您当前的位置：
站内搜索:
有界分形市场假说与中国股市的实证研究（1）
作者：佚名 来源：不详 发布时间：2008-6-13 19:52:59|
收藏 []
">
增大字体 摘 要
本文从现代金融学的一个基本问题:股价运动是否随机出发,探讨了两种不同的理论框架――均衡体系和复杂系统对金融市场本质的不同刻划.我们表明有效市场假说代表的均衡体系与经验证据不符,不能反映金融市场的复杂系统本质.基于前沿的金融研究领域——经济物理学,我们提出了有界分形市场假说,该假说使用复杂系统的一支——分形的方法来研究金融市场的行为.我们主要考察了分形系统的长程相关和统计标度行为,并表明二者作为复杂系统的特征具有内在一致性.在对中国股市进行的实证研究中,我们发现上证指数和深圳指数存在长程相关和统计自相似的有力证据.我们使用R/S分析估计了长程相关的特征量赫斯特指数,并通过0概率的标度行为估计了列维稳定分布的自相似参数.稳定性分析验证了以上结果的稳健性.在R/S分析中我们发现上证指数存在着"非周期性循环",其平均周期约为一年,而深圳成份指数则没有发现类似的非周期循环证据.实证研究的结果一致地表明中国股市存在内在的分形性质,验证了有界分形市场假说.本文的研究对于发展新的理论框架,应用新的数理工具和统计方法认识金融市场的本质提供了令人感兴趣的证据,并为进一步的研究提供了基础.
关键字:有效市场假说,有界分形市场假说,标度行为,长程相关,R/S分析,列维稳定分布
Abstract
Starting from a fundamental issue of modern finance – the random hypothesis of asset price motion, this paper discussed two different theoretic frameworks: equilibrium system and complex system and their different descriptions of the essence of financial markets. We showed that the equilibrium framework, represented by Efficient Market Hypothesis, can not explain empirical findings and failed in reflecting the essence of financial market as complex system. Based on newly developed finance frontier - econophysics, we proposed Bounded Fractal Market Hypothesis. Our discussion focused on two characteristics of fractal system: long term correlation and scaling behavior in statistical distribution. We showed that the two properties are consistent as intrinsic property of complex system. In the empirical analysis, we found strong evidence of long term persistence and statistical similarity on Shanghai Integrated Index and Shenzhen Index. Using R/S analysis, we estimated Hurst Exponent to characterize the long term persistence. Also by analyzing the scaling behavior of zero probability, we estimated self-similar coefficient of Levy stable distribution. We further conducted stability analysis to verify the robustness of estimated coefficients. In R/S analysis we found interesting evidence of "Non-periodic Cycle" in Shanghai Integrated Index with average cycle of one year. However in Shenzhen Index we haven't found similar b
增大字体 本文从现代金融学的一个基本问题:股价运动是否随机出发,探讨了两种不同的理论框架――均衡体系和复杂系统对金融市场本质的不同刻划.我们表明有效市场假说代表的均衡体系与经验证据不符,不能反映金融市场的复杂系统本质.基于前沿的金融研究领域——经济物理学,我们提出了有界分形市场假说,该假说使用复杂系统的一支——分形的方法来研究金融市场的行为.我们主要考察了分形系统的长程相关和统计标度行为,并表明二者作为复杂系统的特征具有内在一致性.在对中国股市进行的实证研究中,我们发现上证指数和深圳指数存在长程相关和统计自相似的有力证据.我们使用R/S分析估计了长程相关的特征量赫斯特指数,并通过0概率的标度行为估计了列维稳定分布的自相似参数.稳定性分析验证了以上结果的稳健性.在R/S分析中我们发现上证指数存在着"非周期性循环",其平均周期约为一年,而深圳成份指数则没有发现类似的非周期循环证据.实证研究的结果一致地表明中国股市存在内在的分形性质,验证了有界分形市场假说.本文的研究对于发展新的理论框架,应用新的数理工具和统计方法认识金融市场的本质提供了令人感兴趣的证据,并为进一步的研究提供了基础.
关键字:有效市场假说,有界分形市场假说,标度行为,长程相关,R/S分析,列维稳定分布
Abstract
Starting from a fundamental issue of modern finance – the random hypothesis of asset price motion, this paper discussed two different theoretic frameworks: equilibrium system and complex system and their different descriptions of the essence of financial markets. We showed that the equilibrium framework, represented by Efficient Market Hypothesis, can not explain empirical findings and failed in reflecting the essence of financial market as complex system. Based on newly developed finance frontier - econophysics, we proposed Bounded Fractal Market Hypothesis. Our discussion focused on two characteristics of fractal system: long term correlation and scaling behavior in statistical distribution. We showed that the two properties are consistent as intrinsic property of complex system. In the empirical analysis, we found strong evidence of long term persistence and statistical similarity on Shanghai Integrated Index and Shenzhen Index. Using R/S analysis, we estimated Hurst Exponent to characterize the long term persistence. Also by analyzing the scaling behavior of zero probability, we estimated self-similar coefficient of Levy stable distribution. We further conducted stability analysis to verify the robustness of estimated coefficients. In R/S analysis we found interesting evidence of "Non-periodic Cycle" in Shanghai Integrated Index with average cycle of one year. However in Shenzhen Index we haven't found similar b
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] 下一页
相关文章
学习第一网·WWW.XueXiNo1.COM
本类热门排行