In this study of how income distributions in developing countries evolve, the authors propose a generalization of the Oaxaca-Blinder decomposition of differences in mean earnings, extending it to entire distributions of household incomes. The method uses a model of household income determination to simulate counterfactual distributions of income.