In February 1991, Saddam Hussein’s military began a chaotic withdrawal from Kuwait, only to find it faced a new conflict: widespread domestic unrest. At the uprising’s height, 14 of Iraq’s 18 provinces were under opposition control. To many observers’ surprise, however, by April the state regained authority throughout the country. Hussein’s ability to re-stabilize his regime and hold onto power for more than a decade after had dire consequences. Citizens’ attempt to rewrite the social contract in Iraq failed and over the course of the 1990s Iraqis suffered in human and economic terms under a corrupt and authoritarian government. Hussein’s regime also threatened security beyond Iraq’s borders. In the years following the uprising the international community attempted to coerce Iraq—via diplomacy and kinetic military operations—to stop stockpiling chemical and more deadly weapons. The 1991 Iraq uprising clearly highlights how citizen aspirations and international security hinge on questions of regime change. Scholars and policymakers thus require a sophisticated understanding of the mechanics of regime change and specifically its primary driver: the armed forces’ defection. What explains variation in soldiers’ compliance with orders to suppress domestic unrest? Amid the Iraqi intifada, of Hussein’s million-man army, an estimated 200-300,000 soldiers defected. Many Kurdish units in the North defected; a number of Shi’a units in the South defected; and the majority of Sunni units remained loyal. My research is motivated by the inability of existing arguments to explain these puzzling loyalty and defection patterns in Iraq. In this paper, I draw on diverse primary sources—including 8 months of field research based in Jordan, 3 months of archival research, and Arabic-language military memoirs—to explain variation in soldiers’ responses to the 1991 Iraqi uprising. I argue that Iraqi soldiers faced significant obstacles to defection, and whether and how these costs were overcome depended on a soldier’s position in the military hierarchy. Specifically, the causes of soldier rebellion varied according to rank because a commander’s interests and constraints differed substantially from those of his subordinates. The paper’s theory explains why commanders consistently repressed demonstrations, and argues that subordinate soldiers fled their posts and even fomented unrest when unit control weakened and personal safety deteriorated. These findings shed light on an under explored yet critical event in Iraqi’s contemporary history. In addition, this research aids practitioners in weighing conflict intervention options and designing meaningful security sector reform in post-conflict settings.