Quamrul Ashraf and Oded Galor argue that “the level of genetic diversity within a society is found to have a hump-shaped effect on development outcomes in both the pre-colonial and the modern era, reflecting the trade-off between the beneficial and the detrimental effects of diversity upon productivity.” Africans have too much genetic variation, Amerindians too little, while Europe and East Asia are just right.
Of course, most genetic variation is neutral, having no significant effect on phenotypes, so the numbers they use are totally irrelevant to the question they’re addressing. One could imagine that it might be better to have more (or less) genetic variation in cognitive or personality traits, but we don’t know enough about the genetic architecture of those traits to say diddly about who has more or less.
Lots of people – not just Ashraf and Galor – seem to think that having more overall neutral variation implies more trait variation. That isn’t the case.
A population with more total (mostly neutral) variation can easily have less variation in a particular trait. For example, hair color and eye color ( both genetically controlled) are more variable in Europeans than in sub-Saharan Africa, even though African populations have more overall genetic variation.
There are particular populations in Africa that have extreme phenotypes – for example, Pygmies are shorter than anybody else – but that is a product of the special selective pressures they have experienced over a long time. It is not as if the standard deviation of height is lots bigger in major African populations than elsewhere.
If we ‘re talking IQ, African-Americans pretty clearly have less trait variation: their standard deviation of IQ is about 12 points, rather than 15 in Europeans.
The difference in mean IQ is far more important – as it is globally.