Wednesday, February 25, 2004

Ethnic Fractionalization and Economic Development - More Evidence

I've just had the good fortune to discover this 2002 Harvard Institute of Economic Research paper, written by (amongst others) William Easterly of "Africa's Growth Tragedy" fame. What I particularly like about the new and more comprehensive data provided by this paper is that it does an even better job of capturing the amount of diversity in a given society, going beyond linguistic diversity to consider other markers of differentiation like race/ethnicity and religion.

We provide new measures of ethnic, linguistic and religious fractionalization for about 190 countries. These measures are more comprehensive than those previously used in the economics literature and we compare our new variables with those previously used. We also revisit the question of the e¤ects of ethnic, linguistic and religious fractionalization on quality of institutions and growth. We partly confirm and partly modify previous results. The patterns of cross-correlations between potential explanatory variables and their different degree of endogeneity makes it hard to make unqualified statements about competing explanations for economic growth and the quality of government.

The mark of a good paper like this one (unlike the sort of junk turned out by race cranks), is that the authors go out of their way to avoid making exaggerated claims about the scope and implications of their work. Nevertheless, one thing is clear from the data provided: the link between heterogeneity and poor growth is, if anything, stronger than was implied by Easterly and Levine's 1997 paper.

By the way, I couldn't resist pointing out the following information, for those who've ever doubted that Nigeria is a far more diverse place than India, in spite of the fact that the latter country has 8 times as many inhabitants as the former. For ethnic, linguistic and religious diversity, the numbers for India were 0.4182, 0.8069 and 0.3260 respectively (with a higher number meaning greater heterogeneity). For Nigeria, the corresponding numbers were 0.8505, 0.8316 and 0.7421, higher in all categories, but tremendously so for both the ethnic and religious measures. Looking over the data, not a single country on earth matches Nigeria for heterogeneity - not even Indonesia or Papua New Guinea!