Gautrain, Elsa 2026 Child fostering is widespread in sub-Saharan Africa, yet its effects on children’s education remain debated. This paper examines educational outcomes across different living arrangements in Ghana, distinguishing fostered children, biological siblings remaining with their mothers, host siblings, and children in non-fostering households. A key challenge is that fostered children observed in host households lack information on their family background in standard surveys. I address this limitation by exploiting information from foster-out children declared by mothers and implementing a donor-based multiple random imputation strategy to recover missing family background for foster-in children. Once mothers and sibship characteristics are accounted for, the initial educational disadvantage of fostered children largely disappears. The results suggest that observed educational gaps mainly reflect selection into fostering arrangements rather than fostering itself. application/pdf https://folia.unifr.ch/global/documents/335197 https://folia.unifr.ch/documents/335197/files/wp545.pdf eng FSES info:eu-repo/semantics/altIdentifier/doi/10.51363/unifr.ewp.377g8x info:eu-repo/semantics/altIdentifier/issn/3042-7517 info:eu-repo/semantics/openAccess CC BY-NC Child fostering Educational outcomes Family background Multiple imputation Extended family Ghana info:eu-repo/classification/udc/33 Left Without Origins? Child Fostering and Education under Missing Family Information http://purl.org/coar/resource_type/c_18ws