摘要
以西德克萨斯中质原油(WTI)为研究对象,采用expectile模型测度原油价格风险,同时引入极值理论,构建exepctile-EVT模型刻画极端风险。研究表明:油价收益序列具有典型的长记忆性和自相关特征;油价的涨跌会影响多头风险,而空头风险仅受油价下跌的影响;引入极值理论后的expectile-EVT模型能很好地刻画极端风险的动态演化规律,其预测结果也比其他模型更合理。
By taking WTI crude oil as research object in this paper,expectile model is adopted to directly calculate the common level risk of price.Meanwhile,a new framework combining expectile and EVT is constructed for extreme risk measurements.The empirical results shows that:oil price exhibits typical long-memory and autocorrelation properties;the rise and fall of price will together influence the long position risk value while short position only influenced by the fall;adding EVT into expectile model can effectively depict the dynamic evolution of extreme oil price risk.The predicting results are therefore more reasonable than benchmark models.
出处
《统计与信息论坛》
CSSCI
北大核心
2018年第1期58-64,共7页
Journal of Statistics and Information
基金
教育部规划基金项目<绿色发展中能源环境政策效应的动态CGE研究>(17YJA790030)
湖南省哲学社会科学基金重点项目<基于市场化机制与手段的湖南省环境治理与生态建设研究>(15ZDB030)
关键词
油价风险
期望分位数
极值理论
oil price risk
expectile
extreme value theory
