The child malnutrition myth in India Arvind Panagariya Oct 1, 2011, 12.00AM IST In the early 2000s, when the 55th (1999-2000) round of the expenditure survey showed a surprisingly sharp decline in poverty over its predecessor survey, the reform critics descended on the finding like a ton of bricks. Their critique eventually led to a healthy debate, important new research and eventual downward revision in poverty reduction numbers by the reform advocates themselves. In total contrast, almost no objections have been raised to the absurdly high estimates of malnutrition in India trumpeted by journalists, NGOs, politicians and international institutions within and outside India. Not a day goes by without some TV channel or newspaper running the headline that the world's fastest growing economy suffers worse malnutrition than sub-Saharan Africa (SSA). In terms of vital statistics such as life expectancy at birth, infant mortality and maternal mortality, India fares better than all except one or two of the SSA countries with comparable or lower per capita incomes. So it is puzzling that, according to World Health Organisation (WHO) statistics, it suffers from higher proportion of underweight children than every one of the 48 SSA countries and higher rate of stunting than all but seven of them. Such countries as the Central African Republic, Chad and Lesotho, which have life expectancy at birth of just 48 years compared with India's 65, have lower rates of stunting and underweight. If you still do not believe the absurdity of these malnutrition numbers, compare Kerala and Senegal. Kerala exhibits vital statistics edging towards those in the developed countries: life expec-tancy of 74 years, infant mortality rate of 12 per 1,000 live births and maternal mortality rate of 95 per 1,00,000 live births. The corresponding figures for Senegal are far worse at 62, 51 and 410, respectively. But nutrition statistics say that Kerala has 25% stunted children compared to 20% of Senegal and 23% underweight children relative to 14.5% of the latter. In Punjab, which has a life expec-tancy of 70 years and is the breadbasket and milk dairy of India, 37% of children are stunted and 25% underweight. To make sense of this nonsense, we must look at how the stunting (and underweight) rates are calculated. To classify a child of a given age and sex as stunted, we must compare his height to a pre-specified standard. The WHO sets this standard. In the early 2000s, it collected a sample of 8,440 children representing a population of healthy breastfed infants and young children in Brazil, Ghana, India, Norway, Oman and the United States. This "reference" population provided the basis for setting the standards. As expected, when comparing children of a given age and sex even within this healthy sample, heights and weights differed. Therefore, some criterion was required to identify stunting and underweight among these children. In each group defined by age and sex, the WHO defined the bottom 2.14% of the children according to height as stunted. The height of the child at 2.14 percentile then became the standard against which children of the same age and sex in other populations were to be compared to identify stunting. A similar procedure applied to weight. The key assumption underlying this methodology is that if properly nourished , all child populations would produce outcomes similar to the WHO reference population with just 2.14% of the children at the bottom stunted and underweight. Higher rates of stunting would indicate above normal malnutrition. So the million-dollar question is whether this assumption really holds for the population of children from which the estimate of half of Indian children being stunted is derived? As it happens, the answer to the question can be found buried in a 2009 study published by the government of India. The latest estimate for stunting in India has been derived from the third National Family Health Survey (NFHS-3). The report draws a highly restricted sample from the fuller NFHS-3 sample consisting of 'elite' children defined as those 'whose mothers and fathers have secondary or higher education, who live in households with electricity, a refrigerator, a TV and an automobile or truck, who did not have diarrhoea or a cough or fever in the two weeks preceding the survey, who were exclusively breastfed if they were less than five months old, and who received complementary foods if they were at least five months old'. If the assumption that proper nutrition guarantees the same outcome as the WHO reference population is true, the proportion of stunted children in this sample should be 2.14%. But the study reports this proportion to be above 15%! The assumption is violated by a wide margin. The implication of this and other facts is that Indian children are genetically smaller on average. A competing hypothesis - which says that nutrition improvements may take several generations - fails to explain how, without a genetic advantage, the far poorer SSA countries, which lag behind India in almost all vital statistics, could have pulled so far ahead of India in child nutrition. Moreover, the trend of the stunting proportions based on WHO standards, available for India since the late 1970s, would suggest that nearly all those born in the 1950s or before - the writer included - are stunted! Either way, the statistic that half of Indian children today are stunted needs to be viewed far more sceptically and investigated more deeply. The right treatment requires a right diagnosis. The writer is a professor at Columbia University.