The Fluidity of Race

Emily Nix and Nancy Qian just put out a paper – The Fluidity of Race – that has gotten some attention. They claim (based on their analysis of US Census records from 1880-1940) that at least 19% of black males ‘passed’ for white during this period, with about 10% switching back.  And this wasn’t a one-time thing: it kept happening for at least several generations, so there was a continuing net black-to-white flow, about 20% of each generation! They don’t talk about women, since their surnames change, but presumably there would be at least some race-switching among black women as well.

So let’s say that 17% of black males permanently passed over into the white category. During the time in question,  the black percentage of the US population was around 11%;  1 13.1% in 1880, 11.6% in 1900,  with a low in 1930 of 9.7% due to lots of recent immigration from Europe.

I think we can assume that half of the black population was male, at least until someone publishes claims of long-secret, industrial-scale parthenogenesis.

The period in questions covers about two generations.  So:

the fraction of the population called white should have absorbed  about

2 generations x 17% of the black male population x 0.5 (male fraction) x 11% (black fraction of the population) = 1.87% .  The white population during this period was about 89% of the population.    So you’d expect that whites in this country, on average would have about 2% black ancestry.  Or maybe less, since blacks average about 75% African ancestry: more like 1.5% African ancestry.

But they don’t : the average amount of African ancestry among self-labeled whites is , according to a recent, massive 23andme study,  0.19%.  The majority don’t have any African ancestry at all. 0.19% is way less, at least ten times less, than suggested by  the Nix-Qian paper.  Considerably less than you’d see in one generation, if they were right, and remember that they thought this was an ongoing process over many decades.  Moreover, for most of those whites that have any detectable African ancestry at all, the amount is small, a percent or two – which suggests the admixture event happened quite a while ago.

In the South, the amount of African admixture is larger among whites: about 5% of self-labelled whites in South Carolina have at least 2% African ancestry.  About 12% have over 1% African ancestry. But that amounts to an average African ancestry well under 1% among whites in South Carolina, a state in which blacks used to be the majority. That strongly suggests that the fraction of blacks that ever successfully passed into the white zone,  over the entire history of the US,  is more like 1 in 100, rather than 20% per generation.

So: what can we conclude about this paper? It’s a classic case of economic imperialism, informed by what ‘intellectuals’ [ those that have never been introduced to Punnet squares, Old Blue Light, the Dirac equation, or  Melungeons] would like to hear.

It is wrong, not close to right.


This entry was posted in Uncategorized and tagged . Bookmark the permalink.

123 Responses to The Fluidity of Race

  1. 2 generations x 17% of the black male population x 0.5 (male fraction) x 11% (black fraction of the population) = 1.87% . The white population during this period was about 89% of the population. So you’d expect that whites in this country, on average would have about 2% black ancestry. Or maybe less, since blacks average about 75% African ancestry: more like 1.5% African ancestry.

    But those designated as black under the 1 drop rule who were able to pass for white would surely have been much less than 75% African ancestry. If “blacks” who were 1/16th African-ancestry blended into the white population, the numbers would work.

    • gcochran9 says:

      There were never were many such people. Except around New Orleans. For a while the Census was trying to make a count of blacks with noticeable white ancestry – in reality almost all have some, but we’re talking visible – maybe 15%. And that doesn’t mean that they could pass.
      If the actual average amount of African ancestry in self-identified whites is 0.2%, and well under 1% even among whites in the South (which it is), fluidity is hardly an accurate description of what was happening. Effectively, African gene flow into whites during this period was about as low as Gentile gene flow into Ashkenazi Jews during the Middle Ages. Throw out the Ellis Islanders: we’re still talking about less than a tenth of a percent per generation. All the way from zero to 2 tenths of a percent in two hundred years.

      About as fluid as the Berlin Wall.

      • There were never were many such people [“blacks” with mostly European ancestry].

        Are there any reliable estimates of the number of such “people” ?

        Nix and Qian’s argument is equivalent to saying there used to be more such people than there are now. Many people with relatively little African ancestry had been designated “mulatto” by Census takers because they lived in black-majority areas and communities. Some of these would have been the whiter-looking members of otherwise black-looking families. Nix & Qian also note that at some point the mulatto category was dropped by the Census.

        By changing residence from south to north, and only then passing, the whitest-looking blacks were “boiled off”. That seems consistent with the fact that blacks in US states outside the old South tend to have more European admixutre than blacks in the South. Migration was self-selective with respect to racial admixture. Those with more European ancestry tended to migrate, and the whitest-looking migrants “changed” race on the Census.

        • that should have been “such people” quotes around both words !

        • gcochran9 says:

          It means that there would have been lots of slaves that were hardly distinguishable from whites: this wasn’t the case.

        • gcochran9 says:

          Except the the African admixture levels among whites in the North are lower than the national average, already only 2 tenths of a percent. Where did these people go? Perhaps they failed to breed. It can be pretty hard to find a mate, when you’re imaginary.

          • But if US whites had received 2% total ancestry from “passers” who were on average 1/16th African, then that’s consistent even with the lowest estimates of African ancestry amongst US whites that you see in Table S2 (page 19) of

            So it just gets back to whether there were enough people designated black or mulatto who could pass and make that 2% contribution.

            According to the 1860 census, which is online, ~10% of the total “coloured” population were designated “mulatto” (about 450,000 people). Even if you assumed all “mulattoes” could pass (a dubious assumption), that’s not consistent with Nix & Qian’s 19% passing estimate either… and that’s just their lower bound !

            • gcochran9 says:

              It’s worse than that: they’re saying that net transfer (from black to white) was ~17% (of males) per generation for several generations, not just one.

              If there were a lot of people in South who were socially black, but had 92.75% European ancestry, observers would have talked bout it. Laughed about it. Mary Chestnut would have written about it in her diary. So would de Tocqueville. The Confederate generals would have subjected them to conscription – there was a war on & they were losing. I could go on. Southerners would have constantly worried about people passing: they would have been as hard to detect as sleeper agents.

              But none of those things happened.

              I noticed someone saying that I have to explain what Qian and Nix did wrong, rather than simply pointing out that their result is impossible. Au contraire.

          • JayMan says:

            “I noticed someone saying that I have to explain what Qian and Nix did wrong, rather than simply pointing out that their result is impossible. Au contraire.”

            That’s a work-load saver for me. I come across a good bit of stuff that was apparently laboriously research but nonetheless patently silly, because of other evidence.

            Why go about furiously disassembling the entire twisted mass when simply knocking out a few key legs brings the whole structure down?

            • Reminds me of a problem encountered in the Manhattan Project. While working on the problem of firing sub-critical masses together to get to the required critical mass one scientist raised the issue of wear on the barrel, saying it would affect the accuracy of the desired result, and that this should be studied. Another pointed out that this particular projectile would be fired once, and once only. Once a line of reasoning can be shown to be wrong, its wrong. (I assume that the team are of sufficient intellect to understand when something is obviously wrong).

  2. jamesd127 says:

    Half black looks black
    One quarter black looks sort of blackish
    One eighth black, might pass as white, maybe not.
    One sixteenth black, generally passes.

    So the crossovers are going to be among the one eighth to one sixteenth population. Those are fluid.

    Further, if the passing population mostly marries among themselves, the passing population could randomly tell the census that they were black or white, even with near zero gene flow between blacks and whites.

    • Dale says:

      I’ve not made a study of this, but as I recall from reading: half African is called “mulatto”. quarter African is called “quadroon”. one-eight African is called “octoroon”. And there isn’t a word for one-sixteenth black, which I’ve seen reported as about the point at which they could pass for white.

  3. sprfls says:

    Of course you’re right and their paper is laughable. But I do wonder if the numbers are a tad higher; the 23andme study is based on their own users, and the set of people who pay for personal genetic testing isn’t exactly representative…

    • gcochran9 says:

      Whyever wouldn’t be that sample be representative?

      It’s ok to put out a laughable paper if you’re an economist. Like Emily Oster’s thesis about the real cause of the funny sex ratio in China: I guess that hepatitis B was spread by mobile ultrasound machines. And yet she has a job.

      • dearieme says:

        “It’s ok to put out a laughable paper if you’re an economist.” That’s the point in a nutshell. These buggers are too dim to do the sort of back-of-the-envelope calculation that my father taught me to do when I was young.

        • Back-of-the-envelope is exactly what I think when people accuse Big Oil of cutting prices of gasoline to favor one candidate or another the few months before an election. At a half-billion a week, it would be much simpler to just give the money to candidates. Or even voters.

      • To be fair, Oster retracted.

        • gcochran9 says:

          To be even fairer, she didn’t make the slightest effort to understand the subject. Sex-selective abortion in China was hardly a secret, while hepatitis B was decreasing with time.

          • But she also argued hepatitis B was a minor cause of the skewed sex ratio in India, which implies she understood something about sex-selective abortion, infanticide, etc. Besides isn’t the bias of a researcher interested in gender issues to rail against misogynistic practises ?

            A much fairer point is that she made an epidemiological and medical argument (linking hepatitis B infection to male birth) which she had no business making. In that sense it’s very similar to that Ashraf-Galor paper asserting that neutral genetic distances must have something causal to do with modern outcomes.

            • gcochran9 says:

              China: past expiration pas d’ennemis à gauche?

              Hep B was going down in China, sex ratio distortion was going up. They were wheeling ultrasound machines around in China, even far back in the sticks. For that matter getting rid of girl babies in China goes back a long, long time. I have no idea what Oster’s problem was, but functionally, she’s a pinhead.

              The Ashraf-Galor thesis is the sort of idea you would come up with if your sole exposure to genetics was your mother, while pregnant, coming upon and being deservedly frightened by a copy of The Metabolic and Molecular Bases of Inherited Disease, 4 volume set (6,338 pages!).

      • BurplesonAFB says:

        Why would you think every group in the US is equally likely to pay $100 to spit in a tube?

        I think there’s a decent case to be made for african gene flow to the white population occurring overwhelmingly in the south and disproportionately among lower class whites. If you overlaid the zipcodes that 23andme sends results to, I think you’d find geographically they’d trend northeast and pacific. I have no evidence of this.

        What we do know is that by 2011 those who had answered the survey reported a median income above 100k. This is not representative.
        Could not locate more recent data. I believe prices have fallen, so that may be a factor as well as the usual reporting bias.

        • gcochran9 says:

          I really have to start using that sarcasm font.

          Lower class whites: people named Ewell?

          • BurplesonAFB says:

            Oh I see. So you agree that there are ‘white’ demographics where considerably more prevailing AA ancestry lies, which are going to be undersampled by consumer genomics. What would be your intuition of magnitude? Maybe another million or two with detectable AA ancestry, but tending to be on the order of 5% rather than 0.5%?

            Lower class whites: Well if they weren’t before admixture, they’re going to tend to be now.

        • Boris Bartlog says:

          Could be. On the other hand, when you’re talking about fractions like 0.2% African ancestry, you also have to consider that it could be from older admixture or measurement error rather than recent US admixture. I have 23andme results that show 0.1% African ancestry; all four of my grandparents were German and my parents didn’t come to the US until 1963. Maybe I have an 18th century African ancestor, maybe 23andme goofed … but in any case, counting that fraction as evidence of racial fluidity in the US would be an error.

        • Anonymous says:

          Here’s something: look how white the Hispanic 23andme users are. That can’t be representative.

          I wouldn’t say it’s relevant to the question of white 23andme users being discussed here, but I thought it was interesting.

        • Dale says:

          Of course it’s more complicated than just income. But the percentage of people who pursue any particular avocation tends to vary a lot by region, age, ethnicity, income, social class, etc. Of these facts much of the marketing profession is made. OTOH, there are also ways of correcting observed statistics to remove this sort of sampling bias.

      • T says:

        “Whyever wouldn’t be that sample be representative?” Very strong self-selection. There were 5269 blacks in the sample, 148789 whites, and 8663 hispanics. If this sample is not biased due to self-selection then 23andme has proven that only 3% of Americans are black. Well we know that isn’t true, so we know that the 23andme database is whiter than average. There is no reason to believe that this doesn’t apply within races. The 23andme whites are whiter than average and the 23andme blacks are whiter than average.

        Now of course you are right that this 17% per generation result is ridiculous. Most whites don’t have much, if any black ancestry. However I would not be surprised if instead of .19% the true number was more like .59%.

    • JayMan says:

      Well, I can’t really afford 23andMe, as much as I’d like to get tested, if that says anything.

  4. George says:

    They paper was pulled from the site, so here’s an archived version:

  5. sprfls says:

    Pushkin was 1/8th African but to me he sill looks like an upper class Russian (cough German cough), just with outlier curly hair. If anything it made him look vaguely French. Or Jewish. 😉

    It is funny that lots of aristocrats have at least a little African blood since Pushkin’s daughter married some prince.

    • Dipitty do says:

      I thought maybe we had Jewish blood, too, until I figured out the origins of the curly hair in my family tree.

      • syon says:

        “I thought maybe we had Jewish blood, too, until I figured out the origins of the curly hair in my family tree.”

        What does curly hair have to do with anything?Lots of Euro Gentiles have curly hair.For that matter, lots of Jews have straight hair.

        • Anonymous says:

          There’s a certain type of curly hair that’s prominent in Jews that not many Europeans have. Kinko’s founder Paul Orfalea (not Jewish, but Semitic) describes it thusly:

          The Name Kinko’s: “I called it Kinko’s because of my nickname — because I had this really kinky hair. If you think about it, the first thing a baby learns is ‘Googoo, gaga,’ and if you think of good businesses like Kodak, Xerox, Google, people remember consonants — which was why Kinko’s was a good name. But really I had this big head of curly hair and before being called ‘Kinko’ I was ‘Pube Head.’ So I thought Kinko’s was better than Pubo’s.”

          • syon says:

            “There’s a certain type of curly hair that’s prominent in Jews that not many Europeans have. Kinko’s founder Paul Orfalea (not Jewish, but Semitic) describes it thusly:”

            Yeah, but large numbers of European Jews don’t have that kind of hair.Also, quite a few Gentile Europeans do have it.Just take a trip through Italy some time.The number of “Pube Heads” there is pretty big.

    • Peter Lund says:

      Alexandre Dumas, père et fils.

    • Jim says:

      In the pictures of Alexander Dumas pere he definitely appears part black. Based on the pictures of Alexander Dumas fils I guess he could pass for white.

      Based on a photograph of him I’m surprised that Anatole Broyard could pass for white. However I think many Creole probably could pass for white.

      • Patrick Boyle says:

        Do you suppose it means anything that all four of these famous mulattos who ‘passed’ for white – Pushkin, The two Dumas, and Broyard – were all prominent in literature? I read a lot of Pushkin (in translation) when I was young and the younger Dumas a year or two ago (again in translation). I never detected a ‘black’ voice but maybe it was ‘lost in translation’.

        If the effect of some black ancestry was beneficial you would expect it to be better distributed – career wise. Why no chemists or mathematicians? Could it be that fiction has a lower threshold for accomplishment? Wikipedia is no help except to the extent that it does seem to be the case that mathematicians and chemists are indeed smart. No one seems to have researched the question – how smart are the writers of best sellers?

        No doubt the very fact that I imagined a connection between these factors prove that I am an unreconstructed racist.

        • jIM says:

          I don’t think it was a secret that Alexandre Dumas was part black. I’m don’t believe that he was hiding his African ancestry.

        • sprfls says:

          Abram Gannibal, Pushkin’s great-grandfather and the African in question, was probably a pretty smart guy. That’s why he was “noticed” in the first place and brought to Peter the Great, who further acknowledged his intelligence and adopted him as a godson. As a child he served as Peter’s valet and then was sent off to France to be educated where he studied math and engineering and by all accounts did well. Of course, all of this might have been orchestrated like some historical version of Trading Places. Mooortimer!!

          Abram’s oldest son Ivan (Pushkin’s great-uncle) also went on to an illustrious career in military and government. Unlike his father there exists a definitive portrait of him. He barely looks black to me.

          It’s also worth noting that Pushkin’s unfinished novel loosely based on his great-grandfather’s life was incorrectly translated to English. The Russian title is Арап Петра Великого — Peter the Great’s ‘Arap,’ not Peter the Great’s Negro. Arap refers to north/east Africans. I’m not sure such a distinction exists in English — the dictionaries say ‘blackamoor’ but I don’t think that’s exactly correct.

        • Steve Sailer says:

          Here’s an African-American who was a heavyweight statistician:

          He looks maybe 40% black.

          • Patrick Boyle says:

            Well I won’t trade racial/facial observations with you Steve. You’re a pro. But I would just note that all these prominent and accomplished part blacks – including our President – are praised for their achievements as if they were caused solely by their fraction of black ancestry. If someone is 7/8s Swedish and one eighth black he will be described in Wikipedia as black. All the ‘black history’ books and web site will cite him as an example of black accomplishment. He will be listed among the famous blacks in history. He will be shown on PBS and praised on NPR. Is this not what has happened with Pushkin?

            Maybe we should say that these fractional blacks were successful despite their black heritage.

            As I remember there are currently about three hundred SNPs that each account for a small proportion of the IQ variance. Presumably these sort independently of those SNPs which influence skin color, hair form and facial features. So we should expect some dark, smart hybrids among the more common not so smart hybrids.

            Disregarding environmental effect in general if a person is smart that means that they have received mostly favorable alleles at those three hundred SNPs. If they are part white they would have most likely have gotten them from their white ancestors. So we should see Pushkin as a man with a largely white brain cursed with a black physiognomy. To imagine that intelligent mulattos got their smarts from their black fraction would require a whole series of spontaneous favorable mutations – a statistical impossibility..

            Part of the confusion is that many still consider heredity to be an analog mixing of fluids. Whereas it is, of course, a discrete process. The difference is that if you mix two liquids you always gat an intermediate result. For example regular milk and chocolate milk when mixed together always yield a kind of beige milk. That’s what people expect. Anything but an intermediate result seems like a violation of the laws of thermodynamics. So Pushkin is seen as very odd.

          • Jim says:

            He looks like Milton Friedman.

        • Dipitty do says:

          Alexander Dumas’s father, half black, was notable for his military career.

          I have spent a lot of time around writers, and most of them are extremely intelligent–they are, at the very least, folks with extremely high verbal IQs. Verbal and mathematical IQs tend to correlate, and my experience with these folks is that they are generally well above-average at math; a few I know are quite above average. Writers are some of the few people with whom I feel at home.

          Whether or not the folks you cite could have been chemists or mathematicians, they were probably very well situated to become writers–they were wealthy, if not outright nobility like Pushkin, and so had more to benefit by cultivating themselves as interesting people for other nobles to be around than to hole up in laboratories. From a genetic standpoint, being slightly different from the people around you probably helps one come up with different sorts of ideas than everyone else is coming up with. In Chemistry, your ideas have to also be accurate, but in literature, just being creative is sufficient to be beloved.

  6. Dipitty do says:

    I have a small amount of black ancestry from the 1800s. I can pinpoint it both in my genome and in family records (an ancestor recorded as being “as dark as an Indian”, among other things.)

    So these things did happen. Just not very often.

    • gcochran9 says:

      Yes. Sure, it happened. I can think of well-known examples like Anatole Broyard. But not at anything like the rate postulated by Nix and Qian. If ~20% of the boys in each black generation were turning socially white, people could hardly have missed it, unless they were like the students in Sunnydale High School that somehow don’t notice all the kids being eaten.

      • Michael says:

        Maybe many of the black to white conversions ended up marrying a spouse in a similar situation. It is probably much easier to persuade a census taker than your girlfriend. This would also explain switching back, because some children might look more African.

        But didn’t the 23andme study suggest (based on the size of the segments) that most of this admixture occurred in the 1600-1700s?

      • syon says:

        “Yes. Sure, it happened. I can think of well-known examples like Anatole Broyard.”

        Maybe the best known antebellum example involves Sally Hemings’ son Eston:

        “Passage of the Fugitive Slave Act in 1850 increased pressure on the black communities in Ohio and other free states bordering slave states. In towns along the Underground Railroad, slave catchers invaded the communities, sometimes kidnapping and selling into slavery free people as well as fugitive slaves.[17] In 1852 the Hemingses decided to move their family further north for security, and migrated to Madison, Wisconsin. There they took the surname Jefferson, and they passed into the European-American (white) community.[14] Eston Hemings Jefferson died in 1856.

        John Wayles Jefferson
        Colonel John Wayles Jefferson.gif
        Their eldest son John Wayles Jefferson served as a white officer in the regular United States Army during the American Civil War, achieving the rank of colonel. John W. Jefferson led the Wisconsin 8th Infantry. He was wounded twice in battle. During the war, he published letters home, and after the war, published articles about his experiences. Before the war, John Jefferson ran the American House hotel in Madison, which was taken over by his younger brother Beverley. After the war and the end of slavery in the U.S., John Jefferson moved to Memphis, Tennessee. He became a successful cotton broker, supported his mother, and left a considerable estate at his death in 1892. He never married or had known children.[5][14]

        Both Anna and Beverley Jefferson married white spouses, and their descendants have identified as white. Anna married Albert T. Pearson, a carpenter who was a captain during the Civil War. Their son Walter Beverly Pearson became a wealthy industrialist in Chicago.[18]”

        • Indiana Jack says:

          Judging by this account, the Hemings/Jeffersons seem to have been a fairly accomplished group. I couldn’t help being reminded of Gregory Clark’s ‘The Son Also Rises’ – especially striking for a son that was never publicly acknowledged by Jefferson and who inherited none of his estate.

      • Dipitty do says:

        I agree that 20% is an absurd number. I suspect that someone looked at a bell curve of “% African ancestry of African Americans” and took the rightmost 20% of the paper and mistook that for 20% of the total population.

        As it is, I find it kind of hilarious that apparently a bunch of people in my family either couldn’t figure out where the mysteriously dark kid (now ancestor) came from, or just didn’t bother saying anything.

  7. Perhaps it is the percentages which confuse people, as they seem to have confused the authors of the study. Might it be better to describe races in terms of drops of blood (base 1000)? Gerd Gigerenzer found that people can deal with statistics when they are presented in terms of “natural frequencies”. So, the 23andme white clients have 2 drops of African blood, 998 drops European blood. As already observed, probably as fluid as the Berlin Wall.

    • “Perhaps it is the percentages which confuse people, as they seem to have confused the authors of the study.”

      But they don’t talk about admix percentages in their Census analysis. They simply note that the US census categorised people as black or mulatto, but at some pointed dropped the mulatto category altogether.

    • BurplesonAFB says:

      The measurement should be an expression of blood quantum and the unit should be milliquants

  8. amac78 says:

    My recollection is that 23andMe’s estimation is saying, “By our analysis, your alleles at the subset of SNPs we believe are informative gives your ancestry as about [X]% of [ethnic/racial group A], [Y%] of [ethnic/racial group B], and so on, to 100%. They don’t claim that their algorithm is perfect, so that some small percentage might either be some small contribution from a racial/ethnic group — or an artifact of the analysis.

    So it’s not clear to me that a 23andMe finding of SNPs suggesting, say, “0.5% African ancestry” should be taken to mean “0.5% African ancestry” as opposed to “very small and possibly zero African ancestry.”

    Perhaps a better-informed reader can confirm or correct this impression.

  9. JK says:

    Yeah, I too found the results of this study highly implausible. Their findings are dependent on complex algorithms that attempt to match individuals across different census years based on their names, ages, birth states, and parents’ birth states and then checking if their racial categorization has changed in the intervening years. However, there are lots of missing data, changes in the spelling of names, and unreliable reporting of birth years. This means that their “matches” are often pairs of individuals with phonetically similar but non-identical names, no data on birth state, and age differences of up to three years.

    For example, their algorithms may conclude that Ravale Thibodeaux of Louisiana, born in 1870, as reported in the 1900 census, is the same person as Raval Tibido of Louisiana, born in 1867, as reported in the 1910 census. If Thibodeaux is classified as black and Tibido as white, they conclude that Thibodeaux/Tibido has “passed” as white.

    There’s one source of bias that they don’t consider. If blacks with lots of white ancestry were more educated than less admixed blacks, as was (and is) the case, then we would expect that they would be more likely to correctly report their names, birth dates, and places of birth across different census years. Those with more white ancestry would also be much more likely to pass as white, which means that those blacks who were most prone to passing are easier to identify across multiple census years than those blacks who were less likely to pass. This biases their estimates of passing upwards.

    • I was told (sorry, no reference) that many freed slaves wound up with the last names of their erstwhile owners. That might muddle the matching a little more.

      • JK says:

        Yes, black names, whether last names or first names, weren’t distinctive in those days — no one was called DeShawn or LaTrine. So any false match in the study is likely to be a white person, considering that whites were much more numerous.

  10. Razib Khan wrote an article on this subject a while back, here

    I will let Razib speak for himself, He takes issues with a New York Times headline that proclaims “White? Black? A Murky Distinction Grows Murkier.” Nooooo. What Razib goes on to say is this
    “What genetics is showing is white Americans are shockingly European to an incredibly high degree for a population with roots on this continent for 400 years.”

    Regarding African Americans Razib says being “about 28% proportion (black) is about where african ancestry becomes salient or not.” So….are 19% of African Americans below this percentage African American? Nope, not even close. To quote Razib from the same link ” A small minority of Black Americans are more than 50% European in ancestry and a majority are less than 25% European in ancestry.”

    The above article by Razib destroyed a silly notion I had of my own ancestry. I primarily hail from New England puritan stock with 8 plus generations of ancestors in North America. 2 to the 8th power is 512 direct ancestors who met their mates here. I thought anyone with that many ancestors here must be a mutt, meaning they are part american indian and likely many nationalities and races. I was wrong. It is no big deal today for mixed marriages but it sure was in the past.

    • CaoMengDe says:

      “2 to the 8th power is 512” only works if none of your ancestors are related. Longer term, the assumption of distinct lineages is apparently false since this would imply 4000 years ago, there were staggering amount of people on this earth.

      • gcochran9 says:

        Didn’t the Old Race lose a ship out that way a few half-deaths of Thorium-230 ago?

      • Nador says:


        • dave chamberlin says:

          Yea, I did a lousy job making my point. Let me try again. I have no stinkin idea how many direct ancestors I have who have selected mates in the United States. Because the overwhelming percentage of the ancestors that I know of go back to the revolution it is a huge number. Waaay over 200, it could zoom over 1000, easily.

          I’m plus 90% WASP. My kids are half that percentage WASP and my grandchildren are one quarter that percentage WASP. My point is this. There is a dramatic change in this country in just the last few generations in folks sticking to their own kind in mate selection. My personal story is confirmed statistically by the link to Razib’s thread over at GNXP. Ain’t no way no how 19% of blacks can choose to be white because it is a far smaller percentage of blacks than that that have that choice.

  11. As far as I can tell, “passing” seemed to be a temporary phenomenon. Often people did it to work a different job, or even just ride the “white” bus for the afternoon and come back from the other side of town. So it seems that uprooting one’s entire life and ignoring family was something difficult and unnecessary that not many people wanted to do (who would have thought?).

    I read Anne Bernays memoir of the fifties and she dated Anatole Broyard for a period of time. Their social circle knew he was a “passing Negro.” I would imagine most people with black admixture know about their backgrounds because it would have been important at some point, but many Americans, particularly white ones, don’t know poop about their backgrounds and can imagine anything is in there.

  12. IC says:

    Try to rationalize the difference.

    Fake white suffered higher rate of murder (due to KKK not tolerating such fake), or lower rate of marriage (due to looking strange for white who select them low and fake white unwilling to settle for black partner since they believe they can marry white), or higher mortality (due to health reason), or lower fertility rate )due to immuno problems).

    These are just speculations without fact or data. At end, treat this as half serious joke.

    • IC says:

      I do personally know a white looking black lady in Massachusetts never get married. She never want to date black men. But her white skin can not hide her some African features and have some problems dating white guys. Single in her 60s. Never married.

    • CBurd says:

      Although a deplorable organization, the KKK were not the equivalent of the Nazi Einsatzgruppen.

  13. RCB says:

    Some relevant social data: the Gallup poll shows that black-white marriage was widely disapproved of only one generation ago.

    Among whites, the approval only appears to have surpassed 50% in the mid-90’s. (Was much more commonly approved by blacks.) So approval (at least verbal approval) is a surprisingly recent sentiment, with the majority of change occurring in the last two generations.

    • Stan D Mute says:

      Keep in mind that there is a difference between answering a poll question, “I’m okay with miscegenation,” and saying to your daughter, “That’s great honey, when will my mulatto grandkids be born?”

  14. Asher says:

    What I find interesting is the “black” people who could easily pass as white but who strongly identify as “black”. You’d almost think people had an advantage by having others see them as “black” …

  15. Asher says:

    Is anyone familiar with Gunnar Myrdahl’s study of segregation in the South from about fifty years ago? Apparently, the number one reason verbalized by southern whites for segregation was simply to keep black men from impregnating black women.

    You’d almost think male sexual competition was some sort of universal phenomenon … thank God I have social scientists to tell me differently.

    • RCB says:

      Do you mean “impregnating white women” ?

      • Asher says:

        Doh. Yes. What I find irritating is that there was a very real, objective reason behind southern segregation, one that no longer exists today. The objective reasons for segregation’s existence is gone, making it resurgence highly improbable.

        You’d think that Marxists would have grasped this since Marx’s central claim was that objective conditions create ideology … guess not, though.

  16. Craig Palsson suggested a test of the Nix-Qian results : “take a subset of switchers and see how many are confirmed false matches e.g. using spouse/children info…By Cochran’s estimates, the false matches should be about 90%…”

  17. ” The Confederate generals would have subjected them to conscription – there was a war on & they were losing.” An exceptionally good point, which had never occurred to me. Whether it is original to you or simply remembered from some other smart person who everyone else ignored, +1 Cochran

    • gcochran9 says:

      There was a big debate about drafting blacks: Patrick Cleburne, one of the South’s better generals, was for it early, but others strongly rejected the idea – arguing that if slaves made good soldiers, the underlying theory of their state was wrong. [this is an example of why it’s best for your state not to bother having any underlying theory: proposition nations can be mistakes, and what then?] But too late, after Lincoln’s re-election, Lee said it had to be done. Laws were passed, a few slaves recruited, but the South surrendered before any got into combat.

      It seems to me that if a substantial fraction of ‘blacks’ existed that couldn’t be told from whites, there would have a been strong temptation to conscript them: “close enough for government work”. And there were some such people, particularly around New Orleans – but they generally weren’t slaves, and they didn’t make up a significant fraction of the population of the Confederacy.

      • Sean II says:

        A good analogy can be found in the Waffen SS. When things started going not so well, they conveniently discovered a bunch of Aryans hitherto unknown to racial science.

        • ursiform says:

          I seem to recall the Nazis proclaimed the Japanese “honorary Aryans”. So much easier to ally with them.

          When QE II visited Saudi Arabia she was proclaimed an honorary man. It would have been awkward, otherwise. (Prince Philip went along and was, as usual, awkward …)

          • Sean II says:

            To be fair, some of the Duke’s gaffes are funny.

            The SS at one point became desperate enough to recruit Bosnians. They even got the Grand Mufti of Jerusalem (who bore an odd resemblance to Ryan Gosling, see here: to write a little booklet on their behalf.

            Now if they could bring themselves to do that, you gotta figure Cochran’s right about how the Confederacy would have handled guys who looked like Tom Jones.

  18. Cracker1 says:

    They didn’t track people from one census to the next. They tracked names (using Soundex) with a mathematical model. There is absolutely no way that this will provide accurate information using census data. It can be difficult tracing a known family from census to census even when you have additional non-census information to aid you. For many reasons the use of a “census name” is not reliable. Not the least being that for most of the censuses that they used, most of the Southern states were not issuing birth certificates. People weren’t as choosey over their lifetime about their name as we are today. Now, if an additional question is how many mixed race people were there in 1870, then just go to the 1870 census and look in the race column for W, B or M and add them up. The only problem one might have is determining if the letter is a W or an M, but since the same person transcribed several pages one can using get that straight. If you don’t want to bother with that you can just take my word for it. In those areas where there were a lot of former slaves, there are a lot of M’s.

    • Anonymous says:

      In that regard, I had trouble getting a passport because my parents ignored census takers (didn’t answer the door or told them to get lost?). I was told that even though I go straight back to the May Flower and my father was in the military and hence had regular old passport that I needed to “prove” that my “claim” to be American had merit… It was quite bizarre. At any rate, tracking individuals across multiple census years sounds rather dubious! Especially way back then.

  19. pyrrhus says:

    Academics in the social sciences can’t do simple math–shock…..

  20. DJ says:

    “So you’d expect that whites in this country, on average would have about 2% black ancestry. Or maybe less, since blacks average about 75% African ancestry: more like 1.5% African ancestry.”

    No; I would not expect that whites have 2% black ancestry on average. Rather, I would expect 2% of whites to have African ancestry. Those are not the same, in large part because the blacks who would pass as whites in 1940 and before would not be the ones who have 73% black ancestry — they would need to have much less.

    Now a direct quote from the 23andme study:

    “Consistent with previous anecdotal results,32 the frequency of European American individuals who carry African ancestry varies strongly by state and region of the US (Figure 3A). We estimate that a substantial fraction, at least 1.4%, of self-reported European Americans in the US carry
    at least 2% African ancestry. Using a less conservative threshold, approximately 3.5% of European Americans have 1% or more African ancestry (Figure S8). Individuals with African ancestry are found at much higher frequencies in states in the South than in other parts of the US: about 5%of self-reported European Americans living in South Carolina and Louisiana have at least 2% African ancestry.”

    Seems consistent to me.

    Finally, your analysis assumes that the classification criteria remains constant over time since 1940. It can well be that, in modern days, people who have significant black ancestry are not as likely to report themselves as whites because the incentives to do so are much smaller than in 1940. So descendants of people who “misreported” in 1940 and before might not “misreport” today. That would translate into the whites today being less likely to have African ancestry compared to 1940, because the “whites” group does not include the descendants.

    Maybe the paper is wrong, but your points do not warrant that conclusion.

    • gcochran9 says:

      There never were many people classified as black that had super-high levels of European ancestry. They were the only ones that could have ‘passed’.

      The idea that 17% of each generation of black males moved from the social category ‘black’ to the ‘white as a continuing process – it’s ridiculous. The paper is ridiculous. So are you, but that by itself is not interesting.

      • Sam says:

        Well, it is a working paper – so not yet published (I haven’t read the paper – link not working). Hopefully, they will see your comment, incorporate and update the paper. So I wouldn’t be too hard on them yet.

        • gcochran9 says:

          They don’t know anything at all about genetics or American history. You think that’s going to change?

          • AnonymousForGoodReason says:

            What might need to change earlier still is their campus ideological bell-weather.

            More interesting than the paper itself, is that nobody from the Yale econ monoculture has tried to correct them. Off-stage, who knows what outright encouragement they’ve been subjected to. Will such a bell-weather be derailed by the disapproval of a scoundrel like Greg Cochran, who is suggesting boring work, which can only lead to a vehicle less capable of sailing the winds?

          • Sean II says:

            They know one thing about American history: it can be used.

          • dearieme says:

            WKPD helps out.

            “A bellwether is one that leads or indicates trends.

            The term is derived from the Middle English bellewether and refers to the practice of placing a bell around the neck of a castrated ram (a wether) leading his flock of sheep. The movements of the flock could be noted by hearing the bell before the flock was in sight.”

      • DJ says:

        So I come and point to a counter-argument to your post (without even saying I support the authors), and your response is to tell I’m “ridiculous” and use as justification your prior, for which you have shown no evidence.

        You must be really proud of yourself. Such a thoughtful response!

        “There never were many people classified as black that had super-high levels of European ancestry.”
        Care to cite any support for that? I don’t know about this subject. Wouldn’t it be possible, say, that 1 out of 5 black males at the time where sons or grandsons of whites?

        • gcochran9 says:

          A white grandfather would not be enough to allow someone to pass as white, nor would a white father.

          If you looked at the 1860 census, they talked about both blacks and mulattoes: blacks defined as not obviously having any white ancestry (although they surely did have some) and mulattoes as having noticeably more (from visual inspection). Mulattoes made up about 10% of the black population. Most of the mulattoes couldn’t have passed, and none of the other blacks could have.. While the Nix/ Qian paper has an impossibly large fraction switching (19%) – not just in one generation, but on a continuing basis.

          And, if there was that much gene flow into the white population, where did it go? It sure isn’t there today. To be exact, there is some (0.2%)- but not nearly as much as would resulted from the high levels of passing suggested in the Nix/Qian paper.

          Anyone with a reasonable knowledge of current genetics would have realized that their paper was wrong. Anyone with a reasonable knowledge of American history would have known it was wrong: it implies that about 20% of the slaves in the South couldn’t be told from white.

          I know both of those topics pretty well. Nix and Qian don’t know jack about either one. Of course they’re going to be wrong on a subject like this: it is not just that they’re ignorant – effectively, they’re nuts. In that paper, they also mention how prominent blacks, political leaders and writers and such, obviously have much more European ancestry than average blacks. They attribute that to better nutrition: which must explain how light-skinned all those running backs in the NFL are.

          To be fair, if they didn’t say nutty things on this subject, they would be fired.

          Your comment wouldn’t have survived a one-line quantitative estimate. So why should I play nice?

  21. jef says:

    Bar fight! Bar fight!

  22. Steve Sailer says:

    How many 20th Century celebrities passed from black to white during their lifetimes? There’s the famous case of Greenwich Village man of letters Anatole Broyard:

    But I’m not familiar with many other cases. Carol Channing found out her dad had passed from black to white. But, in general, I know a lot of celebrity anecdotes and I don’t know many about passing. It was a difficult thing to do: you had to cut public ties with at least one side of your family.

    Maybe these shifting Census designations aren’t the same as passing.

  23. Steve Sailer says:

    Recently, I noticed a case of one set of Americans passing from black to American Indian: the Shinnecock triracials in the Hamptons of Long Island were described as black in golf articles in the 1980s because of their long involvement with the famous Shinnecock Hills golf course. But recently the Obama Administration declared them to be an Indian tribe so that they could open a casino, and now you never hear about them being part black anymore.

    • ironrailsironweights says:

      I actually had some dealings with several Shinnecock members when I was selling life insurance a few years back. Most had a fairly distinctive look about them, somewhat Latin American, somewhat Middle Eastern, but not really like either group. Only one of them might have been taken for black.


  24. L says: least now we know it’s not White people’s fault black people choose to identify as black; instead of checking the White identifier and reaping the full benefits of their Whiteness. Can we please call off the whole affirmative action thing as well since this is clearly more Whites identifying as AA to get a pass at undeserved social benefits?

  25. Lot says:

    While you’re right that the study is a gross overestimate, your commenters are right that your reductio goes too far. Both because the passing blacks would not be the average 73% africans, and because you did not account for the substantial flow of 0% blacks coming from Europe.

    Also, 23andme’s whites are heavily Jewish and from California, the Northeast, and the Upper Midwest: the whites least likely to have African ancestors.

    • gcochran9 says:

      If there had been lots and lots of blacks passing each generation, they would have had to be fairly typical ones: there was no bottomless well of “high yaller” men.

      • BB753 says:

        Would “black” basketball player Jason Kidd have passed back then? He looks pretty white but still displays some negroid features. And frizzly hair was harder to conceal before shaved heads became fashionable . In order to pass no questions asked, you need to have a black ancestry of no more than 1/16.

        • ziel says:

          I would think Derek Jeter could pass as well – but not sure his sister could.

          Anyway, if it were that easy, it would be a lot more common. But as Steve mentioned above, it pretty much means cutting yourself off from half your family.

          And you wonder what’s the point? If you lived down south, it was probably too risky to try. And up north, I doubt there was a lot to be gained. Obviously there is a great deal of prejudice in the North, but from my observations it was very facultatively applied – “Oh he’s really ‘white'” would be a typical comment about someone who, shall we say, conformed better to the prevailing behavioral norms.

          If there’s any field where we should see “passing” be a commonplace occurrence, it would be in music, where ability seems to be evenly distributed by race (and perhaps more prevalent among blacks). Are there any prominent musicians from the 1st half of the 20th century who passed as white? Can’t think of any myself.

          • Dipitty do says:

            Not necessarily cutting yourself off from any of your family if you never met your father in the first place. If you looked close enough to your mother’s family to possibly fit in with them, you’d have gone with that.

        • ironrailsironweights says:

          I doubt Kidd could have passed for white back in the 1800’s. He might have been able to claim part American Indian ancestry, if he didn’t want to be taken as black.


  26. Hipster says:

    I am one of the White Americans with about 3.5% Sub-Saharan African Ancestry.

    My family always thought we were some small part Indian. We have worked out that a great-great-great Grandfather of mine was Native American, had worked out the tribe and the area he was from and all. We are all from the North-East, no slave-states in my family history at all.

    23andme told me I wasn’t Native American at all, but about 3.5% African American, which works out to the contribution of a Great-great-great Grandfather.

    So I assume that somewhere along the line in the 1800s, someone who was 1/4 black or 1/8 black made up the lie that his dark complexion was due to Native American ancestry.

    I do wonder about the historical circumstances around this. It means that I’m 1/32 SSA, my mother 1/16 and my grandmother 1/8, and my great-grandfather 1/4. None of them look in the slightest bit black.

    It made me think that perhaps there wasn’t 1 single black ancestor but several. But my 23andme data shows long chunks of SSA ancestry on my chromosomes, so it does not seem like I had a lot of mixture from different SSA sources.

    If anyone here is a genetics wiz I would love to forward my 23andme data and see if anyone can help me make sense of it.

    • Dipitty do says:

      IIRC, some Indian tribes had quite a few people of African origin in them, for a variety of reasons. So your ancestry could have actually come via the claimed tribe.

  27. Pingback: Outside in - Involvements with reality » Blog Archive » Chaos Patch (#47)

  28. SF says:

    Well, I know three children whose mother is 1/4 African and 1/8 native. They are white, unless applying for affirmative action. Mom could pass for something Mediterranean. Also have a niece by adoption whose biological father was African. She looks white.

  29. Pingback: Economic History Link Dump 15-01-2015 | Pseudoerasmus

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