We examine whether and how the educational background of political leaders matters for policy choices and outcomes. Using data on municipalities in Brazil from 2000-2008, we estimate the effects of electing a more educated leader in a regression-discontinuity design whereby policy inputs and outcomes in municipalities where a highly educated candidate barely won the election are compared with those of municipalities where a highly educated candidate barely lost. Our results indicate that highly educated mayors make different choices regarding the allocation of public funds and inputs in critical sectors when compared to non-highly educated mayors, yet they do not produce better indicators on a variety of measurable outcomes. Furthermore, our estimates suggest a negative impact of educated mayors on local economic growth and children's health. We additionally document the existence of heterogeneity in the effects of highly educated leaders along political ideology and age of the candidate. Lastly, highly educated leaders are not more likely to be reelected, suggesting that they are not perceived as better politicians.