Abstract: Based on the extreme theory, a new approach is presented for measuring systemic financial risk. Using extremal quantile regression, 33 listed financial institutions'contributions to the systemic risk of financial system are estimated and systemically important institutions in China' s financial system are recagnjzed. It is found that, the distributions of the growth rate of market valued total assets and the co-movement of the tail risk can be accurately estimated using the extremal quantile regression. The level and variance of the systemic risk contribution in bank sector are the highest. The top ten systemically important financial institutions are almost in bank sector. The levels of systemic contribution in security sector, insurance sector and trust sector are relatively low. Moreover, systemic contributions of joint-stock commercial banks are higher in our study. An empirical tool is provided in this paper for further macro-prudential regulation of systemically important institutions.

Key words: systemic financial risk, CoVaR, extremal quantile regression, risk measure