Abstract: In recent years, China's financial sector has encountered more and more unstable factors than before. A number of factors, domestic and foreign, such as the increasing fiscal deficit of the central government, the heavy debt of local government, and the rising bad loans of banks, the trade conflicts between China and the United States and other developed countries, directly or indirectly pose a potential threat to the normal operation of China's financial sector and enhancing the expectation of financial instability of China. Due to China's current macroeconomic environment and uncertainty faced by China's financial sector, the decision-making authorities of Chinese government attach great importance to the prevention and resolution of systemic financial risks. How to continuously improve the accuracy of the model for risk measurement and prediction, and provide more reliable, multi-level and multi-perspective reference for the formulation and implementation of effective regulatory policies is the direction of joint efforts of academia and regulatory authorities. In view of this, this paper takes 33 listed financial institutions in the four industries, banking, securities, insurance and comprehensive financial services in China, as samples to empirically study their systemic risk spillovers, their systemic risk contributions and the determinants. The rolling window dynamic copula model is applied to fit both the time-varying structures and the time-varying coefficients of the dependence between financial institutions and the financial system. The systemic risks of sample financial institutions are measured with this dynamic dependent structure fitting framework, and the determinants of the systemic risk contributions of these financial institutions are investigated through macro, industry and institution-level analyses. The main conclusions are as follows. First, the systemic risk contributions of financial institutions in securities industry are the largest, while banking institutions are the biggest potential threat to the financial system. Second, the dependence structures of institutions and the system determine the size of institutional risk spillovers, and the degree of dependence between institutions and the system, and the volatility of the system significantly affect the systemic risk of the system as a whole. Third, there is a phenomenon of mutual transformation between financial risk and fiscal risk in China. The fundamental way to resolve the systemic financial risk is sticking to the improvement of the financial ecological environment. The contributions of this paper are as follows. Firstly, an improved dynamic copula model with dual time-varying characteristics is established to make the model settings and estimation closer to the real. Secondly, the determinants of systemic risk contributions of the financial institutions are analyzed from macro, industry and institution levels, which helps prevent and control risks from the source of systemic risk and can be used as a reference for regulatory decision-making.

Key words: systemic risk, rolling window dynamic Copula model, dependence structures