指数基金与指数的相依性研究

发布时间：2018-08-31 09:59

【摘要】：本文主要探讨了我国指数基金与指数的相依结构,相依性分析是金融理论和实践的一个基础性问题,然而传统的线性相依衡量有其局限性,因此引入Copula模型来测度变量之间非线性和尾部相依结构。指数基金的投资采用被动管理策略,通过分散投资标的指数的成份股,力求股票组合的收益率与目标指数的平均收益率保持一致。然而实际情况是指数基金的走势与标的指数并不一定同步,追踪同一标的指数的指数基金在走势方面也存在的偏差,指数基金跟踪指数的效果千差万别。因此本文拟通过研究指数基金与指数的相依性来评价指数基金追踪指数的效果优劣,这有助于基金管理者构建投资组合更加有效、监管者在市场制度建设更加合理以及对基金投资者在选择基金投资种类方面提供指导作用。本文首先总结了国内外在Copula理论与应用取得的成果,说明了Copula在衡量变量相依性存在的优势；其次在概述Copula理论的基础上,详细给出了Copula建模的一般步骤。静态相依结构的衡量利用了阿基米德族Gumbel、Clayton、 Frank和这三种Copula的线性混合模型,在此基础上使用滚动窗口的方法,利用AR(2)-GARCH (1,1)模型对参数实现动态演化拟合,从而将静态衡量模型拓展到动态衡量。最后对追踪沪深300指数的LOF基金嘉实300、银华300与鹏华300指数基金与指数的相依结构进行实证研究,通过Copula模型从静态与动态两方面深入研究变量之间的相依程度和相依结构。通过实证研究可得出各指数基金与指数的相依结构呈现不同的特征,利用Copula模型在衡量变量之间相依性上确实存在不少优势,而且本文从静态与动态两方面的建模对于深入分析指数基金追踪指数的效果具有全面性,也使得解释性更强,可信度更高,使用相依性分析也为选择指数基金提供了新的视角。
[Abstract]:This paper mainly discusses the dependence structure of index fund and index in China. Dependency analysis is a basic problem in financial theory and practice. However, the traditional linear dependency measurement has its limitations. Therefore, the Copula model is introduced to measure the nonlinear and tail dependent structures between variables. The investment of index fund adopts passive management strategy, and the return rate of the stock portfolio is consistent with the average return of the target index by diversifying the component stocks of the underlying index. However, the actual situation is that the trend of the index fund is not necessarily synchronized with the target index, and there is also a deviation in tracking the trend of the index fund with the same target, and the effect of the index tracking index is very different. Therefore, this paper intends to evaluate the effectiveness of index tracking index by studying the dependence of index fund and index, which is helpful for fund managers to build a portfolio more effective. Regulators are more reasonable in the construction of market system and provide guidance to fund investors in the choice of fund investment types. This paper first summarizes the achievements in Copula theory and application at home and abroad, and explains the advantages of Copula in measuring the dependence of variables. Secondly, on the basis of summarizing the Copula theory, the general steps of Copula modeling are given in detail. The static dependent structure is measured by using the linear mixed model of the Archimedes family Gumbel,Clayton, Frank and the three kinds of Copula. On the basis of this, the dynamic evolution fitting of the parameters is realized by using the AR (2) -GARCH (1K1) model using the method of rolling window. Thus, the static measurement model is extended to dynamic measurement. Finally, this paper makes an empirical study on the dependence structure of the LOF funds, Jiashong 300, Yinhua 300 and Penghua 300 index funds, which track the CSI 300 index, and studies the dependence degree and structure of variables from static and dynamic aspects through the Copula model. Through the empirical study, we can find that the dependent structure of each index fund and index presents different characteristics, and there are many advantages in measuring the dependence of variables by using Copula model. Moreover, the static and dynamic modeling in this paper has a comprehensive effect on the in-depth analysis of the index tracking index of index funds, and makes the explanation stronger and the reliability higher. The use of dependency analysis also provides a new perspective for the selection of index funds.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.51;F224

【参考文献】