The Bitter Bean: Chocolate's Child-Labour Supply Chain

Most of the chocolate eaten in the world begins its life in a hot, humid belt of West Africa, on smallholdings in Ivory Coast and Ghana that between them grow around six in every ten cocoa beans on Earth. The pods grow straight from the trunk and branches of the cacao tree, heavy and ridged, and they are harvested with a machete or a long-handled blade, then split open to scoop out the pale, pulp-covered seeds inside. It is skilled, exhausting, seasonal work, done for very little money on farms often no larger than a few hectares. And on a great many of those farms, some of that work is done by children — sharpening blades, hauling loads heavier than they should carry, spraying agrochemicals without protection, missing school. This is not a rumour. It is one of the most thoroughly documented facts in the entire global food economy, and it has survived twenty years of solemn promises to end it.

The promise that keeps being renewed

The modern chapter of this story opens in 2001. Investigative journalism in the late 1990s and around the turn of the millennium — reporting by news organisations in the United States and Britain that traced the beans in Western chocolate back to West African farms worked by trafficked and unpaid children — created enough public pressure that the American chocolate industry faced the prospect of legislation. A congressman, Eliot Engel, and a senator, Tom Harkin, drafted a measure that would have required a federal “no child labour” label on chocolate.

The industry, facing a mandatory law, negotiated a voluntary one instead. The result was the Harkin-Engel Protocol of September 2001, signed by the major chocolate and cocoa companies. It committed the industry to eliminate the worst forms of child labour from West African cocoa. The original target date was 2005. When 2005 arrived and the problem had not gone away, the deadline moved to 2008. Then to 2010. Then a fresh framework set a goal of reducing the worst forms of child labour by seventy per cent by 2020. That target, too, was missed. More than two decades after the protocol was signed, the practice it promised to end is still there, at a scale the industry’s own commissioned research keeps confirming.

The repeated slippage is the part worth holding onto, because it is more revealing than any single statistic. This was not a case of a hidden problem that no one had gauged. The industry knew, the public knew, deadlines were set in writing and signed by chief executives, and still the beans kept coming. Whatever the obstacle was, it was not ignorance.

What the counting actually found

The most authoritative measurements come from a series of studies carried out by researchers at Tulane University and, later, by the survey organisation NORC at the University of Chicago, commissioned in part with United States government funding to check whether the protocol’s goals were being met.

The headline finding of the NORC study published in 2020 was stark. In the cocoa-growing regions of Ivory Coast and Ghana, roughly 1.5 million children were engaged in child labour on cocoa farms — and the great majority of them in what the International Labour Organization defines as hazardous work: using sharp tools, carrying heavy loads, applying agrochemicals, clearing land with machetes, working long hours that kept them out of school. The proportion of children in cocoa-growing households doing such work had, if anything, risen over the preceding decade rather than fallen, even as cocoa production expanded. The seventy-per-cent reduction that had been promised for 2020 had not been delivered; the numbers had gone, broadly, the wrong way.

It is important to be precise about what these numbers describe, because the precision is where honest reporting parts company with the slogan. The overwhelming majority of this is child labour on family farms — children working, often alongside parents, in conditions that are hazardous and that rob them of schooling. A much smaller and separate category is outright trafficking and forced labour: children brought across borders, sometimes from Mali or Burkina Faso, and made to work without pay under coercion. Both exist. Both are real. But they are not the same thing, and the difference matters enormously for what any remedy could look like. The word that collapses them — “slavery” — is emotionally true to the trafficking cases and misleading about the far larger number of children working on their own families’ land because the family is too poor to hire help or pay school fees.

The courtroom and the limits of the law

In June 2021 the question reached the highest court in the United States. In Nestlé USA v. Doe, six men from Mali who said they had been trafficked as children and forced to work on Ivorian cocoa farms sued the American chocolate companies whose supply chains, they argued, had been sustained by that labour. The case turned on a technical but consequential point of law: whether the plaintiffs could bring their claim in a US court under the Alien Tort Statute, an eighteenth-century law occasionally used to sue over human-rights abuses committed abroad.

The Supreme Court ruled, by a wide majority, that they could not — because the specific conduct alleged against the companies inside the United States amounted, in the Court’s view, to general corporate decision-making rather than to the abuses themselves, which had happened overseas. The ruling was about jurisdiction and the reach of a particular statute; it was not a finding that the child labour had not occurred, nor an exoneration of the industry’s conduct. But its practical effect was to close one of the few avenues by which trafficked workers might have held the buyers of their labour accountable in a Western court. It is a good illustration of a recurring feature of this whole story: the gap between what is known to be true and what any institution is actually able or willing to do about it.

What certification does, and does not, promise

Faced with all this, the well-meaning shopper reaches for the reassuring label — Fairtrade, Rainforest Alliance, UTZ, a brand’s own “sustainably sourced” seal — and here the story asks for more care than either the industry or its fiercest critics tend to offer.

Certification schemes are not nothing. At their best they pay farmers a premium, fund community projects, run monitoring systems designed to catch and remediate child labour, and give a company a reason to know where its beans come from. But they do not, and cannot honestly, guarantee that no child touched the cocoa in a given bar. Cocoa from certified and uncertified farms is routinely mixed in the supply chain; “mass balance” accounting means a company can buy a certified quantity without every bean in a specific bar being traceable to a certified farm. Monitoring systems catch a fraction of what happens on hundreds of thousands of scattered smallholdings. Investigations over the years have found certified farms using child labour, and have found the audit systems stretched far too thin to police the territory they claim to cover. A label is a signal of effort and a mechanism for improvement. It is not a warranty of a clean conscience, and treating it as one is its own small self-deception.

The poverty at the root of it is not vague. For much of the past decade the typical cocoa smallholder in Ivory Coast or Ghana has earned well below the World Bank’s extreme-poverty line from the crop, a farm-gate price squeezed at the bottom of a chain whose profits accumulate far downstream, in the grinding, branding and retailing done elsewhere. In 2019 the governments of the two countries tried to force the issue directly, jointly imposing a “Living Income Differential” — a fixed premium of 400 dollars a tonne that buyers would have to pay on top of the market price, explicitly to lift farmer incomes. The major traders and manufacturers resisted, sought ways around it, and the mechanism has delivered less than its architects hoped, partly because a premium on the price does little if the underlying price then falls or if middlemen absorb the difference. The attempt is instructive precisely because it aimed at the actual cause. Child labour on cocoa farms is, above almost everything else, a symptom of farm-gate poverty, and the interventions that move the needle are the ones that move money towards the farmer.

The fork: from real problem to hopeless slogan

Everything above is documented and sober. Now the point where clear eyes start to blur.

Because the underlying facts are so grim and so well-established, the story is unusually prone to a particular kind of overshoot, and the overshoot travels faster than the nuance. The slogan version says: chocolate is slavery, every bar is soaked in child suffering, the situation is hopeless, and the only ethical response is to boycott chocolate altogether. Each of those moves is understandable and each is, in an important way, wrong.

“Chocolate is slavery” flattens the distinction the evidence insists on — between the large population of children doing hazardous work on their own families’ poor farms and the smaller number who are genuinely trafficked and coerced. Both demand action; conflating them makes the problem easier to feel and harder to fix. “It’s hopeless” is contradicted by the fact that we can measure the problem at all, that some interventions — paying farmers enough that they can afford both labour and school fees, monitoring-and-remediation systems that actually track children back into education — demonstrably reduce it where they are properly funded. And “boycott chocolate” is the move most likely to backfire, because the root cause is poverty. A farmer earning below the poverty line for a day’s work uses child labour partly because he cannot afford anything else; a Western boycott that collapses the price of his crop makes him poorer, and a poorer farmer leans on his children’s labour harder than before. The interventions that have actually helped run in the opposite direction: paying more, buying more transparently, and staying engaged rather than walking away.

The cost the label never mentions

What makes this story so uncomfortable is that it is not a conspiracy in any of the usual senses. Nothing is hidden. The studies are public and government-funded; the deadlines were announced with fanfare; the companies file sustainability reports thick with the language of concern. The scandal is not concealment. It is the strange, durable gap between a problem that everyone acknowledges and a solution that no one, in twenty years, has been willing to pay the full price of — because the full price is fair prices to farmers, which means lower margins or costlier bars, and the appetite for that has never quite arrived.

The believer who insists that your chocolate is soaked in suffering is closer to the truth than the shopper who trusts the label and thinks no more about it. The correction the record asks for is not to look away but to look more exactly: to see that the cost is real, that it lands on children, that it is measured and named and unresolved — and that the honest response, past both despair and the comfort of a seal on a wrapper, is a stubborn attention to who actually gets paid. The same opacity that lets cocoa hide its cost is what lets olive oil hide its origins and honey hide where it was really made: a supply chain long enough that, somewhere in the middle of it, the truth about a thing quietly goes missing. With chocolate the missing truth is not what the product is. It is who paid for it, and how much they were spared.