Innovation2026-05-04

We link U.S. patent and inventor records to individual voter register files and map politically polarized policy issues to related technologies. Compared to Republicans, Democrats are one-third more likely to patent technologies addressing climate-change mitigation or women's reproductive health, and one-third less likely to patent weapons and related technologies. These gaps are not explained by differences in inventive ability or by sorting across organizations or teams. Party-technology alignment has strengthened over the past two decades, a period of rising political polarization in U.S. society. Technology diffusion is also politically polarized: Democrats are more likely than Republicans to cite aligned technologies and less likely to cite misaligned ones. Together, these findings are consistent with political polarization and societal views being important drivers of the direction and diffusion of technological change and operating, at least in part, through inventors' technology choices, with implications for innovation policy.Gaia DossiMarta MorandoDiffusion, Innovation, Partisanship, Polarization, Technology2026-03This paper offers a conceptual review of the relationship between the knowledge economy and innovation, challenging simplistic, linear assumptions about how new ideas are generated. Given their increasing global significance, the study focuses on Fourth Industrial Revolution (4IR) technologies. Specifically it examines the nature, evolution, and defining features of the knowledge economy. Such an economy relies on increasing specialization, research, innovation, and continuous learning, with learning and experience being its most critical sources. Furthermore, the paper argues that innovation constitutes a fundamental dimension of the knowledge economy, noting that knowledge production is strictly related to innovations. Rather than a sequential chain, innovation is presented here as a complex, systemic process characterized by multiple feedbacks and loops. This analysis highlights the systemic, non-deterministic nature of innovation and its strong relationship with knowledge. However, the capacity of companies to innovate depends heavily on the innovation ecosystem—the framework where stakeholders interact and collaborate—as well as the regulatory and legislative framework. Consequently, several factors, including the availability of sufficient human capital with appropriate education and advanced skills, the presence of robust infrastructure, and the role of institutions, are necessary to make companies' innovations effective. While this paper does not claim to provide definitive answers, it seeks to offer new insights for future research.Schilirò, Danieleknowledge economy; knowledge; learning; networks; innovation; technological progress; competitiveness2025Why would a market leader choose not to patent an innovation? We study Samsung's decision to forgo patent protection for dual SIM technology in the Indian mobile handset market. Using a structural model of demand and supply estimated on quarterly product-level data from the Indian mobile handset industry, we document that rival firms' dual SIM products generated a preference discovery externality. Rival firms' widespread adoption of the dual SIM technology allowed consumers to discover the value of the technology, also benefiting Samsung itself. Counterfactual simulations show that a patent would have suppressed this externality, reducing Samsung's equilibrium profits despite holding monopoly rights. Voluntary non-patenting was therefore privately optimal. Our findings shed light on wider debates about open-sourcing in software and other markets.Debi Prasad MohapatraVatsala Shreetiinnovation, patenting, telecom, preference discovery2026-04We use a stylized model of a dynamic innovation tournament to show that the effectiveness of monetary incentives depends on whether contestants receive cardinal, ordinal, or no information about their rival’s performance. The model’s main implication is that performance information acts as a substitute for prize money in creating incentives to invest in new ideas: The investment-maximizing information policy switches from no to ordinal to cardinal information as the tournament’s prize is reduced. A laboratory experiment provides support for our theory but also unveils an unpredicted pattern of behavior capable of overturning the model’s conclusions concerning optimal policy.Martina Bossard, Marc Möller, Catherine RouxInnovation Tournaments; Performance Information; Rank Information; R&D Investment.2026-03Governments already spend large sums to promote innovation through grants, tax credits, loans, equity instruments, incubators, prizes, and advisory programs. Yet public procurement is vastly larger than conventional innovation-policy budgets. In OECD economies, procurement is roughly 13 percent of GDP, while direct support and tax relief for business R&D together are only a fraction of one percent of GDP. This asymmetry matters. Even a very small innovation-oriented tilt in procurement can represent a material increase in the effective scale of innovation policy. Yet procurement systems are rarely used this way. Most public procurement organizations are designed to secure timely delivery, preserve integrity, ensure equal treatment of suppliers, and obtain value for money. They are not designed to explore technological uncertainty, nurture early markets, or orchestrate experimentation with new solutions. Procurement officers are typically judged on compliance, continuity of service, and avoidance of visible failure. Under those incentives, the safe equilibrium is predictable: detailed specifications, strong threshold requirements, large established suppliers, price-dominant competitions, and risk transfer to vendors wherever possible. This report argues that governments do not need to choose between lawful procurement and innovation policy. They can make procurement more innovation-friendly without abandoning core procurement principles. The relevant question is not whether procurement law should be suspended in the name of innovation. The relevant question is how familiar and lawful procurement tools can be reframed so that public buyers learn about technological possibilities, reduce uncertainty, validate solutions, and scale what works. This is the purpose of The Cube.Ricardo HausmannYariv GabayInnovation, Industrial Policy2026-04The advances of artificial intelligence (AI) are built on the groundwork laid by researchers. We study the labor market competition between academia and industry for AI researchers and its consequences for public knowledge production. Using data on 150, 000 computer science researchers, we document a major reallocation of AI talent toward top technology firms between 2005 and 2020. Publications at AI conferences predict transitions to top firms more strongly than to academia. Exploiting acceptance decisions at a leading AI conference, we compare accepted authors with similar rejected authors and find that a publication increases the probability of moving to a top firm by 2-6 percentage points in the next 1-3 years. Sorting to top firms is stronger for male researchers, whereas female students and postdocs are more likely to get tenure-track positions following a publication. Researchers who move to top firms subsequently publish fewer papers, resulting in approximately 1, 000 fewer AI papers and 2, 000 fewer papers in other computer science areas per year in the public domain.Francesca MiserocchiSavannah NorayAlice WuSorting, Productivity Signals, Labor Market Concentration, Innovation2026-04We study the global footprint and real effects of Chinese overseas corporate ownership. By assembling a comprehensive micro-level dataset of 161, 773 firms across 159 countries (2012–2021), we independently reconstruct multi-layered ownership chains to trace capital through offshore tax havens to its ultimate origin. This approach reveals a global footprint substantially broader than official FDI statistics. Chinese-controlled foreign assets expanded at 20% annually, reaching $2.1 trillion or roughly 3% of global corporate assets by 2021. Chinese investors—particularly state-owned enterprises (SOEs)—strategically target R&D-intensive and supply-chain-linked firms. Following acquisition, target firms increase capital stock and R&D expenditures, yet these inputs fail to generate higher patent output and are accompanied by a significant decline in profitability. We document a novel 'innovation spillback' mechanism: while target innovation remains stagnant, Chinese parent firms experience a sharp acceleration in granted patents following their first developed-economy acquisition. Furthermore, a greater Chinese presence crowds out R&D at non-target peer firms, though aggregate industry-level innovation remains unchanged. China thus represents a distinctively state-driven model of global ownership that accepts weaker near-term performance to internalize technological capacity at home.Jennie BaiLuc LaevenYaojun KeHong Ru2026-04Research assessment frameworks play a central role in shaping the priorities, direction, and culture of scientific research. Yet there are growing concerns over their misalignment with evolving policy priorities, public expectations, and new demands from science. Over-reliance on narrow performance measures has generated perverse incentives and undesirable behaviours. At the same time, these measures tend to undervalue critical research practices and outputs, such as collaboration, openness, societal engagement, and support to policy making. This paper provides a system-level overview of research assessment, highlighting key tensions, mapping the main actors and drivers, and identifying a set of common reform principles through a comparative review of the literature. It makes the case for developing new assessment frameworks that are better aligned with the evolving expectations and demands placed on science.OECDOpen science, Research assessment, Research evaluation frameworks, Science policy2026-04-29In this paper, we use data for more than 5, 000 Chilean companies to investigate whether participation in business association increases the probability of R&D investment. Dealing with the endogeneity of participation through a bivariate Probit model with an exclusion variable that captures the trust environment among firms, we find that this probability increases by about 27%. This effect is heterogeneous across firms. Participation increases the probability of R&D investment by 30.8% for SMEs and by 43.9% for those companies with severe financial constraints. Our evidence is consistent with the idea that associativity may help SMEs to close the innovation gap and/or to alleviate financial problems.Felipe AguilarRoberto Alvarez2025-12In this study, we present the first systematic evidence of the impact of cartel recidivism on innovation. Combining data from an international price-fixing cartel database with the structural characteristics of the US manufacturing sectors at the six-digit NAICS level, we analyze how cartel recidivists influence subsequent innovation outcomes. Using a staggered difference-in-differences (DiD) framework for 110 US cartel cases over the period 1979-2016 and a novel heterogeneous estimator, we find that cartel recidivists lead to a significant and sustained decline in innovation progress. We argue that cartel recidivists, rather than single offenders, drive the negative impact of collusion on innovation. The results of this study are vigorous to several robustness tests, justifying the absence of pretreatment effects and endogeneity.Fotis, PanagiotisPolemis, MichaelCartel; Recidivism; Innovation; Antitrust; Difference in Differences2026-02-20Technological disruptions often generates political conflict. Artificial intelligence (AI) is widely expected to transform labor markets and economic systems, yet it has not become a strongly polarizing political issue in advanced democracies. This paper investigates why, by fielding a preregistered survey experiment with 11, 418 respondents in the United States, Germany and Italy. We examine factual knowledge on AI and automation, beliefs over its economic effects, demand for policy intervention and signatures of online petitions on Change.org. We document limited knowledge, widespread pessimism on their labor-market impact, substantial demand for government intervention and considerable potential for political mobilization, pointing to an unmet demand for policy responses. We then test the mobilization power of competing political narratives on the economic effects of AI and automation. Overall, across countries and institutional contexts, politicizing AI shifts policy preferences in the expected directions but reduces engagement in political mobilization. In addition, it decreases support for the extreme petitions, thereby reducing polarization. These findings suggest that emerging technologies characterized by high uncertainty and large distributive effects may not follow the historical pattern of polarization associated with past economic shocks. Our results rationalize politicians' hesitation towards increasing the salience of AI and automation.Giacomo BattistonFederico BoffaEugenio LeviAlberto ParmigianiSteven StillmanArtificial Intelligence; Automation; Political Polarization2026-03We examine the long-term labor market impact of the steam engine, an early general-purpose technology, by linking newly digitized 19th-century records from Prussia to modern German labor market data (1975-2019). Regions with a higher concentration of steam engines per worker in 1875 exhibit higher wages today, primarily because of higher firm productivity and a more skilled workforce. These regions also exhibited greater skill diversity in 1939 and generated more innovations between 1877 and 1918, a pattern that persists to this day. Our findings highlight a lasting, self-reinforcing cycle between technology and skills, set in motion by the steam engine, offering a novel explanation for regional income disparities and their persistence.Sascha O. BeckerChristian DustmannHyejin Kusteam engine, technology adoption, diversity, innovation, human capital, productivity2026-01Using two waves of nationally representative Danish firm surveys linked to employer-employee administrative registers, we study how adoption varies across artificial intelligence (AI) and related advanced technologies. We show that AI adoption is highly technology-specific. While firm size and digital infrastructure predict adoption broadly, workforce composition operates through distinct channels: STEM-educated workforces predict core AI adoption, whereas non-STEM university-educated workforces are associated with generative AI adoption, indicating different human capital complementarities. The factors associated with adoption differ from those predicting deployment breadth: firm size and digital maturity matter for both, whereas workforce composition primarily predicts adoption alone. Machine learning and natural language processing are deployed across multiple business functions, whereas other advanced technologies remain concentrated in specific operational domains. Individual-level evidence provides a foundation for these patterns, with awareness of workplace AI usage concentrated among managers and high-skilled workers. Self-reported AI knowledge is higher among younger and more educated individuals. Finally, commonly used occupational AI exposure measures vary substantially in their ability to predict observed adoption, with benchmark-based measures outperforming patent-based and LLM-focused alternatives. These findings show that treating AI as a monolithic category obscures economically meaningful variation in who adopts, what they deploy, and how well existing measures capture it.Giuseppe PulitoMariola PytlikovaSarah SchroederMagnus LodefalkArtificial Intelligence; Technology Adoption; Digitalisation; Human capital; AI Exposure Measures2026-03