Job Market Paper:

Creative Construction: Knowledge Sharing in Production Networks

Abstract: Knowledge flows between firms are often measured using patent citations. I show that even the most cited patents on average receive the majority of citations from one firm only, and this concentration has significantly increased since 2000. Using the movement of inventors across companies, I show that the concentration is primarily driven by firms rather than inventors. I develop a theory of knowledge sharing between firms that accounts for these citation patterns. Citations are correlated with the sharing of trade secrets that are complementary to patented technologies. They are concentrated because only a limited set of firms gets access to private knowledge of a patent owner. Firms have incentives to share their secrets with producers of complementary products such as suppliers and customers but to conceal them from competitors. In turn, competitors can obtain private knowledge from each other through their common suppliers and customers if the latter did not sign confidentiality agreements. The model predicts contractual arrangements and patterns of knowledge sharing (citations) in a production network based on the degree of industry competition and firms' bargaining positions vis-a-vis their suppliers/customers. Using the network data for the U.S. publicly traded firms and the variation across industries in the exposure to import competition from China, I provide empirical evidence supporting the predictions of the theory.

Research in progress:

Regulation of Confidentiality Agreements: Growth and Welfare Implications (joint with Alice Wu)

Abstract: There has been an increase in the use of non-disclosure agreements (NDAs) by U.S. companies. We use data on litigation of trade secrets to estimate how breaches of confidentiality agreements affect firms’ growth, investments in R&D, and patent citations. Next, we develop an endogenous growth model in which firms decide whether to share their private knowledge with business partners and employees and use NDAs to prevent knowledge leakages. We use the model and the empirical results on litigation to evaluate growth and welfare implications of restrictions on the use of NDAs.

Global Sourcing and Multinational Activity: A Unified Approach (joint with Pol Antràs, Teresa Fort, and Felix Tintelnot)

Abstract: We link US Census data on firms’ domestic establishments to BEA surveys on foreign direct investment (FDI) to enhance our understanding of the interaction of the domestic and global operations of US based manufacturing firms. We first document that multinational firms operating in the US account for the vast majority of exporting and global sourcing in the US: more than eighty percent of US manufacturer’s exports and imports are associated with sales and purchases of these US-based multinationals. Furthermore, these same firms follow remarkably global assembly strategies: sales of foreign affiliates of multinational firms headquartered in the US are close in magnitude (74 percent) to the worldwide sales (domestic sales and exports) of their US establishments, and these affiliates’ sales are more than seven times larger than US exports by these same US establishments. Leveraging our merged dataset, we document a series of novel facts regarding the global assembly and global sourcing strategies of U.S. firms, and unveil systematic interdependencies in these firm decisions. We then develop a multi-country model in which firms decide on the location of their assembly plants (i.e., their assembly strategy) as well as the source of the inputs used in their plants worldwide (i.e., their global sourcing strategy), and use the model to interpret our novel facts.