金融危机传染的非线性相似模型

Empirical analysis of financial contagion based on nonlinear similarity

采用相空间重构和改进符号的非线性动力学相似性模型,研究了2007—2012年全球金融危机前后美国与英国、法国、德国、日本、中国金融股指的相似性.计算数据表明,各国的市场指数都在不同时段和不同程度对美国指数表现出动力学相似性,即经济繁荣阶段,美国经济影响其他国家;金融危机时期,危机也确实从美国传染到了他国市场.模型计算的走势及反应的现象证明使用改进符号的非线性动力学相似性模型适用于金融市场,是一种研究金融危机传染的可行新方法.

We modify the dynamical similarity method of nonlinear time series, and apply C-C method of phase space reconstruction and the sign-modified dynamical similarity model to estimate the similarity index between the financial markets of US and Britain, France, Germany, Japan and China respectively. The results show that nonlinear similarity between each country＇s stock index and that of US exists in different time and different degrees, which means US economy influences other countries in periods of economic prosperity, and also infects these countries in financial crises. The trends of model calculations and the phenomenon they present indicate that the sign-modified nonlinear dynamical similarity model is suitable for the financial market, and it is a practicable new method to analysis financial contagion.