Abstract: In this paper we propose an innovative method for the empirical analysis of cross-country economic growth, based on the application of a maximum likelihood clustering algorithm, proposed by Giada and Marsili (2001 and 2002), to a dataset on economic growth. This method uses the maximum likelihood principle to obtain a partition of multidimensional objects, without imposing a specified number of clusters from the outset, and has the advantage of allowing the evaluation of the statistical significance of the cluster structure obtained. .