Institutions of the Offensive: Domestic Sources of Dispute Initiation in Authoritarian Regimes, 1950–1992 Brian Lai University of Iowa Dan Slater University of Chicago What are the most important sources of institutional variation among authoritarian regimes, and how do such institutions influence these dictatorships’ propensity to initiate military disputes? This article argues that most existing studies in both comparative politics and international relations employ a flawed conceptualization of authoritarian institutions. Excessive focus on the personalization or institutionalization of authoritarian regimes’ decision-making procedures has distracted attention from the more critical issue of what institutions these regimes deploy to enhance social control and secure political incumbency. Since military regimes are systematically less effective than single-party regimes at developing these types of authoritarian institutions, they more frequently resort to desperate measures to fend off domestic challenges to their power. In particular, we find compelling empirical support for our hypothesis that military regimes are more likely than single-party regimes to initiate military disputes, irrespective of whether those regimes are highly personalized or not. S tates fight wars, but governments in power make the fateful decision whether or not to start them. Warfare thus results not merely from what students of international relations have tended to see as the bench- mark interest of states: maximizing relative gains in a Hobbesian world (i.e., Waltz 1979). It also arises from what students of comparative politics typically argue is the benchmark interest of government officials: staying in power (i.e., Bates 1981). Since prospects for retaining power are influenced by domestic political institutions, we expect such institutions to exhibit a powerful effect on the conflict propensity of different types of governments. Unfortunately, we still know much too little about how or even which institutions make governments more or less likely to initiate military disputes. This problem is particularly acute in our study of authoritarian regimes. There is considerable scholarly consensus that dictator- ships are more likely than democracies to instigate wars; but why are some dictatorships more belligerent than oth- ers? Like many students of the democratic peace, we ar- Brian Lai is assistant professor of political science, University of Iowa, 341 Schaeffer Hall, Iowa City, IA 52242 (Brian-lai@uiowa.edu). Dan Slater is assistant professor of political science, University of Chicago, 5828 S. University Avenue, Chicago, IL 60637 (slater@uchicago.edu). Thanks to Emily Broeckling, Dan Reiter, Terry Chapman, Scott Gartner, Sara Mitchell, Kelly Kadera, David Clark, and the editors and anonymous reviewers at AJPS for their comments and assistance with this article. Previous versions were presented at the 2002 American Political Science Association Annual Meeting, the 2003 Midwest Political Science Association Meeting, and the 2003 Iowa Political Science Workshop. gue that variation in international aggression arises from variation in domestic political institutions. But what are the most important sources of institutional variation among authoritarian regimes? And how do such institu- tions influence these dictatorships’ propensity to initiate military disputes? We address these questions by applying a novel insti- tutional typology of authoritarian regimes to the question of militarized dispute initiation (Slater 2003). We argue that authoritarian institutions influence conflict propen- sity through their effect on regime legitimacy and gov- ernment tenure. The less legitimate the regime and the less secure the government in power, the more likely the political leadership will be to initiate military conflict. We argue further that the critical institutional fac- tor influencing an authoritarian regime’s legitimacy and security in office is whether it is ultimately backed by the military or by a ruling party. What matters in institu- tional terms is not how these regimes make decisions (per- sonalized vs. collective procedures), but how they enforce American Journal of Political Science, Vol. 50, No. 1, January 2006, Pp. 113–126 C 2006, Midwest Political Science Association ISSN 0092-5853 113