Chetty et al. (2014) show that children from low-income families achieve much better adult outcomes, relative to those from higher-income families, in some places than in others. I use data from several national surveys to investigate whether children’s educational outcomes (educational attainment, test scores, and non-cognitive skills) mediate the relationship between parental and child income. Commuting zones (CZs) with stronger intergenerational income transmission tend to have stronger transmission of parental income to children’s educational attainment, as well as higher returns to education. By contrast, the CZ-level association between parental income and children’s test scores is only weakly related to CZ income transmission, and is stable across grades. There is thus little evidence that differences in the quality of K-12 schooling are a key mechanism driving variation in intergenerational mobility. Access to college plays a somewhat larger role, but most of the variation in CZ income mobility reflects (a) differences in marriage patterns, which affect income transmission when spousal earnings are counted in children’s income; (b) differences in labor market returns to education; and (c) differences in children’s earnings residuals, after controlling for observed skills and the CZ-level return to skill. This points to job networks or the structure of the local labor and marriage markets, rather than the education system, as likely factors influencing intergenerational economic mobility.