Resource Allocation in Extended Sibships: An Empirical Investigation for Senegal

Using data from an original nationally representative survey in Senegal, we estimate the effect of the widespread practice of child fostering on host children's educational outcomes. We find that host girls aged six to nine at the arrival of the foster sibling significantly benefit in terms of school enrolment. We do not find any effect on host boys. Although cross-sectional, our survey data enable to implement different estimation strategies to measure the effect of interest limiting estimation bias due to the potential joint determination of decisions to host a foster child and to enrol one's child. Furthermore, we address a second issue commonly encountered by the literature: the identification of the foster children. Two groups are identified: children formally fostered-in and children informally fostered-in. Extending the group of foster children to include children informally fostered-in affects our results. One possible channel for the observed effect on host girls could be the release of liquidity constraints. Indeed the labour supply of mothers increases with the presence of foster girls, likely because foster girls replace them in their household tasks.