股票价格行为分形特征的实证分析

The Empirical Study on Fractal Price Behavior of Stock Indexes

价格的随机游走和市场有效是主流金融计量理论的重要理论基石。然而,长期以来,主流的有效市场假说就不断受到市场实际运行状况和相关研究的争议。20世纪80年代以来,分形市场研究,作为非线性经济学的一个子集,从非线性的角度,对主流金融计量理论线性模式的基础假设:随机游走及与之相应的正态分布,提出了争议和挑战。本文将分形市场研究的理论和方法,应用于中国股票市场的价格行为研究,从非线性的角度,分析中国股市的价格行为。本文采用重标极差法和消除趋势波动分析法对上证综合指数、深证成分指数进行分析。实证结果表明,两市指数收益率的分布相对正态分布来说,有显著异常的峰值和偏斜系数,因此不能用正态分布来近似表示;日数据的重标极差分析和日内数据的消除趋势波动分析结果显示,两市证券指数的价格行为具有持久性、记忆性和标度不变性等非线性特征。

Issues on stock price behavior and pricing mechanism are widely concerned and constantly debated in the field of finance and investment research.The hypotheses of random walk and efficient market are important foundation stones of mainstream finance theories.For long time,however,the Efficient Market Hypothesis（EMH）has been disputed by factual status of market＇s running and relevant investigations.Since 1980s＇,Fractal Market Analysis,as a subset of nonlinear economics,has brought forward challenge and controversy on the key assumptions of EMH＇s linear paradigm that focuses on random walk of price movement and normal distribution of asset return.This paper studies the price behavior and efficiency of China＇s stock markets from the angle of nonlinearity by applying the theory of Fractal Market Analysis.This study examined the behavior of Shanghai Composite Index and Shenzhen Sub-Index with the methods of rescaled range analysis and detrended fluctuation analysis.The empirical results show that the distributions of the logarithmic returns of the two indexes are greatly different from normal distribution because of the significant skewness and kurtosis.The empirical research on day data with R/S method,on intra-day data with DFA method,show the memorial,coherent and persistent properties in China＇s Stock Markets.