摘要
随着经济全球化的发展,金融市场间的风险传染问题越来越受到学术界的关注。本文利用高斯混合自回归(MAR)模型估计收益率的边际分布,其独立性等检验表明,MAR模型能很好地描述"尖峰厚尾"和"有偏"的现象。通过利用权重时变的混合Copula描述收益率间的动态非线性和非对称的动态相关性,本文对中国股市与国际主要股市的相关性进行实证研究,结果表明,无论是从下尾相关性还是上尾相关性来看,中国股市与美、英、日主要股市之间的相关性均还处于较低水平,与香港股市的相关程度相对较高。而当异常事件发生时,美、英、日和香港股市对中国股市的风险传染效应则十分显著。
With the development of economic globalization, scholars have paid more and more attention to the problem of risk infection between financial markets. By using Gaussian Mixture Autoregressive model （ MAR ） , this paper estimates the marginal distribution of yields. Test on independence indicates that MAR can befit well to describe the hiasness and the phenmnena of ＂sharp peak and thick tail＂ . In the meantime, the paper also describes nonlinear dynamic charaeleristic and dynamic asymmetric dependence between different yields in the use of mixed Copula weighted by time-varying. The results of empirical tests between Chinese stock market and that of international main markets show that, no matter it is the lower-tail dependence or upper -tail dependence, China is little dependent on the U.S., Britain and Japan. However, a strong dependence appears between the China's Mainland and Hong Kong＇ s stock market. When exceptional events take place, risk infection effects on China's Mainland which are conducted by all the above four places are significant.
出处
《南方金融》
北大核心
2013年第10期80-84,共5页
South China Finance
基金
浙江工商大学研究生创新基金(项目编号:1060XJ1512086)的资助
