We propose a novel framework to assess financial system risk. Using a dynamic factor framework based on state-space methods, we construct coincident measures (‘thermometers’) and a forward looking indicator for the likelihood of simultaneous failure of a large number of financial intermediaries. The indicators are based on latent macro-financial and credit risk components for a large data set comprising the U.S., the EU-27 area, and the respective rest of the world. Credit risk conditions can significantly and persistently de-couple from macro-financial fundamentals. Such decoupling can serve as an early warning signal for macro-prudential policy.
credit portfolio models financial crisis frailty-correlated defaults state space methods systemic risk