New paper: India’s Struggle Against Malnutrition—is the ICDS Program the Answer?

A recent paper by Monica Jain (World Development, Vol 67, pp72-89, 2014) “India’s Struggle Against Malnutrition—is the ICDS Program the Answer?” tries to answer this question.

The title promises more than it delivers. The paper only looks at one of the ICDS programme elements—daily supplementary feeding to children 0-2 (actually 6-23 months). 

The paper also does not have the benefit of a true comparison group (which would require matching on observable characteristics and on unobservable characteristics such as maternal priority given to infant nutrition). To match on observable characteristics, the paper uses propensity scoring to adjust outcomes between the two groups (those mothers opting in to receive supplementary feeding for their 0-2 year olds and those who do not).  To match on unobservables the paper assumes that any differences in heights seen between 4-5 year olds whose mothers opt in and those who do not, is linked to unobservable effects.  It is not linked to observables because they have been taken care of and it is not linked to the programme because other studies are cited to show that supplementary feeding for 4-5 year olds is ineffective.  

So treating 0-2 year olds as the treatment group and 4-5 year olds as the control, these unobservables are “differenced” out. There are lots of assumptions here, but the author must have done enough to convince 3 skeptical World Development reviewers to let it through. It must have taken a great deal of perseverance.

Findings?  

The “impact” of daily supplementary feeding in the previous year (the survey is the nationally representative 2005-6 NFHS3) is 1cm, or about 28% of the 3.5cm difference between the height of women in the lowest and highest wealth groups.  The results are less robust for boys, but of similar magnitude.

Despite this potentially positive result, the author reports that only 6% of children 0-2 received supplementary feeding daily in the previous year of the survey. The author suggests this might be because of high private time costs; the lack of appreciation of potential benefit; poor delivery of the program; non-provision of supplementary feeding; and the prioritization of the 3-6 year old group for feeding.  It seems plausible that the potential is there for a much bigger impact of ICDS, not through spending more money but through better targeting of it.

The results in the paper are from 2005-6 data and we think much has been achieved since then in improving ICDS delivery and we also have prioritized the 0-2 age group more.  And open defecation rates will be lower today than in 2005-6. 

So even under the least favourable of circumstances the much maligned ICDS seems to have an effect on the 0-2 age group. If more children 0-2 actually received the supplements on a daily basis the total effect of ICDS would improve even more.  In the absence of randomized controlled trials of ICDS (difficult with universalization, although given uptake rates this low a randomization of encouragement to use the services could be undertaken and an impact estimate derived) the paper represents a valiant effort by the author to estimate impact. 

It would be good to see this analysis done for the new national 2013-14 RSOC data (see Global Nutrition Report). I bet the results would be even more positive.