We are back again! What can artificial intelligence and machine learning models tell us about why countries knock at the door of the IMF?

Abstract

This paper examines the factors that predict an IMF bailout. In doing so, we use a large dataset from 1993 to 2021 with 6550 observations and 138 features and adopt recent advances in machine learning and artificial intelligence models such as tree-based, boosting and artificial neural network techniques. We find that apart from traditional indicators such as debt and macroeconomic factors; agricultural, energy, health and social factors are strong predictors of an IMF bailout. These factors have hitherto not received much attention in the literature.