基于时变高阶矩的碳市场风险预测研究  

Forecasting VaR and ES of Carbon Market Based on the Time-Varying Higher-Moments Model

作　　者：杜坤海 黄迅 Du Kunhai;Huang Xun(Economics School of Xihua University;Business School of Chengdu University)

机构地区：[1]西华大学经济学院,610039 [2]成都大学商学院,610106

出　　处：《环境经济研究》2022年第2期52-65,共14页Journal of Environmental Economics

基　　金：四川省金融学会2022年度重点课题研究项目“能源市场对碳市场的金融风险传染研究”(SCJR2022078);成都哲学社会科学规划资助项目"成都系统性金融风险的大数据智能预警研究"(2021BS028);四川矿产资源研究中心资助项目“后疫情时代四川省战略性矿产资源安全评价体系研究”(SCKC-ZY2021-YB005)的研究成果。

摘　　要：对碳市场实际波动特征及风险状况作出尽可能准确的描述,是碳市场发展过程中一个关键性的课题。本文以欧盟碳市场和北京碳市场为例,首先运用GJRSK模型全面考察了碳市场方差、偏度和峰度的时变性,然后基于严谨的后验分析探讨了考虑时变高阶矩信息的模型在碳市场VaR和ES预测中的适用范围和精确程度,并与GJR模型的相应结果作对比研究。研究发现,与条件方差一样,碳收益的条件偏度和条件峰度也具有十分显著的时变性,并且三者的变化具有同步性;相比GJR模型,能够刻画碳收益时变高阶矩特征的GJRSK模型取得了明显更高的VaR和ES预测精度。最后,本文为全国碳市场建设发展提出有效建议,为我国顺利完成3060双碳目标提供重要经验借鉴。It is a key issue in the development of carbon market to accurately describe the market’s volatility characteristics and market risk. In this paper, the EU ETS and Beijing carbon market are taken as examples. Firstly, the GJRSK model is used to comprehensively investigate the time-varying characteristics of the variance, skewness and kurtosis of the carbon return. Then, based on rigorous backtesting method, we compare the forecasting accuracy in the VaR and ES between GJR model and GJRSK model. The results shows that the variance, skewness, kurtosis of two carbon returns have notable time-varying characteristics, and the volatility of the variance, skewness and kurtosis are synchronized. Compared with GJR model, GJRSK model, which can describe the time-varying characteristics of higher-moments of carbon return, have significantly higher VaR and ES forecasting accuracy. Finally, this paper puts forward effective suggestions for the construction and development of the national carbon market and provides important experience reference for China to successfully achieve the 3060 dual carbon target.

关 键 词：碳市场 时变高阶矩 风险预测 后验分析 

分 类 号：C32[社会学] G1[文化科学] P28[天文地球—地图制图学与地理信息工程]