A growing literature suggests that stressful events in pregnancy can have negative effects on birth outcomes. Some of the estimates in this literature may be affected by small samples, omitted variables, endogenous mobility in response to disasters, and errors in the measurement of gestation, as well as by a mechanical correlation between longer gestation and the probability of having been exposed. We use millions of individual birth records to examine the effects of exposure to hurricanes during pregnancy, and the sensitivity of the estimates to these econometric problems. We find that exposure to a hurricane during pregnancy increases the probability of abnormal conditions of the newborn such as being on a ventilator more than 30min and meconium aspiration syndrome (MAS). Although we are able to reproduce previous estimates of effects on birth weight and gestation, our results suggest that measured effects of stressful events on these outcomes are sensitive to specification and it is preferable to use more sensitive indicators of newborn health.
Copyright © 2013 Elsevier B.V. All rights reserved.