Working Paper
Political Sources of Humanitarian Emergencies

This diagnostic study explores the political conditions that are associated with humanitarian emergencies. It employs a risk rather than cause-effect methodology. Humanitarian emergencies are not random events. They occur most frequently in states having the following characteristics: i) more than two distinct ethnic, language, and/ or religious communities; ii) recent (since 1945) independence; iii) government exclusion and often persecution of distinct social groups; iv) rule by kleptocrats or entrenched minorities; and v) weak government legitimacy. The study examines 17 non-random cases of humanitarian emergencies and finds a strong fit between the risk model and the patterns of armed violence. Contrary to many analysts who have emphasized 'ethnic conflict' as a major source of bloodshed, these cases show that it is governments rather than spontaneous explosions of ethnic hatred that usually launch the violence. In several cases, politicides and ethnocides had been planned and organized long before the humanitarian emergency began. External agents often become involved in domestic violence, but in the 17 cases foreign intervention was a causal factor in only two. In one other, withdrawal of foreign aid after the end of the cold war precipitated an economic crisis that led to the collapse of political authority and to subsequent killing and refugee flows. The study finds that organized politicides by governments have resulted in far greater casualties than civil wars, rebellions, and armed secessions. The presence of early warning indicators regrettably has not helped to mobilize the international community to prevent humanitarian emergencies.