Trade, Innovation and Optimal Patent Protection (with David Hémous, Simon Lepot, Ralph Ossa, and Thomas Sampson) ― Revise & Resubmit: The American Economic Review  

Intellectual property rights are a recurrent source of tensions between developed and developing economies. This paper provides the first quantitative analysis of optimal patent policy in trading economies. We develop a new model of trade, growth and patenting in which patent protection affects both innovation and market power. The model is estimated using data on patent applications to calibrate patent protection by country and the geography of innovation. Counterfactual analysis yields three main results. First, the potential gains from international cooperation over patent policies are large. However, achieving these gains requires more innovative economies to offer stronger protection. Second, only a small share of these gains has been realized so far. And third, by pushing towards policy harmonization, the TRIPS agreement hurts developing countries without generating global welfare gains. Overall, there is substantial scope for policy reforms to increase efficiency.

Market Power when Ideas get Harder to Find: A Theory of Directed Innovation

I provide a theory of directed innovation that connects the decline in research productivity (Bloom et al. 2020) to rising market power (De Loecker, Eeckhout, and Unger 2020) and various other macroeconomic trends, explaining these outcomes as consequences of lower research productivity. In a Schumpeterian model, innovators choose a search intensity that determines the sample size of potential markets they can target. I establish that a higher search intensity implies more directed innovation, which in turn leads to selection of firms based on market power. I calibrate the model to match the TFP growth rate, average markup, and firm entry rates of the US economy prior to 1980. The firm selection process has macroeconomic implications, and the model predicts rising markups when ideas are getting harder to find. As the stakes in the innovation process rise, the incentives to conduct market search increase; this leads to stronger selection on market power and an upward shift of the whole markup distribution. Moreover, the model performs well on untargeted moments: The predictions for factor shares, average firm size, productivity growth, the real interest rate, and the R&D-to-GDP ratio all align closely with the data.

Ricardo Meets Pareto: Dispersion and the Gains from Trade

Dispersion is pivotal in Ricardian trade: Dispersion governs Ricardian comparative advantage, which in turn shapes trade patterns. It is uncontested that a general increase in dispersion results in greater gains from trade. What are the sources of these gains? Do countries primarily benefit from their own dispersion, or are the benefits of dispersion shared globally? To address such fundamental questions, I provide the first tractable Ricardian theory that accommodates country-specific dispersion parameters. In an environment à la Eaton and Kortum (2002), I show that Pareto-distributed cost draws allow for country-specific scale and dispersion parameters. The model can readily be characterized and calibrated to the data. I consider counterfactual exercises where I increase dispersion parameters both unilaterally and uniformly, and my quantitative results suggest that trading partners would contribute to between one-third and one-half of the potential welfare gains from additional dispersion. Unlike in gravity models, trade flows do not need to obey the independence of irrelevant alternatives (IIA) property, and I indeed find asymmetrical effects of a small China shock on other countries. Moreover, increased dispersion implies higher gains from trade than what the canonical formula by Arkolakis et al. (2012) can account for. Finally, my model can explain trade zeros with Ricardian forces.

Creative Obstruction: Industry Dynamics with Dominant Firms (with Elie Gerschel and Paolo Mengano) (draft available upon request)

We investigate the long-term effects of dominant firms on industry dynamics and aggregate productivity growth. To achieve this goal, we construct a Schumpeterian firm dynamics model that incorporates heterogeneity across two dimensions: (i) the ability of firms to defend their technologies against improvements by competitors, and (ii) their innovation efficiency. We characterize the steady state and show that dominant firms gain significant market shares and become, on average, significantly larger than the rest of the firms. In contrast, the effect on aggregate growth crucially depends on the type of dominant firms: While firms with superior innovation efficiency are conducive to growth, defensive dominant firms hamper growth, which calls for policy intervention. We simulate firm-level trajectories and find systematic differences between defensive and innovative firms. This demonstrates the potential of our framework to deduce firm types from firm-level data.