Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Political Science


Since 1979, one out of six South American presidents have failed to complete their terms of office. These "failed presidencies" occur when democratically elected presidents are forced to leave office early, but without compromising the democratic order. This dissertation seeks to solve the puzzle of what drives presidents out of office. Previous studies have found that institutional and political factors, economic issues, and social mobilizations are powerful forces affecting presidential failures. In this research, I examine the impact these factors have on the likelihood of presidential failures.

Additionally, I argue that previous works have failed to find a significant relation between democracy and presidential failure because they have used an inadequate conceptualization and measurement of democracy. I postulate that using a country's current levels of democracy does not fully capture democracy's effects, and that thinking of democracy as a legacy may better enable us to see its true effects on presidential failures. Drawing upon the literature of regime legacies, I hypothesize that there is a significant negative causal relation between a country's democratic tradition or legacy and the occurrence of presidential failures. I employ a mixed-method research approach to test and examine the effects of these factors on the occurrence of presidential failures. Using survival analysis, I quantitatively analyze 65 all presidencies in South American countries between 1979 and 2012. In addition, I use a case-study approach for the qualitative, in-depth analysis of two failed presidencies: Fernando de la Rúa in Argentina (1999-2001) and Lucio Gutiérrez in Ecuador (2003-2005).

The major finding of this dissertation is that political institutions clearly matter for presidential survival. The survival analysis and the two case studies show that partisan support and democratic tradition play a pivotal role when it comes to surviving in office. Surprisingly, economic recessions, political scandals, general strikes, and anti-government demonstrations appear to have no substantial effects on presidential survival. Moreover, this dissertation offers new evidence for the "institutions vs. street" debate by showing that, when the role of congress is accounted for, the effects of social mobilizations significantly weaken.

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.