In this paper, we investigate whether voters hold local politicians accountable for the performance of local schools. We examine this effect for the 2013 and 2017 Danish local elections using register data and polling station-level voting records. We find robust evidence of retrospective voting from pooled and fixed effects estimations. Exploiting the micro-level character of our data, we present evidence that higher-income citizens are more sensitive to changes in school performance, while other demographic and political characteristics do not appear to have mattered. •We study how changes in school results affect the mayoral party’s electoral success.•Contributes to retrospective voting literature on local politics.•School outcomes affect vote shares in FE and repeated cross-sectional analysis.•Heterogeneity analysis shows that higher-income voters react more to school results.