Now, I have had a chance to look into the Ashraf and Galor (2012) paper on genetic diversity and economic development. I went through a huge pile of papers on diversity (as research for Gligorov 2012) and at least two additional ones rely on DNA evidence to explain stability and development, Desmet et al. (2011) and Spolaore and Wasziarg (2009). I am not impressed.

Some, e.g. Desmet et al. (2011), use the DNA data as an indication that the population is more or less diverse due to it being more or less closed to interactions with the other populations; so it is an indicator of, for example, ethnicity, but it is not what constitutes it. So, a sample of DNA drawn from ethnic nation X will have less diversity than the sample drawn from the cosmopolis, but the nation X will be ethnic for other reasons, e.g. language, culture, religion or whatever: DNA diversity could be the consequence of ethnic nationalism, rather than the other way around. The problem is that we have quite good data on these ethnic characteristics, so it is not clear what is the added value to using the genetic data – why use indirect rather than direct measure if the latter is available.

The others, e.g. Ashraf and Galor (2012), explain differences in development by the diversity of the DNA of particular populations. They relate diversity to economic performance in two ways. On the one hand, more diversity provides for richer set of endowments (traits) to draw from and should, simply by the greater richness of the implied combinations of the endowments of that population, lead to more innovations and thus to increases in total factor productivity. However, on the other hand, greater diversity brings in institutional costs, which can be expressed in a number of ways (e.g. higher distrust, whatever that is). This part of the argument is hard to make consistent, but it is appealing to those who believe that it is hard to manage, for instance, ethnic differences, irrespective of the fact that there is no clear way in which DNA diversity is connected with institutional frictions or with the distribution of economically meaningful endowments for that matter. This is why the first approach is what the other group of researchers prefers.

If one is looking for a justification for these assumptions of the positive role of diversity when it comes to innovations but negative when it comes to the cooperation within more diverse groups, there is nothing to be found in the Ashraf and Galor paper. They develop a model which states these two claims as hypothesis in the mathematical form and proceed to test for them. They find significant effects of varying levels of diversity on economic development: highly homogeneous populations are less developed as are the most diverse ones, while those in between manage to have higher levels of development. Original African cosmopolis from which populations migrated out was and still is characterised by high genetic diversity and low economic development and the tribes that migrated too far out to mix with others are too homogeneous and thus less developed, while those in Europe and Asia that have intermediate levels of diversity are also more developed. Diversity per se is not a problem except that it increases costs of cooperation in all shapes and forms. But all these claims are either not justified in any theoretical way or rely on other scientific disciplines where it is not clear that those claims were developed for the same purpose and with the needed justification to be applied to economic development. In any case, we do not know why richer set of endowments will be supportive of more innovative economic behaviour or why should members of more diverse populations have problems in trusting each other or in general transacting or cooperating with each other. So, there is yet to be shown why diversity is both good and bad in the hypothesized manner. It is also not clear how it is measured, but that is another story.

Is there a way to argue that it will be rather difficult to found the theory of development on genetic diversity? Let me look into the endowment part of the argument assuming people differ in traits and functionings in general. One can start with the Plato approach and assume that people differ in what they are good for. Plato thought there were three classes of people: warriors, labourers, and philosophers, now maybe more appropriately called scientists. Less than three classes, and there is no political society. More than three classes, e.g. add traders or bankers, and that is too much. The good society is the one in which everybody does what they are best for, that is employ their inborn endowments and avoid their misallocation (the latter to an extent being the job of the philosophers to ensure by advising on the best institutions that the society should adopt). The key is in the advantage of specialization which follows inborn physical and intellectual endowments.

Now, enter Descartes. Assume that we differ as res extensa (physical abilities) but not as res cogitans (intellectual abilities). If that sounds hard to believe, assume that apart from res cogitans there is also Plato’s res philosophans (I am unsure about the Latin here, but it sounds good). In other words, we have different physical and abilities to innovate and those are innate, so diverse DNA distribution increases the probability of there being more of both. So, diverse, but not too diverse, populations can combine res extensa to do more things and res philosophans to invent more new things. But, too much diversity may lead to conflicts, possibly violent, between various groups of res extensa and to acrimonious disputes between the various res philosophans, perhaps about the meaning of res publica. There will be, to revert to Plato, military entrepreneurs and too many schools of philosophy competing for, for instance, public money or honour or approval and diversity will prove unmanageable.

Ergo, we should expect that properly diversified DNA across a population, let us call it Platonic distribution, will lead to better performance of the economy, society and polity, i.e. that ethnic nations thus defined will be supportive of enlightenment and improvement.

Back to Cartesius. What about res cogitans? Clearly, there are differences in physical abilities, but what if lumen naturale is evenly distributed? Even if we cannot all light the fire of knowledge or carry the light, we can all see. Scientific innovation is individual, but we can all absorb even the most advanced scientific knowledge. After all, almost everybody can speak fluently and can use mathematics and learn even the most difficult scientific discoveries. On that assumption, liberal education and in particular the enlightened university learning is based.

So, let us assume that cosmopolis, initial one in Africa or the potentially existing one, is more diverse than an ethnic nation and the latter is more diverse than perhaps a tribe and that proper level of diversity is conducive to innovation and thus fasters growth of total factor productivity with the consequence that tribes and a cosmopolis will be less developed than optimally heterogeneous or homogenous nation states. Does that last step follow? No. One needs to explain the speed with which the enlightenment spreads, because even if not everybody innovates, everybody can learn. So, even the most diverse or most homogenous societies can do as well as anybody else.

So, with Cartesian rather than Platonic distribution, genetic or any other diversity is not the reason for divergence in development. There may be institutional or distributional reasons why knowledge does not diffuse or does not diffuse fast enough, but that is not a new insight and connecting it to genetic diversity begs the question. Spolaore and Wasziarg (2009) attempt to answer that question by measuring genetic distance on the assumption that those who are genetically closer learn easier from each other. So, genetic distance is a barrier to the diffusion of innovations and thus of development. This theory is equally ad hoc as that of Ashraf and Galor, but at least it identifies diffusion of knowledge as important for development, which implicitly assumes that knowledge can be evenly spread across the cosmopolis if for instance schools and universities are invented.


Ashraf, Q., O. Galor (2012), “The ‘Out of Africa’ Hypothesis, Human Genetic Diversity, and Comparative Economic Development”, American Economic Review, forthcoming.
Desmet, K., M. Le Breton, I. Ortuna-Ortin, S. Weber (2011), “The Stability and Breakup of Nations: A Quantitative Analysis”, Journal of Economic Growth 16: 183-213.
Gligorov, V. (2012), “Why Unions Fall Apart?”, unpublished.
Spolaore, E., R. Wasziarg (2009), “The Diffusion of Development”, Quarterly Journal of Economics 124: 469-529.


Pešč, 17.09.2012.