基于多元GARCH模型的流动性溢出效应研究

发布时间：2018-01-09 12:35

本文关键词：基于多元GARCH模型的流动性溢出效应研究　出处：《浙江工商大学》2012年硕士论文　论文类型：学位论文

【摘要】：随着全球经济一体化进程加快,世界经济可谓风起云涌变幻莫测。我国继改革开放之后,走出了一个不平凡的三十年,九十年代之后开始走市场化道路,国内生产总值更是以每年两位数的发展速度高速增长,令世界各国叹为观止。然而在大跨步向前进的过程中,道路是坎坷的,1998年的亚洲经济危机、2008年美国次信贷危机所引爆的全球金融危机,都深刻的对中国这个新兴的经济体给予了极大的考验。与此同时,尤其是08年金融危机之后,流动性风险在各国之间的传递效应暴露在中国及世界各国面前,在毫无防备情况下,流动性的缺失使世界各国经济像多米诺骨牌一样纷纷倒下。在此背景基础上,本文提出针对以中国、日本、美国和英国证券市场之间的流动性溢出效应为研究对象。我们分别考虑了流动性度量指标、样本数据选取以及一定的模型选择,从流动性水平和流动性风险溢出效应两个角度,来度量他们之间的流动性生溢出效应。流动性水平溢出效应的刻画,我们建立VAR向量自回归模型,并在此基础之上应用格兰杰因果关系检验、脉冲响应分析、方差分解分析了变量之间的溢出效应。流动性风险溢出效应的刻画,我们从流动性指标二阶矩的估计上,最终选择多元GARCH模型中具有代表性的BEKK. DCC-GARCH.GO-GARCH三个模型,在高维情况下进行了流动性风险的实证分析,从横向和纵向对危机前后以及各个模型之间得出的结论进行了分析比较,一方面应用多元GARCH模型分析流动性风险溢出效应,另一方面将多元GARCH的三类模型进行对比分析比较,以期进步推动存在维数灾难情况下多变量的研究。经实证研究之后得到以下结论,危机前中美证券市场流动性互为格兰杰因果关系,中国证券市场为日本证券市场的单向格兰杰英国关系；危机后,D工L_FS100、DIL_RJ225均为DIL_HS300的格兰杰原因,而此时其它市场之间流动性并没不存在显著的因果关系。各市场对来自他们自身一个标准差新息响应的时候,总是可以很快平复到零,对来自其它变量一个标准差新息的响应的时候,他们都是以零为中心上下震荡,并且在经过十天之后回复到零。不管是危机前还是危机后,各国证券市场中流动性波动的贡献因素主要为自己市场本身。波动风险方面,危机后有了更高程度的波动持续性,且受各国在应对危机上经济政策的一致性,各国之间的流动性风险溢出效应都表现为一定程度的正相关,其中SP500与FS100正相关程度最高,原则上流动性的系统性风险仍然存在。
[Abstract]:With the acceleration of the process of global economic integration, the world economy can be described as ups and downs of unpredictable. After the reform and opening up, China has been out of an extraordinary 30 years, after 90s began to take the road to marketization. Gross domestic product (GDP) is growing at a double-digit rate every year, which is amazing to the world. However, the road is bumpy in the process of big leap forward, the Asian economic crisis in 1998. In 2008, the global financial crisis triggered by the US sub-credit crisis gave a profound test to China, the emerging economy. At the same time, especially after the 2008 financial crisis. The transmission effect of liquidity risk between countries is exposed in front of China and other countries all over the world. Under the condition of defenseless, the lack of liquidity causes the economies of all countries to fall like dominoes. On the basis of this background, this paper proposes to study the liquidity spillover effects between China, Japan, the United States and the United Kingdom stock markets. We consider the liquidity metrics respectively. Sample data selection and certain model selection, from the liquidity level and liquidity risk spillover effect, to measure the liquidity spillover effect between them, the characterization of liquidity level spillover effect. Based on the VAR vector autoregressive model, we apply Granger causality test and impulse response analysis. Variance decomposition analyzes the spillover effect between variables. The characterization of liquidity risk spillover effect is based on the estimation of the second moment of liquidity index. In the end, we choose the representative BEKK. DCC-GARCH.GO-GARCH three models in the multivariate GARCH model, and carry on the empirical analysis of liquidity risk in the case of high dimension. The conclusions before and after the crisis and among the models are analyzed and compared horizontally and vertically. On the one hand, multiple GARCH models are used to analyze the liquidity risk spillover effects. On the other hand, three kinds of multivariate GARCH models are compared and analyzed in order to promote the multivariate research in the presence of dimensionality disaster. After the empirical study, the following conclusions are drawn: before the crisis, the liquidity of China and America stock market is Granger causality, and the Chinese stock market is the one-way Granger British relationship of Japanese securities market. After the crisis, LFS100 / Dil RJ225 is the Granger cause of DIL_HS300. At this point, there is not a significant causal relationship between other markets. When markets respond to a standard deviation innovation from themselves, they can always flatten to zero quickly. When responding to a standard deviation innovation from other variables, they oscillate around zero and return to zero after ten days, either before or after the crisis. The main contribution factor of liquidity volatility in the securities market of various countries is their own market itself. In terms of volatility risk, there is a higher degree of volatility after the crisis, and the consistency of economic policies in responding to the crisis. The spillover effect of liquidity risk between countries is positive correlation to some extent, in which SP500 and FS100 have the highest positive correlation. In principle, systemic risk of liquidity still exists.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F831.7

【参考文献】