LAUSR

ABSTRACT This article contributes both empirically and methodologically. Empirically, we seek to advance our understanding of an important puzzle: does oil cause ethnic war? Methodologically, we seek to identify more precisely the different weaknesses and strengths of the quantitative approach and case studies with process-tracing by explicitly comparing results from these two approaches on the same empirical question. We thus subject the statistical association between the ethnogeographical location of oil and the onset of ethnic war to test with process-tracing. Examining several pathway cases, we find that oil has rarely been a deep cause of ethnic war. Instead, the ethnogeographical location of oil either reignites dormant conflict that has deeper roots in ethnic resentment and hatred or intensifies ongoing conflict, mostly by facilitating the operation of two interconnected mechanisms. Our study echoes the notion that quantitative exercises alone often cannot establish specific causal mechanisms or how contextual factors impact the operation of these mechanisms, and it is precisely on these two key fronts that qualitative exercises possess critical advantages. Hence, quantitative methods and qualitative methods are complementary rather than competitive. Our study also yields important policy implications for preventing and managing ethnic conflict in countries with rich mineral resource.