【作者】 江洲；

【导师】 谢赤；

【作者基本信息】 湖南大学 ， 管理科学与工程， 2008， 博士

【摘要】 证券市场收益率是金融经济学中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。一方面,证券市场作为一个复杂系统,其内部因素有着一定的内在统计运行规律,从而其收益率具有一定的分布特征,因此对证券市场收益率的描述和预测必须考虑其自身的运行特征。另一方面,证券市场的运行势必受到外来信息以及因素的冲击,其中包含多个变量的宏观经济因素是对证券市场最直接、最有影响的外在冲击来源,所以在对证券收益率进行描述与预测的过程中,必须综合考虑证券市场内在运行规律以及外在宏观经济因素的作用机制。本文的研究目的是对证券市场收益率的描述与预测进行新的探索,综合证券市场发展的内因和外因,即全面考虑证券市场收益率本身的分布特征和宏观经济因素的影响,探索一种更有效率的组合预测模型,以便更精确地预测证券市场的走势和理解证券市场的内在运行机理。本文首先对沪深证券市场进行有效性实证检验,发现中国证券市场已经达到弱式有效,但是还远没实现半强有效。在中国证券市场上,内幕交易者通过掌握的内部信息能够获取超额利润,但是中小投资者在市场上并没有得到应有的回报。为了真实地反映中国证券市场的运行状况,本文分别从不同的角度,采用不同的方法对沪深证券市场收益率分布特征进行研究。发现上海和深圳两个证券市场的关联程度非常大;中国证券市场中的总体日收益率分布不服从正态分布,而是呈“尖峰态”,并且具有“厚尾”的特征;证券市场收益序列呈现稳态特征与多标度分形特性,表现为长期记忆特性、厚尾特性、标度特性和易变性。证券市场收益率长期记忆性的存在意味着利用更多的历史信息可以显著提高其预测的效果。本文对中国证券市场收益率的长期记忆特征及其成因做了深入的研究,同时分别选择上证综合指数与深圳成份指数代表沪深两个市场的总体变动趋势,运用多种方法刻画了沪深证券市场收益率的长短记忆性,在此基础上确定了刻画其长短记忆性的合适模型。证券市场是经济发展的产物,因而一国证券市场的发展与该国经济发展紧密相关。本文从理论和实证两方面研究了宏观经济对证券市场收益率的影响,发现上海、深圳证券市场收益率均与宏观经济变量之间存在协整关系,说明中国证券市场与宏观经济的发展是基本一致的;上海、深圳证券市场收益率的短期波动的变化也受到工业增加值、货币供给量、通货膨胀率、利率、储蓄短期变化的影响。政策是影响证券市场价格、回报及其波动性的重要因素。本文分析了证券市场政策市的含义及其形成原因,并利用政府相机治理模型对政策市进行了的解释,在此基础上实证分析了中国重大政策事件的证券市场效应。最后,本文考虑证券市场收益率的分布特征与长记忆性,利用GARCH族模型预测了中国证券市场的月度收益率;考虑宏观经济变量与政策对证券市场收益率的影响,利用VAR模型预测了中国证券市场的月度收益率;在此基础上综合考虑证券市场收益率本身的分布特征和宏观经济因素的影响,利用VAR-GARCH族模型对证券市场收益率进行了描述与预测,并比较了各模型的预测效果,发现VARGARCH族模型的效果最好。更多还原

【Abstract】 The security market return is an important conception in the area of the financial economics, and the accuracy of the description of the return volatility generally affects the correctness of the security portfolio, the validity of the risk management and reasonability of the option pricing. For one side, as a complex system, security market has some instinct running pattern, thus the return follows certain distribution characteristic, consequently, it is essential to consider the instinct running pattern to describe and predict the return of the security market. For the other side, the running of the security market is undoubtedly shocked by the external information and other factors. Among these factors, the macroeconomics which contains series of variables is the most direct and influencing factor, therefore, in the process of the describing and predicting the security return, it is necessary to consider the instinct running rules and the mechanism of the external macroeconomic factors.The thesis is to explore the new ways to describe and forecast the security return combining the internal factors and the external factors of the security market development. By integrating effects of the distribution characteristics of the security market return and the macroeconomic variables, the thesis explores a more efficient forecasting combination model to predict the trend of the security market and the instinct mechanism of the security market.The thesis firstly empirically examines the effectiveness of the Chinese security market, and finds the security market has achieve the weak form efficiency, but there is still a long way to reach the semi-strong form efficiency. In Chinese security market, inside investors often get abnormal return by using the inside information, while medium and small investors doesn’t acquire the deserved return in the market.To authentically reflect the running pattern of the Chinese security market, the thesis adopts different methods to analyze the distribution characteristics of the security market in Shanghai Stock Exchange and Shenzhen Exchange. The results show three outstanding characters: these two stock markets are closely correlated; the overall day return doesn’t follow the normal distribution, but it is characterized with leptokurtic and fat tail; the security market return has stable characteristics and multi-fractal characteristics which shown as the long memory, fat tail, scale properties, and variability. The long memory of the security market return means that the predicting ability can be enhanced by adopting more history information. The thesis deeply studies the long memory characteristic and its causes of the security market in China. Meanwhile, choosing composite index of Shanghai stock exchange and component index of Shenzhen stock exchange to represent the overall change trend of the markets, the long memory characteristic of the security market returns is described by using different method, and the appropriate models are found by the empirical study.Security market is the outcome of the development of the economics, thus the status of the security market is closely connected with the economic development of the country. This thesis theoretically and empirically studies the effects of macroeconomic on the security market return, and the results indicate that there exists a long run equilibrium relationship between security market return and various macroeconomic variables. The development of security market is consistent with the development of macro economy in principle, and the short-run fluctuation of security market return is affected by added industry value, money supply, inflation rate, interest rate, the short run change of the saving and so on.The government policy is the important factor that affects the security price, security return and the volatility. The thesis analyzes the implication of China’s Policy Market and the cause of its formation, explains the China’s Policy Market by taking the contingent governance model, and empirically detects the market effect of the important polices.Lastly, considering the distribution characteristics and the long memory characteristics, the monthly return of the Chinese security market is predicted by using GARCH models. Similarly, considering the effects of the macroeconomics on the return of the security market, the monthly return of the Chinese security market return is predicted by taking VAR models. Meanwhile, integrating the distribution of the security market return and the effects of the macroeconomics, VAR-GARCH models is build to prescribe and predict the security market return. By comparing the predicting effects, the VAR-GARCH models are superior to the other models.更多还原