The Research Detective — The Great Indian Height Puzzle and missing 21 million girls
Seema Jayachandran recent op-ed and related research on India’s missing girls and son preference is a must read for all interested in social science research.
Related:
Rohini Pande and I found that eldest son preference helps explain India’s unexpectedly high rate of stunting among children under age five.3 Height is a commonly used measure of child well-being, not because we care about height per se, but because height reflects the nutrition and illness a child has experienced during the critical early years of life. Childhood stunting has been linked to low cognitive ability, poor health, and low earnings later in life. India is a poor country, so it is unsurprising that stunting is more prevalent than in a rich country such as the United States, but researchers have been puzzled by the high level of stunting in India compared to other countries at a similar level of economic development. Pande and I compare India to a set of sub-Saharan African countries which, though they are poorer than India, have lower levels of stunting.
We find that child height varies considerably among siblings, and specifically that there is a strong drop-off with birth order. This birth order gradient is observed in almost all societies, but it is especially strong in India. Firstborn Indian children are no shorter than their sub-Saharan African counterparts. The India height puzzle is mostly concentrated among later-born children, as shown in Figure 1.
How do we detect this “less wantedness” or “unwantedness”? Here is a common pattern of childbearing. A couple wants to have two children, ideally one son and one daughter, but it’s especially important to them to have at least one son. If they have two daughters in a row, they will keep having children until they get a son. (Meanwhile, if they have two sons in a row, they might regret having no daughter but not enough to expand their family.) A son might arrive on the third birth, and their children will be girl, girl, then boy.
It might take two tries: Girl, girl, girl, then boy. Notice that in both cases, the last child in the family is a boy. If we aggregate all families, we’d notice that the sex ratio of the last child (SRLC) is male-skewed. SRLC is thus a revealing measure of parents wanting sons. A subtle but important point is that these fertility “stopping rules” do not skew the populations’ overall sex ratio. I used the SRLC measure in my research to show that the fervent desire for sons in India is not a feature of all less economically developed societies. For example, in the historical US, there wasn’t a male-skewed SRLC.
The Economic Survey built on this work and took it further. Its analysis revealed that even Kerala and Assam have a male-skewed SRLC; if we only tracked missing women, these states would look problem-free. Importantly, the report also calculated the India-wide total of 21 million. In the figure below, taken from the report, the right panel shows that last children are disproportionately male. The left panel shows that non-last children are more female; that’s because the child being female led the parents to keep having children in their quest for a son. The 21 million unwanted girls can be calculated using the left panel: They form the gap between what would occur naturally — the dashed horizontal line at a sex ratio of 1.05 — and what we actually see.
3. The Roots of Gender Inequality in Developing Countries
