Measuring systemic risk of the banking industry in China: A DCC-MIDAS-t approach

This paper measures systemic risk in the Chinese banking sector by using the CoVaR approach. First, we introduce the student's t distribution into the standard DCC-MIDAS and propose a DCC-MIDAS-t model, particularly suitable for processing fat-tailed financial returns. We then apply the proposed DCC-MIDAS-t model to measure systemic risk following the idea of CoVaR. The empirical studies on the Chinese banking industry show that the DCC-MIDAS-t model is preferred in terms of the accuracy of volatility prediction and CoVaR measure. Macroeconomic information are proved to be able to improve the accuracy of systemic risk measure and M2 is most useful in comparing with IP and PPI. Several banks, such as BOC and BONB, suffered a lot from the stock market crash in June 2015 and some of them still maintain high risk spillovers during the after crash period. Our findings suggest that these banks should be the focus of supervision even after the crash.