Abstract: Mainly due to data unavailability, time with the mother is usually not included as an input in empirical papers trying to find the determinants of child's achievement. This paper attempts to do so by using unique data that collects a child's time-use diary, cognitive and non cognitive test scores. We implement the program evaluation of continuous treatments methodology. The main conclusion is that the returns to maternal time are not monotonic: increasing maternal time affects positively children's cognitive skills only up to a certain level. This is true both for young and old children. Once we divide the sample according to age and race we find that, at least for cognitive outcomes, young white children seem to benefit particularly from maternal time, while black young children from school/care time. For older children, there is some evidence that white children benefit from spending more time with mothers in all domains while black children present more behaviour problems if they receive high maternal time.