Links this week:

1. Andrew Berry guest blogs at Why Evolution is True on Alfred Russel Wallace’s unfortunate end to his Amazon expedition. The equivalent of accidentally erasing your PhD thesis the day before submission?
2. I am enjoying some of the press coming out about Marlene Zuk’s new book Paleofantasy: What Evolution Really Tells Us about Sex, Diet, and How We Live (and the reactions to it); this week an interview in the Los Angeles Review of Books.
3. An interview with Gary Becker on rationality, behavioural economics and the use of mathematics in economics.
4. A profile of Ingela Alger, another economist looking to bring some evolutionary biology into the picture.

In two recent posts I examined the causative mechanisms underlying Ashraf and Galor’s hypothesis linking genetic diversity to economic growth (innovation and conflict). In those posts, I avoided examining whether genetic diversity could be considered a proxy of phenotypic diversity unrelated to that genetic diversity (such as language).

Part of the reason for this is that Ashraf and Galor do not indicate in their paper or web appendix that they intended to use genetic diversity as a proxy in this way. As I have posted about before, the language of the paper is focused on genetic diversity and the phenotypic expression of that genetic diversity. In dissecting the paper, I wanted to focus on what the paper states.

However, given that Ashraf and Galor have now made an argument [Update: the response is no longer online] that genetic diversity is a proxy in their response to a critique of their paper, the argument is worth assessing. Ashraf and Galor write:

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. …

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.

The more relevant question to ask therefore is to what extent the measure we use can reasonably be considered a proxy for diversity in unobserved phenotypic or socially-constructed characteristics. There is indeed an emerging body of scientific evidence that establishes remarkable correlations in this regard.

Ashraf and Galor refer to two articles on this point – one on diversity in head shape, which I noted in my post on innovation, and a second diversity in language. On the second, they cite a Science paper from 2011 in which Atkinson reports a finding that diversity in phonemes – perceptually distinct units of sound that differentiate words – declines with distance from Africa. This pattern reflects that found for genetic diversity, and Atkinson suggests that similar forces were acting on each. There are still significant hurdles to show a causative link between diversity in phonemes and innovation and conflict, but the persistence of the phenotypic diversity leaves opens this possibility. The task is to identify what forms of phenotypic diversity might be relevant.

Ashraf and Galor build the case further in a new paper that will be published in the American Economic Review Proceedings and Papers in May. They propose that:

Building on the role of deeply-rooted biogeographical forces in comparative development, this research empirically demonstrates that genetic diversity, predominantly determined during the prehistoric “out of Africa” migration of humans, is an underlying cause of various existing manifestations of ethnolinguistic heterogeneity.

Rather than proposing that genetic diversity is a proxy for other diversity shaped by the Out of Africa event, they propose that genetic diversity is a cause of ethnolinguistic heterogeneity. Depending on whether they use a modern global or old world sample, they find that genetic diversity is responsible for between 7 and 11 per cent of ethnolinguistic heterogeneity.

An alternative approach would have been to attribute genetic and ethnolinguistic diversity to a common cause in the founder effect. I asked Ashraf and Galor by email why they preferred an explanation of genetic diversity causing ethnolinguistic heterogeneity and they replied as follows:

[T]he ethnicities located at greater migratory distances from East Africa (especially those outside of the African continent) DO NOT represent a subset of the ethnicities in Africa. Had the serial founder model (as applicable in generating the global distribution of genetic diversity) been equally applicable in explaining global spatial variation in ethnic diversity, we should observe that ethnic groups extant outside of Africa are also present within Africa. This is clearly NOT the case.

Thus, our hypothesis is that of quasi-random migrant selection from the origin in each step of the “out of Africa” demic diffusion process, with the migrants engaging in endogenous group selection (or endogenous sorting) upon reaching their destination in that step of the diffusion. Moreover, this group selection process would take into account the trade-off associated with intragroup diversity (i.e., diversity across individuals WITHIN the new group), and possibly, also reflect the interaction of intragroup diversity with location-specific geographical factors.

They further expand on the mechanism in their paper:

Following the “out of Africa” migration, the initial level of genetic diversity in indigenous settlements presumably facilitated the formation of distinct ethnic groups through a process of endogenous group selection, based on the tradeoff between the costs and benefits associated with heterogeneity and scale. While heterogeneity raised the likelihood of disarray and mistrust, reducing cooperation and thus adversely affecting group-specific productivity, complementarities across diverse productive traits and preferences stimulated productivity. Since in a given environment, diminishing marginal returns to diversity and homogeneity entail an optimal size for each group, higher initial genetic diversity would have positively contributed to the number of groups, and thus to the degree of fractionalization. Further, to the extent that higher initial diversity did not lead to an excessively large number of groups, it would have positively contributed to the degree of polarization as well.

The causative argument relies in part on the conflict between genetically dissimilar individuals limiting group size, which would in turn affect ethnic fractionalisation, which would then affect inter-group conflict. I need to think about this argument more, but one possible implication of this group-sorting argument is that groups with less genetic diversity would become larger. Part of the benefit to lower diversity would be to enable a scale effect – that is, more people leading to more ideas. A larger group would be more innovative simply through having more innovators.

So, to answer the question of whether genetic diversity can be a proxy for phenotypic diversity beyond phenotypic expression of that genetic diversity, yes. And as a plausible causative link may exist between ethnolinguistic heterogeneity and conflict, this builds the case for the link between genetic diversity and conflict. But, as their new paper suggests, Ashraf and Galor propose a more direct relationship than through a common cause in the founder effect, at least as it relates to ethnolinguistic diversity. Rather than pursuing the proxy argument, their new paper builds the case that the relationship between genetic diversity and conflict is more direct.

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? (this post)
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?

Earlier debate on this paper can also be found here, here and here.

Being a naturalist in the Victorian era was a different exercise to today. From Darwin’s The Descent of Man:

Many kinds of monkeys have a strong taste for tea, coffee, and spiritous liquors: they will also, as I have myself seen, smoke tobacco with pleasure. (6. The same tastes are common to some animals much lower in the scale. Mr. A. Nichols informs me that he kept in Queensland, in Australia, three individuals of the Phaseolarctus cinereus [koalas]; and that, without having been taught in any way, they acquired a strong taste for rum, and for smoking tobacco.) Brehm asserts that the natives of north-eastern Africa catch the wild baboons by exposing vessels with strong beer, by which they are made drunk. He has seen some of these animals, which he kept in confinement, in this state; and he gives a laughable account of their behaviour and strange grimaces. On the following morning they were very cross and dismal; they held their aching heads with both hands, and wore a most pitiable expression: when beer or wine was offered them, they turned away with disgust, but relished the juice of lemons. An American monkey, an Ateles, after getting drunk on brandy, would never touch it again, and thus was wiser than many men. These trifling facts prove how similar the nerves of taste must be in monkeys and man, and how similarly their whole nervous system is affected.

Links this week:

1. A good profile of Joe Henrich and WEIRD people.
2. Nassim Taleb and Daniel Kahneman discuss antifragility.
3. A gene for cooperation in response to punishment.
4. I’ve uploaded an updated version of my working paper Conspicuous Consumption, Sexual Selection and Economic Growth. There are no earth-shattering changes, but we’ve simplified the second model and generally sharpened it up.

Should Ashraf and Galor’s paper The ‘Out of Africa’ Hypothesis, Human Genetic Diversity, and Comparative Economic Development have been published? I planned to place this post last in my sequence of posts on the paper, but some comments in response to my earlier posts (links at the bottom of this page) have encouraged me to discuss this question now.

The major critique about whether the paper should have been published comes from the conclusion of an article in Current Anthropology:

Social scientists seeking to explain economic behavior through genetics must exercise particular caution. As Benjamin et al. (2012:656) point out, “researchers in this field hold a special responsibility to try to accurately inform the media and the public about the limitations of the science,” especially in studies intended for “social-scientific interventions” (Benjamin 2010:1). Without proper methodology and data analysis standards, false positives are likely to be misunderstood as facts, and these can then be mobilized in the political arena. Ashraf and Galor’s (2013) paper is based on a fundamental scientific misunderstanding, bad data, poor methodology, and an uncritical theoretical framework. While the attempt to create interdisciplinary studies that link anthropology, genetics, and economics is laudable, economists should consult with specialists in those fields to avoid making such uninformed blunders. The same should be true of the peer-review process for such interdisciplinary articles.

More egregiously, this study has the potential to cause serious harm. By claiming a causal link between the degree of genetic heterogeneity and economic development, their thesis could be interpreted to suggest that increasing or decreasing a nation’s genetic (or ethnic) diversity would promote prosperity. Ultimately, this can provide fodder to those looking to justify policies ranging from mistreatment of immigrants to ethnic cleansing (especially by groups with real political power, e.g., Golden Dawn in Greece).

We are not concerned here with the authors’ own social or political attitudes. Rather, we wish to emphasize the irresponsibility of bad science. In the social sciences, scientific methods are an extremely powerful tool for analyzing trends in an empirically demonstrable manner and thus have the important opportunity to guide political action. When used improperly or when it is of dubious quality, however, science can become a justification for reactionary policy. The dismal nature of economics is often appealed to when facts contradict a desired reality. However, we are not arguing a case for blissful ignorance. What we see in Ashraf and Galor’s study is the worst of all worlds: something false and undesirable.

This conclusion has two major threads. First is the responsibility of the authors to consult with specialists in other fields and for the peer review process to do likewise. I don’t know who reviewed this article for the American Economic Review, but Ashraf and Galor acknowledge five anonymous referees. The acknowledgements section of the paper also lists participants from 33 seminars, 13 conferences and 3 lectures. They list 24 people by name, including the population geneticist whose data they use. The list does not include any anthropologists, but this is not through hiding the paper. They posted the paper on IDEAS in May 2010, on the Social Science Research Network in January 2008 and again in May 2011 (with regular updates to those submissions), and as an NBER working paper in July 2011. If anything, this paper epitomises open peer review.

The issue seems to be that it was not noticed by some people until it was mentioned in Science late last year. If not for the Science mention, it may have been published without any fuss. This is indicative of a broader problem that I am not sure how to solve. In working on a paper of this nature, how do you get it seen by everyone who might object and want to comment? My own experience of seeking a broad variety of feedback on my working papers is that it is not easy. Everyone is busy and has limited time. It is only when accepted by a major journal that people become interested, by which time it is too late. (On that note, here is my SSRN page and feel free to critique).

The second argument by the authors of the Current Anthropology piece concerns the potential for harm. But for this argument to hold much weight, we need to assume that those who would misuse such research are not gathering evidence from other sources. A quick google search will show the depth and breadth of discussion on the topic of “human biodiversity”. Much of the discussion is serious and data rich (some of it is horrible). Barring work that implies genetic differences between populations does not stop these discussions. All it prevents is the ability to examine openly whether the hypotheses have merit and to build evidence against those that don’t. As an example, research by Arthur Jensen, who faced similar claims about the responsibility of his research, triggered James Flynn’s important work on IQ. As Flynn wrote at the end of his recent book:

Psychologists should thank Jensen for pursuing his life-long mission, against great odds, to clarify the concept of g. In addition to intellectual eminence, he had the courage to face down opposition often political rather than scientific. If I have made a significant contribution to the literature, virtually every endeavor was in response to a problem set by Arthur Jensen.

A further issue arises where the work is empirically strong. Should it then be published? If you base your defence against unethical use of the hypothesis on the hypothesis not being true, your defence falls away as soon as the evidence mounts in support of the hypothesis. An ethical framework of human equality or liberty, or something of that nature, is a stronger place to build the barricades if you are worried about the ethical implications of research such as this.

As a reading of my other posts on this paper suggest, I am not convinced that the conclusions drawn by Ashraf and Galor are correct. But I am glad it was published. During the three years in which Ashraf and Galor presented the paper and sought feedback, you could see the econometric analysis evolve and be strengthened. They take the hypothesis seriously, and we should treat the empirical result with some respect. If their robust statistical result is not evidence of a causative link, why? This paper provides a basis for further thinking and research. Personally, the process of going through this paper and the debate surrounding it has been (and is continuing to be) an important learning experience.

Scientific progress is not made by publishing the perfect answer the first time. We should not see peer-reviewed economics articles as truth handed down from above (if you do, you would become confused very quickly). A serious effort can provide a springboard for other work, even if you consider it flawed. In that sense, Ashraf and Galor’s work is a springboard that we should make use of, rather than complaining about whether it should be published at all.

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? (this post)

Earlier debate on this paper can also be found here, here, here and here.

James Flynn of Flynn effect fame has a relatively new book out, Are We Getting Smarter? I have found Flynn’s earlier books to be easy but not great reads, and this book followed that pattern. However, reading them is worthwhile as they tend to provide a comprehensive update on the latest in IQ testing from around the globe. Flynn is also not afraid to throw in some interesting arguments.

The question in the title of the book has two elements. First, does the Flynn effect mean that we are getting smarter? Flynn prefers to say that we are not necessarily smarter, but that we are more modern. We are born with the same mental hardware, but in a more complex world humans are becoming more “scientific” in their thinking. We are better able to characterise and abstract.

The second is whether this trend is continuing, and generally it is. Except for the Scandinavian countries, IQ is still going up in the United States, England, Germany and other developed countries. IQ is also increasing at a slightly faster rate in developing countries, but not fast enough to close the gap with developed countries in the near future. There are also some notable exceptions, such as Sudan.

One of the more interesting chapters is Flynn’s discussion of IQ testing of death row inmates. If an inmate has a tested IQ of 65 or below, they are spared execution. But consider two inmates of the same intelligence who were both tested for intelligence in 1975, but one was tested using a 1972 normed test, while the other completed a test normed in 1947-48 (tests need to be normed regularly because of the Flynn effect). Given the different dates on which the tests were normed, the former can have their IQ measured at 65 and be spared, while the latter would be measured at 73 and face the death penalty. The view of courts on this point appears mixed. More broadly, however, it hints at an important discipline in considering IQ test results. If you are confronted with an IQ test result, you should ask when was the test normed and when was the test taken. It is only in that context that the result can be meaningful.

Flynn spends some time revisiting old debates with Arthur Jensen about whether the Flynn effect is measuring anything meaningful. In particular, Flynn accuses Jensen of psychometric obsessions by only caring about whether the Flynn effect is relevant to g, and not about whether there are any real world implications.

The relative cognitive complexity of the tasks (or their relative g-loadings) is beside the point. If you do not care about anything but finding an absolute measure of our ability to deal with cognitive complexity, an absolute measure of intelligence if you will, you will not be interested. Since you cannot correlate IQ gains with g, you dismiss them as “hollow” (Jensen, 1998). But that is only because you have been blinded to social significance by psychometric obsessions.

Despite their disagreements, however, Flynn credits Jensen with triggering his interest in the area, and dedicates the book to him. Flynn writes:

Psychologists should thank Jensen for pursuing his life-long mission, against great odds, to clarify the concept of g. In addition to intellectual eminence, he had the courage to face down opposition often political rather than scientific. If I have made a significant contribution to the literature, virtually every endeavor was in response to a problem set by Arthur Jensen.

The book also contains come interesting ideas about of the growing gap in language between parents and children (the problem teenage years) and the rise of single motherhood among black women in the absence of eligible men. Flynn also explores what he calls the bright tax and bright bonus – the relatively steeper decline in analytical ability but slower decline in verbal ability as those of high intelligence age (which I have posted about before).

Links this week:

1. A week ago I posted about a piece in the Economist on Greg Clark’s work using surnames to estimate social mobility. A debate followed the Economist article, with contributions by Miles Corak, Fransicso Ferreira, Greg Clark (and again), and Jason Long. All are worth reading.
2. Jonathan Last notes three issues that he would include in a second edition of What to Expect When No One’s Expecting: America’s Coming Demographic Disaster. The third relates to my recent working paper with Oliver Richards on natural selection driving an increase in fertility.
3. The anthropology wars are in full swing at the moment, moving from Diamond to Chagnon. The two most interesting articles I read this week were by Alice Dreger in the Atlantic and John Horgan in Scientific American. I’d be interested in hearing a response from the gang of five to Horgan’s claims.
4. Half a year late, I have finally read this article by Dierdre McCloskey on the measurement of happiness. Its full of straw men, but many passages are fantastic.

Ashraf and Galor’s hypothesis linking genetic diversity to economic growth has two limbs. The first, which I posted about last week, is that genetic diversity pushes out the production possibility frontier through increasing the range of traits in the population for developing and implementing new technologies. The second, the subject of today’s post, is that genetic diversity decreases trust and cooperation between people, increasing social disorder and conflict.

The measure of genetic diversity used by Ashraf and Galor is expected heterozygosity, which is a measure of the probability that two people selected from the population will have the same allele (variety of a gene), averaged across all measured genes. Genetic diversity is often confused with genetic distance, a measure of the time since two populations had a common ancestor. Genetic distance can be calculated using data of the type used by Ashraf and Galor, with that measure being the probability that two alleles from a given genetic locus selected from two different populations will be different. Genetic distance has a resemblance to the genetic diversity measure, but is across populations, not within them. Spolaore and Wacziarg used genetic distance in their 2009 paper where they proposed that genetic distance (or other measures proxied by genetic distance) hindered the transfer of technology, leading to technological and income differences between countries.

Genetic diversity should also be distinguished from relatedness. Relatedness is the genetic similarity of two individuals, relative to average similarity of all individuals in the population. Since relatedness is measured relative to average similarity, relatedness does not increase in a less genetically diverse population. However, the average similarity between population members is higher where diversity is low.

This distinction is important, as Ashraf and Galor’s argument for the causative pathway for genetic diversity and its effect on cooperation is via relatedness. They state:

[T]o the extent that genetic diversity is associated with a lower average degree of relatedness among individuals in a population, kin selection theory, which emphasizes that cooperation among 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.

It may be possible to craft an argument that relatedness is higher in a less genetically diverse population if you considered the relevant population for measuring relatedness to be the global population. Across the global population, two individuals from a less diverse sub-population would have a relatedness marginally above zero. However, this is a stretch. A more feasible argument would be to take Ashraf and Galor’s use of the term relatedness to refer to genetic similarity.

Ashraf and Galor dedicate little time to building the evolutionary basis to their argument in the main paper, but give it some focus in the web appendix. They note a study in which long tailed tits (a bird) provided breeding support to kin, and an analysis of 18 vertebrate species that found a strong correlation between brood rearing assistance and relatedness. They also describe a study in which juvenile spiders cooperate with kin while feeding to increase feeding efficiency.

The disconnect between these studies and Ashraf and Galor’s argument is the degree of relatedness involved. The studies referenced in the web appendix involve relatively close kin, with relatedness a relative measure within the population. For the genetic diversity hypothesis to hold, humans would need a very fine tuned sense of relatedness in the broad sense noted above. Are people more likely to cooperate with those who they are more genetically similar, despite no immediate reference group for comparing that similarity? The problem is that, beyond close kin, there is poor empirical support for kin recognition.

One possible angle in support of Ashraf and Galor’s hypothesis might be to use evidence of the detection of genetic similarity and heterozygosity in mating decisions (such as here, here and here). However, the mating preferences are usually for more dissimilar or heterozygous individuals, suggesting diversity of this nature has a positive effect.

At the close of the article, Ashraf and Galor seek to build their case by using their genetic diversity dataset to examine whether genetic diversity affects trust. Across 58 countries for which a measure of trust can be gleaned from the World Values Survey, they found a significant relationship in the required direction. Increasing genetic diversity by 1 percentage point is associated with a 2 percentage point decrease in the prevalence of trust. In obtaining this result they included controls for geography, OPEC, and sub-Saharan Africa, and used continent fixed effects, meaning that the effect is within continents.

As was the case for the regressions of genetic diversity on measures of scientific output, I find this result unconvincing. The range of controls used and the question of whether there are other relevant variables (such as IQ), combined with the lack of causative pathway, leaves me needing much more evidence.

So, where to from here? As I noted in the discussion of innovation, cross-species comparison is one potential avenue for further research, as is examination of isolated human populations. Inbred human populations might also be an interesting source of evidence, although it can be difficult to separate relatedness, diversity and inbreeding effects. One consideration is that genetic diversity in particularly low in humans relative to other species due to some bottlenecks in our past. Could that low diversity be linked to the generally high levels of cooperative behaviour in humans?

As was the case for my post on the effect of genetic diversity on innovation, I have deliberately avoided the question of whether Ashraf and Galor were directly relating genetic diversity to economic development, or whether genetic diversity is a proxy for phenotypic diversity unrelated to that genetic diversity (such as language). As you can see below, my thoughts on that point are forthcoming.

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? (this post)
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?

Earlier debate on this paper can also be found here, here, here and here.

I am re-reading Fisher’s The Genetical Theory of Natural Selection and was reminded of this passage that predates modern economic arguments about the evolution of the rate of time preference by over 50 years. For those who want to follow the maths, m is the Malthusian parameter (the relative rate of increase or decrease of a population), l_x is the number living to age x, and b_x is the rate of reproduction at age x:

In view of the close analogy between the growth of a population supposed to follow the law of geometric increase, and the growth of capital invested at compound interest, it is worth noting that if we regard the birth of a child as the loaning to him of a life, and the birth of his offspring as a subsequent repayment of the debt, the method by which m is calculated shows that it is equivalent to answering the question—At what rate of interest are the repayments the just equivalent of the loan? For the unit investment has an expectation of a return l_xb_xdx in the time interval dx, and the present value of this repayment, if m is the rate of interest, is e^-mxl_xb_xdx; consequently the Malthusian parameter of population increase is the rate of interest at which the present value of the births of offspring to be expected is equal to unity at the date of birth of their parent. The actual values of the parameter of population increase, even in sparsely populated dominions, do not, however, seem to approach in magnitude the rates of interest earned by money, and negative rates of interest are, I suppose, unknown to commerce.

Fisher’s result that time preference should reflect the population growth rate matches that of Hansson and Stuart, about which I have posted before. Fisher also notes that this outcome is not what we see. This is where we need to call on other ideas, such as aggregate risk.

Links this week:

1. Matt Zimmerman of Biased Transmission reviews Ashraf and Galor’s theory of genetic diversity and economic development. He points out that the hypothesis can explain any observed pattern in the data. I recommend subscribing to Matt’s feed.
2. Jason Antrosio takes on Jared Diamond’s arguments about violence in hunter-gatherer societies. An excellent read.
3. David Sloan Wilson provides another critique of a straw man version of the invisible hand. The interesting aspect of this critique, as for many other of Wilson’s takes on economics, is that the group selection framework he wants to bring into economics doesn’t even have the evolutionary biologists onside.
4. Geoffrey Miller presents on sexual selection and runway consumerism.
5. The Santa Fe Institute MOOC on complexity has kicked off.