Global Hunger Index: A lesson in how not to measure hunger

21 Oct 2022

Opinion: Sonalde Desai:

Global Hunger Index is riddled with inadequate and poorly described data and a lack of conceptual clarity. The problem with indices of this type is that it directs governmental attention to cross-national comparisons, sometimes resulting in the rejection of underlying issues and sidetracking the public discourse

The world seems to have gone index-happy. Happiness Index, Ease of Doing Business Index, Calmness Index, you name it. The problem with this mini-industry of index creation is that it has the potential to both influence government action and generate aggressive pushback. The experience with the Global Hunger Index provides a salutary lesson in the latter.

Recently, Concern Worldwide released the Global Hunger Index. India ranked 107 out of 123 countries, dropping from the rank of 101 in 2021. The government has responded sharply to the publicity surrounding this, rejecting the methodology employed by the researchers and noting the substantial efforts made by the government to improve access to foodgrains by India’s poor. The rebuttal by the government rests on valid grounds — about a third of the index rests on the Food and Agricultural Organisation’s estimates of the proportion of undernourished in the population. Digging deep, we see that these estimates are based on Gallup World Poll’s survey of 3,000 households in India (and 1,000 households in smaller countries).

In addition to its small size, the Gallup sampling methodology does not follow the usual processes used in India. This suggests a need to evaluate the representativeness of the sample. Unfortunately, we cannot easily do this because the underlying data are located behind a paywall. To ensure transparency, it is essential that international agencies only use data that are freely available in the public domain along with key characteristics such as education, residence and age of the respondents. In this case, the uncritical use of questions is particularly problematic because FAO has not released standard errors for their estimates, making it difficult for us to evaluate whether the growth in the proportion of households experiencing hunger in India, from 14.8 per cent in 2013-15 to 16.3 per cent in 2019-21, is statistically significant. This is very important given the difficulties in collecting data during the pandemic.

However, quibbles about this one indicator obfuscate the larger question: Is this index genuinely measuring hunger, or is it lumping together various indicators with only a weak relationship with hunger? The index rests on four indicators: Proportion of undernourished in the population, under- five mortality rate, prevalence of stunting (low height-for-age) and wasting (low weight-for-height) in children under five. The last three indicators come from the National Family Health Survey for India. Proportion undernourished and child mortality contribute 1/3 each to the index, while stunting and wasting contribute 1/6 each.

How good are these indicators in picking up on hunger? While the first, if well collected, could presumably identify the proportion experiencing hunger, the latter three are only partially related to hunger.

Child mortality depends heavily on a country’s disease climate and public health systems. Today, 40 of 1,000 children in India die before their fifth birthday; 27 of these deaths occur in the first month of life. This suggests that many child deaths are associated with conditions surrounding birth, congenital conditions, or delivery complications. These are not necessarily markers of hunger.

Similarly, the relationship between stunting (low height-for-age), wasting (low weight-for-height), and hunger is not apparent. As UNICEF notes in an article titled ‘Stop Stunting’, poverty is not a clear cause of stunting as there are stunted children even among the wealthiest households. Various factors contribute to stunting, such as infant and child care practices, hygiene, dietary diversity and cultural practices surrounding maternal diet during pregnancy. Food insecurity contributes to child stunting, but its relative importance in determining stunting is not established. Wasting is associated with both recent illnesses and low food intake. The two are closely related; children suffering from diarrhoea are less likely to eat, and poor nutritional status makes them more susceptible to disease.

Thus, while all three indicators of child health are related to poor food intake, none of them is solely determined by hunger. Moreover, trends in all three reflect somewhat different patterns. Between 1998-99 and 2019-21, National Family Health Survey 2 and 5 show that the child mortality rate fell from 95 deaths per thousand to 40 per thousand. This is a significant improvement attributable to improved immunisation coverage and increased hospital delivery. Child stunting decline was also substantial, from 51.5 per cent to 35.5 per cent, possibly due to improved water and sanitation systems. Wasting has not changed, barely budging from 19.5 per cent to 19.3 per cent.

Despite this progress, we need to continue our effort to reduce child mortality and find ways of reducing stunting and wasting. However, whether the reduction in hunger is either necessary or sufficient to improve nutrition remains unclear.

In an intriguing article published in 2009, Angus Deaton and Jean Dreze try to reconcile these puzzles and find that average caloric intake has severe limitations as a nutrition indicator. They argue that “close attention needs to be paid to other aspects of food deprivation, such as the intake of vitamins and minerals, fat consumption, the diversity of the diet, and breastfeeding practices.” My research using data from the India Human Development Survey, organised by the National Council of Applied Economic Research and the University of Maryland, supports this. This study found that holding household incomes constant, with access to the public distribution system, skewed consumption towards cereals, reduced dietary diversity, and failed to improve anthropometric outcomes.

Thus, the Global Hunger Index is riddled with inadequate and poorly described data and a lack of conceptual clarity. The problem with indices of this type is that it directs governmental attention to cross-national comparisons, sometimes resulting in the rejection of underlying issues and sidetracking the public discourse. In a way, this episode illustrates the concern that Amartya Sen, one of the principal consultants to the Human Development Report, 1990, expressed. He has argued that concentrating too much on the Human Development Index or any other index would be a great mistake. The Global Hunger Index is one example in which the weapon has backfired, detracting attention from the very real challenges of improving nutrition and reducing child mortality.

The writer is a professor of Sociology at the University of Maryland and professor and Centre Director of the National Council of Applied Economic Research. Views are personal

Published in: The Indian Express, 21 Oct 2022