This paper provides a comprehensive early warning system (EWS) that balances the classical signaling approach with the best-realized machine learning (ML) model for predicting fiscal stress episodes. Using accumulated local effects (ALE), we compute a set of thresholds for the most informative...

We measure the impact of frequent exogeneous shocks on small ECCU economies, including changes to global economic activity, tourism flows, oil prices, passport sales, FDI, and natural disasters. Using Canonical-Correlation Analysis (CCA) and dynamic panel regression analysis we find significant...

This article discusses election forecasting in the Caribbean through the reflective lens of model-based election forecasting in Jamaica between 2015 and 2020 to chart the progress made and to increase public understanding of the issues. Poll-based election forecasting is the dominant modality...