VoxEU Column Gender Health Economics

Son preference drives India’s high child malnutrition rates

Indian children are more likely to be malnourished than their counterparts in Sub-Saharan Africa, despite higher standards of living. This column uses data on child height – an anthropometric measure of net nutrition – from Africa and India to explore how parental gender preferences affect the likelihood of children being malnourished. Indian firstborn sons are found to have a height advantage over African firstborn sons, and the height disadvantage appears first in second-born children, increasing for subsequent births. This suggests that a preference for a healthy male heir influences fertility decisions and how parents allocate resources between their children.

It has long puzzled researchers that when you look at child height – the best anthropometric measure of net nutrition – the average child born in India is more likely to be stunted than her counterpart in Sub-Saharan Africa, even though her mother is more likely to survive her birth, and her parents are probably richer and more educated (Gwatkin et al 2007). In 2005, 40% of Indian children under the age of five remained stunted (IIPS 2010). Thus, despite a GDP per capita that is higher than over 60 countries’, India has the fifth-highest stunting rate in the world (UNICEF, 2013).

Ramalingaswami et al (1997) dubbed this apparent contradiction ‘The South Asian enigma,’ and explanations offered since have included South Asia’s breastfeeding and weaning conventions, its use of well water, its poorer treatment of women, and health problems associated with open defecation – which is more widespread in India than Africa and can cause children to suffer malnutrition even when they are well fed (Ramalingaswami et al 1996, Smith et al 2003, Spears 2013).

Yet these explanations fail to account for one important fact – Indian firstborns are taller than African firstborns. In a working paper, we analyse data on over 174,000 children from 25 Sub-Saharan African countries and India, drawn from recent Demographic and Health Surveys (DHS), and reveal that the Indian height disadvantage emerges with the second child and then increases with birth order. We propose that a preference for eldest sons in India (encompassing both a desire to have at least one son and for the eldest son to be healthy) leads families to dramatically reduce resources devoted to mothers and children over successive pregnancies, and that this drives the country’s high malnutrition rates.

The evidence for this comes by comparing the height of children of different birth orders in India and Africa. The Indian height disadvantage appears for second-born children and increases for third and higher order births, at which point Indian children have a mean height-for-age lower than that of African children by 0.3 standard deviations of the worldwide distribution. Figure 1 shows this in terms of the stunting rate, using the World Health Organization’s definition of having child height-for-age that is 2 standard deviations or more below the worldwide reference population median for one’s gender and age in months (WHO 2006). We see the same pattern – a much steeper birth order gradient in child height in India than in Africa – when looking only at children within the same family. Thus, the effect cannot arise from wealth or other differences in background of smaller versus larger families.

Figure 1.

We also look at an array of health inputs – prenatal and postnatal check-ups, iron supplements, vaccinations – and as with height, we observe a steeper drop-off with birth order in India than Africa. The same holds true for children’s blood haemoglobin levels.

Boys over girls

While Indian children of both genders exhibit a sharper drop-off in height relative to African children, it is only among boys that we observe the advantage in height for Indian firstborns. Consistent with this fact, averaged across birth orders, the Indian height deficit only holds among girls. This suggests that prenatal investments are made in the form of nutrition and medical care as long as the parents believe the child might be a boy. When a girl’s gender is revealed at birth, these investments drop off, and so does the height advantage she might have built up in the womb.

Furthermore, eldest son preference will generate the observed birth order gradient among boys simply because a lower birth order son is more likely to be the family’s first son. Consistent with this argument, a son born at birth order 2 is taller in India than Africa if and only if he is the family’s eldest son.

These patterns suggest that the prevalence of malnutrition in India is not an artefact of using child height to measure malnutrition, in which case low child height in India would simply reflect genetics. Genotypes do not vary with birth order or siblings’ gender, so a genetic predisposition to be short would not cause the effects that we see. Other health- and environment-related factors that distinguish India from Africa such as India’s worse maternal health and worse sanitation are potentially important in explaining the shortfall in height that is common across all Indian children but fail to account for the observed birth order patterns. Finally, the reason Indian later-borns are so malnourished cannot be that family income declines over the lifecycle in India relative to Africa. We find that it is only among pregnant women that women’s health and nutrition in India and Africa have different time profiles; Indian women do relatively worse as family size grows, but only when they are pregnant, whereas declines in family income would be evident even when women are not pregnant.

It appears that families allocate inordinate resources – nutritious foods, iron supplements, tetanus shots and prenatal check-ups – to a pregnant woman as long as there is a possibility that she is carrying the family’s firstborn son. Once a male heir is born, prenatal investments drop off.

Sibling rivalry, but with a prospective sibling

Among girls, eldest son preference generates a birth order gradient through a more subtle mechanism, namely fertility stopping rules. Indian parents who start off only having daughters are likely to keep trying for a son, and in the process exceed the number of children they would otherwise have desired. Household resources dwindle and a later pregnancy is less well-resourced. If the later pregnancy yields yet another daughter, then she can receive even fewer resources than her older sisters because her parents have revised their fertility plans to keep trying for a son. Consistent with this, the India-Africa height gap is particularly large for daughters who only have girls as elder siblings.

Economists are familiar with the notion of ‘sibling rivalry,’ according to which siblings compete for household resources and girls who do not have a brother will show improved outcomes (Garg and Morduch 1998). However, the pattern we observe stands in contrast to that prediction – in a household with only daughters, these girls will not benefit from being brother-less because their parents are keeping back resources as they keep trying for a male heir.

Religious and regional differences

When we leave aside comparisons with Africa to focus on patterns within India, we find more evidence supporting our interpretation that eldest son preference drives the country’s high malnutrition rates. First, the height gradient is absent in Kerala, an Indian state with strong matrilineal traditions. Second, on average, Hindu children are taller than Muslim children, but not by the amount one would expect given Hindus’ higher income. Figure 2 compares stunting rates between Indian Hindus and Muslims over birth order, and shows that Hindu firstborns are markedly taller than Muslim firstborns. It appears that Hindu parents seem to invest so little in their later-born children’s health that, overall, they only show modest overall height advantage. For third and subsequent births, Hindu children fare worse than Muslim children in terms of height.

Figure 2.

The difference may be due to religious as well as cultural norms. Hinduism prescribes a patrilocal and patrilineal kinship system – meaning, ageing parents live with their son, typically the eldest, and bequeath property to him (Dyson and Moore 1983, Gupta 1987). Also, Hindu religious texts stipulate that only a male heir perform certain post-death rituals, such as lighting the funeral pyre, taking the ashes to the Ganges River, and organising death anniversary ceremonies (Arnold et al 1998).

One might expect these household inequalities to reduce as India develops. The hope is that with greater financial resources, all children might be well nourished enough to achieve their height potential. However, when we compare households by wealth, the Indian birth order gradient in height is actually relatively larger among wealthier households. This echoes other forms of gender inequality that resist vanishing with development in South Asia as they do in other parts of the world. Richer and more educated Indian women are less likely to work, and richer families are more likely to practice sex-selective abortion. Indeed, the problem of malnutrition already seems resistant to the forces of development – between 1992 and 2005 India’s economic growth exceeded 6% per year, yet stunting declined by just 1.3% annually (Tarozzi 2012).

The beneficial forces of prosperity are strong, but social and religious norms may well be stronger. Unless policies are put in place to counteract the social and economic forces that support son preference, Indian families may continue to have more children than their ideal number, and Indian daughters may continue to receive less than their fair share.


Arnold, F, M K Choe and T K Roy (1998) “Son preference, the family-building process and child mortality in India,” Population Studies 52: 301–315.

Dyson, T and M Moore (1983) “On kinship structure, female autonomy, and demographic behavior in India”, Population and Development Review, 9(1): 35–60.

Garg, A and J Morduch (1998) “Sibling rivalry and the gender gap: Evidence from child health outcomes in Ghana”, Journal of Population Economics, 11(4): 47–493.

Gupta, M D (1987) “Selective discrimination against female children in rural Punjab, India”, Population and Development Review, 13(1): 77–100.

Gwatkin, D, S Rutstein, K Johnson, E Suliman and A Wagstaff (2007) “Socio-economic differences in health, nutrition and population within developing countries: An overview”, Health, Nutrition and Population Network Informal Paper Series, World Bank.

IIPS (2010) District Level Household and Facility Survey (DLHS-3), 2007-08, International Institute for Population Sciences, Mumbai.

Jayachandran, S and R Pande (2015) “Why are Indian children so short?”, CEPR Discussion Paper 10503.

Ramalingaswami, V, U Jonsson,and J Rohde (1997) Malnutrition in South Asia: A regional profile, S Gillespie, Ed., UNICEF Regional Office for South Asia, Kathmandu, pp. 11–21.

Ramalingaswami, V, U Jonsson, and J Rohde (1996) “Commentary: the Asian enigma”, The Progress of Nations, UNICEF.

Smith, L C, U Ramakrishnan, A Ndiaye, L Haddad and R Martorell (2003) The importance of women’s status for child nutrition in developing countries, IFPRI, 131, chap. 7.

Spears, D (2013) “How much international variation in child height can sanitation explain?”, Mimeo, Princeton University.

Tarozzi, A (2008) “Growth reference charts and the nutritional status of Indian children”, Economics and Human Biology, 6(3): 455–468.

UNICEF (2013), Improving Child Nutrition: The Achievable Imperative for Global Progress, UNICEF, New York.

WHO (2006), Child Growth Standards: Methods and Development, World Health Organization, Geneva.

2,310 Reads