Job Market Paper:

Creative Construction: Knowledge Sharing in Production Networks

Abstract: Knowledge spillovers between firms are often measured using patent citations. I show that even the most cited patents granted before 2000 received around 50% of citations from one firm only, and this concentration has increased up to 77% by 2014. I provide additional evidence consistent with the following hypothesis: instead of spillovers citations reflect intentional sharing of trade secrets between business partners. I develop and test a theory of knowledge sharing between firms in a production network that partially explains the observed citation patterns. The theory highlights that knowledge flows depend on the market structure and contractual arrangements between firms.

Coverage: The Visible Hand, Episode 30

Working Papers:

Global Sourcing and Multinational Activity: A Unified Approach (joint with Pol Antràs, Teresa Fort, and Felix Tintelnot), NBER Working Paper 30450

Abstract: Multinational firms (MNEs) accounted for 42 percent of US manufacturing employment, 87 percent of US imports, and 84 of US exports in 2007. Despite their disproportionate share of global trade, MNEs’ input sourcing and final-good production decisions are often studied separately. Using newly merged data on firms’ trade and FDI activity by country, we show that US MNEs are more likely to import not only from the countries in which they have affiliates, but also from other countries within their affiliates’ region. We rationalize these patterns in a unified framework in which firms jointly determine the countries in which to produce final goods, and the countries from which to source inputs. The model generates a new source of scale economies that arises because a firm incurs a country-specific fixed cost that allows all its assembly plants to source inputs from that country. This shared fixed cost across plants creates interdependencies between firms’ assembly and sourcing locations, and leads to non-monotonic responses in third markets to bilateral trade cost changes.

Coverage: Becker Friedman Insitute

Work in Progress:

1. Knowledge Sharing and Organizational Structure: Evidence from Trade Secret Litigation (joint with Alice Wu)

Abstract: We apply deep learning language models on trade secret litigation records to classify legal cases based on whether the (claimed) misappropriation of trade secrets happened through employees or business partners (e.g., input suppliers) and to compare technologies involved in trade secret litigation with inventions disclosed in patent files. We use this information to study how firms change their patenting, employment, and input sourcing decisions following misappropriation of trade secrets.

2. Concentrated Citations and Private Value of Patents

Abstract: The number of patent citations is often used to measure patent quaity. I compute the distribution of citations across firms and show that conditional on the number of citations more concentrated citations are associated with higher growth in sales/profits of patent owners and greater stock market responses on the grant days of patents. I argue that instead of social value concentrated citations are more likely to reflect private value of a patent to the patent assignee.