The Systemic Risk Index Of My Country's Commercial Banks

The outbreak of the global financial crisis in 2007-2009 has made people realize that the capital shortage of large financial institutions may cause a painful negative spillover effect to a country’s real economy.This requires a reasonable and effective tool for regulators to measure the degree of capital shortage that a financial institution will experience under the severe impact of the entire financial system.Based on various considerations,this paper uses SRISK to measure the share of a single financial institution in the overall systemic risk.SRISK is defined as the expected capital shortage of a financial institution under the condition of long-term market downturn.It is a function of LRMES(Long Run Marginal Expected Shortfall)under the condition of financial institution size,leverage level and market downturn.This paper focuses on the selection of LRMES estimation method in SRISK calculation.First,the SRISK calculation method is explained.The calculation of SRISK requires balance sheet information and reasonable LRMES estimation.Then,two methods of estimating LRMES(GARCH-DCC model and static bivariate normal model)are introduced.Then,the GARCH-DCC model and static bivariate normal model are used.Sixteen listed commercial banks in China have carried out empirical research.Finally,two methods of estimating LRMES are evaluated by predictive regression.The results show that:(1)From 2008 to 2017,the five banks contributed the most to systemic risk.The contribution of joint-stock commercial banks to systemic risk lies between five major banks and city commercial banks.(2)The SRISK of commercial banks is constantly changing.During the financial crisis of 2007-2009,the stock market disaster of 2015 and 2017,the systemic risk of commercial banks is higher.(3)Although the five major banks are generally recognized as systemically important financial institutions,the systemic importance of China Merchants Bank,Shanghai Pudong Development Bank and Societe Generale Bank is also on the rise.(4)According to Brownlees and Engle’s research on SRISK in 2017,this paper evaluates two methods of estimating LRMES by using predictive regression method.It is found that estimating LRMES based on GARCH-DCC model is better than estimating LRMES by using static bivariate normal model.