The role played by endogenous growth models in growth literature might be analyzed from two perspectives. In the first place, is it emphasized the necessity to replace the hypothesis of perfect competition with monopolistic competition in every mathematical model. Secondly, there is no scientific argument to accept the assumption of unconditional income convergence among countries. Taking into consideration empirical evidence and theoretical arguments, we tried to demonstrate that, once we accept the existence of increasing returns, there is only a little place for convergence all over the world. From this perspective, we can accept the hypothesis of convergence only for some categories of countries characterized by homogeneity as regards institutional arrangement and geographical position.