Innovation and Knowledge Transfer: the Role of Inventors’ Mobility and Collaboration Networks

This article studies the contribution of different channels of knowledge transfer on innovation output, measured by patent production. We deal with informal channels of knowledge transmission that arise from working relationships established by inventors, and study their impact on firm economic performance in terms of innovation production. The study’s contribution to the literature is twofold. First, differently from most of the existing literature on the topic, we consider those relationships not formally settled in co-inventorships. These relationships, although informal, can be responsible for the creation of networks of knowledge sharing and transfer (Almeida and Kogut, 1999; Power and Lundmark, 2004). Specifically, we consider those interactions that are born through inventor multi-collaborations and mobility. Second, rather than studying the relationship between social interactions and the diffusion of knowledge, we analyse their effect on company innovation production. The analysis is carried out on the Italian region of Veneto. Resuming the debate on the localization of knowledge transfer and the existence of productivity lags, the paper focuses on the role played by geographical proximity, and on the dynamic effects of knowledge flows. The geographical issue is faced at a very detailed level; we measure knowledge spillovers observed within the same Local Labor System (LLS), between different LLSs of the region and, also, from extra-regional LLSs. The work is based upon the original OECD REGPAT database on patent applications filed at the European Patent Office. The dataset has been cleaned by using a manual procedure that allows to correctly identify inventors and applicants, to avoid false mobility and overestimation of knowledge connections. The estimates, exploiting the panel dimension of the dataset, show that inventor-firm networks and inventor mobility have highly localized effects on productivity. However, while the productivity effects induced by inventor mobility are almost simultaneous in time, the effect of inventor-firm networks is lagged in time.