Harvard academics on genetic diversity and economic development

A group of Harvard academics have penned a short response to Ashraf and Galor’s forthcoming American Economic Review paper, The Out of Africa Hypothesis, Human Genetic Diversity and Comparative Economic Development.

Ashraf and Galor argue that economic development is affected by genetic diversity, which increases innovation but also increases conflict and distrust. This leads to an optimum “goldilocks” level of diversity, with genetically diverse Africans and less genetically diverse native Americans falling on either side of that optimum.

The Harvard academics suggest that the findings of the paper are scientifically flawed and that Ashraf and Galor “misuse genetic, evolutionary, archaeological, historical and cultural data”. They question the causal mechanism proposed by Ashraf and Galor and their statistical treatment. I will give my views on the causative mechanisms when I write a full post on the paper (probably to coincide with its publication). However, the statistical issue is interesting. The academics write:

The argument is also statistically flawed by treating genetic data as each population having an entirely independent history both from a genetic and from a historical point of view, when in fact, they are highly correlated and inextricably entangled with genetic population structure and with contingent historical events. Such haphazard methods and erroneous assumptions of statistical independence could equally find a genetic cause for the use of chopsticks.

This argument is similar to the statement that the various independent origins of agriculture are not actually “independent”. Populations in each region may have developed the specific idea of agriculture themselves, but they had a shared cultural and evolutionary history and their state at the time of the development of agriculture reflected elements of that shared history. However, in the limit, there is little of interest in the social sciences that is truly independent, so the question is what level of independence is required and whether statistical techniques can draw out relationships that are not spurious. Still, the risks presented by population structure in research such as this is very real.

In the last part of the letter, the Harvard academics reflect on the implications of the research and raise the common argument that this area should not be investigated due to “the potential to be misused with frightening consequences to justify indefensible practices such as ethnic cleansing or genocide.” I do not find this argument compelling, and consider that there is a need for robust, scientific exploration of an area that will continue to be debated with or without that research.

*Postscript: I somehow missed it, but Ashraf and Galor have written a response (Update: which now seems to have been taken down). Thanks Vincenzo for the pointer.

My posts on Ashraf and Galor’s paper on genetic diversity and economic growth are as follows:

  1. A summary of the paper methodology and findings
  2. Does genetic diversity increase innovation?
  3. Does genetic diversity increase conflict?
  4. Is genetic diversity a proxy for phenotypic diversity?
  5. Is population density a good measure of technological progress?
  6. What are the policy implications of the effects of genetic diversity on economic development?
  7. Should this paper have been published?

Other debate on this paper can also be found hereherehere and here.

Comments

  1. says

    Having not read the original paper I am not really qualified to comment. On the other hand some snippets I have read suggest that things are gone goofy on both sides of the net.

    First, on the Harvard side, most of the authors are from the anthropology department there. This suggests immediately that we should not pay a lot of attention to them, and indeed their arguments do not seem very solid. The response by the authors of the original paper does a good job of taking them down. Strange that Reich and Patterson signed on to the critique since they ought to be above this kind of thing.

    My take on the original paper is that they assume that the relevant kind of genetic diversity is heterozygosity. This is a new argument. Since there is only one history of our species, one would ordinarily expect that they would bring to us evidence from other organisms in support of their very strange hypothesis.

    In terms of evolutionary theory diversity usually means diversity in kinship. How different from me are people whom I encounter walking down the street? Hamilton’s theory and its descendants are based on kinship, which can be measured with SNPs and which is completely independent of and unrelated to heterozygosity. Certainly strife and distrust ought to be up with ethnic diversity within a group, but not with heterozygosity.

    I would expect respectable economists to bring comparative evidence to bear in support of their bizarre claim. Are regular Chimps more heterozygous than Pygmy chimps? Are highly inbred human isolates especially peaceful and cooperative? The ethnographies of lowland South American semi-isolates suggests no such thing.

  2. Nick Patterson (Broad) says

    I’m confused: You say the claim of the paper is bizarre (I agree) but that
    I ought to be “above this kind of thing” ; It is really hopeless statistically to
    find meaningful correlations with something like economic development and a genetic quantity
    such as heterozygosity given how few independent populations there are to look at, and
    how many genetic quantities could be chosen for analysis.

    This for me is the main point.

    • says

      Hi Nick:

      My remark was too cryptic: by “this kind of thing” I refer to the moralizing angst in the essay, like “In today’s economic climate, it is easy – and irresponsible — to blame the economic status of entire regions or peoples on uncontrollable factors like genetic variation.”

      Much of anthropology is not to be taken seriously because of just this failure to distinguish between what is true and what is desirable. I hate to see serious players like you and Reich and Danny Lieberman and David Pilbeam indulging.

      The objections to the statistical analysis of spatially autocorrelated data are right on the money—no disagreement with that at all.

    • Anke Mueller says

      After reading the entire exchange, I have to conclude that your criticism is subjective and is not based on the evidence presented in this exchange. According to the Nature article:

      “Sohini Ramachandran, a population geneticist at Brown University who provided the genetic data for the study… adds that Galor and Ashraf used estimates of genetic diversity that she and her colleagues [Cavalli-Sforza et al.] specifically developed to overcome many of the confounding factors caused by the overlapping genetic and cultural histories of neighboring countries.”

      Ashraf and Galor in their response further argue:
      “First, on the statistical front, our critics have falsely suggested that we treat socioeconomic and genetic data as if populations are independent of one another. On the
      contrary, our empirical analysis accounts for the possibility of spatial dependence across
      observations, including analytical methods that correct for spatial autocorrelation in “error
      terms” and bootstrapping. This criticism of our work thus reflects either a misunderstanding of the techniques that we employ or a superficial reading of our work.”

  3. Jade d'Alpoim Guedes says

    Mr. Harpending, Anthropologists do not have a problem with the truth. The question is can you demonstrate that this kind of argument is indeed true. If it relies on flimsy evidence such as the article in question does, I believe that it is irresponsible to make statements such as

    “cross –country migrations (linked to colonialism) altered genetic diversity and hence composition of human capital in colonized countries (…..) the level of diversity that existed in these locations during the pre-colonial era changed substantially, towards the optimum level for development” or that the “the high degree of diversity among African populations and the low degree of diversity among Native American populations have been a detrimental force in the development of these regions”.

    We would like to point out that an additional problem with this analysis is precisely their failure to consider the anthropological literature (well actually the literature in general) on complex behaviors like cooperation, trust, innovation. A problem that none of the economists commenting on this issue have picked up on. To support the claim that genetic heterozygosity “increases the likelihood of miscoordination and distrust, reducing cooperation and disrupting the social order”, Ashraf and Galor rely on data from a single question concerning trust on the World Values Survey and several animal cooperation studies described in their Appendix H. The studies selected by Ashraf and Galor emphasize the importance of kin selection in encouraging cooperative behaviors, that is, individuals cooperate more with close genetic relatives.The authors fail to account both for the animal literature in which kinship does not predict cooperation and for the vast literature on human cooperation.

    To demonstrate the genetic heterogeneity is associated with increased propensity to innovate, Ashraf and Galor use the number of scientific articles published per year, per capita, from the World Bank’s World Development Indicators, and a set of studies based on Drosophila and honey bees. The insect studies link genetic diversity to disease resistance in Drosophila, and to several aspects of hive performance in honeybees. The use of these animal studies to justify their claim that genetic homogeneity is causal to a human propensity to innovate is troubling. Even if disease resistance is linked to overall genetic heterogeneity, this is not indicative of innovation. Neither is hive performance in honeybees, a eusocial insect with a different genetic structure than humans. Perhaps Ashraf and Galor were inspired to use this data because there is no research demonstrating that genetic heterozygosity at the population level is associated with capacity to innovate. In addition, using the number of scientific publications per capita as a cross cultural indicator of innovation is in itself problematic. The number of scientific articles published by a nation is closely tied to the amount of government funding allocated to research, number of individuals able to attend secondary education etc.

    I find it quite interesting that you call on economists to discuss literature on heterozygosity in chimps… Why not call on evolutionary biologists or Biological anthropologists to discuss this? How many economists take classes on human evolution and genetics in their PhD programs? I think your dismissal of other fields because they exercise caution when dealing with such ideas based on their intimate knowledge of the data is troublesome. It’s almost like saying: even though these people have dealt with these issues for most of their careers, we will ignore what they say, because it does not agree with the uninformed conclusion we came to. Might I suggest a slice of humble pie??

    • says

      I have no argument with your empirical criticisms of the paper: I endorse and agree with all of them.

      My complaint is about the overlay of pious talk about “irresponsible” to criticize a paper about genetics and human variation.

      Let me put out an analogy. A really awful paper shows up about evolution. At its heart is a mildly interesting idea, out of left field, that might merit investigation. So I write a critique of the paper, pointing out that there is neither theoretical nor empirical support for the idea. Then I also point out that Genesis tells us all we need to know about evolution and that this paper directly violates the truth of Genesis. Not only would you ignore my response, you would never pay any attention to anything I did after that. The “irresponsible” prose in your rejoinder, citing (not you, someone else) S.J. Gould, is precisely like my reliance on Genesis to critique a bad paper about evolution.

      I am not against anthropology: indeed my day job is teaching anthropology and I even have my doctorate from (gulp) mother Harvard. It breaks my heart that my beloved discipline has been overrun by pious talking twits who claim that they have something to say. You have something to say, if you wrote most of your critique, but why put in the religion stuff? What this kind of thing has done is to have made anthropology a laughing stock in the sciences.

      So why the attention this manuscript is getting? Why would an editor at SCIENCE pick it out? Interesting question. Certainly evolution and economics are hot topics these days, and a lot of the focus is on fallout from _IQ and the Wealth of Nations_ by Lynn and Vanhanen. So a way of seeming to be au courant is to pick up a pseudo-reputable paper and write about it as a way of avoiding, even burying, Lynn and Vanhanen.

      Or am I just being paranoid?

      • Jade d'Alpoim Guedes says

        When did I ever bring up religion? I said its irresponsible to make such claims and suggest (in other venues) that genoeconomics could be used to make policy.
        I believe what I said was if you are going to justify colonialism in the way Ashraf and Galor do and suggest that one can optimize genetic diversity you better have some damn good proof behind your claims and preferably an experiment you can repeat. This goes for any scientific claim. There is no religion behind this.

      • says

        Things like “the suggestion that an ideal level of genetic variation could foster economic growth and could even be engineered has the potential to be misused with frightening consequences to justify indefensible practices such as ethnic cleansing or genocide” read a lot like religion or magical thinking to me.

      • Evans says

        Would the creationists, religiously fanatic statements, of the ayatollahs from the Anthropology department at Harvard be featured prominently in the Teheran Daily?

      • Jeffrey Shyu says

        Frankly, I fail to see how pointing out the normative consequences of research makes it “religion” or “magical thinking” at all. Research has real-world ramifications. And often, research is carried out with the real-world consequences in mind. This is the case for nearly all of biomedical research, for example. Dr. Harpending, your point would seem to suggest that you also think that medical researchers engage in “religious” or “magical” thinking when they advocate for a new breast cancer drug, based on strong evidence that it saves lives.

        Being aware of the normative context of a research study, its benefits or its harms, is hardly magical thinking. In fact, one probably has to have an irrational blind faith to believe that research has no normative consequences at all.

      • Jeffrey Shyu says

        Look, if it’s okay for economists to point out the potential positive outcomes of a research study, then why is it not okay for anthropologists to point out the potential negatives?

      • Very Conscientious Scientst says

        Dr. Shyu, I cannot agree more with you. Indeed, _all_ normative aspects of _any_ research always need to be taken into an account. For instance, the use of ultrasound technology in obstetrics has been irresponsible for decades! Ultrasound has been used in selective abortions in India and China thereby directly causing a loss of hundreds of thousands of unborn women’s lives. Yet, hardly anyone mentions Gendercide, the scale of which exceeds many known instances of genocide. “In this economic climate”, when our rocket ships are conquering outer space, it is irresponsible to ignore the greatest violation of human rights happening right here, on this planet!

      • Nathasa says

        It is mind-boggling to me how someone can claim that stressing the benefits of genetic diversity (celebrating diversity actually) can lead to genocide.

  4. Jeffrey Shyu says

    Setting aside the fact that it’s pretty laughable to use insect data to support a hypothesis about human behavior (I mean, seriously), one of the more troubling things about Ashraf and Galor’s paper and their subsequent reply is the significance they attach to the genetic data.

    In the paper, they are clearly arguing for the much stronger interpretation that genetic heterozygosity is a causal force in economic development. The quotes highlighted by Dr. Guedes clearly demonstrate this.

    However in their reply, they make the much more modest claim that genetic diversity is merely intended as a “proxy”, or some wastebasket measure for other phenotypic, sociocultural factors that actually influence said development. I find this very problematic on two counts.

    1) If the genetic data is merely proxy, then why study it? After all, we have plenty of more direct evidence from phenotypic and sociocultural data, from both anthropology and economics, on why some societies are more successful. It would make their research rather useless.

    2) I take that the “proxy” talk is intended to shield them from criticism of their actual, original claims. This sleight of hand is intended to win the skeptics, while the original claims are for the converted. But this kind of deception is deeply troubling as it confers a false legitimacy to a pseudoscientific thesis that could be easily co-opted for malignant purposes.

    • J.G. Frazer says

      “1) If the genetic data is merely proxy, then why study it? After all, we have plenty of more direct evidence from phenotypic and sociocultural data, from both anthropology and economics, on why some societies are more successful. It would make their research rather useless.”

      Actually, you hit the problem on the head – unfortunately, there is a dearth of direct measures of diversity.

      On a theoretical level, there is little difference between use of phenotypic data and genetic data. Anthropologists would equally object to it (you can ask Ms. Jade D’ about it). Also, phenotypic differences are harder to objectively measure as they are socially selected for.

      Sociocultural data – unfortunately, data on sociocultural traits is too diverse and cannot be properly standardized. Direct sociocultural comparison is a tricky topic that lends itself better for qualitative rather than quantitative analysis and precludes drawing larger scale projections. For example, what kind of sociocultural data could one use to compare the Dobe Ju/’Hoansi and Italian-Americans without being reductionist?

      So, no, I would not consider their study useless or redundant. It does raise an important topic of cross-cultural long-term comparison.

      “2) I take that the “proxy” talk is intended to shield them from criticism of their actual, original claims. This sleight of hand is intended to win the skeptics, while the original claims are for the converted. But this kind of deception is deeply troubling as it confers a false legitimacy to a pseudoscientific thesis that could be easily co-opted for malignant purposes.”

      Well, I would not demonize the authors by ascribing them conniving logic (even though they are economists). Some of their claims are indeed outrageous but should not overshadow the rest of their work and their conclusions. In the end of the day, any research in comparative politics or even quantitative sociology that links, say, ethnic fractionalization and conflict could be accused of malignant intentions and could be used to incite violence. Does it mean that comparative poli sci should cease to exist?

  5. J.G. Frazer says

    When in 1957 “Dr Zhivago”, a novel by Boris Pasternak, was smuggled out of the USSR and published, the Soviet elites were unanimous in their rhetoric: “We haven’t read ‘Dr Zhivago’, but vehemently condemn “. It seems that much of discussions about Ashraf and Galor’s paper follows a very similar trajectory. Observers confess that they haven’t read the paper beyond snippets in “Nature” or an abstract in “Science” and yet they take a firm stance.

    I believe that Ashraf and Galor’s paper deserves closer attention. Its critics are largely missing a crucial point. The authors’ main argument is a little bit less and at the same time a little bit more than what everybody thinks it to be.

    First of all, as claimed by the authors, Ashraf and Galor’s argument is NOT about genetic diversity. Indeed, they do use genetic data, but only as a proxy for a multitude of other ways that create differentiation in a society. As I see it, it is not about genes making humans cooperate or distrust each other. It is not about our biology driving humans towards innovation or conflict. It is about diversity (that could be physical, cultural, or anything in between) and its effects on social organization. And anthropologists are the first to argue that specifics of social organization play a role in cooperation and conflict.

    Secondly, the measure of diversity that the authors chose is very much in line with research on ethnicity and differentiation in anthropology and sociology. Indeed, ethnicity and race as constructed phenomena have long been established in social sciences. Yet, the question of methodological difficulties of studying differences in a society (especially in a comparative, cross-cultural context) remains as strong as ever. How do you study “ethnicity without groups” or “ethnicity, inc” if you would like to do so across countries? Ashraf and Galor’s study appears to offer a solution to this predicament.

    As a cultural anthropologist working on issues of ethnicity and conflict, I find their study intriguing and hope the dialogue that the authors have initiated will continue in a more productive manner.

  6. Anke Mueller says

    After reading the entire exchange, I have to conclude that your criticism is subjective and is not based on the evidence presented in this exchange. According to the Nature article:

    “Sohini Ramachandran, a population geneticist at Brown University who provided the genetic data for the study… adds that Galor and Ashraf used estimates of genetic diversity that she and her colleagues [Cavalli-Sforza et al.] specifically developed to overcome many of the confounding factors caused by the overlapping genetic and cultural histories of neighboring countries.”

    Ashraf and Galor in their response further argue:
    “First, on the statistical front, our critics have falsely suggested that we treat socioeconomic and genetic data as if populations are independent of one another. On the
    contrary, our empirical analysis accounts for the possibility of spatial dependence across
    observations, including analytical methods that correct for spatial autocorrelation in “error
    terms” and bootstrapping. This criticism of our work thus reflects either a misunderstanding of the techniques that we employ or a superficial reading of our work.”

  7. Vincenzo says

    here some other fuel for the debate: http://www.vwlmac.rwth-aachen.de/blog/2012/10/a-controversy-between-oded-galorquamrul-ashraf-and-harvard-anthropologists/

    Mainly, there are two points, as I saw the debate.
    First, there is a technical, scientific ground on which I find the reply of Galor and Ashraf convincing. What does it mean regression, or less, correlation? This appears to me confusing in the eyes of the anthropologists of the letter; the Ashraf/Galor reply should have set the field. For instance, the use of the Maya-Aztech story indicates that the authors of the letter clearly missed the point of what does it mean to have a series of data points. The reply about the smokers-cancer is something astonishing for the simplicity and for the clarity of the argument. The debate still continue on the spatial correlation, notwithstanding that Ashraf/Galor explained in the paper and in the letter the methodology they use to correct for spatial correlation and all the related issues.
    As still explained, not only genetic diversity is used a proxy for more complex diversity, but moreover in the regressions an IV strategy, that exploits the migratory distance, has been used to further ensure about the non-existence of counfounding factors behind diversity and economic development. It is exactly this strategy that emphasizes that the authors recognize the possibility of spatial correlation, for which do they correct.

    Further, there is a second point. This appears to me more political. Is an economist allowed to enter in a literature (i.e. anthropology) without consulting, before, a group of anthropologists. Here the link on the top I suggest. This an empty debate, I think, because miss, or induce to miss, the main point, which is related to the theoretical and methodological content of the paper.

  8. Nick Patterson (Broad) says

    I’ll confess to not understanding this proxy argument.
    Genetic heterozygosity is (anti)-correlated with distance form Africa and with phenomic diversity.
    That’s true because they are all related to an known historical process (the out of Africa migration).

    What has this to do with “diversity” in a general sense? I have no idea.
    For instance, Yoruba Nigerians, a people of Nigeria, with a common language have greater genetic diversity
    (say as measured by heterozygosity) than the whole of Europe. Do they have more “diversity”?

    So heterozygosity is a proxy for what exactly?
    ~

    • J.G. Frazer says

      “For instance, Yoruba Nigerians, a people of Nigeria, with a common language have greater genetic diversity
      (say as measured by heterozygosity) than the whole of Europe. Do they have more “diversity”?”

      You are absolutely correct – Nigeria is one of the more diverse countries in the world. It is home to over 300 different ethnic groups that speak over 500 languages. There are 3 ethnic groups that represent a statistical majority (Yoruba, Igbo and Hausa-Fulani). However, even those groups recognize finer internal divisions. So, as you can imagine, the question of ethnicity is a tricky one there (that reflects the past colonial history of ethnicity construction and contemporary political struggles, but that’s a different story). All in all, one thing is certain – Nigeria is far from being uniform or homogeneous.

      Disclaimer: I am not conflating ethnic differences with heterozygosity. Ethnicity is a symbolic construction of identity that is influenced by geography, history, institutions, intra- and in-group interactions, etc, etc, etc. So, it is as different from heterozygosity as it could be. These are two absolutely different categories and concepts.

      But you are also correct in your question – heterozygosity is a proxy for what? I think Ashraf and Galor’s argument is that this needs to be studied further. As they say, “what are the different potentially-measurable dimensions of diversity that are important for social and economic outcomes?” As of now, it is only possible to state a statistically significant correlation between hyterozygosity and diversity in at least one cultural trait (phonemic language structure). Evidence does seem to point at other aspects of cultural diversity (Nigeria is a perfect example of that), but more careful study is necessary. And this is consistent with their call for further research on this topic.

    • says

      I gather that they are taking their heterozygosity calculations from the old HGDP SNP tables. These always had, it seemed to me, a bad ascertainment problem. I have seen mention of a properly ascertained set of SNPs, at Harvard. From your shop I presume.

      Do new correctly ascertained data change the picture of global diversity clines very much? I do not keep very close touch on genetics any more.

      Thanks,

      • Nick Patterson (Broad) says

        There are some ascertainment issues lurking here, and the Reich lab has indeed made a public
        data set with ascertainment controlled. But I don’t think this would change very much.
        This is a very minor issue, compared with the major ones.

  9. Jade d'Alpoim Guedes says

    Hi Anke and everyone,

    Regarding the example we used for the Americas, our logic was as follows. We of course understand that there are multiple data points in the study, we were questioning simply how these data points are used.

    Let’s just take one example from the paper: the figure where they look at results at 1000 AD.
    Seeing that we are talking about the Americas, let us focus on one point in particular, the USA. Notice its particularly low population density. Why might this be the case. For one, when you look in their appendix, they clearer state that they are using modern nation boundaries which date to 1975. To do so for North America makes very little sense. At this period in time there are a large number of Native American groups, many of which had high population densities. However if one takes the population density of each of these groups and divides it over the entire surface area of the USA this is highly misleading. It would be the same as counting the whole of Europe as one nation. Each of these groups should be counted as a separate nation.

    Regarding this figure, it is also worth pointing out the following: In Appendix E and table A.3 the peripheral centers of Athens, Constantinople, Cordoba (and the entire country of Russia!)are taken to stand for “Europe”. (Also, note that the same city, Constantinople, is considered “Europe” in 1000 CE and “Asia” in 1500 CE.) Constantinople, Turkey and Cordoba Spain were under North African occupation until 1236 CE as proxies for European Populations. By the same token, the productivity of Cordoba should be counted as an African Nation. If these are the two largest populations in Europe (and one is not in Europe and the other is occupied by Moors), then the authors have by default demonstrated that in 1000 CE Europe does not conform. These are just a few out of many examples where this data is used in a faulty fashion. In fact, I see one very interesting trend in how the data is used. While modern boundaries might give us OK estimates for Europe, they are consistently underestimating pop. density in the Americas and in Africa. We bring up the example of the Aztecs and Mayas because the same problem occurs here. For the Aztecs the authors use the entire surface area of Mexico to calculate pop. density. No archaeologist would argue that the boundaries of Aztec control conform to that of modern Mexico. Indeed one should be calculating population based on a much much smaller surface area. To add in territory outside of the basin of Mexico seems completely irrational. There are very few Aztec archaeological remains outside of the basin of Mexico. I also doubt one could replicate their results at say 4000 BC or 8000 BC.

    Regarding their misuse of the genetic data, its worth taking a look at how many populations they used for the Americas (only 4). The authors cite a paper (Wang et al. 2007) which contains data for an additional 24 yet they fail to use this data. In fact, one sample, the Brazilian one, appears to play a large role in anchoring their line of best fit. The Brazilian population they use are outliers within the Americas itself. Using this point to estimate genetic diversity in the rest of the continent seems ridiculous. Had they used other more representative Brazilian populations this point should would no longer be at 0.44.

    There are also major issues with things like start date for the Neolithic transition (their sources in most of the areas of the world I have fact checked are totally off!!). Same with the reference Ollson et al on domesticable plants. The numbers refer to the geographical distribution of the world’s 56 heaviest wild grasses (Blumler, 1992). Assuming that the only domesticable plants are wild grasses is simply ridiculous. As someone who works in ecology, I also have major issues with how they have “corrected” for land suitable for agriculture. They use a measure very familiar to me Ramankutty’s 2002 measure of land suitability which uses growing degree days and soil suitability. It is well known that this measure cannot account for instance for coverage by tropical forests. Using this measure makes the Amazon look like prime agricultural land, something it is never been.

    I could go on and on pointing out all the examples, but I will stop here.

    Some final points:

    There is indeed a lot slippage about how the term “diversity” is used. In the authors reply, they claim they refer to diversity in general. What is meant by “Diversity more broadly defined”. The authors seem to be unable to explain.Yet in the abstract, introduction, and conclusion of their paper they clearly state that they are dealing with genetic diversity. In fact whether they now like it or not, or whether or not they reworded these parts of their paper to reflect their new claim, they clearly are. The very measure they are using is heterozygosity is derived from alleles.

    Regarding the statement about religion, we have never claimed that just because a scientific find is not politically correct that it should be suppressed. What we argue is then when such finds are made, because of their potential dangerous implications, some caution are appropriate. I think Mr. Shyu pointed this out quite well. Only rarely do we carry out academic research in a complete void of real world effects. By keeping the normative effects of our research in mind, we are also forced to be more critical of our own findings. Indeed if Ashraf and Galor (who now change their argument to say that they are talking about diversity in general) had thought about the normative effects earlier, they may not have been inclined to make statements like “Specifically, the results indicate that the migration to the New World in the course of European colonization significantly altered the GENETIC DIVERSITY (my emphasis) and, hence, the composition of human capital in New-World countries. In particular, the suboptimal level of diversity that existed in the New World during the precolonial era increased substantially, towards the optimal level, in the post-1500″.

    In most fields with deal with human subjects research (medicine, biology, anthropology, genetics etc etc) ethics classes are generally required and people think actively about these questions. Why should economics be exempt???

  10. Henry Harpending says

    @Jeffrey Shyu:

    No disagreement with what you say, but research in oncology about new cancer drugs is not quite the right analogy. Let me propose what I think is a closer analogy.

    In this imaginary world academic oncology is committed to the idea that lack of natural foods is the cause of cancer and that appropriate treatment is the right kind of natural food. There is some diversity of opinion: some feel that whole wheat bread, enough of it, will stop cancer. Others say avoid any wheat but eat large quantities of asparagus. A prominent group advocates alfalfa sprouts. While there are minor differences in detail, all agree that pharmacological treatment of cancer is unnatural and dangerous and a major threat to public health. No pharmacological oncology should be published without including a statement that drugs should only be used as a desperate last resort in cancer cases and only then with full contemplation of the risks of mass death if these drugs are not rigidly controlled.

    The natural food schools of oncology have taken over academic departments by claiming the moral high ground, and they enjoy a near monopoly on federal funding. They maintain this status quo, in part, not only by proclaiming their moral superiority but also by insisting that such boilerplate always be inserted in publications.

    I would call this religion, others might call it Soviet, or even magical thinking.

  11. Jade d'Alpoim Guedes says

    @Natasa Herein lies the problem. Ashraf and Galor are not just saying more diversity is good. They are saying there is an optimal level, which just so happens to correspond to the levels that Europeans and Asians have. “the high diversity of African populations and the low diversity of Native American populations have been detrimental for the development of these regions”. They are essentially blaming population structure (genetic diversity levels) for a “lack of development” in these regions. But they don’t stop here, they go further in their conclusion to suggest that colonialism actually had a positive effect on these continents. “Specifically, the results indicate that the migration to the New World in the course of European colonization significantly altered the genetic diversity and, hence, the composition of human capita in New-World countries. In particular, the suboptimal level of diversity that existed in the New World during the precolonial era increased substantially, towards the optimal level, in the post-1500 time period.”

    I hope in reading my comments above and once you read our full rebuttal (forthcoming you will understand why firstly their measures of population density in these regions are fundamentally flawed. Secondly, there truly is no evidence to support Ashraf and Galor’s claim that these populations had a “lack of development”. The concept itself of development is very tricky, but even more so, is how their own estimate of “development”= population density is calculated. Just think for a second about how population density was calculated for the USA in AD 100, AD 1000 and AD 1500. You are taking modern national boundaries within which there were many “nations” and dividing their population by total surface area…….I’ll be happy to point by point for the Americas demonstrate why every single one of their pop. estimates (and estimates for diversity after they carry out their regression) is flawed. Thirdly regarding colonialism, when Europeans arrived in the New World, they decimated Native American populations.. we all know how and why so I won’t go into this. How exactly can almost completely killing off entire populations and replacing them with European transplants and African slaves can in anyway be interpreted as a positive increase in diversity. This had no positive effects for two of the populations involved. It only benefited European settlers.

    I’ll explain the genocide connection and why social scientists (myself included) have to use great care in their claims. Let’s imagine for a second a hypothetical scenario. Lets say a politician, from say Ethiopia as an example, reads Ashraf and Galors papers… even better…politicians aren’t going to read an academic paper, they are probably just going to read an interview, like this one:

    http://www.voxeu.org/article/out-africa-hypothesis-human-genetic-diversity-and-comparative-economic-development

    They see the following statement: ” In particular, increasing Bolivia’s diversity to the optimum level prevalent in the US would increase Bolivia’s per capita income by a factor of 4.7, closing the income gap between the US and Bolivia from 12:1 to 2.5:1, and decreasing Ethiopia’s diversity to the optimum level of the US would increase Ethiopia’s per capita income by a factor of 1.7 and, thus, close the income gap between the US and Ethiopia from 47:1 to 27:1.”

    Now lets create a hypothetical situation and imagine, for instance, that Ethiopia is going through an economic down turn, and in recent years has had high levels of immigration. People are upset, they are jobless and they are looking for an explanation. They are offered one by a government who decides to point the finger at their immigrants and which states “Scientists have proven that in order to help our economy we need to get rid of our genetic diversity, hence immigrants”. (They are not going to read or understand any of the subtler points of the argument).

    Sound familiar, yes it is..The claims of social scientists at the beginning of the century were used to justify the Third Reich’s policy against Jews… while I do not mean to be alarmist about Ashraf and Galors article, it is important to remember how easy it is for the statements we make to be used by individuals who have an ulterior motive. There are moments in history that I, for one, hope we never forget. When we forget is when they are most likely to be repeated.

    I’ll leave you with a news article about what is going on in Greece right now…largely as a result of the economic down turn.
    http://www.nytimes.com/2012/10/01/world/europe/amid-greeces-worries-the-rise-of-right-wing-extremists.html?pagewanted=all
    I hope that Ashraf and Galor don’t recommend that Greece at all change its genetic diversity…..I’m sure the Golden Dawn would be thrilled to have scientific proof backing up their violence against immigrants.

    Henry, Natasja and others, Do you really think that as social scientists, by definition careers which deal with man, we are not responsible for dealing with the human consequences of the research statements/ recommendations we make? Should we not consider the ethical implications of the policies we recommend? Our research has normative effects and we cannot NOT consider these.

    We should always be in pursuit of the truth. Yet when we discover a “truth” which serves to rationalize why peoples who have been oppressed and who have suffered are somehow deserving of their condition because of their inherent population structure, then I think it is highly appropriate to ask if we are sure of our claims beyond any doubt. There is no being a thought policeman here, no religion, its simply asking “Is this really a reasonable conclusion to come to? Have I really examined all other alternatives???” This is the same question a medical researcher would ask about the safety of a drug before he allows it to be tested on human subjects.

    In Ashraf and Galors case, their ideas are both unproven, unfounded and have potentially worrying conclusions.

    Does anyone care to respond our criticism of the model and the data?

  12. says

    Both Ashraf and Galor and their Harvard critics are confused:

    http://isteve.blogspot.com/2012/10/the-latest-car-crash-in-trendy.html

    The key point is that population geneticists prefer to look at mutations to neutral genes that don’t have much function and thus aren’t much subject to natural selection. This has little to do with what most people think of as “diversity.”

    Thus the economists end up making odd statements like: “the most homogeneous country, Bolivia, placed at 0.63 and the most diverse country, Ethiopia, at 0.77.”

    In what economically meaningful sense is Bolivia the most genetically homogeneous country in the world?

    Even before the Spaniards’ arrival, the indigenous peoples of the highlands had evolved a genetic mutation for dealing with thin atmosphere, which the lowland indigenous peoples seldom shared. (Google “Cynthia Beall” for details of this genetic adapation.) Now, _that_ is an economically important example of genetic diversity and it already existed within the boundaries of modern Bolivia in 1491.

  13. Jade d'Alpoim Guedes says

    Hi all, please note a typo in my first sentence above. I meant to say “Ashraf and Galor are not just saying more diversity is good.” [Amendment made to your earlier comment. JC]

  14. Christina Warinner says

    In reading this blog thread, it strikes me that several of the respondents who have left comments don’t seem to have read the Ashraf and Galor article, but instead seem to be responding primarily to the Guedes et al. criticism and the Ashraf and Galor rebuttal. Having read all three documents, I feel that the Ashraf and Galor rebuttal grossly mischaracterizes the original article (and its criticism) by downplaying, softening, and even contradicting some of its bolder and more absurd claims. I hope those participating in the discussion who have not read the original article will benefit from the following three examples:

    Rebuttal Point 1:
    “This brings us to the issue of correlation vs. causality. Our critics claim that we have erroneously ascribed a causal interpretation to the statistical relationship that we uncover between economic development and diversity…The key is that the measure of intra- population genetic diversity that we employ should be interpreted as a proxy (i.e., a correlated summary measure) for diversity amongst individuals in a myriad of observable and unobservable personal traits that may be physiological, behavioral, socially-constructed, or otherwise.”

    Article Quote 1 (Abstract in full):
    “This research advances and empirically establishes the hypothesis that, in the course of the prehistoric exodus of Homo sapiens out of Africa, variation in migratory distance to various settlements across the globe affected genetic diversity and has had a persistent hump-shaped effect on comparative economic development, reflecting the trade-off between the beneficial and the detrimental effects of diversity on productivity. While the low diversity of Native American populations and the high diversity of African populations have been detrimental for the development of these regions, the intermediate levels of diversity associated with European and Asian populations have been conducive for development.”

    My Response 1:
    Contrary to their claims in the rebuttal, there is no equivocation in this abstract. The words “correlation” or “proxy” do not even appear. Instead the causal relationship is clearly implied by the strong and conclusory language: X “has had” Y effect. Finally, in fact, the word “proxy” appears only six times in their entire 109 page manuscript: twice in a literature review, three times in footnotes relating to characterization of subsistence mode, and once stating that aerial distance is a proxy for the actual distance traveled along migration routes. Not once is it used in reference to their interpretation of the genetic data in this study. Later they restate their points in more detail, again making it clear that they are drawing cause and effect relationships:

    “Consistent with the predictions of the theory, the empirical analysis finds that the level of genetic diversity within a society has a hump-shaped effect on development outcomes in the precolonial as well as in the modern era, reflecting the trade-offs between the beneficial and the detrimental effects of diversity on productivity. While the low degree of diversity among Native American populations and the high degree of diversity among African populations have been a detrimental force in the development of these regions, the intermediate levels of genetic diversity prevalent among European and Asian populations have been conducive for development.”

    The language here is very clear. Correlations do not produce effects; causes do.

    Rebuttal Point 2:
    “A careful reading of our research should make it apparent that our use of the measure of genetic diversity from the field of population genetics does not imply that our hypothesis is one of biological determinism, nor does it imply that DNA material is directly important for economic outcomes or that some genes are more important than others for economic success… In conclusion, it is apparent that all of the criticisms raised are scientifically baseless, reflecting a misunderstanding of our empirical methodology, potential unfamiliarity with the statistical techniques that we employ, and a misinterpretation of our findings. Moreover, our work should not be misconstrued as being suggestive of biological determinism and, thus, of any distressing inferences that one may derive from such potential misinterpretation.”

    Article Quote 2:
    “The direct effect of genetic diversity on contemporary income per capita, once institutional, cultural, and geographical factors are accounted for, indicates that: (i) increasing the diversity of the most homogenous country in the sample (Bolivia) by 1 percentage point would raise its income per capita in the year 2000 CE by 41 percent, (ii) decreasing the diversity of the most diverse country in the sample (Ethiopia) by 1 percentage point would raise its income per capita by 21 percent, (iii) a 1 percentage point change in genetic diversity (in either direction) at the optimum level of 0.721 (that most closely resembles the diversity level of the U.S.) would lower income per capita by 1.9 percent, (iv) increasing Bolivia’s diversity to the optimum level prevalent in the U.S. would increase Bolivia’s per capita income by a factor of 5.4, closing the income gap between the U.S. and Bolivia from a ratio of 12:1 to 2.2:1, and (v) decreasing Ethiopia’s diversity to the optimum level of the U.S. would increase Ethiopia’s per capita income by a factor of 1.7 and thus close the income gap between the U.S. and Ethiopia from a ratio of 47:1 to 27:1.”

    My Response 2:
    Again, there is no equivocation about proxy or correlation in the language above. They very clearly refer to the role of genetic diversity as the “direct effect of genetic diversity on contemporary income per capita,” and after accounting for “institutional, cultural, and geographical factors,” they provide specific examples of biological determinism, namely that a 1% increase in genetic diversity in Bolivia would result in a 41% increase in income per capita. Similar examples are provided for Ethiopia and the United States. Note, however, that these statements are based entirely on predicted genetic diversity in Bolivia, Ethiopia, and the U.S., as no actual genetic data from Bolivia (“the most homogenous country in the sample”), Ethiopia (“the most diverse country”), or the U.S. (the “optimum”) are included in the study.

    Rebuttal Point 3:
    “The fact that the measure of genetic diversity we use is based on variation across individuals in non-protein-coding regions of the genome (and, thus, in genomic characteristics that are not necessarily phenotypically expressed so as to be subject to the forces of natural selection) is clear reason why our findings should be interpreted through the lens of our measure serving as a proxy for diversity more broadly defined.”

    Article Quote 3:
    “The hypothesized channels through which genetic diversity affects aggregate productivity follow naturally from separate well-established mechanisms in the field of evolutionary biology and from experimental evidence… The benefits of genetic diversity, for instance, are highlighted in the Darwinian theory of evolution by natural selection… On the other hand, to the extent that genetic diversity is associated with a lower average degree of relatedness amongst individuals in a population, kin selection theory, which emphasizes that cooperation amongst genetically related individuals can indeed be collectively beneficial as it ultimately facilitates the propagation of shared genes to the next generation, is suggestive of the hypothesized mechanism through which diversity confers costs on aggregate productivity… According to this influential theory, the indirect fitness gains of genetic relatives can in some cases more than compensate for the private fitness loss incurred by individuals displaying altruistic or cooperative behavior.”

    My Response 3:
    In their rebuttal, the authors state that they did not argue in their article that genetic diversity is under selection. As evidence, they state that the CEPH data is from non-protein-coding regions and so doesn’t necessarily have a phenotypic expression upon which selection can act. However, in their article they very clearly argue that the “hypothesized mechanism through which diversity confers costs on aggregate productivity” is kin selection theory. They then go on to discuss the fitness and costs of genetic diversity within a framework of selection. Their positions in the rebuttal and in the article are at odds. Despite what the authors state in their rebuttal, the article contains a lengthy discussion of Darwinian evolutionary theory and kin selection theory precisely because they draw on these theories of evolutionary biology to make the link between genetic data and aggregate productivity. The authors thus present selection as the mechanism explaining the cause and effect relationship between genetic diversity and aggregate productivity.

Comments, thoughts, suggestions?