Systemic Risk Diagnostics

A macro-prudential policy maker can manage risks to financial stability only if currentand future risks can be reliably assessed. We propose a novel framework to assessfinancial system risk. Using a dynamic factor framework based on state-space methods, we model latent macro-financial and credit risk components for a large data setcomprising the U.S., the EU-27 area, and the rest of the world. Controlling for global,region-specific, and industry effects, we construct coincident measures ('thermometers')and forward looking indicators of financial distress and the likelihood of financial melt-down. We find that credit risk conditions can significantly and persistently de-couplefrom macro-financial fundamentals. Such decoupling can serve as an early warningsignal for macro-prudential policy.