Finally, Stefano and me managed to get a working paper version out! For details, look on the research page or directly at SSRN.
Abstract We study the diffusion of knowledge from scientists to firms within scientific communities. We build on a unique dataset on conference proceedings as “paper trail” of almost all relevant conference series in computer science since 1996. More than 5000 firms appear as conference sponsors or as affiliations in proceedings. Their participation is concentrated in the highly ranked conferences and their scientific contributions are on average highly cited. We exploit direct flights as an instrumental variable for the participation choice of scientists between a conference where a firm participates and other similar conferences. The participation in the same conference has a positive causal effect on knowledge diffusion to the firm’s scientific and inventive activities. Additional analyses suggest that interactions and collaborations with scientists that remain external to the firm are likely a key mechanism of this diffusion. The effects are remarkably stronger the larger the firm’s investments in participation.