An "extended" knowledge production function approach to the genesis of innovation in the European regions

The paper looks at the genesis of innovation in the EU regions in order to shed light on the link between innovative inputs (R&D and Human Capital) and the genesis of economically valuable knowledge. The "traditional" regional Knowledge Production Function (KPF) is innovatively developed in three complementary directions. First, the KPF is "augmented" in order to control for all possible "unobservable" and "immeasurable" time varying factors that influence the genesis of innovation (i.e. localised institutional and relational factors, regional innovation policies). Second, a semi-parametric approach that relaxes any arbitrary assumption on the "shape" of the KPF is adopted. Finally, the assumption of homogeneity in the impact of R&D and Human Capital is relaxed by explicitly accounting for the differences between "core" and "peripheral" regions. The econometric results confirm the importance of accounting for time varying unobserved heterogeneity through the adoption of a "random growth" specification: R&D efforts exert a significant influence on innovation only after controlling for regional specific time varying unobserved factors. In addition the semi-parametric approach uncovers significant threshold effects for both R&D expenditure and Human Capital and highlights a strong complementary between these two factors. However, "core" regions benefit from a persistent advantage in terms of the "productivity" of their innovation inputs. This has important implications for the EU innovation policies at the regional level.