Indian children are among the shortest in the world, and the country’s preference for sons might be to blame.

Globally, one in four children under the age of five is stunted – that is, they grew at a slower rate than a healthy child would. This stunting is manifest in shorter than average height. About half the stunted children live in Asia and another one-third live in Africa. India has the fifth-highest stunting rate in the world – nearly 40% of the children were stunted in 2005. This is a worrying proportion, even if you didn’t know that by 2020 India is projected to have the world’s youngest population.

Shorter children are not just less healthy, but also perform poorly in cognitive tests and, thus, eventually earn less.

Asian engima

Economic growth in West has been associated with rapid growth in height. For instance, the Dutch are, on average, nearly 20 cm taller than they were only 200 years ago. But economic growth and height haven’t increased in tandem in Africa and Asia.

India’s economic growth exceeded 6% per year, between 1992 and 2005, yet stunting declined by just 0.6% per year. But while India performs better on most health and economic factors, such as life expectancy, food security and educational attainment, when compared to Sub-Saharan Africa, Indian children are, on average, shorter than African ones.

Previous work has suggested that widespread practice of open defecation in India is partly to be blamed for this height discrepancy. This happens because disease-causing bacteria, released in the open because of this practice, find an easier route into the guts of children. Height is highly heritable, so another reason may be that Indian genes are biased towards shorter heights.

However, Northwestern University’s Seema Jayachandran and Harvard University’s Rohini Pande thought that it can’t be the full explanation behind the discrepancy. In a working paper published by the American non-profit, National Bureau of Economic Research, Jayachandran and Pande find that stunting in Indian children could also be blamed on the cultural preference for sons.

In India, on average, the first child – if he is a son – doesn’t suffer from stunting. But, if the first – and so the eldest – child of the family is a girl, she suffers from a height deficit. And, then, if the second child is a boy, and hence the eldest son of the family, he will not be stunted. This happens, according to Jayachandran and Pande, because of an “unequal allocation of resources to the first child”.

India has one of the worst gender ratios in the world. The reasons behind preference for a male child are economic, cultural and religious. Males bring in more money into the family (mostly because females are given few opportunities to work).

Boy child is king, girl child is nothing

The height not only varies between sons and daughters, but also tends to differ according to the order of the birth. Basically, the first is the tallest; the second is shorter than the first, but taller than the third, and so on.

The birth order dependency of a child’s height also downplays the effects that genes may have. If genes were causing children to be shorter, then all children within a family would suffer – and not the ones lower in the birth order.

When Jayachandran and Pande compared India and Africa results through this lens, they found that the Indian first and eldest son tends to be taller than an African firstborn. If the eldest child of the family is a girl, and a son is born next, the son will still be taller in India than Africa.

For girls, however, the India-Africa height deficit is large. It is the largest for daughters with no older brothers, probably because repeated attempts to have a son takes a beating on the growth of the girls.

While the study is yet to peer-reviewed – that is, checked by experts in the same field for validity of results – the authors believe “the results are robust.” They are based on data from extensive health and demographic surveys on more than 174,00 children in 25 sub-Saharan countries and India.

This article was originally published on qz.com.