This post is on
Why Do Mothers Breastfeed Girls Less Than Boys? Evidence and Implications for Child Health in India by
Seema Jayachandran (Stanford University) and
Ilyana Kuziemko (Princeton University), NBER Working Paper No. 15041,
http://www.nber.org/papers/w15041, June 2009
Seema Jayachandran and Ilyana Kuziemko’s answer to the question in their title is that (a) breastfeeding acts as a natural contraceptive, (b) Indian mothers are aware of this, (c) they have a preference for male babies, and, therefore, (d) after the (somewhat disappointing) birth of a female baby they cut the nursing period short in order to try again ASAP.
The authors do not speculate on the reasons for Indian mothers’ preference for male babies; they are empiricists par excellence and their data—from the
National Family Health Survey (NFHS) in India—do not allow them to go any farther.
The existence of this preference for male babies did not surprise me. What did strike me as valuable about this paper is that it is a brilliant demonstration of how wide the reach of economics can be, how exciting good empirical work in economics can be, and how important the policy implications of such work often is.
First, let us look at how Jayachandran and Kuziemko sleuthed out the answer to their title question. The key to figuring out the reason behind a phenomenon (such as the observed gender disparity in breastfeeding) is variation. If the phenomenon under study is found to exist to the same extent in all manner of situations, we’ve hit a brick wall; its causes cannot be deduced. But if the phenomenon exists strongly in some situations and not so strongly (or perhaps not all) in other situations, then we have in our sights an explanation for the phenomenon.
First, J and K find that higher birth-order children—that is, babies with more siblings that are older—are breastfed more. There is no known biological reason why this should be so; it looks like a deliberate choice. But what’s behind this deliberate choice? J and K conclude reasonably that a mother who already has had many children and has just given birth to another, is more likely to have reached or exceeded her desired number of children and would therefore be less inclined to cut breastfeeding short in order to have yet another baby. In other words, the observed fact that higher birth-order children are breastfed more tells us something important: mothers keep in mind the contraceptive properties of breastfeeding and adjust their breastfeeding behavior according to their personal goals.
Second, J and K find that “daughters are weaned sooner than sons”. What can explain this? We have already seen that Indian mothers regulate weaning to attain a desired family size. So, it is not a stretch to conclude that they wean female babies quicker in order to attain a higher proportion of male children.
Third, “for both sons and daughters, having few or no older brothers results in earlier weaning”. If you are a newborn baby—male or female—and you have no older brothers, you are out of luck in the breastfeeding department. Your mom is not quite done having boys yet. Therefore, she will cut off your supply soon so that she can quickly try to get pregnant again; breastfeeding is nature’s contraceptive, remember.
Finally, Jayachandran and Kuziemko find that gender disparity in breastfeeding has an inverted-U shape when graphed against family size. That is, gender disparity in breastfeeding is low for small families and for large families, and peaks in the middle, close to each mother’s desired family size. What does this pattern of variation tell us?
To an Indian mother, the number of male children matters a lot, but the overall number of children—both male and female—matters too. For a mother who has just given birth to her first child, her desire to have another baby will not depend much on whether her newborn is a boy or a girl. Therefore, her breastfeeding of her first-born will be largely independent of the baby’s gender.
Now assume that this mother’s desired number of children is three and she has just given birth to her second daughter. She will be desperate to try quickly for a boy, and will, therefore, cut the breastfeeding short. Had she given birth to a boy instead, there would have been less of an urge for another baby, and the breastfeeding would have gone on for a while longer.
Finally, if our hypothetical mother—who, remember, would ideally like to have three children—has just given birth to her fourth child, she would have little urgency to cut breastfeeding short to get pregnant again, irrespective of the gender of the newborn.
In this way, J and K are led by the observed variation in breastfeeding behavior across family size and family sex composition to infer the following theory of fertility choice: an Indian mother cares about how many children she has and how many of those children are male, and she will adjust breastfeeding according to the newborn’s birth order, the newborn’s sex, and the sexes of her other children in order to achieve her ideal family.
Okay, so we now know why Indian mothers breastfeed girls less than boys. But why should we care about this?
Apart from the well known health benefits of breast milk compared to its substitutes, “breastfeeding is thought to have greater health benefits for infants who would otherwise consume unsafe drinking water and contaminated food”. J and K argue that the gender disparity in breastfeeding accounts for 14% of the gender gap in mortality rates of Indian children. Tellingly, J and K find that the excess mortality of girls is small for firstborns and for the first six months of a baby’s life and those are exactly when the gender disparity in breastfeeding is small as well. Moreover, J and K find that most of the excess mortality of female children is in families with no access to piped water. In other words, the preference for male babies has a cost: female babies are being denied breast milk and are instead being fed infected water that is killing them. In the measured tones of the best academic research, Jayachandran and Kuziemko write: “Back-of-the-envelope calculations suggest that breastfeeding accounts for 14 percent of the gender gap in child mortality (deaths between ages one and five) in India, or 22,000 missing girls each year.”
This is as good a time as any to remind readers of the economist Amartya Sen’s famous 1990 article in The New York Review of Books, “
More Than 100 Million Women Are Missing”. Sen had quantified the extent of this grisly excess mortality of women and speculated that its roots lay in women being denied basic access to the resources that are essential for life. J and K provide a slightly nuanced explanation for (14 percent of) the problem identified by Sen: “Son preference is the underlying cause of this excess female mortality, but in a subtle way: Rather than resulting from parents' explicit decisions to allocate more resources to sons, the missing girls are mainly an unintended consequence of parents' desire to have more future sons.” (Please see
this Wikipedia entry on the whole missing women debate.)
Okay, now that we have seen why Indian mothers breastfeed girls less than boys and why it matters, what if anything should we do about all this? What, in other words, are the policy implications of the findings of Jayachandran and Kuziemko?
Perhaps Indian mothers are not entirely well informed about the effects that their breastfeeding behavior can have on their female children. (Actually, I probably should have said this with greater force and certainty because the alternative is the ghoulish idea that Indian mothers are knowingly and deliberately hobbling the lives of their daughters.) In that case an education or advertising campaign would be an obvious solution to try. Other equally obvious remedies to try would be to expand access to clean water (so that weaned babies have healthy alternatives to mothers’ milk) and to provide better medical care for newborns in general and female babies in particular.
A campaign to reduce total fertility may also be in order. The more children a mother seeks to give birth to during her finite reproductive period, the more inclined she may be to cut short the breastfeeding of her children, with all the attendant health problems discussed above.
In case some women are giving birth to more children than they want, expanded access to artificial contraceptives—perhaps through subsidies—may help. But one needs to watch out for the law of unintended consequences. Some women who breastfeed their babies because breastfeeding is nature’s contraceptive may reduce breastfeeding when artificial contraceptives become available. This could worsen the health of children. So, Jayachandran and Kuziemko dutifully attach a warning label to their recommendation: if you expand access to synthetic contraceptives, be sure to expand access to piped water and medical care at the same time, because breastfeeding might decrease.
This is a really nice paper and I am glad I took the time to read it.
One final thought: How many people do you think know that economics may just as easily be about breastfeeding behavior as about the gross domestic product, the inflation rate, or the Dow Jones Industrial Average? Economics is the study of what rational people choose to do when faced with trade-offs and the overall consequences of those choices. Those choices could be involve deciding what stocks and bonds to buy with one’s savings, or how much time to devote to sleep, or when to stop breastfeeding a newborn.
