Geography, Strategic Ambitions, and the Duration of Civil Conflict
Halvard Buhaug, NTNU , Trondheim
Scott Gates, PRIO, Oslo
Paivi Lujala, PRIO, Oslo
Both the control of territory possessing natural resources used to finance armed conflict and the distances an army must travel to project force affect how a civil war is fought and who will prevail. In this paper, a model based on a contest success function designed to explicitly account for distances is employed to model the ability to project force and sustain conflict. The strategic ambitions of the rebel group will determine whether the conflict is focused on territorial secession or conquest of the government. These goals, in turn, affect where the war is fought, how it is fought, and the likelihood of one of the parties succeeding militarily. Using both Cox regression and parametric survival analyses, specific propositions regarding the duration of conflicts derived from the formal model are analyzed. Special problems of repeated events, competing risks, non-proportionality, unmeasured heterogeneity (a general problem related to omitted-variable bias) will be addressed. Using a precisely dated armed civil conflict duration data (Gates & Strand, 2004), which is based on the Uppsala Armed Conflict dataset (Eriksson & Wallensteen, 2004), and using data regarding the location of conflict (Buhaug & Gates, 2002) and natural resources (Buhaug & Lujala, 2004) we are able to assess the role geography has on the duration of armed civil conflict, especially in terms of the location of the conflict zone and location of lootable natural resources. We are also able to differentiate armed civil conflicts according to the strategic ambitions of the rebels.