The Geography of Knowledge Spillovers Between High‐Technology Firms in Europe: Evidence from a Spatial Interaction Modeling Perspective

Abstract

The focus in this article is on knowledge spillovers between high‐technology firms in Europe, as captured by patent citations. The European coverage is given by patent applications at the European Patent Office that are assigned to high‐technology firms located in the EU‐25 member states (except Cyprus and Malta), the two accession countries Bulgaria and Romania, and Norway and Switzerland. By following the paper trail left by citations between these high‐technology patents we adopt a Poisson spatial interaction modeling perspective to identify and measure spatial separation effects to interregional knowledge spillovers. In doing so we control for technological proximity between the regions, as geographical distance could be just proxying for technological proximity. The study produces prima facie evidence that geography matters. First, geographical distance has a significant impact on knowledge spillovers, and this effect is substantial. Second, national border effects are important and dominate geographical distance effects. Knowledge flows within European countries more easily than across. Not only geography, but also technological proximity matters. Interregional knowledge flows are industry specific and occur most often between regions located close to each other in technological space.

Knowledge Spillovers Across Europe: Evidence from a Poisson Spatial Interaction Model with Spatial Effects

“…Geographic distance is measured in terms of the great circle distance [in km] between the economic centres of the regions i and j, country border effects in terms of the existence of country borders between regions i and j, language barriers in terms of the existence of different languages in the regions i and j, and technological distance in terms of the technological proximity index developed by Maurseth and Verspagen (2002). The Maximum Likelihood estimate of dispersion is 0.725 and highly significant (see Fischer, Scherngell and Jansenberger 2006) which points to the presence of overdispersion that is modelled by the Bayesian Poisson spatial interaction model with spatial effects.…”

“…It is this model that distinguishes the current study from prior work by Fischer, Scherngell and Jansenberger (2006) which produces more conventional maximum likelihood estimates based on a heterogeneous Poisson spatial interaction model. Their model specification arises from introducing multiplicative heterogeneity in the mean of the Poisson model as a proxy for fixed effects parameters.…”

Barriers to cross-region research and development collaborations in Europe: evidence from the fifth European Framework Programme

“…Two regions that patent exactly in the same proportion in each subclass have a proximity index equal to one, while two regions patenting only in different subclasses have an index equal to zero. This proximity index is appealing because it allows for a continuous measure of technological distance by simple transformation (see Fischer et al 2006). Table 1 about here Table 1 reports the parameter estimates, the associated p-values and standard errors for the three model versions A, B and C. The parameters are estimated by the NBPML estimator described in Section 2.…”

“…Note that Belgium has French nological barriers between regions measured in terms of technological distance. We follow Fischer et al (2006) to use regional patent data from the European Patent Office and construct a 630-by-1 technological vector for each region that contains its share of patenting in each of the 630 technological subclasses at the third level of the International Patent Classification System. Technological proximity between two regions is measured in terms of the uncentred correlation between their technological vectors.…”

“…An alternative, however, that we follow in this study is to specify the variance in a more accurate way. The negative binomial spatial interaction model provides an obvious model specification (see, for example, Fischer et al 2006) to handle the extra variance. This probability distribution can be written as…”