Human Development

Has the ICDS helped reduce stunting in India?

  • Blog Post Date 09 October, 2015
  • Articles
  • Print Page
Author Image

Monica Jain

International Initiative for Impact Evaluation (3ie)

mjain@3ieimpact.org

While stunting has declined sharply in India, the levels remain disturbingly high at 38.7%. This column evaluates the impact of the supplementary feeding component of ICDS – India’s flagship programme for early child development - and finds sizable positive effects on heights of 0-2 year olds. However, these gains are achievable only if the programme is focused on this age group and if food is delivered regularly.

According to the fact sheet released by the Ministry of Women and Child Development, the recent provisional estimates from the Rapid Survey on Children 2013-14 indicate that stunting1 has declined sharply in India. However, even now, child stunting levels remain disturbingly high at 38.7%. In this column, I present new research which aims to show the impact of supplementary feeding through the Integrated Child Development Services (ICDS) programme on children’s physical growth using data from the 2005-06 National Family Health Survey (NFHS) (Jain 2015).

The ICDS programme was launched in India in 1975 and is today one of the world’s largest programmes for early child development with more than 1 million centres nationwide. The programme provides a package of services ranging from supplementary nutrition and immunisation to pre-school education, to 0-5 year old children. However, the flagship component of the ICDS programme is supplementary feeding to 0-72 month old children for 300 days in a year (25 days a month). Children aged 3–6 years are generally fed at the anganwadi2 itself, while children below age three receive take-home rations that last for a week or a month depending on the frequency of distribution3.

Unravelling the benefit of ICDS food on physical growth of girls and boys

I analyse the impact of receiving supplementary food through the ICDS on physical growth of children. I focus on 0-2 year old children4, as this age is widely recognised as the “critical window” when the maximum growth faltering takes place, and it is very difficult to reverse stunting after this age (Martorell, Khan and Schroeder 1994).

Evaluation of benefits of supplementary feeding on physical growth of children is a cumbersome task. All children ages 0-2 years are eligible to receive supplementary feeding under ICDS. However, only a small proportion of children are actually receiving feeding daily. To evaluate the benefits of feeding on these children, we need a group of children to compare them with, who have the same characteristics in all aspects, except that they did not receive the feeding daily. In the technical impact evaluation language, we need a ‘control’ or comparison group.

I do this in two ways. First, from the group of children not receiving supplementary feeding daily, I manufacture a comparison group, which is the closest match to the group of children receiving supplementary feeding daily, on the basis of observable demographic and socioeconomic characteristics, and geographic location. I then evaluate the differences in various physical growth measures between the two groups of children: those who got the supplementary feeding daily and those in the comparison group.

In the above matching method, I assume that there are no differences between the two groups of children on the basis of unobservable characteristics. But it is possible that this assumption does not hold. Children who are receiving supplementary food daily might have mothers who are highly motivated to take care of their children and hence do everything possible, including going to the ICDS to get food. So we need a control group which is based on observable characteristics plus the motivation level of mothers (and other such unobservable characteristics). However, we do not observe the motivation level of mothers.

To address this problem of unobservable characteristics, I take advantage of evidence from nutrition studies (Bhutta et al. 2008) indicating that supplementary food often has no or minimal effect on the heights of older children. Though there is no obvious fixed age threshold for this, following a recent review by Bhutta et al. (2008), I use children aged 4–5 years as a comparison group. I estimate the difference in heights of 4-5 year old children who go to ICDS centres and those who do not. Since this difference is not due to supplementary food through the ICDS, it can be attributed, to unobservable characteristics (observable characteristics are controlled for). I use this estimate to control for difference in heights of children aged 0-2 years due to unobservable characteristics, between those who get food from the ICDS centres and those who do not5.

Findings

Small children eat poorly, even among richest and most educated households

Complementary feeding6 is a process that starts when breast milk is no longer sufficient to meet the nutritional requirements of children. In addition to breast milk, 6-24 month old children require frequent, nutritious, and energy-dense foods that can be easily digested. The NFHS 2005-06 found that only 21% of 6-23 month old children in India are fed appropriately according to the recommended complementary feeding practices. Though feeding practices improve with age, income and education of women, only 29% of children in the highest income quintile families, and only 32% of those whose mothers had at least 12 years of education were being fed properly. It is not surprising, therefore, to find that 25% of the children of the highest income quintile were stunted. Poor child feeding practices is a South Asia wide-phenomenon. In a recent paper, using longitudinal data from 1996 to 2006-07, I find that the complementary feeding practices for 6-23 month old children in Bangladesh are poor and have not improved significantly even over a decade (Jain 2014).

ICDS provides a sizable portion of required food for children under two

The ICDS provisions 300 calories and 8–10 g of protein for all 6-72 month old children, and 500 calories and 15-20 g of protein for pregnant and lactating mothers. For 6-23 month old children, this covers 50–150% of their required complementary energy needs and 70–100% of the recommended protein needs. For children above age two, it can fulfil 20–30% of the energy requirements of children and 50–70% of their recommended protein needs.

Anganwadis exist widely in rural India, but ICDS food does not

According to NFHS 2005-06, 91.5% of Indian villages had an ICDS centre. However, only 27% of 0-2 year old children and 33% of 3-5 year old children got any food from the centres in rural India. The condition is even worse when we consider the percentage of children who got the food most intensely: only 6% of 0-2 year old children and 13% of 3-5 year old children got the food daily. There are no major differences in these statistics by gender; however, there are large inter-state variations. In states like Assam and Uttaranchal no 0-2 year old child got ICDS food daily, whereas 20% in Maharashtra got it. Even within states, there are large differences in ICDS food uptake between the younger and older pre-school children. Like in Karnataka, while only 5% of 0-2 year old children got ICDS food daily, 28% of the older children (3-5 years of age) got it. These differences by age reflect the traditional emphasis of the ICDS programme on the supplementary feeding of older pre-school children.

Children from weaker socioeconomic groups more likely to get ICDS food

Children from weaker socioeconomic groups, such as scheduled caste and tribes (SC/ST), were more likely to receive ICDS food daily. Children receiving ICDS food daily were concentrated in some states like Tamil Nadu, Maharashtra, West Bengal, Himachal Pradesh and Mizoram, and in some villages within these states.

Girls who got ICDS food daily at least 1 cm taller than those who did not

I find that in rural India, 0-2 year old girls who received ICDS food daily were at least 1 cm taller that those who did not. The effect on boys is similar but less robust. Given that these height differentials are most likely irreversible, supplementary nutrition could potentially bridge the height gap between the richest and poorest girls by at least 28%, and for boys by 19%, at adulthood.

I find additional evidence consistent with the programme effects presented above. My hypothesis is that the benefits of the programme are driven only by supplementary food and not the other ICDS services like immunisation. The data confirms this hypothesis. Also, If the gains in height are driven by ICDS supplementary food, then the intensity of feeding would make a difference. I do find that ICDS supplementary food benefitted only those children who received it daily and there was no benefit for those who received it less often. I also expect that the benefits would be bigger in states that have a better ICDS performance record. I do find the benefits to be bigger for children in the five top performing states: Tamil Nadu, Maharashtra, West Bengal, Himachal Pradesh and Mizoram.

Implications for policy

My simple back-of-the-envelope calculations indicate large long-run economic benefits of supplementary feeding through the ICDS programme for 0-2 year old girls and boys. However, all of these benefits are achievable only if the programme is targeted toward 0-2 year old children and if supplementary feeding is delivered regularly, which did not seem to be the case in 2005-06. There has been restructuring of the ICDS programme in recent years, including attempts to focus supplementary feeding on children below three years of age and pregnant and lactating mothers. These might partly explain the sharp drop in stunting in recent years. But there is still a long way to go.

Notes:

  1. Children whose height-for-age z-score deviates by more than two standard deviations from that of an international reference group of healthy children of the same age who are raised in a healthy environment, are considered stunted (World Health Organization (WHO)). Standard deviation is a measure in statistics that quantifies the amount of variation or dispersion of a set of data values.
  2. Anganwadis are the child care centres at which ICDS services are provided, and are usually located within the village itself.
  3. The NFHS survey asked about the frequency of supplementary feeding in the last one year. The surveyors were specially instructed to probe this, so that the even if the take-home rations were distributed at weekly or monthly intervals, but were meant for daily consumption, they were recorded accordingly.
  4. The age groups of children in years correspond to the months as follows: 0–2 years mean 0–24 months; 3-5 years mean 25-59 months; 4–5 years mean 37–59 months.
  5. There could be two problems with this method. First, the mothers of 4-5 year old children getting food from ICDS may be relatively less motivated than those of 0-2 year old children. If this was true, then we should find a higher percentage of 0-2 year old children getting supplementary food daily in comparison to 4-5 year old children. But this is not the case, as a significantly higher percentage of 4-5 year old children report getting ICDS food daily. Second, it is possible that the 4–5-year olds benefitted from ICDS food when they were young. However, if this is the case, then the impacts would be underestimates. Also, we do not know who and how many received ICDS food, but assuming that the share of 4-5 year olds who received it when they were younger is the same as today’s 0– 2 year old group, then the bias in estimates would be small because only 6% children report getting it in this younger age group.
  6. Complementary feeding can be provided privately by the family and/or publicly through supplementary feeding programmes like the ICDS. The private/public ratio would depend on the quantity provided by the public programme and its uptake by the family.

Further Reading

No comments yet
Join the conversation
Captcha Captcha Reload

Comments will be held for moderation. Your contact information will not be made public.

Related content

Sign up to our newsletter