International Organization

Research Article

Rage Against the Machines: Explaining Outcomes in Counterinsurgency Wars

Jason Lyalla1 and Isaiah Wilson IIIa2

a1 Princeton University, E-mail: [email protected]

a2 United States Military Academy at West Point, E-mail: [email protected].


During the nineteenth century, states routinely defeated insurgent foes. Over the twentieth century, however, this pattern reversed itself, with states increasingly less likely to defeat insurgents or avoid meeting at least some of their demands. What accounts for this pattern of outcomes in counterinsurgency (COIN) wars? We argue that increasing mechanization within state militaries after World War I is primarily responsible for this shift. Unlike their nineteenth-century predecessors, modern militaries possess force structures that inhibit information collection among local populations. This not only complicates the process of sifting insurgents from noncombatants but increases the difficulty of selectively applying rewards and punishment among the fence-sitting population. Modern militaries may therefore inadvertently fuel, rather than deter, insurgencies. We test this argument with a new data set of 286 insurgencies (1800–2005) and a paired comparison of two U.S. Army divisions in Iraq (2003–2004). We find that higher levels of mechanization, along with external support for insurgents and the counterinsurgent's status as an occupier, are associated with an increased probability of state defeat. By contrast, we find only partial support for conventional power- and regime-based explanations, and no support for the view that rough terrain favors insurgent success.


Author names are alphabetical. We thank Steven Biddle, Max Boot, Chris Butler, Robert Cassidy, Audrey Cronin, Rozlyn Engel, Stacy Haldi, Karen Long Jusko, Stathis Kalyvas, Catherine Kelleher, Andrew F. Krepinevich, David Lyle, Michael Meese, John Nagl, Judith Reppy, Thomas Ricks, Elizabeth Saunders, Todd Sechser, Norton Schwartz, Don Snider, Shankar Vedantam, Elizabeth Wood, Steven Zaloga, and Ivan Arreguín-Toft, along with three anonymous reviewers and the editors of IO, for helpful comments and conversations. We also thank the soldiers and marines who agreed to be interviewed. Sara Evans, Raymond Hicks, and Samantha Lomeli provided excellent research assistance. We are also grateful for critical feedback on earlier versions from seminar participants at Columbia University, the U.S. Military Academy at West Point, Princeton University, the Defense Science Board, U.S. Transportation Command (USTRANSCOM), the University of Maryland, the Sante Fe Institute, and the 2006 Annual Meeting of the American Political Science Association. Data and supplemental analyses can be found at xs27E8 The views expressed in this article do not reflect the official policy or opinion of the U.S. Army, Department of Defense, or the U.S. Government.