Political violence and civil war have become a widespread phenomenon at the beginning of the new millennium and substance of violence, conflict and anarchy affect the life of many people. While the dominating political economy paradigm on civil wars has stressed the dichotomy of greed versus grievances as explanatory variables for the incidence and protracted duration of civil wars, this paper argues for a more contextual approach that investigates the nexus of greed and grievance in its space-time relation. An institutionalist political economy perspective can provide important insights into understanding the institutional logic of warfare and violence in its local context. I introduce two different approaches in the new institutionalism, namely the contract (transaction costs) and the distributional school of thought. Based on empirical studies from the war zones of Sri Lanka, I delineate the comparative advantages of the contractual and the distributional theory of institutional change in explaining 'real-life' phenomena on conflicts over property rights to local resources.