We exploit exogenous variation in the risk of waterborne disease created by implementation of a major water reform in Mexico in 1991 to investigate impacts of infant exposure on indicators of cognitive development and academic achievement in late childhood. We estimate that a one standard deviation reduction in childhood diarrhea mortality rates results in about a 0.1 standard deviation increase in test scores, but only for girls. We show that a reason for the gender differentiated impacts is that the water reform induces parents to make complementary investments in education that favor girls, consistent with their comparative advantage in skilled occupations. The results provide novel evidence of the potential for clean water provision to narrow test score gaps across countries and, within countries, across gender.