Thin Children in a Fed Country: The Gap Rice Cannot Close

Executive Summary. Bangladesh's child nutrition problem is not a food supply problem, and the two are routinely treated as one. Cereal yield was 5,321 kg per hectare in 2023, third of six Asian economies and 1.464 times the peer median of 3,634. The share of the population undernourished was 10.4%, third of six, down from 15.5% in 2001 and below Pakistan's 16.5% and India's 12%. Those are the readings of a country that has largely solved grain. Now look at its small children. Wasting, the acute measure of a child too light for their height, ran at 10.7% of under-fives in 2022, the highest of the four countries reporting and 1.486 times their median of 7.2%. Stunting was 23.6%, also the highest of the four, even after falling from 51.1% in 2000. Childhood overweight was 1.6%, the lowest of the four and about 15% of the peer median of 10.9%. Together those three readings describe children whose diets are thin rather than merely lean, in a country whose fields are not.

The fields are not the constraint

Start with the part that works.

Cereal yield in Bangladesh was 5,321 kg per hectare in 2023, third of six, 1,687 kg above the peer median of 3,634 and 1.464 times it. Vietnam (5,995) and Indonesia (5,829) are ahead. India (3,602), Pakistan (3,634) and Thailand (3,072) are behind. A country of 173.6 million people is growing grain at a rate three of its five comparators do not match.

The population-level hunger measure agrees. Prevalence of undernourishment, which estimates the share of people whose dietary energy supply falls short of requirements, was 10.4% in 2023, third of six. Pakistan reports 16.5% and India 12%. Bangladesh has come down from 15.5% in 2001, a fall of 5.1 percentage points, or 32.9% of the starting level. It remains 4.1 points above the peer median of 6.3%, at 1.651 times it, so this is not a finished job. It is a mid-table position in a group containing two considerably worse cases.

Hold those two readings, because what follows is harder to reconcile with them.

The children are thinner than the country is hungry

Wasting is the acute measure: weight too low for height, measured now. It responds to recent illness and recent intake rather than to a lifetime of them, which makes it the closest thing in nutrition statistics to a live reading.

In 2022, wasting affected 10.7% of Bangladeshi children under 5, the highest of the four countries reporting. Indonesia recorded 7.5%, Thailand 7.2%, Vietnam 4.4%. The peer median is 7.2%, putting Bangladesh 3.5 percentage points above it with 1.486 times the median share of wasted children. More than one child in ten.

Source: World Bank World Development Indicators, prevalence of wasting (% of children under 5, SH.STA.WAST.ZS), 2022.

The limit on that claim belongs in the same breath as the claim. India and Pakistan have no reading in this vintage, so the child anthropometry comparison rests on four countries, not six. They are also the two comparators that report more undernourishment than Bangladesh (16.5% and 12%) and more anaemia among women (53.7% and 46.5%). If they measured their under-fives in the same year, Bangladesh might well not be first.

So the defensible statement is narrower than the headline. Bangladeshi under-fives are the most wasted of the four countries with a 2022 reading, by 3.5 points over the median of the other three, and that sits beside a national hunger rate of 10.4% that is third of six rather than the group's worst. The second half of that sentence is the surprising part, and it does not depend at all on who is missing from the first.

The least overweight children of the four, which is the same finding twice

Childhood overweight in Bangladesh was 1.6% of under-fives in 2022, the lowest of the four countries reporting: 9.3 percentage points below the peer median of 10.9%, and about 15% of it. Vietnam recorded 11.8%, Thailand 10.9%, Indonesia 3.6%.

Source: World Bank WDI, prevalence of overweight (% of children under 5, SH.STA.OWGH.ZS), 2022.

This is usually filed as good news, and in one narrow sense it is. Vietnam at 11.8% and Thailand at 10.9% are carrying a real cost that Bangladesh is not. But the reading does not sit on its own. Within these four countries the two ends line up: Vietnam has the lowest wasting (4.4%) and the highest childhood overweight (11.8%), Bangladesh the reverse on both. Four observations are not a relationship and no causal claim can be built on them. What they do show is that Bangladesh is not a country eating carefully. It is a country whose children's diets have almost no slack in either direction, which is what a thin diet looks like when you measure both tails.

The honest boundary: this pack carries no adult obesity, birthweight or non-communicable disease series. Whether children who are this light at two become metabolically vulnerable adults is a live question in the literature and it is not one these seven indicators can answer. The essay stops where the data does.

Stunting halved, and stayed at the top of the four

Stunting is the chronic measure: height too low for age, the accumulated record of everything that went wrong in the first years of life.

Bangladesh recorded 23.6% of under-fives stunted in 2022, the highest of the four reporting, against Indonesia 21.6%, Vietnam 18.9% and Thailand 12.4%. That is 4.7 percentage points above the peer median of 18.9% and 1.249 times it, a much narrower gap than the one on wasting.

The trend is the strongest result in this essay. Stunting was 51.1% in 2000. The fall to 23.6% is 27.5 percentage points, a decline of 53.82%, leaving the 2022 rate at 0.4618 of the 2000 level. Just under a quarter of Bangladeshi children under 5 are stunted today. In 2000 it was just over half.

Source: World Bank WDI, prevalence of stunting (% of children under 5, SH.STA.STNT.ZS), Bangladesh against comparators, 2000-2022.

The divergence between the two child measures is the whole argument. Stunting is a stock: it falls as each successive birth cohort spends less of its early childhood in deficit, and it responds to slow things: household income, sanitation, maternal schooling, birth spacing. Wasting is a flow, generated fresh every season by illness, by the price of an egg, by what is actually put in front of a two-year-old this month. A country can drive the stock down by more than half over two decades and still be running the flow at 10.7%. That is the configuration in the data, and it means the machinery that produced the stunting decline is not the machinery that would close the wasting gap.

Which of those slow drivers did the work is not something this pack can settle: none of them are in it. The claim here is only that the two child measures moved differently, which they demonstrably did.

Where the constraint plausibly sits

Two indicators point past the grain balance.

Anaemia among women aged 15-49 was 37.6% in 2023, third of six, behind India's 53.7% and Pakistan's 46.5%, and above Indonesia's 26.7%, Thailand's 20.7% and Vietnam's 20.5%. Bangladesh is 10.9 percentage points over the peer median of 26.7%, at 1.408 times it. The mechanism that matters for children is upstream of them: a woman's nutritional status before and during pregnancy sets the birthweight her child starts from, and a child born small begins the wasting count already behind. This pack contains no birthweight series, so that chain is imported from outside the data rather than shown by it.

Food imports were 13% of merchandise imports in 2018, the highest of the six, 5.07 percentage points above the peer median of 7.93% and 1.64 times it. Indonesia 10.21%, Pakistan 10.04%, Vietnam 7.93%, Thailand 6.0%, India 3.85%.

Two qualifications, both material. The denominator is the whole merchandise import bill, so a country with a narrow import basket posts a high food share while importing modest quantities per person; this indicator measures the composition of imports, not food dependence per head. And 2018 is the latest year the series covers here, four years before the child readings and five before the hunger and anaemia ones. What survives is a pattern rather than a proof: the economy with above-median cereal yields is also the one whose import bill leans most heavily toward food, which is what you would expect if the grain is grown at home and much of the rest of the diet is bought.

Put the pieces in order and the shape is consistent. Cereal output per hectare is above the peer median and the share of people short of dietary energy is mid-table. The foods that carry protein, iron, zinc and fat into a small child, fish, eggs, pulses, dairy, oil, are the ones a poor household buys last and drops first when money is tight. A national energy balance can look adequate while the diet of the household member with the smallest stomach and the largest requirement per kilogram of body weight does not.

What would change this conclusion

A specific, checkable test. This essay claims aggregate food supply is not the binding constraint on Bangladeshi child nutrition. If a future WDI vintage shows undernourishment falling from 10.4% toward the peer median of 6.3% and wasting falling with it, below 7.2%, then the two were one problem, supply was the constraint, and the argument here named the wrong mechanism. If undernourishment closes while wasting stays near 10.7%, they are separate problems, and only one of them currently has an owner.

Three moves, with owners and signals.

1. Measure wasting often enough to see it. Wasting is acute and seasonal, so a single survey round cannot distinguish a lean-season spike from a trend, and a prevalence figure pinned to one fieldwork window will be over-read in both directions. Owner: Bangladesh Bureau of Statistics, with the national nutrition surveillance function. Success signal: a published wasting series carrying at least two readings a year with fieldwork month recorded, which would make the 10.7% figure interpretable for the first time.
2. Treat maternal nutrition as a child nutrition instrument, not a women's health line item. At 37.6%, anaemia among women aged 15-49 runs 1.408 times the peer median, and it sits upstream of the birthweight that starts the wasting count. Owner: Directorate General of Health Services. Success signal: the anaemia reading below the current peer median of 26.7% in a later WDI vintage.
3. Fund the non-cereal side of the diet. Cereal yield at 5,321 kg per hectare is 1.464 times the peer median, so the grain side of the food system is not the part that is short. Fish, eggs, pulses and dairy carry the protein and micronutrients that grain does not. Owner: Ministry of Agriculture with the Ministry of Fisheries and Livestock. Success signal: wasting below the current peer median of 7.2% in the next vintage that reports it.

The counterargument

The strongest objection is that the comparison is rigged by missing data. The three child indicators carrying this essay report for four countries, and the two absent comparators are precisely the ones most likely to displace Bangladesh from the top: India and Pakistan report more undernourishment (12% and 16.5%) and more anaemia among women (53.7% and 46.5%) than Bangladesh does. On that reading, "highest of the four reporting" is an artefact of who filed a survey, and Bangladesh has been convicted on a technicality.

Concede it, because it is correct as far as it goes. Nothing here establishes that Bangladeshi child wasting is the worst in South and Southeast Asia, and this essay does not say so. The regional superlative is not available from four countries and should not be claimed from them.

But the objection does not touch the argument, because the argument is internal. Bangladesh's own aggregate measures and its own child measures diverge, and that divergence needs no comparator at all. Undernourishment fell 32.9% between 2001 and 2023 to a mid-table 10.4%. Cereal yield stands at 1.464 times the peer median. Stunting fell 27.5 percentage points since 2000. Against all of that, 10.7% of under-fives are wasted and childhood overweight sits at 1.6%. Whether that wasting rate would rank first or third in a fully reporting group of six changes the headline and changes nothing about the mechanism: a country cannot grow its way out of a deficit that is about composition, timing and the household's youngest member, and grain yields are the one thing here that is already working.

Two final limits. Six countries is a comparison, not a sample, and none of the associations in this essay identify a cause. And the readings span 2018 to 2023 across different survey instruments, so any cross-indicator statement here is a statement about a period, not about a single moment.

---

Data sources: World Bank World Development Indicators, retrieved from the BDPolicyLab data lake, 2026-08-10. Comparators are India, Pakistan, Indonesia, Thailand and Vietnam; "peer median" is the median of the comparators reporting in the stated year, excluding Bangladesh. Years differ by indicator because coverage does: wasting, stunting and childhood overweight are 2022, undernourishment and anaemia are 2023, cereal yield is 2023, food imports are 2018. Wasting, stunting and childhood overweight report for four countries in 2022 (Bangladesh, Indonesia, Thailand, Vietnam); every comparison drawn from them is stated on that basis. Population is 2024.

Sources

1. World Bank WDI, Prevalence of wasting (% of children under 5), SH.STA.WAST.ZS: https://data.worldbank.org/indicator/SH.STA.WAST.ZS
2. World Bank WDI, Prevalence of stunting (% of children under 5), SH.STA.STNT.ZS: https://data.worldbank.org/indicator/SH.STA.STNT.ZS
3. World Bank WDI, Prevalence of overweight (% of children under 5), SH.STA.OWGH.ZS: https://data.worldbank.org/indicator/SH.STA.OWGH.ZS
4. World Bank WDI, Prevalence of undernourishment (% of population), SN.ITK.DEFC.ZS: https://data.worldbank.org/indicator/SN.ITK.DEFC.ZS
5. World Bank WDI, Prevalence of anemia among women of reproductive age (%), SH.ANM.ALLW.ZS: https://data.worldbank.org/indicator/SH.ANM.ALLW.ZS
6. World Bank WDI, Cereal yield (kg per hectare), AG.YLD.CREL.KG: https://data.worldbank.org/indicator/AG.YLD.CREL.KG
7. World Bank WDI, Food imports (% of merchandise imports), TM.VAL.FOOD.ZS.UN: https://data.worldbank.org/indicator/TM.VAL.FOOD.ZS.UN