What this covers

Gregory Clark speaks with Tyler Cowen about social mobility across centuries and societies, arguing that relative social position has barely budged despite radical changes in institutions, geography, and governance. The core claim rests on surname studies showing that status persistence rates are nearly identical in medieval England, modern Britain, and Denmark—suggesting that where you end up depends far less on the institutions around you than on inherited traits and chance. The conversation moves from England's medieval and industrial history through modern Scandinavia, the American immigrant experience, and India under British rule, each case deployed to test whether mobility actually responds to policy or institutions.

Clark distinguishes sharply between relative vs absolute mobility: when Britons migrated to Australia or Canada, their material living standards rose dramatically, yet their rank relative to others stayed constant. He extends this logic to argue that assortative mating in social status—people consistently matching with partners of similar standing—perpetuates inequality across generations not by changing average ability but by widening its distribution. He addresses the O-ring Model of Production and why 19th-century cotton mills succeeded despite low-skilled labor, uses Hungarian communist records to show that regimes cannot suppress inherited status, and turns surprising scrutiny on whether standard mobility studies actually measure what they claim. The discussion includes Clark's claim that pre-industrial Northern Europe underwent "Survival of the Richest"—a genetic sorting that made meritocracy, by his definition, largely true centuries before the modern era—and his argument that true equality would require abandoning the high-ability concentrations that drive growth.

Sharpest takeaway

Clark argues that relative social mobility is remarkably low and stable across centuries and across radically different societies, because social status is transmitted largely through genetics and randomness rather than environment or institutions—meaning most societies, including pre-industrial England, were effectively meritocratic long ago.

  • Surname studies show status persistence rates that are nearly identical in medieval England, modern Britain, and modern Denmark
  • Migrants preserve their relative social position even as their absolute living standards rise dramatically
  • Assortative mating in social status is a constant feature that widens ability distributions over generations

The argument · threads10 threads · 39 claims
0.78

Assortative mating based on social status widens ability distribution and drives both growth and inequality.

6 pointscentrality 5/5
  • Records of 1.7 million English marriages from 1837 onward show people consistently match with spouses who are as close to them genetically as their siblings, selecting on social status rather than looks, youth, or height; this does not change the average ability level but, repeated over generations, widens the distribution of ability and increases inequality.

    they're matching with people that are as close to them, essentially genetically, as their siblings in marriage

  • If you believe the engine of a high-tech society is the top 1-5% of the ability distribution, more assortative mating raises growth; but assortative mating also produces more inequality, so the only way to compress inequality would be much more random matching—revealing a fundamental tradeoff between growth-driving extremes and equality.

    But it does produce more inequality, so if you're worried about inequality in society, you don't want assortative mating. The one way to correct a lot of inequality would just be to have much more random matching.

  • Despite Denmark's compressed income distribution, free education, and subsidies, Nordic countries have not reduced the inequality of PISA test-score measures compared to much more unequal societies like the United States, because people in Nordic societies still mate very strongly assortatively.

    Nordic countries have not reduced the inequality of PISA measures compared to much more unequal societies like the United States.

  • Cousin marriage is poor for assortativeness because most people lack many cousins of the right age and gender, forcing essentially random cousin matching; systematically carried out, it should reduce the distribution of abilities within a society compared to the tightly assortative Northern European pattern.

    most people don't have a lot of cousins at the right age and gender, so it ends up that you have to marry a random cousin

  • Contrary to the common idea that women marry up because men value physical appearance more, analysis of English marriages back to 1837 (Clark and Cummins, PLOS One) finds no hypergamy—men and women end up matching in social status, showing complete dominance of social status in determining unions.

    for England, going back to 1837, we claim that there is no hypergamy, that essentially men and women end up matching in social status in marriage

  • Although the general pattern of strong assortative mating lets Clark predict the kind of person someone will marry, it provides no basis for marriage advice because we don't know whether highly assortative marriages make people happy—perhaps everyone in such marriages is unhappy and would be better off choosing on romance.

    Do we know that that pattern makes people happy? Maybe everyone with highly assortative marriage is actually unhappy

0.70

Social mobility rates have not improved in three centuries despite radically different institutions.

4 pointscentrality 5/5
  • Relative social mobility rates in a society like England have not improved in the last 300 years; rates in medieval England were just as high (or low) as they are now, and dimensions like social neighborhood and group affiliation show no improvement either.

    if we even go back to medieval England, rates of social mobility were just as high as they are now

  • Despite radically different social institutions—Denmark's free education to age 24, living subsidies, and compressed income distribution—social mobility rates in modern Denmark are no higher than in Britain, showing that institutions have little influence on the connection between parents' and children's social outcomes.

    social mobility rates in modern Denmark are no higher than in Britain. It's not correct to think of Britain as an immobile society.

  • Standard mobility studies overstate mobility because error in measuring status (e.g., crude years-of-education proxies) reduces parent-child correlation and thus appears as higher mobility; once you correct for these errors, social mobility rates are much lower and do not vary between Scandinavia, modern England, and 18th-century England.

    if there is error in measuring the status of somebody, it tends to increase social mobility because there's less correlation

  • Denmark proves that a society can impose very high tax rates and a large amount of income compression and redistribution while remaining a highly productive, technologically innovative society—so heavy redistribution need not destroy economic functionality.

    tax rates on people are so high, but it's still a highly productive, technologically innovative society

0.66

Physical migration between societies preserves relative rank while raising absolute material welfare dramatically.

3 pointscentrality 5/5
  • When people move from Britain to Australia, Canada, or the United States, their material welfare increases very significantly, but their relative social position does not change and is somehow preserved across these different societies—showing physical mobility raises absolute outcomes without altering relative rank.

    we can actually trace people who move from Britain to Australia, Canada, and the United States. What's interesting is that their material welfare increases very significantly, but their relative social position doesn't change and somehow is preserved across these different societies.

  • Among migrants from Latin America to the United States, the further away the origin country is from the US, the better the migrants tend to do and the more positively selected they are (e.g., Chileans are strongly positively selected)—consistent with a distance-based selection effect.

    the further away the country is from the US, the better they tend to do, and so, the more positively selected they tend to be

  • Moving people between societies is an enormous source of economic gains; if America takes in more immigrants it will benefit the world at no cost, so a great deal of substantial economic gain is available simply by moving people around, and this is where economic policy thinking should focus rather than trying to change relative outcomes within a society.

    if America takes in more immigrants, it will benefit the world at no cost

0.56

Genetic inheritance and randomness determine social position; meritocracy was achieved by 1700.

4 pointscentrality 4/5
  • A truly meritocratic society is one where the only things determining social position are genetic inheritance and randomness; by this definition England already achieved meritocracy by around 1700, meaning meritocracy was reached hundreds of years before the modern era.

    the only things that really determine social position are going to be people's genetic inheritance and randomness. That status seems to have been achieved in England already by 1700.

  • A detailed genealogical study of 300 years of English families testing all predictions of a genetic model (birth order, gender, mother vs father, family size, parental death, grandparent and interaction effects) concludes we are mainly the product of our genetics and randomness, with randomness a huge component—implying England has largely been a meritocracy for 300 years and, half-jokingly, that spending lots of time with your children doesn't matter.

    The conclusion in all of these studies is, we're mainly the product of our genetics and randomness. Randomness is a huge component as well

  • Pre-industrial Northern European societies, not just England, exhibited 'survival of the richest' for hundreds of years before the Industrial Revolution, where genetic differences between those at the top and bottom of the social ladder gradually changed the underlying genetics of these societies.

    all of these societies show this phenomenon of survival of the richest, going on hundreds and hundreds of years before the Industrial Revolution

  • Recognizing that many outcomes are the product of immutable inheritance and randomness—accidents of family history we don't control—implies that a good society is one that does not magnify those differences but compresses them through redistribution; hereditarianism and a Denmark-style redistributive welfare state go together.

    a good society is one that doesn't magnify those differences between people, but rather can seek to compress them

0.53

Relative position persists across different regimes because it is determined in a physics-like way.

3 pointscentrality 4/5
  • In Hungary under Communism, surname analysis shows descendants of the previous elite preserved their status in spheres like medical school admission; in the political sphere the elite disappeared temporarily but re-emerged fully once the free market was restored, demonstrating that relative position is determined in a physics-like way regardless of regime.

    in things like admission to medical school, it didn't change under the system of Communism. The same elites preserved themselves.

  • Although the North of England is much less successful on education, output, employment, health, and longevity, a comprehensive genealogical database controlling for parents' status shows very little independent cost to being born in the North; outcomes are overwhelmingly predicted by parental status, and the South simply contains more high-status people—bearing on the Heckman-Chetty debate over family versus neighborhood effects.

    This new book says the growth of the South and the decline of the North is all attributable to self-selection within Britain.

  • While economics achieves little predictive power about economies, social mobility is stunningly deterministic at the group level: in English data you can predict what will happen to a surname group 300 years later from its current position, which Clark regards as the only path forward and a genuine scientific gain.

    in the English data, you can predict what will happen to people 300 years later based on their position now

0.44

Only mechanistic, empirically testable explanations count; humanities-based cultural theories do not qualify scientifically.

8 pointscentrality 2/5
  • Pluralistic, humanities-based cultural explanations for why nations succeed or fail (e.g., 'they got their act together') do not work because they are not empirically testable; the explanations ordinary people offer (too many people, corrupt politicians, education) are easily testable as false, so only mechanistic, closed-form, empirically testable explanations count as more than mere conversation.

    We can only explain things if we have these mechanistic, closed-form, empirically testable explanations. Otherwise, it's just conversation.

  • Surname analysis shows nearly all European-descended groups in the US are very average, indicating 19th-century migration out of Europe was not strongly selected; the evidence even shows Scandinavian and possibly Irish emigration was negatively selected, while Quebec French-speakers became a US underclass—contrasting with the strong positive selection of Chinese, Indian, and Jewish immigrants who form the new American elite.

    Basically, all European groups within the United States look very average.

  • Societies can be ranked by how old they are (China extremely old, much of Southeast Asia younger), and Chinese laborers exported to plantations across Indonesia, Malaysia, and the Philippines became the entrepreneurial class everywhere—suggesting the deep history of a society, shaped by thousands of years of Malthusian selective pressure, predicts how its people perform in modern economic competition.

    laborers exported from China to work on plantations elsewhere in Southeast Asia have become the entrepreneurial class in all of those societies

  • Scotland is doing relatively well economically only because of generous subsidies from the South of England paid to keep it in the UK—including free university education that forces Scottish universities to charge English (but not, under EU rules, German) students—so Scottish independence would be like a parasite detaching from its host.

    when Scotland was considering independence, this would be like a parasite detaching itself from the host, because in order to keep Scotland within the UK, they have been paying out these very generous subsidies

  • Post-1960 immigration is dramatically transforming English society: white British are becoming the lower class while Africans, Bangladeshis, Pakistanis, and particularly Hindu Indians and Chinese form substantial new upper classes; because immigrant fertility is higher, a large fraction of 15-year-olds now come from these origins, which complicates calculating social mobility rates but has been an enormous benefit to English society.

    the white British now are the lower class of English society. There's emerging very substantial new upper classes, which are Africans, Bangladeshis, Pakistanis, and particularly Hindu Indians and Chinese.

  • People, particularly in Britain, have a deep fascination with relative position in society—who attends elite universities, who becomes members of parliament, and the backgrounds of such people—making relative social mobility an important feature of societies even if material living standards have risen fifteenfold since 1800.

    people do have this fascination with relative position in society, and people are obsessed about who's going to the elite universities, who are the members of parliament

  • The gain in material living standards is in some ways far less important than commonly believed; Clark says he would be perfectly happy to move into Dr. Johnson's early-18th-century London house today, with healthcare being the main genuine improvement.

    I've gone to Dr. Johnson's house in London from the early 18th century. I'd be perfectly happy to move in and live in that house now.

  • Successful immigrant groups assimilate culturally with surprising plasticity and speed—Huguenots driven from France in 1685 (about 1% of England's population) were completely assimilated within one or two generations as seen in naming and marriage patterns, yet their descendants remain four times more likely to attend Oxford or Cambridge—while groups that fail economically face barriers to assimilation.

    You can see them within one or two generations completely culturally assimilated.

0.39

Pre-industrial Malthusian equilibrium could support high living standards without poverty through population control.

3 pointscentrality 3/5
  • Malthusianism is widely misinterpreted as being about low living standards; properly, a Malthusian society only requires a positive association between fertility and living standards plus limited material resources. Post-Black-Death England (under 2 million people in 1450) and Tahiti were rich Malthusian societies—Tahiti maintained high living standards through murderous internal warfare and infanticide killing a third to half of children—proving the equilibrium need not imply poverty.

    This is a widespread misinterpretation of Malthus, that Malthus is about low living standards.

  • In agriculture-dependent sub-Saharan societies like Malawi with a quasi-Malthusian link, economic progress gets eaten up by population growth so living standards have not risen for 50-60 years; aid still benefits people's living standards and is worthwhile, but a substantial portion induces more population growth, undercutting its ultimate effects—so transforming societies off limited agricultural resources would be far more effective.

    There are societies like Malawi in sub-Saharan Africa where there is consistent economic progress, but that progress gets eaten up in terms of population growth.

  • Higher pre-industrial material living standards in Northwest Europe versus Asia show up biologically today: Europeans consumed expensive calorie sources like alcohol and dairy, which now manifests as greater European tolerance for alcohol and ability to absorb milk, while Asians who consumed little of these lack that tolerance.

    people at higher living standards consume alcohol. It's an expensive form of calories. In Asia, there's very little consumption of alcohol, and that actually shows up now in terms of people's tolerance for alcohol.

0.35

Britain's economic decline since World War I stems from upper-class fertility collapse and elite shrinkage.

2 pointscentrality 3/5
  • From about 1850 to the 1920s-30s, an astonishingly low fertility (one to one-and-a-half children per family for 50-60 years) emerged among the British upper classes, compounded by 20% of upper-class men dying in World War I, dramatically shrinking the relative size of the upper class and offering a potential explanation for Britain's lukewarm economic performance and declining entrepreneurship since WWI.

    There was a period from 1850 till the 1920s or '30s when an astonishingly low fertility emerged amongst upper classes in Britain.

  • The growth of the South of England and decline of the North is entirely attributable to self-selection within Britain: more talented and able people out-migrate from places like Glasgow and the North, while less skilled people migrate the other way, so apparent regional differences reflect who moved rather than place effects.

    This new book says the growth of the South and the decline of the North is all attributable to self-selection within Britain.

0.25

India's failure to industrialize despite institutional reforms and a preserved Brahmin elite defies explanation.

3 pointscentrality 2/5
  • India remained stunningly stagnant even after the British introduced civil and legal institutions, free trade, free capital markets, low interest rates, and welcomed entrepreneurs—and despite a thousand-year preserved Brahmin elite that becomes a super-elite when migrating to the US—and the failure of India to industrialize under the British defies serious explanation; in 1940 not one of 450 Indian textile mills reached the average productivity of English mills.

    In 1940, there were 450 Indian textile mills. There was not a single one that achieved the average level of productivity of the English mills.

  • The major industries of the 19th century like cotton textiles were failure-resistant: an entire cotton mill could run with three skilled managers and low-level tasks required no literacy (in Japan performed by teenage girls); errors were observable and repairable, unlike modern high-tech manufacturing where worker failure imposes huge costs—undermining the O-ring explanation for India.

    You could run an entire cotton mill with three skilled managers, and the tasks of people at the lower level there are extremely simple, don't require literacy.

  • There are essentially only two great problems in the economic history of the world: the incredibly long delay in the arrival of the Industrial Revolution, and the incredible divergence since it, producing differences of about 50 or 60 to 1 in material living standards between societies.

    there's essentially only two great problems in the economic history of the world. The first is the incredible long delay in the arrival of the Industrial Revolution. Then the second puzzle is the incredible divergence of society since the Industrial Revolution

0.21

Measurement error in status proxies obscures true social mobility rates in standard mobility studies.

2 pointscentrality 2/5
  • Living standard series in A Farewell to Alms use a bare-bones basket (food, housing, clothing) because they reflect what the average person actually consumed; this creates a 'tyranny of the majority' where significant technological advances in goods like books (a tiny share of consumption) don't register, which is why the Industrial Revolution appears abrupt—advance finally reached a mass-consumed commodity like clothing.

    if we weight technological change in economic terms, there's this tyranny of the majority, which is what do the majority want to consume? That's determining what the measured rate of technological advance would be.

  • Genome-wide association studies, even with up to 4 million genotyped people and a thousand identified loci affecting educational outcomes, still cannot mechanically locate the genetic basis we know exists from twin studies; combinatoric effects appear largely ruled out (effects are mostly additive), implying you may need samples larger than all of humanity to identify the rare permutations involved, and small populations like Native Americans may never be studyable this way.

    even with something like, is it up to 4 million people that they have genotyped... still don't identify that well what the genetic basis of this is that we know exists from these twin studies