Penalized maximum likelihood estimation of logit-based early warning systems

Panel logit models have proved to be simple and effective tools build early warning systems (ews) for financial crises. But because crises are rare events, the estimation of ews does not usually account country-specific fixed effects, so as avoid losing all information relative countries that never face a crisis. I propose using penalized maximum likelihood estimator fixed-effects logit-based where observations retained. show including country while preserving entire sample, improves predictive performance ews, both in simulation out with respect pooled, random-effects standard models.