How Does Regional S&T Planning Affect Urban Green Innovation Diffusion in China? A Spatial Econometric Analysis

This study examines how Regional Science and Technology Planning (RSTP) relates to urban green innovation diffusion (UGID) using a panel of 284 Chinese cities from 2008 to 2020. UGID is measured by registered green patent transfers aggregated by the transferor’s city. We combine a spatial difference-in-differences framework with a BERT–DNN model that classifies policy sentences into obligatory goals (OG), anticipatory goals (AG), and other categories. In the baseline estimates, RSTP is positively associated with local transfer activity. The coefficient on innovation-network policy exposure is positive and statistically significant at the 10% level, whereas the coefficient on geographical policy exposure is not statistically significant. OG shows a positive own-city coefficient and a negative geographical exposure coefficient. The contemporaneous AG coefficients are not statistically significant, which does not establish the absence of longer-term effects. Regressions show that OG is positively associated with technology-service personnel density, fiscal science and technology expenditure intensity and relative local research and development (R&D) personnel advantage, supporting innovation-resource concentration as a plausible channel. The findings highlight the importance of distinguishing local transfer activity from geographical and innovation-network relationships when evaluating regional planning. They also suggest that target design should give greater attention to cross-regional cooperation and the longer-term evaluation of anticipatory goals.