Hidden hunger is a supply problem before it is a diet problem

In 2024, an estimated 673 million people, 8.2 percent of the world, experienced hunger. In Africa the proportion passed 20 percent, affecting 307 million people. Around 2.3 billion people were moderately or severely food insecure.¹ Those are the figures that make headlines, and they describe an insufficiency of food.

Underneath them sits a second problem that headlines handle badly. A modelling analysis published in The Lancet Global Health in 2024 estimated inadequate intake of 15 micronutrients across 185 countries and 34 age and sex groups, and found that more than half the global population consumes too little of several of them. Inadequate intake was estimated at 68 percent for iodine, 67 percent for vitamin E, 66 percent for calcium and 65 percent for iron, with riboflavin, folate, vitamin C and vitamin B6 each above 50 percent.²

These are estimates of dietary intake rather than measured deficiency in blood, and the authors are explicit about that limitation. The figures nonetheless describe something that the word hunger does not capture. A person can meet their energy requirement, not be counted in any undernourishment statistic, and still be short of the micronutrients that determine cognitive development, immune function and maternal outcomes.

The numbers are larger than the framing

The term hidden hunger is useful shorthand and slightly misleading. It suggests a condition that is concealed, when in most cases it is simply not looked for. Micronutrient status is rarely measured at population scale because biomarker surveys are expensive, which is precisely why modelling studies like the Lancet Global Health analysis exist. The estimates are imperfect and they are the best available picture of a problem that would otherwise be invisible by default.

The analysis also found that the pattern is not uniform by sex. Women were estimated to have higher inadequate intakes of iodine, vitamin B12, iron and selenium, while men had greater inadequacy for calcium, niacin, thiamin, zinc, magnesium and vitamins A, C and B6.² That matters for programme design, because a single fortification vehicle aimed at a household does not deliver the same correction to everyone in it.

Why this is not mainly about choice

The reflexive framing for nutrition shortfalls in high-income media is behavioural. People should eat more vegetables, more pulses, more fish. The advice is not wrong and it is close to irrelevant at the scale these numbers describe.

A diet's micronutrient density is largely determined before anyone chooses anything. It is determined by what the local food system produces, by what the soil those crops grew in contained, by what survives processing and storage, by what is in the market and by what it costs relative to a staple. A household that meets its calories from maize meal is not making a nutritional error. It is buying the cheapest available energy, which is what people do when energy is the binding constraint.

A household meeting its calories from maize meal is not making a nutritional error. It is buying the cheapest available energy.

This is where my own field background changes how I read the literature. The crops that dominate smallholder systems dominate because they are reliable, storable and tolerant of poor soil, not because they are nutritionally complete. Diversifying a diet means diversifying what somebody can produce or buy, and both of those are agricultural and economic questions before they are dietary ones.

The affordability line

The affordability data makes the point more directly than any argument. In 2024, 2.60 billion people could not afford a healthy diet, down from 2.76 billion in 2019. But that global improvement conceals a divergence. In low-income countries the number rose over the same period, from 464 million in 2019 to 545 million in 2024.¹

Read that against the micronutrient estimates and the mechanism becomes hard to avoid. Where a nutritionally adequate diet costs more than a household earns, intake shortfalls are a price outcome, not a preference outcome. Nutrition education aimed at such a household is asking it to solve an arithmetic problem with information.

What actually moves the number

Three categories of intervention have a plausible claim on the problem at scale, and they operate at different points in the chain.

1. Fortification of widely consumed staples. The advantage is that it requires no behaviour change at all, which is also the reason it is the intervention most consistently associated with population-level shifts. Its limit is that it reaches only people who buy the fortified vehicle, which in subsistence-heavy contexts can exclude the households most at risk.
2. Biofortification and production diversification. Changing what is grown changes what is available before price enters the picture. This is slower and depends on varieties that farmers will actually adopt, which returns the question to agronomy and to whether the biofortified variety yields competitively.
3. Income and price policy. Unfashionable in nutrition writing and probably the largest lever, given that the affordability figures move with prices and incomes rather than with knowledge.

None of these is a complete answer and the evidence on each is contested in ways a short piece cannot settle. The point is that all three act on supply, availability or price. The behavioural layer sits on top of them and can only redistribute what the system has already made available and affordable.

A note on how this gets written about

Two failures recur in coverage of this subject. The first is treating modelled intake estimates as measured deficiency prevalence, which overstates what the data can support. The 2024 analysis estimates inadequate intake, and inadequate intake is a risk indicator rather than a diagnosis. The second is the slide from a population-scale structural finding into individual-scale advice, which is how a paper about 185 countries becomes an article about eating more leafy greens.

Both failures are avoidable by saying what the study measured before saying what it means. That is a low bar, and a surprising amount of nutrition journalism clears it only occasionally.