Systemic Financial Risk Inference in a Global Setting
29 Pages Posted: 16 Aug 2014
Date Written: August 15, 2014
We propose a new top-down approach to measure systemic risk in the financial system. Our framework uses a combination of macroeconomic, financial and rating factors in representative regions of the world. We formulate a mixed-frequency state-space model to estimate macroeconomic factors. To derive financial risk factors, we use Moody's/KMV expected default frequencies after accounting for ratings of major financial institutions in the considered regions. The estimated factors are combined to derive probabilities for systemically relevant defaults in the financial industry. Regional macroeconomic factors are significant predictors of the existence and number of systemically important defaults, while regional financial risk and ratings factors are relevant for the existence only. For major events, global credit risk also matters.
Keywords: Systemic financial risk, Factor models, Mixed frequency models, Kalman filter, State-space model, Hurdle model
JEL Classification: C33, E44, G01, G17
Suggested Citation: Suggested Citation