A Study on the onset of civil conflict

Abstract

The author broadly categorizes the models of civil war literatures on conflict onset into five dimensions: government, individual, geo-political, regional, and international. From empirically testing the models from each dimension, this study was able to find largely consistent results with major works on the topic. The following indicators were significant from the testing: on government dimension, the government’s capacity to repress and contain dissents; on individual dimension, the private incentives such as lootable primary commodities and opportunity costs proxied by education and income; on geo-political dimension, natural resource sites and the history of conflict; on regional dimension, regional spill-over effects through population movement and neighboring conflicts; finally, on international dimension, peacekeeping operations on the prevention of recurring civil war. This study used panel IV regression with an endogenous variable, GDP per capita instrumented with the access to electricity (% of household). Additionally, this study used dynamic panel models for the estimation of the effect of peacekeeping operations on the recurrence of civil war to supplement more commonly used survival analysis, which does not include temporal variations of independent variables. The findings suggest that there are specific conditions that favor the onset of civil war, and that civil wars are not necessarily accidental or idiosyncratic. Based on this findings, the author argues that the international efforts should be a two-tracked approach: one with immediate intervention to halt conflicts and regional spill-over; and the other with a more targeted approach to economic and social development.