The high cost of misdirected airport security (and why we no longer need sky marshals)
We often see problems where people fail to take any appropriate action to prevent bad things from happening. After an unfortunate event, we often see a flurry of activity, because there is a belief that we need to be doing something! Anything! Unfortunately, taking inappropriate action to solve a problem can be more destructive than taking no appropriate action, or even taking no action at all.
As an example, a consultant we knew was asked how to turn off telephoning to certain area codes at certain hours. The idea was to prevent employees wasting precious morning time calling home for idle chit-chat when they got in. It was explained to management that this could easily be done, but that instead of losing 10 minutes of time when the employee, who had left home somewhere between 6am and 7am, called home, you would lose a half hour when they went downstairs to a pay phone. The cost of the solution would exceed the benefit of the solution.
The same holds especially true after a major incident. As we have discussed many times, putting security measures into place without knowing the threat they are to deal with, the consequences of failing to meet the threat, and the costs of meeting – or failing to meet – the threat is generally counterproductive. This is not new: After the crash of TWA Flight 800 we saw activity suspiciously similar to what we are now seeing.
For those of you who do not remember 17 July 1996, 230 people died when Flight 800 fell into Long Island Sound. The NTSB final report said “The National Transportation Safety Board determines that the probable cause of the TWA Flight 800 accident was an explosion of the center wing fuel tank (CWT), resulting from ignition of the flammable fuel/air mixture in the tank. The source of ignition energy for the explosion could not be determined with certainty, but, of the sources evaluated by the investigation, the most likely was a short circuit outside of the CWT that allowed excessive voltage to enter it through electrical wiring associated with the fuel quantity indication system.” This prompted the inspection – and repair – of aircraft, with continuing mandated inspection to prevent it from happening in the future.
It also prompted a flurry of anti-terrorism legislation based on the Gore Commission report, the outcome of which was, in the words of then-President Clinton, that “as a result of these steps, not only will the American people feel safer, they will be safer.”
The Cost of Antiterrorist Rhetoric, by Robert W. Hahn from the Cato Institute (http://www.cato.org/pubs/regulation/reg19n4e.html), written after the TWA flight 800 crash points out some lessons that apply equally after the events of September 11th, starting with the fact that eliminating acts of terrorism on aircraft is fairly trivial: Simply ground all the aircraft.
We will not close down all our airlines, independent of how effective the measure would be, but the thought should prompt us to ask what else we should not be doing. In effect, how we respond to the view of one friend who noted that we can’t afford to have another Twin Towers incident, and so we have to do something.
Flight 800 crashed as the result of a specific technical problem, and the recurring inspection requirements will, in fact, prevent this specific accident from happening again.
However, new and unproven security measures, no real threat analysis, and unsubstantiated security claims will likely result in:
• Increasing the number of deaths
• Increasing cost and inconvenience
•Reducing civil liberties
Increasing the number of deaths
In some cases we can reduce deaths through reasoned actions. In other cases the best we can do is shift the deaths from one arena to another, often causing more deaths than we would have had absent any action.
Let us see how policy can shift and increase the number of people who die. When The Pill was introduced, there was a movement to get it banned because roughly two out of every 100,000 women taking it died of embolisms: If the pill were banned these women would live. On the other hand, none of the 100,000 women got pregnant, but if the pill were not available to them, and they used diaphragms, 5000 of them would get pregnant, and 5 would die from the pregnancy. In addition, because the pill required seeing a doctor, non-related but potentially-lethal problems may be discovered and dealt with, preventing deaths. The bottom line (we ignore here any theological issues) is that fewer women die when taking the pill.
The same phenomenon holds true in transportation. As air travel costs and delays increase, some people will choose to drive, even on long trips. More of the people who switch from flying to driving will die in automobile accidents than would have died from terrorist and accidental airplane deaths.
Increasing cost and inconvenience
Security, like other protective disciplines, is supposed to be about increasing productivity and profit through managing risk. This implies
1. Identifying threats and vulnerability
2. Calculating its probability
3. Gauging its impact
4. Reducing vulnerability, transferring the risk through insurance, or living with a known level of risk.
In some cases the probability is high but the total impact is low, so we ignore it: Employees taking pens, for example.
Airline terrorism has an extremely low probability of happening, and, putting aside the fluky World Trade Center collapse (which we are sure surprised the incident’s planners as much as it surprised those of watching it out our windows), the risk has been relatively (if you or someone you loved wasn’t directly involved) low: There were few bombings, and most hijackings ended relatively satisfactorily.
Hahn calculated that the cost per person saved after Flight 800’s antiterrorist measures were implemented was in the $200 million to $300 million range. Let’s say that he was off by an order of magnitude, and that the cost per life saved would only be $20 million to $30 million.
An expenditure this large should prompt us to as if it worth the investment and inconvenience for events this unlikely? It should also prompt us to ask if we can reduce airline terrorism further through technology, and whether current or planned approaches to airport security will work better than they did on September 11th? If adding cost and inconvenience will not improve safety, these measures should be removed or not implemented.
Scanning for bombs
Let’s look at scanning for bombs, where we can accept some false positives but no false negatives. Technological approaches currently give a false- positive rate of about 20%. We don’t know about the false negative rate because there are have been so few bombs in the history of aviation, but let’s assume that the devices were, collectively, 99.99% accurate. Is this useful? Not if you have as many flights as we have in this country, because:
1. The false positives will so overwhelm the inspectors that they will ignore most of them, so even the real positives will likely be passed.
2. We have so much baggage that, if people actually started bringing bombs on board, at 99.99% we would have false negatives.
3. Cargo isn’t checked, because, not having the technology or capacity to check it, we trust the shippers.
There are some additional significant problems:
• In reality we have no inclusive technology, and virtually nothing that is affordable, and absolutely nothing reliably usable by people who are poorly trained, paid minimum wage, exhausted from going to the other jobs they need to have in order to survive, and who are likely to quit in a few weeks.
• The cost is staggering. An x-ray CT system costs over a million dollars, and you need an awful lot of them to keep up the 6-seconds- per-bag flow needed to keep the system from bogging down. Then you also need to use sniffers to find those things not visible on x-ray, but these can be overwhelmed by strong smells and require reset time, so you also really need to use nuclear testers, which go for $10 million and up. We don’t know the price of a Nuclear Quadrupole Resonance device. We don’t have a price on the decompression equipment El Al uses to decompresses luggage before being loaded to make sure there is not an altitude sensitive detonator.
• Most significant of all, we can only test for the things we know we need to test for, and we can’t test for new and inventive things we don’t know we need to be testing for!
Other current security measures
The same holds true for other current airport security efforts: Even if rigidly enforced, most would have had little or no impact on past threats, clearly did not have much preventive effect on the events of 11 September 2001, and most likely will not have any impact on new threats, designed to be outside the security paradigm.
• Baggage matching has little value in deterring a suicide bomber.
• Keeping all but ticket-holders out of the terminal is inconvenient for loved ones and passengers, and serves no security purpose, since a bad guy wanting to get on the plane would likely just buy a ticket rather than trying to sneak on without being noticed.
• Banning small knives (even El Al didn’t ban small knives) serves no valid security function, and we traditionally carried one designed specifically to cut seat belts in an emergency. Keep in mind that in the September 11th incidents the problem was not the knives, but, rather the unexpectedly-outdated paradigm that led passengers to believe that the hijackers intended to negotiate for something, rather than kill themselves. And think of the embarrassment we will face in the unlikely event we ever somehow discover that the passengers on United Flight 93 overpowered their attackers using their Swiss Army Knives!
• Requiring passengers to produce identification has no security benefit: The 11 September criminals had carefully crafted identities and personas and, save for the fake pilot credentials, passed with no problem. While positive passenger identification has value to airlines because it eliminates resale of non-refundable, non-transferable tickets, which is not a security issue.
Policies tend to be unevenly enforced
• On a recent flight, before the ruling was clarified, our nail clippers and lighter were taken away, forcing us to replace them – and get a letter opener! – from the magazine shop inside the secured area!
The sad thing about fruitless measures is that the large sums of money wasted on unresponsive security measures could better serve society.
Reducing civil liberties
In addition to producing pointless physical security measures for the sake of doing something, incidents always cause loss of personal liberties, with the less-prudent willing to give up some liberty to keep “this” from happening again, while failing to realize that the while the liberty might be lost, future threat will not be reduced. We are seeing a renewed effort to make sure government can intercept all communications, with keys escrow systems again being proposed. Putting aside the usual significant issues of abuse and loss of keys, key escrow tacitly assumes that bad guys will give their keys to the government, not develop their own encryption, and that they will even use encryption at all: Usama Bin Laden stopped using all communications technology when he read that his whereabouts was being tracked by his SatCom usage.
Why we no longer need sky marshals
It has been estimated that a full cadre of American sky marshals would cost something in the neighborhood of $4.8 billion a year, a figure that would doubtless diminish as the pass-along of the cost reduced air travel. As it is, however, we no longer need sky marshals.
We went through a period when hijacking became more common, and there was some concern that one of the things one might hear onboard was “Take me to Cuba,” which induced us to put sky marshals on certain planes. This phase passed into history as hijacking destinations decided they didn’t want these people. Hijackings happened increasingly rarely, and when they did happen nobody was, in general, hurt.
On the surface it would appear that the events of September 11th, 2001, have changed this, and that there is once again the need for sky marshals. We believe this is not the case. Our reasoning is that on September 11th we saw two sets of incidents.
1. In the first set of incidents, the passengers were under the impression that this was another ho-hum hijacking, and sat back for the ride, expecting to land and be released, not bothering to fight back against the low-level threat of mat knives. This (that is to say, nothing) is exactly what we would have done if we had been on those flights.
2. In the second set of incidents, on United Flight 93, the passengers knew that their lives were at risk, as were the lives of others, and took appropriate action. We like to think this is exactly what we would have done if we had been on that flight. While, unfortunately, the passengers and crew died, their heroic actions foiled the plans of the hijackers.
At this time, there is no longer an expectation that hijackings will be benign. Because of this, one can reasonably expect that in any future such incident the passengers will fight back. Because of this, there is no longer a need to implement an unaffordable sky marshal program.
Other options have been suggested involving weapons on planes, some of which would serve to increase uncertainty in the minds of bad-guys as to whether there was someone armed on a plane, thus reducing the likelihood of on-plane attacks, and some of which are silly.
• Pilots are agitating to be allowed to carry guns as they used to (critics of this plan say they would prefer pilots to be locked in the cockpit and stick to the more-traditional tasks of avigate, navigate, communicate, which misses the point that the purpose of guns in the hands of pilots is to produce uncertainty in the minds of potential hijackers).
• Senator John McCain suggested that pilots carry stun guns, which overlooks the fact that stun guns must be held against the cooperating attacker – preferably in the stomach – for a period of up to 5 seconds before the device might – though more probably would not – work. (This editor was attacked in the not-too-distant past by three assailants: One to hold from behind in a bearhug, one to apply the stun gun, and one to get the goods. None of the attackers were injured in their unsuccessful attempt.)
• Others have suggested that cops be allowed to carry guns on planes as they used to (critics of this plan – all cops – point out that the average department has under four hours of firearms training a year, and that most cops don’t shoot all that well, which misses the point that the purpose of guns in the hands of unidentified cops is to produce uncertainty in the minds of potential hijackers).
• Still others have proposed that selected civilians be allowed to carry guns on planes as they used to (critics of this plan, ignoring the implications of United Flight 93, point out that this is a job for the police or military, not private citizens, which misses the point that the purpose of guns in the hands of unidentified civilians is to produce uncertainty in the minds of potential hijackers).
At the time this is written, we have been apparently been prudent and thoughtful in dealing with the external threat of terrorism in its larger sense, attempting to build a coalition to find some workable approach to dealing with the nature of terrorism, in an attempt to solve the problem without causing unnecessary deaths. We can only hope the same will apply to airline security, and that we will not face needless cost, needless inconvenience, needless loss of privacy, and needless deaths in an effort do demonstrate activity rather than get results.