| With the development of economy,environmental problems are becoming more and more serious.The emission trading policy is the first large-scale market-based environmental protection policy in my country,and its emission reduction effect has always been valued by the government.Existing studies have used the Differences-in-Differences method to analyze the implementation effect of the emission rights trading policy,but the policy conclusions drawn by the DID method have always been controversial.The traditional Differences-in-Differences method needs to satisfy the strict linear parallel trend assumption,which is difficult to hold in practice,which leads to biases in the estimated effect;And existing studies have assumed that the treatment effect for different groups is the same,but due to differences in the characteristics of different groups,the effect is expected to be different;DID can only obtain the average treatment effect of policy implementation,but cannot obtain the distribution treatment effect,that is to say,it cannot tell which regions the policy is effective for.Industrial sulfur dioxide as one of the main air pollutants,its harm has been affecting people.Then how about the effect of emission trading policy on industrial sulfur dioxide emission reduction? Is the implementation of policies more effective in dirtier areas? In order to promote the implementation of the system in our country more effectively,it is necessary to carry out a comprehensive inspection and analysis.Based on the panel data of 30 provinces and autonomous regions（excluding Tibet,Hong Kong,Macao and Taiwan）from 2001 to 2019,the random forest algorithm is used to select control variables from 18 control variables in a data-driven way,and finally selects five control variables add into the model of this paper.The effect of industrial sulfur dioxide emission reduction by emission trading policy in China is analyzed empirically by constructing Nonlinear Difference-In-Differences model.At the same time,the treatment effects of treatment group and control group were given respectively,and the differences between them were observed.Further more,the effect of policy implementation is analyzed from distribution effect.The robustness test was carried out by replacing dependent variables,replacing the time of policy intervention,and eliminating the influence of policies in the same period.Finally,the Nonlinear Difference-In-Differences method is used to test and analyze the regional heterogeneity.The results show that :（1）Emission trading policy has a significant effect on the reduction of industrial sulfur dioxide in China.（2）The treatment effect of the treatment group and the control group was not exactly the same,and the treatment effect of the control group was higher than that of the treatment group.（3）Emission trading policies are not more effective for areas with higher industrial sulfur dioxide emission intensity.（4）The emission trading policy shows obvious regional heterogeneity,and the implementation effect is more obvious in western China,but not in eastern and central China.Finally,some suggestions are put forward according to the research results:（1）Improve the market trading system in the pilot areas.（2）To strengthen the total amount control and quota allocation of industrial sulfur dioxide emissions in areas with a high degree.（3）The regional heterogeneity of industrial sulfur dioxide emissions caused by emission trading policies should be fully emphasized. |
PDF Full Text Request