基于日间高频交易数据的股市波动研究

Shares flutuation research based on day high frequency trade data

本文利用上证综指在2005～2009年内的日间高频数据,通过已实现波动率这一概念对我国股市在这5年间的波动特性做了研究。进一步地,根据已实现波动率序列的统计特征,对其进行长记忆建模,并对模型的波动率预测效果与常规GARCH模型的预测效果做了对比分析。基于上证综指的研究结果表明,利用了日间高频信息的波动率模型在波动率预测上,比仅利用了收盘信息的GARCH模型更有优势。

This paper studies the volatility characteristics of China stock market using the intraday high-frequency data of the Shanghai Stock Exchange Composite Index（SSEC） from year 2005 to year 2009 via the idea of Realized Volatility.Moreover,according to the statistical properties of RV series of SSEC,this paper builds two long memory models on logarithmic RV and compares them with the conventional GARCH model in the domain of volatility forecasting.The analyses results show that the models using intraday high-frequency data are better than the model using only daily close price information in forecasting volatility.