An Early Warning System for Systemic Banking Crises - A Robust Model Specification

Using a panel of 27 developed economies estimated quarterly over the period 1980-2011 we develop a flexible systemic banking crisis early warning system (EWS). Clear evidence is provided that fitted multivariate logit... [ view full abstract ]

Using a panel of 27 developed economies estimated quarterly over the period 1980-2011 we develop a flexible systemic banking crisis early warning system (EWS). Clear evidence is provided that fitted multivariate logit probabilities, estimated recursively, yield more informative crisis signals than any single macroeconomic, credit aggregate or asset price variable does on its own. When the model robustness techniques of Young and Holsteen (2017) are applied even stronger crisis signals are generated. Deciding which of the above variables to include in the EWS is determined by adopting a signals-based approach to each prospective indicator with the most informative yet robust variables identified in terms of their performance according to noise-to-signal ratios, weighted noise to signal ratios and the "usefulness''  measure. The latter two measures take policy-makers' relative preferences for false versus missed signals into account. The overall approach ensures a parsimonious yet effective specification yielding forward-looking indicators that outperform all raw input indicators examined in crisis-signaling terms.