The effectiveness and prediction analysis of China’s agricultural low-carbon production policy from the perspective of policy tools

Government intervention through policy measures to curb carbon emissions from agricultural production is an essential pathway toward achieving sustainable agricultural development in China. However, does the intervention effects of Agricultural Low-carbon Production Policy vary depending on the policy tool types employed? This study compiles 884 policy texts and carbon emission datasets at both national and regional levels in China (1993–2022). Using the Latent Dirichlet Allocation to classify policy tools and Support Vector Machine Regression to predict the effectiveness of policy combinations. We find that (1) Agricultural low-carbon production policies exhibit a lag effect; (2) Coercive policy tools are the most prevalent, accounting for 40.72%, while incentive policy tools are the least common, making up only 16.06%; (3) Empirical results demonstrate that incentive policy tools yield the most effective intervention outcomes from agricultural low-carbon production, followed by coercive, directive, and voluntary policy tools. (4) Predictive results show that a high-growth model combined with incentive policy exerts the most significant suppression effect on agricultural carbon emissions. The findings of this study offer the following insights: Building upon existing research on policy tools, future policy formulation should be guided and informed by strengthening effectiveness, highlighting significance, and enhancing feasibility across three key dimensions.