基于双成分已实现EGARCH模型的VaR度量研究

Measuring VaR Based on Two-Component Realized EGARCH Model

在对资产收益率和已实现波动率测度同时建模的已实现EGARCH模型的基础上,将资产收益率的波动率分解为两个成分:长期成分与短期成分,并引入偏斜t分布描述资产收益率的分布,构建了双成分已实现EGARCH模型对VaR进行测度.构建的模型充分利用了高频与低频数据信息,能够迅速捕捉大的市场波动,同时能够捕获波动率非对称性(杠杆效应)与长记忆性,充分刻画资产收益率的偏斜、尖峰厚尾分布特征,具有较高的建模灵活性,且易于实现.采用上证综合指数和深证成分指数日内高频数据,对双成分已实现EGARCH模型进行了实证研究,结果表明:沪深股市波动率具有高度的持续性以及显著的杠杆效应,且杠杆效应主要体现在短期波动率成分中;双成分已实现EGARCH模型相比单成分已实现波动率模型—已实现GARCH模型和已实现EGARCH模型不仅具有更好的样本内数据拟合效果,而且具有更为优越的样本外VaR预测效果.

Ba^sed on the realized EGARCH model considering both asset returns and realized volatility measure,this paper proposes the two-component realized EGARCH model that decomposes the volatility of asset returns into two components,a long-run component and a short-run component,and includes skewed t-distribution for the return distribution to measure VaR.The proposed model makes full use of the information of high-frequency and low-frequency data,can quickly capture severe market fluctuation.Moreover,the model can account for volatility asymmetry(leverage effect)and long-memory volatility and capture skewness,leptokurtosis and fat tails of return distribution.In addition,the model is flexible and easy to implement.We apply the proposed two-component realized EGARCH model to intraday high-frequency data of the SSE Composite Index and SZSE Component Index.The results show that the volatilities of Shanghai and Shenzhen stock markets are highly persistent,and significant leverage effects exist in both the markets and mainly exist in the short-run variance component.Compared with the traditional one-component realized volatility models,the realized GARCH model and the realized EGARCH model,the two-component realized EGARCH model performs better both in in-sample data fitting and out-of-sample VaR forecasting.