This paper employs a spatial Durbin model for analyzing the impact of human capital on regional productivity using for 198 NUTS-2 European regions for the sample period from 1995 to 2004. The study provides evidence for the existence of spatial externalities and interactions of the sort as emphasized by new growth theory. To interpret results meaningfully, we calculate summary measures that account for the simultaneous feedback nature of the underlying model. By sampling from the parameter distribution we present measures of dispersion, revealing that it is relative regional advantages in human capital that matter most for productivity growth. (authors' abstract)