摘要：We estimate the harm from smoking during pregnancy upon child birth outcomes,
using a rich dataset on a cohort of mothers and their births. We exploit a fixed
effects approach to disentangle the correlation between smoking and birth weight
from the causal effect. We find that, despite a detailed set of controls for
maternal traits, around one-third of the harm from smoking is explained by
unobservable traits of the mother. Smoking tends to reduce birth weight by 1.7%,
but has no significant effect on the probability of having a low birth weight
child, pre-term gestation or weeks of gestation. Exploring heterogeneity in the
effect on birth weight, it is mothers who smoke for the 9 months of gestation
that suffer the harm, whereas there is an insignificant effect for mothers who
chose to quit by month 5. Additionally, there is evidence of potential
complementarity in investment of human capital, as the impact on birth weight of
smoking is much greater for low educated mothers, even controlling for the
quantity of cigarettes they smoke. We suggest policy should target the low
educated mothers, offering a more holistic approach to improving child health,
as quitting smoking is only half of the battle.
