Measuring Malnutrition and Dietary Diversity: Theory and Evidence from India
OLDIGES, CHRISTIAN (2016). 'Measuring Malnutrition and Dietary Diversity: Theory and Evidence from India' Paper presented at the annual conference of the HDCA, Tokyo 2016.
“[H]ealth is among the most important conditions of human life and a critically significant constituent of human capabilities which we have reason to value” (Sen, 2002). Within the capability space of health, being well-nourished to enjoy a life free of hunger and starvation is certainly the most basic functioning.
There is widespread consensus that merely meeting standardised calorie norms, as set for example by the FAO, does not translate into adequate nutrition. Instead, a diverse diet is favoured and necessary. Studies on nutrition in India show that diets become more diverse with increasing income levels (Deaton and Drèze, 2009) and studies for Europe, the United States and China reveal similar preferences. Besides the intrinsic value of a diverse diet, there is ample evidence for the functional link between a diverse diet and health outcomes, and between a diverse diet and economic performance.
A widely used method to capture the simultaneous consumption of food groups is “a simple count of food groups over a given reference period” (Ruel, 2002). This can be summarised in the dietary diversity index (DDI ). The latter sets a minimum number of food groups for a diet to qualify as nutritionally diverse. In this way, the DDI identifies individuals consuming a diverse diet or not.
However, there are several drawbacks with such an approach. For one, the DDI reflects merely the incidence of the malnourished and neglects the extent of malnourishment. In doing so, the DDI treats the absence of a diverse diet in such a way that the extent of nutritional deprivation is not accounted for. For instance, individuals consuming very few diverse food groups are considered equally deprived as those consuming just below the required minimum. The second weakness relates to minimum requirements of a food group. By not considering the quantity consumed of a food group but merely counting incidences, the DDI may underestimate the number of malnourished individuals. The third weakness is related to the previous one. The DDI neglects idiosyncratic variations in food requirements. While every person is assumed to be in need of a diverse diet, the extent of minimum requirements varies greatly by age, gender, health status, and occupation besides other factors. Therefore, a dietary measure would ideally apply person-specific thresholds for each food group.
In this paper, I develop a framework for a Nutritional Deprivation Index (NDI) to measure access to diverse diets using individual level data. An alternative framework is also defined for cases when only household level data are available. I apply this framework to compute an NDI using household level data on food consumption from India’s National Sample Survey (NSS) for the year 2011-12.
The NDI overcomes the three weaknesses of the DDI by adapting and extending the Alkire-Foster counting approach (Alkire and Foster, 2011), which is a technique widely used in multidimensional poverty measurement. The NDI addresses the first two weaknesses of the DDI by accounting for the actual amount consumed of each food group as well as the number of under-consumed food groups. By doing so, the framework yields both the incidence and intensity of nutritional deprivation. The absence of idiosyncratic thresholds in the DDI is also addressed by the NDI. It allows for minimum food group requirements which vary by food group as well as by individual characteristics such as age, gender, and occupation.
The NDI is more effective in identifying the nutritionally deprived than existing measures because it simultaneously reveals the incidence, extent and kind of food deprivation. This is demonstrated in the paper by applying the DDI and NDI framework to the NSS data on food consumption. The analysis using the DDI approach finds that one per cent of India’s rural population are deprived in at least five of eight food groups. On the other hand, the NDI yields that more than 30 per cent are nutritionally deprived. Further, the NDI highlights that the nutritionally deprived are primarily deprived of leafy vegetables, fruits, and dairy products. Finally, the NDI reveals that the average intensity of nutritional deprivation amongst those experiencing the lack of diverse diets is nearly 70 per cent. Decomposing the NDI by Indian states highlights considerable variation in the kinds of food deprivation. For instance, in the northern state of Punjab, nutritional deprivation in cereals contributes to some extent to overall nutritional deprivation. In the most populous state of Uttar Pradesh, however, cereals are sufficiently consumed while the consumption of leafy vegetables and fruits is insufficient. Further decompositions by social groups such as caste and religion shed further light on the kinds of nutritional deprivation.
In this manner, the NDI addresses the gaps in existing measures and proves to be a more accurate tool to quantify access to diverse diets. It is a step forward in measuring the most basic functioning of human well-being in the capability space of health.