论文部分内容阅读
本文以中国商品期货市场的3类主要期货价格指数为研究对象,探讨了当期货收益同时存在波动聚集性、非对称杠杆效应以及条件“有偏”和“尖峰胖尾”特征状况下的市场风险价值(VaR)计算模型。同时,运用更加严谨和稳健的动态分位数回归(Dynamic quantile regression)检验法,对各类不同波动模型和收益分布假定下的VaR估计精度进行了全面的后验分析(Backtesting)。实证结果发现,中国商品期货市场的价格波动不存在显著的杠杆效应。但是,假定条件收益服从有偏学生分布(Skewed student distribution)的波动率模型具有较好的风险测度精度。
In this paper, we take the three major futures price indices of China’s commodity futures market as the research object, and discuss the characteristics of volatility aggregation, asymmetric leverage and conditional conditions of “partial” and “sharp fat tail” Under the market VaR (VaR) calculation model. In the meantime, using a more rigorous and robust Dynamic quantile regression test, the VaR estimation accuracy under various volatility models and return distribution assumptions was comprehensively verified by backtesting. The empirical results show that there is no significant leverage effect on the price volatility of China’s commodity futures market. However, it is assumed that the conditional returns obey the volatility model of Skewed student distribution and have better risk measurement accuracy.