EN
This paper focuses on the testing of income convergence of the EU regions using both non-spatial and spatial approaches. The main motivation for this analysis was the fact that the classical income convergence models suffer from a misspecification due to omitted spatial dependence among regions. Our empirical results provide support for the absolute beta-convergence modelling from spatial econometric perspective in our sample of 252 NUTS 2 regions over the period 2000 – 2011. Another serious finding is that the assumption of a single steady-state for all regions often mismatches with the reality. We found the club spatial beta-convergence models to be more appropriate for analysed data.