Here is the paradox that should alarm every development economist: India is richer than sub-Saharan Africa, yet Indian children are shorter, lighter, and more developmentally delayed. A child in India is 36 percent likely to be stunted—shorter than expected for their age, a marker of chronic malnutrition that dims cognitive capacity and earning potential across a lifetime. In sub-Saharan Africa, the figure is 34 percent. How can India, with greater per-capita wealth, produce smaller bodies?
The conventional answer has been poverty, insufficient nutrition programs, poor sanitation, lack of maternal education. These are real. But they explain only part of the damage. A study published this year in the Journal of Economics, Race, and Policy has uncovered what development economists and public health authorities have been missing: the "stunting paradox" dissolves the moment you disaggregate by caste.
Ashwini Deshpande and Rajesh Ramachandran, examining National Family Health Survey data from 2019–21 covering 200,000 children under five, found that forward-caste Hindu children in India experience a 27 percent stunting rate—placing them at lower risk than sub-Saharan African children. Children from Scheduled Castes and Scheduled Tribes, by contrast, are approximately 50 percent more likely to be stunted than forward-caste peers. The gaps between these groups within India dwarf the India–sub-Saharan Africa gap entirely.
This is not incidental. This is the visible truth of caste as embodied hierarchy: discrimination does not operate only through social exclusion or ritual pollution. It operates through the body itself. And that inscription begins at conception.
Untouchability—though declared illegal by the Constitution—functions through material deprivation so systematic it has become invisible to policymakers. In districts where untouchability practices are most widely reported, Dalit children's height disadvantage sharpens substantially. The mechanisms are documented. Dalit households have access to piped water at half the rate of upper-caste households. Health-care workers conduct fewer home visits to Dalit mothers with malnourished children. Dalit women face barriers to prenatal and postnatal care that leave their children born into biochemical deficit. And the most basic deprivation: in communities practicing food denial and social boycott of Dalit families—documented in IHDS-II survey data—children are systematically fed less, later, and with less nutritional density than upper-caste peers.
This is not poverty alone. Research by Sukhadeo Thorat has demonstrated that when you control for household income and maternal education, the caste stunting gap persists. Something independent of class is killing these children's potential. That something is discrimination embedded in water access, health worker refusal, social boycott, and the denial of basic dignity that compounds across pregnancy and infancy into permanently shortened stature.
It is worth pausing on what this means. A forward-caste child born into a poor household has a better chance at full physical development than a Dalit child born into a middle-class one. Caste is not a residual category to be controlled for. It is a direct mechanism of physical harm.
Economists and public health authorities did not arrive at this understanding by accident. For decades, the international development literature treated India's stunting as a function of generalized poverty, sanitation deficits, and female education levels—categories that, while real, happened to be color-blind to caste. The World Bank, the WHO, Indian Ministry of Health documents all circled the puzzle without naming its center. This was not deliberate dismissal. It was silence. Caste simply did not appear in frameworks otherwise sophisticated about socioeconomic gradients. The assumption, conscious or not, was that India's internal hierarchies were historical, that modern income and modern infrastructure would dissolve them. The body knew better.
B.R. Ambedkar understood this architecture long before biostatistics confirmed it. His concept of graded inequality—the idea that caste operates not as binary oppression but as a multi-layered hierarchy where each group simultaneously dominates and submits—explains precisely how discrimination persists even as formal legal equality takes hold. In a graded system, nearly every group enjoys some advantage over those beneath. This fractures potential for unified resistance. But it also means that benign policies—growth, sanitation infrastructure, vaccination programs—trickle downward through the hierarchy, leaving the lowest rungs still behind. A piped water scheme reaches the village but stops short of Dalit hamlets. A health worker makes a home visit but not to a Dalit house. A ration distribution reaches the food-secure first and the socially boycotted last, if at all. Graded inequality survives modernization because it is not primarily a feature of scarcity. It is a feature of systematic exclusion justified by hierarchy itself.
What makes the Deshpande-Ramachandran findings urgent is not that they reveal something unprecedented. It is that they make visible, through the irreducible fact of child height and weight, what decades of policy failure have obscured. You cannot argue with a centimeter. You cannot rationalize away a two-year-old who is the height of an eighteen-month-old. The stunting data has converted the question from philosophy to biology, from debate to diagnosis.
The diagnosis demands something more than universal programs. Analysis of untouchability practices shows that generalized improvements in sanitation, healthcare access, and nutrition subsidies do not automatically reach communities where discrimination operates. Food denial is illegal but continues. Water exclusion is illegal but continues. And the legal apparatus designed to protect Dalit communities—the Scheduled Castes and Scheduled Tribes (Prevention of Atrocities) Act—remains so underutilized in health contexts that it has created no deterrent. The institutions charged with child health improvement have not been made to see caste discrimination as their problem.
This is where India's 2027 Census, the first to enumerate caste since 1951, becomes a watershed. The 1951 decision to discontinue caste enumeration was meant to erase caste from the statistical record. It succeeded—which is precisely why development policy could ignore caste while touting poverty reduction. A census that counts caste, as 2027 will do, makes erasure impossible. It forces the state to see what it has been trained not to see.
But counting alone is not redemption. The data must become the foundation for policy that is ruthless about causation. Not programs that treat stunting as a poverty problem with a caste coincidence. Programs that treat untouchability as the stunting crisis and strip away the mechanisms of exclusion—the refusal of water, the rejection of health workers, the social boycott of food—with the same force that the state deploys against other forms of violence.
Ambedkar wrote that dignity cannot be conferred from above. It must be claimed and defended through autonomy, freedom from exploitation, and substantive access to the resources and care that sustain life. A child's height is not a luxury. It is evidence of whether a nation has granted its lowest members the freedom to exist without denial. The hidden divide reveals what we have failed to grant. The next phase—the harder one—is deciding whether to change.