Measuring the Private and Social Returns to R&D: Unintended Spillovers versus Technology Markets

Abstract

The canonical approach to measuring private and social returns to R&D assumes diffusion through spillovers. We develop a general framework that allows R&D to diffuse through both spillovers and voluntary technology transfers. To operationalize the framework, we create a data set of interactions in the market for technology between publicly held US firms. We use changes in tax incentives for R&D to identify causal effects. Taking market channels into account decreases the observed difference between social and private returns to R&D as compared with the canonical approach. The analysis suggests significant gains from trade in markets for technology.We thank Chad Syverson and four anonymous referees for helpful and constructive comments. We are grateful to

Patent remedies and technology licensing: Evidence from a supreme court decision

“…We then complemented this dataset by compiling information on license agreements. The license agreements of US firms were extracted from ktMINE, which provides a comprehensive database of license agreements in the United States (Arque‐Castells & Spulber, 2022; Fosfuri, Helmers, & Roux, 2012) gathered from the Securities and Exchange Commission, the United States Patent and Trademark Office reassignment database, and federal records. We also hand collected information on European firms' license agreements via a company name search and a manual match in the Factiva database.…”

“…One way of generating revenue from innovation is licensing internally developed ideas in technology markets (Arora, Fosfuri, & Gambardella, 2001b;Arora, Fosfuri, & Ronde, 2013;Gans & Stern, 2003). A recent study estimated that the revenue gains from technology transactions for the period from 1990 to 2014 were larger than $1 trillion per year, corresponding to over 10% of the total revenue generated by US public firms (Arque-Castells & Spulber, 2022). Technology commercialization via licensing is argued to work best when the appropriability regime is tight (Teece, 1986(Teece, , 2006.…”

Firm Matching in the Market for Technology: Business Stealing and Business Creation*

“…Our analysis suggests that conventional spillover estimates are potentially contaminated with internalized technology transfers because the technological proximity metric used to capture externalities is correlated with matching in the market for technology. In Arqué-Castells and Spulber [2022], we provide a framework for estimating the private and social returns to R&D that allows the benefits of R&D to diffuse both through spillovers and technology markets. Our other study also draws on the newly constructed dataset on interactions in the market for technology but conducts a considerably different analysis that addresses different issues.…”