Last night I took part in an interesting panel discussion at the Asia Society in New York examining lessons from Vietnam that might apply to the current war in Afghanistan. (The video is here.) One of my fellow panelists, Gordon Goldstein, author of Lessons in Disaster: McGeorge Bundy and the Path to War in Vietnam, a book widely read in the Obama White House during the deliberations over Afghanistan last fall, claimed in the course of our discussion that counterinsurgencies are only successful 25 percent of the time. He was challenged on this claim by the third panelist, Rufus Phillips — an old counterinsurgency hand who first went to Vietnam in 1954 with the legendary Edward Lansdale. Goldstein replied that he had got this figure from the journal International Security.
Today my colleague Rick Bennet searched that journal’s database and could find no such figure. The closest he could find was this recent article: “The Myth of Military Myopia,” by Jonathan D. Caverley. In it, Caverley, citing another International Affairs article by Alexander Downes, claims that democracies win 47 percent of counterinsurgencies and non-democracies win 58 percent.
That’s broadly in line with this article, which ran last August in Foreign Policy. Based on a sample of 66 20th-century counterinsurgencies, the authors find that “militaries succeeded about 60 percent of the time” but that after 1945 “this rate dropped to 48 percent.” But the authors also found that adopting of a hearts-and-minds strategy — of the kind the U.S. is now employing in Afghanistan — had a much higher probability of success: 75 percent. They add, however, that “Our analysis indicates that all foreign states that shifted to a hearts-and-minds strategy after eight years of counterinsurgency ultimately failed to defeat the insurgents, a pattern that does not bode well for Afghanistan.”
I take all such quantitative analyses with a big grain of salt because I’ve looked at the data sets employed by political scientists and they are invariably skewed in scope and full of dubious assumptions. (Downes’s International Security article notes that one such data set claims that Britain and America initiated World War II in Europe and that the United States started the Vietnam War!) Generating such numbers often gives an aura of faux certainty to what are in fact highly subjective judgments. That said, even in this flawed data, I could find no support for the extremely gloomy claim that counterinsurgencies win only 25 percent of the time. The actual figure generated by social scientists is at least twice as high.