Human Capital Reallocation and Agglomeration of Innovation: Evidence from Technological Breakthroughs

Abstract

This paper identifies the reallocation of human capital as a key channel of agglomeration spillovers for innovative firms. To measure agglomeration spillovers, I study how R&D labs in different local labor markets respond differently to scientific breakthroughs, which create large and unexpected shocks to innovation productivity in certain technology categories. Taking advantage of U.S. Census longitudinal establishment data matched with patent records, I systematically locate R&D labs in all local labor markets for each firm. I document four main findings. First, following scientific breakthroughs, affected labs in thicker local labor markets (i.e., commuting zones with more inventors innovating in a certain field) produce more patents and higher-quality patents, consistent with positive agglomeration spillovers. Second, the increase in patenting is mostly attributed to new hires rather than incumbent inventors. Third, the thick labor market effect is concentrated in states and industries where there is lower enforceability of non-compete agreements and labor is more mobile. Finally, using textual analysis to identify lab-level exposure to scientific breakthroughs, I find that inventors are reallocated to labs that are more favorably affected by shocks, which helps labs in thicker labor markets to more easily bring in inventors working in the same niche fields and having a diverse knowledge base. Taken together, these results point to labor mobility as a key force in explaining why innovative firms cluster, and suggest that the clustering of firms in thick labor markets can foster corporate innovation by facilitating productivity-enhancing reallocation of human capital following scientific breakthroughs.