This article will examine the link between ethnic diversity and economic success at the national, local, and group level. As will be seen, though not definitive, evidence at each level of aggregation suggests that ethnic diversity has a negative impact on economic prosperity. Before turning to the empirical evidence, let’s take a moment to see why, theoretically, ethnic diversity might be good, or bad, for the economy.
Ethnic diversity might be bad for the economy because it may cause people to work less well together. In-group and out-group preferences may also cause decisions within firms to be based on tribalism rather than meritocracy. Further still, ethnic differences can lead to political tribalism which reduces the efficiency of government. On the other hand, ethnic diversity may improve the economy by introducing a greater variety of skills and abilities. Either possibility is conceptually plausible. To decide between them, we must turn to empirical evidence.
Before doing so, two things are worth noting. First, when looking at national data sets it is not clear that statistical significance should be given the same weight it generally is in statistical research. The statistical significance of a result is supposed to tell us the probability of finding an association as strong as the one we found in our sample, due to random sampling error, when there is no association in the total population. In general, it may be true that a given effect size is small enough to be produced by chance in a sample size of only, say, 150. However, when this is a sample of 150 nations the sample includes nearly every nation on earth and, therefore nearly the entire population. Almost any association found in a sample of 150 nations is surely not due to sampling error. Given this, I don’t think we should be particularly troubled when national regression results come up as non-significant. Instead, their practical significance should be what concerns us.
Secondly, a quick word about how ethnic diversity is measured is warranted. Ethnic diversity is conceptualized as the probability that two randomly selected people from a population will be of different ethnicities. Unfortunately, this measure does not account for how different a given pair of ethnicities are. In other words, this measure will often count an area which is half British and half Somalian and an area which is half German and half Austrian as equally diverse. Worse still, because this data is based largely on government data and every government census does not count the same ethnicities. There is variation in which groups are counted as ethnicities across countries.
Some people might imagine that this makes current measures of ethnic diversity useless, but this is wrong. Because there is no reason to suspect that these flaws will tend to be worse in more or less wealthy societies, they introduce variation in the diversity metric which is not related to our outcome variables (economic progress) or what we would like to measure (meaningful ethnic diversity). That is, it introduces random error. Random error inevitably does one and only one thing: it biases associations towards zero and so makes associations weaker (though it does not change their direction). The high level of random error present in current measurements of ethnic diversity gives us yet another reason to be skeptical of the statistical significance of results given that significance is partly a function of effect size.
With all that said, let’s look at the empirical evidence. Studies have consistently shown a negative link between economic success at the national level and ethnic diversity. Poorer nations tend to be more diverse.
Moreover, diverse nations tend to have economies which grow more slowly than homogenous nations. This was the finding of Alesina et al. (2002) who found that ethnic diversity correlated at -.471 with economic growth between 1960 and 2000 in a sample of 119 countries. This association persisted when controlling for initial income, education, public debt, whether a nation was in sub-Saharan Africa, whether a nation was in Latin America, and other potentially confounding variables. The relationship’s strength was such that going from perfect homogeneity to perfect heterogeneity was predicted to decrease a nation’s growth rate by 1.9%.
Diversity was also found to have a negative association with business climate (not statistically significant, NS), tax compliance (NS), education rates, indexes of democracy and political rights, and overall government quality, and to have a positive association with corruption (NS), size of government (NS), and illiteracy rates. Controls in these analyses include population size, dummy variables for regionality (Sub Saharan Africa, East Asia, Latin America) the origin of the legal system (Socialist, French, German, or Scandinavian), latitude, and the prevalence of Islam and Catholicism.
Easterly (2001) found that institutional quality interacts with diversity’s effect on economic growth. His measurement of institutional quality includes the following: “(a) freedom from government repudiation of contracts, (b) freedom from expropriation, (c) rule of law, and (d) bureaucratic quality into an overall index of institutional quality.” It is highly correlated with other measures of corruption and quality of business environment. Easterly showed that diversity more strongly predicted low economic growth the poorer a nation’s institutions were. In fact, in Easterly’s model diversity had no significant effect on growth at the maximum possible value of institutional quality. Interpreting this data is complicated by the fact that, as just reviewed, ethnic diversity likely has a negative effect on institutional quality itself.
Substantial research has also been done at a more localized level. Looking at US counties between 1970 and 2000, Alesina and Ferrara (2003) found that ethnic diversity is negatively associated with economic growth even after controlling for the size and wealth of the county in 1970. Alesina and Ferrara also provided two reviews of other important studies in this literature which I think are worth quoting at length:
“An explicit focus on ethnic heterogeneity and economic performance is in the stud by La Ferrara (2002b). She uses an original dataset on production cooperatives in the informal settlements of Nairobi, and has information on all members of the surveyed groups, which allows to construct exact measures of group composition in terms of income, education, age and ethnicity. She finds that ethnicity matters for gaining access to group resources, especially in the form of cheap loans: members who share the same ethnicity as the chairperson are 20 to 25 percentage points more likely to borrow from the group or from other members. Ethnic heterogeneity also seems to influence the organization of production: members of more ethnically heterogeneous groups are less likely to specialize in different tasks and more likely to all do the same job. Also, ethnically fragmented groups more often adopt remuneration schemes in which every worker gets the same fixed amount, rather than being paid on the basis of the amount of work put in. These choices on division of labor and wage structure may be due to the relative difficulty of reaching consensus on “unequal” task allocations and remuneration schemes in ethnically heterogeneous groups…
A recent application to lending groups is provided by Karlan (2003). He uses data on members of a Peruvian micro finance organization, and exploits the random selection of people into groups to estimate the effect of group composition on repayment performance. He finds that members of more “homogeneous” groups, both in terms of geographical proximity and of cultural affiliation, are more likely to save and to repay their loans.”
Thus, research on more local levels of aggregation than nations are consistent with the national data in indicating that ethnic diversity damages a community’s economic performance.
Finally, there are studies, often experiments, which look at how well small groups perform. Hulsheger and Anderson (2009) meta-analyzed 8 studies and found background diversity correlated at -.133 (95% CI: -.318:+.052) with innovation.
Williams and O’Reilly (1998) came to this conclusion in their review of the literature:
“There is substantial evidence from both laboratory and field studies conducted over the past four decades that variations in group composition can have important effects on group functioning. These studies show that increased diversity, especially in terms of age, tenure, and ethnicity, typically have negative effects on social integration, communication, and conflict.”
Thus, research done at the national, local, and group, level all suggest the same thing: diversity is bad for the economy. While causality cannot be completely proven, the impressive list of confounding variables included in the longitudinal national research and the experimental nature of the group level research, are about as good as we can ask for in social science. Some of the effect sizes in this literature are small but, as I pointed out before, the measurement of ethnic diversity typically used has a lot of random error. At the very least, this evidence shows that the burden of proof is on the advocate for ethnic diversity to show that it does not harm the economy.