Armed Groups in Conflict: Competition and Political Violence in Pakistan

Abstract

This paper studies how an increase in the number of armed groups operating within an area affects the amount of organized political violence. We use plausible exogenous variation in the number of armed groups in Pakistan, by exploiting the split of a major group due to the natural death of its leader. Employing difference-in-difference and instrumental variable regressions on geocoded incident and fatality data allows us to derive a causal effect: more groups lead to more political violence. By combining different data sources and implementing a new approach to deal with potential double-counting, we provide a proxy for counter-insurgency efforts by the government. We show that the increase in violence is primarily driven by the armed groups and not by responses of the government.