The definition of terrorism that I have always favored is quite different from those employed by most government or  academic sources. Terrorism, as I see it, is the use of terror as a force-multiplier to have an effect (political, economic, or, strategic) that is beyond the capacity of the terrorizing entity to achieve through the application of  conventional elements of power (i.e. political, economic, military, or cultural) under their disposal. By this definition it seems that terrorism may be having some measure of success.

There is an ongoing debate about how successful a strategy terrorism is. There are those whose views are like that of Alan Dershowitz who famously wrote a book entitled “Why Terrorism Works”, and others have written about why it is usually a failed strategy (e.g. Caleb Carr.) Such varying views on something as clear cut a success and failure means that differing definitions of success and failure, and perhaps of what constitutes terrorism, are likely in play.

Osama bin Laden has made calls for attacks on the US economy. His rationale apparently being that if the US were faced with diminished economic resources, American policymakers would be incapable of carrying out “foreign adventures” in the [predominantly] Muslim world. So far these calls may not seem to have been heeded in that there are many ways to cause economic damage that are not particularly high risk and of which terrorists have not taken advantage. However, we are spending enormous amounts in the fight against terrorism both at home and abroad. The Department of Homeland Defense budget alone was almost $70 billion in 2006 (if DHS were a country it would be slightly richer than Ecuador.) This does not count the half-a-trillion Defense Department budget that would make it the 27th largest country in the world if its budget were equated to GDP at purchasing power parity (an economy about the size of the Netherlands.)

The question that arises is, what exactly is the opportunity cost? Opportunity cost is the value of the most highly valued alternative use of the resources. Could we save more lives by putting our resources elsewhere? Do we value the feeling of security gained through such expenditures at an sensible rate or not?

This is, of course, a tricky question in that it is difficult to know what would happen if we reduced these expenditures. While our defense spending could be more wisely distributed, it may not be unreasonable to expect that without substantial military spending many countries might be encouraged to act up in ways detrimental to US interests. (I am not certain that this is true. After all, the entire world knows the US is currently stretched to capacity, and it doesn’t seem there are great efforts to rob the candy store while the constable is preoccupied.) Would there be a rising tide of terrorism if we diverted some of the resources we use to counter the threat of terrorism? On the other hand, we do know that there are vulnerabilities that exist day in and day out that are not taken advantage of by terrorist groups for whatever reason, though they could be at relatively little cost, and, in some cases, with relatively little expertise required.

The question that forms the title of this article is largely rhetorical. I don’t know, and I’m not confident that useful metrics exist to conduct policy analysis on the basis of cost-effectiveness, cost-utility, cost-benefit studies. A number of difficulties exist in trying to capture the benefit of counter-terrorism spending?

First, terrorist events are exceedingly rare by nature (at least in most of the world.) We have all heard success claimed on the basis of the fact that there have been no major terrorist events on US soil since 9-11. There is a strong and a weak version of this argument. The stronger version is to point to foiled plots in order to show that there is some return on investment. The weak version is seen in statements of the nature of: ”Everybody expected more events to occur after 9-11, and the fact that we have not seen them indicates progress.” Given that the time interval between major terrorist attacks in the US has been measured in years, there is no reason to have expected a wave of attacks. One thing we know in behavioral social sciences is that people systematically overestimate the probability of an event happening in the wake of a similar event.

How does one analyze the likelihood of such rare events in a manner that would allow one to calculate meaningful expected value of loss due to an attack? In particular, if a given class of events has not occured yet, one might give it a probability of zero; and, therefore, the expected value of loss would be nil. However, this could be an obvious and catastrophic error. One way to deal with events that have not occured is to calculate the number of days that that type of event could have occured but didn’t, add one, and then put one in the numerator. For example, if one has lived under vulnerability without attack for 2,268 days (say 9-11 defined a new epoch), then one could take 1/2,269 as a worst-case scenario, and could say the probability is less than or equal to that value. Of course, taking this worst-case scenario bound may greatly overestimate the risk. On the other hand, the fact that the probabilities will be tiny means that the we may radically underestimate the value of loss. Does terrorism lay outside the bounds of what can be dealt with by means of conventional policy analysis?   

Second, how do we get a sense of what the indirect effects of a terrorist attack might be? For example, there are papers that try to disaggregate the effects of an event like 9-11 such as reduced demand for air travel, New York City tourism, or increased demand for American flags. However, it seems that any attempts to accomplish this must be based on taking some forecast as a given, but we know that economic forecasts are usually not spot on. For example, the reduced demand for air travel is partly a result of terror attacks, partly a result of rising fuel (and thus ticket) prices, and partly a result of dissatisfaction with the diminishing size of peanut and pretzel bags (no doubt airlines are employing nanotechnology scientists to perfect an even smaller snack pack.) Of course, there are timing clues, but there are also lags and coincidences.

There are also many items that are difficult to value. Valuing life has always been troublesome. Using lost income as a measure doesn’t work well because it totally discounts the value of those who don’t work for financial compensation. Willingness-to-pay methods where one considers how much people are willing to pay to reduce threats to life and limb are troublesome both because one may not have a good idea of the risk attitudes of those questioned and because people may not have a good idea of the likelihood of loss the face. There is often an assumption of risk neutrality that is used just to make things manageable. Consider an individual who is willing to pay $5 to prevent a one in a million probability of being killed by a terrorist. If this individual were risk neutral we could say that they valued their life at $5,000,000. However, a risk adverse person values there life at some amount less that that, and a risk seeking person some amount more.

In conclusion, I don’t know how much we should be spending on counter-terrorism (and I have no reason to think anyone else does either.) However, it does seem likely that we are paying disproportionately in comparison to the risk, and I suspect there are alternative uses of those resources that would net more gain(traffic safety, disease research,  reduced tax burden etc.) In at least this sense, terrorism seems to be an effective strategy.