In much of the world, human development and democratic consolidation are threatened by the presence of armed, non-state groups. Important progress has been made in the study of armed conflict, especially civil war, but as our understanding has advanced, the phenomenon itself has receded (Blattman and Miguel 2010, 18). The end of the ColdWar thankfully brought a sharp decline in the prevalence of civil war (Kalyvas and Balcells 2010), but it also saw the rise of a new security threat: militarized drug war.

Over the last thirty years, sustained armed conflict involving large and well-equipped drug cartels has ravaged Latin America's three largest countries, spilled into Central America, and even threatened US border towns. Especially puzzling is that in Colombia, Mexico, and Brazil, cartels attacked not only one another but the state itself. In fact, in Colombia cartel–state conflict erupted while cartels were at peace with one another, and overshadowed not only subsequent inter-cartel warfare but ongoing guerrilla insurgencies as well. In Mexico, turf war accounts for the numerical bulk of violent events, leading a flurry of recent scholarship to focus almost exclusively on the causes and consequence of violence among cartels. But cartel–state conflict has an outsized impact on society, directly challenging state authority and the rule of law. It can begin as an unexpected response to state crackdowns intended to restore state authority, and its duration can make a mockery of leaders’ original state-building ambitions.

As our understanding of inter-cartel turf war slowly advances, cartel–state conflict remains a puzzle. This study represents a first analytic cut at the problem, and an attempt to systematize our broader thinking about militarized drug wars as a class of conflict. It suffers from many of the weaknesses endemic to the study of violence in general: poor and overly aggregated data (Kalyvas 2008), too few cases, and too many critical variables—including the prevalence of corruption and the extent of successful extortion—that are difficult or impossible to observe. In addition, Mexico and Brazil are ongoing affairs, with major developments of both substantive and analytic importance happening over the course of this study.