This study aims to examine whether digitalization promoted regional convergence across the European Union between 2015 and 2023 or reinforced a persistent Digital Matthew Effect. This study focuses on how regional Digital Intangible Capital, combining research capacity, specialized human capital, digital adoption and institutional maturity, relates to economic performance and whether digital advantages spread beyond established innovation hubs.
The study analyzes 233 EU NUTS-2 regions in 2015, 2019 and 2023. It integrates New Economic Geography with a technological condensation framework, constructs an entropy-weighted Digital Intangible Capital Index and estimates time-stratified Spatial Durbin Models. Robustness is assessed using alternative index weights and k-nearest-neighbor matrices ranging from k = 2 to k = 15.
The results indicate a persistent Digital Matthew Effect and a stable core-periphery structure. Although aggregate dispersion in the index declines, this largely reflects convergence in basic digital access, while innovation-intensive capabilities remain concentrated. Direct effects of Digital Intangible Capital are positive and significant, whereas indirect spatial effects remain statistically insignificant across years and specifications.
The analysis relies on three cross-sectional snapshots and therefore identifies conditional spatial associations rather than causal effects. Missing regional observations required hierarchical imputation, which may reduce variation in some indicators. Future research should use panel and micro-level data to identify knowledge-diffusion mechanisms more precisely.
EU Cohesion Policy should complement infrastructure investment with support for specialized skills, R&D capacity, institutional quality and durable cross-regional innovation networks.
Persistent concentration of digital innovation may deepen regional inequality, brain drain and unequal access to high-value employment and wealth creation.
The study combines spatial econometrics, multidimensional intangible-capital measurement and a technological condensation framework to distinguish local returns from network-mediated effects and demonstrate that the identified pattern is robust to alternative weighting and spatial specifications.
