Using data from 18 countries, we investigate the effects of child marriage bans on infant and under-5 mortality. We use variation in mothers' exposure to the ban across cohorts within each country and regional variation in "treatment intensity," calculated based on child marriage prevalence and marriage age prior to the ban. We find that child marriage bans reduced infant and under-5 mortality, with magnitudes of 14.3 and 19.9 percent corresponding to a one standard deviation change in treatment intensity. Reductions were driven by low-income countries and less wealthy households, primarily due to increases in age at first marriage and birth.