Cascading failure, financial network and systemic risk

How to accurately measure the systemic risk is one of the fundamental and challenging problems in the field of risk management. Most previous studies do not fully consider the cascading failure mechanism caused by risk co-contagion and network effects, leading to misestimation of systemic risk. We construct financial institution tail risk networks by LASSO technique and then simulating the cascading process of risk contagion by ΔCoES on the networks. By developing a general cascading failure model, this paper proposes a novel indicator, ESRank, to measure systemic risk. We apply ESRank to analyze Chinese financial institutions and the empirical results suggest that: (i) during the crisis periods, especially the 2015–2016 stock crash period, the Chinese financial system manifests a higher ESRank in comparison to normal periods; (ii) the securities sector is the largest risk contributor before the stock crash, while the diversified financial institutions have displayed increasing risk contributions afterwards; (iii) compared with the traditional systemic risk indicators such as VaR, CoVaR and SRISK, the proposed ESRank demonstrates the outstanding characteristics of better predictive and explanatory capabilities regarding institutional profitability.