Ethnic and State History as Determinant of Ethnic Fractionalization

Abstract

This paper presents a new dataset which codes for the approximate founding date of the largest ethnic group in the world's states as well as the wave of state creation in which the particular country emerged. All contemporary questions of nationalism and ethnic conflict begin with the imperfect overlap between ethnic/national communities and political units. Ethnic fractionalization indices provide a quantitative measure of one manifestation of this: the degree of ethnic pluralism in a state. Ethnic fractionalization is produced by the overlay of political fragmentation atop ethnic fragmentation. There are currently close to 200 states overlaid on top of some 6000 ethnic groups producing a variegated ethnic grid that varies from state to state. Fractionalization will tend to decrease as ethnicity and politics move toward congruence. This can occur for three major reasons: a) the number of political units increases; b) the number of ethnic groups decreases or certain large ethnic nuclei attract members away from smaller groups; and/or c) the fit between ethnicity and politics improves. Unfortunately, with the partial exception of Philip Roeder's 1961 and 1985 dataset (Roeder 2001), only cross-sectional ethnic fractionalization data is available. This means that we cannot adequately test proposition a). However, we can derive indicators for the arguably more important processes b) and c). (Bockstette et al. 2002) have already demonstrated a significant negative association between the date of state formation (and degree of indigenous control of the state at the time) and the ethnolinguistic fractionalization index (ELF). Older indigenous political histories tend to be associated with less fractionalized states today. However this was not the primary purpose of their work, so the association was simply noted as an interesting univariate relationship. This paper attempts to build on this state history finding, but goes beyond it by parsing out ethnic and state formation variables. The foundation date of the largest ethnic group in a state serves as a measure of b) above, because we assume that older state-dominant ethnic groups have had more time to assimilate subaltern groups than newer state-dominant groups. According to ethnosymbolist theory, modern nations typically form around premodern ethnic 'core' groups which often engaged in social and ethnic assimilation prior to the modern epoch (Smith 1986). For modernists, modern state creation begins the process of nation-building and diversity reduction. Regardless of which approach one favours, it is worth noting that ethnic groups and states serve as assimilationist actors - large fractionalization-reducing ethnic nuclei - within multi-ethnic states. One question this paper seeks to adjudicate is whether ethnic pluralities or states are more important actors in this process. At one time, dominant minorities were common. With the decline of empire, dominant minorities give way to dominant pluralities or majorities as popular sovereignty and democratization spread (Kaufmann 2004). (Connor 1994a: 39) notes that while fewer than 10 percent of states are ethnically homogeneous, a substantial majority of them have an ethnic majority. All but a handful of states have a significant plurality group. How to measure the founding date of the largest ethnic group? This date is defined as the first imagining of the group by a putative member of the group and is determined through consultation with historians and social scientists who possess expertise on the appropriate country. To tap c) above, the paper distinguishes five waves of state formation: a) pre-1815; b) 1815-1917; c) 1918-45; d) 1946-89; e) 1990-present. Here we posit a connection between the date of state formation and the degree of ethnic-nationalist self-determination. This is not a continuous or ordinal variable as there have been discrete waves of state formation punctuating long periods of relative calm. The relative balance between ethnic self-determination, desire for territorial aggrandizement and other-determination of boundaries governs the congruence of borders. We assume that states which largely emerge through secession from a modern state (1990-present) are most homogeneous. Those springing from the transformation of an early modern state (pre-1815) would be predicted to be next in ethno-national congruence. These should be followed by those which emerged in the wake of Versailles (1918-45) and which were animated at least in part by Wilsonian principles of ethno-national congruence, though largely by Great Power motives and the domestic desire for territory. Early secessions from land empires (many of 1815-1917) should follow, and, finally, decolonizations, where ethnic self-determination is least evident (most of 1946-89 state creations). We also test other popular determinants of ELF such as climate, terrain, colonizer, state size and various indicators of modernization and economic development from several leading datasets.