非对称随机波动率模型的偏度影响分析  

The Impact on Skewness of Asymmetric Stochastic Volatility Model

作　　者：孔继红[1] 

机构地区：[1]南京师范大学商学院

出　　处：《金融学季刊》2017年第2期103-121,共19页Quarterly Journal of Finance

基　　金：南京师范大学2016年教改项目:金融专业实验课程实施大纲与规范化建设研究(批准号:18122000091627)的支持

摘　　要：本文考虑了资产收益率遵循广义双曲有偏学生t分布（GHST），及其与波动率误差项存在相关性等的随机波动率模型，以分析其对收益率条件偏度的作用。采用上证综指实证显示，遵循GHST分布且存在滞后相关的SV模型，能更好地捕捉和复制上证综指的关键特征。其中，超额峰度的生成归因于厚尾分布；GHST分布的偏斜参数对收益率负偏性提供了解释；理论上占优的滞后相关性设定并不存在偏度生成机制，但其与偏斜分布的结合，使模型绩效表现更好。Under the specifications of the stochastic volatility model with asset return following generalized hyperbolic skew student＇s t distribution （GHST） and correlation between it and volatility error, the paper analyses the mechanism of skewed distribution and correlation＇s impacts on the asset distribution, especially on the condition skewness. Based on the MCMC estimation, comparative evaluation and simulate forecast of the return data of the Shanghai Stock Composite Index （SSCI） from 2001 to 2015, the paper concludes that the stochastic volatility model with GHST and subsequent correlation can capture and reproduce the characteristic of volatility clustering, negative skewness and excess kurtosis of SSCI. In detail, the excess kurtosis attributes to the fat-tailed return distribution, skewed error can explain the condition and unconditional skewness of return, and when the theoretical dominant subsequent correlation specification which cannot produce the skewness combines with skewed distribution, the comprehensive model can ensure better performance.

关 键 词：偏度 随机波动率模型 广义双曲学生t分布 上证综指 

分 类 号：F830.91[经济管理—金融学]