Previous studies suggest the relationship of an inverted U-curve between regime type and civil war, implying that countries that fall in the middle range between pure democratic and pure autocratic governance have the highest risk of violent challenges to their authority. Yet this finding conveys little information about the institutional determinants of civil war within the vastly heterogeneous category of non-democratic regimes. The use of a one-dimensional measure, or a simple dummy variable, to capture the type of governance, conflates very different configurations of democratic and autocratic institutional arrangements. This paper aims to contribute to the study of the institutional determinants of civil war by examining whether some kinds of non-democratic regimes are more vulnerable to civil wars than others. It uses new, disaggregated data on type of non-democratic regimes, covering 191 countries between 1972 and 2003, which distinguishes between monarchical, military, no-party, one-party, and limited, multiparty authoritarian regimes. If democratic influences in authoritarian regimes are mere window-dressing and a sign of institutional weakness, as some previous research suggest, then the more liberalized autocracies should have a higher risk of civil war. However, if venues of popular influence are a sign of cooperation and institutions help the government make credible promises to the population, then the more liberal and institutionalized authoritarian regimes should be expected to be less vulnerable to violent challenges to their authority.
Fjelde, Hanne (2008) Non-Democratic Governance and the Onset of Civil War, 1972–2003, presented at 49th Annual Convention of the International Studies Association, 26–29 March.