Targeted Killing: Drones

Empirical Analyses of Drones in Pakistan

By Jack Goldsmith
Tuesday, October 9, 2012, 5:29 AM

As a counterpoint to the much-noted critical study of drones in Pakistan from the clinics at NYU and Stanford, I recommend this piece by Joshua Foust of  the American Security Project, which is sympathetic to the NYU-Stanford study but maintains that it “has some serious bias issues.”  I also recommend this (seven-month old) article by Patrick B. Johnston and Anoop Sarbahi (of RAND and Stanford, respectively), which argues very tentatively that drone strikes in northwestern Pakistan reduce militant violence there.  The NYU-Stanford study covers many topics (ranging from civilian casualties to the impact of drones on medical assistance, education, social and cultural activities, and the like), and focuses almost exclusively on the undoubted costs of the drone program without much consideration of possible benefits.  The Johnston-Sarbahi study focuses very narrowly  on the relationship between drones strikes and militant violence in northwestern Pakistan.  The authors actually find a positive correlation between drones and violence in that region, but after “controlling for local factors and time trends,” they find a negative correlation, though only a “modest” one.  The takeaway for me from the Johnston-Sarbahi piece is the genuine difficulty of measuring and assessing the impact of drones on violence in Pakistan, much less their impact on violence in Afghanistan and their other second- and third-order effects.