【作者】 王欣；

【导师】 方兆本； 赵定涛；

【作者基本信息】 中国科学技术大学 ， 管理科学与工程， 2009， 博士

【摘要】 随着金融市场对规避系统性风险的需求日益强烈,股指期货合约这一金融衍生产品应运而生。从其诞生以后二十多年的发展看,可以说取得了巨大的成功。2006年9月中国金融期货交易所挂牌成立,中国金融衍生品的发展在经历了长时间的停滞后,终于正式拉开了序幕。做为中金所成立后即将推出的第一个金融衍生品——股指期货合约自然成为国内理论与实务界关注与研究的焦点,近年来在这一领域的研究成果层出不穷。股指期货所具有的卖空机制改变了投资者仅能在股价的上涨中获取盈利的单边盈利模式,所以对组合优化策略的研究是目前研究的热点之一,因为有效的组合优化可以使投资者即使在利用股指期货进行套期保值时,也能够获得一定的低风险收益。其次,随着未来股指期货合约品种的日益丰富,投资者究竟应当如何度量各种不同合约的风险,选择适合套期保值操作的合约类型同样是非常值得研究的方向。最后,套期保值策略的最优化问题,也是套期保值研究中最核心的内容,这一课题的研究可以对提高套期保值效果,为套期保值在投资实务中的应用提供理论上的支撑。本文以股指期货在投资组合管理中的套期保值研究为切入点,分别研究了上面提到的组合优化,合约风险评价以及套期保值策略最优化等几个层面的问题。本文共分六章,行文的具体安排及主要创新点如下:第一章对股指期货市场的发展和研究概况进行了系统的回顾。第二章在无指导学习的研究框架下,运用分位数回归模型结合变点检验,通过分析持股比例变动与股价收益率间协同演化关系的异常,对存在异常交易行为的股票进行甄别,进而在构建股票池和调整投资组合时及时发现并剔除价格异常波动的股票,有效规避可能存在的投资风险,实现组合优化。第三章首先运用决策树和Logistic回归将基本面因素引入股价指数成份股的选择和权重的确定,编制引导基本面投资理念的基本面股价成份指数(JOYFI300),之后基于这一指数构建相应的指数化组合,论证了该指数化组合在超额收益方面表现出良好特性。具体来说,我们首先通过决策树引入净利润和成交金额变量作为成份股的选样标准,选择成份股初始样本;然后,建立Logistic模型分析基本面变量对成份股权重的影响,通过数据分析客观确定指数成份股及权重:最后实证检验JOYFI300指数组合各项特性,与沪深300、中信标普300等其他指数组合的比较表明JOYFI300组合具有更好的超额收益。由于目前度量股指期货合约套期保值效率的模型大多仅仅考虑了现货价格的风险减少,因此在本文第四章中首次提出了一种度量股指期货合约整体风险变化的合约选择与评价模型。运用这一方法使投资者或交易所可以根据期货合约推出后基差风险、流动性风险和交易操纵风险的变化,对套期保值者的风险暴露状况作出评价与比较,并为下一步的投资和风险管理提供决策依据。在第五章中,首先运用极大交迭离散小波变换对新加坡新华富时A50股指期货合约原始数据进行逐尺度分解,在不同时间尺度下以半方差最小化为套期保值目标对最优套期保值率进行估计,并与最小小波方差套期保值率进行比较。实证结果表明随着时间刻度的增加,期现货收益率间的相关性及套期保值率均相应递增;以半方差作为套期保值目标可以使套期保值组合获得更好的超额收益性质,并且随着套期保值期限长度的增加,超额收益性质的相对表现更为优良。其次,为提高估计的精度,对最小半方差模型提出了进一步改进,首次引入小波密度估计的方法拟合期现货收益率的联合密度函数,并与经验和正态核密度估计方法进行了比较。实证分析的结果表明,小波密度估计对于存在非线性相关结构的数据具有更好的拟合效果;基于小波密度估计的最小半方差模型较其他方法能够更有效地消除套期保值组合的下侧风险,提高套期保值有效性。本文第六章对全文进行了总结,并针对本文的研究范畴与不足提出了一些未来值得进一步研究与改进的方向。更多还原

【Abstract】 The financial derivative-Stock Inedx Futures Contract emerged with the increasing demand to avoid systematic risk in financial market.In view of the development during the past twenty years or more from its birth,the product is proved to be a tremendous success.After a long-time stagnation in the growth of Chinese financial derivatives,the curtain rises at the time when the Chinese Financial Futures Exchange was established in September,2006.The theory and practice circles pay much attention to Stock Index Futures Contract which is the first financial derivative authorized after the establishment of the Exchange.Furthermore, the study in this field has come to bear fruits abundantly in recent years.The short mechanism possessed by Stock Index Futures Contract changes the unilateral profit mode in the case that investors can only profit when stock price goes up.Consequently,the research on portfolio optimization strategy is one of the central issues.The effective portfolio optimization can help the investors to gain low risk earnings even if they hedge with Stock Index Futures Contract.Secondly,confronted with a large variety of Stock Index Futures Contracts in the coming years,the investors should learn how to measure different contractual risks and then hedge with proper contract,so the research in this regard is quite worthwhile.Finally,the study on the core of hedge research-i.e,the optimum hedge strategy can provide theoretic support in improvement of hedge efficiency and its application in investment practice.Taking the hedge research of Stock Index Futures in investment portfolio management as the entry,we make some study on portfolio optimization,contractual risk evaluation and optimum hedge strategy respectively.The frame and main innovations of this dissertation are described as follows:In Chapter One,the research and development of Stock Index Futures market are summerized.In Chapter Two,based on Unsupervised Learning Theory,in order to avoid potential investment risk and realize portfolio optimization,with Quantile Regression Model and Change Point Test,we detect outlier transaction behavior,and then recognize and eliminate timely abnormal fluctuation of stock price when we construct stock pool and adjust investment portfolio by analyzing outlier in coherently evolutionary relationship between changes of shareholders’ owner-ratio and stock price return.In Chapter Three,we select constituent stocks and determine their weights with fundamental factors with Decision Trees and Logistic Regression.Fundamental Stock Price Constituent Index(JOYFI300) is created as a guidance to fundamental investment.The indexing portfolio based on this index is testified better on excess return.To Speak in details,firstly,we choose net profit and trading value as two selection measures by using Decision Trees,and then initial samples of constituent stocks can be found out by these two factors.Secondly,the influence of fundamental factors on weights of constituent stocks can be analysed and determined by Logistic Model.Finally,after test and comparison,we find that JOYFI300 Indexing Portfolio is better on excess return than Shanghai-Shenzhen 300 and CITIC/S&P 300 Indexing Portfolio.In Chapter Four,a Contract Selection and Assessment Model is used for the first time to measure the change of overall risk of Stock Index Futures Contract, because most of the conventional models for valuating hedging efficiency of Stock Index Futures Contract are based only on reduction of cash price risk.The investors or exchange can assess risk exposure of hedger according to the change of basis risk, liquidity risk and trading manipulation risk after introducing Stock Index Futures Contract by using this model,and it also intends to provide a reference of decision-making for risk management.In Chapter Five,we firstly decompose original data involved in Singapore Xinhua/FTSE A50 Stock Index Futures Contract on scale-by-scale basis with Maximum Overlap Discrete Wavelet Transformation.Optimal hedge ratio is estimated under different time scales by taking minimum semi-variance as hedge target.In comparison with minimum wavelet variance hedge ratio under each scale, the empirical result indicates that hedge ratio and correlation of the rate of return between futures and spot go higher along with time scale.Taking semi-variance as hedge target can lead to a better excess return on hedge portfolio.The longer the length of time horizon is,the more excellently the excess return performs.Secondly, we improve Minimum Semi-variance Model with fitting the joint probability density function of rate of return between futures and spot prices based on wavelet density estimation method for improving goodness of fit for the first time.In comparison with empirical and kernel function density estimation method,the empirical result indicates that fitted results of wavelet density estimation is better for data with nonlinear correlation structure.Minimum Semi-variance Model based on wavelet density estimation can eliminate downside risk of hedge portfolio more effectively than other methods.This approach also can lead to a better excess return on hedge portfolio than Minimum Variance Model.In Chapter Six,we summarize this dissertation and point out the directions for further research and improvement in the future in terms of the research content and defect.更多还原