In many OECD countries income inequality has risen, but surprisingly redistribution as well. The theory attributes this partly to the redistributive effect of education spending. In the model income inequality and growth depend in an inverted U-shaped way on education. To maintain a given level of human capital it is shown that a less efficient schooling technology requires more resources, which lowers pre-tax and post-tax income inequality as well as growth. Using consistently defined income data from the Luxembourg Income Study suggests that there is a negative relationship between growth and income inequality in rich countries. It is argued that using some unadjusted inequality measures in growth regressions may yield estimates that are biased upwards. The evidence suggests that a rich country would raise growth with lower pre-tax and post-tax inequality if it spent more on education.