In African countries, children often combine school and work. This article exploits Senegalese panel data to assess the relationship between child labor and learning, measured by test scores. Test scores from the beginning of primary school control for children’s cognitive abilities, and children’s past time allocation decisions are instrumented by changes in rainfall and distance to primary school. Some of the tests were administered verbally in order to pick up effects for children who had only attended school very briefly. I do not find that children’s past participation in economic activities is associated with lower adolescent cognitive achievement, but rather that it is associated with higher oral mathematics scores. This association is stronger when I control for years of schooling, which suggests that work does displace schooling but does allow children to acquire some skills. From a paper by Dumas (2012).And the paper concludes - from an early draft:
The positive impact of work vanishes though if the child worked more than 4 hours a week on average when he started his activity or even strongly depletes accumulation of human capital if he was employed outside of the household.Child labor is extremely resiliant, especially in the case of boys. See here (Tanzania) and here (Brazil).