Financial systemic risk measurement based on causal network connectedness analysis

We construct the causal complex network of financial institution based on the Granger-causality network as well as the principal component analysis, and further analyze the network topology structure characteristics using centrality indicators. From the perspective of time dimension and space dimension, the causal complex network among financial institutions is analyzed with the dynamic variation of systemic risk measured. One improvement is that the CAPM is employed to filter out the market risk. Moreover, we study the contribution of individual financial firm to the systemic risk through three well known systemic risk measures containing the systemic risk index (SRISK), the marginal expected shortfall (MES) and the conditional Value at Risk (CoVaR). The proposed methodology framework is applied to the banks, securities and insurance companies in Chinese financial markets. Empirical researches find that the causal network of Chinese financial entities possesses small world and scale free properties, with the number of connections increasing dramatically in turmoil periods, indicating stronger interconnectedness in the financial system during crisis. Moreover, the network based connectedness measures through topology structure analysis can identify and quantify the financial crisis, serving as significant systemic risk indicators. In addition, different systemic risk measures employed in our framework can identify similar rankings of risky financial firms quantitatively, which provide references for regulations.