The Roles of Birth Inputs and Outputs in Predicting Health, Behavior, and Test Scores in Early Childhood

The goal of this study is to address directly the predictive value of birth outputs, particularly birth weight. Specifically, in the context of four birth inputs (maternal smoking, drinking, prenatal care, and maternal weight gain during pregnancy), three birth outputs (gestational age, birth length, and birth weight), and 32 exogenous covariates observed during pregnancy, we investigate the absolute predictive importance of birth weight in forecasting child health (as indicated by height and weight), child behavioral problems, and a child math/reading test score at age five or six. We also explore the relative predictive importance of birth weight, compared to a contemporaneous aggregate assessment of the family child-rearing environment at age five or six. Strikingly, birth outputs have virtually no structural/casual effects on early childhood developmental outcomes, and only maternal smoking and drinking during pregnancy have some effects on child height. Not surprisingly, family child-rearing environment has sizeable negative and positive effects on behavioral problems index and math/reading test score, respectively, and a mildly surprising negative effect on child height. Despite little evidence of a structural/causal effect of birth weight on early childhood outcomes, our results demonstrate that birth weight nonetheless has strong predictive effects on early childhood outcomes. Furthermore, these effects are largely invariant to whether family child-rearing environment is taken into account. Family child-rearing environment has both structural and predictive effects on early childhood outcomes, but they are largely orthogonal and in addition to the effects of birth weight.