基于条件极值模型的上证综指尾部风险研究

Tail risk research for Shanghai composite index based on conditional extreme value model

在传统ARMA-GARCH时间序列模型的基础上,介绍条件极值模型并运用这些模型对近十几年来上证综指进行VaR和ES样本外预测与事后检验.研究表明假设新息序列为偏t分布的ARMA-GARCH模型与条件极值模型在预测VaR和ES方面均具有出色效果.

Based on traditional ARMA- GARCH time series models, conditional extreme value theory（EVT） model were introduced and applied to do out of sample prediction and back testing about VaR and ES for Shanghai composite index. Research showed that ARMA -GARCH model with skewed t innovations, and conditional EVT models perform well for VaR and ES prediction.