The Hershey Company

Hershey makes the chocolate and sweets that sit by the checkout in almost every American supermarket, petrol station and pharmacy: Reese's peanut-butter cups, Hershey's bars and Kisses, Kit Kat (under licence from Nestlé in the US), York, Almond Joy, Jolly Rancher, Twizzlers and Ice Breakers. Since 2018 it has added salty snacks — SkinnyPop popcorn, Dot's pretzels and, in November 2025, LesserEvil. It sold $11.7bn in 2025, about 88% of it in the United States. A ten-year-old's version: Hershey makes the chocolate you grab while your parents are paying, and the chocolate you hand out at Halloween.

The business has three distinctive features. It is seasonal: Halloween, Christmas, Valentine's Day and Easter account for a large share of sales and an even larger share of the special packs that carry the best prices. It is concentrated: one distributor, McLane — which supplies Walmart, and which happens to be owned by Berkshire Hathaway — accounts for about 27% of its sales. And it is controlled: the Hershey Trust, which runs the school that Milton Hershey endowed with his fortune in 1918, holds about 79% of the votes.

In 2024 and 2025 the price of cocoa, Hershey's one indispensable ingredient, rose from about $2,500 a tonne to an average of more than $8,000, with a daily peak near $12,900. Hershey's reported earnings fell 60% in 2025. In 2026 they are recovering fast — the company guides adjusted earnings up a third — but the shares, at $162.74 on 9 October, are back near their low and 32% below the high they reached in the spring. This report asks why, and whether the market or the company has it right.

Two lessons for an owner. The first is that Hershey has been tested by cocoa before and passed: the price of the bean has spiked several times since 1930, and each time Hershey raised prices, lost some volume for a year or two, and came out with its margins intact. The 2024–25 spike was the most violent ever — a tripling in eighteen months — and the recovery is following the same script, faster than the market seems to credit. The second is that the Trust is both shield and sword. It has twice stopped Hershey being sold, which protects the long-term franchise and denies shareholders a takeover premium; and when it decides to diversify, as it did this year, its selling weighs on the price for everyone else.

Why chocolate is a wonderful business in an ordinary year. A Reese's cup costs Hershey cents in peanuts, sugar, milk and cocoa; the shopper pays a dollar or two for it at the checkout without thinking. In 2024 Hershey's gross margin was 47.3% and its North American confectionery segment earned about 32 cents of every dollar of sales — margins comparable to good software companies, earned on a product unchanged for a century. The business needs little capital (capex of $425–475m planned for 2026, about 4% of sales) and turns most of its profit into cash.

Why it is a hard business in an extraordinary year. Cocoa and its derivatives are the largest single input, and Hershey cannot substitute for them in a Hershey's bar. When the bean tripled, the company had three tools: hedges, which delayed the blow by a year and produced violent mark-to-market swings in reported earnings ($2.26 a share of gains in 2024, $2.08 of losses in 2025); price, which it raised by low double digits through 2025 and into 2026; and pack size and mix. In 2025 those were not enough: commodity costs rose $737m and tariffs and supply-chain costs another $287m, and the gross margin fell to 33.5%.

The recovery arithmetic. In the first half of 2026 cocoa averaged about $3,970 a tonne, half its 2025 level, and Hershey's adjusted gross margin rose to 41.6% in the second quarter. Management guides 2026 sales up 4.5–5%, adjusted EPS of $8.36–8.52 (+32.5–35%), and expects "good visibility into deflation next year from cocoa". The catch is the bean's rebound since March, which trims how much deflation 2027 brings.

Buffett learned the economics of brands from a box of chocolates. Of See's Candies, bought in 1972, he wrote in 1991 that "we saw that the business had untapped pricing power", and in 2007 he called it "the prototype of a dream business": $32m of reinvestment had produced $1.35bn of pre-tax earnings. His definition of an economic franchise, from the same 1991 letter, is a product that is needed or desired, "thought by its customers to have no close substitute", and not subject to price regulation. A Reese's cup passes all three. The 2024–26 cocoa shock was the first real test of that idea in a generation, because Hershey had to raise prices far more than inflation. The evidence is mixed in the way a genuine moat's evidence usually is: shoppers bought less, but they did not switch to cheaper chocolate — private label barely registers in the category — and every large competitor faced the same bean. Where Hershey has lost share in 2026 is to newer products and flavours from rivals, not to price. We score the moat 8: as wide as it was, with less room to raise prices again soon.

What happened, in order. West African cocoa harvests failed in 2023–24 — disease, old trees, bad weather and smuggling — and New York cocoa futures rose from about $2,500 a tonne to a record near $12,900 in December 2024. Hershey's hedges delayed the pain, so its 2024 margins were near record highs while the cost of beans it would use in 2025 was locking in. In 2025 the bill arrived: gross margin fell 14 points, adjusted EPS fell from $9.37 to $6.31, and Hershey did not raise its dividend. Then supply responded. Two seasons of surplus pushed cocoa below $4,000 in January 2026 and to about $2,850 in March, and Hershey's shares peaked near $239 as investors pencilled in a bumper 2027.

Why the shares fell back 32%. Four things, all since March. The bean bounced: an El Niño and forecasts of smaller Ivory Coast and Ghana crops took cocoa back to about $6,000 in July and $5,670 on 9 October, so the "deflation tailwind" for 2027 shrank — Piper Sandler cut its target on exactly that. Volume fell harder than planned: in the second quarter North American confectionery volume was down about 10 points against a 14-point price rise, and Hershey reported share losses. The Trust began selling on 17 March. And analysts reset: Stifel cut to $180 citing "worsening elasticities and declining market share trends"; consensus is now 9 buys, 23 holds and 3 sells.

Our reading: the brand survives the bean; the market is pricing the bean, not the brand. The 2025 earnings collapse was a cost shock, not a demand collapse — sales rose 4.4% that year — and the recovery in 2026 is following the cost curve with a lag, exactly as hedging implies. At $5,670 a tonne cocoa is roughly 30% cheaper than Hershey's 2025 average cost, so 2027 should still bring lower bean costs, just less dramatically than the March futures promised. What deserves real attention is the volume: a 10-point drop is larger than in past cocoa cycles, and the slow drags — weight-loss drugs, lower food-stamp benefits, consumers trading down — make it plausible that some of it does not come back. That would leave Hershey a business with a 42–44% gross margin rather than 47%, still an excellent one. The next test is Halloween, the biggest selling season of the year, reported with the third quarter in late October.

The pattern. US chocolate is an oligopoly of three — Hershey, Mars and Ferrero — and all three faced the same bean, so none could undercut the others on price for long. That is why Hershey's volumes fell but its brands did not lose to cheaper copies. The competition that matters is innovation: new flavours, formats and limited editions that win the retailer's attention and the impulse shopper's eye. Ferrero in particular has spent heavily on renovating the American brands it bought from Nestlé, and Hershey's own admission of share loss in 2026 points there. Tanner's answer — that competition is not about price discounts and buy-one-get-one-free — is right for a brand owner, but it puts the burden on Hershey's new-product pipeline, historically not its strength. The comparison on our board is PepsiCo, Tanner's old employer, whose snack franchise Hershey's salty business is trying to emulate at a twentieth of the scale.

Integrity: no veto. Hershey reports its hedge gains and losses separately and plainly, and its disclosure of share losses in the July release is the kind of candour many consumer companies avoid. Capital allocation: conservative, with one misstep. Pausing the dividend raise, keeping leverage under two times and the A rating, and buying a growing brand are what an owner would do; repurchasing $439m of shares at ~$196 while the Trust sold at the same prices is not — the company and its controlling owner were, in effect, on opposite sides of the same trade. Owner mentality: real at the top. Tanner and two senior colleagues bought shares in November 2025; no insider has bought in 2026. We score management and capital 6, held back by how new the team is.

Which earnings to believe. Hershey's reported earnings are unusually noisy because it marks its cocoa hedges to market every quarter: in the second quarter of 2026 reported EPS was $2.26, up 629%, while adjusted EPS was $1.90, up 57%. Neither is wrong; the adjusted figure, which books hedge results when the cocoa is actually used, is the better guide to the business, and it is the figure management guides to. Over the cycle the two converge — 2024's gains and 2025's losses roughly cancel.

The cash is better than the earnings right now. Operating cash flow in the first half of 2026 was $888m, against $509m a year earlier, and free cash flow over the last twelve months about $2.2bn — more than adjusted net income — because inventory bought at lower cocoa prices releases working capital. That flatters 2026; it will not repeat once costs stabilise. Over a full cycle Hershey turns roughly 100% of adjusted net income into free cash flow, which is how we count owner earnings below.

The verdict on the dividend: safe, and growing again. This is the dividend test a cocoa crisis was designed to fail, and Hershey passed it: through the worst ingredient shock in its history the payout stayed covered 1.7 times by free cash flow, the balance sheet stayed single-A, and the board's only concession was a year without a raise — the first since 2009. At $1.452 a quarter the yield is 3.6% — well above the 2–3% Hershey has usually offered over the past decade — and it takes about 69% of this year's adjusted earnings. Expect raises roughly in line with earnings from here, with the next declaration most likely alongside the third-quarter results. The comparison on our board is Coca-Cola, whose 2.8% yield and ~75% payout come with a cheaper input bill — and no trust selling shares.

What is it worth? We discounted owner earnings of about $9.00 a share at 10% for ten years. If cocoa settles around $5,000–6,000, volumes come back slowly and earnings grow ~7% a year for five years (the recovery plus management's 6–8% algorithm) and 5% thereafter, ending at 18 times, Hershey is worth about $187. If the bean stays expensive and weight-loss drugs bite — $8.00 of owner earnings growing 3% a year, 15 times — about $119. If cocoa normalises toward $4,000 and margins return to 2024 levels — $9.50, 9% then 6%, 21 times — about $245, roughly the spring high.

At $162.74 the price sits about 13% below our central value, for a franchise whose moat we score 8 and whose dividend survived the worst year in its history without a cut. That is a fair price for a wonderful business at a cyclical low, and our rules say not to demand a bargain for it: begin accumulating, and add more below $150 if Halloween disappoints or the Trust starts converting Class B shares. An investor who buys here earns a 3.6% yield, growing again, while waiting for the cost curve to finish turning.