Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This paper studies the spatial agglomeration dynamics of a sample of 224 European regions in the period 1991–2012 by a dynamic extension of the Moran scatterplot, consisting in a nonparametric estimate of the joint dynamics of GDP per worker and its spatial lag. We find evidence that the twin-peaked distribution of GDP per worker in 2012 is the result of the existence of three spatial clubs: the first mainly populated by regions belonging to the former Eastern Bloc countries, the second by regions of the Southern Mediterranean countries (Portugal, Italy, and Spain) and the third by regions of other EU countries (notably Germany, France, UK and Northern European countries). Spatial externalities appears to be mainly driven by geographical, technological and social proximities, and only marginally by institutional proximities. Local spatial dependence is a pervasive phenomenon, and only the low-GDP per worker regions with similar neighbours seem to be immune. The 30-year ahead forecast of GDP per worker distribution suggests a weak convergence among the three clubs and a persistent spatial agglomeration in the future.