Cocoa costs about four times what it did in 2020. That figure shows up in most of the stories about why the Easter chocolate costs more this year. It is the wrong number to stare at.
The number worth staring at is the thirty years before it. From the early 1990s through 2023, cocoa mostly sat in a dull band of a few thousand dollars a tonne. It barely moved while the world around it did. A raw material that holds still for three decades looks like a settled, healthy market.
That stillness was the strange thing, not the spike. The cheap years were not proof the system worked. They were a loan taken out against the farm, the tree, and the weather. The cruel detail is that the money now flooding into cocoa does not reach the people who grew the beans, at least not cleanly and not in time to repair what wore down while the price held still.
Start with where a chocolate bar's money actually goes. Studies of the French and German markets put roughly 70 percent of a bar's value with brands and retailers, and about 11 percent with the farmers who grow the beans. The person doing the hardest, least substitutable work sits at the thin end of the chain. The beans are still cut from the tree by hand, split, then fermented and dried on mats until their moisture falls from about 55 percent to 7.5 percent, work no factory has taken over. That split matters, because resilience is not free, and in any supply chain it has to be paid for by someone. Renewing trees before they age out, controlling disease, keeping growers solvent enough to reinvest: all of it costs money, and the cheapest way to keep a price low is to skip it.
For thirty years, that is roughly what happened. Call it the resilience discount. A price is running one when three things are true. The low price depends on deferred upkeep of the thing that produces the good. The party absorbing that deferral captures the smallest slice of the price. And the calm is not strength, only the absence of a bad year. Cocoa met all three.
The upkeep was deferred in plain sight. Analysts date the last major replanting push in West Africa to the early 2000s, which left much of the crop growing on tired, past-peak trees. Growers earning a sliver of the final price had little to reinvest, so the trees aged, and a virus called swollen shoot, spread by mealybugs that no pesticide handles well, moved through the farms. One brutal defense is to cut the sick trees out. Ghana alone has lost more than 254 million cacao trees in recent years. None of that was a secret. It simply did not show up in the price, because a discount by definition hides the thing you stopped paying for.
Then the bad year arrived, and it arrived as several at once. Heat above the level cacao likes, around 32 degrees Celsius, stretched across more of the main-crop season. The rain came in the wrong amounts: too much late in 2023, floods that drowned some plantations and delayed the harvest, then long dry stretches where some growers counted only a couple of rains across whole months. Disease and old trees turned a stressed crop into a short one. Because roughly two-thirds of the world's cocoa grows in a handful of West African countries, a regional failure became a global price. The 2023/24 season ran a deficit of somewhere between 440,000 and 480,000 tonnes, and world stocks fell to around 27 percent of yearly grinding, the thinnest cushion in about forty-five years. Futures did the rest, roughly quadrupling, and swinging so hard that even traders who were hedged had to post more cash just to hold their positions.
The same mechanism lands on three very different players, and who can absorb a shock is the whole story.
The big manufacturers are the party best placed to absorb the shock, which is exactly why their near miss is instructive. They are large branded firms with the cash to hedge, sitting up at the end of the chain where brands and retailers together take most of a bar's value. Going into the spike they had bought themselves time, and the raw cost took months to reach the shelf. It reached it anyway. One of the largest reported its quarterly gross margin falling six and a half points on what it flatly called unprecedented cocoa inflation, and guided its earnings down about 10 percent for the year. Hedging did what a buffer does: it delayed the hit and softened it. It did not cancel it. What money bought at the top of the chain was time, not immunity. The manufacturers were never really carrying the discount; they were paying to postpone a price shock. The growers had been paying it down for thirty years, in trees.
At the bottom of the buying side is a shop with none of that. In Mequon, Wisconsin, a chocolatier who opened her small-batch store in March of 2020 now pays about 200 percent more for cocoa than she did that first year. She cannot hedge a year of futures, and she cannot reformulate her way out, because the ingredient is the product. Her words: "I gotta buy the chocolate, that is kinda a fixed cost for me. I need the chocolate. I can't go without it." There is a ceiling on what her customers will pay, so her margins are thin, and her real adjustment is quieter than a price sticker. She watches which chocolates sell, and stops making the ones that no longer earn their place in the product line. A global commodity shock reaches her as a decision about which of her own creations should stop existing.
Then there is the party that carried the discount the whole time: the grower. This is where the intuitive story falls apart. A record world price sounds like a windfall for whoever grows the crop. On the farm, it does not feel like one.
In southwestern Cote d'Ivoire, a farmer cooperative leader describes land that once yielded about 600 kilograms of beans per hectare now barely giving 50, with even four or five hectares struggling to add up to a single tonne. In Ghana, a grower watched his 15-hectare plot fall from 50 bags of cocoa in 2015 to seven. He is not thinking about a futures screen. He put it plainly: "Before God and man, if they come asking for my farm to mine, I will sell it." That is the logic the cheap years built. After decades of earning too little to reinvest, the land is finally worth more torn up than planted. One 52-year-old woman found that out directly. Illegal gold miners pressured her to sell, and one day in 2023 she found her farm cordoned off by armed guards while bulldozers tore out the trees. Of nearly 6,000, fewer than a dozen were left. "This farm was my only means of survival," she said. "I planned to pass it on to my children."
The growers who keep their trees do get more than they did, but not cleanly, and not on time. Both big producers raised what they pay farmers before Easter. Ghana lifted its farmgate price twice inside a year; Cote d'Ivoire set a record of about $3.09 a kilogram. Real raises, and in each country still far under what the beans were fetching abroad, for different reasons. Ghana sells much of its crop forward, up to 70 percent of it nine to twelve months before harvest, so the farmgate number is locked in before any rally shows up. That structure protects farmers when prices crash. The same structure strands them when prices spike, because the grower is paid under a system priced at yesterday's number. Cote d'Ivoire fixes its price a different way, by state decree, and even at a record many growers there called it too low against what the world was paying. So the raise is real, and it is also late, partial, and thinner than the headline, and it arrives at farms the cheap years already hollowed out. A higher price per bag does not buy back a mature tree already gone.
The stability you are enjoying might just be the sound of a bad year that has not happened yet.
The honest counterargument is that none of this needs a grand theory. This is what commodity cycles do: weather and disease cut supply, prices scream, demand falls, high prices pull in new investment, and the market cools. And by early 2025 it was cooling. Prices came well off their peak, one big bank cut its deficit estimate as shoppers bought less, and forecasters began pencilling in a surplus for the new season. All true. But the cycle explains one thing and leaves another untouched. It explains the spike and the expected relief. It says nothing about why the price sat so low for thirty years, or who was quietly funding that low price, or why the correction reaches growers late, long after the trees that needed the money aged out or came down. A cycle moves a price up and down. It does not decide who carried the risk while the price was low, or who collects while it is high.
So here is a test you can carry out of this essay and use on almost anything you buy cheaply. A price that has not moved in years, sitting on top of a producer at the thin end of the chain, in a system with no slack for a bad year, is not necessarily a good deal. It may be a resilience discount: a cost someone else is absorbing, quietly, until they can't.
The chocolate on the Easter table this year is a small thing. It is also one of the first clear looks at what a cheap thing really costs once the discount expires. Most of the prices in your life still look calm. But calm and cheap are not the same as sound, and the difference only shows up on the day the bill comes due. Somewhere behind the next cheap pleasure you reach for, someone is holding the bad year so you don't have to, and not being paid enough to keep holding it.