Air Pollution and Innovation
(w/ Felix Bracht)
Journal of Environment Economics and Management, 2025

If air pollution harms innovation -- and therefore future productivity -- existing assessments of its economic cost are incomplete. We estimate the effect of fine particulate matter concentration on inventive output in 977 European regions. Exploiting thermal inversions and weather-induced ventilation of pollutants for identification, we find that a decrease in air pollution equivalent to the average yearly drop in Europe leads to 1.2% more patented inventions in a given region. A back-of-the-envelope calculation suggests that accounting for the effect on innovation increases the economic cost of air pollution as assessed in prior work by about three quarters.

Firm Markups and the Economic Value of Innovation.
(w/  Ralf Martin, Jenniffer Solorzano Mosquera, and Catherine Thomas)
International Journal of Industrial Organization, 2025

We examine the relationship between firms' markups and the economic value of their innovation, including both the private value captured by the innovating firm and the knowledge spillovers that benefit other firms. Using a sample of over 14,500 EU firms and 2,400 US firms granted patents between 2005 and 2014, we find that innovation by high-markup firms is more valuable privately and also creates more external value. These associations are robust to controlling for the stock of past innovation and to estimating innovation value in various ways.

Radical, Disruptive, Discontinuous, and Breakthrough Innovation: more of the same?
(w/ Adrian Kovacs, Cristina Marullo, Alberto Di Minin and Bart Van Looy)
Industrial and Corporate Change, 2025

This study revisits concerns about the lack of conceptual clarity in the literature on exceptional innovations – commonly labelled as ‘radical’, ‘disruptive’, ‘breakthrough’, or ‘discontinuous’. A bibliometric analysis of articles using at least one of these labels shows a dense network of five highly overlapping thematic clusters that largely build on similar scientific foundations. The labels are scattered across different clusters, suggesting they do not coincide with distinctive topics. Definitions, if given, rely on two underlying dimensions: (ex ante) novelty and/or (ex post) impact. None of the labels are consistently defined throughout the literature, and similar definitions appear for different labels. This conceptual ambiguity creates difficulties for scholars and practitioners exploring the literature. It also obscures the theoretical link between novelty and impact. To address this issue, we advance a typology of innovation trajectories that allows us to relate novelty to impact. We argue that the type of novelty introduced by a trajectory – that is, radical vs. architectural – has important implications for the type of firm-level impact it may exert. Thus, our framework has the potential to inspire future work on the origins and effects of exceptional innovation.

Measuring Technological Novelty with Patent-based Indicators
(w/ Reinhilde Veugelers and Jurriën Bakker)
Research Policy, 2016

This study provides a new, more comprehensive measurement of technological novelty. Integrating insights from the existing economics and management literature, we characterize inventions ex ante along two dimensions of technological novelty: Novelty in Recombination and Novelty in Knowledge Origins. For the latter dimension we distinguish between Novel Technological and Novel Scientific Origins. For each dimension we propose an operationalization using patent classification and citation information. Results indicate that the proposed measures for the different dimensions of technological novelty are correlated, but each conveys different information. We perform a series of analyses to assess the validity of the proposed measures and compare them with other indicators used in the literature. Moreover, an analysis of the technological impact of inventions identified as novel shows that technological novelty increases the variance of technological impact and the likelihood of being among the positive outliers with respect to impact. This holds particularly for those inventions that combine Novelty in Recombination with Novelty in Technological and Scientific Origins. Overall, the results support our indicator as ex ante measure of technological novelty with the potential to drive radical technological change. 

Patent citation indicators: One size fits all?
(w/ Jurriën Bakker, Lin Zhang, Bart Van Looy)
Scientometrics, 2016

The number of citations that a patent receives is considered an important indicator of the quality and impact of the patent. However, a variety of methods and data sources can be used to calculate this measure. This paper evaluates similarities between citation indicators that differ in terms of (a) the patent office where the focal patent application is filed; (b) whether citations from offices other than that of the application office are considered; and (c) whether the presence of patent families is taken into account. We analyze the correlations between these different indicators and the overlap between patents identified as highly cited by the various measures. Our findings reveal that the citation indicators obtained differ substantially. Favoring one way of calculating a citation indicator over another has non-trivial consequences and, hence, should be given explicit consideration. Correcting for patent families, especially when using a broader definition (INPADOC), provides the most uniform results. 

Knowledge Spillovers from Clean Innovation: A Tradeoff between Growth and Climate?
(w/ Ralf Martin)
R&R Nature Climate Change

Innovation policy faces a tradeoff between growth and climate objectives when the knowledge spillover externality from clean innovation is low compared to other sectors. To make such a comparison, we use patent data to estimate field-specific spillover returns generated by R&D support. Supporting Clean presents itself as a win-win opportunity, yielding global returns one-eighth higher than those of an untargeted policy. Nevertheless, only a modest portion of the returns stays within country borders, raising the question of whether national interests distort efficient allocation. Our policy simulations underscore the benefits of supranational coordination in clean innovation policy, potentially boosting returns by approximately 25% for the EU and over 60% globally. Moreover, the EU benefits strongly from US Clean innovation spillovers, impacting the debate on the Inflation Reduction Act. Overall, we identify no explicit innovation policy tradeoff in tackling the twin challenges of economic growth and climate change but emphasize the necessity for international cooperation.

Do Broader Teams Lead to More Innovation?
Rej&R Management Science

Organizations often combine high-skilled workers of different expertise in broad teams to feed the knowledge production process with diverse inputs. This paper asks whether interventions that change team breadth lead to more innovation. Using quasi-random variation in inventor team breadth induced by the death of a team member, I find no evidence of such an effect. This result is at odds with prior findings of a positive relationship between team breadth and invention quality. I show that this positive relationship is not robust to accounting for sorting -- positive selection in particular -- of inventors into broader teams. My results temper optimism about interventions by managers or policymakers that promote team breadth beyond current levels.

Efficient Industrial Policy for Innovation: Standing on the Shoulders of Hidden Giants.
(w/ Charlotte Guillard, Ralf Martin, Pierre Mohnen and Catherine Thomas)
Under Review

We quantify the efficiency gains from targeting R&D subsidies to fields with the highest knowledge spillover externalities. By combining new measures of spillovers and private returns with field-specific structural estimates of R&D costs, we develop a framework to estimate the spillover return rate of a marginal subsidy in any given field. Our results show substantial welfare gains from targeted innovation policies: within-country returns from optimal targeting are at least 40% higher than those from uniform subsidies. Moreover, coordinating field-specific subsidies across OECD countries could increase returns by an additional 30%.

Intellectual Mobility Frictions
(w/ Jordan Bisset)
Working paper

Despite repeated findings that intellectually mobile inventors - those with experience in multiple technological fields - have a comparative advantage in the production of the highest value inventions, aggregate evidence suggests that the share of intellectually mobile inventors is falling. In this paper, we propose the presence of allocation frictions which increase the cost a firm must pay to allocate inventors to a technological field in which they have not previously invented. We create a simple framework to estimate these frictions at the micro-level and find them to be very high in magnitude, and heterogeneous across firms and the technological space. Counterfactual estimates suggest that even moderate decreases in these allocation frictions could increase idea generation by 50%.  

The Private Value of Clean Energy Innovation
(w/ Ralf Martin)
Working paper

We examine the distribution of the private value of clean and dirty innovation using new methods based on patent data. We document that the value of clean innovations is higher and more dispersed. We find an overall decline in the variability of private values and returns in the wake of the Great Recession. This is consistent with the idea that financial restrictions have made investors and innovators more risk averse. Because clean and dirty innovations show different exposures to such risk aversion, the recession could have contributed to the decline of clean relative to dirty innovation. We develop a method to quantify counterfactual clean and dirty innovation that would have prevailed if the distribution of private values (or returns) would have stayed fixed. The results suggest that shying away from risky R&D in the wake of the Great Recession has considerably depressed the relative share of clean innovation, but is not responsible for the clean drop in absolute terms. More broadly, our results suggest that financial constraints in the wake of crises may be an important barrier on the path to the clean equilibrium. 

Configurations of Coordination Mechanisms to Sustain Value Appropriation from Innovation
(w/ Giulia Solinas)
AOM Proceedings, 2019

The issue of value appropriation from innovation is a centerpiece of firms’ innovation strategies. Appropriability through successful patent protection occurs due to a hierarchical structure in the R&D decision-making, a certain level of cross-functionality between the organizational units involved, and the codification of information in the application process. However, the interaction between these coordination mechanisms remains open to debate. This study explores this interplay by adopting a configurational perspective and analyzing 20 cases using the fs/QCA approach. Our findings explore the synergies derived from the intersection of coordination mechanisms for appropriability and identify which are core versus peripheral. The results suggest that centralizing both decision-making and cross-functionality are core to appropriability when combined in an organization. However, both mechanisms become ancillary when in a configuration with formalization through planning.

Spillovers from Science to Technology
(w/ Ralf Martin, Arjun Shah and Anna Valero)

Measuring the Private Value of Innovation
(w/ Alfonso Gambardella)

Hidden Figures? Peer Effects and Female Scientists' Productivity
(w/ Thomas Schaper)