Maternal Fasting and Pregnancy Outcomes

Put this in my folder of "cool identification and interesting questions". Douglas Almond and Bhashkar Mazumder attempt to measure how important in utero health is on observed health outcomes upon birth and adulthood by using the Muslim holiday, Ramadan, as a natural experiment. During Ramadan, Muslims fast for several days (including women, and apparently, including pregnant women). If fetal health is an outcome produced by some underlying health production function which takes maternal food as inputs, then we might expect the holiday to lower fetal health, all things equal. Almond and Mazumder examine using different data and different outcomes what happens to child birthweight and gestation length, as well as using Census data adult outcomes. Here's what they find.

Abstract

We use the Islamic holy month of Ramadan as a natural experiment for evaluating the short and long-term effects of fasting during pregnancy. Using Michigan natality data we show that in utero exposure to Ramadan among Arab births results in lower birthweight and reduced gestation length. Preconception exposure to Ramadan is also associated with fewer male births. Using Census data in Uganda we also find that Muslims who were born nine months after Ramadan are 22 percent (p =0.02) more likely to be disabled as adults. Effects are found for vision, hearing, and especially for mental (or learning) disabilities. This may reflect the persistent effect of disruptions to early fetal development. We find no evidence that negative selection in conceptions during Ramadan accounts for our results. Nevertheless, caution in interpreting these results is warranted until our findings are corroborated in other settings.

Very interesting. They appear to find the effects in most places you'd expect, and not just in one place. For instance, they find a reduction in male births. This is a first order effect of reduced fetal health, but not one you may think of. The body produces statistically more males than females, but males also have a slightly lower survival rate (hence probably why more are produced). If there is a common shock to all pregnant women experienced by Ramadan fasting, then it should lower the fetal survival odds of children, but it will lower it slightly more for males than females since more males will be at that survival margin. More specifically, their results suggest fasting: (i) lowers birthweight, (ii) increases premature babies, (iii) increases miscarriages (that's my interpretation of the fewer male births), (iv) and increases vision, hearing and mental/learning disabilities among adults conceived during Ramadan. That last result is going to be the toughest to sell, as the baby is exposed to fasting during any of the nine months, but they only focus on people born nine months after Ramadan. Perhaps they use a regression discontinuity to examine both those born 10 months after Ramadan and 9 months, as someone born 10 months after Ramadan would not have gotten the same early fetal treatment of reduced fetal health inputs as the one born 9 months after. Their interpretation of the result is that early inputs matter more than later inputs for reasons similar to "compounding interest" in development.

All in all, this looks like a fascinating and important contribution. If you needed some evidence that fetal health inputs mattered for survival and both short and longrun health outcomes, this one should give you a run for your money.