Growing rural-to-urban and international migration flows have sparked concerns about the investments in the education of the children left behind in Cambodia. We draw on a panel household-level survey conducted in rural villages in 2014 and 2017 to analyse the relationship between parental migration and children's schooling. The analysis shows that children of migrant parents complete less years of schooling than children of non-migrant parents. We find a bigger effect for children whose parents migrated abroad, for children aged 12 to 17, and for maternal migration. The effect persists over time, with parental migration in 2014 influencing schooling in 2017. We exploit the longitudinal dimension of the data to estimate a placebo, which greatly reduces the concerns related to the possible confounding effect of timeinvariant unobserved heterogeneity. The negative effect that we find appears to be driven by the reduced parental input in children's education rather than by an increase in children's work.