What this covers

James Heckman, Nobel laureate economist, sits down with Russ Roberts to challenge the premise that welfare-state generosity and equalized resources can break cycles of inequality. Drawing on his research in Denmark with Rasmus Landersø, Heckman argues that despite the Danish state's provision of free education, subsidized childcare, and universal pre-K—resources far exceeding what the United States offers—intergenerational economic mobility there matches American levels. The conversation ranges across family influence, educational policy, and neighborhood effects, with Heckman pressing a single point: family dynamics and parental capacity matter far more than money or location in determining whether children escape their parents' economic position.

The argument centers on three overlapping claims. First, Chetty's influential work on sorting and neighborhood effects confuses family-quality differences with true neighborhood influence—when family attributes are properly controlled for, neighborhoods lose their predictive power. Second, universal programs that appear egalitarian often amplify inequality by design: educated parents navigate them more effectively, turning universal access into a mechanism for advantage. Third, interventions that work—teaching parents how to engage children, or programs linking subsidies to school attendance—succeed through family engagement, not cash transfers. Heckman also critiques the assumption underlying Chetty's causal claims about neighborhoods, challenges the statistical significance of findings in the Atlas of Opportunity, and argues that recent declines in American mobility partly reflect one-time post-Depression recovery rather than a stable underlying trend. The conversation extends into questions of human flourishing, work as meaning-making rather than mere disutility, and whether policy centered on utility as consumption misses what people actually value.

Sharpest takeaway

Heckman argues that Denmark's welfare-state policies equalize material resources but fail to break the powerful intergenerational transmission of skills and outcomes through families, so social mobility is no greater than in the US; he contends that family influence, not money or neighborhoods, is the fundamental driver of inequality.

  • Intergenerational skill mobility in Denmark is about as low as in the US despite free education, childcare, and universal pre-K
  • Universal programs disproportionately benefit advantaged families who know how to use them, reducing rather than increasing relative mobility
  • Chetty's 'neighborhood effects' are largely proxies for unmeasured family-quality differences because families sort into neighborhoods

The argument · threads7 threads · 33 claims
0.76

Family interaction and parental guidance shape child outcomes far more than money can; direct family work succeeds where cash transfers fail.

5 pointscentrality 5/5
  • You can tax and transfer money from successful people, but you cannot tax and transfer them into being more or less effective parents; the early family years and parental guidance are not compensated by preschool, nurseries, childcare centers, or formal education, so a deeper approach than money is required to provide true social opportunity for children.

    “even though you can tax and transfer successful people, but you can't tax and transfer them and make them into more effective or less effective parents. Those are things that actually need a somewhat deeper approach”

  • Interventions in the US, China, and Jamaica that work with families—simply teaching parents how to interact with and engage their children—produce real, carefully evaluated benefits in traits like grit, demonstrating that the route to improving child outcomes is changing family interaction rather than transferring money.

    “these are interventions, by the way, that work with families. And, they encourage the families to interact with the child. That's all they do... we have a lot of studies on this now, and they're being watched and being carefully evaluated and finding real benefits. But it's not money.”

  • Educating mothers or caregivers about how powerful their role is—simply telling them they have an important role and can do something about it—works surprisingly well; many disadvantaged mothers rise to the occasion because the child is the one thing they value most, their creation and legacy, so showing people the consequences of responsibility and then leaving the choice to them is the most effective approach.

    “it's very surprising how many disadvantaged mothers will rise to that occasion. They will see it. Right? Because the child is the one thing they value the most in life. This is their creation. This is something that they--this is their legacy.”

  • Family life plays a central, fundamental role in shaping whether children succeed or fail through guidance, advice, molding of preferences, and the shaping of the child's sense of self; economic and social policies that ignore the family ignore a fundamental source of inequality, and material transfers cannot substitute for what families provide.

    “Family life plays a central role, and economic and social policies that ignore that are going to ignore a fundamental source of inequality.”

  • Inequality and child outcomes are not merely a matter of money, scale, or any measurable numerical dimension; there is an element of quality—the family environment that shapes a child's preferences, sense of self, and motivation—that is under-valued and largely absent from public datasets and policy discussions.

    “it's just not a matter of money. It's just not a matter of numerical, this, that, or the other thing... There's an element called quality. And, in this case, the family is playing a very important role.”

0.72

Chetty's neighborhood-effect research rests on false assumptions and methodologically flawed causal claims about geographic sorting.

4 pointscentrality 5/5
  • Magne Mogstad and coauthors examined the small zipcode-level neighborhoods in Chetty's Atlas of Opportunity and asked whether mobility patterns in one neighborhood are statistically significantly different from another; they found the differences were not statistically significant and could arise by chance, meaning the atlas is essentially numerology rather than hard evidence that neighborhoods are predictive.

    “just asking: Are the mobility pattern in one neighborhood statistically significantly different from those in another?... the answer was: Yes, it could. These were not statistically significant differences. These atlases were just basically artifacts. It was numerology.”

  • The key identifying assumption in Chetty's neighborhood work—that families decide where to live essentially randomly with respect to the age of the child—is crazy because parents intensely sort on school quality and crime when children are young (often settled by age five to seven), so the assumption of random assignment contradicts how families actually behave and invalidates the causal interpretation.

    “the key identifying assumption in the Chetty work... is that basically people decide where they live... basically randomly in terms of the age of the child... Anybody who has had kids knows that when you have, like, three- and four- and five-year-olds, you start thinking: What school is this kid going to go to?”

  • What Chetty's work labels 'neighborhood effects' are actually family-quality effects resulting from sorting: disadvantaged families with lower education, criminal histories, and lower incomes cluster into the same neighborhoods, so when family sorting is properly controlled for—as is possible with Danish data—the apparent neighborhood effects disappear, because a neighborhood is just an agglomeration of families.

    “what we call 'neighborhood effects' are really family quality effects. And, as a result of sorting that goes on in the labor market... we find that once we control for that kind of sorting process, these neighborhood effects go away.”

  • If a parent has committed a crime, especially gone to prison or been treated adversely by the criminal justice system, the child—usually the boy—is much less likely to avoid that problem in the next generation; there is a strong, well-documented intergenerational transmission of criminal activity that Chetty's data cannot control for.

    “If a parent--has committed a crime, especially, has gone to prison... the child--usually the boy--is much less likely to avoid that problem in the next generation. There's a very strong intergenerational transmission of criminal activity. Well-documented.”

0.63

Denmark's extensive welfare state, despite high taxation and universal services, produces intergenerational mobility no better than the US because family influence remains dominant.

5 pointscentrality 4/5
  • The roughly $2 trillion policy packages proposed under Biden—unlimited childcare, universal pre-K, free college tuition—are reincarnations of the Danish welfare state almost line by line, and since those policies failed to break family-based inequality in Denmark, they will produce material improvement but will not change the dependence of child outcomes on family quality.

    “you realize that these Biden formulas are basically reincarnations of the Danish welfare state, almost line by line. Unlimited childcare. Universal pre-K. Free college tuition.”

  • Denmark's celebrated low inequality is measured only after taxes and transfers; before taxes and transfers there remains a fair degree of wage inequality because people with different skills are paid different amounts, so the state has not eliminated inequality but only allowed high earners to keep less of their income while shaping the final consumption bundle.

    “Before tax and transfer, there is a fair degree of inequality--what the wages are before taxes and transfers... It's just that they're allowed to keep less of it.”

  • Despite Denmark's extensive welfare state providing free education, subsidized childcare, universal pre-K, and equalized school expenditures, intergenerational mobility measured in terms of skills and how the next generation actually fares is pretty much the same as in the United States, because family influence remains very strong.

    “intergenerational mobility--properly measured in terms of skills, in terms of how the next generation is really faring--is pretty much the same. Family influence is very strong in Denmark, about as strong as it is in the United States.”

  • When Denmark switched mid-20th-century from targeted educational programs for the disadvantaged to universal across-the-board programs, the social mobility gains of the earlier era vanished, because universal policies are a 'check' that more educated and affluent parents are better able to draw on to advise and reinforce their children, turning universality into a vehicle for relative social immobility.

    “around the middle of the century, the educational policies became universal. So, instead of going after the disadvantaged, they became across the board. And, what happened was, the social mobility that had been witnessed in the first half of that century, started to vanish.”

  • Denmark is best described not as a socialist state but as a social welfare state: very high taxation combined with very large provision of public services, with more than half of the benefits people receive coming from the state in one way or another.

    “It's not a Socialist state... A social welfare state, they would say: Huge tax agent; but huge provision of public services. More than a half, I think, of all benefits that people receive come from the state”

0.63

Policies promising material benefits without effort and conditions erode personal agency, accountability, and the dignity of autonomous striving.

5 pointscentrality 4/5
  • The traditional labor/leisure trade-off in economics treats work as a disutility offset by leisure's value, but many people genuinely enjoy their jobs and derive meaning from them, so framing income replacement plus leisure as welfare-neutral misrepresents how people actually experience work.

    “I remember Victor asked me a very good question... he says, 'You have this idea of leisure, but you realize that people really enjoy their jobs.' And, he's right.”

  • Modern electoral politics, in which competing groups can only promise more and more goodies achievable without effort, is building a dependency state that, through entitlements, slowly erodes people's ability to operate in an autonomous fashion.

    “we really are building this dependency state and with the entitlements and with this whole sense, we're slowing eroding this ability to sort of operate in an autonomous fashion.”

  • There is a widespread loss of personal accountability that has been sanctioned and reframed as liberating—seen in people having children without taking responsibility for their wellbeing—and this same erosion of duty to the larger population manifests in resistance to public-health measures like masks and vaccination, which is fundamentally about externalities, not liberty.

    “There's a loss of accountability that somehow has been sanctioned. It's been sanctioned, and it's viewed as liberating”

  • A Danish program targeted at disadvantaged dropouts gave a subsidy that effectively induced participants not to go to school and not to work, so while they were equalized at age 17-18, by age 35-36 they lacked the skills of comparable people who had not participated, demonstrating powerful disincentive effects from guaranteed income.

    “the subsidy that they were given was a subsidy that more or less induced them more or less not to go to school and not to work... yes, they were equalized when they were 18 and 17, but when they were 35 and 36, they lacked the skills of those who had not participated in the program.”

  • What ultimately matters to human flourishing is agency, dignity, and striving—the process of responding to and getting to a challenge rather than the final goal itself—so the happiness movement's hedonic focus on simply 'being happy' misses that people derive deep pride from accomplishment and the instinct of workmanship, and policies that promise goodies without effort erode this autonomous capacity.

    “it's about the value in life is responding to the challenge. It's not the final goal: it's getting to the goal. It's striving.”

0.57

Universal programs become tools for advantage when affluent parents navigate them better, while targeted programs for disadvantaged groups prove more effective.

7 pointscentrality 3/5
  • Despite Denmark's reputation as a free-love country with high cohabitation since the 1960s-70s, cohabitation there is actually a very stable form of relationship, and once children are born to a cohabiting couple they frequently marry or stay together, so children grow up in stable two-parent homes with substantial support—the family is a much more stable entity than the 'free-love' reputation suggests.

    “what you'll see is that in Denmark, there's a tremendous--even the cohabitation turns out to be a form of a relationship that's very, very stable. And, once children are born to a cohabiting couple, frequently they marry or they stay together”

  • Government should not try to force stable two-parent families through coercion, but the one defensible government policy is voluntary programs that build the role of parents as guides and cultivators of their children's morals, motives, and academic achievement—entering homes that families are free to refuse—because these tap into parents who care and can be very successful.

    “the only government policy that I favor in this regard... a government policy that kind of builds the role of parents as a guide, as a cultivator of the morals and motives, and academic achievements of the children--programs that kind of encourage parents or caretakers... can be very successful. That's completely voluntary.”

  • Even though Danish teachers are paid the same regardless of location, higher-quality teachers (measurable by their own school grades and college transcripts) tend to show up teaching in more affluent neighborhoods, because they can't be paid more money but can be given higher-quality students who make their lives more enjoyable—a non-market response to a market imperfection.

    “in more affluent neighborhoods, you find that the teachers who show up teaching there are higher quality... You can't pay them any money, but you can give them higher-quality students, which makes their lives much more enjoyable. So, it's a non-market response to a market imperfection.”

  • Chris Walters's study of school choice and charter schools in Boston found charter schools were particularly beneficial for the most disadvantaged children, yet among those eligible the children who actually enrolled were the most advantaged because their parents sought out the schools, illustrating how universal access creates inequality when the bottom of the distribution lacks informed parents.

    “he found that school choice was particularly beneficial--charter schools--very beneficial, and for the most disadvantaged children. But, and this is the interesting part, who are the kids that were going to charter schools? Among those eligible, the most advantaged.”

  • Sorting is an important part of any free-market system, and if anything sorting has increased over the last 20, 30, and 50 years, with much more sorting on income, education, and other traits that influence families.

    “if anything, sorting has gone up in the last 20, 30 years, the last 50 years. There's much more sorting on income, on education, on a bunch of traits”

  • When Heckman raised data and econometric criticisms directly to Chetty in front of hundreds of people at a Princeton lecture, Chetty refused to respond with any real give and take, which Heckman characterizes as a strategy that works when one has prestigious media support like the New York Times.

    “I was sitting there in a lecture hall in Princeton with hundreds of people in the room asking him, specifically... Which he refused to do by the way. There was very little response, very little give and take, more or less refused.”

  • Chetty's finding that the share of children out-earning their parents fell from ~70-80% to ~50% is misleading because the earliest cohorts were born/raised in the 1940s-50s coming out of the Great Depression while the US was still highly agricultural, so much of the measured mobility reflected the one-time recovery of the economy and the rural-to-urban migration as agriculture became productive, not a stable underlying mobility process.

    “the first cohorts in these Chetty studies start from people who were born or so were raised... in the 1940s. In 1940, we were coming out of a Great Depression, and we are leaving agriculture... a huge amount of it had to do with the recovery of the economy”

0.36

Conditional cash transfers and income subsidies work through incentive structures, not through money itself; unconditional transfers create disincentives.

2 pointscentrality 3/5
  • A study cited to support giving people money—showing that people starving to death get lower IQ scores than those who aren't—is a demented misuse of evidence, comparable to the trivial claim that people who decide too quickly make more mistakes than those who decide too slowly; it does not establish that cash transfers transform cognitive capacity.

    “one of the powerful pieces of evidence in support of giving people money by this guy was that the people who are starving to death get lower IQ scores than people who aren't starving. That's viewed as the importance of money. Come on.”

  • Programs claimed to show that giving families money improves child development often actually work through attached conditions—such as casino-revenue programs that paid only if children were sent to school, or Mexico's Progresa conditional transfers—so the apparent effect of money is really the effect of incentive conditions, especially in rural areas where children are agricultural labor.

    “the opening of casinos on Indian land apparently did lead to more Indian children going to school, except for one little detail: that some of these programs have had the feature that they could only collect the money if they sent their kids to school. So, it was basically what's called in Mexico, Progresa.”

0.32

Historical poverty reduction relied on cultural adoption of moral values, not government mandate; similar cultural force is needed today.

1 pointcentrality 3/5
  • Historically the response to 19th-century poverty was not a government mandate but a cultural adoption of values—Victorian norms promulgated partly by churches and religious revivals like the Great Awakenings—that motivated people to reform their lives and recognized that values like respecting laws, civility, hard work, and saving benefit society; achieving that moral force, rather than legislation, is what is needed today.

    “these Victorian values played a very important role, I think, in kind of motivating people... there was kind of the cultural force suggesting what appropriate behavior would be... how do you achieve that moral force? How do you achieve the recognition of those values? And, I think that is what probably needs to be attacked”