Abstract: It is very important for market managers and participants to accurately predict the volatility of financial markets. On the basis of the standard GARCH model, this paper decomposes multiply conditional variance into two parts: long-term variance and short-term variance, and constructs the long-term variance equation and short-term variance equation including lever function, respectively, to capture the long-memory and short-term micro-fluctuation of volatility. Empirical analysis using high frequency data such as daily closing price, realized variance and realized kernel of Shanghai Composite Index and Nikkei Index shows that the two-factor GARCH model of the two indexes has achieved greater likelihood estimates in samples than the standard GARCH model, and the two-factor GARCH model has also realized better performance through out-of-sample error function analysis and DM test.

Key words: two-factor realized GARCH model, realized variance, realized kernel, volatility prediction

摘要： 准确地预测金融市场的波动率对市场管理者和参与者而言都是至关重要的。本文在标准已实现GARCH模型基础上,将条件方差乘性分解为长期方差和短期方差两部分,分别构造包含杠杆函数的长期方差方程和短期方差方程,用以捕捉波动率的长记忆性和短期微观波动。运用上证综指和日经指数的日收盘价、已实现方差和已实现核波动此类高频数据进行实证分析,结果表明:与标准已实现GARCH模型相比,两指数的双因子已实现GARCH模型在样本内表现出更大的似然估计值;通过样本外误差函数分析和DM检验,双因子已实现GARCH模型也取得更好表现。

关键词: 双因子已实现GARCH模型, 已实现方差, 已实现核波动, 波动率预测