ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846
Reader Mode
-A A +A

Child Undernutrition in India

Assessment of Prevalence, Decline and Disparities

Sunny Jose ( and Bheemeshwar Reddy A (bheem@ teach at and Mayank Agrawal (f20150499@ is a graduate student at the Birla Institute of Technology and Science Pilani, Hyderabad.

Analysing the latest National Family Health Survey-4 (2015–16) data, an assessment of the prevalence and decline in child undernutrition in India between 2005–06 and 2015–16 is undertaken. Despite a moderate decline in child undernutrition during this period, more than one-third of children under five years are stunted and underweight. A large, graded socio-economic disparity in child undernutrition continues. Arunachal Pradesh, Himachal Pradesh, and Mizoram emerge as better performers in reducing child undernutrition. While north-eastern states have done well in reducing underweight prevalence, Tripura, Punjab, and Chhattisgarh have performed better in reducing stunting. About 80% of high stunting prevalence (above 40%) districts belong to eight states, that also house 90% of high underweight prevalence districts.

Figures 1 to 5 can be accessed from the online version of this paper.

The authors would like to thank the anonymous reviewer and Padmini Swaminathan for many insightful comments, and Kopal Khare and Prakar Gupta for assistance in data processing.

In the avoidance of endemic undernourishment and hunger, India has done worse than nearly every country in the world,” wrote Amartya Sen in 2001. Has the situation improved during the last decade? Specifically, has India made a decent progress in reducing child undernutrition during this period of robust economic growth? Have the socio-economic disparities in child undernutrition declined? Does the performance of states convey any specific, unique pattern of decline? Have the states with high undernutrition fared better during this period? Also, does the district-level data, available for the first time for 640 districts of India, reveal which districts remain most affected by child undernutrition and whether they are dispersed across or confined to a few states? We try to address these questions by analysing the unit-level data of the National Family Health Survey-4 (NFHS-4) (2015–16).

Decline in Child Undernutrition?

Not long ago, child undernutrition was endemic and high across South Asia, and was considered an enigma (Ramalingaswami et al 1996). This is not the case anymore. Bangladesh and Nepal have made significant strides in recent years in reducing child undernutrition (Nisbett et al 2017; Cunningham et al 2017). Sri Lanka, with an impressive performance, has become an island of exception. Only a few countries of South Asia are still lagging behind, including India. India alone housed 38% of the world’s stunted children in 2011 (UNICEF 2013: 9), and “there were more undernourished children in India than in all of Africa” (Headey 2013: 80). Hence, it was imperative for India to measure up to its well-faring neighbours. Such an improved performance, besides bettering the country’s position globally, would also help reduce the magnitude of un-freedom and illness for which undernutrition is a cause in India (Shekar et al 2016).

Is such an expectation of improved performance realistic for India given its poor record so far? It appears to be eminently possible, especially due to its better performance in economic growth. This raises an immediate question: Can economic growth help reduce child undernutrition? The answer appears to be mixed. Headey (2013) concludes from a cross-county analysis that 5.5% increase in per capita economic growth per year is likely to reduce stunting by 1 percentage point. By contrast, Subramanyam et al (2011) and Joe et al (2016) conclude that economic growth plays only a marginal role in reducing child undernutrition in India. These findings require of us to neither play up nor decry the role of economic growth in reducing child undernutrition.

Studies also identify the crucial role of a specific set of factors, especially enhancement in education and nutrition of women, better child feeding practices, improvement in household wealth, dietary diversity, and better sanitation for improvement in child nutrition (Aguayo and Menon 2016; Corsi et al 2016). Nonetheless, progress in these aspects is dependent on another important factor: strong leadership and commitment for nutrition. Analysis suggests that commitment and actions of leaders, especially their role in mainstreaming nutrition in the development discourse and policy formulation, is important for the decline in child undernutrition (Nisbett et al 2015). Recall, here, that India’s dismal record in child undernutrition at the peak of economic growth, brought to light by the NFHS-3 in 2005–06, had both created some expectation and mounted pressure on the state to act. The Prime Minister in his Independence Day address in 2008 unequivocally expressed a need for definitive action: “the problem of malnutrition is a curse that we must remove.” The present analysis, by implication, is also an attempt at examining the follow up on that avowal.

In this regard, it would be instructive to assess the performance of Bihar, Chhattisgarh, Odisha, Madhya Pradesh, and Uttar Pradesh, which have had high levels of child undernutrition. Most of these states, barring Uttar Pradesh, had the fortune of not only having continuity in policy and political regimes for more than a decade, but also having their leaders hailed for effective governance and commitment to development. Hence, some of these states rank high in governance parameters (Mundle et al 2016). These states have also emerged as a “centre of rapid growth” during the last 10 years (Sharma 2013). Whether the synthesis of these factors has enabled these states to register an impressive decline in child undernutrition is worth examining.

This paper, therefore, aims to (i) assess the decline in child undernutrition and socio-economic disparities in India between 2005–06 and 2015–16, (ii) analyse the performance of states in general and of the states which have had high undernutrition in particular, and (iii) identify specific patterns from the analysis of child undernutrition prevalence in districts of India. To examine these issues, we analyse the unit-level data of the recently released NFHS-4, conducted during 2015–16 in all the states of India with a sample size of 6,01,509 households. Specifically, we analyse the prevalence of stunting (low height-for-age) and underweight (low weight-for-age) among children aged 0–59 months at the national, state, and district levels. To assess the progress made in reducing child undernutrition between 2005–06 and 2015–16, we attempt a comparative analysis between NFHS-3 and NFHS-4 at the national and state levels. An analysis of the performance of 640 districts of India, complemented by a multivariate analysis carried out in 606 districts, follows.

The NFHS-4 reveals that in 2015–16 in India, about 38% of under five children are stunted and 36% are underweight. These figures are close to double of that of wasting (Figure 1, available on the EPW website). In rural India, the figures mount by 10% as compared to urban India, with over 40% of children being stunted and 38% being underweight. Stunting has declined by 10 percentage points and underweight conditions by 7 percentage points between 2005–06 and 2015–16 in India. However, assessed against the cross-country analysis of Headey (2013) as mentioned before, even a 10 percentage points decline in stunting in the last 10 years appears inadequate, as India’s growth rate has been well above 5.5% during this period.

Socio-economic Disparities

The NFHS-3 brought out marked socio-economic disparities in child undernutrition in India. The NFHS-4 reiterates that the disparity between wealth groups remains quite high in India (Table 1). Half of the children from poorest households are stunted and underweight, the percentage of prevalence of these conditions being double than the richest households. The prevalence decreases progressively along with an increase in the wealth order. While the poor and the middle groups have registered a relatively larger decline in underweight, it has gone up marginally among the richest wealth group. In stunting, the middle three groups (poor, middle, and rich) have performed relatively better than the richest and poorest groups.

Assessment of inter-temporal progress in reducing child undernutrition primarily through a simple measure like difference (decline measured in percentage points), though useful, may not be robust enough to capture the complexities of varied progress. Specifically, the difference measure is level-insensitive, and thereby ignores the hugely varying base-level differences in child undernutrition prevalence across socio-economic groups. Therefore, it is important to employ measures that are level-sensitive (Kakwani 1993; Mishra and Subramanian 2006; Joe and Mishra 2016; Nathan and Mishra 2017). Hence, we employ a level-sensitive index proposed by Nathan and Mishra (2017) to substantiate the analysis using difference. Interestingly, the Nathan and Mishra (NM) index1 corresponds to the broad patterns emerging from the simple rate difference. The notable difference the NM index brings out is the relatively better performance of the rich over the middle group in stunting and the middle group over poor in underweight.

To assess the inter-group disparities, we use both rate difference (RD) and rate ratio (RR), which are respectively the simple absolute and the relative measures of health disparity.2 The RD is the difference between group i and the reference group (the group with the lowest stunting rate). The RR is ratio between the stunting rate of group i and the reference group. Both the RD and RR suggest a decline in disparity in stunting and underweight among wealth groups.3

Further, we examine in both years the relative contribution of different wealth groups to overall stunting and underweight. Following Jayaraj and Subramanian (2002), we define relative contribution RC=pisi/S, where piis the share of the wealth group i in the child population, siis the stunting (or underweight) prevalence rate ofgroup i, and S is the overall stunting (or underweight) prevalence rate. Through this measure, we consider if the temporal decline in stunting (or underweight) rate of a given group is desirable, if the group’s contribution to overall stunting (or underweight) level declines. Two observations could be made in this regard from Table 1. One, the relative contribution of the poorest and poor groups to overall stunting and underweight continues to be higher than their relative population share, whereas the reverse holds good for the other three groups, especially the richest. Two, the relative contribution of the poorest group and richest group to the overall stunting and underweight prevalence has increased between 2005–06 and 2015–16.

The decline in stunting and underweight across the social groups is the highest among the Adivasis and the lowest in “Others” (used here for the advantaged social group) based on simple rate difference (Table 2). However, the level-sensitive NM index brings out a contrary pattern in stunting decline: the decline is the highest among Others and lowest among Adivasis. However, in underweight, Dalits have the highest decline and Others have the smallest decline. Despite the larger decline in child undernutrition, the relative contribution of disadvantaged social groups (Adivasis and Dalits) to overall stunting and underweight has increased between 2005–06 and 2015–16. The Others seemed to have fared better, especially in stunting. Both the RD and RR measures suggest that social disparity in underweight has declined between 2005–06 and 2015–16. However, the absolute measure RD suggests a decline in social disparity in stunting, whereas the relative measure RR indicates an increase in underweight. The above discussion suggests a marginal decline in socio-economic disparity in child undernutrition in India during the last 10 years. However, despite such a decline, the disadvantaged socio-economic groups still bear the disproportionate burden of child undernutrition in India.

The above analysis does not inform whether the poor from all social groups are equally worse off. Three broad patterns emerge from Table 3 which present the intersection of social and wealth groups. One, a large wealth-based disparity exists among all social groups in India. Among Dalits and Adivasis, the percentage of stunted poorest children is close to twice that of children from richest households. Two, children from all social groups are almost equally worse off under the poorer category and equally better off under the richer category. Three, the social gradient among wealth groups is much smaller than the wealth gradient among social groups. These suggest that wealth disparity in child undernutrition is larger than social disparity in India. The relative contribution of each wealth group within the social group also confirms the above.

Performance of States

Kerala maintains its lead position with the lowest stunting prevalence, followed by Goa, Tripura, Punjab, and Haryana (Table 4, p 66). Bihar, with 48% of stunting, remains at the bottom, followed by Uttar Pradesh, Jharkhand, Meghalaya, and Madhya Pradesh; all with above 40% of stunting. Rajasthan and Gujarat are two other states with stunting prevalence higher than the national average. By contrast, Kerala is the only state having below 20% stunting, though Goa comes quite close. This also implies that the bottom five states have more than double the stunting prevalence of both Kerala and Goa. Thus, what we find is both continuity and change. While the bottom states continue to have the same status, the relative positions of many other states seem to have changed over the years. For instance, Tripura, Himachal Pradesh, and Chhattisgarh have improved their positions marginally, whereas Manipur, Sikkim, Karnataka, Rajasthan, Jharkhand and Tamil Nadu have slid down from their previous positions.

States which registered a larger decline in stunting—Chhattisgarh, Arunachal Pradesh, Gujarat, Himachal Pradesh, and West Bengal—do not present any regional pattern. Instead, the common factor among them is the higher levels of stunting in 2005–06. That said, Bihar, Jharkhand, Madhya Pradesh, and Rajasthan, which also had higher levels, have not performed equally well. Uttar Pradesh and Meghalaya have done better than these states. Oddly, Goa, Kerala, and Tamil Nadu (previous better performers), along with Jharkhand, Rajasthan, and Bihar (previous poor performers) constitute the group of low decline states in stunting. Himachal Pradesh emerges as an exception in both aspects, as it registered a larger decline despite its previous lower levels.

As discussed earlier, the assessment of inter-temporal decline in child undernutrition across the states with hugely varying levels could be misleading through simple rate difference because of its level-insensitivity. Application of the level-sensitive NM index yields rather different patterns. Tripura and Himachal Pradesh, which had both a low base and a larger decline, emerge as the best performers, followed by Arunachal Pradesh, Punjab, and Mizoram. Interestingly, Kerala and Goa, which had the lowest decline in percentage points, have improved their positions and do not belong to the bottom ranks. Chhattisgarh and Gujarat, which have done well in terms of percentage points decline, have slipped from their rankings. Jharkhand, Rajasthan, Bihar, and Madhya Pradesh retain their bottom positions, surprisingly, along with Tamil Nadu.

In underweight, Jharkhand has the highest prevalence of all the states in 2015–16 (Table 5). Bihar, Madhya Pradesh, Uttar Pradesh, and Gujarat comprise the bottom group, with more than 40% of underweight, led by Jharkhand. Meghalaya, with the largest decline (in percentage points), has improved its position significantly, whereas Karnataka and Maharashtra with very little decline (in percentage points) came close to bottom league. Mizoram, Manipur, Sikkim, Kerala, and Jammu and Kashmir, with the least underweight prevalence, occupy the top positions. Punjab and Goa, which were top performers in 2005–06, have slipped from their previous positions.

Madhya Pradesh, Tripura, Himachal Pradesh, and Arunachal Pradesh, besides Meghalaya, have seen large declines (in percentage points) in underweight. Among these, Meghalaya and Madhya Pradesh had high prevalence in 2005–06, whereas Himachal Pradesh and Arunachal Pradesh had somewhat low prevalence. Maharashtra, Goa, and Karnataka team up with Uttar Pradesh and Rajasthan as states with the least progress (in percentage points) in underweight. The use of the NM index revises the ranks here too. Mizoram, which was ranked 14th in incremental improvement (percentage points), emerges as the top performer. Similar upward revision also happens to Manipur and Kerala, whereas Arunachal Pradesh and Himachal Pradesh undergo only a marginal but upward revision of ranks. A previously top-ranked Meghalaya has slid down to the fifth rank. A much larger sliding of ranks has also happened to Madhya Pradesh, Bihar, and Jharkhand. There is hardly any revision in the bottom ranks: Maharashtra, Goa, Karnataka, Uttar Pradesh, and Rajasthan. Clearly, based on the level-sensitive NM index, Arunachal Pradesh and Mizoram emerge as better performers both in reducing stunting and underweight conditions, accompanied by Himachal Pradesh. Note that, except Himachal Pradesh, the remaining four better performers in underweight are the north-eastern states.

Have the states with high undernutrition performed well over time? In stunting, Uttar Pradesh, Bihar, Meghalaya, Chhattisgarh, and Gujarat were the bottom five states with over 52% of stunting in 2005–06. Madhya Pradesh and Jharkhand, with 50% of stunting, closely followed them. Of these states, only Chhattisgarh and Gujarat have done well to an extent that they no longer belong to the bottom five states (Table 4). Though Meghalaya has also done well in reducing stunting (11 percentage points), its level is higher than the national average. Madhya Pradesh had 60% of underweight, both Jharkhand and Bihar had 56%, followed by Meghalaya (49%) and Chhattisgarh (47%) in 2005–06. Most of these states, except Meghalaya, continue to have levels of underweight higher than the national average (Table 5).

Relative Disadvantage

Table 3 suggests that there is a huge, graded wealth-based disparity, which is larger than the social disparity, in child undernutrition in India. We assess the wealth-based disparity in child undernutrition across the states using the relative disadvantage index (RDI) proposed by Jayaraj and Subaramanian (2002). The RDI, invoking the concept of equity in relative terms, presumes that there is no relative deprivation for a group i, if its contribution to the overall child undernutrition equals its share in the child population. In other words, group i is relatively advantaged (or disadvantaged), if its contribution to the overall child undernutrition is lesser (or higher) than its share in total child population. We computed the RDI for the poorest wealth group (bottom quintile) for the states4 (Table 6). The higher the positive value of RDI, the higher the relative deprivation of the poorest wealth group. This analysis is complemented with the use of the concentration index, proposed by Erreygers (2009).5

Uttar Pradesh stands out as the state where the relative deprivation of the poorest group in stunting is the highest. It is followed by Goa, Odisha, Gujarat, and Mizoram. By contrast, Kerala emerges as the state where the relative deprivation of the poorest in stunting is the lowest of all the states, followed by Meghalaya, Arunachal Pradesh, Chhattisgarh, Punjab, and Tamil Nadu (Table 6). In underweight, the relative deprivation of the poorest is the highest in Maharashtra, followed by Rajasthan, Telangana, Odisha, and Gujarat. On the contrary, Himachal Pradesh appears as the state where the relative deprivation of the poorest is the lowest; Manipur, Meghalaya, Tamil Nadu, Nagaland, and Punjab come next.

As is well known, it is a standard practice to examine wealth-based inequality in health outcomes using Concentration Indexes (CIs), as the CI indicates the disproportionate burden of undernutrition among the poor compared to the rich group. It is important to note that the level of inequality across the states may vary based on the choice of CI. Here, we use the CI proposed by Erreygers to complement the above analysis based on relative disadvantage index. The CI takes the value from -1 to +1. A higher negative value indicates higher undernutrition rates among the poorer groups.
Interestingly, the patterns emerging from the CI both in stunting and underweight broadly correspond to the patterns emerging from the relative disadvantage index. There appears to be a substantial overlap between the high prevalence of undernutrition and the relative disadvantage of undernutrition among the poorest.

Van de Poel et al (2008) has proposed three groups to classify countries based on wealth-based inequality. These are mass deprivation (prevalence is very high among all the groups except the top), exclusion (prevalence is low among most of the groups except the bottom group), and queuing (average prevalence is lower than the mass deprivation group, but richer groups are better off while those at the bottom wait for improvement). While all the Indian states do not neatly belong to these three groups, many states constitute mass deprivation group in stunting (Uttar Pradesh, Bihar, Jharkhand, Rajasthan, Madhya Pradesh, Chhattisgarh, Meghalaya, Gujarat, and Karnataka) and underweight (Uttar Pradesh, Madhya Pradesh, Bihar, Chhattisgarh, Jharkhand, Rajasthan, and Gujarat). Surprisingly, the same seven states, except Meghalaya and Karnataka, form the mass deprivation group in both aspects of undernutrition in India.

District-level Analysis

At least three points emerge from the analysis of prevalence of child undernutrition at the district level. First, out of the 10 districts with lowest stunting prevalence, five are from Kerala.6 By contrast, six of the 10 districts with highest stunting prevalence are from Uttar Pradesh, including Deoria with the third lowest stunting prevalence. Two districts of Karnataka also belong to the bottom 10 districts. Second, six of the 10 lowest underweight prevalence districts are from Jammu and Kashmir. The rest are from the north-eastern states. The highest underweight prevalence districts are spread across nine states, including Gujarat, West Bengal, and Karnataka. Third, the difference in prevalence between the top and bottom districts is huge. Stunting prevalence in Bahraich district, Uttar Pradesh is five times the prevalence in Ernakulam district, Kerala. Pashchimi Singhbhum (Jharkhand) has 11.5 times higher underweight prevalence than Mokokchung (Nagaland). Surprisingly, the Dang district (Gujarat) with 60% of underweight comes next, followed by Purulia (West Bengal) and Gulbarga (Karnataka) with about 55% of prevalence.

Figures 2 and 3 (available on the EPW website) present the relative intensity of stunting and underweight across 628 districts in India.7 Of these 628 districts, 27 (or 4%) districts have less than 20% of stunting and 48 (or 8%) districts have stunting prevalence between 51% and 65%, while 99 (or 16%) districts have less than 20% underweight prevalence, and 38 (6%) districts have underweight prevalence between 51% and 65%. While 168 (27%) districts fall under 2% to 30% prevalence in stunting, in underweight 158 (25%) districts belong to the same category, 196 (31%) districts have stunting prevalence ranging from 31% to 40%, and the remaining 189 (30%) districts belong to the next range of 41% to 50%. In underweight, 168 (27%) and 165 (26%) districts belong to 31% to 40% and 41% to 50% categories, respectively.

We attempted to identify the districts where child undernutrition is high and the states to which these districts belong. To do so, we made a broad grouping of districts where prevalence of stunting and underweight is more than 40%; marginally higher than the national average of 38% of stunting and 36% of underweight. Figure 4 (available on the EPW website) indicates that 237 districts, constituting more than 37% of districts of India, have high stunting prevalence of above 40%. Of these 237 districts, 128 (54%) districts belong to three states alone: Uttar Pradesh (60), Bihar (36), and Madhya Pradesh (32). Another 29% of districts (68) are from the five states of Jharkhand (19), Gujarat (15), Rajasthan (13), Maharashtra (11), and Odisha (10 districts). Thus, eight states account for 83% of high stunting prevalence districts in India. The remaining 17% of high stunting prevalence districts are spread across 11 states.

Figure 5 (available on the EPW website) presents similar results for underweight. About 32% (203) of districts in India have underweight prevalence more than 40%. Of these 203 districts, 64% belong to four states, of Madhya Pradesh (37), Uttar Pradesh (34), Bihar (34), and Jharkhand (24). These are followed by 26% underweight prevalence in districts in another four states of Gujarat (15), Maharashtra (13), Odisha (13), and Rajasthan (12). Thus, 90% of high underweight prevalence districts belong to the same eight states which also house over 80% of high stunting prevalence districts. It is clear from the above that 80% of high stunting and 90% of high underweight prevalence districts are clustered in eight states. Our previous results reveal that these are the same states with both high prevalence rates and higher levels among almost all wealth groups, except the richest, that comprise the mass deprivation group.

Multivariate Analysis

Many studies have identified a close association between a specific set of factors and child undernutrition both in India and other developing countries. Of late, these findings have led to a polarising of views on what matters more in reducing child undernutrition. While some studies attribute open defecation and the associated disease environment as the primary causal factors for child undernutrition (Chambers and Medeazza 2013; Spears and Lamba 2013), others establish the primacy of maternal nutrition and education and find the role of the former factors either insignificant or at best marginal (Corsi et al 2016). With this context in mind, we attempt to identify, through a multivariate regression model, the influence of socio-economic, health and demographic factors that can possibly explain the variation in stunting and underweight across the districts in India.

Our choice of explanatory variables is based on a careful scrutiny of empirical studies that examine the influence of varied range of factors on child undernutrition in India and elsewhere. We considered a specific number of variables that emerge consistently as significant in most of these empirical studies.8 Hence, we do not elaborate on the reasons underlying the choice of the explanatory variables due to want of space. We carried out regression analysis using the ordinary least squares (OLS) method of estimation and estimated separate models for stunting and underweight at the district level.9 We restrict our regression analysis only to 606 districts due to non-availability of data for some of the explanatory variables for 34 districts.

The results from regression analysis10 on stunting and underweight are in Table 7. The OLS estimates reported in the first column of Table 7 suggest that there is a statistically significant association between stunting and factors such as women’s undernutrition, anaemia among children, children receiving adequate diet, poverty level, women’s mean years of schooling, access to improved sanitation facility, access to improved drinking water source, and sex ratio in the district. Among all the explanatory factors, women’s mean years of schooling appears to be the most important variable that is strongly associated with stunting. On an average, a 10% decrease in undernutrition in women leads to a 2.7% decrease in stunting in children. A 10% decline in the below poverty line (BPL) population could lead to a 1% decline in stunting.

The estimated regression coefficients of six explanatory variables are statistically significant in underweight. We find that women’s undernutrition and poverty have strong effects on children’s underweight prevalence in a district. The results indicate that, on an average, all else held constant, 10% decrease in women’s undernutrition is associated with roughly a 6% decrease in the child underweight. Also, as the BPL population falls by 10%, the proportion of underweight children in a district drops by around 1.1%. Other explanatory factors seem to exert only a limited influence on underweight prevalence. For example, if proportion of households using improved sanitation facility rises by 10% in a district, underweight prevalence, on an average, is likely to fall by 0.7%.

Our regression analysis suggests the significant and positive role of improvement in women’s nutrition and poverty reduction in the reduction in prevalence of child undernutrition in India. These findings are consistent with previous research which established the significant effects of these factors both in India and other countries (Corsi et al 2016; Nisbett et al 2017). It appears that the same set of factors may not work equally for stunting and underweight. It can also be concluded that, though access to improved sanitation can possibly help reduce child undernutrition in India, its relative role appears to be small.

In Conclusion

India has managed to make a moderate decline in stunting and underweight prevalence in children under five years between 2005–06 and 2015–16. However, not only is the extent of progress deficient against its economic growth, but also close to 40% of children under five years remain undernourished even today. There has been a marginal decline in socio-economic disparity in child undernutrition in this period. Yet, a large, graded socio-economic disparity still prevails.

Arunachal Pradesh, Mizoram, and Himachal Pradesh have performed well in reducing child undernutrition between 2005–06 and 2015–16. Some of the states with high undernutrition in 2005–06, have reduced either stunting or underweight significantly. However, the same group of states—Uttar Pradesh, Bihar, Madhya Pradesh, Jharkhand, Gujarat, and Rajasthan—with high undernutrition prevalence persist at the bottom. These six states with high levels of undernutrition among all wealth groups, except the richest group, also constitute the mass deprivation group. District-level analysis validates that these six states, along with two other states, account for over 80% of districts in India having high undernutrition prevalence (above 40%). The regression analysis at the district-level suggests that enhancement in women’s schooling and nutrition and poverty reduction can play a significant role in reducing child undernutrition in India.

As noted already, some of these states, especially Bihar, Madhya Pradesh, Jharkhand, and Gujarat, had all the vital ingredients required for better performance: policy and political continuity, and leadership committed to development and robust economic growth. Yet, their performance is not only inconsistent on aspects of undernutrition, but also far from sufficient to lift them up from their bottom positions of child undernutrition. What additional aspects, beyond these, would enable these states to attain progress, which will also profit the poor more? The performance of some of the north-eastern states might provide some insight in this regard. Since Bihar, Uttar Pradesh, Jharkhand, Madhya Pradesh, Odisha, Maharashtra, Rajasthan, and Gujarat, house 80% of the districts in India with high child undernutrition prevalence, their stellar performance is a sine qua non not only for their own sake, but also for removing the “curse” of child undernutrition from India.


1 The NM index is computed as follows: NM=(1-Hb/Ha)(1-µHb), where Ha depicts stunting/underweight rate in 2005–06 and Hb is stunting/underweight rate in 2015–16. The NM is reported for µ= 0.5. For details, see Nathan and Mishra (2017).

2 Use of any measure of health inequality involves implicit or explicit value judgments about notion of equality. Outcomes of inequality often depend on the choice of inequality measure (Asada 2010; Kjellsson and Gerdtham 2013).

3 When wealth inequality is measured using standard concentration index, we find that inequality in undernutrition increased between 2005–06 and 2015–16.

4 Following Jayaraj and Subramanian (2002), we define RDI= (pi/(1- pi)))((ui/U) -1) if pi ≥ U; RDI= ((ui- U)/(1-U)) if pi <U; where, pi is share of population of group i, ui is stunting/underweight rate of group i, U is overall rate of stunting/underweight.

5 The Erreygers concentration index captures “an absolute value judgment” (Kjellsson and Gerdtham 2013).

6 Our estimates from the NFHS-4 unit-level data on stunting and underweight prevalence do not match with the estimates reported in district-level fact sheets for a few districts in some states. Hence, we used estimates reported in the fact sheet for district-level analysis.

7 Union territories of India are not included in the district-level analysis.

8 Dalit and Adivasi population figures are from Census 2011, district-level poverty headcount ratio is from Mohanty et al (2016), and the rest of the explanatory variables are from the NFHS-4 data.

9 Using the same set of explanatory variables, we estimated quantile regression model at each of four quantiles (that is, 20th, 40th, 60th and 80th quantiles) of stunting and underweight. The estimates from quantile regression give results similar to our multivariate regression analysis and, hence, are not reported here.

10 The regression results may not be interpreted as causal relationships between explanatory variables and dependent variables because of potential endogeneity bias.

Aguayo, Victor M and Purnima Menon (2016): “Stop Stunting: Improving Child Feeding, Women’s Nutrition and Household Sanitation in South Asia,” Maternal & Child Nutrition, Vol 12, No S1, pp 3­­­­–11.

Asada, Yukiko (2010): “On the Choice of Absolute or Relative Inequality Measures,” Millbank Quaterly, Vol 88, No 4, pp 616–22.

Chambers, Robert and Gregor Von Medeazza (2013): “Sanitation and Stunting in India: Undernutrition’s Blind Spot,” Economic & Political Weekly, Vol 48, No 25, pp 15−18.

Corsi, Daniel J, Ivan Mejia-Guevara and S V Subramanian (2016): “Risk Factors for Chronic Undernutrition among Children in India: Estimating Relative Importance, Population Attributable Risk and Fraction,” Social Science and Medicine, Vol 157, No 5, pp 165−85.

Cunningham, Kenda, Derek Headey, Akriti Singh, Chandni Karmacharya and Pooja Pandey Rana (2017): “Maternal and Child Nutrition in Nepal: Examining Drivers of Progress from the mid-1990s to 2010s,” Global Food Security, Vol 13, No 2, pp 30−37.

Erreygers, Guido (2009): “Correcting the Concentration Index,” Journal of Health Economics, Vol 28, No 2, pp 504−15.

Headey, Derek (2013): “Developmental Drivers of Nutritional Change: A Cross-country Analysis,” World Development, Vol 42, No 2, pp 76–88.

Jayaraj, D and S Subramanian (2002): “Child Labour in Tamil Nadu in the 1980s: A Preliminary Account of Its Nature, Extent and Distribution,” Economic & Political Weekly, Vol 37, No 10, pp 941−54.

Joe, William and Udaya S Mishra (2016): “Reckoning Level Differentials in the Measurement of Progress: An Application in the Context of Deliveries Attended by Skilled Health Personnel Across Least Developed Countries (LDCs),” Journal of International Development, Vol 29, No 3, pp 330–50.

Joe, William, Ramaprasad Rajaram and S V Subramanian (2016): “Understanding the Null-to-small Association between Increased Macro­economic Growth and Reducing Child Undernutrition in India: Role of Development Expenditures and Poverty Alleviation,” Maternal and Child Nutrition, Vol 12, No S1, pp 196−209.

Kakwani, N (1993): “Performance in Living Standards: An International Comparison,” Journal of Development Economics, Vol 41, No 2, pp 307–36. 

Kjellsson, G and U G Gerdtham (2013): “On Correcting the Concentration Index for Binary Variables,” Journal of Health Economics, Vol 32, No 3, pp 659−70.

Mishra, Udaya S and S Subramanian (2006): “On Measuring Group-differentials Displayed by Socio-economic Indicators,” Applied Economics Letters, Vol 13, No 8, pp 519−21.

Mohanty, S K, D Govil, R K Chauhan, R Kim and S V Subramanian (2016): “Estimates of Poverty and Inequality in the Districts of India, 2011–2012,” Journal of Development Policy and Practice, Vol 1, No 2, pp 142–202.

Mundle, Sudipto, Samik Chowdhury and Satadru Sikdar (2016): “Governance Performance of Indian States: Changes between 2001–02 and 2011–12,” Economic & Political Weekly, Vol 51, No 36, pp 55–64.

Nathan, Hippu Salk Kristle and Srijit Mishra (2017): “Group Differentials for Attainment and Failure Indicators,” Journal of International Development, Vol 29, No 4, pp 520–32.

Nisbett, Nicholas, Elise Wach, Lawrence Haddad and Shams El Arifeen (2015): “What Drives and Constrains Effective Leadership in Tackling Child Undernutrition? Findings from Bangladesh, Ethiopia, India and Kenya,” Food Policy, Vol 53, No 4, pp 33–45.

Nisbett, Nicholas, Peter Davis, Sivan Yosef and Nazneen Akhtar (2017): “Bangladesh’s Story of Change in Nutrition: Strong Improvements in Basic and Underlying Determinants with an Unfinished Agenda for Direct Community Level Support,” Global Food Security, Vol 13, No 2, pp 21–29.

Ramalingaswami, Vulimiri, Urban Jonsson and Jon Rohde (1996): “The Asian Enigma,” The Progress of Nations, New York: United Nations International Children’s Emergency Fund, pp 11–17.

Sen, Amartya (2001): “Hunger: Old Torments, New Blunders”, Little Magazine, Vol 2, pp 9–13, viewed on 27 June 2012,

Sharma, Ruchir (2013): “The Rise of the Rest of India: How States Have Become the Engines of Growth,” Foreign Affairs, Vol 92, No 5, pp 75–85.

Shekar, Meera, Julia Dayton Eberwein and Jakub Kakietek (2016): “The Costs of Stunting in South Asia and the Benefits of Public Investments in Nutrition,” Maternal and Child Nutrition, Vol 12, No S1, pp 186–95.

Spears, Dean and Sneha Lamba (2013): “Effects of Early Life Exposure to Sanitation on Childhood Cognitive Skills: Evidence from India’s Total Sanitation Campaign,” Policy Research Working Paper 6659, Water and Sanitation Programme Unit, World Bank, Washington, DC.

Subramanyam, Malavika A, Ichiro Kawachi, Lisa F Berkman and S V Subramanian (2011): “Is Economic Growth Associated With Reduction in Child Undernutrition in India?,” PLOS Medicine, Vol 8, No 3, pp 1–15.

UNICEF (2013): “Improving Child Nutrition: The Achievable Imperative for Global Progress,” United Nations International Children’s Emergency Fund, New York.

Van de Poel, Ellen, Ahmad Reza Hosseinpoor, Niko Speybroeck, Tom Van Ourti and Jeanette Vega (2008): “Socioeconomic Inequality in Malnutrition in Developing Countries,” Bulletin of the World Health Organization, Vol 86, No 4, pp 282–91.

Updated On : 7th Dec, 2018


(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top