We analyze whether a redistributive government should provide ex ante insurance against unfortunate outcomes or whether it should instead rely on transfers for redistributing income ex post. To this end, we develop a model of education in which individuals face educational risk and wage dispersion across two types of skills. Successful graduation and working as a skilled worker depends on individual effort in education and on public resources, but educational risk still causes (income) inequality. We show that in a second-best setting, in which learning effort is not observable, improving the quality of education by public funding of the educational sector has a significant effect and that this increases efficiency in comparison to a pure (linear) income tax with income transfers from skilled to unskilled workers. Compared to a first-best solution, providing ex ante insurance significantly gains importance relative to traditional ex post redistribution, because it simultaneously alleviates moral hazard in education. These results are strengthened when a (distortionary) skill-specific tax can be implemented.
human capital investment endogenous risk learning effort optimal taxation public education