Structural Inertia in Regional Industrial Concentration: Persistence and Mobility across European Regions, 2017–2023

Regional industrial structures are shaped by path dependence and resist rapid change, yet fine-grained evidence on their dynamics through major shocks remains scarce. This paper analyses the persistence and mobility of industrial concentration across 1,320 NUTS3 regions in the EU-27 and UK over 2017–2023, a window bracketing the COVID-19 pandemic. Using a purpose-built firm-level panel of approximately 2.04 million firms drawn from Orbis, aggregated to NACE4 × NUTS3 resolution, the analysis combines Herfindahl-Hirschman concentration indices, Theil inequality decomposition, and Markov transition matrices to characterise structural change at a granularity not previously available. Three findings emerge. First, industrial concentration is highly persistent: approximately half of all regions retained their relative concentration quintile across the period, and rank ordering was strongly preserved even through the pandemic shock, confirming the core prediction of evolutionary economic geography. Second, change operated through distributional compression rather than rank disruption, with mean reversion concentrated at the extremes of the distribution while the relative ordering of regions held. Third, structural mobility is systematically patterned by regional typology: capital metropolitan regions diversified faster than their national context, while island regions moved toward greater concentration. The island result reveals a structural concentration trap with direct implications for Smart Specialisation policy, since the regions least able to diversify are precisely those for which place-based innovation strategy is documented to be weakest. The paper introduces a coverage filter benchmarked against Eurostat Labour Force Survey data, separating genuine structural change from database artefacts.