This study uses longitudinal data from South Africa to estimate the relationship between early childhood nutritional status and schooling outcomes five years later. Preferred estimates from the full sample aged zero to five, which treat
prior nutritional status as endogenous, show no impact of past nutritional status on current schooling, in contrast to a recent article in this journal using data from Pakistan. However, we find significant estimates for children who
were malnourished, as well as among children younger than three years of age in the base year. These results suggest that the relationship between health and cognitive achievement is complex, and the effects may be sensitive to time
between measurements and the timing of malnutrition itself.