Why Does Ethnic Partition Foster Violence? Unpacking the Deep Historical Roots of Civil Conflicts

Abstract

Recent advances in political science and economics highlight the role of historical events in shaping contemporary political and economic outcomes. This article joins the growing literature by utilizing disaggregated data and mediation analysis to explore the causal mechanisms bridging ethnic partition by colonial borders and the risks of postcolonial civil conflicts in Africa. I argue that split ethnic groups are more likely to experience armed conflicts with central government during the postcolonial age, but the conflict-escalating effect is particularly acute for large-sized split groups. When coupled with sizable demographic forces, ethnic partition heightens the political salience of the corresponding ethnic cleavage while generating greater information and commitment problems. The empirical results provide considerable support for the theoretical claim: first, ethnic partition increases the likelihood of armed conflicts between politically excluded ethnic groups and incumbent; and second, a major part of the conflict-escalating effect is attributable to the indirect and causal interaction effects induced by contemporary group size. The empirical analysis also reveals that the role of a primary alternative mechanism, political discrimination against split groups, in generating the historical treatment effect remains limited.

Keywords: Africa, civil war, colonial institutions, conflict, geographic information systems (GIS), mediation analysis, spatial data

JEL Classification: D74, F51, F54, H56, N47, Z13