Among political economists, the prevalence dictatorship and civil conflict captures the central features of societies afflicted with political violence. Accordingly, this chapter adopts this terminology and develop an analytical framework to better understand why many Muslim societies are inflicted with varying and heightened levels of political violence. The chapter identifies patterns in dictatorship and civil war in Muslim societies, relative to non-Muslim societies; and how they vary across oil and non-oil producing countries. These empirical patterns motivate a discussion of how variation in preexisting institutional structures and sources of nontax government revenue (rents) may explain these patterns in political violence. This discussion identifies tensions in existing theoretical accounts, leading the author to develop a formal model. The model shows that increases in rents can entrench dictatorship, while a decline in rents may facilitate peaceful transitions to democracy or lead to the outbreak of civil war. Societies whose preexisting institutional structures encourage “sharing” government resources with the opposition are more likely to transition to democracy. The model’s hypotheses form the basis of book’s subsequent empirical analysis.

Chapter 2 presents the (formal) theoretical framework and advances four arguments. First, whether or not majority support for public provision of insurance exists depends on the distribution of information. Second, some insurance is best provided via pay-as-you-go systems, but these involve a difficult time-inconsistency problem that private markets cannot solve. Some social insurance is therefore bound to remain almost entirely within the purview of the state. Third, people’s preferences regarding social insurance are also a function of the availability of private options. If social insurance is the only game in town, even those subsidizing the system will support it, leading to broad cross-class majorities. With private alternatives, the better-off will lower their support for public spending, which erodes the broad support public social policy programs traditionally enjoy. Fourth, parties continue to matter because they represent different risk groups, and we expect that partisan conflicts will extend into new areas, most importantly the regulation of information and how it may be used.