The political economy of factions: Voting, bargaining and coercion

This dissertation develops game theoretic methods to study political interactions between internally fragmented groups. Each of the central chapters focuses on a major puzzle associated with each of three forms of political interaction: voting, bargaining, and coercion. The field of application of the results is wide, including peace negotiations, corporatist bargaining, international trade negotiations, and ethnic voting and ethnic violence.; In the chapter on voting I consider contexts where decisions are made by multiple groups, each using majority rule. Formal models suggest that a set of stable policies (the “core”) generally exists for majority rule decision-making within a single group if and only if there is only a single dimension of contention. Finding analogous conditions for core existence when an n-dimensional action requires agreement from m fragmented groups has been an open problem. I provide a solution to this problem by establishing sufficient conditions for core existence and characterizing the location and dimensionality of the core whenever n ≤ m. In higher dimensions, however, the core is generically empty.; In the chapter on bargaining I study the conjecture (the “Schelling conjecture”) that negotiators benefit from their differences with their constituencies. Since internal fragmentation may sometimes make bargaining difficult, the problem is to know when heterogeneity helps and when it hinders a sides ability to negotiate. Traditional approaches assume that negotiators are sophisticated while ratifiers are not. I demonstrate that these approaches misidentify bargaining outcomes in a wide class of cases. When an appropriate methodology is applied, the conjecture does not stand up well. In response, I allow, more realistically, for ratifiers to be sophisticated—acting as political agents rather than as technical constraints. In general when this occurs, the Schelling conjecture is supported.; In the chapter on coercion I study cases where political actors cannot depend on majority decisions to be enforced or expect bargained agreements to be honored. Instead they use coercion to attain their objectives unilaterally. For such contexts I develop a model that provides a mapping from cleavage structures and group size to aggregate levels of coercion. A key finding is that measures of ethnic polarization presently employed in empirical work are incapable of capturing—even qualitatively—the relationship between coalitional structures and social polarization.