Multidimensional Distance-to-Collapse Point and Sovereign Default Prediction

This paper focuses on predictability of sovereign debt crisis proposing a twostep procedure centered on the idea of a multidimensional distance-to-collapse point. The first step is non-parametric and devoted to construct a generalized early warning system that signals a potential crisis every time a group of indicators exceeds specific thresholds. The second is parametric and tries to contextualize the country default within a theoretical-based process depending on the distance from the thresholds estimated in the first step. Such regression approach helps to understand how nonparametric distances could be mixed together in order to reduce the n-dimensional measures of fundamentals, then obtaining a “generalized” distance to normality which summarizes all relevant information to better predict the likelihood of default for a given country across time. Empirical evidence on a broad sovereign default database proves that our methodology fit the data quite well, being able to deliver prediction of future events around the 80 per cent of the actual defaults. * We thank Chuck Stone (University of Berkeley) and seminar participants at Bicocca University and University of Brescia for useful comments and suggestions. We also thank Paolo Manasse (University of Bologna) for providing the database, Dan Steinberg (Salford Systems) for providing CART software. ** Corresponding author.