主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2012, Vol. 20 ›› Issue (6): 1-8.

• 论文 •    下一篇

中国燃油期货市场的VaR与ES风险度量

王鹏1, 魏宇2   

  1. 1. 西南财经大学金融学院,四川 成都 610074;
    2. 西南交通大学经济管理学院,四川 成都 610031
  • 收稿日期:2011-05-05 修回日期:2012-07-10 出版日期:2012-12-29 发布日期:2012-12-28
  • 基金资助:
    国家自然科学基金资助项目(71071131,71101119);教育部新世纪优秀人才支持计划(NCET-08-0826);西南财经大学"2011工程"项目(211QN10110)

Backtesting Risk Models for Chinese Fuel-oil Futures Market

WANG Peng1, WEI Yu2   

  1. 1. School of Finance, Southwest University of Fenance and Economics, Chengdu 610074, China;
    2. School of Economics and Manayement, Southowest Jiaotong University, Chengdu 610031, China
  • Received:2011-05-05 Revised:2012-07-10 Online:2012-12-29 Published:2012-12-28

摘要: 近年来,我国燃油期货市场取得快速发展,但有关该市场波动特征和风险状况的研究却非常缺乏。以上海期货交易所燃油期货价格指数为例,分别在多头和空头两种头寸状况以及5种不同分位数水平下,运用条件覆盖检验、非条件覆盖检验等后验分析方法,实证对比了不同风险测度模型对VaR和ES两种不同风险指标估计的精度差异。研究结果表明:在我国燃油期货市场的风险测度估计中考虑国际燃油价格波动因素有助于获得更为精准的风险测度精度;在综合考虑了模型对价格变化动力学的刻画效果以及对极端风险的测度精度等因素后,FIGARCHCST-SST模型是一个相对合理的风险测度模型选择。

关键词: 中国燃油期货市场, VaR, ES, 有偏学生t分布, 后验分析

Abstract: By taking four representative indices of Chinese fuel-oil futures market as sample, VaR predicting is used for eight risk models. Furthermore, two robust backtesting methodologies, unconditional coverage test and conditional coverage test, are introduced to estimate the accuracy for VaR predictions produced by different models. The main results show that adding international fuel-oil price volatility as explanatory variable in typical models is helpful to improve risk estimation accuracy of Chinese fuel-oil futures market. In addition, FIGARCHCST-SST is moderately good in overall consideration of description efficiency and estimation accuracy to extreme risk.

Key words: Chinese fuel-oil futures market, value at risk, excepted shortfall, skewed student-t distribution, backtesting analysis

中图分类号: