Monday, July 11, 2005

previous entry | main | next entry | TrackBack (0)

Prospect theory and homeland security

In the wake of the London transport attacks and calls in the United States for protecting our infrastructure,, it is worth remembering one of the most important results from the work on prospect theory in economics is that human beings overestimate the likelihood of rare events actually occurring. One political implication of this fact is that governments will be asked to overinvest in measures designed to regulate and curb low-risk events.

In the wake of the London transport bombings, there has been a lot of chatter on television about what must be done to boost homeland security. However, prospect theory offers an important corrective to this natural response -- we exaggerate the cost of terrorist attacks on U.S. soil.

Keep this in mind when reading Benjamin Friedman's article in Foreign Policy on the myths and realities of homeland security. Here's how it opens:

The odds of dying in a terrorist attack are minuscule. According to the U.S. Centers for Disease Control, the odds are about 1 in 88,000. The odds of dying from falling off a ladder are 1 in 10,010. Even in 2001, automobile crashes killed 15 times more Americans than terrorism. Heart disease, cancer, and strokes are the leading causes of death in the United States—not terrorism.

People overestimate risks they can picture and ignore those they cannot. Government warnings and 24–hour news networks make certain dangers, from shark attacks to terrorism, seem more prevalent than they really are. As a result, the United States squanders billions of dollars annually protecting states and locations that face no significant threat of terrorism. In 2003, Tulsa, Oklahoma, received $725,000 in port security funds. More than $4 million in 2005 federal antiterror funding will go to the Northern Mariana Islands. In 2003, Grand Forks County, North Dakota, received $1.5 million in federal funds to purchase trailers equipped to respond to nuclear attacks and more biochemical suits than it has police officers.

These small expenses add up. Federal spending on first responders grew from $616 million in 2001 to $3.4 billion in 2005, a 500 percent increase. Homeland security spending will approach $50 billion this year, not including missile defense—roughly equal to estimates of China’s defense spending. Yet pundits call for more. A 2003 Council on Foreign Relations report hyperbolically titled, Emergency Responders: Drastically Underfunded, Dangerously Unprepared, recommmends increasing spending on emergency responders to $25 billion per year. To his credit, the new secretary of homeland security, Michael Chertoff, wants to trim the pork from the department’s budget. But efforts in congress to link funding with risk have failed largely because haphazard spending is consonant with the current U.S. strategy that tells all Americans to be afraid.

It’s true that al Qaeda’s attacks on Sept. 11, 2001, may be a harbinger of a more destructive future. But it is also true that parts of the war on terrorism are working. Tighter U.S. entry requirements, more aggressive European policing, the destruction of al Qaeda’s Afghan sanctuary, and refined intelligence operations have crippled al Qaeda’s ability to strike the United States. Most of al Qaeda’s original leadership is dead or in prison. Few other Islamist terrorists—even the most wanted terrorist in Iraq, Abu Musab al–Zarqawi—are as capable or organized as al Qaeda once was....

Most Americans are safe from terrorist attack. And the most likely forms of attack remain conventional. The fact is, all terrorist attempts to use chemical and biological weapons have failed to cause mass casualties. True, a successful biological weapons attack could kill hundreds of thousands of people. But manufacturing, controlling, and successfully dispersing these agents is difficult—probably too difficult for today’s terrorist groups. Synthesizing and handling chemical agents such as the deadly nerve agent VX, sarin, or mustard gas is complicated and extremely dangerous, often requiring access to sophisticated chemical laboratories. Most experts agree, for instance, that al Qaeda does not possess the technical capability necessary to produce VX. And even if terrorists procure and deploy chemical weapons, they are unlikely to kill many people. The 1995 sarin attack in Tokyo’s subway system was limited to only 12 deaths. Official U.S. government reports, including that of the Gilmore Commission, which examines domestic responses to terrorism, show that it would take one ton of chemical agent, favorable weather, and considerable time to kill thousands of people with chemical weapons.

Read the whole thing -- and then check out this 2003 primer on "Prospect Theory and its Applications for Disaster and Emergency Management."

posted by Dan on 07.11.05 at 05:37 PM


Some of the Homeland Security spending is purely pork -- it has nothing to do with actual security threats, or the public fear of terrorism. Rather, it's about the dividing up of public money in a manner politically acceptable to Congress. It's simply not true that money for "port security" is being sent to Tulsa because the American people have become unreasonably fearful of terrorist acts.

Also, people tend to be more accepting of certain risks than others. I would think that most of us can "picture" terrible automobile accidents, terminal diseases, and falls off ladders more willingly than we can picture terrorists attacks. The argument that the American public overemphasizes spending on homeland security is acceptable if we accept that one way of dying is the same as any other. But I think most of us have the sense that deaths from terrorism are worse than deaths from more typical causes. We call it a "war on terror" for a reason. If I die tonight by falling off a ladder, the rest of you won't be affected. If I die tonight because terrorists blew up my bus, well, that has implications for you too.

We should attack pork-barrel spending -- money to secure Tulsa's ports is obviously ridiculous -- but we shouldn't get to actuarial about this.

posted by: Andrew Steele on 07.11.05 at 05:37 PM [permalink]

At least in the excerpt, the author doesn't mention the possibility of a nuke attack of some kind. That would be far more devastating.

Keeping in mind the usual cavaets, see Al-Qaida nukes already in U.S

And, read up on the corrupt corporations that are profiting off our lax border security here.

If you want to have real homeland security - instead of the fake kind provided by Bush - take the actions described at the last link.

posted by: The Lonewacko Blog on 07.11.05 at 05:37 PM [permalink]

Its an interesting topic. I read something once describing how the human mind interprets various odds. Apparently we are very good at anything to 1 in 10, but beyond that it gets fuzzy. 1 in 100 is interpretted the same as 1 in 1,000,000. That explains the boneheads that hold me up in line at the gas station buying lotto tickets.

posted by: Mark Buehner on 07.11.05 at 05:37 PM [permalink]

This kind of probabilistic analysis is a serious abuse of statistical reasoning. Would the CDC conclude, at an early stage of an epidemic, that your chance of dying of disease X is only .0034%, because that is the probability to date? Of course not, because epidemics spread and the infected population increases as a % of the total population. Would the FAA conclude, given a problem of structural failure in airplanes, that the probability of a particular structural member failing is only .002% per flight, because that's been the experience to date? Not if all the failures occurred at around 6000 hours of flight time, and most of the planes in the fleet have not yet reached that number of hours.

The spread of terrorism may well be similar to these cases. You can't use static probabilities without evidence that the probabilities will, in fact, remain static.

More here:

posted by: David Foster on 07.11.05 at 05:37 PM [permalink]

This is one of those arguments that sounds smart but is actually beside the point. What we have learned from Al-Qaeda is that they feel emboldened when we do not respond to terror. WTC, Kenya-Tanzania, Cole etc. all convinced them that the U.S. was a great passive target.

We also learned from Al Qaeda that they know how to take advantage of even small gaps in our security to tremendous effect (box cutters and airplanes). According to the logic of the guy in the Foreign Policy article we should have some perspective and see that only 3,000 people out of 290 million died on September 11th. However, that attack fundamentally altered the world geo-politically and has cost the world entire in ways that cannot even be measured. This is about an anti-capitalist ideology that is going to gain steam over the course of the next 30 years and their most effective weapon will always be terror since they don't produce anything.

This guy who is telling us that we are overreacting to terrorism would probably be the first to criticize Bush for the routine terrorist bombings in Iraq that are "destabilizing" that country despite the fact that most Iraqis are not likely to be affected.

We have many Islamic states that are fragile or in jeopardy of being taken over by militants, and a whole generation (the youth bulge) in the Middle East that most likely will not find employement but will find comfort in Al Qaeda's fantasy ideology and join the fascist militant cause.

How destabilizing would it be to our global economies if a number of Islamic groups in France, Britian and other countries routinely carried out attacks like the ones in London. The costs would be very high--not just in the First World but the Third World. It's vital that the West, and particularly the United States continue creating technology that lessens the threat of all forms of terrorism.

Perhaps the odds of being killed by terrorists are incredibly low, but the odds of our current era of globalization getting destabilized if our governments do not take seriously the regular threats and plots (how many have been foiled in Europe now?) aimed at us by fascist groups residing within our borders would be very high indeed.

I'll bet this guy thinks the US is overreacting to the terrorist threat but at the same time also believes that US forces are getting absolutely annihilated in Iraq--another Vietnam.

posted by: Patrick on 07.11.05 at 05:37 PM [permalink]

Mark, join the 21st century: pay at the pump.

posted by: rastajenk on 07.11.05 at 05:37 PM [permalink]

Two points:

1. Of course it is true that people over-estimate the probability of rare events. This is precisely why many people are more afraid of flying than driving, even though it is well established that flying is far safer. We can estimate the probability of a airline crash during the coming year because the technology and flying patterns have not changed from last year. In other words, the population from which the random data are drawn is not changing appreciably.

On the other hand, we do not know the probability of a terrorist attack during the coming year because the factors that control the randomness are changing in unpredictable ways. This causes people to instinctively err on the side of overestimating even further.

2. It is not irrational to expect the government to do something about the risk of a terrorist attack, while doing nothing about those dangerous ladders that Andrew cited. That is because there is not much the government can or should do about the "ladder threat" short of mandating design changes that would make ladders unpractical as a household tool. It's a matter of bang for the buck.

Are we getting a good bang for the buck in terms of threat mitigation? No. But the reason has little to do with peoples' overestimating the threat and everything to do with political power struggles. Our leaders are interested in APPEARING to do a better job of protecting us than their political rivals, rather than actually using our tax money effectively to optimize the threat mitigation from various sources.

posted by: Larry on 07.11.05 at 05:37 PM [permalink]

"Mark, join the 21st century: pay at the pump."

Lol. I do, but I have to admit my junkfood intake is gas station driven, and they havent found a way to deliver pringles to the pump on demand yet. There's nothing worse than trying to get your salt fix when some numbskull is reciting supposedly lucky numbers to a register jockey. On more than one occasion i've interjected my unsolicited advice that the odds of winning the lotto if you dont play are only marginally worse than if you do. That statement is universally received with blank stares, at which point i open my pringles, to hell with gas station edicate.

posted by: Mark Buehner on 07.11.05 at 05:37 PM [permalink]

Mark, Why are you trying to discourage people from buying lottery tickets? I love the lottery because it's a tax on people who don't understand math. The more tickets they buy, the less likely the state is to increase other taxes.

posted by: Larry on 07.11.05 at 05:37 PM [permalink]

See the aforementioned pringles dilemna. Abstractly i could care less that people play lotto, but in practice it impacts my life in a negative way. And I have no faith that the government wont raise my taxes one way or another. I think there is a tax threshold that the citizenry is willing to bear and the tax code always approaches that limit. Spending is directly related to that limit, not whatever budget fantasy the bureacracies claim are necessary to meet the crisis de jure. I dont doubt that if our taxes were 30% higher or lower spending would match that level (a bit higher actually) with no noticeable change in services either way.

posted by: Mark Buehner on 07.11.05 at 05:37 PM [permalink]

There is another factor here. Attacks by people increase if not stopped. This makes them more fearsome. Even though more firemen did in the line-of-duty than policemen, people think of police as being at greater risk. Malevolent risks are always seen as more worthy of attention than impersonal risks.

posted by: levi from queens on 07.11.05 at 05:37 PM [permalink]

"The odds of dying from falling off a ladder are 1 in 10,010."
So, in a town of 100,000, there are 10 people a year going to die falling off a ladder?
Hmmm, better check those figures...

posted by: steve (the anti-math) on 07.11.05 at 05:37 PM [permalink]

Post a Comment:


Email Address:



Remember your info?