中国股票市场的非线性结构实证研究
|
摘要
传统的资本市场理论研究一直是在新古典经济学的研究构架下进行,作为新古典经济学核心的有效市场假定(EMH)是资本市场理论研究的基石,许多经典资本市场理论都是在EMH基础上发展起来或与之有密切联系。从新古典主义经济学关于资本市场的研究框架可知,传统资本市场理论大多是在一个线性范式上的分析。但现实资本市场是一个复杂的非线性动态系统,线性分析并不能很好地揭示资本市场的真正本质特征,国内外许多实证研究都表明了有效市场假说(EMH)解释资本市场的不足。因此,对资本市场的研究客观上需要从非线性的方式来进行。全文分五个主要部分:第一部分介绍了论文写作的背景和意义,回顾了国内外有关的研究成果,并简单描述了本文的结构和创新点;第二部分详细介绍了分形市场理论和与有效市场的比较,分形R/S分析方法的来源和理论基础;第三部分介绍了运用分形R/S方法开展实证分析的相关理论;第四部分为实证部分,采用R/S分析方法分析沪深两市,选取国内应用最为广泛的上海综合指数和深圳成份指数的周收益率作为研究的对象,将它们分成94年5月前后和2001年1月前后共三段样本。我们根据分析结果得出以下结论:(1)中国沪深两市综合指数的对数收益率存在较为明显的非线性特征;(2)从全区间看,上证指数的非线性特征比深成指强,存在280天左右的平均周期,而深成指的周期约为350天左右;(3)通过对周对数收益率的非线性结构BDS检验,结果表明沪深两市2001年后到现在的周对数收益率的随机程度有显著的提高;(4)分析发现我国股票市场从建立到2001年间,市场存在明显的低维混沌过程。同时,周对数收益率的赫斯特指数也显示其具有反持续性非线性特征。
Traditional capital market theories are studied in the frame of neoclassical economics. As the core of neoclassical economics, Efficient Market Hypothesis (EMH) has always being the foundation of capital market theories. Many classical capital market theories are developed on the basis of EMH or on some bases closely related with it. Viewing from this, traditional capital market theories are studied by means of analysis on which is based linear model, but the practical capital market can be viewed as a complicated nonlinear dynamic system, the linear analysis couldn't be used to precisely describe the practical capital market. Many empirical analysis, home and abroad, show the limitation of EMH. So it's an inevitable selection to study capital market in nonlinear ways instead. The paper mainly consists of five portions and the main content is as follows: The first portion, as problem bringing forward, expounds the origin and meaning of this paper. Reviews the Domestic and international relevant research result, simply expounds the structure and innovations. The second portion particularly expounds the compare of the EMH and FMH, and shows the origin and theory basis of the R/S analysis. The third portion expounds the relevant concepts about content that the R/S analysis empirically researches the stock market in china. The fourth portion is the empirical analysis, The paper by the R/S analysis empirically researches the stock market in china, and gives the empirical analysis of the weekly logarithm returns of Shanghai General Index and Shenzhen Component Index and conduct comparative analysis of three portions before and after May 1994 and December 2001. The fifth portion is the conclusion. According to the analysis result we can receive the following conclusion:(1) Both Shanghai and Shenzhen Stock Market, from the whole extent, logarithm returns have the obvious nonlinear character.(2)From the whole extent, Shanghai Stock Market has much obvious nonlinear character than Shenzhen Stock Market and Shanghai Stock Market have average circle length about 280 days and Shenzhen Stock Market average circle length is 350 days (3) By using the BDS test of both weekly logarithm return series shows that the randomicity of weekly logarithm returns of Shanghai and Shenzhen Stock Market is obvious improved since 2001. (4) From the establishment of our stock market to 2001, the analysis shows that the market has the obvious low-dimension chaos process, and the Hurst index of the weekly logarithm returns also reveals that it has the anti-continuity character.
Traditional capital market theories are studied in the frame of neoclassical economics. As the core of neoclassical economics, Efficient Market Hypothesis (EMH) has always being the foundation of capital market theories. Many classical capital market theories are developed on the basis of EMH or on some bases closely related with it. Viewing from this, traditional capital market theories are studied by means of analysis on which is based linear model, but the practical capital market can be viewed as a complicated nonlinear dynamic system, the linear analysis couldn't be used to precisely describe the practical capital market. Many empirical analysis, home and abroad, show the limitation of EMH. So it's an inevitable selection to study capital market in nonlinear ways instead. The paper mainly consists of five portions and the main content is as follows: The first portion, as problem bringing forward, expounds the origin and meaning of this paper. Reviews the Domestic and international relevant research result, simply expounds the structure and innovations. The second portion particularly expounds the compare of the EMH and FMH, and shows the origin and theory basis of the R/S analysis. The third portion expounds the relevant concepts about content that the R/S analysis empirically researches the stock market in china. The fourth portion is the empirical analysis, The paper by the R/S analysis empirically researches the stock market in china, and gives the empirical analysis of the weekly logarithm returns of Shanghai General Index and Shenzhen Component Index and conduct comparative analysis of three portions before and after May 1994 and December 2001. The fifth portion is the conclusion. According to the analysis result we can receive the following conclusion:(1) Both Shanghai and Shenzhen Stock Market, from the whole extent, logarithm returns have the obvious nonlinear character.(2)From the whole extent, Shanghai Stock Market has much obvious nonlinear character than Shenzhen Stock Market and Shanghai Stock Market have average circle length about 280 days and Shenzhen Stock Market average circle length is 350 days (3) By using the BDS test of both weekly logarithm return series shows that the randomicity of weekly logarithm returns of Shanghai and Shenzhen Stock Market is obvious improved since 2001. (4) From the establishment of our stock market to 2001, the analysis shows that the market has the obvious low-dimension chaos process, and the Hurst index of the weekly logarithm returns also reveals that it has the anti-continuity character.
引文