The paper incorporates some recent developments from the unsupervised machine learning literature to re-evaluate the cross-country convergence hypothesis in a context beyond GDP. The application of a distribution-based clustering algorithm suggests the formation of three local convergence clubs.