This study develops an early-warning model designed to anticipate sovereign debt crises, based on macroeconomic and financial indicators from 41 countries between 1990 and 2014. Using panel logit estimations, the analysis aims to identify the factors that increase the probability of default and to assess the predictive capacity of the model under different temporal scenarios.
The results show that it is possible to construct a model with strong explanatory power, capable of accurately distinguishing between periods of stability and episodes of crisis. Through temporal validation exercises, the model demonstrates robust out-of-sample performance, even when using cut-off points that emulate real-time prediction conditions.
Overall, the findings confirm the usefulness of early-warning models as complementary tools for monitoring public debt sustainability, while also highlighting the importance of considering structural and political factors in their interpretation.
Keywords: Sovereign debt crises, Early-warning models, Panel data, Macroeconomic indicators, Fiscal vulnerability. JEL classification codes: C33, F34, H63, E62.
Autor: Mauricio Alejandro Alvarez
Directores: José J. Bercoff y Víctor D. Iajya
