An Early Warning System for Bankruptcy Prediction: lessons from the Venezuelan Bank Crisis

During 1993-94 Venezuela experienced a severe banking crisis which ended up with 18 commercial banks intervened by the government. Here we develop an early warning system for detecting credit related bankruptcy through discriminant functions developed on financial and macroeconomic data predating the crisis. A robustness test performed on these functions shows high precision in error estimation. The model calibrated on pre-crisis data could detect abnormal financial tension in the late Banco Capital many months before it was intervened and liquidated.