How to know if we’re winning in Afghanistan

By Peter Feaver

There is a revealing story in today’s New York Times on the Obama administration’s efforts to identify meaningful metrics in the Afghan war. It has a quote from your humble blogging servant on how difficult it was to identify useful metrics in the Iraq war. The quote is accurate and in context (David Sanger is a pro), but it may be useful to provide even more context. The challenge is finding metrics that are valid and reliable and accurate.

Valid metrics correspond to the true situation on the ground and the true prospects for the war; when those numbers are trending up, the war is really trending up, and vice-versa. An "invalid" metric in this sense might be U.S. combat deaths. That figure could be trending up because you are losing the war (as was happening in 2006), or it could be trending up because you are finally taking costly steps to reverse the trajectory (as was happening in 2007).

Reliable metrics are ones that cannot be artificially raised (or lowered) by our actions or enemy actions. An unreliable metric might be enemy suicide bombing attacks. If that is your primary gauge of how things are going in the war then you can be misled in the short run because the enemy could launch a desperate flurry and lead you to draw the wrong inferences about the situation on the ground. The classic example of this was the Tet offensive in Vietnam, which convinced the recently deceased Walter Cronkite that the Viet Cong could never be defeated when, in fact, Tet marked the end of the VC.

Accurate metrics are truthful. Even valid and reliable metrics are no good if they are inaccurate — that is, if the reporting is bogus. During the Vietnam War, the Pentagon inflated enemy body counts so that they became more a metric of Pentagon mismanagement of the war than a metric of progress on the battlefield.

But even that may not be enough. In the story, I mention that we thought the "rat rate," the number of good tips we got from Iraqis reporting on Al Qaeda or insurgent activity in their neighborhoods, was a valid, reliable, and accurate metric. I still think it is a good metric because it tapped into the heart of the pre-surge strategy: were Iraqis "standing up" in the fight so we could begin "standing down." But, by itself, it was obviously insufficient because it did trend positively in 2006 just as the war was trending negatively.

One final word: as I have argued before, establishing good metrics is only a part of a successful strategy. It must be accompanied by a good political-military strategy on the ground in Afghanistan, of course. And it must be accompanied by a sustained commitment of presidential political capital and political persuasion, engaging in the elite debate on the topic and explaining to the American people the president’s side of the case. We are seeing that expenditure of capital and persuasion on health care — in truth, an extraordinary commitment of Chicago-style campaign efforts — but nothing even remotely comparable on the war front. In coming months, I expect the administration will have to ramp up its efforts to explain its war policies. When it does so, I hope it has valid, reliable, and accurate metrics to offer.