我国上市银行间相关性及风险溢出研究

发布时间：2018-11-01 17:29

【摘要】：随着金融全球化及金融创新步伐不断加快，金融机构间的联系逐渐变得更加紧密。在我国，银行业是金融系统的核心。银行与银行之间通过业务往来等具有相互关联性的行为联系变得更加紧密。如何刻画银行间的相关性，以及如何测度银行间的风险溢出强度，是一个不容忽视的问题。 Copula理论在分析变量间的相关结构时具有很多优点，可以较好地刻画变量间非线性、非对称的尾部相关关系，同时变结构Copula可以精确找到变量间相关结构的变点，为研究风险溢出提供依据。本文选取我国上市银行中十家银行作为样本，实证分析了上市银行的相关性特别是时变相关性和风险溢出效应。在实证研究过程中，首先，金融时间序列普遍存在尖峰、厚尾等现象，研究结果显示采用GARCH(1,1)-t模型对拟合我国上市银行间收益率序列的边缘分布是合适的。其次，基于常相关Copula模型和时变相关Copula模型对我国上市银行间的相关性进行了详细的研究，结果表明我国银行间常相关性很强，而银行间的相关系数是不断变化的，而且围绕着某一固定值上下波动，走势非常相似。再次，在时变相关的研究基础上，利用Z检验法对上市银行间的相关结构进行变点检验，结果表明在2008年9月18日，大多数银行间的相关结构发生了变化。因此，本文以2008年9月18日作为分水岭，基于CoVaR结合分位数回归技术研究美国金融危机前后我国国有银行对股份制商业银行的风险溢出强度的变化。研究结果表明：在q=0.05的情况下，危机后，我国国有银行对大部分股份制商业银行的风险溢出强度增大，特别是对民生银行，，风险溢出强度达40%以上。当前，监管当局在确定系统重要性银行时提出不仅要考虑银行的规模因素，更要考虑由于银行间的相互关联性，考虑单个银行陷入困境对其他银行的风险溢出及风险溢出强度。本文对我国上市银行间的相关性研究及量化风险溢出强度首先有助于金融监管部门及时捕获银行间的风险强度，进而对高风险溢出机构进行监测和管理，从而维护金融市场的稳定；其次为微观金融主体进行投资分析和投资组合提供依据。
[Abstract]:With the acceleration of financial globalization and financial innovation, the relationship between financial institutions has gradually become closer. In China, banking is the core of the financial system. Banks and banks become more closely connected through interrelated behaviors such as business transactions. How to depict the correlation between banks and how to measure the risk spillovers between banks is a problem that can not be ignored. Copula theory has many advantages in analyzing the correlation structure between variables, which can describe the nonlinear and asymmetric tail correlation between variables. At the same time, the variable structure Copula can accurately find the variation point of the correlation structure between variables. To provide the basis for the study of risk spillover. In this paper, ten banks of listed banks in China are selected as samples, and the correlation of listed banks, especially the time-varying correlation and risk spillover effect, is analyzed empirically. In the process of empirical research, first of all, financial time series generally exist peak and thick tail phenomena. The results show that the GARCH (1 ~ 1) -t model is suitable to fit the marginal distribution of the return series between listed banks in China. Secondly, based on the regular correlation Copula model and the time-varying correlation Copula model, the correlation among the listed banks in China is studied in detail. The results show that the correlation coefficient between the banks in our country is very strong, and the correlation coefficient between banks is constantly changing. And around a fixed value up and down, the trend is very similar. Thirdly, based on the research of time-varying correlation, we use Z test method to test the correlation structure of listed banks. The results show that most of the interbank correlation structures have changed on September 18, 2008. Therefore, this paper takes September 18, 2008 as the watershed, based on CoVaR and quantile regression technology, to study the change of risk spillover intensity between Chinese state-owned banks and joint-stock commercial banks before and after the American financial crisis. The results show that after the crisis, the risk spillover intensity of state-owned banks to most joint-stock commercial banks increases, especially to Minsheng Bank, and the risk spillover intensity is more than 40%. At present, when determining systemically important banks, regulators should consider not only the scale factors of banks, but also the risk spillover and risk spillover intensity of a single bank into a dilemma to other banks due to the interrelationship between banks. This paper studies the correlation between listed banks in China and quantifies the risk spillover intensity. Firstly, the financial supervision department can catch the risk intensity among banks in time, and then monitor and manage the high risk spillover institutions. In order to maintain the stability of the financial market; Secondly, it provides the basis for investment analysis and portfolio analysis.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F832.3;F224

【参考文献】