The determinants and dynamics of regional convergence in the EU

In this study, we employ the pairwise stochastic convergence approach to identify the pairs of NUTS2 regions for all 28 EU Member States that exhibit co-movement in their growth dynamics, over the period 1980-2018. We then use the observed stochastic convergence trajectories to assess the role of first nature geography, which is defined by variations in physical geography, locations and proximities, and second nature geography, corresponding to the economic interactions between partners, in causing economic growth convergence patterns. We find that western and northern parts of Europe have higher pairwise stochastic convergence (and lower intra-country convergence) rates than regions in East and Southeast Europe. Focusing on the converging NUTS2 regions, we find strong evidence that first and second nature geography drive cluster-like convergence dynamics. Regions with common locational characteristics (metropolitan, coastal, islands, and mountainous) tend to converge to each other, while they do not converge with dissimilar regions. Regardless of national borders, contiguity and accessibility are significant drivers of convergence. Congruence in sectoral specialisation results in divergence that could be driven by competing economic interests within the common market. The opposite holds for dissimilarities in specialisation, which could be explained by complementarity in the production process. Overall, we find strong evidence for stochastic club convergence at the top of the EU. In contrast, bottom regions with low market dynamism and poor economic development, do not converge to each other, and collectively lag significantly behind top European regions. Finally, we find evidence of EU Cohesion Fund payments facilitating the observed convergence dynamics across the EU, which highlights the importance of targeted regional policy interventions in reducing persistent structural regional disparities within the EU.