Garett Jones — Immigration, national IQ, & less democracy
What this covers
Garett Jones argues in conversation with Dwarkesh Patel that immigration policy and democratic governance rest on a shared error: treating individual-level effects as proxies for national welfare. Jones contends that migrants reshape host nations through cultural and institutional channels, not just labor markets, and that average cognitive traits carry externalities large enough to justify strict selection and a partial retreat from democracy in wealthy countries. The conversation moves across his framework: how national IQ predicts long-run productivity through positive externalities, why Secondary and Tertiary Consequences of immigration matter more than immediate fiscal gains, and how technocratic bodies like the Federal Reserve outperform elected institutions in rich democracies precisely because they are insulated from voter pressure.
Jones develops these claims through economics, history, and psychology. He grounds national productivity in Deep Roots—State history, Agricultural history, and Tech history together predicting institutional quality—and argues that migrants carry these traits into host countries via what he calls "spaghetti theory": cultural assimilation is bidirectional, with migrants and natives meeting in the middle rather than one simply adopting the other's values. He contrasts King and Council Equilibrium (oligarchy, not monarchy) with modern electoral democracy, and defends empowering elites only when constrained by competition, free press, and Amartya Sen's other markers of rich democracy. He deploys adoption data on IQ gaps, bond markets as signals of rent seeking, and the distinction between upper-tail intelligence (generating spillover innovations) and lower-tail effects (mostly contained through policy). On fiscal policy, Jones forecasts that bondholders will enforce austerity on welfare rather than permit high inflation. Throughout, he emphasizes uncertainty over outcomes—second-, third-, and fourth-order effects carry wide variance and potentially large downside tail risk, warranting caution regardless of first-order gains.
Jones argues that the long-run prosperity of nations is shaped less by individual migrant skills than by the average cultural traits and cognitive capital migrants carry with them, and that this externality logic justifies highly selective immigration and a modest reduction of democracy in favor of insulated technocratic institutions.
- Migrants reshape host-country culture and institutions through both votes and soft cultural contact ('spaghetti theory'), not just micro-level division of labor
- Mean/median IQ and deep-roots traits drive national productivity via positive externalities, so lowering the mean imposes long-run costs even when first-order migration effects look positive
- Insulated, expert-run institutions (like the Fed) and bondholder discipline outperform more directly democratic bodies in rich democracies
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Neoreactionaries romanticize monarchy mistakenly, because in equilibrium kings are almost always 'king and council'—something more like a corporate board than either a sole entrepreneur-owner or a stationary-bandit monarch—so the successful so-called monarchies aren't really monarchies but oligarchies.
“in equilibrium kings are almost always king and council. It's worth thinking through why king and council is the equilibrium.”
Cultural assimilation is bidirectional: when migrants arrive, host and migrant values 'meet in the middle' rather than migrants simply converging to natives, so apparent assimilation actually reflects migrants changing the host culture (as Italians made Americans eat spaghetti), and this happens through soft cultural contact, not just democratic channels.
“I call it spaghetti theory because when Italians moved to America, that got Americans eating more spaghetti... But migrants assimilate us.”
Within the normal range of variation, the upper tail of IQ matters more for national productivity than the lower tail, because elites generate scientific and intellectual breakthroughs with positive externalities that spill over to the whole world, while crime risk from the lower end is only probabilistic and prosperous nations have tools to limit the influence of the least informed voters.
“I think the upper tail is gonna matter more than the lower tail in the normal range of variation.”
The link between IQ and cooperation in repeated games is driven by Machiavellian intelligence that becomes Coasean intelligence—smart people find cynical, self-interested ways to grow the pie—rather than inherent prosociality; in the last turn of an iterated prisoner's dilemma the smart walk away, so it is endogenous, not exogenous, prosociality, which is a reason to worry.
“in the last turn they walk away. I think there is no correlation between intelligence and normal psychological agreeableness.”
The Fed functions as a non-corrupt technocracy because its officials have guaranteed long terms and it prints its own money to spend, forcing Congress to make a real effort to change anything—'a room of their own'—whereas the FDA and CDC must come to Congress for money annually and serve at the president's pleasure, leaving them in a realm of slack and reflecting Congress's muddled disarray rather than a culture of raw technocracy.
“In the case of the Fed, it really does matter that the people who run it have guaranteed long terms and they print their own money to spend. So that means that Congress has to really make an effort to change anything about the Fed. They really have the kind of independence that matters.”
The deep roots literature uses three predictors of modern prosperity—State history (S), Agricultural history (A), and Technology history (T, measured from 1500)—reflecting how many thousands of years one's ancestors lived under organized states, settled agriculture, and what fraction of the world's technology they used, and these predict modern prosperity especially when adjusted for migration.
“Two are S and A. That is State History and Agricultural History... And then the T-score is the Tech history score. I use the measure from 1500.”
Moving children from poor to rich countries (Swedish adoption data) closes about half the IQ gap between country averages in the short run, but since half the gap remains and IQ has large externalities, this still lowers the host country's mean test scores in the long run, which on average degrades institutional quality, productivity, and savings rates—a classic short-run versus long-run distinction.
“If you close half the gap between the poor country and the rich country, half the gap is still there. And if I'm right, that IQ has big externalities, then moving people from a lower scoring country to a richer scoring country and closing half the IQ gap still means on net you're creating a negative externality”
Handing power to elites works well in moderately-or-more democratic rich countries (Amartya Sen's definition: competitive parties, free press), but below the level of rich democracies, empowering elites tends to produce old-fashioned Gordon Tullock rent seekers who steer everything toward themselves and ignore the masses—which is why '10% Less Democracy' targets only rich democracies.
“Once we get too far below the rich democracies, I figure once you put elites in charge, they really are just gonna be old-fashioned Gordon Tullock rent seekers and steer everything toward themselves and not give a darn about the masses at all.”
America is more libertarian than the median voter's stated preferences because politicians facing reelection trade off between what voters say they want and the economic growth that actually helps them win, mediating voters' contradictory demands—'I don't want you to pull off the bandaid, but I want my wound to get better'—so delegating authority to elected politicians yields some benefits of elitism even within a democracy.
“politicians facing reelection have this tradeoff between giving voters what the voters say they want and giving the voters the economic growth that will help the politicians get reelected.”
Bondholders will win in the long run, so super-high inflation is not the future of the US; instead the fiscal gap will be closed by a mixture of tax hikes (potentially a large overt or disguised value-added tax) and slower spending growth—especially making Medicaid and Medicare worse over time—balancing the long-run budget on the backs of the poor and middle class rather than via hyperinflation.
“the way the bond holders are gonna win is that there's gonna be a mixture of tax hikes and slower spending growth, especially hurting the poor. And that's how the US is gonna close its fiscal gap.”
Even when first-order effects of moving people from poor to rich countries look good, policy should not simply do the obvious thing, because a compelling reason to think the second-, third-, and fourth-order effects are negative with wide variance means adding large uncertainty to outcomes that have a sizable negative tail—especially for the whole planet—which is not worth it.
“if you have a compelling reason to think that the direction of strength of the second and third and fourth order things are negative and the variance is really wide, then you're just adding a lot more uncertainty to your outcomes.”
Most corporations shouldn't live long, and a brutal bond market is a good market signal that moves capital from low-productivity to high-productivity projects—consistent with creative destruction research showing low-productivity firms exiting frees workers and capital—whereas the alternative is stereotypically Japanese zombie firms kept alive by politically pressured banks, tying up human and physical capital in low-productivity projects.
“having a lot of companies fade out is actually a good sign... low productivity firms exiting, just as naive laissez-faire predicts, means that those workers and that capital can get reallocated over to a more productive firm.”
Ethnic conflict's real source is when private cultural values are correlated with ethnic groups; if values are uncorrelated with ethnicity there is nothing to fight over, which is why many old US ethnic battles faded—but humans seek focal points and will use visible appearance as a horrifying focal point for conflict because our brains hunt for visual patterns.
“the real source of ethnic conflict happens when private values are correlated with ethnic groups. So if cultural values are basically uncorrelated with ethnicity, then basically there's nothing to fight over.”
Returns to national IQ are log-linear, which looks crudely like increasing or exponential returns, because human capital relationships are generally multiplicative rather than additive—tied to Adam Smith's pin factory and better organization of the division of labor between people.
“My findings were that it was all basically log linear. And log linear looks crudely like increasing returns. It looks exponential. So yeah, there's increasing returns to National IQ.”
The EU, despite annoying regulations, on net pulls Europe toward markets because countries getting ready to join improve their economic-freedom and laissez-faire scores on the path to accession (and may increase them slightly after joining), nagging less market-friendly post-Germany eastern members toward greater economic freedom—like a Rocky training montage.
“countries that are getting ready to join the EU improve their economic freedom scores, their laissez faire-ness, on the path to getting ready to join the EU.”
A simple change-on-change regression—increase in migrants from countries substantially more corrupt than average against subsequent change in economic freedom—yields consistently negative relationships (sometimes significant, always negative), a transparent test that immigration optimists like Powell and Nowrasteh never ran.
“we looked at migrants who came from countries that were substantially more corrupt than the country's average and we looked at the relationship between an increase in migrants from corrupt countries, and subsequent changes in economic freedom.”
The mean IQ of a nation is normally the best predictor of its productivity, and the exceptions where elite IQ predicts better are non-democracies and South Africa, such as Gulf states (Qatar, UAE, Bahrain) with large, well-educated migrant communities, suggesting the median matters more in democracies.
“if you look at the few countries that are exceptions to the rule, where the mean IQ is the best predictor of productivity compared to elite IQ. The exceptions are non democracies and South Africa.”
Bryan Kaplan's claim that deep-roots scores don't predict China and India's underperformance only tested the S and A measures, never the T (technology history) score; when T is tested, the deep-roots measure becomes statistically significant again and the puzzle disappears, since T has always been the best of the deep-roots scores.
“When Brian Kaplan made the claim that the ancestry scores... don't predict the low performance of India and China he only checked the S and the A in the S A T scores. Which letter did he not predict?... He never tested the T.”
Governments and productivity are shaped by the citizens of a nation even under authoritarian or totalitarian regimes, because rulers can only diverge so far from the masses on average, as recently seen in China where even a strong dictatorship had to listen to the people.
“even in a very strong authoritarian dictatorship, which some would call totalitarian, even there the government has to listen to the masses. So the government can only get so far away from the masses on average even in an autocracy.”
Frugality is good for the economy overall because being frugal builds up the capital stock and gives an invisible gift to future generations; leaving a large inheritance means producing a lot while consuming little, so Mormon frugality (and large bequests) is socially valuable even if individually one might optimally spend more for personal wellbeing.
“being frugal means you're building up the capital stock and you're giving a sort of invisible gift to future generations.”
Mormon high trust and tight-knit communities result from selection: harsh Western pioneer environments required high-trust communities, and over early Mormon history those willing to trust and be trustworthy stayed while free riders were weeded out, with the medium-high commitment of the religion (and the polygamy era) continuing to select for in-group prosociality.
“there was a lot of selection over the first few decades of Mormon history where those who were willing to sort of trust the group stayed in and those who weren't willing to trust the group ended up leaving.”
Persistent income differences between ethnic groups across the world reflect partly privatizable cultural traits—frugality, personal responsibility—that show up in surveys, persist across generations, and affect long-run family productivity, alongside the hive-mind relationship between test scores and productivity; not everything is literally an externality (e.g., education raises individual earnings).
“If you think about differences in frugality, differences in personal responsibility, which show up in the surveys, that are persistent across generations, those are likely to have an effect on long run productivity for yourself and your family.”
Low-skilled immigration simultaneously increases variance and lowers the mean and median, and the cost of pulling down the mean/median—especially in democracies—very likely swamps the small-probability benefit of admitting a few high-elite individuals.
“You didn't just increase the variance, you simultaneously increased the variance and lowered the mean and median, right?”
Effective altruists can raise national IQ through a 'Flynn cycle': public health and K-6 education interventions (iodine, lead reduction, childhood nutrition, preventing mild infections that cause 'health insults') boost test scores, raising productivity, which funds more public health in a self-reinforcing virtuous cycle, since part of the Flynn Effect reflects nutrition and brain health.
“This is what I call a Flynn cycle. I'm hoping for a world where there are enough public health interventions and probably K-6 education interventions to boost test scores in the world's poorest countries.”
While higher-IQ people are on average more cooperative, the evils of super-smart people are the largest risk to humanity—humanity is not threatened by people with sticks and stones but by nuclear weapons that only smart people could invent—and smart people committing fraud (Enron, FTX) tend to be very good at it.
“the evils of super smart people are obviously a huge risk to all of humanity. I don't have to worry about humanity being wiped out by a bunch of people with sticks and stones. I have to worry about humanity being wiped out by nuclear weapons, which could only be invented by smart people.”
It is fine for smart people to leave places like Kentucky for San Francisco because agglomeration matters: even with the internet making remote work easier, being in the room is still important, and finding role models to imitate (Hansonian and Girardian) is probably important for productivity.
“I'm a big agglomeration guy. The internet makes it easier but being in the room is still important.”
Voluntary genetic selection for intelligence is already a widespread cultural practice—anyone who has had amniocentesis has done a form of it—and welcoming voluntary embryo selection, likely later supported by government subsidies or tax credits, will probably have good effects on the future.
“We already do a lot of genetic selection for intelligence. Anybody you know who's in their mid thirties or beyond who's had amniocentesis, they've been doing a form of genetic selection for intelligence.”
Since the benefits of diversity are greatest when search and vetting costs are low, online rating systems (Yelp, Google) dramatically lower entry costs into non-elite professions, so the most important policy is not banning disclosure of data—including making it less risky to leave bad reviews of professionals like doctors—because the benefits of information flow outweigh the malicious-review problem.
“sites like Yelp and Google and all these online ratings have given us tools for checking these things out... 10-20 good Google reviews, and you can actually enter.”
The Liz Truss bond-holder revolt that ousted her after her tax reform demonstrates that, contrary to MMT and pop-Keynesian views that debt is no barrier, bondholders pay attention to long-term signals of fiscal policy credibility and will take action—we already live in a 'bondholder utopia.'
“That was a nice reminder that contrary to the MMT view and the pop keynesian view, the debt is no barrier at all. I think that showed that the bond holders are actually paying attention to long-term signals of fiscal policy credibility and they'll take action.”
Multi-level marketing schemes proliferate in high-social-trust communities like Utah because people predate upon that trust, making such schemes a rent-seeking downside of high social trust.
“Yeah, because people predate upon it. It strikes me as a total rent seeker sort of thing.”
Even though new ideas are getting harder to find, the awesome ideas still being produced are worth so much that we should not be less worried about impinging on innovation engines—if research still adds roughly a month of lifespan per year of effort (a year of life per decade), that is hugely valuable.
“new ideas are getting harder to find but the awesome ideas that we're still getting are still worth so much. If we're still increasing lifespan a month for every year of research we're doing, that just seems great.”
China is the poorest majority-Chinese country primarily because communism is a major mistake with decades-long negative effects on institutional quality, leaving China stuck in an institutional cul-de-sac; the issue is not non-linear ethnic composition effects, and North Korea would be an even bigger deep-roots outlier than China.
“I think it's just that communism is dumb and it has terrible decades-long effects on institutional quality”
On average we should presume that the skill level of voters and the ancestral traits we carry affect current national productivity, because everyone is an externality—positive or negative—on everyone else, though particular countries can be exceptions to the rule.
“It follows from thinking through the fact that we are all externalities, positive or negative... we should presume that the average skill level of voters, the average traits that we're bringing from the nations of our ancestors is having an effect on our current productivity.”
Like a Manhattan Project that puts only a small number of the smartest people in a room, you exclude the 20th smartest person because that person degrades the productivity of the other smart people—illustrating that adding anyone less able imposes a relative externality on the most able.
“part of the reason you don't want the 20th smartest person in the room is because that person's gonna ruin stuff for the other smart people. It's amazing how your worldview changes when you see everybody as an externality.”
Immigration policy should select only the very best people rather than also admitting those merely equivalent to current Americans, because the world depends on US innovation and every additional less-than-best person imposes a relative negative externality on the most productive people—the median and mean skill level shape everyone's productivity, so the second-best turns hundred-dollar bills into eighty-dollar bills.
“the entire world depends on US innovation. So we should make America as good as possible, not just slightly better than it is.”
FTX's collapse was likely old-fashioned financial fraud—commingling funds by lending FTX customer money to a hedge fund that lost it—driven partly by lacking a good board or oversight.
“So it's old fashioned financial fraud partly driven by not having a really good board over them or a good oversight.”
The thesis is not that migration is bad or migrants are bad, but that migrants bring cultural values from their old country to their new one—values that may be better, worse, or debatable relative to the host's.
“My story is that migrants bring cultural values from their old country to their new country. Sometimes those cultural norms are better than what you've got, and sometimes they're worse than what you've got.”
If open-borders advocates are right that mass migration has no effect on institutional quality, the experiment should be run in a lightly populated, high-institution country like Iceland—which could reach Singapore-level density at about 300 million people—rather than in America, so as to preserve the institutional quality of the innovation-powerhouse 'I7' (China, Japan, South Korea, US, Germany, UK, France).
“I would prefer to have open borders for Iceland. If the Open borders advocates are right and we'll have no noticeable effect on institutional quality, then it's great to have our open borders experiment run in a country that's lightly populated, has a lot of open land, and has good institutional quality.”
Poor countries with low deep-roots SAT scores can raise their prospects by creating a welcoming environment for large numbers of persistent Chinese migrants, following the Southeast Asian model, because China uniquely combines a high-SAT-score population, large population size, and enough lingering economic failure that emigrants might still move to a poorer country.
“One thing they could do is create a welcoming environment for large numbers of Chinese migrants to move there persistently... It's the one country in the world with large numbers of high S A T score, a large population and it's enough of an economic failure for at least a little longer that folks might be interested in moving to a poor country with lower s a t scores.”
Rather than increasing presidential power, the power of independent committees should be increased: FDA and CDC reform should make them more like the Fed, with appointed experts on long terms who feel free to ignore Congress and build their own technocratic culture.
“The power of independent committees should be increased. More Congress should be like the Fed. My plan for FDA or CDC reorganization would be to make them more like the Fed, where they have appointed experts who have long terms”
If society enters a zero-sum mode with little economic growth, where increasing one's share of the pie requires cutting bigger slices for oneself, you have to watch out for high-IQ people becoming predatory—like the Middle Ages.
“Then you gotta watch out, right? It's like the Middle Ages right there.”
Because large language models can simulate moderately complex human interactions (ChatGPT measuring around IQ 85), people whose work consists of words, dispensing simple advice, and solving simple problems face rough news for their careers, at least during the transition.
“I think that's gonna be rough news for a lot of people whose life was in the realm of words and dispensing simple advice and solving simple problems. That's pretty bad news for their careers.”
The value of open borders derives from exogenous institutional quality, not from agglomerations of talent and capital: moving more people into places with moderately laissez-faire institutions raises their productivity without reducing the productivity of incumbents, which is why the destination's institutions—not its existing talent clusters—are what matter.
“My version of what I've been exposed to as open borders, is that institutions exogenously exist. Some places have moderately laissez-faire institutions in their country and moving a lot more people there will not reduce the productivity of the people who are currently there”
America is more laissez-faire than Europe partly because of selection in migration: the Europeans with the most 'gumption,' openness, and adventurousness left for America, making the US population more market-friendly than the one they left behind.
“the Europeans with the most, what my dad would call gumption, are those who left and came to America. Some openness, some adventurousness.”
Market signals for allocating intelligence aren't terrible but innovation is full of market failures, so smart people should not solely follow market signals; they should be more Hansonian, playing a variety of status games—the academic and intellectual status game is worth a lot personally and generates positive spillovers for society.
“I wouldn't encourage them to solely follow market signals. I'd encourage them to be more Hansonian and play a variety of status games because the academic and intellectual status game is worth a lot, both personally and then it leads to positive spillovers for society.”
The impact of migrants on a host country divides roughly 50-50 between formal democratic influence (voting) and non-democratic cultural influence, since citizens place limits on government anyway and also work in firms establishing the complex exchange networks crucial to high productivity.
“I'm gonna call it 50-50. The way citizens influence a country through formal democracy is important but citizens end up placing some kind of limits on the government anyway.”
The IQ externality estimate (e.g., a million cloned von Neumanns raising US GDP by about 1.26%) is plausible but would not happen overnight, since IQ externalities operate over two to three generations and resemble what economists call organizational capital.
“I tend to think of the IQ externalities as being two, three generations. I lump it in with what economists call organizational capital. That sounds about right.”
An immigration system should not yet use deep-roots scores as hard quotas or cutoffs because the literature is roughly where Friedman's monetarism was in the late 1960s—full of good ideas but not policy-ready—so after a 10-20 year research project one might at most use deep roots as a small plus factor in a points-based system.
“Right now with the deep roots literature, we're about where Milton Friedman's Monetarism theory was in the late sixties.”
Roughly 20-30% of the world's country averages have cultural values equivalent to or better than the average of the United States, an off-the-cuff estimate by country averages.
“Equivalent to or better then? Just off the top of my head, maybe 20%? I dunno, 30%?”