This paper studies whether voters in Denmark hold local politicians accountable for the performance of local schools. Using register data and polling-station-level voting records, we study this effect for Danish municipal elections in 2013 and 2017. Register data allows us to representatively assess which schools are relevant for every polling-station district by linking residing voters to their children's institution. We find robust evidence of retrospective voting in 2013. In 2017, however, incumbents' electoral success was independent of changes in school performance. Moreover, our results for 2013 show that the effect is strongest for relatively richer and more educated voters, in areas with more intensive political competition, and applies only to right-wing incumbents.