【作者】 卢方元；

【导师】 黄登仕；

【作者基本信息】 西南交通大学 ， 管理科学与工程， 2005， 博士

【摘要】 金融资产收益率是金融经济学中的一个非常重要的概念,能否对收益率的波动状况进行正确描述直接关系到证券组合选择的正确性、风险管理的有效性、期权定价的合理性。本文对上海综合指数收益率和深圳成分指数收益率的波动性进行实证研究,主要研究内容为: 1.中国股市收益率基本统计特征研究。对沪深股市收益率的基本统计量、独立性与相关性、正态性检验、厚尾性检验、平稳性检验、ARCH效应等进行分析研究。所得到的结论为:沪深股市收益率序列不独立而具有长期的相关性、不服从正态分布而具有明显的尖峰厚尾性;沪深股市收益率序列具有一定的平稳性、存在ARCH效应;沪市的波动性大于深市的波动性,沪市存在明显的杠杆效应,而深市的杠杆效应不显著。 2.中国股市收益率的长期相关性研究。在对时间序列长期相关性分析方法评述的基础上,利用较适合中国股市实情的DFA方法对沪深股市三种形式的收益率(一般收益率、绝对值收益率、平方收益率)序列进行整体相关性分析、局部长期相关性分析以及标度不变性分析。所得到的结论为:沪深股市三种收益率序列均不同程度上存在长期相关性,其中绝对值收益率的长期相关性程度最高;沪深股市三种收益率序列的标度指数不具有一致性,而是时间标度的复杂函数,具有明显的多标度特征;沪深股市收益率序列不遵循纯粹的随机游走过程,沪深股市过去的消息在很长一段时间内会影响未来,因此沪深股市不是弱式有效市场,利用数学模型对收益率进行预测是可行的,但要做出准确预测是困难的。 3.中国股市收益率多重分形研究。使用最新的多重分形分析方法——多重分形消除趋势波动分析法(简称MF-DFA)对中国股市收益率的多重分形性进行分析并探讨多重分形形成的原因。所得到的结论为:沪深股市收益率序列均呈现出明显的多重分形;沪深股市收益率的多重分形是由收益率的长期相关性和厚尾分布共同作用的结果。 4.中国股市收益率分布特征研究。使用两种比较典型的能刻画尖峰、厚更多还原

【Abstract】 Return of financial assets is a very important concept in financial economics. The right description to the fluctuation of stock returns is related to the exactness of choosing securities association, the validity of risk management, the rationality of pricing options. This article mainly aims to empirical analyses to the fluctuation of Shanghai Synthesis index returns and Shenzhen composition index returns. The following are the main contents.The basic statistical characteristics on the returns of stock market in China. Through studying the basic statistic, independent and correlation, test of normality , test of fat tail , test of stationary, ARCH effect of Shanghai and Shenzhen Stock market etc, some conclusions can be acquired: The returns series of Shanghai and Shenzhen Stock market is not independent but has a long-range correlation, and is disobedient normally distribution but has obvious peak and fat tail; The returns series of Shanghai and Shenzhen Stock market has certainly stationary and ARCH effect; The fluctuation of the Shanghai stock market is over and above the one of Shenzhen. The former has more obvious lever effect than the later.The long-range correlation of the returns in Chinese stock market. Onthe basis of the review of the analysis method on the long-range correlation of the time series, as for the series of the three kinds of forms of the returns of Shanghai and ShenZhen Stock Markets, we adopt the DFA method relatively appropriate for realistic stock market of China to carry out the whole long-range correlation analysis, partial long-range correlation analysis and mark a scale constant analysis.Then we get draw a conclusion: there is an evidence of long-rang correlations for the three returns, and the intensity of correlation of the absolute returns is the strongest. Neither of the scale indexes of the three kinds of forms of the returns of Shanghai and Shenzhen stock markets complies with a consistency, but both appear a complex function of time scale and show a multi-scale character. The series of the returns of Shanghai and Shenzhen stock更多还原