Small children but big numbers: Estimating the economic benefits of addressing undernutrition
Small Children but Big Numbers
Estimating the Economic Benefits of Addressing
Harold Alderman IFPRI Nov. 30 2018
Putting Nutrition Goals on the Finance
The 2015 Sustainable Development Goals (SDGs) call for
improvements in all forms of malnutrition, including a 40% reduction
However, there are 17 SDGs compared to the eight MDGs.
And 169 targets (compared to 18 in the MDGs)
Prioritization is assisted by analysis of benefits and costs of reaching
Advocacy is enhanced by providing such estimates on a global as
well as national scale
The task is not new but the data are
Since, at least, 1972 economists have been estimating the
contribution of improved nutrition to economic growth
With increasing availability of longitudinal studies and controlled
trials much more is known on the efficacy of specific interventions
But going from these studies to estimates of costs and gains at scale
still requires myriad assumptions
Explorations, however, indicates that the economic benefits to
nutrition are not overly sensitive to the assumptions employed.
Method: Benefit Cost Analysis
Advantages: Based on productivity impacts and context specific
outcomes and costing data, often at moderate scale. Clear economic
interpretation when benefits > costs.
Drawbacks: Full calculation of benefits requires either longitudinal data
or linking proximate outcomes to additional data such as returns to
They also require assumptions about the discounted value of future
Moreover, placing mortality reductions into the calculation of benefits
requires heroic assumptions
Finally: extrapolation from underlying research requires information on
how costs change as scale increases. That information is generally not
Method: Estimates of GNP Loss Due to
Undernutrition of Current Population
Advantages: Converts easily observed current stunting rates to
productivity loss using a best practice estimate of impact of nutrition on
schooling and earnings. Can be compared to the current budget for
Drawbacks: Does not explore a direct path between specific investments
Examples: Horton and Steckel find global GNP loss to be 8%
Another study found a loss of between 1.9 and 11.4 percent of GNP for
countries in the Caribbean and Central America
Similarly, the African Union estimated that Ethiopia lost 16.5% of GNP and
Method: Estimates of Future GNP Loss Due to
Neglecting Undernutrition in Today’s Children
Advantages and disadvantages similar to estimates based on
As there are fewer individuals under 5 than those in the entire
current work force and health has generally improved since the
current workforce was in their most vulnerable years, these
losses are much lower than those mentioned previously.
And as these children will not enter the labor force for years, any
estimates of lost productivity require discounting
Method: Aggregate Budgetary Cost of
Prominent example: 2013 Lancet estimated that scaling up 10 proven
interventions to cover 90% of at-risk children would reduce stunting by
20% and cost $9.6B. The World Bank has updated these estimates
arriving at a similar $10B/yr.
Advantages: Outcomes can be calculated both in terms of age specific
mortality or nutritional status and confidence intervals provided. Costs can
be calculated using a region specific ingredient approach.
Drawbacks: Underlying software are often based on meta-analyses of
efficacy trials with high supervision rates and small samples. It is not clear
that these outcomes or costs will match what might be seen if the projects
were scaled up.
Going one step further to combine
productivity gains and costs at scale
This approach models the impact of changes in nutrition using an
assumed share of the current cohort of stunted children who will no
longer be stunted and an assumption of the value of this
improvement in terms of GNP.
The larger the GNP per capita the greater the value of improvements.
Similarly, the faster GNP is growing, the greater the future value of
The World Bank recently estimated the benefit : cost ratio to be 15: 1.
Returns were highest in India due to low program costs and high GNP
Can we increase confidence in such results?
It is a fair question to ask how much any of these estimates are based on
favorable assumptions or results from limited data.
For example, a global estimate of returns to nutrition by Hoddinott et al.
was based on changes in consumption 3 decades after a single RCT in
I followed the general direction of Hoddinott et al. to test whether
conclusions are robust to the assumptions used in that study.
The core results replaced the data from Guatemala with longitudinal
evidence of stunting and schooling from Zimbabwe and then link these
with estimates on the returns to schooling.
I progressively increase estimated costs and/or reduce the expected
benefits and showed that under a wide set of assumptions estimated
benefits greatly exceed costs.
Core Results for South Asia
Economic gains are calculated for the population born between 2015 and 2030 and
assume that all individuals enter the workforce at age 20 and continue to work for
The increment of earnings is based on the labor share of GNP (assumed to be
50%) which, as in Hoddinott et al., is projected to grow at rates reported in World
Bank models for the region.
The base line stunting rate in South Asia is 35.6%. Gains if current trends in
improvement in nutrition continue are $344 billion at a 6% discount rate over the
lifetime of this cohort.
Base case costs assume that to achieve a reduction in stunting for one child, 5
children receive the package of interventions
The B:C ratio is then 26:1 at 6% discount.
If investments are accelerated to reach 40 percent stunting reduction by 2030 the
benefits would be $497 billion or 18% of 2015 GNI
But the purpose of this study is to explore
various alternative assumptions
➢Increased costs of schooling (since enrollment increases)
➢Also assumed that as education increase wage premiums would be
half of current rates (dropping to 3.5% per year completed)
➢Replaced longitudinal results with global association of height and
wages (from Horton and Steckel).
All of these are the opposite of cherry picking parameters and all show
substantial benefits far greater than cost.
Bottom Line: In all scenarios the Benefits far
exceeded the Costs
Moreover, these substantial benefits do not include: i) reduced anemia; ii)
reduced health costs iii) reduced chronic disease and most important iv)
Assumptions are unavoidable for any estimates of the gains from improved
nutrition. There are uncertainties about extrapolating estimated impacts
derived from a few well-designed trials and the challenge of trying to look
forward for a number of decades.
However, the uncertainties in the estimates are not substantial from a policy
perspective. It is hard to imagine reasonable variations of the assumptions
used here that would invalidate the main conclusion that there is significant
under-investment in nutrition. There is no logical case that accelerate
investments should be delayed because the data have uncertainties.
➢Alderman, Behrman and Tasneem. Big Numbers about Small Children:
Estimating the Economic Benefits of Addressing Undernutrition
➢Hoddinott et al. The economic rationale for investing in stunting reduction.
Maternal and Child Nutrition 9(Suppl. 2): 69–82.
➢Bhutta et al. Evidence-Based Interventions for Improvement of Maternal and
Child Nutrition: What Can Be Done and at What Cost?Lancet 382(9890): 452-
Alderman, Behrman and Tasneem. Big Numbers about Small
Children: Estimating the Economic Benefits of Addressing
Hoddinott et al. The economic rationale for investing in stunting
reduction. Maternal and Child Nutrition 9(Suppl. 2): 69–82.
Bhutta et al. Evidence-Based Interventions for Improvement of
Maternal and Child Nutrition: What Can Be Done and at What
Cost?Lancet 382(9890): 452-477.