Green Industrial Park Policy and Urban Resilience in China: Evidence from Double Machine Learning

Abstract

As industrial green transformation advances and urban risks become more complex, this study explores the relationship between China’s national green industrial park policy and urban resilience. The empirical analysis draws on panel data for 293 Chinese cities at the prefecture level or above between 2011 and 2024 and uses a Double/Debiased Machine Learning (DML) framework to flexibly account for observed city characteristics and possible nonlinear relationships. A positive association is observed between the green industrial park policy and the composite urban resilience measure, with this pattern remaining broadly consistent under a range of robustness and sensitivity analyses. Positive coefficients are also obtained across all four resilience components; however, the empirical support is comparatively stronger for the ecological and infrastructure components than for their economic and social counterparts. The pathway results are compatible with potential roles for environmental governance investment, residents’ income, and new-quality productive forces. Differences in the estimated relationships are more pronounced when cities are grouped by resource dependence and industrial-base characteristics, whereas the east–central–west classification does not yield statistically significant between-group variation. By shifting attention from park-specific environmental performance to broader resilience outcomes at the city level, this study extends the scope of green industrial park policy assessment. Given the assumptions underlying the empirical design, the reported coefficients represent conditional relationships rather than conclusive causal evidence.