Testing Multiple Conjunctural Causation in Theories of Interrupted Presidencies: A Replication Study

• In Event: Re-thinking the Latin American Presidency in the 21st Century

Abstract

Why are the causes of a surge of civilian removals of presidents in Latin America in the 1990s and 2000s? Pérez-Liñán (2007), Hochstetler (2006), Marsteintredet (2013), and others posit causally complex explanations based on a combination of political, social and institutional detonates, for example, that executive scandal begets social protest which begets legislative action. While these theories are convincing, scholars have either been unable to test them statistically or used econometric methods that do not consider the nonadditive and combinational aspects of the explanations. In short, these causally complex theories that have become accepted wisdom in the discipline have not fully demonstrated their statistical robustness. In this article, we replicate Pérez-Liñán’s existing statistical analyses, apply Braumoeller’s (2003) more suitable Boolean probit procedure for multiple conjunctural causation to the data, and compare the results. Outcomes are not as robust when the more methodologically appropriate econometric model is applied, indicating that understanding the causes of presidential instability requires further theory refinement and testing

Authors

• Santiago Basabe-Serrano, FLACSO Ecuador

• John Polga Hecimovich, University of Pittsburgh