基于SWARCH-GED模型的极值VaR度量  

Extreme Value at Risk Measurement Based on SWARCH-GED Model

作　　者：周孝华[1] 张保帅[1] 李强[1] 

机构地区：[1]重庆大学经济与工商管理学院,重庆400030

出　　处：《系统工程》2013年第1期29-36,共8页Systems Engineering

基　　金：重庆市自然科学基金资助项目(CSTC2011BB2088);中央高校基本科研业务费项目(CDJXS11021112)

摘　　要：针对金融资产收益率序列的尖峰厚尾和结构变换特征,用SWARCH-GED模型拟合收益率序列,然后以标准残差序列为媒介与极值理论有机结合起来组建基于SWARCH-GED-EVT的动态VaR模型,最后对模型的精度和有效性进行验证。研究表明,SWARCH-GED-EVT模型能够有效识别上证综指的尖峰厚尾和结构变换特征,并且能精确有效地测度上证综指收益风险,尤其随着置信水平的提高效果更明显。The volatility and structure transformation characteristics of the return series are captured by the SWARCH-GED model, then the return series is transformed into the standard residuals series. On this basis, the articles use SWARCH-GED models incorporating Extreme Value Theory to establish a financial risk measure model-the dynamic VaR model based on SWARCH-GED-EVT. The results show, the SWARCH-GED-EVT model can effectively identify SSEC Index volatility regime characteristics, and it is more advantageous than the other several models about SSEC Index return risk measure, especially in the high confidence level.

关 键 词：SWARCH—GED模型 极值风险 门限值 

分 类 号：F830[经济管理—金融学]