This thesis asks why some communities experience much lower rates of violence, and associated costs of war than surrounding areas in the midst of state collapse, examining the case of Butembo in North Kivu province in the Democratic Republic of Congo. This paper builds on two emerging themes in the literature surrounding civil wars and failed states. The first is a growing trend in examining intra-conflict dynamics and variations in patterns of violence over geographic space. A number of GIS-driven mapping projects have been conducted, as well as probing historical research. However most such studies examine the geographic dispersion of violence and armed actors within a given civil war context. The focus on explaining why violence breaks out in particular areas inevitably leads to a focus on the distribution of supposed violence-driving factors: lootable resources, inequalities and grievances, strategic considerations and capabilities of armed groups. This study examines the same phenomenon of radically different levels of violence within a relatively small area through a different lens, by focusing on areas that are relatively peaceful, and examining the presence of violence-constraining factors, most notably local institutional development.