基于Copula理论对中国投资者情绪的研究

发布时间：2018-01-01 02:21

本文关键词：基于Copula理论对中国投资者情绪的研究　出处：《复旦大学》2013年硕士论文　论文类型：学位论文

【摘要】：投资者情绪作为行为金融学的支柱之一,近年来已经成为研究的热点问题。本文将从Copula理论的视角对投资者情绪进行探讨,通过Copula函数来重点研究中国A股市场和B股市场投资者情绪与股市收益的相依结构形式以及对比两者之间的异同,之后对样本按熊牛市分为两段,分析这两对相关结构在熊牛市表现是否有所不同。在此基础之上通过选取同时在A股和H股上市的公司作为研究对象,进一步分析了境内外投资者的投资者情绪的差异,得到了较多有益的结论。 Copula函数可以很好的反映变量之间的非对称性、时变相关性、尾部相关性等的特征,较之Granger检验和线性相关系数能反映更多的相关结构的信息。同时,Copula函数的特点使得可以分步进行变量边缘分布的估计和相关结构的估计,不需限制变量的边缘分布形式以及Copula函数的形式,使得Copula函数在相关性分析中有着独到的优势,应用范围广泛。本文在第一部分实证分析中选用A、B股新开户率作为投资者情绪的指标研究新开户率和股市收益的相关结构,边缘分布通过ARMA-GARCH模型进行建模,新息项(innovation)的分布选用更能反映不对称及厚尾信息的Hansen's Skewed-T分布。之后运用9种无条件Copula函数和3种时变Copula函数进行相关结构的建模分析。在后两部分的实证中边缘分布不需要进行时间序列处理,故采用非参数核密度(Kernel)估计进行建模,最后一部分的实证分析中,选用同时在A股和H股上市的公司年报公布的EPS与分析是一致预期的差值作为其中一个变量,选用这些公司在A股和H股公告日附近的累积异常收益率(CAR)作为另一个变量来代表两地投资者的投资者情绪反映。文章最后阐述了本文的主要结论和贡献。
[Abstract]:Investor sentiment as one of the pillars of behavioral finance has become a hot issue in recent years. This paper will discuss investor sentiment from the perspective of Copula theory. Through the Copula function to focus on the Chinese A-share market and B-share market investor sentiment and stock market returns of the dependent structure and the contrast between the similarities and differences between the two, and then the sample is divided into two sections according to the bear bull market. On the basis of analyzing whether the performance of the two pairs of related structures in the bear bull market is different, the companies listed in A-shares and H-shares are selected as the research objects. The difference of investor sentiment between domestic and foreign investors is further analyzed, and some useful conclusions are obtained. Copula function can well reflect the characteristics of asymmetry, time-varying correlation, tail correlation and so on. Compared with Granger test and linear correlation coefficient, it can reflect more information of related structure. The characteristics of Copula function make it possible to estimate the marginal distribution of variables and related structures step by step without restricting the form of marginal distribution of variables and the form of Copula function. The Copula function has unique advantages in correlation analysis and has a wide range of applications. In the first part of the empirical analysis, we choose A, B share new account opening rate as an indicator of investor sentiment to study the structure of new account opening rate and stock market income. The edge distribution is modeled by ARMA-GARCH model. Innovation). Hansen's, which can better reflect asymmetric and thick tail information. Skewed-T distribution. Then, nine kinds of unconditional Copula functions and three kinds of time-varying Copula functions are used to model and analyze the correlation structure. In the last two parts of the empirical analysis, the edge distribution does not need to advance. Time series processing. Therefore, the nonparametric Kernel estimation is used to model the model, and the last part of the empirical analysis. The EPS published in the annual report of A shares and H shares listed at the same time is consistent with the expected difference as one of the variables. Select these companies in A shares and H shares near the announcement of the cumulative abnormal rate of return carr). As another variable to represent the investor sentiment of the two places, the paper finally expounds the main conclusions and contributions of this paper.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F832.51

【相似文献】