Mountains block and seas move: The impact of geographical environment on the China's Carbon Emissions Trading Scheme in reducing urban PM2.5 concentrations

Abstract

The dispersion of PM2.5-related pollutants in the atmosphere is intricately connected to the natural geographical environment. Nevertheless, previous evaluations of environmental policies regarding urban PM2.5 concentrations have rarely taken into account the influence of the natural geographical environment. Utilizing panel data and DID models for 266 prefecture-level cities in China from 2006 to 2019, this study thoroughly examines the impact of geographical environment on the China's Carbon Emissions Trading Scheme (ETS) in reducing urban PM2.5 concentrations. The results suggest that (1) Neglecting the impact of natural geographical environment, although the ETS effectively reduces the PM2.5 concentration in the pilot cities from an overall perspective, there are anomalies in the robustness test and the heterogeneity analysis, which are mainly reflected in the decrease in the significance of the PSM-DID results, and the positive haze-enhancement effect of the ETSs of Fujian, Beijing, Shanghai, etc. (2) Considering the influence of geographic environment, combined with the spatial correlation between China's topographic distribution and PM2.5 concentration, four control groups of Basin, Coastal, etc. are screened for group regression, and the results show that geographic environment factors significantly affect the emission reduction effect of each ETS on PM2.5 concentration. (3) The regression results of the three types of spatial econometric models all indicate that there is a significant spatial spillover effect of ETS to reduce PM2.5 concentration. Considering the influence of geographic environment on the spatial spillover effect and the fact that China's provincial division is mostly based on geographic environment factors, in order to more reasonably assess the effect of the policy on the PM2.5 concentration, the researcher should select sample cities with no spatial correlation of PM2.5 and clustering robust standard errors to the provincial level.