Systematic risk measurement in the global banking stock market with time series analysis and CoVaR

Motivated by the growing importance of systemic risk in the global banking system, the authors measure the risk of the system and the marginal contributions of the institutions in several ways in terms of stock markets. The undiversifiable risk appearing in specific market sectors is called systematic risk rather than systemic risk. The paper focuses on global banking stocks comprising global systemically important financial institutions (G-SIFIs), and discusses the global systematic risk measurement. To forecast future joint distribution of returns, the authors utilize the multivariate autoregressive moving average generalized autoregressive conditional heteroscedasticity (ARMA-GARCH) model with the multivariate normal tempered stable (MNTS) distributed and multivariate normal distributed innovations. This work statistically demonstrates that the ARMA-GARCH model with the MNTS distributed innovations is a more realistic model for G-SIFI stocks. In line with previous studies, the authors estimate four systematic risk measures: joint probability and conditional probability of negative stock return movements, ΔCoVaR, and ΔCoAVaR. It is found that the joint probability of negative movements is a good indicator for a significant increase in systematic risk. Subsequently, the authors investigate the relationship among the other three measures and find the following. Crosssectional linkages between AVaR and ΔCoAVaR are few, if any, but there is a strong time series linkage. On the other hand, the conditional probability of negative movements and ΔCoAVaR show similar cross-sectional magnitude relations, though their time series linkage is not clear. Thus, both AVaR and conditional probability of negative movements reinforce each other and serve a useful reference for ΔCoAVaR-based systematic risk measurement.