Revisiting the determinants of sovereign debt ratings in Europe through artificial intelligence techniques

Abstract

In papers using artificial intelligence (AI) techniques, little attention has been paid to the determinants of sovereign debt ratings. We propose a reduced set of variables regarding the economic performance of a country that are consistent with the idea of debt sustainability. The robustness of this set is supported by the results obtained with different well-known AI techniques using data from EU-15 countries during the 2002–2017 period as the experimental setting. The variables are publicly available, allowing a quick and reliable assessment of the creditworthiness of a sovereign and providing useful information for decision-makers and investors.