Regional Growth, Innovation and Latent Nonlinear Effects

This paper studies the link between knowledge, innovation and growth in European regions using nonparametric methods. Our findings suggest that knowledge inputs and the share of innovative firms have a heterogeneous and nonlinear relationship with growth. This evidence has been exploited to examine the consequences of alternative policies using a counterfactual estimation setup, the results of which imply that increasing the formal knowledge base may be optimal in most regions. Less knowledge and innovation intensive regions will also benefit from a higher innovation potential and from a trustworthy and entrepreneurial economic environment. As a byproduct of our analysis, we show that the gradients obtained from a local-linear kernel regression can be used to identify clusters of spatial association