How asymmetric is the response of CO2 emissions to economic restructuring in China? Evidence from NARDL approach

China has experienced a remarkable economic progression since the implementation of its reform and opening -up policy. Yet, it is crucial to confront the challenge of achieving sustainable and high-quality economic growth while concurrently enhancing environmental quality. This study adopts the nonlinear autoregressive distributed lag (NARDL) approach to explore how economic structure adjustments impacted China's carbon dioxide (CO2) emissions from 1980 to 2020. The empirical results reveal some asymmetries in the influence of agricultural progression, industrial development, and foreign direct investment (FDI) on CO2 emissions in short and long-term contexts. Specifically, in the long run, a 10% positive change in agricultural development increases CO2 emissions by 1.55%, nearly double the mitigation effect of a negative change (0.78%); a 10% positive (negative) change in industrialization is associated with a 0.6% decrease (0.23% increase) in CO2 emissions; a 10% positive change in FDI results in a 2.30% increase in CO2 emissions, supporting the "Pollution Haven" hypothesis, while a 10% negative change contributes to a 4.51% increase. In the short run, a 10% negative change in agricultural and industrial sectors increases CO2 emissions by 0.27% and 0.21%, respectively, while a 10% negative change in FDI leads to a 2.0% increase in CO2 emissions. Additionally, short-term positive changes in agricultural and industrial sectors and positive shocks in FDI exert negligible influence on CO2 emissions. These findings highlight the necessity of economic structure transformation and refinement of current environmental policies.