基于SV-POT-TDRM的沪深300股指期货尾部风险研究

发布时间：2018-05-04 16:03

本文选题：尾部扭曲风险 + 极端风险　；参考：《系统管理学报》2017年05期

【摘要】：使用随机波动率模型修正沪深300股指期货收益率序列的波动聚集效应,并在残差服从正态分布和极值分布的假设下,分别计算了度量尾部风险的VaR、ES及尾部扭曲风险测度(TDRM)值。研究发现:股指期货日收益率序列呈现负偏、尖峰厚尾及波动聚集的形态;使用随机波动率模型可以较好地预测波动率的变化;假设残差分布服从极值分布的模型结果优于假设残差分布服从正态分布的模型结果,说明极值模型在尾部分析上比正态分布更加适用;使用扭曲尾部风险测度估计尾部风险,通过扭曲函数的选取与风险厌恶系数的不同设定,调整尾部风险发生的概率,反映了投资者的主观风险偏好,在相同置信水平下,得到的尾部风险估计值比VaR更精确。
[Abstract]:Using the stochastic volatility model to modify the volatility aggregation effect of Shanghai and Shenzhen 300 stock index futures return series, and under the assumption of normal distribution and extreme value distribution of residual service, the VaRNES and TDRM of tail risk are calculated respectively. It is found that the daily yield series of stock index futures show negative deviation, thick tail of peak and aggregation of volatility, and the random volatility model can better predict the change of volatility. The model result of assuming residual distribution from extreme value distribution is better than that of assuming residual distribution from normal distribution, which shows that extreme value model is more suitable for tail analysis than normal distribution, and use distorted tail risk measure to estimate tail risk. Through the selection of distortion function and the different setting of risk aversion coefficient, the probability of tail risk occurrence is adjusted to reflect the subjective risk preference of investors. Under the same confidence level, the estimated value of tail risk is more accurate than that of VaR.
【作者单位】：大连理工大学管理与经济学部;
【基金】：国家自然科学基金资助项目(71471026,71171032)
【分类号】：F224;F724.5

【相似文献】