Whether user fees for basic health services should be charged or abolished for the
poor has recently been debated. This study examines the impact on child health
status of removing user fees in South Africa. Our main innovation is to exploit
plausibly exogenous variation in access to free healthcare, due to the fact that
black Africans under apartheid could exercise little political power and residential
choice. By looking at ex ante similar children, we find substantial improvements
in health status among children. Falsification exercises confirm no preexisting
trend in the pre-reform period or no treatment effect among non-eligible children
in the post-reform period.