Fight the Power : Effect of Post-Election Anti-Government Action upon the Probability of Incumbent Replacement

Abstract: The aim of this thesis is to answer the research question as to whether post-election antigovernment actions lead to increased probability of incumbent being replaced. The theoretical argument is that the anti-government groups can choose to use violence or non-violence to force the government to give concessions, the ultimate one being replacement of the incumbent. Both violence and non-violence are argued to have a positive effect on the probability of this to happen, but the latter should be stronger. Two hypotheses capture this: H1: Anti-government violence increases the incumbent’s probability of being replaced. H2: Anti-government non-violence increase the incumbent’s probability of being replaced more than anti-government violence does. Using a logistic regression on 550 elections compiled from the NELDA and ECAV datasets, the first hypothesis is not supported due to lack of statistical significance across the models. The second hypothesis is supported as the non-violent independent variable receives statistically significant results, but these do not hold in the robustness test. Taking into account the suggestions from previous literature, the results do hint towards non-violence being the more successful option of action, but it cannot be concluded with certainty.