http://www.cnn.com/2010/POLITICS/01/05/obama.terror.meeting/index.html)) But in order to discuss the nature of the failures, isn’t it important to try to understand the nature of the event that the alleged failure permitted?
How Dangerous was Abdulmutallab’s Attempt?
I did some Googling to try to find out how much damage the 80 grams of PETN that Umar Farouk Abdulmutallab was carrying could inflict. More specifically, I wanted to know if 80 grams of PETN could bring down an Airbus A330-323E, the aircraft type of Northwest Flight 253. I couldn’t find a precise answer, but I did find a story of a government demonstration showing 50 grams of PETN blowing a hole in the side of an airplane. ((http://abcnews.go.com/Blotter/northwest-airlines-bomb-photos/story?id=9436297&page=1)) But the demonstration offers no explanation of whether the resulting hull breach would cause the aircraft to crash. ((A Quantas 747 landed safely after losing a piece of its fuselage the size of a car http://www.nytimes.com/2008/07/26/world/asia/26qantas.html)) Further, there is some dispute over the veracity of the cited tests. ((http://willyloman.wordpress.com/2009/12/29/abcs-50-gms-of-petn-video-is-propaganda-and-i-have-the-proof/)) I don’t have the expertise to credibly suggest what kind of damage this quantity of PETN could inflict on the 0.024 inch thick aluminum fuselage of the Airbus 330 ((“On an Airbus A330, the fuselage is a 0.042-inch aluminum sheet. The aircraft skin itself, including the plastic window, wiring, and paint, is 6 or 7 inches.” http://www.slate.com/id/2240155/)) much less conclude either that the explosion definitely would have resulted in a crash or definitely would not have. Nonetheless, while I’m not an explosives expert, 80 grams of PETN seems like it could be a dangerous amount of the material.
Was Flight 253’s Fate Determined by Luck or Security Measures?
I think the questions this begs is (1) was the fact that Abdulmutallab’s attempt failed a result of outlandish luck or (2) did the various security measures in effect force Abdulmutallab into a particular method that significantly reduced his likelihood of success? After the incident, it seems that many assumed that Abdulmutallab’s failure was a result of pure bad luck (or good luck, depending on the perspective). But is it possible that Abdulmutallab was never likely to succeed in the first place?
Did Abdulmutallab’s Attempt Violate our Risk Tolerance?
One natural response to this is that we don’t care precisely what Abdulmutallab’s likelihood of success was. Our risk tolerance for exploding airplanes ((In this context, I’m referring broadly to terrorist acts that ultimately result in an airplane crash and substantiall loss of life.)) is zero. Thus Abdulmutallab’s perceived probably of success only needs to be greater than zero to violate our risk tolerance. The mere possibility of an airplane being brought down by a terrorist act is beyond our risk tolerance. These sentiments may represent our best wishes, but are they realistic? While we may wish for zero airplane incidents, we recognize that we have limited resources to allocate across competing objectives. Thus, while we may wish for perfect air travel safety, we recognize that the benefits of perfect air travel safety do not justify the cost of assuring such safety. We could eliminate all air travel deaths by entirely eliminating air travel. But giving up air travel represents too high of a cost for the marginal increase in safety. We could get very close to achieving 100% success at the prevention of bombs on airplanes by enforcing extremely costly and invasive security measure like full body cavity searches of all passengers, requiring all passengers to disrobe and wear TSA supplied smocks and flying all baggage and clothing in a separate airplane with no passengers on it. Or we could allocate 50% or more of our federal budget to intelligence gathering and attack intervention. But again, the costs of these measures are too high for the marginal additional safety they afford. We wish that all of our activities could be perfectly safe, but in reality we make trade offs and accept what we believe is a reasonable amount of risk. This is true in our homes, cars, hospitals and in airplanes. We understand that air travel safety is a balance between the cost of security measures and the benefit of the additional safety they provide. And thus, we are willing to endure some reasonable amount of risk in air travel in the spirit of controlling the costs of security.
That being said, I’m not sure how to precisely quantify our risk tolerance for bombs on airplanes. Our tolerance for exploding airplanes is not precisely zero. Rather it seems to me that it is something very very close to zero. If our tolerance for exploding airplanes is non-zero then a single fizzled attempt, such as Abdulmutallab’s, does not represent a violation of our risk tolerance simply because there was some unknown probability that he could have succeeded. Rather, that attempt must actually represent a quantity of risk that rises above some threshold. More specifically, the risk of explosion represented by the attempt, along with a variety of other factors, must indicate that our security measures are not, in the long run, affording the level of protection that we demand.
The thought of 80 grams of PETN on an airplane is scary. But does it represent a breach of security that violates our risk tolerance? ((We demand automotive safety, but not every car accident causes us to question automotive safety. But automobile incidents are frequent and airplane incidents are rare. Automotive incidents rarely make the news and airplane incidents are front page stories. Automobile accidents kill people in small increments but with high frequency, while airplane accidents kill people in large increments but with low frequency. Automotive accidents are something that people are familiar with. Airplane accidents are violent explosive events caused by factors that are rather unfamiliar to the laity. The rarity of airplane incidents means that each time one occurs, we find the data that it provides us to be of interest. It provides us with an opportunity to evaluate our procedures. We can look at why the incident occurred and determine whether it reflects upon a failure in our security measures, or is an acceptable infraction. Or perhaps the incident leads us to re-evaluate our balance of the cost and benefit of additional security. But also, because of the above factors, airplane incidents stir emotion and that emotion, if not controlled, can lead to irrational decisions.)) It seems part of the answer lies in what Abdulmutallab’s probably of success was. But since he was indeed unsuccessful, how does one conclude that his attempt nonetheless carried some given probability of success? Notwithstanding that he was able to get 80 grams of PETN onto the aircraft, security measures forced him to use certain methods such as carrying a certain small quantity in his underpants and using a particular method of detonation. Did his specific method significantly reduce his chances of success? And if Abdulmutallab was forced into those methods by security measures, could the security measures be said to have effectively reduced his chance of success? And if so, is it possible that the risk was diminished to a level that is within the bounds of what we consider acceptable? Stated differently, does the mere fact that 80 grams of PETN was brought onto an airplane represent a gross security breach or is it necessary to understand the degree of danger these circumstances presented to the flight before quantifying the magnitude of the security breach? As a lay person my gut reaction is that the mere presence on an aircraft of a substance that Wikipedia refers to as “one of the most powerful high explosives known.” ((http://en.wikipedia.org/wiki/PETN)) scares me and makes me feel like a terrorist attempt thwarted security and sheer luck saved the lives of the passengers of Northwest Flight 253. But does this kind of conclusive reaction lead to progressive thought and conversation about our air travel security policies?
In SuperFreakonomics, authors Steven Levitt and Stephen Dubner suggest:
What makes terrorism particularly maddening is that killing isn’t even the main point. Rather, it is a means by which to scare the pants off the living and fracture their normal lives. Terrorism is therefore devilishly efficient, exerting far more leverage than an equal amount of non-terrorist violence. ((http://www.scribd.com/doc/22949242/Superfreakonomics))
Based on that description Abdulmutallab’s attempt may have been a success.