Strategic Choice Models of Political Change in Latin America

Abstract

Models based on rational analysis and game theory have come to play ap important role in political science and comparative politics.' Despite this development, analysts of Latin America often view these models as an alien tradition of research.2 Yet, in fact, an important variant of the rational perspective, choice analysis, has been usefully applied to research on the region. The strategic approach explored in this article emphasizes the methods through which actors pursue goals by shaping the context in which other actors make choices. The discussion focuses on three models. Each grew out of a particular historical and analytic context, though each has broader implications. Albert O. Hirschman's model of reform-mongering, from his classic book Journeys to Progress, reflects the widespread concern of political research in the 1960s with the politics of reform in Latin America. Adam Przeworksi's threshold model of regime transition grew out of the concern of the late 1970s and early 1980s with democratization and was vital to the major collaborative project on Transitions from Authoritarian Rule. Guillermo O'Donnell's model of democratic consolidation in Latin America, an important component of his writing on democracy, deals with a central political issue of the 1980s and 1990s.3 The goals of this article are twofold. First it examines the building blocks of these models. We show how the models treat actors, preference distributions, and thresholds, and we explore the strategies presented in the models: building coalitions through issue linkage, modifying subjective probabilities of success,4 modifying costs, and modifying perceptions of costs, especially by shaping information through the manipulation of signals and communication. Second, we address three questions about the contribution of these models. How can one determine whether such models fit the specific Latin American contexts to which they are applied? How different are these models from other, more familiar, analytic perspectives in the Latin American field? Finally, what are the implications of these models in understanding the relation between causal regularities, uncertainty, and in the study of development?