Cyclicality in non-state violence: theory and evidence

What explains recurrent fluctuations in violence within militant groups? Temporal swings in the number of attacks perpetrated are common, but lack a general explanation. I provide evidence of cycles in attacks, and provide a theory of why organizational levels of violence vary serially. First, due to the contradictions of non-state violence, there exist, simultaneously, upward and downward pressures on its use. By "contradictions," I mean that violence is both beneficial and detrimental to an organization. It can secure public support and government concessions, but can also tarnish a group's reputation and invite retaliatory violence. Second, since the relative strength of each type of pressure (upward or downward) varies inversely with how much it is accommodated, organizations face disincentives to pause at high or low levels of violence. Groups that shift tactics to accommodate one pressure find they have activated another. For example, a group that curtails violent attacks to avoid government repression finds it lacks media attention. These pressures incentivize a reprise in violence, causing the up-and-down pattern to repeat. Finally, the cyclical pattern is reinforced by endogenous interactions between militants' political environments and their decision-making processes. I provide suggestive support for the arguments by examining the Basque group ETA.