The production of scientific and technical knowledge is mostly concentrated in specific locations (high-tech clusters, innovative industry agglomerations, centres of excellence, and technologically advanced regions). Knowledge flows very easily within regions; however, scientific and technical knowledge also flow between regions. The aim of this paper was to analyse how knowledge flows between regions, and the effect of these flows on the innovative performance, measured by patent applications. We estimate a regional knowledge production function, and, using appropriate spatial econometric estimation techniques, we test the effect of both geographical and relational autocorrelation (measured by participation in EU funded research networks as part of Fifth Framework Programme). We model unobservable structure and link value of knowledge flows in these joint research networks. We find that knowledge Previous versions of this paper were presented at the 7th European Meeting on Applied Evolutionary Economics, Pisa, February 2011; Dimetic Conference on 'Regional Innovation and Growth: Theory, empirics and policy analysis', Pècs, March 2011; Eurolio Conference on the Geography of Innovation, Saint Etienne, January 2012. We thank participants for useful comments and suggestions. In depth, discussions with E. Bergman, P. Elhrost, S. Beretta, B. Dettori, K. Frenken, G. Fagiolo, E. Marrocu, S. Usai and the recommendation from three referees for a thorough revision of the paper have substantially improved its structure and the arguments. The usual caveats apply. flows within inter-regional research networks, along non-symmetrical and hierarchical structures in which the knowledge produced by network participants tends to be exploited by the network coordinator.
JEL Classification