Civil conflict, gender-specific fetal loss, and selection: a new test of the Trivers-Willard hypothesis.

A sizeable economics literature explores the effect of prenatal shocks on later health or socioeconomic status. Work in other disciplines, following the seminal contribution of Trivers and Willard (1973), suggests that prenatal shocks may increase fetal loss and reduce the number of boys relative to girls at birth. This has been largely ignored in the economics literature and could affect the interpretation of estimates of the effect of prenatal shocks and that of gender in other applied economics contexts. This paper analyzes the effect of in utero exposure to a shock - civil conflict in Nepal - on (i) fetal loss, and (ii) gender and (iii) health at birth. Maternal fixed effects estimates show that exposed pregnancies are more likely to result in a miscarriage and in a female birth, but exposed newborns are neither smaller nor more subject to neonatal mortality.