Historical data of crude oil, coal and natural gas in China are used to calculate the carbon emission and carbon tax in recent years. It is proved that the nonlinear model, BP Neural Network, is more suitable for the calculation of carbon tax under the influence of various energy factors. Then BP Neural Network and GM (1, 1) have a good learning ability to forecast the carbon tax revenue from 2020 to 2030. BP Neural Network, however, has less error and shows that Per capita disposable income, GDP and Per capita GDP play a decisive role in the development of carbon tax and environmental protection. Based on the assumption that the carbon tax will be imposed and increased, the carbon emission intensity of each province and the carbon tax acceptance of urban residents will be within the ideal range in the next decade. All evaluation indicators of models and data analysis show that the carbon tax has a good policy guidance for energy saving and emission reduction. This study about carbon tax provides a new solution for domestic environmental protection.
Read full abstract