Dwarkesh PodcastGarett Jones — Immigration, national IQ, & less democracy
CHAPTERS
- 0:00 – 1:09
Framing the core worry: long-run innovation vs. open-borders optimism
Jones opens with a macro-level caution: policies that subtly reduce institutional quality or innovation in top countries could impose global costs that only show up over decades. He argues this long time horizon is often ignored in immigration debates and ties it to his broader skepticism about mass democratic decision-making.
- •Long-run (20–50 year) effects may dominate short-run gains
- •Global dependence on innovation from a handful of top countries
- •Why experimentation should happen in lower-stakes contexts
- •Institutions and productivity as fragile, path-dependent outcomes
- 1:09 – 2:52
Do migrants change countries more through voting or culture?
The discussion turns to how migrants reshape destination countries. Jones gives a rough “50/50” split between formal political influence (voting, constraints on government) and informal cultural transmission through everyday norms that affect firms and cooperation.
- •Estimated 50/50 split: political channel vs cultural channel
- •Even autocracies must track mass preferences to some degree
- •Culture shapes productivity via trust, exchange networks, and norms
- •People affect institutions not only at the ballot box but at work and in civil society
- 2:52 – 5:14
Hive Mind basics: why the smart tail matters—and why democracies care about the median
Jones argues national IQ/test scores correlate strongly with productivity, with the upper tail (elites) especially important for breakthroughs. However, he emphasizes that in democracies the mean/median matters because it influences institutions and policy, making shifts in average traits consequential.
- •Upper tail drives innovation spillovers; elite breakthroughs matter
- •Mean test scores often best predict national productivity
- •Exceptions: Gulf states with high-skill migrant elites; non-democracies
- •Median-voter dynamics make average skill more institutionally relevant in democracies
- 5:14 – 7:26
Variance vs. mean: why low-skill immigration could be costly in democracies
Dwarkesh probes whether increasing IQ variance could be beneficial even if it lowers the mean. Jones rejects this for democratic societies: lowering mean/median skill and norms can swamp any benefit from a thicker right tail because institutional quality is shaped by broad averages, not just elites.
- •Increasing variance while lowering the mean is a bad trade in democracies
- •Mean/median shapes institutions and long-run productivity
- •Elite gains don’t necessarily offset broad institutional degradation
- •Externalities framework: average traits affect everyone’s productivity
- 7:26 – 9:15
“Spaghetti theory”: migrants assimilate natives too (culture as two-way convergence)
Jones introduces his cultural-transplant mechanism: assimilation is bidirectional, with natives changing as migrants arrive. He argues institutions are downstream of culture, so even without voting, large inflows can shift norms that later influence governance and market institutions.
- •Cultural convergence is two-way, not one-way assimilation
- •Institutions are downstream of cultural norms
- •Soft social contact can alter trust, responsibility, and compliance
- •Examples: food and broader social changes across migration waves
- 9:15 – 12:03
Immigration, markets, and corruption: disputing the ‘no harm to economic freedom’ claim
Jones critiques findings (e.g., Powell/Nowrasteh) that immigration doesn’t reduce economic freedom. He describes a “change-on-change” approach: increases in migrants from more corrupt origin countries predict subsequent declines in economic freedom, often negative across specifications.
- •Proposed test: changes in migrant composition vs changes in institutions
- •Finding: increased inflows from more-corrupt origins correlate negatively with economic freedom
- •Critique: prior work didn’t run the most direct ‘change predicts change’ regression
- •Distinction: not ‘migrants are bad’ but origin-country norms can transplant
- 12:03 – 16:54
How selective should immigration be if US innovation is a global public good?
Dwarkesh presses for opening entry to migrants from countries with comparable traits to the US. Jones argues for prioritizing “the best” because second-best inflows can impose externalities on top performers and on the innovation engine that benefits the world, even if short-run gains look large.
- •Argument for ‘Why not the best?’ selection logic
- •US innovation as a global externality worth protecting
- •Deep Roots SAT (state/agriculture/tech history) as country-level signal
- •Tradeoff: bigger population vs lower average positive spillovers
- 16:54 – 21:38
China, communism, and Deep Roots: why ‘T’ (tech history) matters
Jones discusses why China is the poorest majority-Chinese country despite strong diaspora performance: communism as a persistent institutional shock. He defends Deep Roots metrics against critiques by stressing the importance of the tech-history (“T”) component, claiming it restores predictive power in disputed cases.
- •China’s underperformance attributed largely to communist institutional legacy
- •North Korea as an even starker (data-poor) example
- •Critique of Caplan: tested S and A but omitted T
- •Including T makes Deep Roots statistically significant again (per Jones)
- 21:38 – 30:38
Side effects, uncertainty, and the ‘run it in Iceland’ thought experiment
Dwarkesh argues for taking the obvious first-order gains of moving people to rich countries despite uncertainty. Jones insists economics demands attention to second- and third-order effects and suggests running radical open-border experiments in smaller, lower-stakes countries rather than core innovation hubs.
- •First-order gains vs uncertain long-run institutional effects
- •Large negative-tail risk for global outcomes if innovation declines
- •Proposal: run open-borders experiments in places like Iceland/France/N. Ireland
- •I7 concept: protect top innovation powerhouses
- 30:38 – 32:21
Are poor countries stuck? A provocative lever: welcoming Chinese migration (and its risks)
Asked whether low Deep Roots countries can escape, Jones suggests one actionable—if politically fraught—path: sustained inflows of high-SAT-score cultures, especially Chinese diaspora dynamics seen in parts of Southeast Asia. The conversation notes the severe backlash risks, including expropriation and pogroms.
- •Possible development strategy: attract high-trust/high-SAT-score migrant communities
- •Southeast Asia as historical partial model
- •Political risk: backlash, pogroms, expropriation (Indonesia; Uganda case analogy)
- •Deep Roots as persistent but not necessarily destiny
- 32:21 – 39:13
Effective altruism and raising cognitive capacity: health, nutrition, and Flynn-style cycles
Jones outlines interventions he thinks could boost test scores and productivity in poor countries: iodine, nutrition, lead reduction, infection control, and early education. He frames it as a virtuous cycle where better brain health increases productivity, enabling better public health and further gains.
- •Flynn-effect-inspired ‘virtuous cycle’ via health and early education
- •Iodine, childhood nutrition, reliable healthcare, infection reduction
- •Lead abatement as a major ongoing benefit
- •Brain health as neglected human capital policy
- 39:13 – 44:39
Von Neumann clones, talent agglomeration, and why IQ effects look exponential
Dwarkesh explores thought experiments about adding extreme talent and where intelligence should be geographically concentrated. Jones favors agglomeration and argues national IQ effects look roughly log-linear (appearing like increasing returns), plausibly via division-of-labor and organizational capital mechanisms.
- •A million von Neumanns: plausible multi-generation GDP effects
- •Innovation spillovers as a global gift—best located in top hubs
- •Agglomeration beats dispersion; ‘being in the room’ matters
- •Log-linear (exponential-looking) relationship between test scores and productivity
- 44:39 – 46:55
Genetic selection, AI ‘IQ,’ and what intelligence does (and doesn’t) guarantee
Jones supports voluntary embryo selection as a continuation of existing practices and expects policy incentives could emerge over time. They also discuss measured AI “IQ” and the labor-market disruption to word-based jobs, plus a key caveat: IQ correlates with strategic cooperation, not necessarily agreeableness or moral prosociality.
- •Voluntary embryo selection: ethically acceptable and already implicit in practice
- •Potential future subsidies/tax credits after experimentation
- •AI systems simulate many ‘moderately complex’ language tasks; transition shocks
- •IQ linked to Machiavellian/Coasian strategic cooperation, not inherent kindness
- 46:55 – 49:42
10% Less Democracy: why the Fed works better than the FDA, and the case for independent committees
The conversation shifts to institutional design: Jones contrasts the Fed’s durable independence (long terms, budget autonomy) with agencies like the FDA/CDC that rely on annual appropriations and serve at presidential pleasure. He argues more governance should resemble the Fed—expert-led bodies with real insulation to build strong internal cultures.
- •Institutional performance depends on independence structure and incentives
- •Fed: long terms + self-funding → credible autonomy
- •FDA/CDC: vulnerable to political meddling and congressional ‘slack’ dynamics
- •Proposal: redesign agencies to be more like the Fed (expert committees, long terms)
- 49:42 – 1:14:00
Europe vs America on markets, ethnic conflict over time, bondholder discipline, Mormon trust, and an immigration points system
In a wide-ranging final stretch, Jones argues the EU nudges members toward markets despite Europe’s less market-friendly electorate, and he explains why US policy can be more libertarian than the median voter wants. He then touches on whether ethnic conflict fades, how information tools reduce diversity’s vetting costs, why bond markets constrain fiscal policy (VAT, slower benefit growth), the selection story behind Mormon in-group trust, and his view that Deep Roots could eventually inform—but not dominate—an immigration points system. The episode closes with reflections on SBF/Enron-style fraud and how smart cooperation can be strategic rather than altruistic.
- •EU as a (modest) pro-market ‘nag’ relative to member baselines; selection into America
- •Median voter vs actual policy: politicians trade off pandering vs growth (Caplan)
- •Ethnic conflict: may fade when values decouple from visible groups; vetting aided by reputation systems
- •Bondholder democracy: markets enforce fiscal credibility; likely adjustment via VAT/taxes and slower benefit growth
- •Mormonism: high-trust culture shaped by selection and commitment mechanisms; trust can be exploited (MLMs)
- •Immigration design: Deep Roots not yet policy-ready; could become a small ‘plus factor’ after more research
- •SBF and cooperation: high IQ can enable sophisticated fraud; cooperation often strategic in repeated games
