股票价格序列中随机跳跃存在性的一种检验方法  

A Method of Jump-Existence Test of Stock Price Series

作　　者：张利兵[1] 潘德惠[2] 

机构地区：[1]上海立信会计学院 [2]东北大学工商管理学院,辽宁沈阳110004

出　　处：《数学的实践与认识》2008年第19期27-33,共7页Mathematics in Practice and Theory

基　　金：中国博士后基金(20060390628);国家自然科学基金项目(70701023)

摘　　要：提出股票价格序列跳跃的一种检验方法.假设价格具有连续样本路径,建立一个关于股票价格样本观察的统计量,利用中心极限定理求得该统计量的极限分布为正态分布,这样,当该统计量超出基于极限分布算出的临界水平时,可以拒绝原假设,认为样本中存在跳跃.用此方法来应用于中国股市沪深股票指数,得到了中国股市存在随机跳跃的直接证据.提出的跳跃检验方法无需对连续部分的波动率形式作过多的假设,克服了波动率模型对检验准确性的影响.结果对金融资产的定价、投资和风险管理都具有积极的意义.A method of testing existing of jumps in stock price series is suggested. Assuming a stock price has continuous sample path, a statistic is given and its asymptotic distribution is constructed by center limit theorem, with which the null hypothesis of no jumps can be denied if ＇the statistics of observed data is lager than the critical value. Empirical study of the data of Shanghai and Shenzhen using this method shows that jumps exist in financial data. The method of this paper needs little assumption of the volatility function, thus can overcome the effect of the volatility model on the test accuracy. This method can be helpful of asset pricing, investment and risk management.

关 键 词：波动率 跳跃 统计量 极限分布 

分 类 号：F830.91[经济管理—金融学] F224