Land, history or modernization? Explaining ethnic fractionalization

Abstract

Ethnic fractionalization (EF) is frequently used as an explanatory tool in models of economic development, civil war and public goods provision. However, if EF is endogenous to political and economic change, its utility for further research diminishes. This turns out not to be the case. This paper provides the first comprehensive model of EF as a dependent variable. It contributes new data on the founding date of the largest ethnic group in each state. It builds political and international variables into the analysis alongside historical and geoclimatic parameters. It extends previous work by testing models of politically relevant EF. In addition, this research interprets model results in light of competing theories of nationalism and political change. Results show that cross-national variation in EF is largely exogenous to modern politico-economic change. However, the data are inconclusive with respect to competing geoclimatic, historical institutional and modernist theories of ethnogenesis.

Keywords:

• ethnic fractionalization
• geoclimatic
• theories of nationalism
• ethnogenesis
• dominant ethnicity
• founding date

Notes

1. Ethnic group as used here encompasses both ethnic categories and groups. For the distinction, see Eriksen (1993, 44).

2. Initial research on ELF used data from a 1960s Soviet ethnographic atlas (Bruk, S. I. and V. S. Apenchenko, eds. 1964. Atlas narodov mira.

3. All appendices at: http://www.sneps.net/ethnic/append.htm.

4. Exports encompasses a total for all four African slave trades: transatlantic, Indian Ocean, Red Sea, trans-Saharan. For more, see Nunn (2008).

5. See: http://www.sneps.net/ethnic/ethnicdates.htm.

6. All appendices at: http://www.sneps.net/ethnic/append.htm.

7. These partial models can be viewed at: http://www.sneps.net/ethnic/alternative.htm.

8. Logged variants of plurality ethnic group founding date also improve performance, but not as dramatically. For alternative specification with ethnic group founding date instead of 0–1100 AD origin, see: http://www.sneps.net/ethnic/alternative.htm.

9. Rerunning the four identity-based models (EF, ELF, PLURAL, PCTMAJ) with an interaction term for sub-Saharan African slave exports results in the sub-Saharan Africa dummy and sub-Saharan African slave exports each falling out of two models. Both remain signed in the expected direction across all models. Sub-Saharan African slave exports was not included in the combined model due to restricted degrees of freedom, but this specification is shown in Table 5 in: http://www.sneps.net/ethnic/alternative.htm.

10. See Table 2 in: http://www.sneps.net/ethnic/alternative.htm.

11. See Table 4 in: http://www.sneps.net/ethnic/alternative.htm.

12. See the National Geographic's Enduring Voices project website: http://travel.nationalgeographic.com/travel/enduring-voices.

Additional information

Notes on contributors

Eric Kaufmann

ERIC KAUFMANN is Professor in the Department of Politics at Birkbeck College, University of London.