Estimating intergenerational returns to medical care: New evidence from at-risk newborns

We examine whether intensive early-life government-funded interventions targeted to at-risk newborns are transmitted intergenerationally. Using a regression discontinuity design and administrative data from Chile we follow women up to the age of 25, and document the surprising fact that children of individuals who were treated at birth have worse indicators of health at birth a generation later. We suggest this owes to selective fertility, finding that marginally treated individuals are substantially more likely to give birth. These new stylised facts suggest that in certain circumstances, the long-term implications of public investments within family lineages may be quite different to their short-term implications, placing more weight on necessary reinforcing interventions. • Intensive neonatal care reduces mortality but has negative intergenerational effects. • Treated mothers are 20% more likely to give birth, revealing a fertility channel. • Extensive-margin (fertility) and intensive-margin (health) intergenerational returns exist. • Negative transmission owes to selection: marginal births have weaker health stocks. • First-generation benefits substantially outweigh second-generation costs overall.