【作者】 李道叶；

【导师】 王聪；

【作者基本信息】 暨南大学 ， 金融学， 2007， 博士

【摘要】 经典资本市场理论基本上都是以有效市场假定作为研究基石,理性投资者、市场有效和随机游走过程这三个核心前提假设决定了经典资本市场理论是建立在线性范式上,但线性研究范式下发展而来的经典资本市场理论对股票市场许多现象不能给出合理解释,国内外许多实证研究都佐证了有效市场假定的失效,对股票市场价格行为的研究,客观上需要用非线性角度进行。用非线性方法对我国股票市场价格收益率的研究对于丰富股票价格及收益率行为的认识、管理层对股票市场的规范及管理及未来我国不断推出的基于股票价格的金融衍生产品定价都具有一定的理论与实践意义。本论文以沪深股票市场为研究对象,在非线性研究框架下,运用R/S分析法、频谱分析、ARMA模型、异方差模型、因子分析、分形及混沌理论以及其它各种非线性检验方法对我国股票市场价格收益率的分布特征、相关性与持续性、周期性、异方差性、复杂性等特征进行了全面的实证分析,并且探讨了非线性方法对股票市场的风险衡量、交易制度对股价波动的影响、消息对股价影响的不对称性及股价的可预测性等股票实践问题,主要结论与成果有:(1)传统线性理论不能很好解释沪深股票市场,我国股票市场价格收益率行为不符合有效市场假定,收益率分布存在明显尖峰胖尾行为,收益间存在长期相关性与持续性。(2)我国股票市场价格收益率行为存在周期性现象,具有一定数量的非规则周期。(3)异方差模型在我国股票市场能得到很好的拟合,好坏消息对两指数收益率变化的冲击是不对称的,交易成本的上调对股票市场价格波动性有明显的影响,而下调则影响甚微。(4)如果把股票市场视为一个复杂性系统,分形与混沌等理论检验表明我国股票市场价格收益率具有明显非线性特征,最少可用4个状态变量建立沪深两市大盘指数价格序列系统模型。(5)两市股价波动具有明显的时变性、簇集性及共动性,风险与收益间关系不显著;用ARIMA模型对我国股票市场收益率波动预测效果一般。文章最后提出了相应的政策建议及进一步的研究方向。更多还原

【Abstract】 Many classical capital market theories, which are developed on the basis of Efficient Market Hypothesis, are established under the linear framework, for EMH is established on the following three core presuppositions: rational investor, efficient market and random walk process. Nevertheless, Many empirical studies, home and abroad, show the limitation of EMH and many phenomenona in the stock market can not be reasonably explained by classical capital Market theories. So it’s an inevitable selection to study capital market in nonlinear method instead. The study on the price & return rate of China’s stock market under non-linear framework is very meaningful to enrich the knowing of stock price and return rate behavior, to standardize and supervise the stock market, to measure price for the near future derivative securities.As concerned above, under the non-linear framework, the dissertation gives an empirical analysis to the distribution, correlation and persistence, periodicity, heteroskedasticity, complexity of China’s stock market by some non-linear model, such as R/S analysis, frequency spectrum, ARMA model, ARCH model, and factor analysis, fractal and chaos theories and other non-linear methods. Also some practical concerns such as risk measurement, the influence to the price volatility of the trade institute, the unbalance of price shock under new information and the price predictability, have been discussed using non-linear methods. The main results are listed as follows:(1)China’s stock Market can not be reasonably explained by classical capital theories. The behavior of stock price & return rate, which shows long memory and persistence and thin peak and heavy-tailed feature in the distribution picture, is not conform to EMH.(2)There are some periodical phenomenona, with some unregular periods, in the price & return rate behavior of China’s market.(3)The volatility of China’s stock market can been well estimated by AHCH model. The empirical study by AHCH model series shows that: the shock to the price & return rate is unsymmetrical between good and bad news, the volatility of China’s stock market can been great disturbed by the increase of trade cost ,whereas only little influence by the decrease of trade cost.(4)Regarding stock market as a complicated system and examining through fractal and chaos theories , there are distinct non-linear features in the price & return rate of China’s stock market. The system of Shanghai and Shenzhen index series can be established by at least 4 variables.(5)The volatility in China’s stock market varies timely, jointly and assembly. The relation between risk and return rate is not distinct and the performance is not very good in the prediction of return rate volatility by ARIMA model.Some policy suggestions and the further research trend have been pointed out in the last chapter of the dissertation.更多还原