典型事实约束下的上海燃油期货市场动态VaR测度研究

A Study on Dynamic VaR Predicting Models for Oil Futures Market of Shanghai

期货交易的高杠杆率意味着期货市场的高风险特征,而能源市场因其特殊的战略意义一直以来备受关注,因而对能源期货市场的风险测度对投资者和监管者都极其重要。本文对上海燃油期货构建了四个反映不同交割期限的连续价格序列,基于不同的金融市场典型事实分别运用GARCH、GJR、FIGARCH三个模型对波动率建模,并假设条件收益分别服从正态、学生t、有偏学生t(skst)分布进行动态风险价值(VaR)测度,然后运用严格的似然比(LR)检验和动态分位数回归(DQR)检验对风险测度的可靠性进行后验分析(Backtesting),尝试从中提取出在风险管理中最有应用价值的典型事实。研究发现:(1)基于skst分布的波动模型的动态风险测度准确性明显优于其他分布下的相同模型;(2)基于杠杆效应的GJR模型和基于长记忆性的FIGARCH模型并没有表现出比普通GARCH模型更高的精度;(3)远期合约的市场平均收益更高,风险测度比近期合约更准确。

The high leverage of futures means high-risk,and energy market is always concerned because of its strategic significance.So the risk measure of the energy futures market is very important to both investors and regulators.In this paper,four continuous price series are constructed to reflect different delivery period of oil futures listed in Shanghai.Based on different financial stylized facts,GARCH,GJR and FIGARCH are used to model volatility.Under the assumption of the conditional return obeying normal,student t and skewed student t（skst） distributions,dynamic VaR is measured.Then both LR（Likelihood Ratio） test and DQR（Dynamic Quantile Regression） test are used to backtest the accuracy of these models and try to extract the best valuable stylized facts.The results show that：（1） the dynamic VaR measurement with skst distribution is more accurate;（2） the GJR models based on leverage effect and FIGARCH models based on long memory do not perform better than GARCH model;（3） the average return of far futures is higher and dynamic VaR is easier to measure.