Monetary integration in South America: Elección of candidates through unsupervised machine learning

Abstract

Applying Unsupervised Machine Learning techniques to a set of nominal variables (based on the optimum currency area [OCA] theory and the Maastricht Treaty criteria) and industrial indicators (based on similar production patterns), this paper aims to identify potential candidates for a monetary integration in South America (SA). The main conclusion is that, according to the clustering of the nominal and industrial indicators, the countries in best position for a hypothetical monetary integration in SA are Chile, Colombia, and Perú (and Ecuador to a lesser extent); countries that are generally members of the same cluster. This group of economies, which belong to the Pacific Alliance, are in a better position to meet various criteria for regional monetary integration, such as nominal convergence and similar production patterns.