关键词: 系统性风险, 多条件CoVaR, 高频数据, HAR-CAW模型

Abstract: Based on heterogeneous market hypothesis and high frequency data, taking advantage of the characteristics that the high frequency data is not only rich in volatile information ,but also with higher sampling frequency, this essay construct multivariate joint distribution between banks and system process and taking the dynamic contact between banks into account, in this essay, a high-frequency multi-CoVaR model with rich economic significance and flexible forms is constructed, in addition, we propose a group spillover effects index to capture the systemic importance of each group, there are analysis on risk spillover effects of commercial banks on financial system from a group perspective. The empirical results of listed commercial banks in China show that: in terms of systemic risk contribution, the model accurately characterizes the time-varying characteristics of the risk contributions of three types of banks in China this years, for example, during the times of the subprime crisis, the European debt crisis, "money shortage" period in our country and abnormal turbulent times of our country stock market, due to the infectious features of risk, these three kinds of commercial banks' contribution of systemic risk increased, and at other times are relatively stable; in terms of the spill-over effects to system, Industrial and Commercial Bank of China shows the largest, which is closely related to its huge asset size and its importance in the financial system. When analyzing the risk impact of banking groups with same types, the state-owned firms shows the largest, followed by the joint-stock commercial banks. The systemic risk contribution rate of this two groups are almost the same, both are much larger than the city commercial banks. It shows that the risk spillover effect of joint-stock commercial banks on the financial system should not be underestimated; this model is also applicable when analyzing different types of banking groups. The model built in this essay can achieve measuring the spillover effects of a group of banks, it can provided reference to financial supervisory departments for classified regulation and assistance order with limited abilities.

Key words: Systemic Risk, Multi-CoVaR, High-Frequency Data, HAR-CAW Model