The eternal "does terrorism work" debate

(Posted by John.)

Bruce Schneier (via Mike) links to this paper by Max Abrahms titled, simply, "Why Terrorism Does Not Work".  I haven't gotten all the way through the article yet (kind of a rushed weekend) but there are, I think, some serious objections one can raise to the paper's thesis.

1)  Abrahms writes that the success rate of terrorist organization is only 7%.  This leads me to ask, what's the success rate for modern conventional war?  Even the United States Army, one of the most successful military organizations, has failed to achieve the stated objectives of two of its wars in the last 50 years(Vietnam and Iraq, three if you count the dying war in Afghanistan.)  Most other militaries have been far less successful.  Between 1900-1950, didn't the Germans basically bat .000, despite being one of the most advanced and powerful armies on the planet?

2)  Abrahms says that the idea of terrorism as successful is basically an example of selection error -- people know about the few terrorist groups who succeeded, and not the much larger number who lost.  Again, we can point out the above, and say that the larger odds of failure certainly don't prevent national leaders from choosing war.  But it's also worth pointing out that sometimes, it's worth dropping the low performers from the sample.  Including the performance of, say, the French Army in the spring of 1940 tells you very little about the theoretical strengths of armored warfare.  Concentrating on the successes might bias your results, but it's also a useful analytical tool.

There's more to the paper, and much of it is extremely interesting, but as far as the issue of selection bias goes, I'll just note that Abrahms restricts his analysis only to those groups identified as foreign terrorist organizations by the State Department, noting in what can only be deadpan humour, "The only selection bias would come from the State Department."  Oh, that's all, huh?

Update:  I hadn't read my local newspaper this morning before posting this, but there's something rather relevant -- an article on the work of Patricia Sullivan, on why modern militaries fail to achieve their objectives.

It turns out that a major power is much more likely to fail when its war aim requires some sort of co-operation on the part of the adversary or the citizens on the ground, in order to change a despised foreign or domestic policy, for example, or quell sectarian violence, or prop up a regime that's on shaky ground....

A war aim that requires support of the people or cooperation by an adversary is called "coercive." In such a scenario - having Saddam Hussein co-operate with weapons inspectors, for example - the success rate was only 17 per cent. If the war aim is moderately coercive, such as propping up a government that replaced the one that just toppled, the rate was 40 per cent.

So for spending vastly less resources in terms of money or weaponry, terrorists chalk up a success rate of 7%, while modern industrialized nations have a similar success rate of 17% when it comes to coercive violence.  (Not exactly an apples to apples comparison, but not an unfair one I think.)

This is the single biggest problem with modern Republican Green Lanternism:  these people don't understand that using political violence is by definition a failure-prone endeavour, and clamoring for "more willpower" or yelling about "support the troops" doesn't get around that.