Reed pulls an Unz

Race, Realism, and Race Realists

…Let us consider brown people in Peru, a small, heavily mestizo Andean nation of some thirty million. Let us also consider the International Math Olympiad, an annual contest of high-end mathematical talent around the planet. In 2012, Peru finished 16th. Results from the Olympiad vary considerably by year: In 2013, Peru finished 26th. Australia finished 15th and 27th in those two years (and Mexico 17th in 2013). Yet it is hard for me to see how an inherently stupid people could make it to 16th. This is especially puzzling because Peru does not have the highly developed mechanisms for discovering talent that America has.

I consequently suggest that race realists, at least with respect to South America, have become more racial than realistic and may suffer from a recto-cranial inversion. I hope that Fred on Everything can serve them as salutary forceps….

My comment smackdown:

Regarding the IMO results, on the national level, a two minute google scholar search pulls up Rindermann’s (2011) sophisticated analysis. As expected, on average, Peru performs below (global) average. (Here.) The IMO results thus do not seriously disagree with the contemporaneous international test and IQ based ones varyingly reported by Hanushek and Woessmann (2010), Lynn and Vanhanen (2012), Altinok et al. (2013), etc. Not to mention other indexes of ability going back hundreds of years.


As for intra-national differences, it’s well known that self identifying Peruvian “indigenous” under-perform self identifying “non-indigenous”. Typical discussion:

The background and test score differences between indigenous and non-indigenous students give additional insight into the distinct challenges that indigenous students face. In every country, the test score gap between indigenous and non-indigenous students was greater in Spanish reading exams than in math exams, and the gaps in both subjects ranged between 0.6 and 1.1 standard deviations. (Hernandez-Zavala, M., Patrinos, H. A., & Sakellariou, C. (2006). Quality of schooling and quality of schools for indigenous students in Guatemala, Mexico and Peru (Vol. 3982). World Bank Publications.)

Presumably, self identifying indigenous (in the above sample) have more Amerindian ancestry. They do in other ones, ones which also show that education (negatively) correlates with Amer ancestry.


(From: Pereira et al. Socioeconomic and nutritional factors account for the association of gastric cancer with Amerindian ancestry in a Latin American admixed population.)

But the photos!

Reed’s Peruvian IMO Photo Proof (with composite national images for comparison)


Of the 23 Olympiads, 22 were males (shown) — a situation which confirms at least one stereotype. At least 2 of the 22 males (21 and 22) look distinctively Chinese Peruvian. Probably another over-representation. Of the remaining 20, #1-5 look more Spanish than average, #6-17 look about average, and #18-20 look more Amerindian than average. So 5/20 versus 3/20 which translates into a more-Spanish-ancestry advantage of about 1/3rd of a standard deviation. There we go.

Imaginable, Reed would judge things differently. Readers are free to offer their own opinions. Most, hopefully, would agree that this is a pretty awful method. Perhaps we could count Spanish surnames, instead.

This entry was posted in Uncategorized. Bookmark the permalink.

16 Responses to Reed pulls an Unz

  1. Reader says:

    You should email the link to this rebuttal to Jim Goad (editor of TakiMag):


  2. drew hempel says:

    Some cultures are not so much into math. It’s over-rated. As math professor Luigi Borzacchini states: “(iii) after Dedekind, Cantor, Hilbert, Zermelo, Goedel, Cohen we know that the Aristotelean and Euclidean continuum admits numerable models, that we can not
    give to its modern versions a first order categorical axiomatization, that the
    geometrical continuum can not be proved coincident with the numerical one, that
    it can not be empirically verified, that the place of the numerical continuum in the
    transfinite hierarchy is one of the greatest so far open questions, that it is linked to
    the most disputed axiom of set theory, etc.”

    Luigi Borzacchini, “Music and Incommensurability,” Historia-Matematica, August 18, 1999

    There are many basic logic paradoxes at the foundation of math that remained ignored and unsolved, created a pre-established disharmony within the discipline of mathematics. Some cultures are more logically aware of this problem of math compared to Westerners or those who favor Westernization.

  3. Curzio says:

    I think Fred Reed’s article is downright bone-headed, but your analysis is flawed too.

    You can, ahem, expect to see a full-length article expanding on this issue, but for now I’ll say:

    1. Rindermann’s IMO analysis was conducted in 2011. Peru’s performance at the IMO did not begin to take off until 2005 or 2006. Before then, the country was a mess and devoted few resources to the IMO team. In 1998, 1999, 2000, 2002, 2003, and 2004, Peru didn’t even send the maximum 6 members for its team, which further negatively affected its ranking.

    I’ll note that 2012 was Peru’s strongest performance yet, and they even edged out *Japan*. Their momentum continued in 2013, albeit with a slightly lower ranking. I make no guarantees that they’ll maintain this level of performance, but it would be interesting to reconduct an IMO analysis a decade from now with 2005 as the cut-off point for Peru.

    2. Why the cut-off point for Peru, and why the sudden upturn in their IMO performance? Well, there’s this…

    “Peru has the fastest-growing economy in Latin America, having registered substantial growth nearly every year for a decade. Where that prosperity has touched, the change is remarkable. There is a dynamism, an optimism, that did not exist in the 1980s and ’90s, periods of fratricidal guerrilla violence and divisive political turmoil.

    At the same time, Peru continues to be beset by unresolved issues of poverty and the legacy of war.

    But today, the road is smoother; a lovely chain of parks perches on the cliffs above. In the mornings in wealthy seaside districts such as Miraflores, Peruvians are in those parks, doing calisthenics and walking froufrou dogs.

    That’s the higher end of the income bracket. The tiny, white elite that always ruled Peru has only gotten richer. But in the last couple of decades, a middle class has emerged and moved solidly into business offices and government branches. Mestizo kids attend private schools and colleges where they never were before.

    Miserable shantytowns once ringed Lima, clinging to caked-dirt hillsides, void of electricity or running water, the so-called pueblos jovenes, or young towns, of cardboard homes to thousands of people who fled the Andean countryside in a massive attempt to escape war and poverty in the 1970s, ’80s and ’90s.

    Today many of the pueblos jovenes have metamorphosed into mini-cities with services and real buying power.”

    Obviously, if your country is no longer riven by guerillas and civil war, and you’re experiencing economic growth, and more of your children are attending schools, you can devote more resources toward recruiting talent and training for competitions like the IMO. So, like I said, I make no guarantees, but let’s see where they are ten years from now rather than obnoxiously make rash judgments.

    3. You are NUTS, NUTS if you think #2 looks more Spanish than average. He looks practically native.

    4. I thought #22, Ivan Muñoz Castillo, looked Asian too at first, but I located a video of him:

    And the “Asian” features are much less pronounced. Some Chinese ancestry is still possible, but it’s definitely mixed with a good heap of Indio. That kind of phenotype just doesn’t exist in East Asia.

    5. Of course, phenotype doesn’t determine genotype, especially in highly mixed populations. The only way to settle the matter would be to run DNA tests.

    • Chuck says:

      Let’s review the argument:

      (1) On the national level, Peru doesn’t underperforming as one would expect. given….. Evidence: IMO results.


      (a) Rindermann’s IMO results.

      (b) Peruvian performance in other International Olympiads:

      (c) International achievement test results: Angrist, et al. (2013). An Expansion of a Global Data Set on Educational Quality.

      (d) International IQ results: Compare Peru to Vietnam: Behrman, et al. (2013). Intergenerational Transmission of Poverty and Inequality: Young Lives.

      I will note that the Peruvian GMAT and TOEFL scores were much higher than expected…. But not GRE and English Proficiency Test scores. (It would be informative to look at ETS’ massive international SAT data base.)

      (2) Something about a lack of inter-ethnic differences.

      Two minutes of google scholar searching will cure one of this delusion.

      You said: “I’ll note that 2012 was Peru’s strongest performance yet, and they even edged out *Japan*. Their momentum continued in 2013, albeit with a slightly lower ranking.”

      Rindermann, properly, made a host of statistical corrections to create IMO indexes of national ability; he didn’t simply use relative ranking. His is the type of method which you would want to follow. See also: Henseke (2009) “Country Performance at the International Mathematical Olympiad”. Maybe just contact the authors and see if they will share data.

      You said: “You are: “NUTS, NUTS if you think #2 looks more Spanish than average. He looks practically native.”

      Honestly, I can tell.

      Whatever the case, given the highly attenuated phenotype x geneotype correlation due to hundreds of years of interbreeding and given the plausible magnitudes of differences under consideration e.g., 2/3rds of a SD in genotypic g, you would need an absolutely massive sample size to reliably determine one way or the other using this method.

      You said: “And the “Asian” features are much less pronounced.”

      Having seen the video, I can’t disagree.

      You said: “Edward Alsworth Ross’s South of Panama (published in 1915)”

      So… maybe another 99 years? You have to agree that these gaps are fairly resilient.

      (To convert % differences to d differences, try:
      =NORMSINV(%/100)-NORMSINV(%/100) )

    • Chuck says:

      “You can expect to see a full-length article expanding on this issue”

      Why don’t you see if you can procure either Rindermann’s or Henseke’s data set. You can then expand it by adding other IO values. That way you would make a contribution to our overall understanding.

      • Curzio says:

        “On the national level, Peru doesn’t underperforming as one would expect.”

        NOT my argument. I was referring solely to their IMO performance. One step at a time.

        Rindermann: I already explained why his study tells us nothing about Peru’s potential. At the time of his study, Peru had only participated at the IMO 18 times. Of those, 11 were during the period when the country was still a mess and devoted insufficient resources to their team and produced some abysmal scores which pull down their average. Only around 2005 did they begin performing at what one can assume is closer to their full potential. That’s why I said: give it a decade or so, and then try to find patterns.

        I’d include links, but I don’t want this comment lost in the spam filter. All of this info is easily accessible on the IMO website. It’s stupid to assume that all developing countries start from the same place and have the same resources available to them needed to perform at their best.

        Take Thailand for example. For almost a *decade*, Thailand also posted nothing but abysmal/mediocre scores, *about as bad as Peru’s were during their initial 11 years*, but then their performance gradually improved and now they frequently place in the Top 15. Also remember: Thailand has a much larger population than Peru (without checking, I think it’s around 60-70 million), and that over 50% of the Thai population has some Chinese ancestry, so even that’s not a perfect apples to apples comparison.

        You can see a similar gradual improvement in the scores of Indonesia, though I’ll point out that on average Peru has produced better results in the last 7-8 years with a country of 30 million than Indonesia has with a nation of 230 million, and Indonesia has a Chinese minority as well. Also see: the gradual improvement of Malaysia’s scores. Meanwhile, countries like the Philippines (99 million and again with a Chinese minority) continue to post far worse results. And Peru is performing better than Spain itself EVER has at the IMO. Ditto Argentina. Spain has yet to even win a single gold medal.

        Innate talent is necessary but not sufficient to perform well at the IMO — the right institutions to recruit the best possible talent and to train the team are also required — otherwise we would see no upward trends in the scores of certain developing nations. A country like Peru is just beginning to build the institutions required to prepare adequately for these competitions. It won’t happen overnight. That’s why I keep saying: keep an open mind and don’t jump to conclusions.

        With regard to PISA, I’ll point out that while Peru’s score in 2012 was low, there was a significant improvement from 2002, from an overall score of 292 to 368. Personally, I dislike how PISA is issued every 3 years. I just don’t expect substantial improvements in a mere 3 years. Holding it every 10 years would be much more interesting and productive. Nevertheless, it’s not true that Peru’s score has remained static over the last decade. Has it maxed out? Maybe, maybe not. I make no claims in either direction. The underachievement of a country like Argentina on the PISA also makes me suspicious that there aren’t other factors at play.

        “So… maybe another 99 years? You have to agree that these gaps are fairly resilient.

        I’m looking at your graph alone, and I don’t know what ABCC is, so maybe I’m missing something that additional context would provide, but…numeracy comparisons from the seventeenth/eighteenth century? What exactly would that prove? Most Indios were neither encouraged nor expected to be formally educated under the caste system. We definitely don’t have IQ data from that period so this seems based on rather weak extrapolations. Unless I’m missing something here.

        “Why don’t you see if you can procure either Rindermann’s or Henseke’s data set. You can then expand it by adding other IO values. That way you would make a contribution to our overall understanding.”

        I’ve already told you repeatedly above why this means nothing because the data we have available now is too meager, and Peru has only recently become competitive at these olympiads. If I had told you back in the 2005 that Peru’s IMO team would be competitive with, much less place ahead of, countries like Germany, the United Kingdom, France, Canada, Italy, Australia, India, etc. — countries with every conceivable advantage — you would’ve laughed me out of the room and justifiably so. I would’ve had NO data, none whatsoever, to support my assertion because Peru had never done so. Never. But now that is not true. Peru’s team has placed ahead of all of those countries — multiple times, in some cases — and their scores have consistently hovered around a much higher range than they did in the years prior to 2005.

        Likewise, if I told you back in 2000 that Thailand’s IMO team would place in the Top 10 on numerous occasions, you would’ve had a similar reaction. Looking at Thailand’s scores from 1989-2000, you would conclude that with 60-70 million, they were a nation of semi-retards. Even from 2000-2005, they only gradually improved. Now they have a truly formidable team that places consistently in the Top 15. And we’re talking about a time frame of less than 15 years in Thailand’s case, less than 10 years in Peru’s case. These are very very very recent developments. I have no idea why you think these scores can’t improve — when they *have* shown drastic improvement in a small time frame — and that the data we have now is set in stone.

        It’s generally agreed that the IMO is by far the most difficult and “g-loaded” of all the science olympiads. The others require innate talent, but are even more amenable to improvement through training than the IMO. If Peru can do well at the IMO, they can do well at the others if they apply their resources properly. Looking at the Chemistry Olympiad, they didn’t even send the maximum number of participants there. As for the other olympiads, they haven’t even participated yet in most of them, which again signals that their institutions are still developing. I suspect they’re focusing on the IMO because everyone know’s that it’s the most prestigious.

        For the last time, in case my repetition hasn’t driven this home: I’d have no problem with another Rindermann-type IMO analysis *in other 10 years or so.* Then we can see if Peru maintains its current level, or if it increases or decreases. Until then: too soon to say.

        I don’t see why Peru is a priority anyway. The First World is not getting flooded with Peruvian immigrants and their population is only around 30 million, with no drastic increase in sight. I don’t see the harm in…simply keeping an open mind and letting them develop however they may. Why not focus your efforts on far more populated areas of the world that haven’t been studied as much, like Southern Asia? India, Pakistan, Bangladesh, Indonesia, the Philippines, Malaysia, etc. I think this region of the world poses the greatest demographic problem after sub-Saharan Africa and merits more attention than Latin America, especially given the large immigrant populations from Pakistan and Bangladesh in places like England and Greece, and immigration policies allowing family reunification, etc.

        • Chuck says:

          “With regard to PISA, I’ll point out that while Peru’s score in 2012 was low, there was a significant improvement from 2002, from an overall score of 292 to 368…Holding it every 10 years would be much more interesting and productive.”

          The international testing results for Peru go back to 1997. Maybe compare Peru to Brazil in 1997 and in 2012 to see if there’s been relative improvement.

          “It’s generally agreed that the IMO is by far the most difficult and “g-loaded” of all the science olympiads.”

          I would imagine that IOs loose their g-loadedness over participation time. A sort of test-retest effect. Wild imagination, maybe.

          “NOT my argument. I was referring solely to their IMO performance. One step at a time.”

          I’m a big fan of such incremental approaches. I absolutely abhor long winded arguments which leave nothing determined.

          That said…

          “Rindermann: I already explained why his study tells us nothing about Peru’s potential.”

          I said: ‘His is the type of method which you would want to follow’ — meaning: If you wish to use IMO as an index of national ability and to make an argument that the Peruvian national ability is increasing then use R’s method to transform IMO rankings into national ability indexes.

          “I’m looking at your graph alone, and I don’t know what ABCC is, so maybe I’m missing something that additional context would provide, but…”

          “What exactly would that prove? Most Indios were neither encouraged nor expected to be formally educated under the caste system.”

          age heaping is an index of numeracy; the r (NIQ/PISA x Age heaping) is significantly higher than the r( NIQ/PISA x IMO) (0.70 to 0.85 versus 0.50) — even when the age heaping data is from hundreds of years ago! Insofar as IMO can be taken as a measure of national (and international) CA, age heaping can also. What is proved then is what I said: “these gaps are fairly resilient.” Nothing more. You can always argue that they will rapidly disappear in the near future, but the bayesian probability, based on hundreds of years of data, is now low.

          Of course, NIQ doesn’t determine IMO, just as IQ doesn’t determine reasoning ability. Some peoples and people make more efficient use of their limited raw ability. It’s how I win these arguments, when I do.

      • Curzio says:

        I probably should’ve just linked this in the first place:

        But yeah, the graph at that link clearly illustrates the upward trend in Peru’s scores. The question is: can Peru maintain that level of performance? I’m willing to keep an open mind.

      • Curzio says:

        And for comparison’s sake, here is Thailand’s chart:

        See the similar upward trend with some hiccups along the way, not unlike Peru. Now I’m not saying Peru will match Thailand’s peaks by scoring in the Top 10, but keep in mind the differences I noted above: Thailand’s much larger population + over 50% with some Chinese ancestry.

  4. Curzio says:

    By the way, an interesting observation about Peruvians in particular. This is from Edward Alsworth Ross’s South of Panama (published in 1915):

    “There is a widespread conviction that mixed breeds lack nervous stability, and Houston Chamberlain attributes the proverbial lack of character in the tropical South Americans to recent race mixture. In Lima I talked with a German educator, a shrewdly critical man of science.

    “I came out here,’ he said, ”eight years ago in the firm conviction of the racial inferiority of these peoples to the Germanic peoples. I had read Chamberlain, and I looked upon them as hopeless mongrels. But I have faced about completely. The faults of these Peruvitos root in historical conditions, and can be eradicated. There is nothing wrong with the breed. They have capacity, but they lack the tradition of hard work. The spirit of their past has been one of self-indulgence. What they need is right education and discipline. Even now these Peruvians turn off good work when their pay is adequate and certain. They have, to be sure, a juvenile love of impressing, but the ability is there.”

    “Don’t forget, either, that climate is a handicap, and that, after a few years here, the Anglo-Saxons, too, show less energy and force of character.””

    So, yeah, maybe that indolence is genetic and impairs them, but there must have been some latent talent there to engender such a comment from a source like that. It’s hard to find similar remarks about Negro intelligence before the era of political correctness except from die-hard Marxists/egalitarians or people who were Black/Mulatto themselves.

  5. Julian says:

    What’s your take on Greg Clark’s reported data of African immigrants to the US? Elijah Armstrong’s post:

    “Gregory Clark’s new book reports data on the social standing of African immigrants to America. They are economically elite; conservatively, one might estimate an IQ of 105 from their social standing.

    The social elite in Africa (when tested in Africa) has an IQ of about 80. If the average African who emigrates to America is drawn from this social elite, this implies a 25-point environmentally caused difference between Africans and Europeans. Assuming an IQ of 70 for the average African, this yields an African genotypic IQ of 95.

    Of course, this is a high figure and should be taken with a grain of salt. But it does support Eysenck’s conjecture that African-Americans have lower levels of g than equatorial Africans.”

    • Chuck says:

      I left a comment. In a number of scattered posts, I discussed this issue a while back. It would be nice if you could help me conduct a meta-analysis on migrant Black IQ. Why not contribute to our understanding instead of remaining a spectator?

  6. Curzio says:

    Wow, after doing more research, I’m more convinced than ever that people like you are full of shit.

    “Insofar as IMO can be taken as a measure of national (and international) CA, age heaping can also.”

    Except that’s bullshit. Age heaping does NOT reflect a cognitive/intelligence/IQ gap. It’s so obvious you’re grasping for anything that will confirm your biases. The authors of the paper you cited never suggest that. Never. Nor is numeracy a proxy for IQ, any more than literacy is a proxy for IQ.

    [Remainder of Comment Deleted. Baten and Juif (2013) conducted a similar analysis:

    Our equation for human capital determinants also includes the ABCC-Index values for 1820 – the earliest year for which we obtained a sufficiently large set of country data – to assess the path-dependence of education reflected in numeracy skills (see also Appendix E). A considerable number of recent studies have used this innovative proxy of basic numeracy, which
    was introduced by A’Hearn et al. (2009) and is based on the ‘‘age-heaping’’ technique. The age-heaping phenomenon applies
    to historical populations (as well as people in the poorest countries today) when a substantial share of the people are not
    able to state their exact age and therefore reported a rounded age, such as ‘‘I am 30’’, when they are in fact 29 or 31. The
    ABCC-Index reflects the share of people who were able to state an exact age (or the degree to which the distribution of
    age statements approaches an equal distribution). Crayen and Baten (2010) showed that this proxy of early human capital
    has a strong correlation with other measures such as literacy and schooling…

    …Fig. 3 illustrates the strong correlation between the 1820s measure of numerical skills and the Hanushek-Woessmann
    measure for the late 20th century. The dependent variable cognitive is our measure for the cognitive skills of human capital.
    This output measure of human capital reflects the knowledge and abilities that are most favorable for subsequent success.
    Hanushek and Woessmann (2012a) constructed this variable from test scores in mathematics and science for 77 countries
    between 1964 and 2003.
    More specifically, they calculated the simple average of all math and science scores on International Student Achievement Tests conducted by the Organization for Economic Co-Operation and Development (OECD) and the
    International Association for the Evaluation of Educational Achievement (IEA) during that period. With linear regressions,
    they found that by adding cognitive skills to a growth model with school attainment as a dependent variable, the model explains three quarters of the variance in growth rates (instead of one quarter if only school years are included). Further, the coefficient for school years turns statistically insignificant in the presence of cognitive skills
    (A story of large landowners and math skills: Inequality and human capital formation in long-run development, 1820–2000. Journal of Comparative Economics.)

    See also Baten and Prayon:

    Crayen and Baten (2008) found that the relationship between illiteracy and age heaping for Less Developed Countries (LDCs) after 1950 is very close….. The correlation coefficient with illiteracy was as high as 0.7. The correlation with the PISA (Programme for International Student Assessment) results for numerical skills was even as high as 0.85, hence the age heaping measure Whipple Index is more strongly correlated with numerical skills.

    In short, age heaping is a good index of cognitive skills. If you bothered to do your homework, you would have known this,]

  7. Curzio says:

    And you know, I see this kind of motivated reasoning from HBD “researchers” all the damn time. It’s why I stopped taking you guys seriously.

    [If you’re not going to take my discussions semi-seriously, then I’m not going to waste my time reading your comments and replying to them.]

  8. Curzio says:

    [psychotic comment deleted]

    • Chuck says:


      Are you off of your meds, again?

      I have been deleting your comments because they contain overly rude, crude, or psychotic phrases.

      If you wish to discuss an empirical matter, you can email me.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s