In many low- and lower-middle-income countries, private schools are often considered to offer better quality of education than government schools. Yet, there is a lack of evidence to date on their role in reducing inequalities: namely, the extent to which private schooling improves learning among the most disadvantaged children. Our paper uses household survey data from Kenya, Tanzania, and Uganda to identify whether any observed impact of private schooling on core literacy and numeracy skills differs according to children’s household wealth. We demonstrate wealth gaps in access to private schooling, and use inferential models to account for observable differences between those who do and do not enrol in private schools. In Kenya and Uganda, we find that private schooling appears to improve the chances of children learning relative to their peers in government schools, but the chances of the poorest children learning in private schools remains low and is at best equivalent to the richest learning in government schools. In Tanzania, private schooling does not seem to improve poorer children’s learning, whereas it does for richer children. These findings raise a caution about the extent to which private provision can help narrow learning inequalities.