|
on Innovation |
| By: | Tom Davidson; Basil Halperin; Thomas Houlden; Anton Korinek |
| Abstract: | AI labs are increasingly using AI itself to accelerate AI research, creating a feedback loop that could lead to an intelligence explosion. We develop a general semi-endogenous growth model with an innovation network, where research and automation in one sector increase the productivity of research in other sectors, and derive a clean analytical condition under which growth becomes superexponential (``explosive''). We find that automating research can offset diminishing returns to ideas by activating two reinforcing channels: a technological feedback loop across research sectors, and an economic feedback loop in which higher output finances further research. Growth becomes explosive if the combined strength of technological and economic feedback loops overcomes diminishing returns. In a simple simulation calibrated to trends in AI progress, fully automating software research and modest (5%) automation in other sectors generates a singularity within six years. Bottlenecks do not overturn the result if task automation advances sufficiently fast. |
| JEL: | O31 O33 O40 O41 |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35155 |
| By: | Yanrui Wu (Department of Economics, Business School, University of Western Australia) |
| Abstract: | This paper presents a review of the nexus between digital transformation and innovation using evidence from China. It explores the general literature in this field as well as China-specific studies. The trend of digital transformation and innovation is assessed by using Chinese and global data. It also provides new evidence about the nexus between digital transformation and innovation using Chinese regional data. |
| Keywords: | digital transformation, innovation, measurement, nexus, Chinese regions, global perspectives |
| JEL: | O31 O33 R11 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:uwa:wpaper:26-02 |
| By: | Bertschek, Irene; Erdsiek, Daniel; Niebel, Thomas; Sack, Robin; Zimmermann, Volker |
| Abstract: | Recent literature has increasingly focused on deciphering the modern productivity puzzle, with particular attention given to the link between digital technologies and firm-level productivity. So far, much of this research has primarily focused on large and publicly listed firms. Leveraging a panel dataset covering German small- and medium-sized enterprises (SMEs) over the period 2016 to 2021, we investigate whether digitalisation can help revive the sluggish productivity growth and narrow the gap between productivity frontrunners and laggards. We measure digitalisation through firms' digital capital stocks (DK) that we derive from a broad measure of digitalisation expenditures. Building on an augmented Cobb-Douglas production function, we examine the relationship between DK and labour productivity (LP ). Our findings show that higher DK is positively associated with higher LP levels, with the effect being even stronger for firms that are already more digitally advanced. Moreover, higher digitalisation expenditures appear to be related to narrowing the productivity gap between laggards and the frontier. |
| Keywords: | Heterogeneity of Digitalisation, Productivity, Firm-level Data |
| JEL: | L25 O14 O33 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:zbw:zewdip:340839 |
| By: | Sonja Dobkowitz |
| Abstract: | I study optimal implementation of climate targets in a model with distortionary fiscal policy, learning-by-doing, and directed technical change. The key mechanism is that fiscal constraints link innovation policy to labor allocation, creating a tension between directing research and directing learning-by-doing. Analytically, I show that learning-by-doing shapes the effectiveness of carbon taxation in directing research through an expertise effect: carbon taxes are more effective at steering innovation toward green technologies when green expertise is relatively high. Quantitatively, I calibrate the model to the U.S. economy to characterize the optimal policy mix consistent with climate targets. I find that carbon should be taxed heavily, persistently exceeding the social cost of carbon. While higher carbon prices raise green expertise, they induce an excessively rapid reallocation of researchers from fossil to green technologies, generating persistent innovation misallocation. A welfare analysis shows that learning-by-doing substantially amplifies the cost of distortionary taxation, in particular during the transition to net-zero emissions. |
| Keywords: | Second-best climate policy, directed technical change, learning-by-doing, Ramsey taxation, misallocation of innovation, emissions target implementation |
| JEL: | H21 H23 O38 Q54 Q55 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:diw:diwwpp:dp2161 |
| By: | Luisa Alamá-Sabater (Department of Economics and IIDL, Universitat Jaume I, Castellón, Spain); Joan Crespo (INTECO and Department of Economic Structure, Universidad de Valencia, Spain); Miguel Ángel Márquez (Department of Economics, Universidad de Extremadura, Spain); Emili Tortosa-Ausina (IVIE, Valencia and IIDL and Department of Economics, Universitat Jaume I, Castellón, Spain) |
| Abstract: | This article examines the interaction between innovation and employment and population dynamics through the development of a system of simultaneous equations. The model is applied to a panel dataset of 271 European NUTS-2 regions. The results reveal strong bidirectional feedbacks between innovation and employment, while population dynamics operate indirectly through employment rather than exerting a direct effect on innovation. Innovation is found to follow jobs rather than people, indicating that the concentration of economic activity and labor interactions, not demographic size per se, constitute the primary drivers of regional innovative capacity. These mutually reinforcing dynamics give rise to virtuous and vicious cycles that contribute to persistent regional disparities. By opening the black box of employment–population–innovation interactions, the paper provides a structural foundation for designing more effective population, innovation, and employment policies. In particular, the analysis demonstrates that policies targeting a single dimension, whether business climate, quality of life, or innovation support, are unlikely to succeed in isolation. |
| Keywords: | innovation, population-employment dynamics, European Union, NUTS2, spatial effects, territorial development |
| JEL: | C3 O18 O21 R1 R23 R3 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:jau:wpaper:2026/08 |
| By: | Erik CANTON (European Commission) |
| Abstract: | This paper examines how employment protection legislation shapes firm-level technology adoption across OECD countries using novel survey data. We document a negative association between employment protection legislation and the adoption of artificial intelligence and other restructuring-intensive technologies, while more modular digital technologies display weaker relationships with labor market institutions. The patterns are particularly pronounced among large incumbent firms, while younger and fast-growing firms exhibit higher adoption rates. To interpret these patterns, we develop a general equilibrium model with heterogeneous firms in which workforce restructuring costs raise the productivity threshold for technology adoption. The model predicts heterogeneous firm responses: some adopters expand employment, others contract, and highly productive incumbents may optimally refrain from adoption when restructuring costs rise with scale. The framework is extended to incorporate endogenous technology arrival and diffusion through both adoption by incumbent firms and entry of new firms implementing frontier technologies. The analysis highlights how labor market institutions can affect technology diffusion and, through this channel, influence incentives to develop and commercialize new technologies. |
| Keywords: | Employment protection legislation, technology adoption, artificial intelligence, labor market institutions, firm heterogeneity, restructuring costs, general equilibrium |
| JEL: | O33 J24 J38 L25 O32 |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:eug:wpaper:ki-01-26-053-en-n |
| By: | Caroline Veldhuizen (Centre for Technology, Innovation and Culture (TIK), University of Oslo, Norway) |
| Abstract: | This paper examines how, during a 25-year timeline from 2000 to 2025, the European Union (EU) has used climate, industrial and environmental policy to shape the emergence of an integrated carbon management system (CMS). It assesses the extent to which the evolving framework exhibits characteristics consistent with transformative innovation policy (TIP). The CMS is conceptualised as an emergent, policy-driven ‘interconnected system complex’ centred on carbon capture and storage (CCS), carbon capture and utilisation (CCU) and carbon dioxide removal (CDR), linked to multiple socio-technical systems. Methodologically, the study undertakes qualitative analysis of 114 binding and non-binding EU policy documents, over the timeline. An adapted version of Kanger, Ghosh and Entsalo’s (2025) Intervention Points Framework (IPF), grounded in the multi-level perspective, multi-system interaction literature, policy mix research and TIP debates is used as the analytical lens. The IPF is refined to address formation of an emergent, socio-technical configuration, rather than transition in more stable, established meso-level systems. The paper makes a number of important empirical and conceptual contributions. Empirically, the study reveals both the depth and shallowness of policy leverage across different intervention points and establishes that the transformative potential of the emergent CMS is real but ambivalent and contested. Conceptually, the study illustrates how policy may be used as a vehicle for observing and describing evolving systemic structures and flows and their directionality, and the iterative processes of formation and change which define them. The study also enables insights about the political nature of the ‘landscape’ level of the MLP, and the importance of the policymaking paradigm, for determining the potential of policy to drive transformational change. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:tik:inowpp:20260504 |
| By: | Ryan R. Hill; Carolyn Stein |
| Abstract: | We study how a frontier AI model affects scientific discovery by examining the release of the AlphaFold2 algorithm and its impact on structural biology and related fields of science. Structural biology is the field of science concerned with understanding the structure and function of proteins. Researchers in this field historically devoted substantial time and resources to experimentally solving three-dimensional protein structures. AlphaFold can predict these structures without running experiments. In July 2021, researchers gained access to hundreds of thousands of these AI-predicted structures virtually overnight. Yet, to date, we find that the rate of experimental structure determination has remained almost unchanged. Instead, researchers appear to use predicted structures to facilitate and complement experimental structure determination. Looking at downstream science that builds on protein structures, we find that basic research on proteins that had no structure information prior to AlphaFold increases by 15 to 40% relative to proteins that already had a structure, shifting the direction of research toward less-studied proteins. However, we find no evidence so far that more applied, early-stage drug development is targeting these proteins, though such activity may emerge in the future. |
| JEL: | I23 J24 L65 O31 O33 |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35143 |