摘要
作为GARCH族模型的重要拓展模型之一,PGARCH模型的估计往往采用基于日度数据的拟极大似然估计方法。为了探究高频信息对PGARCH模型估计的影响,基于Visser (2011)的研究,本文使用波动率代表模型来整合高频数据,并使用拟极大指数似然估计方法(QMELE)对PGARCH模型进行估计,同时探究了拟极大指数似然估计的渐近性质和模型估计效率的评判标准。模拟研究和实证分析证实,基于高频数据的拟极大指数似然估计有效地提升了PGARCH模型的参数估计精度,这说明基于高频数据的拟极大似然指数估计具有一定的应用价值。
As one of the important extended models of the GARCH family model, it is more common to use quasi maximum likelihood estimator to fit PGARCH model with daily data. Referring to Visser (2011), this article applies volatility representative models to integrate high-frequency data, and then uses the quasi maximum exponential likelihood estimator (QMELE) to fit the PGARCH model, so that we can explore the impact of high-frequency information on the estimation of the PGARCH model. Meanwhile, we establish the asymptotic properties of QMELE and the evaluation criteria for model estimation efficiency. Simulation research and empirical analysis have confirmed that QMELE with high-frequency data effectively improves the parameter estimation accuracy of the PGARCH model, which indicates that the QMELE based on high-frequency data has a certain application value.
出处
《统计学与应用》
2023年第4期1085-1095,共11页
Statistical and Application
