The US really is a Nation of Immigrants – and Peter Brimelow is wrong

Some time back, I got into a discussion with some commenters on Open Borders. The starting point was a claim by Peter Brimelow who is the editor of the restrictionist website VDARE. In an address to the Philadelphia Society, delivered in 2006, he stated it this way:

„But the last estimate that I saw, when I was researching Alien Nation, was that if there had been no immigration at all after 1790—none at all—the population of the US would still be about half of what it is now, through natural increase.“

This is part of an argument that it is misleading to call the US a “nation of immigrants.”

I was baffled by the claim, and my first reaction was to point out that American population would have grown by a factor of 40 since 1790, while the population of Germany grew only by a factor of less than 4, and world population by a factor of 7. A commenter then supplied an argument that very high fertility in the early US was behind it. This seemed to be an explanation, and so I retracted my criticism, but was still amazed how that could be.

Turns out I gave in too fast because:

The US really is a nation of immigrants, and overwhelmingly so.

I will go into more detail in another post because there are further aspects that are interesting (hint: it’s the momentum effect again). Here I will confine myself to a simple argument that shows why something has to be wrong with Peter Brimelow’s claim. I will also derive a more realistic estimate for the counterfactual.

Let’s first look at where the US population in 1790 had come from: Of the slightly more than 3.9 million inhabitants, about 760,000 were African Americans, mostly slaves. Native Americans were not counted at the time. The rest were of European descent, some 3.2 million people. More than 2.5 million of those of European descent or 78.6% traced their ancestry to Britain (59.7% English, 10.1% Scots-Irish, 5.0% Scottish, and 3.8% Welsh).

The reference year for Peter Brimelow’s claim is 1990, and he asserts that about half of the American population would have been there at that time without any immigration after 1790: 122 million in the counterfactual versus an actual population of 249 million (cf. “Alien Nation – Common Sense About America’s Immigration Disaster”, page 48).

There were 30 million African Americans in 1990. It is certainly an overestimate that all of them would have been there without any immigration after 1790 where this would have to include also forced “immigration.” That’s so because about half of all new slaves were brought to the US after the founding of the Republic.

Let’s be generous and concede 20 million African Americans in the counterfactual. There were also almost 9 million Native Americans who would have been there in 1990. So the number of all those of European descent in the counterfactual works out as 122-20-9 = 93 million people in 1990.

It is reasonable to assume that 78.6% of those 93 million would have been of British descent in the counterfactual, or about 73 million people. Otherwise you would have to explain why other groups (mostly Germans, Dutch, French, and Irish) had vastly diverging fertility over 200 years. I don’t see how you can make that case.

Now, since the counterfactual is just a part of what really happened with immigration, there should have been also at least 73 million actual people of British descent in 1990, or about 29% of the population. Of course, those would only be the same people as in the counterfactual if there had been complete segregation of later immigrants, which was not the case.

What counts here are surnames which anchor a claim to ancestry. Since half of someone’s descendents (usually sons) keep the surname, and half of them (usually daughters) lose it, the shares do not change a lot across generations. That is barring strongly differential fertility or systematic name changes that I find implausible. So the shares for ancestries should be roughly stable.

But then 29% has to be an underestimate because there were another 3.5 million immigrants from the UK between 1820 and 1930 alone. You can get a rough idea from German immigration for the effect of that later immigration in 1990.

There were only 280,000 Americans of German descent in 1790. They should have grown to about 7 million people in 1990 with the same rate as for those of British decent. But there were another 5 million German immigrants between 1850 and 1930. In 1990, there were 58 million Americans of German descent. So the 5 million later German immigrants should have grown to about 51 million people.

Hence it seems reasonable that the 3.5 million later British immigrants from 1820 to 1930 should have grown to some 35 million people. But that means that the share of those of British descent in 1990 would have been about 43% of the total population (108 million out of 249 million).

However, there were only 18.8% in 1990 who claimed to be of British descent or much less than half of what it should have been if Peter Brimelow were right.

There is one objection, though. There were also 6.2% who declared “American”, “US”, “European”, or “white” ancestry in the 1990 census. Probably some of them should be counted as of British descent, too. But even if you include all of them, you only get to 25%, or somewhat more than half of what is required. So the numbers simply don’t add up for Peter Brimelow’s claim, and that is so by a wide margin.

Now let’s derive a more realistic estimate (which has its limitations, but should be much closer to the truth):

If you take the high estimate of 25% for those of British descent (including all those who checked “American” ancestry, “US” ancestry, and so forth), there were about 62 million people in 1990 that belonged to that group. Subtracting the 35 million resulting from immigration after 1790, yields 27 million people, or only somewhat more than a third of the 73 million in the counterfactual.

But that can only mean that also the number of those of European descent in the counterfactual has to be much lower, not 93 million, but only 34 million people. Add in the 20 million African Americans (probably an overestimate) and the 9 million Native Americans, and you arrive at an estimate for the American population in the counterfactual of 34+20+9 = 63 million people.

Hey, that’s not bad, that’s almost the population of France! It is well below that for Germany, though, and only half the Japanese population. But relax, the US would still be more populous than Canada, admittedly not by a lot.

And a population of 63 million people would have been only 25%, and not almost 50% of actual population in 1990. Or in other words: Roughly 75% of the American population were there because of immigration after 1790!

But there is also a silver lining for Peter Brimelow here: The US was taken over by immigrants long ago, and it worked out so well that he is now defending the result as the status quo. Just imagine: No one would have noticed this massive swamping if I hadn’t written my post. And it is a fine example of how a nation of immigrants could become a great country. Make America great again!

Hansjörg is a mathematician by training with a doctorate from the University of Bonn, Germany. After a year at Stanford University as a guest scientist, he went on to work in the financial sector and managed corporate bond funds. Currently, he is building his publishing company Libera Media.

See our blog post introducing Hansjörg, or all blog posts by Hansjörg.

4 thoughts on “The US really is a Nation of Immigrants – and Peter Brimelow is wrong”

  1. This is laughably wrong. Your estimate is that the 3.5 million British people who immigrated from 1820 to 1930 would have ballooned to 35 million people, while the descendants of the original British colonists, who were in the country for much longer, during a time of very high fertility rates in the first forty years of America, when there was hardly any immigration, would have grown by about the same proportion, from “more than 2.5 million” to 27 million.

    You can also look at the relative shares of the population, you claim 20 million Blacks and 34 million Whites descending from populations which were there in 1790. But if the ratio of Whites to Blacks was the same as it was in 1790, 4:1, there would be 80 million Whites.

    Self-reported ancestry is unreliable, you’d have to believe in secret death camps if you took it too seriously. Ethnicities go in and out of fashion, and British ancestry went out of fashion in the 1980s, that decade of supposed conservative triumph. Whereas in 1990 census does show 18.1% of the population reporting British descent(the combined totals of British, English, Scottish, Scotch-Irish, and Welsh), the 1980 census showed 32.56 percent of the population claiming to be either wholly or partially English, Scottish, or Welsh.

    Alien Nation cited the following paper for the estimate:

    1. Thanks a lot for your comments, Jason. To address your points:

      (1) The 2.5 million of British descent in 1790 would have grown to 27 million, that’s a factor of about 11, and that would have been mostly due to natural increase. The 3.5 million later immigrants from Great Britain were new immigrants. That means there was not only natural increase, but also a momentum effect (they brought their not yet born children and grandchildren with them, cf. my post “Misinterpreting Growth of Immigrants Populations”). A rough estimate for the momentum effect is that effectively twice as many people immigrated (or even more). So the factor for natural increase here is at most something like 5. That would be parallel with German immigrants after 1790. I don’t see what’s wrong with that. Or you would have to challenge the same assertion also for German immigrants. But then, where do the Americans of German descent come from if momentum and natural increase cannot be the explanation?

      (2) As far as I know, African Americans have had higher fertility for a long time. With one-drop logic and 20% or 25% European admixture, there should have been also a further increase in this way (Obama is the first black President or the 44th white President?). I also ignored that not all African Americans trace their ancestry to the slaves of colonial times and the early US, some are also descended from immigrants from Africa, Latin America, or the Caribbean. And then most slaves were brought in from the middle of the 18th century on, so there was still some momentum effect left from before 1790 that only played out afterwards. Since this is all hard to disentangle and probably not comparable with European immigration, I granted a very high estimate: what Peter Brimelow claims plus a wide safety margin. (Half of all slaves were brought in before 1790 > 20 (!) out of 30 million African Americans would have been their descendants.) Note also that I took a deliberately high estimate for African Americans of 30 million, while the actual number in the 1990 census was less than 24 million. But you cannot conclude from this that I would also have to overestimate the number of European immigrants and their descendents in the same way.

      (3) Granted that self-reported ancestry is not perfect. But then I don’t need very exact numbers for my argument, just a ballpark. The data for 1980 come out as: 13.4% for single British ancestry. If I add in half of those who reported multiple ancestries and also British among them, then I get 23%. That’s roughly the size that I also get for 1990: 18.8% for British ancestry plus some who report “American” ancestry and the like. Even if it were warranted to count all those with multiple ancestries and among them British as you do (I don’t see why), you still fall short of the expected share by a wide margin.

      (4) I know Peter Brimelow relies on this paper, and I can understand that he made an honest mistake in doing so. I will address the point in a further post as I have announced. In a nutshell, the problem with the paper is this: The author is not aware of the momentum effect (cf. once more “Misinterpreting Growth of Immigrant Populations”). What he does is this: He calculates “natural increase” as total population growth minus net migration. And then he applies these rates from 1790 on. I could almost reproduce his numbers. However, the tricky point is that in this “natural growth” there are also the children and grandchildren of immigrants after 1790. So the natural increase he really needs would have to be corrected for this, and it would be quite a bit lower. But he doesn’t do that. Overestimating natural increase in this way builds up to a large error after two centuries.

      [I made some minor edits: typos, clumsy wording.]

    2. P. S.

      It is usual to claim that there was hardly any immigration to the early US. Hans-Jürgen Grabbe in “European Immigration to the United States in the Early National Period, 1783-1820” estimates that there were still 414,800 immigrants from 1780 to 1820 (cf. page 194, the paper is available at JSTOR).

      The average white population (which is perhaps the relevant reference point here) was about 4.7 million. So a crude approximation for the annual rate would be 415K / 40 = 10,370 immigrants per year or 0.22% of average population. That’s low by later standards, but only slightly less than immigration now (from memory: 0.3%).

  2. Hello Mr. Walther, you seem to me the only one on this site who could be rational. Peter Brimelow is an extremist on one end just like how open border manics are on the other end. US is a nation of immigrants and Peter’s claim can’t refute that objectively and no one can. I think many open border manics rely on emotional arguments and erecting strawman to attack. This website reminds me of communist pages, were all arguments are made from a thin air of theory world and a complete lack of understanding of human nature and how history of nation states developed. The result of all the nonsense here (website level not you) is we have zero countries in the world with open-borders (EU restrict it on EU countries level). You can’t show up from outside of the Continent. And even that is under hammer after the Brexit vote. It’s unreasonable to argue for open-borders. Immigration policy should serve the people of each country first. So the numbers and kinds must be controlled. The Libertarian Party just like the Greens, have zero Presidents, zero US Senators or Congressmen, zero Governors and brace for it, zero state legislature seats from over 6000 seats. Libertariansim is a dogma, and customers (voters) have rejected it in the marketplace.

    Happy new year!

Leave a Reply