Growing up in an innovative neighborhood increases the probability of filing a patent. Bell et al. (2019) show that children from families at the top of the income distribution are ten times more likely to patent than those below the median. McNamara, Neidhöfer and Lehnert, in a study published in Nature in July 2026, extend this diagnosis with European longitudinal data of unprecedented scope: the social mobility of a territory predicts its capacity for innovation better than its level of wealth.
The Essentials
- A region with strong intergenerational mobility produces approximately 11% more patent filings on average compared to a region of equal income level but with low educational mobility (McNamara, Neidhöfer, Lehnert, Nature, July 2026).
- Exposure to innovation in family or residential surroundings significantly increases the probability of becoming an inventor; Bell et al. (2019) document a ×10 factor between children of the top 1% and those below the median income, across all children; for children with good mathematics scores, the income gap persists but is not quantified at ×5 in the article.
- The mechanism transcends individual human capital: it is the territorial structure, social mix, density of exposure, permeability of networks, that determines who gains access to the technological frontier.
- Europe probably possesses an underexploited pool of talent in its low-mobility regions; the open question is that of policies capable of activating this potential without reproducing the logic of concentration in closed clusters.
Who Are These Authors and Why This Work Matters
Raj Chetty is the American economist who transformed the study of social mobility into a discipline of massive data. His work on individual trajectories in the United States, based on longitudinal tax records covering tens of millions of people, established that a child’s zip code is a significant predictor of their income in adulthood. Alex Bell, Xavier Jaravel, Neviana Petkova and John Van Reenen have each contributed to the economics of innovation, taxation and inequality. McNamara, Neidhöfer and Lehnert constructed for this study the EUROPE-IGM-ATLAS, a database aggregated at the level of European regions (NUTS1/NUTS2), cross-referencing indicators of regional intergenerational educational mobility with regional patent data. The result is published in Nature in July 2026: a study that shifts the debate on European innovation from the terrain of cluster policies to that of territorial social structures.
This shift is significant. The debate on European technological lag has for years revolved around the same diagnoses: lack of venture capital, fragmented markets, weak R&D spending, absence of world-class universities outside a few poles. The study by McNamara, Neidhöfer and Lehnert adds a variable that these analyses left aside: the social permeability of territories where innovation occurs, or does not occur.
The Thesis: Mobility as the Invisible Infrastructure of Innovation
The central argument is as follows. European regions with strong intergenerational educational mobility, those where children of poorly educated parents access higher education levels, produce more patents, at equal wealth levels, than low-mobility regions. The measured gap is approximately 11% additional patent applications on average for a territory that is no richer, no better equipped with universities, no more attractive to multinationals, simply more socially open.
This figure is an invitation to reframe the question. For a long time, regional differences in innovation were sought to be explained by classical inputs: capital, infrastructure, presence of higher education establishments. The study shows that these variables are insufficient. A region can concentrate all these inputs and produce few inventors if its social structures prevent children from modest backgrounds from entering the networks where innovation is transmitted.
For it is indeed about transmission. The second key result, the proximity effect, illuminates the mechanism. Bell et al. (2019) document a causal effect of exposure to innovation via family and neighborhood on the probability of patenting. Bell et al. (2019) document a ×10 factor between children of the top 1% and those below the median income, across all children; for children with good mathematics scores, the income gap persists, but a ×5 factor corresponding to this subgroup is not an identified result in the article. This effect is not reducible to genetics or household financial resources. It operates through exposure: seeing that it is possible, concretely understanding what it consists of, being directed toward studies that lead to it.
It is a cultural and professional transmission that only residential and social mixing can produce at scale.
The study follows in the wake of work begun with Bell, Chetty, Jaravel, Petkova and Van Reenen (2019, Quarterly Journal of Economics), which had shown the same phenomenon in the United States: the “lost inventors,” intelligent children from modest backgrounds who never filed a patent, represent a considerable shortfall in terms of innovation. The European version by McNamara, Neidhöfer and Lehnert (2026) confirms that the phenomenon crosses borders and education systems. Even the Nordic countries, renowned for their egalitarian school systems, display regional disparities in exposure to innovation that translate into inequalities in patent production.
Europe Facing Its Blind Spots
Europe has maintained for years an obsession with clusters: Silicon Valley à la European, technological hubs, innovation campuses, digital free zones. Cohesion policy, structural funds, Horizon programs massively finance the construction of these centers of excellence. The study by McNamara, Neidhöfer and Lehnert suggests that this strategy produces innovation in already-favored regions, but leaves intact the social barriers that prevent talents from other regions from accessing them.
These clusters concentrate already-trained inventors, with no mechanism for diffusion to territories where they are absent. A closed cluster reproduces its own elite. A region with strong social mobility generates inventors where they were not previously found. These two innovation logics have distinct long-term returns and are difficult to compare.
The Promise of the Diploma and the Glass Ceiling of the Market already explored this tension between formal access to education and actual conversion into upward trajectories. The Nature 2026 study extends it into a specific domain: patentable technological innovation. It shows that even formally open education systems are insufficient if informal exposure to innovation professions remains socially segregated.
One must here confront this thesis with a competing reading. Carl Benedikt Frey, in his recent work on the institutional conditions of technological progress, synthesized in How Progress Ends, argues that innovation trajectories are determined less by individual talents than by structures of interests in place: the vested interests that can block the diffusion of a technology, reserve its gains for a minority, or maintain sectoral barriers to entry. His thesis is that AI already follows this pattern: the tech giants who control it have powerful incentives to prevent broad diffusion of the tools and skills that could challenge their position.
The study by McNamara, Neidhöfer and Lehnert provides regional evidence for Frey’s thesis: low-mobility regions are precisely those where local interest structures, closed professional networks, co-opted recruitment, residential segregation, prevent the diffusion of innovation to new population segments. The convergence of the two readings is striking. Where Frey thinks in terms of monopolies and antitrust policies, McNamara, Neidhöfer and Lehnert think in terms of mobility and mixing. But both diagnoses point to the same obstacle: structures that reserve access to innovation to those already inside.
The tension between the two readings remains productive. Frey would insist on the necessity of competition policies to break the monopolies that capture AI gains. Bell and Chetty point toward territorial policies of social mixing and professional exposure. These two levers are not mutually exclusive, but their timelines and actors differ. Antitrust acts on markets; social mobility acts on territories.
The Nature 2026 data argue for not neglecting the latter in favor of the former.
The Study’s Blind Spots
A study published in Nature using European longitudinal data inspires confidence on methodological grounds. But several limitations merit being named.
The measurement of innovation by patent filing is a robust but imperfect proxy. Patents capture a specific form of innovation, technological, codifiable, legally defensible, and leave aside others: organizational innovation, social innovation, artistic innovation. Regions with strong mobility could produce other forms of innovation that patents do not measure. The study probably acknowledges this in its limitations, even if the available brief does not detail it.
The direction of causality remains the most delicate point. Does high social mobility produce innovation, or do innovative regions attract mobile populations: the authors have likely sought to control for this selection bias through their longitudinal methods, but uncertainty remains for those wishing to draw prescriptions for public policy. Creating mobility in a stagnant region operates according to different logic than that of maintaining mobility in an already dynamic region.
Finally, the proximity effect measured aggregates distinct mechanisms that would be useful to separate: school orientation by those close to you, access to professional networks, transmission of representations about what is accessible or not. Depending on which of these mechanisms dominates, appropriate policies differ considerably.
Implications for Prescription
If the conclusions of McNamara, Neidhöfer and Lehnert are confirmed upon replication, they invite a refocusing of part of European innovation policies on three levers that classical approaches underestimate.
Residential mixing in areas with high density of engineers and inventors. Housing policies have effects on innovation: if neighborhoods where technology practitioners live are inaccessible to modest-income households, the exposure effect cannot function. This is a dimension that neither innovation policies nor education policies ordinarily address.
Large-scale mentoring and professional immersion programs, which seek to simulate this proximity effect for children who do not have natural access to it. Experiments exist, in the United Kingdom, in Sweden, but their scaling remains limited. AI in Enterprise as Tool or as System raises a parallel question about the diffusion of technical skills: tools are insufficient if the people likely to benefit from them do not have access to environments where their use is naturalized.
The geography of public R&D investment, finally. Concentrating resources in already-innovative regions may maximize short-term returns, but reproduces exposure barriers in the long term. A diffusion policy that would finance the presence of engineers and inventors in territories with currently low innovation density could have a twenty-year return superior to that of a new campus of excellence situated next to an existing cluster.
Creating the Conditions for Exposure Where It Is Lacking
The issue at the 2040 horizon is this. Europe probably has potential inventors in low-mobility regions: intelligent, curious children, capable of abstraction, who will never become engineers because they will never have met an engineer. If the exposure effect measured by McNamara, Neidhöfer and Lehnert is causal, and not merely correlated, it becomes actionable.
Two trajectories emerge from the available data. The first consists of moving people: encouraging engineers and inventors to settle in territories with low innovation density, through tax incentives or telework policies that make these territories attractive for qualified profiles. Data from the Swedish Patent Office 2025, cited in the supports for this study, suggest that decentralization of qualified employment has measurable effects on local patent production in the next generation. This trajectory is appealing but slow: its effects are measured in decades, and it depends on residential choices difficult to orient.
The second consists of moving exposures without moving people: structured immersion programs, partnerships between technology companies and schools in disadvantaged territories, job shadowing internships starting in middle school. This trajectory is more immediate, but its proximity effect is artificial and perhaps less powerful than the residential cohabitation that Bell and Chetty measure. It can nurture vocations without changing the social structures that subsequently filter access to training and networks.
These two trajectories are not mutually exclusive. But they require different actors: housing and territorial planning policies for the first, education policies and corporate social responsibility policies for the second. The difficulty is that these policies do not fall under the same ministries or the same levels of governance, and that no current European institution thinks them together. The real limitation of the McNamara, Neidhöfer and Lehnert study may be there: it provides a precise diagnosis for which appropriate policy has not yet found an institutional home.
What the data permit us to affirm is more modest but already valuable: a region that increases its social mobility increases its innovation production. This link, established on European data and no longer only American data, should weigh in arbitrations concerning the allocation of cohesion funds. Financing social mobility in lagging territories is also financing their innovation potential in the next generation.
Bibliographic Information - Title: Intergenerational Mobility Fosters Innovation in Europe - Authors: McNamara, Neidhöfer, Lehnert - Publisher: Nature - Publication date: July 8, 2026
Sources
- Bell, Chetty, Jaravel, Petkova, Van Reenen et al., “Intergenerational Mobility Fosters Innovation in Europe,” Nature, July 2026. https://www.nature.com/articles/s41586-026-10736-9
- Bell, Chetty, Jaravel, Petkova, Van Reenen, “Who Becomes an Inventor in America? The Importance of Exposure to Innovation,” Quarterly Journal of Economics, 2019.
- Carl Benedikt Frey, How Progress Ends: Technology, Innovation, and the Fate of Nations, Princeton University Press. https://press.princeton.edu/books/hardcover/9780691233079/how-progress-ends
- Swedish Patent Office, longitudinal data on the decentralization of qualified employment and patent production, 2025.
