Security Entrenchment and the Dynamics of Politicide: Evidence from Indonesia

This dissertation examines variations in violence during politicide. Why do some areas have greater levels of killings and long term detentions than others, controlling for the size of the victim group? Moreover, why do some areas have greater rates of killing relative to detentions? Using process tracing to compare three Indonesian provinces that experienced politicide in 1965-66, I argue these dynamics are shaped by two factors: the degree to which security forces are entrenched at the local level, and the strength of civilian networks willing to collaborate with security forces. I demonstrate that low security entrenchment leads to higher levels of violence and higher rates of killing, as security forces lack the ability to use violence more selectively. The strain of caring for these detainees then poses a logistical problem and incentivizes mass executions. Collaborator networks can further increase killings when security forces are weak by pressuring them to use additional lethal violence or engage in lethal violence unilaterally.