A core-periphery comeback? Clustering European Union Social Progress Index (EU-SPI)

Regional cohesion in the European Union (EU) remains a major issue, shaping EU policies, EU budget allocation and regional welfare. We dig into the EU-Social Progress Index published in 2020 to explore its territorial/regional clustering patterns. Are there EU internal frontiers –clusters-? Do they coincide with the member States? Do they have other influences? What determines belonging to different clusters? Is there any resemblance between the GDP per capita and the SPI regional distribution? We divide the EU NUTS-2 map into clusters using hierarchical clustering. We look for the optimal number of clusters, and we compare the outcome with state borders, paying attention to discrepancies or to state combinations. Our main finding is that the optimal clustering is two, and that they are very robustly defined. Of course, this leads towards reassessing core-periphery approaches and the impact of the great recession and the sovereign debt crisis. Another major finding is that the EU-SPI clustering reveals major discrepancies with the per capita GDP clustering. Some NUTS-2 regions perform in SPI terms much better than expected by per capita GDP. On the contrary, some NUTS-2 regions perform in SPI much worse than expected by per capita GDP. The discrepancies suggest major public policy successes and failures.