Content area
Full Text
This paper theorizes the link between ethnicity and conflict. Conventional research relies on the ethnolinguistic fractionalization index (ELF) to explore a possible causal connection between these two phenomena. However, such approaches implicitly postulate unrealistic, individualist interaction topologies. Moreover, ELF-based studies fail to articulate explicit causal mechanisms of collective action. To overcome these difficulties, we introduce the new index N* of ethnonationalist exclusiveness that maps ethnic configurations onto political violence. This formalization is confirmed statistically in regression analysis based on data from Eurasia and North Africa.
Toward the end of the Cold War, a wave of scholarship emerged that associated internal conflict with ethnic factors. Ethnic conflict in the former Yugoslavia and in Rwanda lent such theories considerable credibility. More recently, however, the political-economy literature has generated major studies that challenge this causal link (Fearon and Laitin 2003; Collier and Hoeffler 2004). Their application of econometric methods suggests that ethnic grievance washes away once materialist factors, such as per-capita income, access to raw materials, are controlled for.
Some of the most prominent assessments of the role of ethnicity in internal conflicts rely on various versions of the ethnolinguistic fractionalization index (ELF). Despite its widespread use, however, this index has yet to be supported by a convincing set of causal mechanisms that links it to political violence. Before rejecting ethnic characteristics as determinants of civil war onset, it is therefore reasonable to question whether the ELF index serves as a meaningful operationalization of ethnic politics.
In this paper, we argue that specific ethnonationalist configurations are more prone to generate violence in civil wars. In order to back up this claim, we offer an alternative index of eihnonationalist exclusion called N*, which does a better job of capturing mainstream theories of ethnonationalist violence. It deviates from standard fractionalization measures by introducing state-centric, rather than symmetric, ethnic configurations and by postulating group-level, rather than individual-level micro mechanisms of mobilization.
Using Fearon and Laitin's (2003) well-known insurgency model as an empirical reference point, we compare our new measure W to conventional indicators. Because of coding limitations, we focus on a subset of their global dataset, namely Eurasia and North Africa. The results are very encouraging: for the sub-sample in question, N* has a strong effect which is highly significant,...