Explaining (some) Global Inequality: Genes, Culture, or Luck?

Share on FacebookShare on Google+Email this to someoneTweet about this on Twitter

Economists Wacziarg (Stanford) and Spolaore (Tufts) are using Cavalli-Sforza’s genetic distance data. They find that countries that are genetically different from the rest of humanity tend to be poorer, even after controlling for lots of popular variables (like geography and colonial experience, two recent favorites). Their explanation: It’s easier to get ideas from people who are similar–a contrast with much of the trade literature, where it’s easier to get gains from trade from people who are dissimilar.

As Google Scholar shows, W&S have drawn some attention. Along related lines, Bill Easterly and his coauthors remind us that the countries that were more innovative in 1000 BC tend to be richer today–so not much is new under the sun.

Of course, W&S remind us that it’s essentially impossible to disentangle genetic versus cultural stories when looking at nation-level data–and they note that their genetic measures are based on neutral markers, so it’s “different,” not “worse.”

By contrast, one way to interpret Easterly’s result is mere path persistence: A multiple-equilibrium story where you get it right once through sheer luck, and afterwards you’re likely to stay lucky forever. Difficult to disprove–maybe it’s a fair coin, and the West and East Asia just flipped Heads…

Genes, culture, luck: All three stories deserve some attention over the next few years. Two big problems: Measuring culture and luck. Genetic differences, those we’ll be able to measure, with greater precision every year. But that’s almost as much a curse as a blessing, since it’d be extremely valuable to be able to disentangle these hypotheses.

20 Comments

  1. I’m sceptical. The two most important peoples, if that’s a permitted word here, in World history – as seen backwards from here – are the Ancient Greeks and the early modern British. When the first flourished, the second’s ancestors were various savage tribes. When the second flourished, the descendants of the first were backward peasants. There may be more continuity in who didn’t contribute much than in who did.

  2. But if we learn that culture flows from biology…then what?  
     
    Is there even a website out there discussing what our intellectual superstructure is going to look like when the Blank Slate falls? I think we need one…

  3. But if we learn that culture flows from biology…then what?  
     
    biology isn’t determinative, it is just another parameter, that much is obvious. the key is disentangling the relationship. reading greeks describing the character of the roman people in 200 BCE vs. the italians in 1550 is illustrative.* now, another issue: how does culture affect biology? lactose tolerance and disease resistance are two obvious examples of allele profiles being shifted by a cultural matrix. i suspect population density is very important…evolution does not stand still. genetic distance data will get us only so far (as an example, cavalli-sforza found that south chinese cluster with southeast asians and north chinese with northeast asia. even disregarding that these data are partially corrected by later uniparental markers which suggest some Y lineage introgression north to south, it is notable that for the past 1000-1500 years south chinese have held to the same confucian values as north chinese, and for some periods south china was the confucian redoubt while the north was under barbarian control). 
     
    * ah, you say that genetic replacement has occurred. one, i would be willing to say in general no, that hasn’t happened, the romans favored realism in sculpture during many periods and the resemblance between the busts and modern italians is pretty obvious, so we know that many phenotypic characters persist. additionally, archeoDNA seems to be suggesting a lot of continuity, consider the etruscan data which shows deep time persistence over 2,000 years of a rather alien minority population. the further question is whether italian character has changed because of selection for various personality types, or because cultural evolution. i suspect a combination of both, though i tend to view the latter as a null hypothesis as immigrant groups tend to shift personalities pretty quickly showing how plastic this is.

  4. Economists Wacziarg (Stanford) and Spolaore (Tufts) are using Cavalli-Sforza’s genetic distance data. They find that countries that are genetically different from the rest of humanity tend to be poorer, even after controlling for lots of popular variables (like geography and colonial experience, two recent favorites). Their explanation: It’s easier to get ideas from people who are similar–a contrast with much of the trade literature, where it’s easier to get gains from trade from people who are dissimilar.  
     
    A quick pdf search for ?IQ? ?I.Q.? or ?intelligence? in the linked paper turned up no instances of appearance as separate words. So the authors restored to a correlation between neutral marker distance (which is really time back of separation into relatively non-interbreeding populations) and economic development without first finding any serious inadequacy in the correlation between economic development and that other partly genetically mediated (and much better studied and measured quality) IQ that Lynn argues for. They don?t even discuss that or mention Lynn. 
     
    This strikes me as largely an effort to BEGIN to bring genetic factors into ?RESPECTABLE? areas of wider debate and discussion within the social sciences, building on Cavalli-Sforza?s nervous making, but still respectable research (neutral markers only and all). I.e. I consider it highly unlikely that the authors were unaware of Lynn?s work or consider it so scientifically flawed as to be not worth taking seriously. (Which it doesn?t seem to be. It?s largely avoided especially outside the psychometric community because it?s so radioactively controversial.)  
     
    Instead I suspect they thought it wise to side step that radioactivity but still advance the ball by proposing an in and of itself not very plausible (or rather, unnecessarily vague (think differently due to genetic distance) hypothesis.

  5. yeah, the direction they’re going seems “obvious.” so the fact that they didn’t go that far in a biocultural model likely has to do with strategy and tactics.

  6. p.s. all, since the working paper is available, i ask we all actually read it so that the comments get idiotic more slowly than usual (i’m not commenting again until i read it. 
     
    tx

  7. p.p.s., i don’t mean that you need to read the bibliography & appendices (you’ll see why i say this).

  8. interesting…the paper seems a more econometric form of the argument in cavalli-sforza and feldmen’s book on culture, and mcneill & mcneill’s work on world history. some of the correlations were a little low for my taste, but whatever. got to reread later.

  9. did they publish their data somewhere?

  10. nope. it is in the working paper section.

  11. They find that countries that are genetically different from the rest of humanity tend to be poorer 
     
    I read both papers last year; this is an awkward phrasing. What Spolaore found is that countries that are genetically similar are also economically similar. This is mostly somewhat of an inefficient restatement that Euro, Asian, African, etc, countries cluster by economic development (sans the IQ talk, of course). But one point of interest is that they also compared just the European ethnic groups and found the same relationship. (Some other authors have already published (PDF) a rebuttal, saying the (Euro) relationship is explained by geographic trade barriers. Ok.) 
     
    Another point of interest is that controlling for language distance did not attenuate the correlation, suggesting that ‘barriers’ to information transmission are not the reason.

  12. Thanks to all for commenting on my paper with Enrico Spolaore. Yours is one of the most interesting blog discussions of our paper I have come across. 
     
    It is true that we are aware of Lynn’s work and equally true that we “consider it so scientifically flawed as to be not worth taking seriously”. This had nothing to do with tactics but everything to do with substance. We do not refer to these types of hypotheses because as we argue in the paper our data tells us nothing about the direct effect of genes (particularly data based on neutral markers) on economic performance. Instead, we argue using statistical evidence that genetic distance acts as a barrier to the diffusion of technologies broadly construed. 
     
    The effect of genetic distance is not incompatible with reversals of fortune of the type noted in the first comment. If a new center of innovation springs up for whatever reasons (such as England in the late 18th – early 19th century), it is genetic distance to that center that will determine in part who gets the innovations earlier. 
     
    We intend to release the data we used after the paper is published.

  13. Seems rather strained to me. Who did Japan imitate when they first industrialized, and again after WW2? Not anyone genetically close to them.

  14. Japan is in our sample and the statistical finding that genetic distance and income differences are correlated does not go away, so it must be an exception to the statistical regularity. The paper contains more related to this issue. First we certainly do not claim to explain all the variation in income differences – there is plenty of room for other factors. Second Japan may have adopted the Industrial Revolution earlier than its genetic distance from the English might have predicted, but that is not in contradiction with our theoretical model (which is stochastic) where it is possible for some countries to be far genetically but close in income terms. Just not on average.

  15. The Japanese have a tradition of emulating those cultures which they consider superior to their own (e.g. the American culture, since the US defeated them in WWII).

  16. It is true that we are aware of Lynn’s work and equally true that we “consider it so scientifically flawed as to be not worth taking seriously”. This had nothing to do with tactics but everything to do with substance. We do not refer to these types of hypotheses because as we argue in the paper our data tells us nothing about the direct effect of genes (particularly data based on neutral markers) on economic performance. 
     
    Dr. Wacziarg, thank you for commenting. The important point, that needs to be considered in your data, is that IQ levels also cluster by genetic similarity, and this single variable appears to explain a great deal about crossnational differences in economic development. This work is not flawed (at least no good criticisms have been raised in the journals), see my review from last year. Lynn’s work, like yours, charts differences that could be genetic or environmental in nature. So please don’t dismiss it on that basis. 
     
    I agree with you and your coauthor that vertical transmission is more important than horizontal transmission, but I think there are several reasons to favor parent to child transmission of genes as the active agent over transmission of “culture”
     
    1) Behavior genetic studies do show that parents strongly pass on affiliations to their children, such as religious and political identities, but also show parents do not significantly pass on economically relevant traits (e.g. intelligence and personality) outside of genes. Rowe and Rodgers, for instance, found a heritability of 0.42 for income but a trivial 0.08 for shared environment on income. This is much more consistent with a genetic rather than a ‘cultural’ parent to child effect. 
     
    3) The genetic/economic relationship that operates between countries also operates within countries. That is migrants’ genetic background predict their wages/human capital. (PDF) This extends to the 4th generation, even for Europeans, of which I have more data to add later on. 
     
    4) The genetic/economic relationship that operates between countries also operates within families. That is crossnationally adopted childrens’ genetic background predict their wages/human capital. Again I have much more data to add on this important issue later on.

  17. if anyone is up to it, it would be worth comparing their geographic controls with what Rosenberg et al. found http://dx.doi.org/10.1371/journal.pgen.0010070

  18. The paper by Jones and Schneider (2006) mentions Hendricks (2002) atleast 15 times but they did not list the paper in their bibliography, which makes it hard to find it. 
     
    I’m a little skeptical of their finding though (I’ll probably find their results more believable if I read Hendricks’). The number of developing countries that they have IQ data for is lot smaller than the number of source countries for immigrant workers. This is important because the ranking of countries in terms of national IQ is quite different from the ranking of the US wage of the immigrants from those countries. Borjas’ JEL paper in 1994 gives some info on immigrants’ wages.

  19. Hendricks : “How Important is Human Capital for Economic Development?” American Economic Review , Vol. 92, 1, pp. 198-219, 2002 
     
    http://www.econ.iastate.edu/research/publications/viewabstract.asp?pid=11409

  20. Genetic difference reflects family ties. It would hardly be surprising if money-making ideas tended to proliferate among networks of cousins and brothers-in-laws.

a