We analyse the spatial relationships of economic output dynamics in European regions from 2000 to 2019 using dynamic spatial autoregressive growth models. In contrast to previous studies that rely on exogenous spatial weight matrices based solely on geographical proximity, we use a novel Bayesian approach to fully estimate the spatial weight matrix. Our results show that economic and sectoral characteristics (e.g., sectoral production structure, education, etc.) significantly influence the degree of regional interdependence. The approach thus allows to study the complex dynamics of regional economic development beyond mere distance.