Leaders, accountability, and foreign policy in non-democracies

What explains differences in the foreign policy decisions of different types of authoritarian leaders? Are democratic leaders really different from all authoritarian leaders in their decisions to use military force? When do authoritarian leaders face domestic accountability for their international behavior? In this dissertation, I propose that when domestic actors can overcome their coordination dilemma to remove leaders, even autocratic leaders must make foreign policy decisions in the shadow of domestic punishment. I argue that previous research has underestimated the extent to which non-democratic leaders are domestically accountable for foreign policy. Even within regimes with small "winning coalitions," where the support of only a small group of individuals is required to stay in power, elites often have the means and incentives to coordinate to remove the ruler as long as the leader does not control access to high office and the leader has not tampered with military institutions, using them to both spy on potential rivals, and to reduce the likelihood of coups and other types of removal from office.I also introduce a rich new source of regime type data that allows me to distinguish my explanation from existing theories, building on and augmenting work by scholars in comparative politics. I describe how we can distinguish between two general types of authoritarian leaders. On the one hand, there are "constrained" authoritarians who do not personally control top government appointments, and have not tampered with normal military organization. On the other hand are "unconstrained" authoritarians who have seized personal control of civilian and military institutions, appointed cronies to both civilian and military positions, and taken related steps to insulate themselves from removal at the hands of rivals. Some "semi-constrained" leaders in between these two extremes, I find that most leaders of the leaders in my sample either both control civilian appointments and interfere with military institutions to insulate themselves from coups, or do neither.One chapter analyzes how war outcomes affect leader tenure. I find that constrained authoritarians are approximately as likely to lose office after defeat in war as democratic leaders. In sharp contrast, unconstrained authoritarians manage to hold on to power even in the face of serious military defeat. Another chapter deepens this analysis, and attempts to establish a causal (rather than merely correlational) relationship between authoritarian regime type and war outcomes by examining five historical cases in which a leader initiated a war that ultimately ended in defeat. Next, I build on these insights about the domestic consequences of military defeat to ask whether leaders' expectations about accountability also influence their decisions to get involved in military conflicts in the first place. I first assess how regime type affects patterns of victory and defeat in war. Counter to the "democratic advantage" hypothesis, I find that constrained authoritarians are approximately as likely to win wars as democrats. Unconstrained authoritarians, on the other hand, are much less likely to win their wars than other regime types. I then show that these same patterns hold at lower levels of military conflict, namely in Militarized Interstate Disputes (MIDs), which are militarized conflicts short of war. For both wars and MIDs, I show that alternative explanations do not better explain the patterns I observe, including showing that levels of democracy do not explain the disparity in war outcomes among regime types. A final empirical chapter looks at the implications of non-democratic accountability for a related area of international behavior: signaling and crisis bargaining. I show that leaders who can impede domestic elite coordination have difficulty generating audience costs and credible threats. In contrast, authoritarian regimes in which leaders are more constrained---even if the regime wholly lacks liberal institutions---are able to generate credible threats as effectively as democracies.Finally, I summarize the conclusion that reappears throughout the manuscript: unconstrained dictators who can render their tenure secure from domestic elites are indeed the "rogues" we suspect, picking fights, starting disastrous wars, and surviving in office only to repeat the cycle. In contrast, leaders who do not control access to high office, and who do not have the means to monitor and punish potential challenges to their rule, are nearly indistinguishable from democracies on a number of important dimensions of foreign policy. These are findings of great relevance not only to scholars of international relations, but also for policymakers, who can learn from these findings to better tailor their policies to the incentives of different types of authoritarian leaders. (Abstract shortened by UMI.)...