A Regression Tree-Based Analysis of the European Regional Competitiveness

Regional competitiveness is considered a key factor of development. In this work, with the aim of analysing the main drivers of the competitiveness, a Regression Tree analysis has been performed for the Eurostat Regional Competitiveness Index (RCI) as response variable by taking the 74 basic indicators used in the 2019 RCI edition as explanatory variables. Being a non-parametric method, suitable for the analysis of large data sets via a recursive partitioning procedure, the Regression Tree allowed to identify (a) the 12 most influential indicators, out of the initial 74, for the overall 2019 RCI, and (b) a classification of the 268 European regions into 15 homogeneous groups. Interestingly, the groups are ranked by their predicted RCI values which correspond to the mean observed RCI values within the groups themselves. The almost perfect correlation between the Eurostat RCI and the predicted RCI within groups confirms the key role of the 12 selected indicators as determinants of the 2019 RCI. These evidences could help policy makers to address their strategies towards focused objectives in line with the specific needs of the territories, characterized by an intrinsic heterogeneity and complexity.