Academic conflict research, deeply rooted in Cold War thinking, has traditionally focused on the causes of war between states, with an analytical focus on the role of military alliances, joint land borders, arms races, nuclear deterrence, trade relations, democracy levels, balances and imbalances of power, and the like.1 Control over disputed territory has been the most frequent issue of international violent dissent among states.2 After the end of the Cold War, international relations scholars studying violent conflict noticed that since 1945, the most common and lethal form of organized violence was intra- rather than interstate war. Although in general less severe than interstate conflicts, 90% of all battle deaths since the end of the Cold War have been due to civil wars.
The early 1990s have seen a wave of civil wars in the Balkans, on the territory of the former Soviet Union, in large parts of Asia and in Western, Central and Eastern Africa. By the high water mark of 1994, more than a quarter of the states in the world were experiencing civil war, waging on average for six years.3 According to some scholars, this prevalence of civil wars should be treated as an entirely new phenomenon.4
From 1945 to 1999, about 130 civil wars have killed up to 20 million and displaced up to 70 million people in more than 70 countries worldwide. 5 In the same time period, 'only' 25 interstate wars with roughly 3 million battle related deaths were counted. Civil wars tend to last longer than interstate conflicts, are less likely to end by formal settlement, and often tend to recur.6 In addition, two thirds of these intrastate conflicts are fought along ethnic lines7, meaning that the resulting violence is not indiscriminate in nature, but targeted against (non-)members of certain identity groups, defined for example by language, religion, appearance or territorial attachment.
This development has led to a rapidly growing quantitative literature on civil war to identify correlates of onset, duration and termination in general and globally, both from an International Relations (IR) and a Comparative Politics (CP) perspective in political science. In addition, scholars from the fields of development economics and international political economy have started to systematically research causes and consequences of civil wars, and important empirical patterns have been uncovered.
At least 50 systematic studies on factors related to civil war onset, duration, termination or recurrence have appeared in the last five to ten years8, far outnumbering the output of former decades.9 While agreeing that civil war is mainly a result of poverty and less economically developed countries10, other factors possibly correlated with the outbreak of violence within states, such as ethnicity, remain highly disputed.
Despite the growing academic awareness, civil wars in general and ethnic conflicts in particular are still a poorly understood phenomenon.11 Although ethnic civil wars could be regarded as following a fundamentally different logic to that of non-ethnic ones12, and as complex power struggles between identity groups embedded in a deep historic and geographic context13, conventional quantitative research continues to treat them either as non-existent14 or as country-specific events that can be studied in isolation from each other15.
Most large-N statistical studies of (ethnic) civil war have focused on country-level characteristics to predict the onset of a conflict in a given year16, or have focused on duration as the dependent variable17. This literature has produced insightful empirical patterns on factors related to the outbreak of conflict. However, the global statistical approach to explaining civil wars has met with criticism recently, mainly because of concerns over theoretical and methodological shortcomings. Quantitative studies have been frequently criticized for their failure to capture the internal dynamics of civil wars by aggregating at too high a level of analysis, which has led to the emergence of a second generation of empirically downscaled civil war studies. …