Hepatitis B vaccination and neonatal mortality in the United States

We estimate the plausibly causal effect of hepatitis B birth dose vaccination on neonatal mortality using U.S. county-level panel data (2011–2019) and a shift-share instrument that interacts historical fertility rates with national vaccine price variation. With county fixed effects and a control for the historical fertility share interacted with a linear trend, a one percentage point increase in birth dose coverage reduces 0–3 day mortality by 0.022 deaths per 1,000 live births, representing a 1.04% decline at the mean. The effect is concentrated in the first day of life and is indistinguishable from zero at later infant ages, a profile consistent either with an early life mechanism or with an administrative one, and we identify neither. The estimate is an unweighted county average conditional on a linear fertility trend, and we report the diagnostics that qualify it alongside those that support it.