Estimating Currency Crisis Probabilities in MENA Countries Using Markov Switching Models

Résumé

This paper develops an empirical framework to estimate currency crisis probabilities in selected MENA countries using Markov Switching models. Unlike traditional logit/probit or linear regression approaches, this methodology captures both abrupt and transitory regime shifts and allows for potential asymmetries in regime levels. The model predicts crisis probabilities without assuming prior knowledge of the initial state (“crisis” or “non-crisis”), enhancing robustness. Macroeconomic and financial variables are selected based on literature and data availability. The model’s performance is evaluated and compared with conventional approaches. Results demonstrate the effectiveness of Markov Switching models in identifying periods of heightened speculative pressure on exchange rates, providing valuable insights for policymakers and analysts in managing currency risks.