In this paper education simultaneously affects growth and income inequality. More education does not necessarily decrease inequality when the latter is assessed by the Lorenz dominance criterion. Increases in education first increase and then decrease growth as well as income inequality, when measured by the Gini coefficient. There is no clear functional relationship between growth and measured income inequality. The model identifies regimes of this relationship which depend crucially on the production and schooling technology. Conventional growth regressions with human capital and inequality as regressors may miss the richness of the underlying nonlinearities, but viewed as approximations may still provide important information on the nonlinear relationship between growth and education.

JEL Classification: O4, I2, D31, H2; Erstellt März 2006