Abstract: Recent research on macroeconomic growth has been focused on resolving two key issues: (i) parameter heterogeneity and (ii) the consequences of competing economic growth theories. Typically each one of these issues is dealt with individually. However, a key to an adequate study of growth relationships requires simultaneously dealing with all two facets. This paper moves in this direction by considering points (i) and (ii) in lock-step. We use recently developed nonparametric model selection techniques to deal with nonlinearities and competing growth theories. We also contrast our results with the widely used Bayesian Model averaging method. We find that the inclusion of nonlinearities and parameter heterogeneity significantly changes the landscape on which growth empirics has traditionally rested.