<![CDATA[Environmental regulation is often assumed to work the same way regardless of how it is designed — yet this study shows that command-and-control and market-incentive instruments pull green productivity in opposite directions, and that artificial intelligence (AI) does not treat them equally either. Drawing on panel data from 281 Chinese prefecture-level cities over 2012–2024, this study finds that command-and-control regulation, measured through text analysis of local government work reports, significantly suppresses green total factor productivity (GTFP) by raising compliance costs, while market-incentive regulation, measured as the ratio of pollutant discharge fee revenue to GDP, significantly promotes it by channeling price signals into innovation. GTFP is estimated using the SBM-ML index, which incorporates labor, capital, and energy inputs together with industrial pollution as an undesirable output. Interaction-term estimates further show that local AI development attenuates the negative effect of command-and-control regulation while amplifying the positive effect of market-incentive regulation — AI acts as a “buffer” for coercive regulation and an “amplifier” for market-based regulation, rather than a uniform productivity booster. The findings suggest that AI cannot simply substitute for well-designed environmental policy, but it can make both good and bad policy design matter less — or more — than they otherwise would.]]>
Copyrights © 2024
