基于半鞅过程的中国股市随机波动、跳跃和微观结构噪声统计特征研究

本文基于半鞅过程和非参数统计推断方法,利用已实现幂变差的渐进统计特性,构造检验统计量,在统一的分析框架下,对金融资产价格中随机波动、跳跃和微观结构噪声等问题进行全面系统的研究。并根据上海证券交易所不同行业的股票,上证50 股票指数及其成分股的高频数据进行实证研究。结果表明,我国A 股市场中,噪音交易显著;约43%的风险来源于资产收益过程的随机波动风险,可用股票期权交易对冲;不同来源风险的重要性程度依次为:随机波动的风险、系统性跳跃风险以及异质性跳跃风险;流动性越好的股票越显示出跳跃、尤其是无限小跳的证据。

Analysis of the Finer Statistical Characteristics of China Stock Market Based on Semimartingales Process

In this paper the different asymptotic behavior of the power variations is exploited as the power p, the truncation level and the sampling frequency are varying,and test statistics is developed on the realized power variation, then a systematic econometric analysis of stochastic volatility,jump and noise existing in high frequency financial returns is given based on Semimartingales Process, asset returns sampled at high frequency are decomposed into their base components (continuous, small jumps, large jumps), the relative magnitude of the components is determined, considering market microstructure noise. The methodology is applied to individual stock returns from different industries, those with different liquidity as well as stock index returns and its constituent stocks. Our results show that noisy traders exist widely in CSM; 43% risk results from stochastic volatility risk in asset return process, possibly hedged by equity option; the importance order of risk from different sources is stochastic volatility, systemic jump and heterogeneous jump; more liquid equities have more significant proof of jump, especially infinite small jumps.