I propose a novel theoretical framework and examine whether leadership change can affect the behavior of rebel groups using updated original data on assassination attempts of rebel leaders during the Syrian civil war. The relationship between rebel leaders and insurgent group violence is tested on 15 of the largest rebel groups in Syria’s civil war using a quantitative design that leverages variation in insurgent group violence, as well as leadership transitions within insurgent groups. I exploit the inherent randomness of the success or failure of assassination attempts to identify the causal effect of rebel leadership removal on insurgent group violence. I find that successful assassination attempts reduce rebel group violence, but only when senior leaders are targeted, not mid-level officials. Importantly, successful attempts degrade groups, but do not reduce overall levels of violence in the conflict. This points to defection as an understudied mechanism in conflict with multiple armed groups.