The Network Structure and Driving Mechanisms of Construction Waste Management Efficiency in the European Union

This study measures construction waste management efficiency (CWME) in 27 EU member states (2016–2025) using a three-stage super-efficiency SBM-DEA model, constructs a spatial correlation network via a modified gravity model, and employs social network analysis and random forest to investigate network structure and driving mechanisms. Results reveal persistent cross-country heterogeneity and a widening East–West efficiency divide. The CWME network exhibits small-world characteristics, with CONCOR analysis partitioning member states into four blocks: Northern and Baltic states serve as the dominant Net Spillover bloc, whereas Southern periphery countries remain Net Beneficial recipients whose membership contracted from five to three by 2025. Random forest identifies government environmental protection expenditure, unit labour cost, and population density as the top three driving factors, with unit labour cost and government environmental protection expenditure exhibiting the strongest negative marginal effects on CWME, while urbanization rate shows a near-neutral relationship. These findings provide empirical support for policy interventions targeting the structural economic determinants of CWME. By integrating efficiency measurement, spatial network analysis, and machine learning-based driver identification, this study offers a comprehensive framework for diagnosing regional disparities in construction waste governance and informing cross-regional collaborative policy design.