Civilian self-defense militias in civil war

To mitigate the costs associated with suppressing rebellion, states may rely on civilian self-defense militias to protect their territory from rebel groups. However, this decision is also costly, given that these self-defense groups may undermine control of its territory. This raises the question: why do governments cultivate self-defense militias when doing so risks that these militias will undermine their territorial control? Using a game theoretic model, we argue that states take this risk in order to prevent rebels from co-opting local populations, which in turn may shift power away from the government and toward the rebels. Governments strategically use civilian militias to raise the price rebels must pay for civilian cooperation, prevent rebels from harnessing a territory’s resources, and/or to deter rebels from challenging government control in key areas. Empirically, the model suggests states are likely to support the formation of self-defense militias in territory that may moderately improve the power of rebel groups, but not in areas that are either less valuable or areas that are critical to the government’s survival. These hypotheses are tested using data from the Colombian civil war from 1996 to 2008.