高频GARCH模型的最优抽样分析  

Optimal sampling analysis for GARCH model with high frequency data

作　　者：程凌筠 宋泽芳 张兴发[1,2] 李莉丽 CHENG Ling-jun;SONG Ze-fang;ZHANG Xing-fa;LI Li-li(School of Economics and Statistics,Guangzhou University,Guangzhou 510006,China;Lingnan Research Institute of Statistical Science,Guangzhou University,Guangzhou 510006,China)

机构地区：[1]广州大学经济与统计学院,广东广州510006 [2]广州大学岭南统计科学研究院,广东广州510006

出　　处：《广州大学学报（自然科学版）》2023年第3期75-82,共8页Journal of Guangzhou University:Natural Science Edition

基　　金：国家自然科学基金重点资助项目(11731015);国家自然科学基金青年资助项目(11701116);广州市基础与应用基金研究资助项目(202201010276)。

摘　　要：波动率代表是运用高频数据估计日频GARCH类模型时构造的一个重要统计量,文章针对运用高频数据估计日频GARCH模型的3种方法,基于估计量的渐近结果讨论了最优波动率代表的选择问题,并展开了在日内高频数据抽样中应用的讨论。通过对沪深300指数的高频数据实证分析发现:同一波动率代表在不同抽样频率下的表现有明显差异;在同一频率下,不同波动率代表有优劣之分;在不同估计方法下,每一个波动率代表的最优频率都不同。因此,日内高频数据的最优抽样频率应针对模型所用的不同估计方法加以区别对待。Volatility proxies are an important statistic applied to estimate daily GARCH model by using high frequency data.This paper proposes the criterions for choosing an optimal volatility proxy by the asymptotic properties of three GARCH estimators based on the intraday high frequency data,and the applications of these criterions in high frequency data sampling are also discussed.The empirical study of the high-frequency data of CSI 300 index shows that:for the same volatility proxy,the performance of different frenquencies is obviously different;for the same frequency,different volatility proxies perform differently;each volatility proxy has a different optimal frequency under different estimation methods.Consequently,the optimal sampling frequency for intraday high frequency data should be treated differently among different estimation approaches.

关 键 词：波动率代表 抽样频率 GARCH模型 高频数据 

分 类 号：O212.1[理学—概率论与数理统计]