Unnecessary Roughness: Examining Terrain, Indiscriminate Violence, and Conflict Duration

During the past decade scholars have attempted to identify factors influence conflict by using cross-national quantitative analysis, many of which utilize terrain roughness as an independent variable asserting that it provides an advantage in guerrilla warfare. However, despite the theoretical assumptions, these studies fail to reach consensus regarding how or if rough terrain contributes to conflict. One study in particular, Buhaug and Lujala (2005), found that higher levels of rough terrain in the conflict zone were associated, albeit insignificantly, with shorter conflicts, while higher levels of terrain roughness at the country level were associated with longer conflicts. This thesis seeks to explain this counterintuitive result by proposing a new theory about how terrain roughness impacts the way counterinsurgencies are fought. I argue that terrain roughness which conflict zones geographically separated from the capital experience higher levels of indiscriminate violence from the state which increases rebel resolve and prolongs the conflict. Using GIS analysis to construct terrain roughness measures of the country-level, conflict-zone-level and the area separating the conflict zone from the capital, this hypothesis was tested using Cox Proportional-hazards modeling, Seemingly Unrelated Regression, and Coarsened Exact Matching. The results from these test do not provide direct support for the hypothesis. Rough terrain and spatial separation between the conflict zone and the capital correlates to both shorter conflicts and fewer casualties. However, several of the underlying assumptions do receive strong support, including the relationship between state power and conflict location, cost sensitivity, and the application of indiscriminate violence