Identifying excessive credit growth and leverage

Unsustainable credit developments lead to the build-up of systemic risks to financial stability. While this is an accepted truth, how to assess whether risks are getting out of hand remains a challenge. To identify excessive credit growth and aggregate leverage we propose an early warning system, which aims at predicting banking crises. In particular, we use a modern classification tree ensemble technique, the “Random Forest”, and include (global) credit as well as real estate variables as predictors.