Source: Dissertation Abstracts International, Volume: 75-09(E), Section: A.

Adviser: Steven Wilkinson.

For leaders across the developing world, preventing military intervention in politics is a central challenge of governance. Although coups d'etat have become less common over the past few decades, they remain frequent and destabilizing events. In 2013 alone, soldiers in Egypt, Eritrea, Libya, Comoros, and Chad attempted to seize power through a coup. To prevent such intervention in politics, civilian rulers adopt a number of coup-prevention strategies, from bribing soldiers to restricting their access to arms. One of the most common coup-prevention strategies is "counterbalancing," in which leaders divide military power between the regular army and republican guards, militias, and other paramilitary forces. Despite the prevalence of counterbalancing, we know very little about the conditions under which rulers use it and what the consequences of counterbalancing are. How does counterbalancing work? When do leaders employ it? And is it effective at preventing military coups?

I identify three mechanisms through which counterbalancing affects coups. I argue that counterbalancing creates barriers to communication and coordination between forces that make coups more difficult to carry out. Where counterbalancing forces share ethnic or other ties to the incumbent regime, they are also more likely to use force to defend it. At the same time, however, counterbalancing may inflame military grievances, making them more inclined to attempt a coup. As a consequence, while counterbalancing does not necessarily deter soldiers from attempting to seize power, it reduces the likelihood that a coup attempt will succeed. It also makes for more violent coups.

Yet counterbalancing is not ubiquitous, even among rulers that face the threat of a coup. Why do some rulers counterbalance, while others do not? I argue that international and domestic security concerns interact to determine when and where rulers will counterbalance. The same barriers to communication that make coups more difficult to carry out in the presence of counterbalancing forces also impede battlefield coordination. Leaders employing counterbalancing forces hinder their own military effectiveness. As a result, those that face a high risk of war will thus be less inclined to counterbalance than those that do not. Consistent with this argument, I demonstrate that as the risk of interstate war has declined globally, counterbalancing has become more common; indeed, the incidence of counterbalancing has nearly doubled over the past fifty years.

The spread of counterbalancing as a coup-prevention strategy has had important consequences for domestic politics. I demonstrate that while counterbalancing is not associated with fewer coup attempts, coups that occur in the presence of counterbalancing forces are less likely to succeed in removing the incumbent regime from power than those that do not. In this respect, counterbalancing works. However, counterbalancing also results in coup attempts that are more violent. Since failed coups are a common catalyst for civil war, rulers that counterbalance are also trading the risk of a coup for that of broader social conflict.

I develop and test these arguments with a combination of formal, quantitative, and qualitative evidence. I provide a formal model that shows how counterbalancing and other coup-preventions strategies operate, as well as the incentives that structure the decision-making process around counterbalancing. I examine the plausibility of the model's predictions in a paired comparison of Bangladesh and Jordan, cases selected for variation in the risk of war while holding coup risk constant, and in an over-time comparison of the U.S.S.R./Russia immediately before and after the end of the Cold War, when perceptions about the risk of international conflict declined precipitously. I then test my arguments more broadly with an original panel dataset of 254 security forces in 65 developing states between 1960 and 2010. This dataset is the first to include information on the command, composition, and deployment of security forces, information crucial for capturing variation in the use of counterbalancing over time and space. Collecting this data involved surveying vast secondary literatures on military institutions and civil-military relations in each state, as well as historical news sources, annual defense publications, government websites, and reports from non-governmental organizations. As such, it represents the most comprehensive source on state security forces to date.

The results underscore the importance of opening up the "black box" of the state's security sector in order to understand the sources of political stability in the developing world. They also identify a novel channel through which international conflict affects domestic political development and pinpoint where conflict between military and civilian leaders is likely to escalate to civil war.