Can Horizontal Ecological Compensation Mechanism Promote Regional Inclusive Green Growth? Evidence from Panel Data of 81 Cities in 8 Provinces of the Yellow River Basin

This study examines the empirical association between horizontal eco-compensation and regional inclusive green growth. Using panel data for 81 cities in eight provinces of the Yellow River Basin from 2012 to 2022, it applies a staggered difference-in-differences model, mediation regressions, and a spatial econometric model. Double machine learning with five-fold cross-fitting is used as a robustness exercise to mitigate multicollinearity and functional-form misspecification. Conditional on city and year fixed effects and the included covariates, policy implementation is positively associated with inclusive green growth in pilot cities. The mediation regressions are consistent with three possible channels—total factor productivity, science and technology expenditure, and industrial upgrading—but do not establish definitive causal mediation. The spatial estimates indicate a significant negative association with inclusive green growth in neighboring cities. Within-group estimates are significant for upstream and downstream cities but not for midstream cities, and the midstream estimate is significantly weaker than the downstream estimate. Estimates are also significant within the medium- and high-income groups, although full-sample interaction tests do not establish differences across economic development tiers. Because policy placement is not fully random and time-varying unobservables cannot be completely ruled out, the findings should be interpreted as conditional on the maintained identification assumptions. Policy recommendations, therefore, emphasize basin-specific targeting, stronger implementation capacity in weaker areas, cross-regional coordination, and expansion rules that weigh fiscal capacity, opportunity and transition costs, and marginal benefits.