Financial Systemic Risk Measurement Based On Multivariate Time Series Change Point Detection

Since the global financial crisis in 2008,systemic risk has become an important part of financial regulation,and it is also a hot topic and focus of domestic and foreign scholars.Due to the externality and the contagion and spillover effects of systemic risk,to a large extent,systemic risk even may have a serious impact on the entire financial system.Because complex financial data will undergo structural change in the face of many uncontrollable factors,in order to consider the necessity of modeling the financial time series piecewise in order to improve the measurement results.Through the change point detection of multi-dimensional financial time series,combined with the difference of distribution and model before and after the change point occurs,this paper demonstrates that the existence of structural change point will lead to the change of its applicable model.Therefore,in order to further improve the accuracy of the measurement results,it is necessary to combine the existence of change points with time-interval measurement.In this paper,we select 22 listed financial institutions' daily stock returns,including banks,securities and insurances,to construct a multidimensional financial time series.We use the CUSUM algorithm based on covariance matrix to detect the change points.Based on the four identified change points,the series can be divided into five sub-sequences,and the financial system return series is analyzed and tested by descriptive analysis.At the same time,the rationality of the identified change points is proved by considering the domestic economic situation before and after the change points.In addition,based on the CoVaR method,the systemic risk of 22 financial institutions in five stages is measured and compared with the industry benchmark.The results show that the systemic risk of each stage does exist large differences in value and volatility.Based on the results of this study,it is shown that the structural change of the financial system occurred at some time due to the existence of risk,which caused the changes of data distribution and model before and after the time point,and finally led to the financial change point.By proving the authenticity of the existence of financial change points,it is further shown that the systematic risk measurement of financial time series should be based on the existence of change points to improve the accuracy of the results.This paper proves the correlation between structural change point and systemic risk,which fully reflects the influence of the existence of the structure change point on financial systemic risk measurement.