基于变结构Copula模型的股票市场间波动溢出效应研究

发布时间：2018-01-14 16:39

本文关键词：基于变结构Copula模型的股票市场间波动溢出效应研究　出处：《东北大学》2013年硕士论文　论文类型：学位论文

【摘要】：股票市场是一个复杂的动态系统,经济全球化、金融自由化加剧了股票市场的复杂性和波动性,市场之间的波动相关性显著增强。波动溢出效应是指不同金融市场之间的波动可能存在相互制约,一个地区市场的巨大震荡有可能会传递到其它地区的市场。因此,为了提高金融决策的准确性,降低决策风险,对股市间的波动溢出效应进行研究分析是非常必要的。Copula函数能够描述多个随机变量间的相依结构,是研究不同市场间波动溢出效应的有效工具。本文通过实证的方法,利用Bayes法诊断出各变结构点,利用Copula函数得到不同阶段的秩相关系数,检验看其变化的显著性,进而辨析不同市场间波动溢出效应的存在性。论文的研究主要包括(1)对股票市场收益率残差序列进行ARCH效应检验。结果表明,在5%的显著性水平下,所选取的股票市场时间序列存在显著的ARCH效应,能够利用GARCH类模型来建模。(2)利用GARCH(1,1)-t模型描述各收益率时序的边缘分布。首先估计得到收益率序列的边际分布模型参数,然后进行概率积分变换,运用K-S检验方法检验变换后的序列是否服从(0,1)分布。结果表明,上证综指存在明显的尖峰厚尾现象,并且与其它地区相比,其收益率更有可能出现极端值。(3)利用静态、动态Copula函数对不同地区股市收益率序列的相关性进行分析。静态的结果表明,中亚股市相关性最强,中欧股市相关性次强,中美股市相关性最小。动态的结果表明,各地区与上海股市收益率时间序列的相关程度都存在程度不一的上升趋势。(4)利用“三步法”研究股市间的波动溢出效应。结果表明：深圳股市与上海股市收益率时序间的波动溢出效应现象不但表现得最为明显,而且发生得也最为频繁；其它国家与我国市场在不同时间发生波动溢出效应。(5)利用变结构Copula模型研究不同市场的波动溢出效应的研究结果符合股市的现实中的表现,从而证实了此方法的合理性。
[Abstract]:The stock market is a complex dynamic system, economic globalization, financial liberalization exacerbated the complexity of stock market volatility and volatility correlation between markets increased significantly. The volatility spillover effect refers to different financial market volatility may restrict each other, a huge shock area market is likely to transfer to other areas the market. Therefore, in order to improve the accuracy of financial decision-making, reduce the risk of decision-making, research and Analysis on the volatility spillover effect between stock markets is very necessary.Copula function can describe the dependence structure between multiple random variables, is a useful tool to study the volatility spillover effect of different markets. Through empirical methods, diagnosis the variable structure by Bayes method, using the Copula function to obtain the rank correlation coefficient in different stages, the significant change, and the analysis of different city The existence of field volatility spillover effect. The research includes (1) on the stock market return residual sequence by ARCH effect test. The results show that in the 5% significant level, the stock market time series has significant ARCH effect, can use GARCH model to model (2). The use of GARCH (1,1) -t model to describe the marginal distribution of each yield time series. Firstly estimates the marginal distribution model parameters yield sequence, then the probability integral transform sequences by the test of K-S transform (0,1) is subject to the distribution. The results show that the Shanghai Composite Index has obvious leptokurtic phenomenon, and compared with other regions, the rate of return is more likely to occur in extreme value. (3) the use of static, dynamic correlation Copula function on income rate series in different regions of the stock market. The results show that the static analysis, the Central Asian stock market The strongest correlation, strong correlation between the stock market of China EU, Sino US stock market. The results show that the dynamic minimum correlation, correlation degree of each region and the Shanghai stock market return series are rising in different degree. (4) the "volatility spillover effect between the stock market on the three step. The results show that the Shenzhen stock market and Shanghai stock market return the phenomenon of volatility spillover effect between time series is not only the most obvious, but also occurs most frequently in other countries; and China's market volatility spillover effect in different time. (5) study on the research model using variable structure Copula spillover effects in different markets. The results accord with the reality of the performance of the stock market thus, confirm the rationality of the method.

【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F832.51

【相似文献】