Diversion and Political Survival in Latin America

Abstract

In spite of its long history among scholars of international conflict, empirical evaluations of diversionary theory have produced contrasting—even contradictory— results. We offer three reasons for these differences: choice of unit of analysis; failure to model the reciprocal relationship between threats to the survival of political leaders on one hand, and international conflict participation on the other; and measurement error in operationalizations of the independent variable—the incentive of leaders to divert. We construct an empirical test that addresses all of these concerns. Our analysis of data from Latin America during the period 1960-99 reveals robust evidence in support of diversionary theory. The results also expose the biasing effects associated with the failure to control for reciprocal causation and measurement problems.