| The production and living of the people is intimately linked to electric energy,making it a vital energy source for national economic growth.Electricity consumption can be a powerful indicator of economic operation and progress.With the development of industrialization and new urbanization in China,China has already faced great power demand.As China’s economic growth shifts to a new norm,the power demand’s expansion and composition will also experience drastic alterations.At present,China is dominated by coal-fired power generation,but the pollutants produced by thermal power generation will cause great pressure on the environment.Therefore,China’s energy transformation is imperative.Electricity,as a clean energy source for consumers,is a critical element of China’s energy strategic transformation.At the 20 th National Congress,it was emphatically declared that China must further advance the energy revolution,bolster the clean and effective utilization of coal,and hasten the formation and planning of a fresh energy system.To foster the green growth of China’s power sector,it is essential to assess the potential alterations of China’s power structure,particularly the effects of regulations.In the "Twelfth Five-Year Plan",China intends to gradually set up a carbon emissions trading market,with the aim of achieving energy conservation and emission reduction at the least expense through the use of a market economy.This paper proposes specific research assumptions regarding the influence of power policy and carbon emission rights price on power demand,based on a comprehensive examination of related literature research.After that,all the text data of power news,power forum topics and power big data forum sections of China Energy News Network from August 2013 to December 2022 were crawled,and the news text analysis method of Baker team was referred to and improved,and the uncertainty index of China’s power policy was calculated by word frequency.This paper examines the effect of electricity policy and carbon emission rights price on China’s electricity demand by constructing four mixed-frequency MIDAS models.Additionally,it constructs three types of mixing data models-MIDAS,MF-VAR,and mixing dynamic factor-as well as an ARIMA model to forecast power demand.First,after reading and sorting out relevant literature research,this paper puts forward specific research assumptions on the impact of power policy and carbon emission rights price on power demand.After that,all the text data of power news,power forum topics and power big data forum sections of China Energy News Network from August 2013 to December 2022 were crawled,and the news text analysis method of Baker team was referred to and improved,and the uncertainty index of China’s power policy was calculated by word frequency.After that,this paper constructs four mixed-frequency MIDAS models to explore the impact of electricity policy and carbon emission rights price on China’s electricity demand.Finally,this paper constructs three kinds of mixing data models,MIDAS,MF-VAR and mixing dynamic factor,and ARIMA model to forecast the power demand.The empirical results show that:(1)The index constructed in this paper can sensitively capture the fluctuations in China’s power sector,such as the release of the 13 th and 14 th Five-Year Plan and the power rationing in various places.(2)The electricity consumption of the whole society in China has remarkable memory,and the memory lasts for 4 periods.In addition,there is a positive correlation between electricity policy and electricity demand fluctuation,and a negative correlation between carbon price and electricity demand fluctuation,confirming two assumptions of this paper.(3)The prediction results of the mixed data models are better than those of the same frequency model,and the short-term prediction ability of the MIDAS model can be effectively improved by introducing the error correction term.MIDAS model is better at short-term prediction,In the short term,the prediction ability of MIDAS model with error correction term is improved compared with ordinary MIDAS model.With the increase of prediction periods,the prediction accuracy of ECM-MIDAS model will gradually decline,while ordinary MIDAS model will have better prediction effect than other models,while MF-VAR model has the highest prediction accuracy in six periods and is better at medium-term prediction.It is worth noting that although the prediction effect of mixing dynamic factor model is not outstanding,this model can predict the number of high-frequency periods and the corresponding weekly data for monthly data,which is superior to other models in timeliness and timeliness. |
PDF Full Text Request