Why Ethnicity is Linked to Civil War: The Case for, and Limits of, Perennialism

Abstract

A lively debate in the comparative ethnic conflict literature pits those who argue that ethnic fractionalization is linked to civil conflict against those, most famously Fearon and Laitin (2003), who claim that ethnic diversity is not associated with violent civil conflict. The latter perspective is a staple of the mainstream instrumentalist school of ethnic politics, which seeks economic and political explanations for ethnic conflict. Drawing largely on Fearon and Laitin's (2003) data, we instead claim that while ethnolinguistic fractionalization (ELF) does not significantly predict conflict onset, it does indeed predict the incidence - i.e. presence - of civil conflict. This is because indigenous ethnic diversity, i.e. ELF, is relatively static over time, but varies across countries. Conflict onsets, by contrast, are more dependent on short-run changes over time. Little wonder the two are not associated. However, conflict incidents, like ELF, better reflect spatially-grounded conditioning factors and are thus significantly associated. Finally, we note that violent ethnic rebellion is confined to indigenous, or primary, ethnic groups while violent secession is the preserve of politically-conscious national ethnic groups. These findings are more consistent with perennialist than instrumentalist arguments. But this also exposes the limits of the perennialist paradigm: while it can help us identify contextual conditions for conflict in cross-sectional space, it cannot predict when violence will occur.