Impact of pro-government militias on civilian control over the military: A quantitative causal analysis

This dissertation aims to test the causal effect of pro-government militia (PGM) presence on civilian control over the military. Drawing upon agency theory, it argues that PGM presence causes civilian control to diminish. To test this hypothesis on three empirical criteria of causality: robust statistical association, non-spuriousness, and temporal precedence of the cause over the effect, it incorporates data on 155 countries between 1981 and 2010 from three datasets: Civilian Control Scores Dataset for the dependent variable¬ (dynamic civilian control score), Pro-Government Militias Database for the independent variable (PGM presence), and Varieties of Democracy Database for control variables (democracy, economics, state capacity, and conflict indicators). After shifting the dependent variable one to five years to the future, Statistical association, non-spuriousness, and temporal precedence were tested using Random Effects regression models with each version of the dependent variable, the independent variable, and all the combinations (sized one, two, three, four, …fourteen) of fourteen control variables. In all these 98304 analyses, the coefficient of PGM presence remained negative and statistically significant, enabling the rejection of the null hypothesis on no-association, spuriousness, and no temporal precedence dimensions. Keywords: Militias, Civilian Control, Causality, Agency Theory