Estimating Returns to Birthweight by Placenta Previa

Worse future outcomes of infants with lower birthweight are widely documented, but the underlying causal relationship remains controversial. Recent causal studies rely mostly on within-twin variations, and tend to find significant effects of birthweight on long-term outcomes, but substantially smaller effects on infant health than by standard regression estimates. To improve our understanding of the causal effects of birthweight, we analyze population data for 1.7 million singleton newborns in Denmark since 1981 using a new instrument – the diagnosis of placenta previa, which is an obstetric complication that generates exogenous variation in birthweight. Studying an extensive range of outcomes, we find that: (1) birthweight does have significant positive causal effects on child health and, in sharp contrast to earlier twin studies, their magnitudes are similar to or larger than standard regression estimates; (2) the effect of birthweight on general health diminishes as the child grows older (catch-up effect), although long-term effects on health exist; (3) the birthweight effect on infant health in Denmark has diminished significantly over the last three decades; and (4) a large part of the highly significant correlations between birthweight and non-health outcomes appear to be noncausal. Our results provide a foundation for policies that target birthweight: returns to birthweight can be substantial, but they depend on the standard of medical care