Analysis Of Multiple Structural Changes In Crisis Contagion Based On Nonlinear Dynamics

Since the1990s, the frequent financial crisis have attracted lots of researchers’attention. The financial crisis is a national crisis occurred first, and then spread toother countries through a variety of transmission path and usuallly has a hugenegative impact to the whole economic activities. The financial data during crisisoften show nonlinear deterministic features. The traditional theory of financial crisisis impossible to correctly analysis the nonlinear characteristics of crisis contagion.Therefore, we try to study the financial contagion resorting to nonlinear dynamicsystems theory and characterize nonlinear dynamics of the multiple financial timeseries. We try to provide reasonable suggestions for our country’s financial crisiswarning system and to maintain the stability of the financial system in the context ofeconomic integration and financial liberalization.First of all, this article summarized the theoretical and empirical research resultsof domestic and foreign scholars on crisis contagion and the application of nonlineardynamics in the financial system in recent years. Then we introduced the time seriesprediction theory based on nonlinear dynamic. On this basis, the Multiple StructuralChange model in crisis contagion based on nonlinear dynamics was established. Thisarticle combined the largest Lyapunov exponents and nonlinear interdependece withBai and Perron’s Multiple Structural Change model to build the Multiple StructuralChanges model in crisis contagion’s infection based on largest Lyapunov exponentsand the Multiple Structural Changes model in crisis contagion based on based onnonlinear interdependence.Secondly, this paper analyzed empirically the dynamic changes of nonlinearcharacteristics of American, the crisis contagion’s infection, during the2008globalfinancial crisis based on the Multiple Structural Changes model in crisis contagion’sinfection based on largest Lyapunov exponents. We used the S&P500stock indexdata to carry on the empirical analysis on the basis of the simulation experiments. Wefound the the time node of every period during financial crisis and classified thestages of the financial crisis; Compared the empirical results of the modified MultipleStructural Change model with that of the Bai and Perron’s model; Analyzed thedynamic changes of nonlinear characteristics in different stages during financial crisis.Finally, this paper analyzed empirically the dynamic changes of nonlinearinterdependence of American and other9countries (or regions) during the2008global financial crisis based on the Multiple Structural Changes model in crisiscontagion based on nonlinear interdependence. Firstly, we analyzed empirically the2008financial crisis data based on Bai and Perron’s Multiple Structural Changesmodel. Then we used the original nonlinear mutual prediction and the modified model based on nonlinear interdependence to study the direction, time and strength of crisiscontagion from the United States to nine other infected countries. And on this basis,we tried to provide applicable tools and early warning indicators to our country forpreventing crisis contagion.