Hasbro, Inc. (NASDAQ: HAS) — A Great Franchise, Rented Out to Subsidize a Bad One

Report date: 2026-07-17 | Price reference: $82.09 (close, 2026-07-16) Sector: Consumer Discretionary · Leisure Products (Toys, Games & Trading Cards)

This is an independent analytical piece. Sections 1–15 below carry no recommendation and no price target — they assess embedded expectations and scenarios only. The single exception is the opinion block immediately below (Claude's Take), which is clearly labeled as the author’s own subjective view. Nothing here is investment advice.

⚡ Claude’s Take

The author’s own subjective opinion and general information only — not investment advice, and not a recommendation to buy or sell any security. The analytical body (Sections 1–15) below carries no position and no price target; this block is the one place a view is expressed.

HOLD at $82. A genuinely great franchise, fairly priced, in the hands of demonstrably poor stewards. Accumulate on weakness in the mid-$60s. Not a short — the moat is real and the downside is bounded.

Tag: “The best business in toys, owned by the worst allocator in toys.”

The call. Hasbro is not a toy company with a games division. It is Magic: The Gathering — a 46%-operating-margin franchise that earned $1,006.8M of segment operating profit in FY2025 against $11.1M for the entire company — with a $2.4B toy business bolted on that lost $47.5M at the operating line in Q1 2026 with no impairment to blame, and $3.28B of debt that is the fossil of an acquisition the company has already written off. Magic’s moat is the real thing: not the brand, not the IP, but a 32-year, ~100,000-card backlist that is simultaneously the rules engine, the customer’s collection and a stored asset he cannot liquidate without quitting the game. Valve, Riot, Disney, Lucasfilm and Blizzard have all attacked it with unlimited money and better IP, and all of them failed.

The reason this is a HOLD and not a BUY is that the market is not making the mistake the bull case assumes it is making. Back out the toy stub and the market capitalizes Wizards at roughly 12.5–14.1x FY2025 segment EBIT — which looks cheap against Games Workshop’s ~19.3x until you notice the denominator is a peak. Re-strike it on my normalized durable Wizards EBIT of $750–900M and the multiple is ~14–19x: a reasonable range for a mature franchise-IP publisher (~15–18x) and roughly fair against Games Workshop once you adjust for the three things GAW does not make you buy — a loss-making toy business, $2.77B of net debt, and negative ~$1.17B of tangible equity. My scenario work brackets fair value at roughly $62–101 with a base of $77–85. At $82.09 the stock sits in the upper half of its own fair range. There is no margin of safety here for a company whose management has written off $3.42B in six years against $565.5M of book equity.

The framing: harvest, not compounding — and not a falling knife. Hasbro is monetizing Magic rather than compounding it. Universes Beyond rents other companies’ customer captivity (Final Fantasy, Marvel, The Hobbit, Star Trek) and routes it through Magic for a disclosed and rising fee: management guides Wizards’ margin down from 46.0% to “the low 40% range,” which on the FY2026 guide means selling ~$109M more Magic and earning ~$54M less doing it. Universes Beyond goes from half the slate to four of seven sets in 2026; seven sets, higher prices and less content in the box is the signature of a franchise being harvested. The tape agrees with the diagnosis, not the bull case: the stock’s factor-nearest neighbours are COWZ, SCHD, Aptiv and Lear — a cyclical dividend cash-cow, not an IP compounder. This is not a falling knife (3-year returns are positive, the business is growing, and the drawdown is 22.5% from a five-year high) and not a momentum trade (the Momentum loading is negative and price has broken both the 50- and 200-day EMA). It is a stalled re-rating in a rich-against-nothing, high-idiosyncratic-vol name that is rolling over on stock-specific news while its dominant factor is a tailwind. That last part is the bearish tell: DividendYield was the best-performing style factor of the trailing year (+17.1%, z=1.69) and HAS fell anyway.

What holds me back from AVOID. Three things are genuinely good and I will not pretend otherwise. First, the moat protects share and the base of the profit pool — and that was just tested live: Magic’s own collector boosters fell 60–70% from peak while play boosters held ~$100–120, and Magic still grew +36% in Q1 2026 with backlist at a quarterly record. The speculative tranche has already left, and the franchise grew anyway. That is what a moat is for, and it bounds the downside. Second, the balance sheet is genuinely safe (~2.0x net leverage, undrawn $1.1B revolver, a 4.65% five-year print in March 2026, no near-term wall). Third, the dividend is covered 1.77x by free cash flow at a 3.4% yield. This is not a company that breaks.

What holds me back from BUY. Capital allocation, and it is not close. eOne: ~$4.41B of cash out in 2019, ~$760M recovered, $2.8–3.1B destroyed — roughly a quarter of the current market cap, in one transaction. Six-year cumulative GAAP net income is negative $547.9M. The board deleted ROIC from the long-term incentive plan for the 2024 and 2025 awards — immediately after the largest capital-destruction event in company history — leaving only scale and accrual metrics that an impairment cannot touch. In FY2025 the CEO was paid $18.67M, his highest ever, including a maximum 200%-of-target cash bonus, in a year the company wrote off $1.02B, lost $318.2M and reported −$2.30 of GAAP EPS. Insiders bought $1.12M of stock in five years against $74.5M of sales (66:1); the only three purchases landed in a single five-day proxy-defence week in April 2022, ten weeks after an activist publicly shamed the board for never buying; nobody bought the $38.76 low; and the February 2026 selling cluster (~$38.6M) landed at the 99th percentile of the five-year price range — in the same week the board authorized $1.0B of buyback. Then there is the ~$1B spent on video games since 2018 across six studios with nothing capitalized ever shipped, ~$413M of it still sitting un-amortized on the balance sheet, which the CEO rationalizes as a project where “you probably won’t make billions, but your chances of at least making your money back is much higher” — a ~0% ROIC ambition funded out of a 65–85% ROIC franchise.

Conviction: MEDIUM. The moat call is high-conviction; the normalization is not — Hasbro discloses no Magic units, no players, no price and no player-vs-collector split, so my $750–900M is a reasoned triangulation, not a disclosed number, and reasonable analyses disagree about where in that range it sits.

• Flips me bullish: Wizards’ FY2027 margin guided back toward 46% with the set slate reverting below seven and Universes Beyond back under half — i.e. evidence the franchise is being compounded rather than rented — or the $1.0B authorization actually executed hard into a sub-$70 tape, which would be the first capital-allocation decision in a decade that bought low.
• Flips me bearish: two consecutive periods of declining Wizards Play Network store count or organized-play participation while price and cadence rise. That is the moat unwinding from the periphery inward, and it would show up there before it ever shows up in revenue.

📈 Stock Price Action — Five-Year Event Map

Hasbro has round-tripped violently and arrived nowhere. Over five years to 2026-07-16 the shares fell −12.7% in price ($94.04 → $82.09); the entire +6.7% cumulative total return (+1.30%/yr) came from the $2.80 dividend — earned at ~33% annualized volatility. The five-year path ran from a COVID-boom peak, down 59% to a closing low of $42.88 (2023-11-13), then up ~115% off the April-2025 tariff crash to a five-year high of $105.94 (2026-02-11), and has since given back 22.5% to $82.09. The 52-week range is $70.95–$105.94. One correction matters for anyone reading a chart: February 2026 was a five-year high, not an all-time high — the shares closed at $126.07 on 2019-07-29 and remain ~35% below that. The February “record” is a record only on a dividend-reinvested basis.

Cycle narrative. (1) The July-2021 peak was the top of a COVID toy cycle running on revenue that the eOne acquisition had inflated to $6.4B — a number the company has never approached since. (2) Goldner’s death removed the architect of the “Brand Blueprint” strategy and opened the governance vacuum the activist walked into. (3) Alta Fox’s February-2022 campaign demanded a tax-free Wizards spin-off; it lost the vote decisively (~11.6% support) but drew a 27.6% withhold against its sole target, who retired within three months. (4) The 2023 collapse is the thesis in one bar: a negative pre-announcement, mass layoffs, the eOne liquidation at a $539.0M loss and ~$1.19B of goodwill written off — the market marking the Brand Blueprint to zero. (5) The recovery was real and earned: costs came out and Magic reaccelerated. (6) April 2025’s crash was tariffs, and it produced the $1,021.9M impairment — Hasbro writing off a billion dollars of toy goodwill on a forecast against a tariff regime the Supreme Court then struck down ten months later. (7) The run to February 2026 capitalized Magic’s record year; note that insiders sold ~$38.6M into it and the board authorized $1.0B of buyback at the top. (8) The current drawdown is stock-specific — a cyberattack that cleanly separated the two Hasbros (Magic shipped on cadence; the physical-goods business took the hit) plus sell-the-news on a quarter that had been pre-released on 4/23. Price moves are FACT; the attributed drivers are INTERPRETATION.

1. Executive Summary

Hasbro is two companies stapled together, and only one of them is a business. In FY2025 Wizards of the Coast & Digital Gaming earned $1,006.8M of operating profit on $2,186.9M of revenue — a 46.0% margin — while Consumer Products lost $942.6M on $2,437.6M, and the consolidated company reported total operating profit of $11.1M, a pre-tax loss of $102.0M, and diluted EPS of −$2.30. Strip the $1,021.9M non-cash goodwill impairment and Consumer Products earned $79.3M on $2.4B of revenue — a 3.3% margin. Wizards is 46.5% of revenue and ~87% of adjusted segment profit; on a GAAP basis it is more than 100% of it. Q1 2026 removed the last defence: with no impairment in either period, Consumer Products lost $47.5M at the operating line, against $43.9M in the prior-year quarter.

Magic’s moat is real, and it is one of the best in consumer. In Greenwald’s taxonomy it is demand-side customer captivity reinforced by economies of scale. The mechanism is not brand and not IP: it is a 32-year, ~100,000-card backlist that is at once the rules engine, the collection and a stored financial asset, plus a two-sided secondary market (players ↔ local game stores ↔ TCGplayer/eBay) in which a card’s value is a function of the player base. A rival must beat not Magic’s next set but every Magic card ever printed plus the resale value of the customer’s existing shelf. The test has been run repeatedly by parties who were neither under-resourced nor short of IP — Valve’s Artifact, Riot’s Legends of Runeterra, Lucasfilm’s Star Wars: Destiny, Blizzard’s WoW TCG, and now Ravensburger’s decelerating Lorcana — and all of them failed at the same step: bootstrapping secondary-market liquidity from zero.

But the franchise is being harvested, not compounded, and FY2025 is a peak. Magic’s record year was substantially delivered by IP Hasbro does not own — Final Fantasy (Square Enix) was the best-selling set in the franchise’s 32-year history — at a royalty cost management has now disclosed: Wizards’ margin is guided from 46.0% down to “the low 40% range.” Do the arithmetic on the company’s own FY2026 guide and Hasbro expects to sell roughly $109M more Magic and earn ~$54M less doing it; the rent exceeds the increment. Universes Beyond moves from half the slate to four of seven sets in 2026; MSRPs rise; the sample collector booster is removed from Commander decks; card rotation stretches from 18–24 to 32–36 months; and Hasbro’s own head designer concedes player product fatigue. Our normalized durable Wizards EBIT is $750–900M, not $1,006.8M — a 10–25% haircut, not a collapse.

The Consumer Products segment has no moat on any test, and management’s auditors have signed the verdict. Revenue fell −38.8% from $3,981.6M (2021) to $2,437.6M (2025), including −4.2% in 2025 — a year the US toy market grew +6% and the global market +7%. It did not participate in its own industry’s best year in four, even though that recovery was 92% driven by trading cards and licensed building sets, two things Hasbro nominally owns. Mattel, the best-run pure-play toy company alive, did −1% with margins down 200bp in the same year: this is an industry problem, and Hasbro’s version is a below-average instance of an already-mediocre business, facing Amazon (11%) and Walmart (9%) as monopsony buyers. Hasbro wrote CP goodwill down 65%, from $1,582.0M to $561.0M, and KPMG flagged it as a Critical Audit Matter. When a company writes off a billion dollars because it re-forecast its own industry, the industry has rendered its verdict.

Capital allocation is the defining negative and it is not a matter of opinion. eOne cost ~$4.41B of cash in 2019; ~$760M came back; $2.8–3.1B was destroyed — roughly a quarter of today’s market cap. Cumulative impairments and disposal losses over six years total ~$3.42B against $565.5M of book equity — and every dollar traces to capital deployed outside Wizards, whose own goodwill has never been impaired. Six-year cumulative GAAP net income is −$547.9M. Book value per share fell from $31.43 (2020) to $11.87 (2025). ROIC was deleted from the 2024 and 2025 long-term incentive awards, immediately after the largest write-off in company history; the CEO was paid $18.67M in FY2025 — his highest ever, on a maximum 200%-of-target bonus — on a profit metric that excludes the write-down. Insiders bought $1.12M against $74.5M of sales over five years, never bought the low, and sold ~$38.6M at the 99th percentile of the range in the same week the board authorized $1.0B of buyback.

Valuation: fair, not cheap — and the “richest-ever multiple” story is wrong. At $82.09, EV is $14.55B (re-derived at the live price; the vendor snapshot is stale). That is 11.3x TTM EBITDA against a ~12.2x ten-year average, and ~14x forward adjusted earnings. The own-history valuation composite at the 90.2nd percentile of its own history is a mix artifact, not evidence of excess: the P/B percentile is meaningless (book equity is entirely goodwill against negative $1.17B of tangible equity) and P/S is high precisely because the revenue mix shifted to a 46%-margin franchise — a company whose margin structure quintupled should trade at a record price-to-sales. The real question is the Wizards multiple, and it does not support a cheapness claim: on normalized EBIT the implied multiple is ~14–19x, a reasonable range for this kind of asset and roughly fair against Games Workshop once adjusted for what GAW does not make you buy. The market is already underwriting Magic’s FY2025 profit as a peak. It is approximately right.

The variant perception is that both sides are wrong. Bulls capitalize a rented, peak, price-and-cadence-inflated $1.0B as a base. Bears think Magic is a bubble. The evidence says neither: Magic’s speculative tranche has already deflated by roughly two-thirds — collector boosters −60–70% from peak while play boosters held — and Magic grew +36% anyway with backlist at a record. The moat did its job. The debate that matters is not whether Magic survives; it is the size of the ~$850M base and the rising rent on it.

2. Business Overview

Hasbro, Inc. is a 103-year-old play-and-entertainment company headquartered in Pawtucket, Rhode Island — relocating to a 265,000 sq ft sublease at 400 Summer Street, Boston in Q4 2026 — with approximately 4,520 employees worldwide (58% North America, 19% Europe, 18% Asia Pacific, 5% Latin America; ~2% unionized). FY2025 revenue was $4,701.3M, up 13.7%. The company reports in three segments.

Wizards of the Coast & Digital Gaming — $2,186.9M revenue (46.5% of total), +44.7%; operating profit $1,006.8M, 46.0% margin. This is the entire company’s profit engine and it splits into two lines. Tabletop Gaming ($1,686.6M, +62.2%) is overwhelmingly Magic: The Gathering — the collectible card game Wizards has published since 1993, which exceeded $1.7B of revenue in FY2025 (+59%), its best year in 32 years, driven by +$638.2M of growth from Universes Beyond sets (Final Fantasy, Marvel’s Spider-Man, Avatar: The Last Airbender) plus backlist and Secret Lair. It also contains Dungeons & Dragons, the category-defining tabletop RPG, which Hasbro does not disclose separately and which declined in 2025. Digital and Licensed Gaming ($500.3M, +6.1%) is principally royalties from third-party developers licensing Hasbro IP — dominated by MONOPOLY GO! (Scopely), which contributed $168.0M (FY2024: $112.2M). Critically, ex-Monopoly GO! that line shrank from $359.5M to $332.3M (−7.6%): the segment’s “digital licensing growth” is a single mobile-game royalty.

Consumer Products — $2,437.6M revenue (51.8% of total), −4.2%; operating LOSS $(942.6)M (a $79.3M profit, 3.3% margin, excluding the $1,021.9M impairment). This is the traditional toy business: NERF, PLAY-DOH, MY LITTLE PONY, TRANSFORMERS, BEYBLADE, MONOPOLY and the board-game portfolio, plus inbound-licensed Marvel and Star Wars toy lines made under license from Disney. By region: North America $1,421.7M (−4.8%), Europe $566.0M (+8.9%), Asia Pacific $249.4M (−13.0%), Latin America $200.5M (−18.0%). Declines came from NERF, MY LITTLE PONY (−$40.5M, −47.0%) and PLAY-DOH; growth from BEYBLADE, MARVEL, MONOPOLY and TRANSFORMERS. Most product is manufactured by third-party manufacturers in China, Vietnam, India, Japan, Belgium, the US, Mexico and Indonesia; reliance on China is described as “substantial but decreasing,” with increasing use of ODMs that the 10-K concedes “may reduce our direct control over product design, quality, sourcing and manufacturing processes.”

Entertainment — $76.8M revenue (1.6%), −4.4%; operating profit $0.4M. What survives of the eOne acquisition: the Family Brands (PEPPA PIG, PJ MASKS) plus film/TV licensing of Hasbro IP under an “asset-light” model. Hasbro sold eOne’s Film & TV business to Lionsgate in December 2023 and ceased film co-financing in November 2024. Only $325.2M of goodwill remains allocated here, and at the Q2-2025 test its fair value exceeded carrying value by only ~15%.

How it makes money — and the asymmetry that matters. Hasbro monetizes IP three ways, and the economics differ by an order of magnitude. (i) Direct product sales (Magic cards; toys) — Magic is printed in North Carolina, Texas and Japan, carries under $10M of tariff exposure, and requires $12.9M of segment capex to produce $1,006.8M of operating profit; toys are manufactured in Asia, bear the tariff and inventory and markdown risk, and consume $45.2M of capex to earn $112.7M. (ii) Inbound licensing (Marvel/Star Wars toys) — Hasbro takes every risk and pays Disney a royalty off the top; accrued royalties rose to $207.7M from $160.5M. (iii) Outbound licensing (Monopoly GO! via Scopely; Baldur’s Gate 3 via Larian) — Hasbro bears no development or platform risk and collects a royalty. Note the structural point that first-party game publishers (EA, Take-Two, Roblox) make visible: those publishers cite the ~30% app-store platform tax as the single largest external claim on their profit pool. As licensor, Hasbro sits above that toll — Scopely eats the 30% and Hasbro collects gross. That is a genuinely favourable position versus a first-party publisher, and it is underrated in the Hasbro debate.

Recurring vs. non-recurring. Hasbro has no subscription base of consequence and reports $965.4M of unsatisfied performance obligations (fixed consideration/minimum guarantees only: $223.5M expected in 2026, $164.5M in 2027, $128.9M in 2028, $448.5M thereafter). Revenue quality is nonetheless bifurcated in a way the consolidated line hides. Magic’s backlist — cards printed in prior years, still selling, at a quarterly record in Q1 2026 — is the closest thing to recurring revenue in the company and the single best proxy for genuine player demand, because nobody speculates on backlist. Against that, the toy business is hit-driven and non-recurring by construction, and the Monopoly GO! royalty is a decaying single-title annuity. Customer concentration: Amazon 11% and Walmart 9% of consolidated revenue; because Wizards moves through hobby distribution, concentration against Consumer Products specifically is materially higher than 20%.

Verdict. The reporting structure obscures the reality. On any economic reading, Hasbro is a trading-card company that owns a toy company — and it has been managed, financed and compensated as though the reverse were true.

3. Industry Dynamics

Hasbro straddles three industries with radically different structures. They must be assessed separately; the blended entity’s attractiveness is not the average of its parts.

3.1 Trading card games / hobby gaming — a structurally excellent industry at a cyclical high

Structure and profit pool. Third-party TAM estimates disperse violently — GMInsights puts the global TCG market at ~$8.4B (2025) growing 6.9%; Custom Market Insights at ~$13.0B growing 5.2%; others at ~$13.3B growing ~10%. A 55% dispersion for the same year means these are directional only and must never be a valuation input. What is reliable is the margin evidence: Hasbro’s Wizards segment earns 46.0% operating margins and Games Workshop — the closest listed pure-play — earns 42.3% with a 61.7% ROIC and net cash. Two independent operators at 40%+ across cycles is not an aberration; it is the industry’s structural signature.

Barriers to entry are genuine and have been tested to destruction. The graveyard is deep and, critically, well-capitalized: Valve’s Artifact, Riot’s Legends of Runeterra (wound down), Lucasfilm/Fantasy Flight’s Star Wars: Destiny (2016–2020), Blizzard’s WoW TCG (2006–2013), the 1995–2001 Star Wars CCG, KeyForge. Trade commentary observes that most CCGs see initial success then “start to peter out” after ~2 years — a CCG is “a type of business model, not a type of game.” None of these failures can be attributed to weak IP or thin capital. They failed at the same step: bootstrapping the secondary-market liquidity that makes a card an asset, which makes the collection a switching cost, which is the moat.

But every marker of a late-stage capital cycle is present simultaneously. Applying Marathon’s supply-side lens: (1) High returns — 42–46% margins at both major operators. (2) Capital responding — The Pokémon Company’s Millennium Print Group announced a 1.27M sq ft printing plant in December 2025, full-scale late 2028; PSA announced a $200M infrastructure investment with ~370 open roles and ~700 more planned; Ravensburger, Riot, Bandai and Fantasy Flight are all funding competing TCGs; Hasbro itself is accelerating set cadence. (3) Speculative capital in the asset itself — 26.8M cards graded in 2025, +32%, with TCG/non-sports grading (16.8M) overtaking sports (10.0M) for the first time; PSA went from ~2M cards graded in 2020 to ~19M in 2025, a ~10x in five years; April 2026 set an all-time monthly record of 3.10M. (4) The asset-growth anomaly — capacity is being commissioned at the top on trailing demand, and the largest tranche does not reach full scale until late 2028, two to three years after the demand signal that justified it. Marathon’s central lesson is that supply, not demand, drives returns.

And the first crack has already appeared — in the right place. Ravensburger disclosed that Disney Lorcana sales FELL in 2025, “settling at a high level” after “initial hype,” and was explicit about the mechanism: “Those joining the brand with a primary interest in investment have since withdrawn, while the game continues to enjoy growing popularity among its core target group of players and collectors.” Lorcana is the same product cycle as Magic, started ten years later, and its speculative tranche rolled over first while player demand held. That is precisely the decomposition the market is not applying to Magic — and, as §4 shows, it is now replicating inside Magic itself.

Competitive set. Pokémon is the #1 TCG and it is booming: TPCi printed ~10 billion cards in 2025 — more than 10% of the ~85 billion printed since 1996 — and still could not meet demand; Pokémon did $2.5B in US toy-channel sales, +87%, the first property in 20+ years above $2B in a year. Bandai’s One Piece outsold a member of the “Big Three” for a second consecutive quarter on TCGplayer’s Q1-2026 data; its Gundam game debuted strongly. Riot’s Riftbound launched October 2025 with quarterly cadence planned for 2026 — its launch was marred by undersupply, burred card edges and a collation error under-seeding rares, i.e. it is failing at exactly the step the graveyard says is hard.

Verdict — TCG/hobby gaming: a structurally EXCELLENT industry at or near a CYCLICAL HIGH. It passes every Greenwald test — genuine customer captivity, real scale economics in the installed base and secondary-market liquidity, and barriers validated by a decade of failed, extremely well-funded assaults. It is the best industry Hasbro is in by an enormous distance. But the Marathon overlay cuts the other way on timing: capital is visibly flooding in, the speculative tranche has already begun withdrawing at Lorcana, and 2028–29 is the identified supply-shock window. Hasbro’s own FY2026 guide of +3–5% constant-currency revenue after +13.7% is the company conceding the deceleration. Good industry; bad point in the cycle. Magic’s moat will hold its share. It will not hold the profit pool’s cyclical top.

3.2 Traditional toys — a structurally bad industry, and Hasbro is losing share in it

No growth, and 2025’s “recovery” is a costume. US toy industry dollar sales rose +6% in 2025 to $30.3B (ASP +4%, units +3%); global sales rose +7% after three consecutive down years. But the five-year CAGR versus 2020 is only ~3% — below cumulative US CPI, i.e. real decline. And the critical decomposition, per Circana: Games & Puzzles +37% (driven primarily by Pokémon), Explorative & Other Toys +20% (led by trading cards), and Building Sets +15% (licensed properties) together contributed 92% of the entire US toy industry’s 2025 growth. Dolls, Plush, and Outdoor & Sports posted the steepest declines. Strip trading cards and licensed building sets out and traditional toys were flat-to-down in a +6% year. The “toy industry returned to growth” headline is a TCG story wearing a toy-industry costume.

The customer is shrinking, demographically and behaviourally. Under-10s are 67% of the toy sector but their share of global sales is declining; birth rates are at multi-decade lows in many countries; and children’s engagement with traditional toys decreases from ~age 9 as interest shifts to technology and social interaction. The only growth cohort is 15+, up 111% since 2020, now larger than the 10–14 cohort; nearly 40% of European consumers bought toys for themselves or another adult in 2025. And what the 15+ cohort buys is trading cards and collectibles — so even the demographic escape hatch from traditional toys leads back into §3.1, reinforcing that the TCG boom is doing all the work.

Monopsony on one side, cost capture on the other. Amazon (11%) and Walmart (9%) are ~20% of consolidated revenue — and materially more of Consumer Products specifically. These are the two most sophisticated, most private-label-capable and most margin-extractive buyers in retail. Historical concentration has shifted but not eased (FY2019: Walmart 18%, Target 9%, Amazon 8%). On the buy side sits Chinese contract-manufacturing concentration and a live, legally chaotic tariff regime. Hit-driven, non-recurring demand sits in between.

The verdict is proven by the numbers, not argued. Hasbro’s Consumer Products earned a 3.3% operating margin ex-impairment (down from 4.5%) while shrinking 4.2% in the industry’s best year in four. And Mattel — the best-run pure-play toy company alive, with Barbie, Hot Wheels and a live Hollywood flywheel — did −1% revenue with gross margin down 210bp (48.7% vs 50.8%) and operating income down $148M to $546M (~10.3% reported margin), on 8.9% ROIC and 9.4% ROE. When the best operator in a category cannot grow with a +6% market or hold its margin, the problem is the industry.

Verdict — traditional toys: a structurally BAD industry containing a share-losing participant. No real growth; a shrinking, ageing-out customer; monopsony buyers; tariff-exposed Chinese sourcing; and no barriers to entry of any kind — no switching costs, no captivity, and no scale economics a rival cannot rent from the same ODMs. Hasbro’s auditors signed off on writing $1,021.9M of goodwill to zero here on a forward-looking view. That is the industry rendering its own verdict.

3.3 Digital / licensed gaming — good economics, poor durability, and not Hasbro’s industry

The margin structure is superb: near-100% incremental, no capital, no tariffs, no inventory, and — as noted — Hasbro sits above the ~30% platform tax as licensor. The strategic direction is right. But this is not a business with a moat; it is a toll booth on somebody else’s road, and the traffic is down ~52% year-on-year. Third-party mobile tracking puts Monopoly GO! gross revenue at ~$90M/month in April 2026 versus a peak near $117.8M/month in October 2024. Management characterizes the royalty as a stable ~$12–14M/month run-rate and the CEO has called the game “a juggernaut.”

Both statements can be literally true, and the framing still conceals the mechanism. Royalties are typically struck on net revenue after marketing deductions. As Scopely cut user-acquisition spend into a mature/harvest posture, Hasbro’s effective royalty rate mechanically rose — so the royalty line looks stable while underlying bookings halved. That cushion is finite and non-renewable: once UA spend approaches zero the effective rate cannot rise further, and Hasbro’s royalty then decays 1:1 with a game already down by half. This is a direct contradiction between management commentary and external evidence, and it is flagged as such (§12).

The read-across from Games Workshop is the warning for the broader “asset-light licensing” pivot: GW guided its licensing line DOWN from £52.5M to ≥£30M for FY2026. Even the best hobby-IP licensor cannot annuitize the licensing tranche — it is lumpy, hit-dependent and renewal-cliff-exposed.

Verdict — digital/licensed gaming: attractive ECONOMICS, fragile DURABILITY, and NOT an industry Hasbro controls. Mobile gaming has no barriers to entry and no customer captivity — the player owns nothing transferable, the exact inverse of Magic’s collection-as-switching-cost. Treat Monopoly GO! as a decaying asset, not a run-rate. Not a foundation on which to underwrite a re-rating.

3.4 Regulatory and structural factors

Tariffs — asymmetric, and the impairment trigger. Hasbro guided in April 2025 to a $100–300M gross 2025 tariff hit; actual realized FY2025 tariff cost in COGS was only $44.9M. On 2026-02-20 the US Supreme Court ruled against the IEEPA tariffs Hasbro had paid since April 2025; Hasbro states it is “evaluating the accounting impacts including our ability to apply for a refund on tariffs previously paid,” and has filed a refund claim of roughly $50M, in reconciliation, with no timeline and not embedded in FY2026 guidance. The surviving regime: Section 301 List 4A at 7.5% on Chinese-origin toys (unaffected), plus a Section 122 10% tariff effective 2026-02-24 scheduled to sunset ~2026-07-24 — ruled unlawful by the Court of International Trade on 2026-05-07 but stayed by the Federal Circuit on 2026-06-11, so collection continues pending appeal. Chinese-origin toys currently face ~17.5% combined. (The litigation timeline is sourced to tariff-calculator sites and is flagged as requiring primary verification; the SCOTUS ruling and refund optionality are corroborated by the 10-K.) Mitigation is real: Hasbro targeted ~40% of global sourcing out of China by end-2026 and has moved several hundred SKUs. Critically, Wizards’ tariff exposure is under $10M because Magic is printed in North Carolina, Texas and Japan.

The asymmetry reframes the impairment. Hasbro wrote off $1,021.9M of Consumer Products goodwill in Q2 2025 on a tariff forecast — and the Supreme Court then struck down the very tariffs that drove the forecast, with a possible refund of amounts already paid. Goodwill is never written back up under GAAP, so the impairment is permanent regardless. This is a second, independent reason the two segments should be valued as separate businesses.

Loot-box regulation — a genuine, under-priced tail that points at Magic, not at video games. Belgium banned paid loot boxes outright in 2018 (penalties to €800,000 and possible imprisonment of officers). The Netherlands’ KSA declared loot boxes whose contents are tradeable illegal — a Dutch court overturned the EA/FIFA fine on the ground that the packs did not meet the full gambling definition, and legislation to close that loophole has been considered. Brazil enacted a prohibition on loot-box sales to under-18s effective March 2026. The UK permits them with odds disclosure. The legal hinge is the “money’s worth”/tradeability limb — and a physical Magic booster fails it far more clearly than a video-game cosmetic does: it is a randomized paid product whose contents have a deep, liquid, publicly-quoted secondary market and a professional grading industry that exists specifically to certify their monetary value. Physical TCGs have escaped this via forbearance rather than a principled distinction. The very infrastructure that makes Magic’s moat durable is the same fact that satisfies the gambling test. No major jurisdiction has yet applied loot-box law to physical card packs; this is not a near-term earnings risk. It is a low-probability, high-impact tail on the single asset generating ~100% of Hasbro’s operating profit, and it is not in the price.

Child-safety regulation (CPSIA/CPSC, EN71/EU Toy Safety Regulation, phthalate/lead limits, small-parts rules) is a structural compliance cost and a recall tail — but not a barrier to entry of any consequence: every Chinese contract manufacturer serving Hasbro serves its competitors under the same standards. Compliance is rented, not owned.

4. Competitive Position

The right approach on a multi-segment IP business is to apply the Greenwald taxonomy per asset, not per company. On a business where one segment earns more than 100% of consolidated operating profit, a blended moat verdict is not merely imprecise — it is meaningless.

4.1 Magic: The Gathering — demand-side customer captivity + economies of scale. The real thing.

The mechanism, stated so it can be attacked. The barrier is not the brand and not the IP. It is: (a) a 32-year, ~100,000-card backlist that is simultaneously the rules engine, the customer’s collection and a stored financial asset that cannot be liquidated without exiting the game; (b) two-sided secondary-market liquidity (players ↔ local game stores ↔ TCGplayer/eBay), in which a card’s value is a function of the player base; © organized play — especially the social Commander format — which converts the existing collection into the switching cost. A rival must beat not Magic’s next set but every Magic card ever printed, plus the resale value of the customer’s existing shelf.

The financial outcome that would deteriorate if the moat were absent — and has not. A moat claim that cannot be tied to a financial outcome is not a moat. Here it can: backlist revenue set a quarterly record in Q1 FY2026, and the FY2025 10-K attributes Magic’s growth to “Universes Beyond sets… plus backlist and Secret Lair.” Backlist is the cleanest available player-demand proxy precisely because nobody speculates on backlist — it is what people buy in order to play. A 32-year-old catalogue that still sets records is the moat’s signature.

Greenwald’s ROIC bar (15–25%), cleared by 3–4x. Segment assets are no longer disclosed (ASU 2023-07 dropped the asset column). Building invested capital from the disclosed fragments — WOTC goodwill $370.5M, plus the ~$385.6M of capitalized software that sits in this segment, plus modest working capital ≈ $0.9–1.2B — against NOPAT of ~$775M gives an implied segment ROIC of ~65–85%. The denominator is assumption-heavy; the conclusion is not close to the bar. Wizards spends $12.9M of capex and carries $17.6M of D&A to earn $1,006.8M. And note the corroborating fact from the impairment record: WOTC’s goodwill has never been impaired across a six-year, ~$3.42B write-off history in which every charge traces to capital deployed elsewhere.

Greenwald’s market-share-stability test — passed, and the test has been run by the best. Magic holds ~100% of the Magic-format economy and has held it for 32 years against Valve, Riot, Disney/Ravensburger, Lucasfilm and Blizzard. A precision this analysis must not fudge: Magic’s share of the TCG category is falling (Pokémon at $2.5B US toy sales +87%; One Piece outselling a “Big Three” member) while Magic’s absolute revenue rose +59%. These are not in conflict, and category share is the wrong denominator for the moat test. You cannot play a Pokémon card in a Magic deck, and a Magic collection has zero residual value inside Pokémon. Magic’s relevant market is Magic. Category share is the right denominator for the growth question (§5) and the wrong one for the moat question.

Network-effect pressure-test (mandatory). This is a genuine two-sided network effect — but it is local, and it is more precisely economies of scale + captivity than a Metcalfe network. The mechanism: more players → deeper secondary market → tighter bid/ask → both a lower effective cost of entry and a higher resale value on the existing collection → more players. Is it a network effect or just liquidity? Both — liquidity is the output. The discriminating test is whether a marginal player makes the game better for incumbents, and he does, through two channels: a deeper secondary market (financial) and more opponents at the local store (physical). The second channel is Greenwald’s “think local” maxim in its purest form: paper Magic’s playability depends on the shop ten miles away having twenty players, not on the global count. That locality is why the barrier is strong (an entrant must rebuild it store by store) and why it is fragile (it can decay store by store too). The proof by negation is the graveyard: Artifact, Legends of Runeterra, Star Wars: Destiny and WoW TCG all died because they never bootstrapped secondary-market liquidity — so a card had no residual value, so the collection was not an asset, so there was no switching cost, so there was no moat. That is the same mechanism running in reverse, four times, funded by Valve, Riot, Lucasfilm and Blizzard.

What would break it, rank-ordered. (1) LGS/Wizards Play Network attrition — the load-bearing risk. Organized play is where captivity is manufactured. Management claims 11,000+ WPN stores and +20% participation — a hypothesis, not evidence, but a checkable one. (2) Hasbro itself, via price and cadence — see §4.2; the most probable breaker. (3) Regulatory — and note the irony: the secondary market is the network effect and is the gambling exposure. The moat and the tail risk are the same fact. (4) A format reset — only Hasbro can vaporize the backlist’s value; commercially suicidal, but note Hasbro attempted the analogous act at D&D in January 2023 and was beaten back by its own customers, so the instinct exists in this management. (5) Digital substitution — immaterial.

4.2 The central judgment: peak or base? — and the rented-moat tension

Verdict: Magic’s FY2025 EBIT is a CYCLICAL PEAK sitting on a RISING, MOAT-PROTECTED BASE. Both halves are load-bearing. Normalized durable Wizards EBIT is $750–900M — not $1,006.8M, and not the collapse the bear case requires.

The best evidence in this report is a natural experiment, and it is independent of management. The secondary market has already separated the two tranches inside Magic’s own product line, and the player tranche held while the speculative tranche crashed. Collector Booster boxes: Final Fantasy fell from ~$1,500 to ~$1,000; Marvel’s Spider-Man peaked near ~$1,000 and now trades below $300; Avatar peaked near ~$1,000 and now trades below $400. Over the same window Play Boosters held $100–120 against release-day highs near $140. (Secondary-market pricing is trade/community-grade data — MTGStocks, GamesMarket — not institutional; use directionally, and do not read the specific dollar figures as precise.)

This is the Lorcana precedent replicating inside Magic, in real time. Ravensburger said investment-motivated Lorcana buyers “have since withdrawn, while the game continues to enjoy growing popularity among its core target group of players.” Magic’s collector boosters are down 60–70% from peak while its play boosters are down ~15% — and Magic still grew +36% in Q1 FY2026 with backlist at a quarterly record. The speculative tranche is not “at risk of” rolling over; it has already rolled over, by roughly two-thirds, and the franchise grew anyway. The moat did exactly what a moat is supposed to do: it held the base while the froth left.

This refines — and materially strengthens — the “share, not cycle” formulation. The precise statement is that the moat protects share AND the base of the profit pool, but not the cyclical top of it. That is the difference between a value trap and a fair price for a real asset: Magic cannot be competed away; it can only be de-rated from a speculative peak toward its player-demand base. Games Workshop is the proof of the floor — 42.3% margins and 61.7% ROIC sustained across cycles, because the hobby player base is captive.

The tranche that worries us most is not the speculators; it is Hasbro. 2026 carries seven Magic sets, up from six in 2025 (historically ~4–5). Collector Boosters rise from $24.99 to $26.99; Bundles $53.99 to $57.99; Commander decks $44.99 to $49.99 — and the sample collector booster is being removed from Commander products. Card rotation has been stretched from 18–24 to 32–36 months and reprint lead times from ~6 weeks to 3–4 months. Hasbro’s own head designer has publicly acknowledged product-fatigue complaints from “many players.” More sets, higher prices, less content in the box, and a longer rotation window is the signature of a franchise being harvested, not compounded. The rotation extension is the most revealing: lengthening rotation supports the residual value of cards already sold — good, and evidence management understands the mechanism — while simultaneously reducing the forced-repurchase cadence that Standard rotation exists to create. Management is trading future recurring revenue for present sell-through. Greenwald’s warning applies exactly: scale advantages must be defended move-for-move — and here the enemy is not a competitor, it is the incumbent’s own price and cadence policy. The moat’s principal threat is its owner.

The rented-moat verdict: Universes Beyond is moat-HARVESTING dressed as moat-widening. The steelman for widening is real and deserves to be stated properly. From 2025, Universes Beyond sets became legal in all Constructed formats, identical to in-universe sets — a deliberate captivity design that forces a Final Fantasy buyer who wants to play out of a novelty purchase and into the format economy, where the backlist, the LGS and the secondary market take over. And the funnel appears to be converting: organized play +20%, backlist at a record, play boosters holding price. If UB were a pure speculative sugar-rush, backlist would not be at a record.

We nonetheless conclude harvesting, for five reasons. (1) UB rents somebody else’s captivity. Greenwald’s mechanism is that my collection is worthless outside Magic. UB is orthogonal to that — what it monetizes is Final Fantasy fans’ captivity to Square Enix and Marvel fans’ captivity to Disney. Hasbro is monetizing other companies’ moats, not building its own. A funnel you must rent annually is not a barrier to entry; it is an operating expense that happens to work. (2) The rent is disclosed, large, and compounding. 46.0% → “low 40s” on a ~$2.3B base is a 400–500bp give-back ≈ $90–115M/yr transferred to licensors and video-game investment — and the slate is moving the wrong way: Wizards’ lead designer confirmed half of all future premier sets will be Universes Beyond, and 2026 runs four UB of seven sets — a majority (TMNT, Marvel Super Heroes, The Hobbit, Star Trek). As UB goes from half to a majority, the blended royalty burden rises structurally, not cyclically. (3) Hold-up risk at renewal is now materially worse, and Hasbro created it. Square Enix, Disney, Middle-earth Enterprises and Paramount now know exactly what a Magic set is worth — Final Fantasy was the biggest set in franchise history, and that is a negotiating fact pointing one way. Hasbro has manufactured a demonstration effect that raises the price of its own inputs. The exact error it makes in toys — take the risk, pay Disney the royalty — it is now importing into cards. (4) Brand-identity dilution is slow but attacks the moat’s substance: the captivity mechanism is a coherent 32-year backlist, and a majority-rented, Standard-legal slate at seven sets a year makes the “Magic universe” a minority of new product. (5) The counterfactual settles it — the 10-K attributes the growth explicitly to Universes Beyond. The record year is rented.

The honest nuance: UB is not a mistake. It is the correct commercial response to a maturing franchise and it demonstrably works. The problem is that it changes the shape and quality of the earnings — from a 46%-margin, owned-IP, backlist-driven annuity into a low-40s-margin, licensor-dependent, cadence-driven publishing business with renewal cliffs and third-party hold-up risk. That is a structurally lower-multiple earnings stream than the one the bull case is capitalizing. The moat is not shrinking; the economic rent it captures is being shared away. Those are different things, and only the second is happening — which is exactly why it is hard to see.

4.3 The rest of the portfolio

Dungeons & Dragons — intangible/brand advantage, weakly monetized, and structurally weaker than it looks. Narrow moat at best. D&D is the category-defining brand with genuine captivity (campaign continuity, DM ruleset investment, D&D Beyond’s installed base). It fails the durability test for a specific and damning reason: Hasbro tested its own switching costs in January 2023 and discovered they were thin. The leaked OGL 1.1 proposed a 25% royalty on third-party revenue above $750K; the community revolted (mass D&D Beyond cancellations forced the cancellation page offline; major publishers formed a rival open licence); a Wizards survey found 88% would not publish under OGL 1.2 and 89% were dissatisfied; Wizards fully retreated, left OGL 1.0a in place, and placed SRD 5.1 under Creative Commons — irrevocably, in its own words “a decision we can never change.” Hasbro attempted to convert D&D’s captivity into rent, and the customers demonstrated the captivity did not exist at a level that would bear rent. Worse, the retreat was structural: the core mechanics are now permanently free to clone (Paizo’s Pathfinder is the living proof). Contrast with Magic: nobody can print a Magic card, and the backlist is not open-licensed. D&D’s moat is a brand moat; Magic’s is a switching-cost moat — and Greenwald is explicit that brand alone is not a barrier (the Mercedes-Benz case: the world’s most recognized luxury brand earns average returns). D&D also fails on trajectory: it declined in 2025, and WOTC tabletop revenue fell 22% to $207.0M in Q4-2024 on D&D weakness even as Magic set records. (Hasbro does not disclose D&D revenue; a widely-cited ~$460M figure circulates in trade commentary but is not company-sourced and is not used here.)

Monopoly / board games — brand intangible only. No real barrier. Greenwald’s habit-based captivity requires frequent, automatic, low-consideration purchases and explicitly does not apply to infrequent considered ones. A board game is bought once every several years. Monopoly’s real advantage is 90 years of shelf presence and near-universal rule literacy — a distribution and search-cost advantage that Amazon’s shelf can reprice at will. Monopoly did grow in FY2025 and is the healthiest CP brand. Worth something; not a moat.

NERF / PLAY-DOH / MY LITTLE PONY — no advantage. None, on any test. All three declined in FY2025 and were named as the drivers of CP’s revenue fall; MY LITTLE PONY fell $40.5M (−47.0%). No proprietary technology (foam, plastic, salt dough), no captivity (children age out from ~9), no scale economics a rival cannot rent from the same ODMs — and per the 10-K, even the manufacturing capability is increasingly rented. In Greenwald’s phrase, in the long run everything is a toaster; a foam dart blaster reached the toaster stage decades ago.

Inbound-licensed Marvel / Star Wars toys — rented, not owned. This is the same “rented, not owned” pattern seen in licensed-IP video games like NBA 2K. The moat here is real and it belongs to Disney. Hasbro takes the inventory, tariff, markdown and hit-or-miss risk; Disney takes a royalty off the top and bears none of it. This is not a competitive advantage — it is the absence of one, purchased annually. Licence expiry dates are not publicly disclosed (the last announced extension was January 2022), making renewal a live, undated, unquantified risk.

Monopoly GO! royalty — not a moat; a decaying toll booth. Structurally the opposite of Magic: the player owns nothing transferable, which is precisely the inverse of collection-as-switching-cost. That contrast is the cleanest way to explain what Magic’s moat actually is.

4.4 Consumer Products — the verdict, bluntly

NO MOAT. NONE. On every Greenwald test, and it is not close.

Market-share stability — fails catastrophically. Greenwald: >5pp drift over 5–8 years means no barriers; <2pp means formidable ones. CP revenue fell −38.8% since 2021 and −4.2% in 2025 while its market grew +6–7%. That is not share drift; it is a rout — in a year whose growth was driven by trading cards and licensed IP, two things Hasbro nominally owns.

ROIC — fails. A 3.3% ex-impairment operating margin on $2,437.6M of revenue, against segment invested capital of perhaps $1.5–2.0B (post-impairment goodwill $561.0M plus most of the $1,059.8M of receivables and $259.8M of inventory) implies ~4–6% pre-tax, ~3–5% after tax, against a ~9–10% WACC. Consumer Products destroys capital. And it is worse than low-return: it lost $47.5M at the operating line in Q1 2026 with no impairment, against $43.9M in Q1 2025. The “it’s just the impairment” defence is dead.

No identifiable source of advantage; monopsony captures what is left. And the diagnostic tell is management’s own behaviour: CP advertising was cut $30.3M “to reduce variable expenses to offset the operating profit impact of tariffs” while WOTC advertising rose $28.8M. Greenwald is explicit that scale advantages must be defended move-for-move, because every point of share lost narrows the cost gap irreversibly. Cutting demand generation in a business already losing share is not a defence; it is a harvest. Hasbro’s capital allocation already agrees with this verdict even as its strategy deck does not. The only healthy part of CP is out-licensing — i.e. the part that is not making toys; the 10-K concedes it: “Licensing in our Consumer Products segment continues to drive strong operating profit.”

The clinching evidence is management’s own audited signature: CP goodwill written down 65%, from $1,582.0M to $561.0M, blessed by KPMG as a Critical Audit Matter.

4.5 Head-to-head

§4 Verdict: A DURABLE COMPETITIVE ADVANTAGE — concentrated in one asset, and being harvested rather than compounded. Magic is a genuine, rare, Greenwald-grade franchise whose barrier has been validated by the failure of every well-funded assault on it for 32 years, and whose downside is therefore bounded rather than open-ended. But its FY2025 economics are a cyclical peak on that base; its record year was rented from Square Enix and Disney at a rising and structurally increasing toll; and the rest of the portfolio ranges from narrow (D&D, Monopoly) to nonexistent (NERF, Play-Doh, the licensed toy lines). Hasbro owns one of the genuinely great consumer franchises of the last fifty years, and is renting it out to pay for a toy business that cannot earn its cost of capital and a video-game ambition its own CEO expects merely to break even.

5. Growth History and Forward Opportunities

5.1 Six years, zero growth — and both endpoints are M&A artifacts

The headline. Revenue: $4,720.2M (2019) → $5,465.4M (2020) → $6,420.4M (2021 peak) → $5,856.7M (2022) → $5,003.3M (2023) → $4,135.5M (2024) → $4,701.3M (2025). FY2025 revenue is 26.8% below the 2021 peak and essentially flat against 2019 — six years, zero growth. But the flat line disguises a violent mix shift, not stagnation: Wizards went $906.7M (2020) → $2,186.9M (2025), +141%; Consumer Products went $3,649.6M (2020) → $2,437.6M (2025), −33%. Wizards’ +$1,425.7M was offset almost to the dollar by Consumer Products’ −$1,443.6M. This is a growing high-margin games business fused to a shrinking low-margin toy business, and the sum has gone nowhere for six years.

Both endpoints are acquisition accounting, in opposite directions. Reading the restated segment data in the FY2021 10-K: 63% of the 2019→2021 “growth” was eOne; the core toy business contributed 6% — at the best demand moment for toys in twenty years. Then 47% of the 2021→2024 “decline” was selling the same asset. The Entertainment segment ends the six years $1.0M smaller than it started, after ~$4.41B of cash went out the door. Anyone reading Hasbro’s revenue line as an operating record is reading a chart of one acquisition and its unwind.

The single most clarifying fact in this section: ex-Magic, FY2025 revenue FELL ~$72.4M (−2.4%). Magic was 113% of all company growth. Everything else, in aggregate, shrank.

5.2 The quality of the growth

Hasbro discloses no Magic units, no packs, no price, no player count and no player-versus-collector split — on the asset generating ~100% of its operating profit. That absence is itself a finding, and it means the decomposition below is interpretation, honestly labeled.

Four tranches drove Magic’s +$638.2M / +59%:

(a) Player/gameplay demand — durable, moat-protected, and genuinely growing. Organized-play participation grew ~20% in 2025 and is guided to double digits again in 2026; the Wizards Play Network counts 11,000+ stores; MagicCon Las Vegas sold 23,000+ badges; backlist set a quarterly record in Q1 FY2026. (WPN and organized-play figures are management-sourced and unaudited — a hypothesis, not evidence.) The independent corroboration is the collector-versus-play-booster divergence in §4.2: the speculative tranche fell 60–70% and Magic grew +36% anyway. A base that absorbs that and still grows is not small.

(b) Collector/speculative demand — cyclical, moat-irrelevant, and already deflating. The moat does not protect this dollar at all: Greenwald’s captivity binds a player to his collection; it does nothing to bind a speculator to a sealed box. This tranche is a function of the grading complex (26.8M cards graded in 2025, +32%; PSA volumes 10x in five years; a $200M capacity build) — Marathon’s late-cycle infrastructure, arriving at the top. It cannot be sized from outside, but its price action bounds it: roughly two-thirds of its pricing has already gone, and that took Magic’s growth from +59% to a guided mid-single-digit — not to negative.

© Price and set cadence — self-cannibalizing, and the tranche we worry about most. Set cadence has risen roughly 5x from its historical baseline; 2026 carries seven sets; MSRPs rise across Collector Boosters, Bundles and Commander decks while content is removed from the box. Each lever pulls demand forward and taxes format integrity. Note the litigation overlay: the securities class action (S.D.N.Y., filed 2024-11-13, class period 2021-09-16 to 2023-10-26, motion to dismiss filed 2026-02-06) alleges precisely that Magic’s growth was driven by oversupply and set proliferation rather than the consumer-driven “segmentation” strategy management described. The complaint is an allegation and nothing more — no loss is accrued and its legal merit is unresolved. Its analytical content is that a court is being asked to examine the exact question this section is asking, about an earlier cycle.

(d) Licensed-IP-driven new-buyer acquisition — rented. See §4.2. The record year is rented, and the rent is rising.

A conflation risk worth flagging: Harry Potter, Voltron and Street Fighter are toy licenses, not Magic sets. They are routinely confused in trade coverage.

5.3 The forward arithmetic — management is guiding Wizards’ profit DOWN

This is the most important paragraph in the section, and it is arithmetic, not opinion. FY2026 guidance: consolidated revenue +3–5% cc; adjusted operating margin 24–25%; adjusted EBITDA $1.40–1.45B. By segment: Wizards mid-single-digit revenue growth at “low 40%” operating margins (from 46.0%); Consumer Products low-single-digit growth at 6–8% adjusted margin; Entertainment slightly positive at ~50%.

Take Wizards at +5% revenue (~$2,296M) and a 41.5% margin: operating profit ≈ $953M — below FY2025’s $1,006.8M. Decompose it: profit from growth at a constant 46.0% margin = +$50.3M; profit lost to the margin give-back = −$103.3M; net −$53.9M. Sensitivity across the guide: $910M at 40.0% / $953M at 41.5% / $997M at 43.0% — down on every defensible reading of “low 40%.”

Stated plainly: Hasbro is guiding to sell roughly $109M more Magic and earn about $54M less doing it. The royalty and video-game investment consume more than 100% of the incremental profit from growth. That is the rented-moat thesis expressed in the company’s own numbers — and it means FY2026 puts 100%+ of consolidated profit growth in Consumer Products, the segment that has not earned its cost of capital in five years and lost money in the last two first quarters. Management is handing the growth baton from the engine that works to the one that does not.

5.4 The forward opportunity set, with a number and a skeptical read

• Universes Beyond cadence (Marvel Super Heroes, The Hobbit, Star Trek in 2026): the largest and most certain revenue driver — and per §4.2, the one whose rent rises structurally as UB goes from half the slate to four of seven. Revenue opportunity: real. Profit opportunity: negative on the company’s own guide.
• AAA video games — EXODUS (Archetype, H1 2027) and Warlock (Invoke, 2027). Treat with maximum skepticism. ~$413M of capitalized software sits on the balance sheet at Q1 2026 with zero titles released in 2025, 2024 or 2023 and no amortization ever recognized; there is a $24.4M write-off precedent (two cancelled titles, FY2024); and the CEO has disclosed that Hasbro has spent “nearly $1 billion on video games since 2018” across six studios, with no live-service titles — implying roughly $600M already expensed or written off with nothing capitalized ever shipped. FY2026 — the year being underwritten — carries zero of this cost. This is not a timing question; it is a capability question, and the recognition event lands in 2027, in the one segment carrying the thesis.
• D&D monetization — historically weak, declined in 2025, and now permanently capped: the OGL 1.1 retreat placed SRD 5.1 under Creative Commons irrevocably (§4.3). The monetization ceiling is self-inflicted and cannot be undone.
• Digital and direct — modest; ex-Monopoly GO! the digital/licensing line shrank 7.6% in FY2025 and grew only 3% in Q1 2026.
• Out-licensing / “Partner Scale” — strategically the right trade (Hasbro sits above the platform tax) but structurally lumpy: Games Workshop’s licensing line is guided down by more than 40% for FY2026, and even the best hobby-IP licensor cannot annuitize it.
• International — a headwind, not an opportunity, on current evidence: Latin America −18.0%, Asia Pacific −13.0% in FY2025.
• Consumer Products “return to growth” — guided to low-single-digit growth at a 6–8% adjusted margin. Not credible on the record: CP fell 4.2% in the best toy year in four and lost money at the operating line in each of the last two first quarters. Note also that ~$40–60M of Q2 2026 revenue was pushed into H2 by the cyber incident, and management’s assumed 2H recovery is an assumption.

5.5 Scoring management’s forecasting record

Testing the February-2024 promises against what the filings show — deliberately asymmetric, because the record is:

Management called Magic right and everything else wrong. That is a genuine, dated forecasting win on the asset that matters — and it is the strongest argument for giving Cocks the benefit of the doubt on Wizards. It is also the only one.

5.6 The demographic backdrop

Under-10s are 67% of the toy sector, shrinking on birth rates and exiting to screens from ~age 9. The only growing cohort is 15+, up 111% since 2020, and what it buys is trading cards. Hasbro’s growth engine sits on the right side of that line — and 52% of its revenue sits on the wrong side.

§5 Verdict: LOW-QUALITY GROWTH. The case for high quality is real and stated first: Wizards grew +141% since 2020 at 46% margins and 55.5% incremental margins on $12.9M of segment capex, behind a validated moat. Five things defeat it. (1) Six years, zero consolidated growth — a mix shift financed by attrition. (2) Both endpoints are M&A artifacts: 63% of the 2019→2021 growth was eOne, and 47% of the subsequent decline was selling it. (3) The real growth is rented, and the rent now exceeds the increment: ~$109M more Magic revenue for ~$54M less profit. (4) The composition is undisclosed and every observable proxy — cadence up 5x, price up, content down, Lorcana’s speculators already gone, a class action alleging oversupply — points cyclical. (5) Management concedes it: FY2026 puts 100%+ of profit growth in the toy segment while Wizards’ profit falls. Growth without economics is not investable; here the growth is real, the economics are being transferred to licensors, and the consolidated line has not moved in six years.

6. Financial Quality

6.1 Two numbers, and the wedge between them

FY2025 GAAP: revenue $4,701.3M (+13.7%), gross margin 64.6%, operating profit $11.1M (0.2%), pre-tax loss $(102.0)M, tax expense $216.2M, net loss attributable to Hasbro $(322.4)M, diluted EPS $(2.30). Management’s adjusted diluted EPS: $5.54 (adjusted net earnings $784.4M); adjusted operating profit $1,140.0M; adjusted EBITDA $1,361.5M.

The GAAP-to-adjusted wedge is $7.84/share (≈$1,099M), composed of: goodwill impairment $7.17 (91% of the wedge); acquired intangible amortization $0.25; loss on disposal of business $0.18; strategic transformation initiatives $0.13; restructuring/severance $0.05; eOne divestiture costs $0.03.

Adjudicating the adjustments item by item:

• Impairment ($7.17) — legitimate to exclude from run-rate EPS. Non-cash, non-recurring. But it is not costless: it is the delayed recognition of the 2019 eOne purchase price. Excluding it from earnings is right; excluding it from the capital-allocation verdict is not.
• Acquired intangible amortization ($0.25) — defensible. Purchase accounting on a divested business, running off ($58.4M/yr 2026–27, declining).
• Loss on disposal ($0.18) — defensible as non-recurring, though it is the third consecutive year of disposal losses ($539.0M FY23, $37.4M FY24, $25.0M FY25).
• Strategic transformation $0.13 + restructuring $0.05 + eOne divestiture costs $0.03 = $0.21 — NOT legitimate. These recur. Cumulative restructuring under the Operational Excellence Program is $163.5M to date and the 10-K describes the program as “ongoing.” A cost incurred every year for four years is an operating expense, not an adjustment. Mild but real scrubbing.
• SBC is NOT added back — credit to management. The adjusted EPS is not an SBC-washed number, unlike most of the comp set.

Normalized FY2025 EPS ≈ $5.33 (adjusted $5.54 less the $0.21 of recurring costs) — ~15.4x at $82.09. Never use GAAP −$2.30, and do not use management’s $5.54 unlabeled.

The tax “anomaly” is not an anomaly. The (212.1)% effective rate is arithmetic: the $1,021.9M impairment carried a tax benefit of only $5.4M because goodwill is largely non-deductible (the rate reconciliation shows “Goodwill impairment: $209.8M / (205.8)%”). Secondary drivers: Subpart F $9.5M, GILTI $5.7M, valuation allowances $5.4M, §162(m) officer compensation $9.8M, China withholding $8.0M; offset by foreign tax credits $(14.2)M, R&D credits $(6.4)M, UK credits $(13.4)M and a Swiss rate differential $(22.7)M. On adjusted pre-tax (~$1,027M) the implied rate is ~23%, consistent with FY2024’s 20.7%. A modelling trap: anyone normalizing by adding back $1,021.9M must NOT tax-effect it at 21% — the correct add-back is ~$1,016.5M after the $5.4M benefit. Two things are not optical: Hasbro paid $57.6M of current tax, essentially all foreign, while the US entity lost $160.8M pre-tax — profits are earned offshore and the US structure is a cost centre; and recognized deferred tax assets fell from $424.6M to $286.8M, real erosion of US tax attributes.

A data-integrity warning that matters for anyone screening this name. A third-party fundamentals aggregator reports FY2025 operating income of $1,058.0M and a 22.5% operating margin. The filing reports $11.1M and 0.2%. The vendor reclassifies the impairment and the disposal loss out of operating income into “other non-operating” — which is not the filing’s presentation; the 10-K puts the impairment inside the Consumer Products segment’s operating result. The same defect runs through every year (FY2023: filing −$1,538.8M vs feed $191.4M). The filing governs: the filing wins. The aggregator’s operating income, margin, EBITDA, ROIC and ROE for HAS are a pseudo-adjusted series and must never be quoted as reported figures.

6.2 Segment economics — the hypothesis confirmed, and then some

(GAAP, $M. FY2021–22 are as-originally-reported with eOne Film/TV inside Entertainment; Wizards and CP are broadly comparable across the set, Entertainment is not.)

Wizards’ operating margin: 42.5% → 40.6% → 36.1% → 41.8% → 46.0%. Consumer Products: 10.1% → 6.1% → −2.2% → 4.5% → −38.7% (GAAP).

On management’s adjusted basis, FY2025: Wizards $1,006.8M (46.0%) | CP $112.7M (4.6%) | Entertainment $39.5M | Corporate −$19.0M | Total $1,140.0M. So Wizards is 46.5% of revenue and 86.9% of adjusted segment profit (from 36.5%/75.9% in FY2024); Consumer Products is 51.8% of revenue and 9.7% of adjusted segment profit.

The capital intensity comparison is the whole story. Wizards: $12.9M of capex and $17.6M of D&A to earn $1,006.8M — roughly a 78x return on incremental fixed assets. Consumer Products: $45.2M of capex and $92.7M of D&A to earn $112.7M — 3.5x the capex for 11% of the profit. CP shows negative operating leverage in both directions: adjusted profit fell 72% against a 39% revenue decline from 2021 to 2025. Answer to “do economics improve with scale?” — for Wizards, emphatically yes; for Consumer Products, emphatically no; and the consolidated answer depends entirely on mix.

6.3 Quality of earnings — one significant flaw

The capitalized software is the single biggest quality-of-earnings issue in the file. The $135.0M of FY2025 “acquisition of intangible assets” inside capex is software development (FY24 $110.3M; FY23 $73.8M). $385.6M sits in Other assets (FY24 $264.4M), and the 10-K states: “The Company did not release any software titles during 2025, 2024 or 2023 that were previously capitalized… and therefore there was no amortization or impairments recognized.” At Q1 2026 the balance is ~$413M (the interim 10-Q does not break it out; additions were $27.7M in the quarter).

Three years of growing spend — ~46% YoY — has never touched the P&L. Operating profit, adjusted operating profit ($1,140.0M), adjusted EBITDA ($1,361.5M) and adjusted EPS ($5.54) are all flattered by it. Expensing the FY2025 $135.0M as incurred costs ~$0.74/share after tax, taking cash-real normalized EPS from ~$5.33 to ~$4.59 (~17.9x at $82.09). The fair counter-argument is that this is timing, not permanence — it amortizes once titles release. But the titles are slated for 2027 (EXODUS H1-2027; Warlock 2027), so FY2026 carries zero of the cost, the balance will approach ~$500M by release, and there is a demonstrated write-off precedent. The recognition event lands in 2027, either as amortization against Wizards’ margin or as an impairment — in the one segment carrying the entire thesis.

But free cash flow is honestly stated, and it is the number to trust. FY2025 CFO $893.2M − capex $198.3M ($63.3M PP&E + $135.0M software) = FCF $694.9M (FY24 $649.9M). The software spend IS deducted in capex, so FCF — unlike EBITDA and adjusted EPS — is not flattered by the capitalization. FCF is the honest number in this filing.

D&A now understates the economic charge. D&A $171.3M against capex $198.3M = 1.16x; stripping acquired intangible amortization, capex is 1.88x maintenance D&A, and FY2026 capex is guided to ~$250M = 1.46x current D&A. Rising capex will compress FCF conversion even if adjusted EPS grows.

Everything else is clean, and we say so. Receivables are fine: DSO 82.3 days vs 81.2 in FY2024 and 85.3 in FY2021 — the $150.2M AR build came on $565.8M of revenue growth, a +1.1 day change. Cash conversion cycle (hand-computed on filing balances) 61.0 days vs 60.4 — flat, not the 12.6-day improvement the vendor feed implies. Deferred revenue FELL $46.0M to $190.5M while revenue rose 13.7% — a mild drag, i.e. the top line is not being inflated. Inventory $259.8M is disciplined (from $676.8M in 2022). CFO diverges violently from net income (−2.77x) but the driver is the non-cash impairment plus $158.3M of deferred tax, not accrual manipulation; against normalized net income, CFO/NI ≈ 1.19x. Healthy.

SBC was $80.4M in FY2025 vs $50.8M in FY2024 (+58%) — but the history is $97.8M (2021), $83.4M (2022), $72.4M (2023), $50.8M (2024): FY2024 was the anomaly (depressed accruals on missed targets), not FY2025 the explosion. At 1.7% of revenue and 11.6% of FCF it is modest and below the 2021 level in dollars. It is, however, unoffset: zero buybacks 2023–25 and diluted shares drifting 138.0M → 140.2M, ~0.4%/yr.

6.4 ROIC done properly — and the trap

The vendor’s ROIC/ROE are unusable here (computed on a book equity of $538.5M that is collapsing from impairments and supported by −$1,174.9M of tangible common equity). Hand-computed, with NOPAT = normalized operating profit × (1 − 23%) and invested capital = net debt + total book equity:

• FY2023: normalized operating $256.4M → NOPAT $197.4M | avg IC $5,169.1M → ROIC 3.8%
• FY2024: $727.4M → $560.1M | $3,939.6M → ROIC 14.2%
• FY2025: $1,058.0M → $814.7M | $3,462.6M → ROIC 23.5% (26.7% on year-end IC)

Against a ~9–10% WACC: FY2023 destroyed capital (−570bp); FY2024 modestly exceeded it; FY2025 comfortably exceeded it (+1,400bp).

The trap, and the most important number in this section: FY2025’s 23.5% ROIC is flattered by the very impairment that produced the GAAP loss. Writing off $1,021.9M of goodwill shrinks the denominator, mechanically lifting ROIC. Invested capital fell from $6,330.1M (FY22) to $3,054.2M (FY25) — −52% — of which $2,879.5M is cumulative capital written off in 2023–25. Add it back and FY2025 ROIC on un-impaired invested capital is 13.7% ($814.7M / $5,933.7M): the honest return on the capital shareholders actually committed — barely 400bp above WACC after seven years.

Both statements are true and this analysis carries both. (1) Hasbro as it stands today earns ~23.5% on capital — a genuinely good business, because what remains is mostly Wizards. (2) Hasbro as it was managed earned ~13.7% on the capital deployed. Serial impairers always look high-ROIC afterwards. The gap between those two numbers is the capital-allocation story, and reporting either one alone misleads in opposite directions.

6.5 Balance sheet and liquidity — safe on the numbers, concentrated on the analysis

Cash and equivalents $776.6M plus short-term investments $105.4M = $881.6M; total debt $3,281.9M face, all fixed-rate; net debt ~$2,399.9M (or ~$2,488.7M on a cash-only definition — immaterial to the conclusion). Total equity $565.5M incl. NCI; goodwill $1,256.7M + other intangibles $456.7M ⇒ tangible common equity ≈ −$1,174.9M.

Maturities: 2026 $497.0M | 2027 $475.0M | 2028 $109.9M | 2029 $900.0M | 2030 $0 | thereafter $1,300.0M. Revolver: the Fourth A&R agreement (2026-02-20) extended maturity to February 2031 but reduced the commitment to $1.1B from $1.25B (+$550M accordion) at improved pricing; fully undrawn; the company was in compliance with all covenants. Market access is confirmed: on 2026-03-12 Hasbro issued $400.0M of 4.650% notes due 2031 — against 6.05% in May 2024, i.e. the credit market is repricing the post-eOne balance sheet favourably. Notes carry a 0.25%–2.00% coupon step-up on a downgrade, making ratings financially material; the 10-K does not disclose them.

Leverage: net debt / adjusted EBITDA 1.83x; / normalized EBITDA 2.02x; adjusted EBITDA / interest 8.3x. Near-term liquidity risk is negligible.

But is 2.0x safe given the concentration? Arithmetically yes; compositionally, be careful. Net debt / Wizards’ EBITDA alone is 2.43x. Because Consumer Products contributes $112.7M of adjusted operating profit, essentially the entire $2.5B net debt load is serviced by one segment — and within it, disproportionately by one brand. A 30% Magic decline would take consolidated adjusted EBITDA to roughly $1.05B and leverage to ~2.4x on a deteriorating coverage trajectory. This is a thin-tailed 2.0x, not a fat-cushioned one: leverage is low against the level of EBITDA and high against its concentration.

6.6 The six-year shape

Gross margin went 52.9% (2019) → 64.6% (2025), +1,170bp. D&A collapsed 81%, from $908.7M (2021) to $171.3M (2025) — driver: eOne film/TV program cost amortization, which alone fell $448.9M (2023) → $35.8M (2025).

The shape of this business: Hasbro bought eOne for ~$4.6B in 2019 in a debt-funded pursuit of a content flywheel. That purchase inflated 2020–22 revenue to a $6.4B peak and buried ~$900M/yr of program amortization in D&A, making EBITDA look far better than earnings. It then unwound: eOne Film/TV was sold in December 2023 at a $539.0M loss, the D&A pillar vanished, and revenue fell back to the 2019 level. What is left is a structurally different, far better, far smaller company — gross margin up 1,170bp on the mix shift to cards, D&A down 81%, capex-light, and a single segment generating 87% of the profit on 46% of the revenue.

§6 Verdict: THE ECONOMICS IMPROVE WITH SCALE — but only in one segment, and the consolidated answer is a mix question, not a quality question. The accounting is honest where it counts (FCF is fairly stated, receivables and deferred revenue are clean, SBC is not added back) and flawed where it is convenient (~$413M of capitalized software that has never touched the P&L, plus $0.21/share of recurring costs scrubbed from adjusted EPS). Normalized EPS is ~$5.33, or ~$4.59 cash-real if software is expensed. The balance sheet is safe but the debt is serviced by one brand. And the ROIC gap — 23.5% today versus 13.7% on capital actually deployed — is the cleanest single measure of what management has cost shareholders.

7. Capital Allocation

This is the defining negative of the thesis, and none of it is a matter of opinion.

7.1 eOne — quantified value destruction

Hasbro completed the Entertainment One acquisition on 2019-12-30 for an aggregate ~$4.6B (~$3.8B cash for shares at GBP 5.60/sh, plus ~$0.8B to redeem eOne’s notes and revolver); the FY2020 cash flow statement shows $4,412.9M of cash for acquisitions. It was financed with $2.4B of senior unsecured notes, $1.0B of term loans, and 10,592,106 shares issued at $95.00 (~$1,006M). (The headline price is disputed in the record — Alta Fox cited $4.6B, other reads suggest ~$3.8B; the difference is equity versus enterprise value. Hasbro’s five-year 10-K corpus never restates it.)

Recoveries: eOne Music sold to Blackstone for $385M (closed 2021-06-29); eOne Film & TV sold to Lionsgate for $375M cash plus assumption of production financing (headline ~$500M), closed 2023-12-27. Gross proceeds ~$760M. Retained: the Family Brands (Peppa Pig, PJ Masks), carrying $325.2M of goodwill whose fair value exceeded carrying by only ~15% at the Q2-2025 test — a live impairment candidate.

Explicit eOne write-offs: FY2023 goodwill impairment $1,191.2M; eOne Trademark intangible $65.0M; PJ MASKS intangible $51.0M; and disposal losses of $539.0M (FY23) + $37.4M (FY24) + $25.0M (FY25) = $601.4M cumulative. Total explicit eOne-attributable charges: $1,908.6M.

Paid ~$4.41B in cash. Recovered ~$760M. Retains a ~$325M residual that is ~15% from impairment. Value destroyed ≈ $2.8–3.1B — roughly 25–27% of the current ~$11.5B market cap, incinerated in one transaction. Among the worst large-cap M&A outcomes of the era. (The precise figure is a reasoned estimate, not a disclosed number, and turns on the Family Brands residual.)

7.2 Serial impairments — a pattern, not a sequence of accidents

Power Rangers was itself a failed acquisition: Hasbro bought Saban Brands’ Power Rangers for $522M (closed 2018-06-12); $281.0M — ~54% of the purchase price — was written off within 4.5 years. The Discovery Family Channel JV is now carried at $0.

Cumulative GAAP net income FY2020–FY2025: NEGATIVE $547.9M ($225.4M + $435.3M + $203.0M − $1,487.8M + $394.4M − $318.2M).

~$3.42B of write-downs in six years, against $565.5M of book equity — roughly 6x book, and ~30% of today’s market cap. Six-year cumulative GAAP net income is negative solely because of these charges. And the pattern has a single root: every charge traces to capital deployed OUTSIDE Wizards of the Coast — eOne, Power Rangers, Discovery Family, non-core Entertainment, cancelled game titles. WOTC’s goodwill ($370.5M) has never been impaired. A company that writes off this much this often does not have a bad-luck problem; it has a capital-allocation problem.

One accounting judgment worth scrutiny. The Q2-2025 impairment aggregated the four regional Consumer Products reporting units under ASC 280/350 in the same quarter as the charge, justified by “our ongoing transformation.” Aggregation raises the pooled fair value and can reduce the measured shortfall versus testing a weak North America unit alone — and the 10-K concedes the charge was only “a portion of which related to the North America Consumer Products reporting unit.” Management chose both the aggregation basis and the timing. Note also the sequence: a $1.02B quantitative write-down in Q2, then a purely qualitative all-clear in Q4 six months later, with $561.0M of CP goodwill still on the books. Re-impairment risk is live, not settled.

7.3 The dividend — a real-terms cut, and a constraint

Dividends paid: $372.7M (FY20), $374.5M, $385.3M, $388.0M, $389.9M, $392.5M (FY25) — six-year cumulative $2,302.9M. Per share: $2.72 → $2.71 → $2.78 → $2.80 → $2.80 → $2.80. Nominal growth 2020→2025: +2.9% cumulative — not annual. The rate went to $0.70/quarter in 2022 and has been flat for four years. (FY2024’s $2.10 declared is a declaration-timing artifact of the 52/53-week fiscal year, not a cut.)

US CPI rose ~24% over the span. A +2.9% nominal dividend is a real-terms cut of roughly 17–18% — administered silently, while management takes credit for never having cut it.

Is it safe? Yes. Was it free? No. FY2025 cover: $392.5M / $694.9M FCF = 56.5% (1.77x cover; ~2.6x on working-capital-neutral FCF). Against GAAP earnings the payout is undefined — FY2025 EPS was −$2.30 and FY2023 was −$10.71. Over six years dividends consumed 64% of cumulative FCF of $3,595.5M. In the stress year it was cash-funded: FY2022 FCF $198.7M against $385.3M of dividends plus $125.0M of buyback — a $311.6M shortfall, and cash fell $493.4M.

The dividend is a sacred cow. $2.3B was distributed over six years in which the company cumulatively LOST $547.9M on a GAAP basis, took $2.68B of impairments and carried $3.3B of debt. It is covered by FCF, so it is not reckless — but it consumed the optionality that could have deleveraged faster or bought stock at the 2023–24 lows. Holding it flat in nominal terms was the compromise: management refused either to cut it (admitting the balance sheet was strained) or to grow it (admitting it could afford to).

7.4 Buybacks — dilution without offset, and textbook procyclical timing

The last repurchase was $125.0M in FY2022, at an average $87.46. FY2023, FY2024 and FY2025: zero. Over those three years SBC totalled $203.6M against $0 of buyback, and shares outstanding drifted 138.18M → 140.39M (+1.6%). In February 2026 the board authorized $1.0B, replacing all prior authorizations, with no expiration and no obligation.

Shareholders funded management’s compensation through dilution for three straight years with no offset. And the timing is textbook procyclical: the company bought $125.0M in 2022, bought NOTHING through the 2023–24 trough when the stock traded in the $40s, and then authorized $1.0B in February 2026 after the stock had recovered to the $80s — within days of the five-year high, and in the same week insiders sold ~$38.6M at the 99th percentile of the five-year range. Management did not buy when it was cheap and is authorized to buy now that it is not. This is the opposite of the discipline the 10-K claims: “The Company has a long history of increasing shareholder value through its share repurchase program.”

7.5 Incentives — paid at maximum in an impairment year

CEO Chris Cocks’s total compensation: FY2025 $18,672,521 (salary $1.5M; stock awards $10.5M; non-equity incentive $6,000,000; other $672,422); FY2024 $16,841,413; FY2023 $15,110,869. Three-year total: $50,624,803. CEO pay ratio 165:1 (median employee $113,241).

The 2025 annual incentive paid at 200% of target — the plan maximum — and the Committee “determined to not make individual performance modifications for the NEOs.” The metrics, quoted from the payout table:

• Company Net Revenue (50%) — goal $4,111,000K; actual $4,607,000K; achievement 112%; payout 100%
• Company Operating Profit Dollars (50%) — goal $892,000K; actual $1,114,000K; achievement 125%; payout 100%
• Final payout: 200%

The impairment is invisible to the bonus. That “$1,114M” of Operating Profit Dollars is an adjusted figure that excludes the $1,021.9M goodwill impairment; GAAP operating profit was ~$36M before net interest, producing the reported $102.0M pre-tax loss. Cocks’s target bonus was raised from 175% to 200% of salary “to more closely align with market”; his $6.0M cash bonus is $1.5M × 200% target × 200% payout.

In a fiscal year in which Hasbro wrote off $1.02B of goodwill, reported a $102.0M pre-tax loss, a $318.2M net loss and −$2.30 of diluted EPS, the CEO was paid $18.67M — his highest ever — including a maximum 200%-of-target cash bonus, on a profit metric constructed so that destroying capital cannot reduce the bonus. That is not a subtle misalignment; it is the defining fact of Hasbro’s compensation design.

And it got worse by design. ROIC was DELETED from the long-term incentive plan. Per the FY2025 10-K Note 16: “The 2023 Stock Performance Awards are measured based on achieving targets set for diluted earnings per share and return on invested capital (‘ROIC’)… The 2024 and 2025 Stock Performance Awards are measured based on achieving targets set for diluted earnings per share, in addition to a TSR modifier.” Hasbro carried a ROIC metric in 2023 and removed it for 2024 and 2025 — removing the only capital-efficiency measure in the entire incentive architecture, immediately after the largest capital-destruction event in company history. The Committee also removed the strategic cost-savings metric and “increased the emphasis on net revenue and operating profit dollars.” Every remaining metric — net revenue, adjusted operating profit dollars, cumulative diluted EPS — is a scale or accrual measure that a value-destroying acquisition can flatter and that an impairment cannot touch. The relative-TSR modifier on PSUs is the only genuine link to per-share outcomes, and it is a ±25% modifier on the outcome, not a gate.

Pay versus performance: revenue fell ~16% from FY2020 to FY2025; the stock compounded at ~0.5%/yr over five years and ~3.3%/yr over ten — roughly flat nominally and materially negative in real terms. Over the same window the CEO was paid $50.6M in three years.

7.6 Insider behaviour — conviction is absent

Across 281 Form 4s, 5 Form 4/As and 16 Form 3s (393 transactions, 35 filers, 2021-03-23 → 2026-06-30):

• Open-market purchases (code P): 3 filings, 12,602 shares, $1,124,302. That is the entire five-year record.
• Sales (code S): 32 filings, 749,431 shares, $74,541,661. A 66:1 dollar ratio.
• Every purchase, all three: Cocks 1,302 sh @ $88.76 (2022-04-21); Cocks 8,800 sh @ $89.71 (2022-04-21); Michael Burns 2,500 sh @ $87.70 (2022-04-25). Only 2 of 35 insiders ever bought a share, and all three purchases fall in a five-day window in April 2022 — during the Alta Fox proxy fight, six weeks before the vote, and ten weeks after Alta Fox publicly stated that “none of the current directors have purchased even a single share of Hasbro stock over the last ten years.” The Form 4 record supports that claim for the testable window: zero code-P purchases from 2021-03-23 to 2022-02-09.
• Discretion, not plans: exactly one of 32 sales cites a Rule 10b5-1 plan. ~99.7% of sale dollars were discretionary — insiders chose their moments.
• Nobody bought the low. 2023 — the year containing the $42.88 closing low and the eOne sale — contains zero code-P and zero code-S transactions.
• Selling clustered at the top. Following the 2026-02-11 five-year high, Cocks exercised 377,512 options and sold 377,992 shares at $100.32–$106.48 for ~$38.6M, with Goetter, Sibley, Kilpin, Bunge and Barbacovi selling alongside. (Fair nuance: Cocks’s total holdings ROSE from 249,861 to 303,310 — this was option monetization, not an exit.)

These insiders behave like optionees monetizing compensation, not owners accumulating an asset.

7.7 The 2022 Alta Fox proxy contest — the activist lost the vote and won the argument

Alta Fox Capital (Connor Haley), holding ~2.5% and never crossing 5%, launched a campaign on 2022-02-17 with a 100-page deck at FreeTheWizards.com, nominating five directors and demanding a tax-free spin-off of Wizards of the Coast. Its claims: Wizards was a hidden gem (~$1.29B revenue, +42%, ~47% EBITDA margin in 2021) whose share of Hasbro EBITDA had gone from ~20% (2016) to ~50% (2021); investors were paying only ~11.8x for it; a standalone Wizards would be worth “>$100/share.” It called eOne “the defining moment and greatest failure of the Brand Blueprint era” and noted Hasbro had not disclosed Wizards’ revenue for the 20 years after buying it in 1999.

Hasbro’s board expanded from 11 to 13 mid-contest, concluded a spin “would not create value for shareholders” — but never published that analysis. Both proxy advisers backed the company; ISS nonetheless acknowledged a “limited case for change” and flagged long-term issues in capital allocation, disclosure and performance. Alta Fox cut its slate 5 → 3 → 1. At the 2022-06-08 meeting all 13 company nominees were elected; Alta Fox’s nominee drew ~11.6% of shares represented. But its sole final target, 19-year incumbent Edward Philip, drew 28,850,094 withheld (27.6%) — by far the weakest result on the board.

The aftermath is the verdict. Within three months Philip and Bronfin (the two longest-tenured directors) announced retirement. Within 18 months Hasbro executed essentially the entire activist agenda minus the structural remedy: a strategic review (Oct 2022) → the Operational Excellence Program → ~1,900 headcount cuts → the eOne sales process (Nov 2022) → the Lionsgate sale (Dec 2023). The board that beat Alta Fox has since been about half replaced and rebuilt around gaming operators (Frank Gibeau of Zynga, Owen Mahoney of Nexon, and in January 2026 Douglas Bowser of Nintendo of America and Carla Vernon).

Alta Fox lost the vote decisively and has been substantially vindicated on the analysis. On eOne, the company conceded within 14 months — it did not defend eOne, it liquidated it, at a $539.0M disposal loss plus $1,191.2M of goodwill impairment on the very reporting units eOne created. On Wizards, Hasbro’s own segment footnote now says more bluntly than the deck did what the deck alleged: essentially all of the profit sits outside Consumer Products. The board was arguably right on the narrow mechanical question — a 2022 spin might not have created value, and Wizards’ IP is entangled with Hasbro’s licensing engine — but wrong on the diagnosis it was defending. (Alta Fox’s current position is untraceable; it never filed a 13D/13G. Do not assume it still holds.)

7.8 What management has done right

Three things, and they deserve to be stated plainly.

Debt management is genuinely good. Leverage has gone from roughly $5B post-eOne to $3.28B; the 2024 maturity was retired; 2026/2027 notes were opportunistically repurchased at small gains ($119.9M principal in FY2025, $1.7M gain); the book was termed out via the 2034 notes and the March-2026 4.65% 2031 print; it is 100% fixed-rate at ~2.0x net leverage with $882M of liquid assets and no commercial paper outstanding. Caveat: deleveraging was substantially funded by selling the assets it should never have bought. This is cleaning up a self-inflicted wound competently, not creating value.

The Operational Excellence Program is real and largely delivered — “almost $800 million of gross cost savings” through 2025 against a $1.0B commitment, roughly half of all SKUs eliminated (representing only ~2% of revenue), zero-based budgeting introduced. (The goalpost moved from $350–400M → $750M → $1.0B, and the framing shifted from net to gross — score it ~half delivered.)

The “Playing to Win” asset-light pivot is directionally correct and backed by structural exit, not rhetoric. Hasbro sold the studio, ceased film co-financing in November 2024, and zeroed the cable JV. Licensing partner-funded entertainment — letting a studio carry the production risk while Hasbro collects a royalty on IP it already owns — is genuinely higher-ROIC than owning eOne. Two caveats. First, the pivot is a consequence, not a philosophy: Hasbro did not choose asset-light in 2019 when it had the option; it arrived there after spending $4.4B to learn. Second, “partner scale” spending still shows up — and it is the next eOne in miniature.

7.9 The next mistake, already sized and dated

Hasbro has spent “nearly $1 billion on video games since 2018” across six studios and an internal publishing organization, with zero live-service titles and — per the FY2025 10-K — zero previously-capitalized titles released in 2025, 2024 or 2023. Roughly $600M has already been expensed or written off, and ~$413M sits capitalized awaiting a 2027 recognition event.

The CEO’s own rationale is the tell, quoted verbatim: “If you invest a fair amount of money and give a fair amount of time to a talented team to do a more traditional game, you probably won’t make billions, but your chances of at least making your money back is much higher.”

That is a CEO describing a project he expects to roughly break even on — a ~0% ROIC use of capital, funded out of a franchise earning 65–85%. Under Greenwald’s expansion test, capital must be deployed inside the franchise, where the barriers apply. AAA video-game development is squarely outside it: no barriers to entry, no captivity, no switching costs, and no demonstrated Hasbro capability across eight years and six studios. “If you don’t bring anything to the dance, don’t expect to take anything home.”

And the strategic contradiction is the damning part: Hasbro’s two biggest digital successes — Baldur’s Gate 3 (licensed to Larian) and Monopoly GO! (licensed to Scopely) — were both achieved by LICENSING the IP out and bearing no development risk. Hasbro’s conclusion from its own best data was to build six studios and spend ~$1B doing it first-party instead — reversing the trade and stepping DOWN into the ~30% platform toll it currently sits above. That is the wrong lesson, drawn from its own evidence, and it directly contradicts the “Partner Scale” pillar of the strategy its own proxy describes. Note finally that a console RPG does not extend the tabletop moat: the moat is a physical backlist plus secondary-market liquidity plus local organized play, and a single-player game touches none of the three.

(A smaller item, judged fairly: the Boston HQ relocation — trading ~366,000 sq ft of owned Pawtucket real estate for a 265,000 sq ft sublease, with the Pawtucket properties to be listed for sale — is structurally defensible: a smaller footprint, no ownership, a monetization, and a real talent market for the digital pivot. But the build-out cost and expected proceeds are undisclosed, and it is discretionary cash and management attention spent during a turnaround.)

§7 Verdict: NEGATIVE, and it is the most important verdict in this report. Management destroyed ~$2.8–3.1B in one acquisition, wrote off ~$3.42B in six years against $565.5M of book equity, produced negative cumulative GAAP earnings across the span, froze the dividend into an ~18% real cut, bought no stock through a generational low and authorized $1.0B at a five-year high, deleted ROIC from the incentive plan immediately after the largest write-off in company history, and paid the CEO his highest-ever package at a maximum bonus in the year of a $1.02B impairment — while insiders sold 66:1 into the top. The debt work and the cost program are real and creditable, and the asset-light pivot is the right lesson learned at catastrophic tuition. But capital allocation is the bridge between business value and shareholder value, and on this evidence the bridge is out. Hasbro’s shareholders own a wonderful franchise and have received ~1.30%/yr over five years for it. That gap is not the market’s error; it is management’s record.

8. Changes and Headwinds — Last Two Years

The cybersecurity incident (March 2026) — live, unquantified, and the next catalyst. Unauthorized network access was identified 2026-03-28; systems were proactively taken offline; the Q1 10-Q was filed late under Rule 12b-25 as a direct result. Guided impacts: ~$20M of one-time remediation opex (excluded from adjusted EBITDA); $40–60M of Consumer Products revenue delayed from Q2 into H2 2026 (mostly Q3); and a receivables/cash-flow shift from Q2 into Q3. Management states the situation is “contained” and systems were on track to be fully restored by ~June 2026 — but the Q1 release disclosed that Q2 legal and remediation costs had begun with “the full scope of the costs and related impacts has not been determined,” and Hasbro is “continuing to work with cybersecurity and forensic experts to identify and review any files potentially impacted,” with “any notifications deemed appropriate” still pending. The assumed 2H recovery of the Q2 shortfall is management’s assumption, not a fact. Analytically, the incident did something useful: it cleanly separated the two Hasbros. Magic shipped on cadence (Secrets of Strixhaven shipped in April); the physical-goods business took the entire hit.

The $1,021.9M impairment (Q2 2025) and its tariff reversal. Covered in §3.4/§7.2. The asymmetry bears repeating: Hasbro wrote off a billion dollars of toy goodwill on a forecast built around the April-2025 reciprocal tariffs, and on 2026-02-20 the Supreme Court struck those tariffs down, with a ~$50M refund claim now in reconciliation (no timeline; not in FY2026 guidance). Goodwill is never written back up. FY2025 GAAP results are therefore distorted by a $1.02B non-cash charge predicated on a regime that no longer exists — FY2025 GAAP EPS is meaningless as a run-rate.

The securities class action. West Palm Beach Firefighters’ Pension Fund, S.D.N.Y., filed 2024-11-13, amended 2025-11-26, on behalf of purchasers between 2021-09-16 and 2023-10-26. Alleges §10(b)/20(a) violations — “false or misleading statements regarding the growth and success of Magic: The Gathering card sets, including statements attributing Magic’s growth to a consumer-driven ‘segmentation’ strategy.” Defendants moved to dismiss 2026-02-06; Hasbro is “unable to estimate a reasonably possible range of loss.” A derivative action (Sbriglio v. Stoddart, R.I. Super. Ct.) is stayed pending the MTD. The class period brackets the proxy fight and the allegation strikes directly at the crown jewel’s narrative — the same disclosure quality Alta Fox attacked in 2022.

Governance and leadership. The board went from the defensive 13 of 2022 to 11 nominees today, with Philip, Leinbach and Bronfin gone and five of 11 joining in 2024 or later, rebuilt around gaming operators — Gibeau (Zynga), Mahoney (Nexon), Harris (Jack in the Box) in 2024; Carla Vernon and Douglas Bowser (President/COO, Nintendo of America) in January 2026. That is a materially more credible board for a company whose profit is a games franchise. Executive churn has been real but is often overstated: Eric Nyman lasted 13 months as COO; Cynthia Williams resigned the Wizards presidency in April 2024 — but the seat was filled within three months by John Hight (2024-07-19), a 12-year Blizzard veteran who ran the Warcraft franchise. The “crown jewel leadership is vacant” framing that circulates on this name is stale and wrong.

Financing. The Fourth A&R revolver (2026-02-20) extended to February 2031 at improved pricing but cut the commitment to $1.1B; $400M of 4.650% notes due 2031 were issued 2026-03-12 to refinance the November 2026 maturities. The February 2026 $1.0B buyback authorization superseded all prior authorizations.

Cost and macro headwinds into FY2026. ~$40M of year-on-year below-the-line EPS headwind from higher interest expense and lower non-operating income (including the absence of prior-year Swiss deferred-tax benefits); ~$30M H2 headwind from oil-related input costs (freight, resin, packaging) at ~$100/bbl, largely contained to Consumer Products; ~$15M favourable tariff variance versus the start of the year.

The consumer. Cocks: “we continue to see kind of a tale of two cities. The top 20% of households in terms of wealth are really driving a lot of demand and are staying pretty resilient. The lower quintiles… their pennies are pinched.” That is consistent with the category data: premium collectibles boom while mass-market toys shrink.

§8 Verdict: NET NEUTRAL-TO-NEGATIVE, with the composition mattering more than the balance. The genuinely thesis-strengthening changes are structural exits (eOne gone, film co-financing ceased, the cable JV zeroed), a repaired and cheaply-refinanced balance sheet, and a board rebuilt around people who have actually run games businesses. The thesis-weakening changes are the ones with open ends: an unquantified cyber remediation with an assumed H2 recovery, $561.0M of surviving CP goodwill supported only by a qualitative test, litigation aimed squarely at the Magic narrative, and a Wizards margin guided down. The two-year record is a competent repair job on a self-inflicted wound, bracketed by two shocks (Goldner’s death; the cyberattack) — and it has not changed the central fact that one franchise carries everything.

9. Risk Analysis

The risk that is not on this list, because it is not a risk — it is the thesis: Magic being competed away. No evidence in this file supports it. Valve, Riot, Disney, Lucasfilm and Blizzard all tried. The downside here is normalization, not destruction, and that distinction is what separates a fair price from a value trap.

10. Valuation Discussion — Embedded Expectations

No recommendation and no price target appear in this section.

10.1 The starting point, correctly struck

At $82.09 (2026-07-16), on 143.2M diluted shares and the Q1 FY2026 balance sheet (debt $3,623.3M, cash $857.1M, minority $25.9M ⇒ net debt $2,766.2M): market cap $11.76B; EV = $14.55B. (A vendor snapshot prints $15.47B because it is struck at the 2026-03-31 quarter close, ~$93.6/share — a 12% stale price. Use $14.55B.)

10.2 The “richest-ever multiple” story is wrong for this name

The own-history valuation index reads 90.2nd percentile composite (n=2), with P/B 18.06x at the 90.8th and P/S 2.42x at the 89.5th; P/E is null because GAAP TTM EPS is negative. Taken at face value that says “richest ever.” It is a mix artifact, and this analysis says so plainly.

• The P/B percentile is meaningless. Book equity of $538.5M is entirely goodwill and intangibles against negative ~$1.17B of tangible common equity, hollowed by buybacks and serial impairments; price/tangible book prints −10.0x. Per the standing house caveat, ignore P/B on negative-book-equity issuers.
• The P/S percentile is high precisely BECAUSE the business got better. Consolidated operating margin went from 3.8% (FY2023) to ~22% normalized (FY2025) as the revenue mix shifted to a 46%-margin franchise. A company whose margin structure has quintupled SHOULD trade at a record price-to-sales. Comparing today’s Wizards-led Hasbro to the old toy-led Hasbro on P/S is the cross-sectional error the framework warns against, committed across time.
• On the metrics that adjust for mix, HAS is not at an extreme. Its own ten-year year-end EV/EBITDA history: 10.4x (2016), 11.4x, 19.8x, 14.1x, 12.1x, 9.7x, 9.9x, 11.8x, 11.2x, 11.3x (2025) — average ~12.2x. Today’s 11.3x TTM / ~10.2x forward sits at or slightly BELOW that average. Forward P/E is ~14x.

10.3 Embedded expectations — what is the market actually paying for Magic?

The only honest way to value this company is to value the stub and back into Wizards. Against EV of $14.55B:

(Stub multiples are our own assumptions, anchored on Mattel’s ~11.8x EV/TTM EBIT and discounted for CP’s inferior margin and returns. The implied-Wizards conclusion is robust across the range, so it does not hinge on them.)

Read naively, this says the market capitalizes a 46%-margin, 32-year IP franchise that grew revenue 44.7% and operating profit 59.3% at only 12.5–14.1x EBIT — cheap against Games Workshop’s ~19.3x. That reading is wrong, and correcting it is the central analytical act of this valuation.

The 12.5–14.1x is struck on PEAK EBIT. Per §4.2 and §5.3, normalized durable Wizards EBIT is $750–900M, not $1,006.8M. Re-strike the same implied Wizards EV on the normalized denominator and the multiple becomes ~14–19x: at the midpoint (~$850M), ~14.7–16.7x — a reasonable multiple for a mature franchise-IP publisher (~15–18x P/E, ~14–16x EBITDA, roughly where a standalone publisher like EA has historically traded) — and the discount to Games Workshop narrows from ~30–35% to roughly 15–24%.

And the residual discount is rational, not an anomaly. Three reasons, each of which a Games Workshop buyer does not pay: (1) You cannot buy Wizards. You must buy Hasbro — and take on a structurally declining, currently loss-making toy segment, $2.77B of net debt, negative ~$1.17B of tangible equity and a corporate cost centre. (2) Games Workshop OWNS Warhammer outright and pays no royalties; Magic’s record year was substantially delivered by licensed third-party IP at a rent that is visibly compressing margin from 46.0% to the low 40s and rising structurally as UB goes to a majority of the slate. (3) GAW’s capital allocation is exemplary (61.7% ROIC, net cash, ~87% payout); Hasbro’s is among the worst in large-cap consumer. A conglomerate discount applied to a franchise trapped inside a poor capital allocator is a rational discount.

The Mattel cross-check confirms the stub is not conservative. Mattel trades at ~8.4x EV/TTM EBITDA on a 10.2% operating margin and 8.9% ROIC. Hasbro’s Consumer Products — 3.3% ex-impairment margin, guided 6–8%, loss-making in Q1 — is a materially worse business than Mattel’s whole company. Hasbro’s ~35% EV/EBITDA premium to Mattel is entirely attributable to Wizards, and correctly so.

The factor tape cross-checks the comp set and disagrees with the bull case. HAS’s factor-nearest neighbours by cosine similarity are COWZ (0.914, Pacer US Cash Cows), Aptiv (0.906), Lear (0.902), then RHS, ZGN, RDIV, Visteon, SPGP/PXLG, DSTL, Host, JBHT, SCHD, Knight-Swift, SDY, VFLO. Auto suppliers, truckers and FCF/dividend-value ETFs. Not Mattel, not any games or IP compounder, and not a single high-multiple IP name. The tape’s implied comp set is a cyclical dividend cash-cow, which sits directly against a Magic-annuity multiple. In factor space HAS carries a DividendYield loading of +0.38 and a Momentum loading of −0.31, with Value (+0.03) and Quality (+0.05) approximately zero — the market does not price this as either a value stock or a quality compounder.

10.4 Scenarios

At $82.09 the stock sits in the upper half of that range. The asymmetry is roughly −24% to +23% against a base of ~$77–85 — i.e. approximately symmetric, with the current price already inside the base zone. This is a fairly-valued security, not a mispriced one.

§10 Verdict: FAIRLY VALUED — and the market is not making the error the bull case assumes. The single most important embedded expectation is explicit and correct: the market is already underwriting Magic’s FY2025 profit as a peak rather than a base. That is why the naive implied multiple looks low. Normalize the denominator and the apparent cheapness disappears — the implied ~14–19x sits at a reasonable multiple for exactly this kind of asset, and the residual discount to Games Workshop is fully explained by the toy stub, the rented IP and the capital-allocation record. What is NOT in the price, in either direction: the ~$50M tariff refund (upside), the 2027 capitalized-software recognition event (downside), the Star Wars/Marvel licence renewal (downside), and the loot-box tail (downside). The entire debate reduces to the size of the ~$850M base and the trajectory of the rent on it — not to whether Magic survives.

11. Variant Perception

Consensus. The tape prices Hasbro as a cyclical dividend cash-cow — factor-nearest neighbours are COWZ, SCHD, SDY, Aptiv, Lear and JBHT; the only meaningful positive style loading is DividendYield (+0.38); Value and Quality are ~zero; Momentum is negative. Sell-side framing treats it as a Magic-led re-rating story that stalled on the cyberattack. The stock has compounded at +0.46%/yr over five years and +3.28%/yr over ten at ~33% volatility with a −63.8% lifetime drawdown — a decade of dead money at equity-like risk — and yet in February 2026 it printed a five-year high and now sits at the ~90th percentile of its own P/S history. The re-rating, not the earnings record, is what has been paid for.

The strongest bull case. Hasbro is a 46%-margin, 65–85%-ROIC IP monopoly hiding inside a toy company’s multiple. Magic grew +59% in FY2025 and +36% in Q1 2026; organized play is up 20%; backlist set a quarterly record; the moat has repelled Valve, Riot, Disney, Lucasfilm and Blizzard for 32 years. The market pays only 12.5–14.1x for it against Games Workshop’s 19.3x. The toy business is being fixed (~$800M of savings) or will eventually be separated — Alta Fox’s 2022 sum-of-the-parts case is stronger today than when it was made, and the board has been half-replaced with gaming operators who might act on it. Meanwhile you are paid 3.4% covered 1.77x by FCF, leverage is 2.0x, the balance sheet refinances at 4.65%, and a $1.0B buyback is authorized. A great business at a mediocre-business price.

The strongest bear case. Hasbro is a company that has not grown revenue in six years, has written off $3.42B against $565.5M of book equity, produced negative cumulative GAAP earnings over six years, and pays its CEO a maximum bonus in impairment years on metrics from which it deleted ROIC. Its one good asset just printed a speculative peak on rented IP at a rising toll, and management is guiding that asset’s profit down while handing the growth baton to a toy business that has lost money in each of the last two first quarters. Grading volumes are 10x in five years, Pokémon is commissioning a 1.27M sq ft plant that lands in late 2028, and the first speculative tranche has already withdrawn at Lorcana. Insiders bought $1.12M against $74.5M of sales and dumped $38.6M at the high while authorizing $1.0B of buyback. A melting toy business subsidized by a card bubble, run by people who have destroyed $3B and been paid handsomely for it.

The variant perception: BOTH SIDES ARE WRONG, and the evidence that settles it is a natural experiment nobody is looking at.

The bulls capitalize a rented, peak, price-and-cadence-inflated $1.0B as a base. The bears think Magic is a bubble. Magic’s own product line has already run the experiment: collector boosters fell 60–70% from peak — Final Fantasy ~$1,500 → ~$1,000; Spider-Man ~$1,000 → below $300; Avatar ~$1,000 → below $400 — while play boosters held $100–120. The speculative tranche did not just soften; it left. And Magic grew +36% anyway, with backlist at a record. (Trade-press-sourced pricing; directional, not precise.)

That single fact refutes both cases. It refutes the bear because a franchise that absorbs a two-thirds collapse in its speculative tranche and still grows is not a bubble — the moat did exactly what a moat is supposed to do, and it bounds the downside. It refutes the bull because the froth leaving took Magic’s growth from +59% to a guided mid-single-digit, which means the $1,006.8M was never the base.

So the debate everyone is having — “will Magic collapse?” — is the wrong debate. The right one is about the size of the ~$850M base and the RISING RENT on it: 46.0% → low 40s, with Universes Beyond going from half the slate to four of seven sets, at a toll set by Square Enix, Disney, Middle-earth Enterprises and Paramount — all of whom now know exactly what a Magic set is worth, because Hasbro just showed them. On that framing, consensus — which prices HAS as a cash cow rather than an IP compounder — is closer to right than the bull case is. The five most important assumptions: (1) the size of Magic’s durable player-demand base ($750–900M of EBIT, undisclosed and unprovable from outside); (2) the trajectory of the UB royalty burden as the slate goes majority-rented; (3) whether Consumer Products can reach even 6% margins, having never done so recently; (4) whether the ~$1B video-game program becomes the second eOne; (5) whether this board and management can be trusted with $1.0B of buyback authorization and ~$695M/yr of FCF.

12. Fact vs. Interpretation

13. Open Questions

1. The decisive one: what is Magic’s player-demand versus collector/speculative revenue split? Hasbro discloses no units, no packs, no price, no player count and no split — on the asset generating ~100% of its operating profit. The $750–900M is a reasoned triangulation, not a disclosed number, and reasonable analyses land in different halves of that range.
2. Sell-in versus sell-through. Hasbro reports shipments to distributors/retail, not consumer sell-through. If TCGplayer marketplace volume is softening while reported revenue accelerates, channel inventory is building. This is the single most important falsifiable item on the Magic thesis and we could not obtain the distributor-channel data to settle it.
3. Is the Wizards Play Network genuinely growing? The 11,000+ stores and +20% participation are management numbers. They are load-bearing to the moat verdict and independently unverified; no hard LGS count exists publicly.
4. The ~$413M of capitalized software. What exactly is in it beyond EXODUS and Warlock, and does it survive a quantitative recoverability test? The Q1 10-Q dropped the breakout. The Q2 FY2026 10-Q (~August 2026) and the FY2026 10-K are the checkpoints.
5. Does the surviving $561.0M of Consumer Products goodwill (and $325.2M at Family Brands, only ~15% above carrying) survive a second leg? The Q4-2025 annual test was qualitative only.
6. How large is the IEEPA tariff refund and when does it land? ~$50M is claimed and in reconciliation, with no timeline, and it is not in guidance.
7. When do the Star Wars and Marvel master-toy licences expire, and on what terms do they renew? Not publicly disclosed.
8. Current credit ratings and outlook — not in the 10-K, and financially material given the 0.25–2.00% coupon step-up.
9. What is the total cash cost of the Boston build-out and the expected Pawtucket sale proceeds? Undisclosed.
10. Will the $1.0B authorization actually be executed, and at what prices? This is the key forward test of whether capital-allocation behaviour has changed.
11. Has any regulator formally examined PHYSICAL TCG boosters under loot-box/gambling law? No definitive answer could be sourced.
12. The eOne headline purchase price ($4.6B per Alta Fox versus ~$3.8B on another read — likely equity versus enterprise value). Hasbro’s five-year corpus never restates it.

14. What Must Be True

For the bull case

1. Magic’s ~$1.0B of FY2025 EBIT is close to a durable base, not a peak — i.e. the player-demand tranche is large enough that normalization is shallow. Falsification test: Wizards’ operating profit declining year-on-year in FY2026 as guided, and again in FY2027 — confirming the peak. Note this test is already tracking against the bull: management itself guides Wizards profit flat-to-down on growing revenue.
2. The Universes Beyond funnel converts rented attention into captive players faster than the rent rises. Falsification test: Wizards’ margin guided below 40% for FY2027, or UB exceeding four of seven sets in the 2027 slate — the rent outrunning the funnel.
3. Consumer Products reaches its guided 6–8% adjusted margin and stops shrinking. Falsification test: a third consecutive first-quarter operating loss (Q1 2027), or FY2026 CP revenue declining again against a growing toy market.
4. The board deploys the $1.0B authorization and ~$695M/yr of FCF intelligently. Falsification test: another acquisition outside Wizards, or the buyback executed above ~$90 while the ~$413M software balance grows without a released title.

For the bear case

1. FY2025 was a speculative peak that mean-reverts hard toward $600–700M of Wizards EBIT. Falsification test: *play-booster secondary pricing breaking materially below ~$100/box, or Magic revenue declining year-on-year in a quarter with a normal UB release. This test has ALREADY partially failed for the bears: the collector tranche fell 60–70% and Magic still grew +36% in Q1 2026 with backlist at a record.
2. Magic’s moat is being breached by Pokémon, One Piece, Riftbound and Lorcana. Falsification test: already failed. Category share is the wrong denominator — no competitor has taken a Magic player’s collection, and Riot’s Riftbound is failing at the same liquidity-bootstrapping step that killed Artifact, Runeterra, Destiny and the WoW TCG.
3. The TCG capital cycle turns hard and the profit pool mean-reverts. Falsification test: grading volumes (26.8M in 2025, +32%) declining year-on-year for two consecutive quarters would confirm the speculative cycle rolling over; continued growth through 2027 would delay it. Watch the late-2028 Pokémon plant.

The falsification test that matters most, for either side

The moat’s integrity shows up at the local game store before it shows up in revenue. Test: a reported decline in Wizards Play Network store count or organized-play participation for two consecutive periods while price and cadence continue to rise. That is the moat unwinding from the periphery inward — and because organized play is where captivity is manufactured, it would be visible there first. Conversely, the 2027 slate reverting to five or fewer sets with Universes Beyond back below half, and Wizards’ margin guided back toward 46%, would falsify the harvest thesis and indicate the franchise is being compounded rather than rented.

15. Source Appendix

See Appendix B — Source Appendix below for the full source list with URLs and access dates. Primary sources relied upon: Hasbro FY2025 Form 10-K (filed 2026-02-25, FYE 2025-12-28), FY2023 and FY2022 Forms 10-K, the Q1 FY2026 Form 10-Q (filed 2026-05-13), the DEF 14A (filed 2026-04-17), the 2022 contested-proxy record (DEFC14A / DFAN14A / PRRN14A / DEFA14A), the 56-filing 8-K corpus, and 281 Forms 4 — all read in full from public SEC filings. Management commentary is drawn from the Q4 FY2025 (2026-02-10) and Q1 FY2026 (2026-05-20) earnings calls and treated throughout as hypothesis, not evidence.

APPENDIX A — Standard Diligence Questionnaire

Report date: 2026-07-17 | Price reference: $82.09 (close, 2026-07-16) | Supplemental to the analysis above; contains no recommendation and no price target.

General

What thoughtful questions have other investors asked about this company?

Three serious external interrogations exist on this name, and each asks a question this analysis also has to answer.

1. Alta Fox Capital’s “Free the Wizards” campaign (February–June 2022) — the richest material by far. Connor Haley’s Alta Fox held ~2.5% (never crossing 5%, never filing a 13D) and launched a 100-page public deck on 2022-02-17, nominating five directors and demanding a tax-free spin-off of Wizards of the Coast. Its core questions were structural, not tactical:

• Why does a ~47%-EBITDA-margin games business sit inside a toy conglomerate at all? Alta Fox argued Wizards had gone from ~20% of Hasbro EBITDA (2016) to ~50% (2021) on ~$1.29B of revenue growing +42%, that investors were paying only ~11.8x for it, and that a standalone Wizards was worth “>$100/share.”
• Why did Hasbro not disclose Wizards’ revenue for the twenty years after acquiring it in 1999? A disclosure question, and the same one the securities class action now presses from a different angle.
• Why has no director bought a single share in ten years? FACT: the Form 4 record supports this for the testable window — zero code-P purchases from 2021-03-23 to 2022-02-09. The only three open-market purchases in the entire five-year corpus (Cocks ×2, Burns ×1, ~$1.12M) land in a five-day window in April 2022, ten weeks after the deck and six weeks before the vote.
• Was eOne “the defining moment and greatest failure of the Brand Blueprint era”? Hasbro did not defend eOne on the merits; it liquidated it fourteen months later at a $539.0M disposal loss on top of $1,191.2M of goodwill impairment.

The outcome is the useful part. Alta Fox lost the vote decisively and won the argument. At the 2022-06-08 meeting all 13 company nominees were elected and Alta Fox’s nominee drew ~11.6% of shares represented — but its sole final target, 19-year incumbent Edward Philip, drew 28,850,094 withheld (27.6%), by far the worst result on the board, and both Philip and Bronfin announced retirement within three months. ISS backed the company while conceding a “limited case for change” and flagging capital allocation, disclosure and performance. Within 18 months Hasbro executed essentially the entire activist agenda except the structural remedy: strategic review → Operational Excellence Program → ~1,900 headcount cuts → the Lionsgate sale (Dec 2023). INTERPRETATION: the board was arguably right on the narrow mechanical question (a 2022 spin may not have created value; Wizards’ IP is entangled with the licensing engine) and wrong on the diagnosis it was defending. The 2026 segment footnote now states more bluntly than the deck did what the deck alleged. (Alta Fox’s current position is untraceable — no 13D/13G was ever filed. Do not assume it still holds.)

2. The securities class action — West Palm Beach Firefighters’ Pension Fund (S.D.N.Y., filed 2024-11-13, amended 2025-11-26; class period 2021-09-16 to 2023-10-26; MTD filed 2026-02-06). It alleges §10(b)/20(a) violations for “false or misleading statements regarding the growth and success of Magic: The Gathering card sets, including statements attributing Magic’s growth to a consumer-driven ‘segmentation’ strategy.” The analytical content is the investor question that matters: was Magic’s growth driven by consumer demand, or by oversupply and set proliferation into the channel? FACT: it is an allegation, no loss is accrued, Hasbro is “unable to estimate a reasonably possible range of loss,” and its merit is unresolved; a derivative action (Sbriglio v. Stoddart) is stayed pending the MTD. INTERPRETATION: a framing device, not evidence — its value is that a court is being asked to examine, for an earlier cycle, the exact sell-in-versus-sell-through question we cannot settle from outside for this one.

3. The recurring sell-side/buy-side questions, restated honestly. Is Magic a bubble? (No — the speculative tranche already deflated ~two-thirds and Magic grew +36% anyway.) Will the toy business ever earn its cost of capital? (No evidence it can.) Will the board spin Wizards now that it is half-rebuilt around gaming operators? (Open; nothing in the filings suggests it is on the table.) Is Monopoly GO! an annuity? (No.) The question almost nobody asks, and should: what is the recognition event on the ~$413M of capitalized software that has never touched the P&L?

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low?

A cyclical high, and management’s own guidance concedes it — but the high sits on a rising, moat-protected base rather than on air. FACT: FY2025 revenue $4,701.3M (+13.7%); Wizards revenue $2,186.9M (+44.7%) and operating profit $1,006.8M at a 46.0% margin — the segment’s best margin in the five-year record (42.5% → 40.6% → 36.1% → 41.8% → 46.0%). Magic exceeded $1.7B (+59%), its best year in 32 years. FACT: FY2026 guidance is +3–5% constant-currency revenue after +13.7%, with Wizards’ margin guided from 46.0% down to “the low 40% range.” INTERPRETATION: our normalized durable Wizards EBIT is $750–900M, not $1,006.8M — a 10–25% haircut, not a collapse. GAAP FY2025 (operating profit $11.1M, pre-tax loss $(102.0)M, diluted EPS $(2.30)) is not a cyclical signal at all; it is a $1,021.9M non-cash impairment. Normalized FY2025 EPS is ~$5.33, or ~$4.59 cash-real if capitalized software is expensed.

Consumer Products is at neither a high nor a low — it is in secular decline: revenue −38.8% from $3,981.6M (2021) to $2,437.6M (2025), including −4.2% in 2025, a year the US toy market grew +6% and the global market +7%. It lost $47.5M at the operating line in Q1 FY2026 with no impairment in either period (Q1’25: $43.9M loss). A business that shrinks in its industry’s best year in four is not cyclically trough-ed; it is losing share.

Driven by the external environment or by internal actions?

Both, and the split is unusually clean. External: the trading-card boom (26.8M cards graded in 2025, +32%; TCG/non-sports grading overtook sports for the first time; PSA volumes ~10x in five years), a wealth-skewed consumer (CEO: “the top 20% of households… are really driving a lot of demand… the lower quintiles, their pennies are pinched”), the tariff regime, and the toy industry’s own +6%/+7% recovery — 92% of which, per Circana, came from Games & Puzzles (+37%, mostly Pokémon), Explorative & Other (+20%, led by trading cards) and Building Sets (+15%, licensed). Internal: the Universes Beyond strategy (+$638.2M of Magic growth, Final Fantasy the best-selling set in 32 years), set cadence raised to seven sets in 2026 from a historical ~4–5, MSRP increases (Collector Boosters $24.99 → $26.99; Bundles $53.99 → $57.99; Commander decks $44.99 → $49.99, with the sample collector booster removed), rotation stretched from 18–24 to 32–36 months, and ~$800M of gross cost savings.

INTERPRETATION: the internal actions are the larger driver and the more worrying one. More sets, higher prices, less content in the box and a longer rotation window is the signature of a franchise being harvested. The moat’s principal threat is its owner.

How stable are revenues?

Bifurcated, and the consolidated line hides it. Hasbro has no subscription base of consequence and reports $965.4M of unsatisfied performance obligations (fixed consideration / minimum guarantees only: $223.5M expected 2026, $164.5M 2027, $128.9M 2028, $448.5M thereafter) — thin cover against a $4.7B revenue base. The genuinely stable tranche is Magic’s backlist (cards printed in prior years, still selling, at a quarterly record in Q1 FY2026) — the closest thing to recurring revenue in the company and the best proxy for real player demand, precisely because nobody speculates on backlist. Against that: the toy business is hit-driven and non-recurring by construction, and the Monopoly GO! royalty is a decaying single-title annuity.

Six-year stability at the consolidated level is an illusion of averaging. Revenue: $4,720.2M (2019) → $6,420.4M (2021 peak) → $4,701.3M (2025). Flat across six years — but Wizards went $906.7M (2020) → $2,186.9M (+141%) while Consumer Products went $3,649.6M → $2,437.6M (−33%). Wizards’ +$1,425.7M was offset almost to the dollar by CP’s −$1,443.6M. Both endpoints are also M&A artifacts: 63% of the 2019→2021 “growth” was eOne (the core toy business contributed 6%, at the best toy demand moment in twenty years) and 47% of the 2021→2024 “decline” was selling the same asset.

Outlook for products/services?

FACT (FY2026 guidance): Wizards mid-single-digit revenue growth at “low 40%” margins; Consumer Products low-single-digit growth at a 6–8% adjusted margin; Entertainment slightly positive at ~50%; consolidated adjusted EBITDA $1.40–1.45B and a 24–25% adjusted operating margin. The forward arithmetic is the tell: Wizards at +5% revenue (~$2,296M) and a 41.5% margin earns ~$953M — below FY2025’s $1,006.8M. Sensitivity across the guide: $910M at 40.0% / $953M at 41.5% / $997M at 43.0% — down on every defensible reading of “low 40%.” Hasbro is guiding to sell roughly $109M more Magic and earn about $54M less doing it. That puts 100%+ of FY2026 consolidated profit growth in Consumer Products — the segment that has not earned its cost of capital in five years.

Pipeline specifics: 2026 carries four Universes Beyond sets of seven (TMNT, Marvel Super Heroes, The Hobbit, Star Trek), with the lead designer confirming half of all future premier sets will be UB; D&D declined in 2025 and is permanently capped by the OGL retreat; ex-Monopoly GO!, the digital/licensed line shrank 7.6% in FY2025 and grew only 3% in Q1 2026; the AAA titles EXODUS (H1 2027) and Warlock (2027) carry zero cost in FY2026 and land their recognition event in 2027.

How big will this market be — growing, shrinking, domestic or international?

Three markets, three answers.

• TCG/hobby gaming — growing, but sized unreliably and at a cyclical high. Third-party TAM estimates disperse violently: ~$8.4B (GMInsights, +6.9%) vs ~$13.0B (Custom Market Insights, +5.2%) vs ~$13.3B (+~10%) for the same year. A 55% dispersion means these are directional only and must never be a valuation input. What is reliable is the margin evidence: two independent operators at 40%+ across cycles (Hasbro’s Wizards 46.0%; Games Workshop 42.3% with 61.7% ROIC and net cash).
• Traditional toys — real decline dressed as recovery. US dollar sales +6% to $30.3B in 2025 (ASP +4%, units +3%); global +7% after three down years. But the five-year CAGR versus 2020 is only ~3% — below cumulative US CPI, i.e. real decline — and 92% of 2025’s growth was trading cards and licensed building sets. Strip those out and traditional toys were flat-to-down in a +6% year. Management’s own datapoint concedes it: Cocks says its targeted categories “grew about 22%, while the balance of the toy industry declined 3%.”
• Digital/licensed gaming — attractive economics, decaying traffic. Monopoly GO! gross revenue ~$90M/month (April 2026) versus a peak near $117.8M/month (October 2024): ~−52% YoY.

Geography. North America is ~58% of the workforce and the dominant revenue base (CP North America $1,421.7M, −4.8%). International is currently a headwind, not an opportunity: Europe +8.9%, but Asia Pacific −13.0% and Latin America −18.0% in FY2025.

The demographic backdrop is the structural point. Under-10s are 67% of the toy sector and their share is falling on birth rates; children exit traditional toys from ~age 9. The only growth cohort is 15+, up 111% since 2020, now larger than the 10–14 cohort — and what it buys is trading cards. Hasbro’s growth engine sits on the right side of that line and 52% of its revenue sits on the wrong side.

Business Quality & Competitive Moat

Is the industry getting more or less competitive?

More — in both of Hasbro’s industries, for opposite reasons.

In TCGs, applying Marathon’s supply-side lens, every marker of a late-stage capital cycle is present at once. High returns: 42–46% margins at both major operators. Capital responding: The Pokémon Company’s Millennium Print Group announced a 1.27M sq ft printing plant in December 2025, at full scale late 2028; PSA announced a $200M infrastructure investment; Ravensburger, Riot, Bandai and Fantasy Flight are all funding competing TCGs; Hasbro itself is accelerating cadence. Speculative capital in the asset itself: grading volumes 10x in five years, an all-time monthly record of 3.10M in April 2026. The asset-growth anomaly: capacity commissioned at the top on trailing demand, the largest tranche not at full scale until late 2028 — two to three years after the demand signal that justified it. The CEO describes the setup himself: “trading cards is probably the hottest category in all toys and games… Supply is always a challenge, especially with a bunch of new entrants.” Marathon’s central lesson is that supply, not demand, drives returns.

And the first crack has already appeared in the right place. Ravensburger disclosed that Disney Lorcana sales FELL in 2025: “Those joining the brand with a primary interest in investment have since withdrawn, while the game continues to enjoy growing popularity among its core target group of players and collectors.” Same product cycle as Magic, started ten years later; the speculative tranche rolled over first while player demand held.

In toys, competition is not intensifying so much as the profit pool is being captured — by monopsony buyers on one side (Amazon 11%, Walmart 9% of consolidated revenue, materially more of CP) and by Chinese contract-manufacturing concentration plus a live tariff regime on the other.

How profitable is the business (ROIC, ROE)?

ROIC — hand-computed, because the vendor feed is unusable and ROE is meaningless here.

The trap, and the single most important number in this appendix: FY2025’s 23.5% ROIC is flattered by the very impairment that produced the GAAP loss — writing off $1,021.9M of goodwill shrinks the denominator. Invested capital fell from $6,330.1M (FY22) to $3,054.2M (FY25), −52%, of which $2,879.5M is capital written off. Both statements are true and must be carried together: (1) Hasbro as it stands today earns ~23.5% on capital, because what remains is mostly Wizards; (2) Hasbro as it was managed earned ~13.7% on the capital shareholders actually committed. Serial impairers always look high-ROIC afterwards, and the gap between those two numbers is the capital-allocation story.

ROE and P/B are UNUSABLE — say so plainly rather than compute them. Book equity of $565.5M (incl. NCI) is entirely goodwill ($1,256.7M) and other intangibles ($456.7M); tangible common equity is negative ~$(1,174.9)M; price/tangible book prints −10.0x. Book value per share fell from $31.43 (2020) to $11.87 (2025). The own-history P/B percentile (90.8th) is therefore a meaningless artifact — the correct sector analog for a negative-tangible-equity, capex-light IP business is ROIC and FCF yield, not ROE or P/B.

The capital-intensity comparison carries the argument: Wizards spends $12.9M of capex and carries $17.6M of D&A to earn $1,006.8M — roughly a 78x return on incremental fixed assets, with a 55.5% FY24→FY25 incremental margin. Consumer Products spends $45.2M of capex and carries $92.7M of D&A to earn $112.7M — 3.5x the capex for 11% of the profit, with negative operating leverage in both directions (adjusted profit −72% against a −39% revenue decline, 2021→2025).

How profitable is the industry — how many competitors, what barriers to entry?

TCG: few competitors, genuine barriers, tested to destruction. The relevant set is Pokémon (#1; ~10 billion cards printed in 2025 — more than 10% of the ~85 billion printed since 1996 — and $2.5B in US toy-channel sales, +87%, the first property in 20+ years above $2B), Bandai (One Piece outsold a “Big Three” member for a second consecutive quarter; Gundam debuted strongly), Ravensburger’s Lorcana (declining), and Riot’s Riftbound (launched 2025-10-31; undersupply, burred card edges and a collation error under-seeding rares). The graveyard is deep and, critically, well-capitalized: Valve’s Artifact, Riot’s Legends of Runeterra (wound down), Lucasfilm/Fantasy Flight’s Star Wars: Destiny, Blizzard’s WoW TCG, the 1995–2001 Star Wars CCG, KeyForge. None of these failures can be blamed on weak IP or thin capital. They failed at the same step: bootstrapping the secondary-market liquidity that makes a card an asset, which makes the collection a switching cost, which is the moat.

Toys: no barriers of any kind. No switching costs, no captivity, no scale economics a rival cannot rent from the same ODMs. Child-safety regulation (CPSIA/CPSC, EN71/EU Toy Safety Regulation) is a compliance cost and a recall tail, not a barrier — every Chinese contract manufacturer serving Hasbro serves its competitors under the same standards. Compliance is rented, not owned. Hasbro’s auditors signed off on writing $1,021.9M of goodwill to zero here; KPMG flagged it as a Critical Audit Matter. That is the industry rendering its own verdict.

Can the business be easily understood?

Yes — once you stop reading the segment labels as a description of the company. On any economic reading, Hasbro is a trading-card company that owns a toy company, and it has been managed, financed and compensated as though the reverse were true. Wizards is 46.5% of revenue and 86.9% of adjusted segment profit; on a GAAP basis it is more than 100% of it. The complications are all accounting, not business: a $1.02B impairment, a −212.1% tax rate, ~$413M of never-amortized capitalized software, and a vendor data feed that reports a 22.5% operating margin where the filing reports 0.2%. The business is simple; the financial statements are not.

Can it be undermined by foreign low-cost labor?

The toy segment already has been; the card segment is structurally immune — and this is one of the cleanest lines in the file.

FACT: Most Consumer Products output is made by third-party manufacturers in China, Vietnam, India, Japan, Belgium, the US, Mexico and Indonesia; reliance on China is “substantial but decreasing,” with increasing use of ODMs that the 10-K concedes “may reduce our direct control over product design, quality, sourcing and manufacturing processes.” Low-cost manufacturing is not a threat to Hasbro’s toy business — it is the toy business’s operating model, and it is available to every competitor on identical terms. That is precisely why there is no moat there: the manufacturing capability is rented.

Magic is printed in North Carolina, Texas and Japan, and carries under $10M of tariff exposure. FACT: FY2025 realized tariff cost in COGS was only $44.9M against a guided $100–300M gross — essentially all of it in Consumer Products. The value in a Magic card is not the cardstock; it is the 32-year rules engine, the secondary market and the format economy behind it. No amount of cheap printing capacity reproduces those.

Do brands matter?

Yes — and this analysis’s most important qualification is that brand is not the moat. Greenwald is explicit that brand alone is not a barrier (the Mercedes-Benz case: the world’s most recognized luxury brand earns average returns). The Hasbro portfolio is a natural experiment on the point:

D&D is the decisive case, because Hasbro ran the experiment itself and lost. In January 2023 the leaked OGL 1.1 proposed a 25% royalty on third-party revenue above $750K. The community revolted; a Wizards survey found 88% would not publish under OGL 1.2 and 89% were dissatisfied; Wizards fully retreated, left OGL 1.0a in place, and placed SRD 5.1 under Creative Commons — irrevocably, in its own words “a decision we can never change.” Hasbro attempted to convert D&D’s brand captivity into rent, and the customers demonstrated the captivity would not bear it — permanently, since the core mechanics are now free to clone forever (Pathfinder is the living proof). Contrast with Magic: nobody can print a Magic card, and the backlist is not open-licensed.

What is the nature of competition?

In TCGs, competition is for the customer’s format, not for his attention — and that distinction decides the moat verdict. A precision this analysis insists on: Magic’s share of the TCG category is FALLING (Pokémon at $2.5B US, +87%; One Piece taking shelf) while Magic’s absolute revenue rose +59%. These are not in conflict. You cannot play a Pokémon card in a Magic deck, and a Magic collection has zero residual value inside Pokémon. Magic’s relevant market is Magic. Category share is the right denominator for the growth question and the wrong one for the moat question.

In toys, competition is a commodity share fight — Hasbro loses to Mattel, LEGO and Pokémon simultaneously — decided by shelf, price and licence, and refereed by two monopsonists. The diagnostic tell is Hasbro’s own behaviour: CP advertising was cut $30.3M “to reduce variable expenses to offset the operating profit impact of tariffs” while Wizards advertising rose $28.8M. Greenwald is explicit that scale advantages must be defended move-for-move. Cutting demand generation in a business already losing share is not a defence; it is a harvest — Hasbro’s capital allocation already agrees with the no-moat verdict even as its strategy deck does not.

Customers’ switching costs?

Magic: among the highest in consumer, and mechanically specific. The barrier is (a) a 32-year, ~100,000-card backlist that is simultaneously the rules engine, the customer’s collection and a stored financial asset he cannot liquidate without exiting the game; (b) two-sided secondary-market liquidity (players ↔ local game stores ↔ TCGplayer/eBay) in which a card’s value is a function of the player base; © organized play — especially the social Commander format — which converts the existing collection into the switching cost. A rival must beat not Magic’s next set but every Magic card ever printed, plus the resale value of the customer’s existing shelf.

The network effect is genuine but local — Greenwald’s “think local” maxim in its purest form. Paper Magic’s playability depends on the shop ten miles away having twenty players, not on the global count. That locality is why the barrier is strong (an entrant must rebuild it store by store) and why it is fragile (it can decay store by store too). The proof by negation is the graveyard: four well-funded rivals died because they never bootstrapped liquidity — so a card had no residual value, so the collection was not an asset, so there was no switching cost, so there was no moat.

The moat claim ties to a financial outcome: backlist revenue set a quarterly record in Q1 FY2026 — the cleanest player-demand proxy precisely because nobody speculates on backlist. And the live test: collector boosters fell 60–70% from peak while play boosters held $100–120 — and Magic still grew +36% in Q1 2026. The froth left and the franchise grew anyway. (Secondary-market pricing is trade/community-grade data — directional only.)

Everywhere else: zero. Toys have none (children age out from ~9). Monopoly GO! is the structural inverse of Magic — the player owns nothing transferable, which is the cleanest way to explain what Magic’s moat actually is.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet?

Yes — and this is the single most important unrecognized asset in the file: the Magic and D&D IP itself. FACT: Wizards of the Coast’s goodwill is carried at $370.5M and has never been impaired across a six-year, ~$3.42B write-off history in which every charge traces to capital deployed elsewhere. INTERPRETATION: Magic’s ~100,000-card backlist and the Magic/D&D intellectual property are on the balance sheet at nothing like economic value — a franchise earning $1,006.8M of operating profit on $12.9M of segment capex is, in Greenwald terms, an earnings-power asset whose reproduction cost bears no relation to its carrying value. This is why the negative $(1,174.9)M tangible common equity is not a solvency signal and why P/B must be discarded rather than interpreted.

Three further items:

• ~$413M of capitalized software development (Q1 2026; $385.6M at FY2025; $264.4M at FY2024) sits in Other assets — recognized, but as an asset of highly uncertain value. See the accounting section below; this is as likely a liability as an asset.
• Pawtucket real estate. Hasbro owns ~343,000 sq ft in Pawtucket, RI plus an adjacent ~23,000 sq ft building (~366,000 sq ft total), carried at depreciated historical cost, and will list the properties for sale after moving to a 265,000 sq ft Boston sublease in Q4 2026. Expected proceeds are undisclosed. ASSUMPTION: modest relative to an $11.8B market cap; not thesis-relevant.
• The ~$50M IEEPA tariff refund claim — filed, in reconciliation, no timeline, and explicitly not in FY2026 guidance. Contingent-gain accounting means it will not be recognized until realized. Real, unrecognized, and small.

Off-balance-sheet liabilities?

This question does not map well to Hasbro, and padding it would be dishonest. The correct answer is: there is no material off-balance-sheet exposure of the classic kind, and the analogs that matter are contractual and contingent rather than structural. No securitization, no unconsolidated leverage vehicles, no pension time-bomb disclosed, no production-financing overhang (film co-financing ceased in November 2024 and the Discovery Family Channel JV is carried at $0). Leases are on-balance-sheet under ASC 842, and the Boston sublease is a long-term operating lease, not a hidden obligation.

The correct sector analogs, ranked by materiality:

1. Inbound licence renewal risk (Disney’s Marvel and Star Wars master-toy licences). FACT: expiry dates and terms are not publicly disclosed; the last announced extension was January 2022. Accrued royalties rose to $207.7M from $160.5M. This is a live, undated, unquantified claim on a meaningful slice of CP revenue and is not in the price.
2. Universes Beyond royalty escalation / licensor hold-up. The rent is now disclosed in the guidance itself: 46.0% → “low 40s” on a ~$2.3B base ≈ $90–115M/yr transferred to licensors and video-game investment. And Hasbro has manufactured a demonstration effect that raises the price of its own inputs — Square Enix, Disney, Middle-earth Enterprises and Paramount now know exactly what a Magic set is worth, because Final Fantasy was the biggest set in franchise history.
3. Litigation. The securities class action (no loss accrued, “unable to estimate a reasonably possible range of loss”) and the stayed derivative action.
4. Cyber remediation. ~$20M of guided one-time opex, but the Q1 release states “the full scope of the costs and related impacts has not been determined,” with forensic file review and “any notifications deemed appropriate” still pending.
5. The coupon step-up. Notes carry a 0.25%–2.00% step-up on a downgrade, making ratings financially material — and the ratings are not disclosed in the 10-K.

How conservative is the accounting?

Mixed, and precisely locatable: honest where it counts, flawed where it is convenient.

Genuinely clean, and we say so: FCF is honestly stated and is the number to trust — FY2025 CFO $893.2M − capex $198.3M = FCF $694.9M, and the software spend IS deducted in capex, so FCF, unlike EBITDA and adjusted EPS, is not flattered by the capitalization. Receivables are fine (DSO 82.3d vs 81.2d — a +1.1 day change on $565.8M of revenue growth; CCC 61.0d vs 60.4d, flat, not the 12.6-day improvement the vendor feed implies). Deferred revenue FELL $46.0M while revenue rose 13.7% — the top line is not being inflated. Inventory $259.8M is disciplined (from $676.8M in 2022). SBC is NOT added back to adjusted EPS — credit to management. And the tax “anomaly” is not an anomaly: the −212.1% rate is arithmetic, since the impairment carried a tax benefit of only $5.4M because goodwill is largely non-deductible. (A modelling trap: anyone normalizing must NOT tax-effect the $1,021.9M add-back at 21% — the correct add-back is ~$1,016.5M.)

Not conservative, in four specific places:

1. The ~$413M of capitalized software has never touched the P&L. The 10-K states: “The Company did not release any software titles during 2025, 2024 or 2023 that were previously capitalized… and therefore there was no amortization or impairments recognized.” Three years of growing spend (~46% YoY) flatters operating profit, adjusted EBITDA ($1,361.5M) and adjusted EPS ($5.54) — but not FCF. Expensing the FY2025 $135.0M costs ~$0.74/share after tax (normalized EPS $5.33 → ~$4.59). The fair counter is that this is timing — but the titles slate for 2027, FY2026 carries zero of the cost, the balance approaches ~$500M by release, and there is a $24.4M write-off precedent (two cancelled titles, FY2024).
2. $0.21/share of recurring costs is scrubbed from adjusted EPS (“strategic transformation” $0.13 + restructuring/severance $0.05 + eOne divestiture costs $0.03). Cumulative restructuring is $163.5M to date and the 10-K describes the program as “ongoing.” A cost incurred every year for four years is an operating expense, not an adjustment. Mild but real.
3. FY2024 carried THREE prior-year error corrections — a $31.1M expense for historical environmental liabilities, a $26.7M benefit from over-accrued vendor commitments, and an $18.1M benefit from reversing performance-stock-award compensation (~$13.7M net benefit) — all booked to Corporate and Other rather than the Consumer Products segment they originated in, flattering FY2024 CP operating profit of $115.3M. The true CP run-rate is worse than reported, which makes the FY2024→FY2025 swing look more abrupt than it was.
4. Two impairment judgments deserve scrutiny. The Q2-2025 charge aggregated the four regional CP reporting units in the same quarter as the charge — aggregation raises pooled fair value and can reduce the measured shortfall versus testing a weak North America unit alone. Then the Q4-2025 annual test was QUALITATIVE only, with $561.0M of CP goodwill still on the books six months after a $1.02B quantitative write-down. Re-impairment risk is live, not settled (Family Brands’ $325.2M exceeded carrying value by only ~15% at the Q2-2025 test).

One more flag for anyone screening this name: a third-party fundamentals aggregator reports FY2025 operating income of $1,058.0M and a 22.5% operating margin; the filing reports $11.1M and 0.2%. The vendor reclassifies the impairment and disposal loss below the operating line — not the filing’s presentation. The same defect runs through every year (FY2023: filing −$1,538.8M vs feed $191.4M). The filing wins.

How CapEx-hungry is the business?

Structurally capex-light, and the intensity is where the whole business model reveals itself. FY2025 capex was $198.3M on $4,701.3M of revenue (4.2%) — but the composition matters more than the ratio: only $63.3M is PP&E; $135.0M is software development (FY24 $110.3M; FY23 $73.8M), which is a discretionary growth bet, not maintenance.

Two warnings. First, D&A now understates the economic charge. D&A of $171.3M against capex of $198.3M = 1.16x; stripping acquired intangible amortization, capex is 1.88x maintenance D&A, and FY2026 capex is guided to ~$250M = 1.46x current D&A. Rising capex will compress FCF conversion even if adjusted EPS grows. Second, the reported D&A collapse — −81%, from $908.7M (2021) to $171.3M (2025) — is not an efficiency story: it is eOne film/TV program cost amortization vanishing ($448.9M in 2023 → $35.8M in 2025). Anyone reading Hasbro’s D&A trend as operating improvement is reading the unwind of an acquisition.

Capital Allocation & Management

How much FCF does the business generate, how does management use it, what is the philosophy?

FACT: FY2025 FCF $694.9M (FY24 $649.9M); six-year cumulative FCF $3,595.5M. Uses: dividends $392.5M (56.5% of FY2025 FCF; 1.77x cover); $118.2M of note repurchases; the software build; zero buyback.

The philosophy, stated by revealed preference rather than by the strategy deck: pay the dividend, delever, and fund a video-game ambition out of a card franchise. Six-year dividends of $2,302.9M consumed 64% of cumulative FCF across a span in which the company cumulatively LOST $547.9M on a GAAP basis and took ~$3.42B of write-downs. INTERPRETATION: the dividend is not reckless — it is covered — but it consumed the optionality that could have deleveraged faster or bought stock at the 2023–24 lows.

Significant acquisitions recently?

None recently — and that is the good news. The record before it is the defining negative of the thesis, and none of it is a matter of opinion.

eOne (completed 2019-12-30). ~$4.41B of cash out (per the FY2020 cash flow statement), financed with $2.4B of notes, $1.0B of term loans and 10,592,106 shares at $95.00. Recovered: eOne Music to Blackstone for $385M (2021) and eOne Film & TV to Lionsgate for $375M cash (2023-12-27) — ~$760M gross. Retained: the Family Brands, carrying $325.2M of goodwill only ~15% above carrying value. Explicit eOne write-offs total $1,908.6M. Value destroyed ≈ $2.8–3.1B — roughly a quarter of today’s ~$11.8B market cap, in one transaction. (A reasoned estimate, not a disclosed number; it turns on the Family Brands residual. The headline price is itself disputed — $4.6B per Alta Fox vs ~$3.8B on another read, likely equity versus enterprise value; the five-year corpus never restates it.)

Power Rangers (closed 2018-06-12) was itself a failed acquisition: $522M purchase; $281.0M — ~54% — written off within 4.5 years.

~$3.42B of write-downs in six years, against $565.5M of book equity — roughly 6x book, and ~30% of today’s market cap. A company that writes off this much this often does not have a bad-luck problem; it has a capital-allocation problem. And the pattern has a single root: every charge traces to capital deployed OUTSIDE Wizards. WOTC’s goodwill has never been impaired.

The next mistake is already sized and dated. Hasbro has spent “nearly $1 billion on video games since 2018” across six studios, with no live-service titles and zero previously-capitalized titles released in 2023, 2024 or 2025 — implying ~$600M already expensed or written off with nothing capitalized ever shipped, and ~$413M awaiting a 2027 recognition event. The CEO’s own rationale is the tell: “If you invest a fair amount of money and give a fair amount of time to a talented team to do a more traditional game, you probably won’t make billions, but your chances of at least making your money back is much higher.” That is a CEO describing a ~0% ROIC project funded out of a 65–85% ROIC franchise. Under Greenwald’s expansion test, capital must be deployed inside the franchise where the barriers apply; AAA development is squarely outside it. And the strategic contradiction is the damning part: Hasbro’s two biggest digital successes — Baldur’s Gate 3 (Larian) and Monopoly GO! (Scopely) — were both achieved by LICENSING the IP out and bearing no development risk. Hasbro’s conclusion from its own best data was to build six studios instead, stepping DOWN into the ~30% platform toll it currently sits above. The wrong lesson, drawn from its own evidence.

Buying back shares?

FACT: zero repurchases in FY2023, FY2024 and FY2025. The last was $125.0M in FY2022 at an average $87.46. Over those three years SBC totalled $203.6M against $0 of buyback, and shares outstanding drifted 138.18M → 140.39M (+1.6%). In February 2026 the board authorized $1.0B, replacing all prior authorizations, with no expiration and no obligation.

INTERPRETATION: the timing is textbook procyclical. Hasbro bought $125.0M in 2022, bought NOTHING through the 2023–24 trough when the stock traded in the $40s, and then authorized $1.0B in February 2026 — within days of a five-year high, and in the same week insiders sold ~$38.6M at the 99th percentile of the five-year price range. Management did not buy when it was cheap and is authorized to buy now that it is not. This is the opposite of the discipline the 10-K claims: “The Company has a long history of increasing shareholder value through its share repurchase program.” Whether the $1.0B is actually executed, and at what prices, is the single best forward test of whether capital-allocation behaviour has changed.

Issuing large amounts of new shares to insiders?

No — and this is a genuine, if narrow, credit. SBC was $80.4M in FY2025 = 1.7% of revenue and 11.6% of FCF — modest by any standard, and below the 2021 level in absolute dollars ($97.8M 2021 → $83.4M → $72.4M → $50.8M → $80.4M; FY2024 was the anomaly, depressed by missed targets, not FY2025 an explosion). It is not added back to adjusted EPS — credit to management, unlike most of the comp set. Dilution runs ~0.4%/yr.

The flaw is that it is unoffset. Shareholders funded management’s compensation through dilution for three straight years with no buyback. Separately, 10,592,106 shares were issued at $95.00 to help fund eOne — the largest issuance in the window, and it funded the worst decision.

Compensation policy of directors/management?

The defining fact of Hasbro’s compensation design, and it is not subtle.

FACT: CEO Chris Cocks’s FY2025 total compensation was $18,672,521 — his highest ever (salary $1.5M; stock awards $10.5M; non-equity incentive $6,000,000; other $672,422). FY2024 $16,841,413; FY2023 $15,110,869. Three-year total: $50,624,803. CEO pay ratio 165:1 (median employee $113,241).

The FY2025 annual incentive paid at 200% of target — the plan MAXIMUM — and the Committee “determined to not make individual performance modifications for the NEOs.” The metrics, quoted from the payout table:

That “$1,114M” of Operating Profit Dollars is an adjusted figure that EXCLUDES the $1,021.9M goodwill impairment. Cocks’s target bonus was raised from 175% to 200% of salary “to more closely align with market”; the $6.0M is $1.5M × 200% target × 200% payout.

In a fiscal year in which Hasbro wrote off $1.02B of goodwill, reported a $102.0M pre-tax loss, a $318.2M net loss and −$2.30 of diluted EPS, the CEO was paid $18.67M — his highest ever — including a maximum 200%-of-target cash bonus, on a profit metric constructed so that destroying capital cannot reduce the bonus.

And it got worse by design. ROIC was DELETED from the long-term incentive plan. Per the FY2025 10-K Note 16: “The 2023 Stock Performance Awards are measured based on achieving targets set for diluted earnings per share and return on invested capital (‘ROIC’)… The 2024 and 2025 Stock Performance Awards are measured based on achieving targets set for diluted earnings per share, in addition to a TSR modifier.” Hasbro carried a ROIC metric in 2023 and removed it for 2024 and 2025 — removing the only capital-efficiency measure in the entire incentive architecture, immediately after the largest capital-destruction event in company history. The Committee also removed the strategic cost-savings metric and “increased the emphasis on net revenue and operating profit dollars.” Every remaining metric — net revenue, adjusted operating profit dollars, cumulative diluted EPS — is a scale or accrual measure that a value-destroying acquisition can flatter and that an impairment cannot touch. The relative-TSR modifier is a ±25% modifier on the outcome, not a gate.

Pay versus performance: revenue fell ~16% from FY2020 to FY2025; the stock compounded at ~0.5%/yr over five years and ~3.3%/yr over ten at ~33% volatility. Over the same window the CEO was paid $50.6M in three years.

Motivations of management?

Read the Form 4s rather than the proxy’s stated philosophy. Across 281 Form 4s, 5 Form 4/As and 16 Form 3s (393 transactions, 35 filers, 2021-03-23 → 2026-06-30):

INTERPRETATION: these insiders behave like optionees monetizing compensation, not owners accumulating an asset. (Fair nuance: Cocks’s total holdings ROSE from 249,861 to 303,310 through the February 2026 cluster — that was option monetization, not an exit.)

What management has done right, stated plainly. Debt management is genuinely good — leverage from roughly $5B post-eOne to $3.28B, 100% fixed, ~2.0x, termed out via a 4.650% five-year print in March 2026 (against 6.05% in May 2024). The Operational Excellence Program is real and largely delivered — “almost $800 million of gross cost savings,” ~half of all SKUs eliminated (representing only ~2% of revenue), zero-based budgeting. The “Playing to Win” asset-light pivot is directionally correct and backed by structural exit — the studio sold, film co-financing ceased November 2024, the cable JV zeroed. Two caveats bind all three: the deleveraging was funded by selling the assets it should never have bought (cleaning up a self-inflicted wound competently is not creating value); the cost goalposts moved $350–400M → $750M → $1.0B and the framing shifted from net to gross; and the asset-light pivot is a consequence, not a philosophy — Hasbro did not choose it in 2019 when it had the option, it arrived there after spending $4.4B to learn.

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer?

No, on all three — and there is nothing complicated here. Hasbro, Inc. is a US domestic C-corporation, incorporated in Rhode Island, a Form 10-K filer (CIK 0000046080), with a single class of common stock listed on NASDAQ. Ordinary Form 1099-DIV treatment; no K-1, no partnership pass-through, no ADR ratio, no foreign withholding, no FPI accommodations. Fiscal year ends the last Sunday of December (2025-12-28), so the 52/53-week calendar occasionally produces declaration-timing artifacts — the FY2024 “$2.10 declared” is one, and is not a dividend cut.

Dividend policy?

FACT: $2.80/share, $0.70/quarter — unchanged for four years (rate set in 2022). Per share: $2.72 (2020) → $2.71 → $2.78 → $2.80 → $2.80 → $2.80 (2025). Nominal growth 2020→2025: +2.9% cumulative — not annual. Dividends paid: $372.7M → … → $392.5M (FY2025); six-year cumulative $2,302.9M. Yield 3.41%; cover 1.77x on FY2025 FCF (~2.6x on working-capital-neutral FCF); payout 56.5% of FCF, 51% of adjusted EPS, 61% of cash-real normalized EPS.

INTERPRETATION: US CPI rose ~24% over the span. A +2.9% nominal dividend is a real-terms cut of roughly 18% — administered silently, while management takes credit for never having cut it. It is safe, and it was not free: $2.3B was distributed over six years in which the company cumulatively LOST $547.9M on a GAAP basis, took $2.68B of impairments and carried $3.3B of debt; in the stress year (FY2022) it was cash-funded — FCF $198.7M against $385.3M of dividends plus $125.0M of buyback, a $311.6M shortfall, and cash fell $493.4M. Holding it flat in nominal terms was the compromise: management refused either to cut it (admitting the balance sheet was strained) or to grow it (admitting it could afford to).

How profitable is the business?

Answered in full above. In one line: GAAP FY2025 says the company earned $11.1M of operating profit and lost $2.30 a share; the economics say one segment earns a 46.0% margin and ~65–85% ROIC while the other destroys capital, and the consolidated number is a mix artifact, not a quality signal. Use normalized EPS ~$5.33 (or ~$4.59 cash-real); never GAAP −$2.30, and never management’s $5.54 unlabeled.

Is net income diverging from cash from operations?

Yes, violently — and it is benign. Say so rather than manufacture a red flag. FACT: FY2025 CFO $893.2M against a net loss of $(318.2)M — a −2.77x ratio. The driver is the $1,021.9M non-cash goodwill impairment plus $158.3M of deferred tax, not accrual manipulation. Against normalized net income (~$748M), CFO/NI ≈ 1.19x — healthy. The corroborating evidence is that every accrual channel through which a divergence would normally be engineered is clean: DSO +1.1 days, deferred revenue FELL $46.0M while revenue rose 13.7%, inventory down from $676.8M (2022) to $259.8M, and FCF fully deducts the $135.0M of software spend.

The genuine divergence is the opposite one, and it runs the other way: EBITDA and adjusted EPS are flattered by ~$413M of never-amortized capitalized software, while FCF is not. That is why FCF, not adjusted EBITDA, is the honest number in this filing.

Valuation reference points (no target, no recommendation): at $82.09, EV $14.55B (re-derived at the live price; the vendor snapshot at $15.47B is struck at a stale 2026-03-31 close). EV/TTM EBITDA 11.26x against a ~12.2x ten-year year-end average; ~13.9–14.9x forward adjusted EPS; FCF yield ~5.9%; P/tangible book n/m (TCE negative ~$1.17B). The own-history 90.2nd-percentile composite is a mix artifact, not evidence of excess: the P/B percentile is meaningless on negative tangible equity, and P/S is high precisely because the mix shifted to a 46%-margin franchise — a company whose margin structure quintupled should trade at a record price-to-sales.

Risks & Downside

What factors would cause the stock to decline?

Ranked by this analysis’s risk matrix, most severe first:

1. Magic profit normalization from a cyclical/speculative peak (HIGH/HIGH) — collector boosters already −60–70%; FY2026 guided +3–5% cc after +13.7%. Largely in progress and partly priced.
2. Single-asset concentration (CERTAIN/HIGH) — Wizards is >100% of GAAP operating profit and net debt / Wizards’ EBITDA alone is 2.43x: essentially the entire $2.5B net debt load is serviced by one segment and, within it, disproportionately by one brand. This is a thin-tailed 2.0x, not a fat-cushioned one.
3. Capital-allocation recurrence (MED-HIGH/HIGH) — the ~$1B video-game program; the next eOne.
4. TCG capital-cycle reversal (MED-HIGH/HIGH) — the Pokémon plant at full scale late 2028; PSA’s $200M build.
5. Universes Beyond royalty step-up / licensor hold-up (MED-HIGH/MED) — $90–115M/yr and rising structurally as UB goes to a majority of the slate.
6. Hasbro’s own price/cadence policy degrading the moat (MED/HIGH) — seven sets, higher MSRPs, less in the box, rotation to 32–36 months, and the head designer conceding player fatigue.
7. CP re-impairment (MED-HIGH/MED) — $561.0M of CP goodwill survives on a QUALITATIVE Q4-2025 test only; Family Brands’ $325.2M only ~15% above carrying.
8. Monopoly GO! royalty decay (HIGH/LOW-MED) — bookings ~−52% YoY, held up by a non-repeatable UA-deduction cushion: as Scopely cut user acquisition, Hasbro’s effective royalty rate mechanically rose, so the royalty looks stable while underlying bookings halved. Once UA approaches zero the rate cannot rise further and the royalty decays 1:1.
9. The 2027 capitalized-software recognition event; the cyber cost scope; the class action; the undated Star Wars/Marvel renewal; tariff volatility.

Two things are NOT in the price in either direction: the ~$50M tariff refund (upside) and the 2027 software recognition, the licence renewal and the loot-box tail (downside).

Risk of a catastrophic loss?

Low, and the reasons are specific rather than reassuring. The balance sheet does not break: ~2.0x net leverage, 8.3x interest cover, 100% fixed-rate debt, $881.6M of liquid assets against a $497.0M 2026 maturity, an undrawn $1.1B revolver to February 2031, covenant compliance, and demonstrated market access at 4.650% in March 2026. No maturity wall (2026 $497.0M | 2027 $475.0M | 2028 $109.9M | 2029 $900.0M | thereafter $1,300.0M).

The one genuine catastrophic-loss channel is regulatory, and it is the report’s most under-priced tail: loot-box/gambling reclassification of PHYSICAL card packs (LOW likelihood / HIGH impact). Belgium banned paid loot boxes outright in 2018 (penalties to €800,000 and possible imprisonment of officers). The Netherlands’ KSA declared loot boxes whose contents are tradeable illegal — the legal hinge is the “money’s worth”/tradeability limb. A physical Magic booster fails that test far more clearly than a video-game cosmetic does: it is a randomized paid product whose contents have a deep, liquid, publicly-quoted secondary market and a professional grading industry that exists specifically to certify their monetary value. Brazil enacted a prohibition on loot-box sales to under-18s effective March 2026. The very infrastructure that makes Magic’s moat durable is the same fact that satisfies the gambling test — the moat and the tail risk are the same fact. No major jurisdiction has yet applied loot-box law to physical packs; physical TCGs have escaped via forbearance rather than a principled distinction. This is not a near-term earnings risk. It is a low-probability, high-impact tail on the single asset generating ~100% of Hasbro’s operating profit, and it is not in the price.

A second, lesser channel: a format reset that vaporizes the backlist’s value. Only Hasbro can do this, and it would be commercially suicidal — but note that Hasbro attempted the analogous act at D&D in January 2023 and was beaten back by its own customers. The instinct exists in this management.

Chance of a total loss?

Negligible, and this analysis says so without hedging. Wizards alone earns $1,006.8M of operating profit on $12.9M of capex behind a barrier that has repelled Valve, Riot, Disney, Lucasfilm and Blizzard over 32 years; FCF is $694.9M; leverage is ~2.0x all-fixed with no wall; the dividend is covered 1.77x. The risk that is not on the risk list, because it is not a risk — it is the thesis: Magic being competed away. No evidence in this file supports it. The natural experiment already ran: the speculative tranche fell ~two-thirds and Magic grew +36% anyway with backlist at a record.

The honest framing of downside here is normalization, not destruction — and that distinction is what separates a fair price from a value trap. The plausible bad outcome is a de-rating toward a lower Wizards multiple on a $750M rather than $900M base, not an impairment of the enterprise.

Recent News & Events

Has the business environment changed recently?

Yes, in four ways.

1. The cybersecurity incident (March 2026) — live, unquantified, and the next catalyst. Unauthorized network access identified 2026-03-28; systems proactively taken offline; the Q1 10-Q filed late under Rule 12b-25 as a direct result. Guided impacts: ~$20M of one-time remediation opex (excluded from adjusted EBITDA); $40–60M of Consumer Products revenue delayed from Q2 into H2 2026; a receivables/cash-flow shift from Q2 into Q3. Management says it is “contained” with systems restored by ~June 2026 — but the Q1 release disclosed that “the full scope of the costs and related impacts has not been determined,” with forensic file review and notifications pending. The assumed 2H recovery is management’s assumption, not a fact. Analytically the incident did something useful: it cleanly separated the two Hasbros. Magic shipped on cadence; the physical-goods business took the entire hit.
2. The tariff regime reversed. On 2026-02-20 the Supreme Court struck down the IEEPA tariffs Hasbro had paid since April 2025; a ~$50M refund claim is in reconciliation with no timeline and is not in FY2026 guidance. Surviving: Section 301 List 4A at 7.5% on Chinese-origin toys plus a Section 122 10% tariff effective 2026-02-24 (ruled unlawful by the CIT 2026-05-07, stayed by the Federal Circuit 2026-06-11, collection continuing) — ~17.5% combined. (The Section 122 timeline is sourced to tariff-calculator sites and requires primary verification; the SCOTUS ruling and refund optionality are corroborated by the 10-K.) Realized FY2025 tariff cost was only $44.9M against a guided $100–300M gross, and ~40% of sourcing is targeted out of China by end-2026.
3. Litigation aimed at the crown jewel’s narrative — the securities class action, MTD pending.
4. A wealth-bifurcated consumer — premium collectibles boom while mass-market toys shrink.

Significant acquisitions?

None. The two-year record is the inverse — structural exits: eOne Film & TV sold to Lionsgate (December 2023), film co-financing ceased November 2024, the Discovery Family Channel JV zeroed. The forward capital-allocation question is not an acquisition; it is the $1.0B buyback authorization (February 2026) and whether ~$695M/yr of FCF gets deployed intelligently.

Change in accounting policies?

No policy change of substance — but three reporting/estimate changes materially affect what can be analyzed, and one should be flagged rather than waved through.

• ASU 2023-07 dropped the segment asset column the FY2022 10-K carried. This blocks a clean segment-level ROIC and forces the ~65–85% Wizards estimate to be built from fragments (WOTC goodwill $370.5M + the ~$385.6M of capitalized software + working capital). An assumption-heavy denominator; the conclusion is not close to the bar regardless.
• ASU 2023-09 produced the FY2025 rate reconciliation that makes the −212.1% rate legible (Goodwill impairment: $209.8M / (205.8)%).
• The FY2024 prior-year error corrections (§ above) — three of them, ~$13.7M net benefit, booked to Corporate and Other rather than the Consumer Products segment they originated in. Not a policy change; an estimate correction. It flatters FY2024 CP profit of $115.3M and makes the FY2024→FY2025 swing look more abrupt than it was.
• The Q2-2025 reporting-unit aggregation and the Q4-2025 qualitative-only annual test are management judgments, not policy changes — but both affect the measured charge and the surviving $561.0M.

Recent changes — new markets, facilities, management?

Facilities. The Boston HQ relocation: a long-term sublease executed October 2025 for 265,000 sq ft at 400 Summer Street, move expected Q4 2026, with the ~366,000 sq ft of owned Pawtucket property to be listed for sale. INTERPRETATION: structurally defensible — a smaller footprint, no ownership, a monetization, and a real talent market for the digital pivot. But the build-out cost and expected proceeds are undisclosed, and it is discretionary cash and attention spent during a turnaround; the 10-K concedes it “requires substantial investment and could disrupt operations.”

Management and board — genuinely improved, and one piece of circulating misinformation should be corrected. The board went from the defensive 13 of 2022 to 11 nominees today, with five of 11 joining in 2024 or later, rebuilt around gaming operators: Frank Gibeau (Zynga), Owen Mahoney (Nexon) in 2024; Carla Vernon and Douglas Bowser (President/COO, Nintendo of America) in January 2026. A materially more credible board for a company whose profit is a games franchise. Executive churn has been real but is routinely overstated: Eric Nyman lasted 13 months as COO, and Cynthia Williams resigned the Wizards presidency in April 2024 — but the seat was filled within three months by John Hight (2024-07-19), a 12-year Blizzard veteran who ran Warcraft. The “crown jewel leadership is vacant” framing that circulates on this name is stale and wrong.

Financing. The Fourth A&R revolver (2026-02-20) extended to February 2031 at improved pricing but cut the commitment to $1.1B (+$550M accordion), fully undrawn and covenant-compliant; $400M of 4.650% notes due 2031 issued 2026-03-12. The credit market is repricing the post-eOne balance sheet favourably. Ratings remain undisclosed — and the 0.25–2.00% coupon step-up on a downgrade makes them financially material.

New markets. None of consequence; international is a headwind (Asia Pacific −13.0%, Latin America −18.0%). The genuine “new market” is the AAA video-game entry — six studios, ~$1B since 2018, ~$413M capitalized, first titles 2027 — and on this evidence it is the next capital-allocation event, not the next growth engine.

Cost and macro into FY2026. ~$40M of below-the-line EPS headwind (higher interest, lower non-operating income including the absence of prior-year Swiss deferred-tax benefits); ~$30M H2 headwind from oil-related input costs at ~$100/bbl, largely contained to CP; ~$15M favourable tariff variance.

APPENDIX B — Source Appendix

Report date: 2026-07-17 | Price reference: $82.09 (close, 2026-07-16) | CIK: 0000046080 | All URLs accessed 2026-07-17 unless otherwise stated.

This appendix lists the public sources relied on, ordered by evidentiary authority: SEC filings first, then transcripts, then third-party quantitative data, then industry/trade sources. Section 7 grades the sources and is the most important part of this appendix — it states plainly which claims rest on institutional-grade evidence and which do not.

Standing binds carried throughout. (a) For a US filer, EDGAR and the filing govern; every third-party aggregator is a cross-check, never the authority. (b) Management commentary — transcripts, guidance, press releases, investor-day framing — is a hypothesis, not evidence, and is labelled as such at every point of use.

1. Primary sources — SEC filings (highest authority)

The trailing 60-month corpus (filings since 2021-07-01) was reviewed via SEC EDGAR. Every filing below was read in full. Corpus composition: 5× 10-K, 15× 10-Q, 56× 8-K, 281× Form 4 + 5× Form 4/A + 16× Form 3, 4× DEF 14A, 20× DEFA14A, 2× DEFC14A, 29× DFAN14A, 1× PREC14A, 1× PRRN14A, 1× NT 10-Q, plus ARS/S-3ASR/S-8/SD/11-K. The Q4 FY2025 and Q1 FY2026 earnings-release exhibits (EX-99.1 — where the segment “as adjusted” tables and non-GAAP reconciliations live) were retrieved directly from sec.gov/Archives/edgar/data/46080/.

1.1 Annual reports (Form 10-K)

1.2 Quarterly reports and late-filing notice

1.3 Proxy statements

1.4 The 2022 contested proxy record (Alta Fox Capital Management)

The full contested record — 29 DFAN14A, 20 DEFA14A, 2 DEFC14A, 1 PREC14A, 1 PRRN14A, clustered 2022-02-17 → 2022-06-06 — is public on SEC EDGAR. The entire “Free the Wizards” thesis and its outcome were reconstructed from this public record plus press corroboration (§6.9).

Also relied on: DFAN14A 2022-04-18 (governance questions; demand that the board publish its spin analysis), DFAN14A 2022-05-19 (slate cut 5 → 3), and 8-K 2022-06-08 (Item 8.01, preliminary results).

1.5 Current reports (Form 8-K) — the material-event timeline

56 8-Ks were reviewed across 2021-07 → 2026-06. The ~18 material ones:

Negative findings from the 8-K corpus, recorded because they are evidence: zero dividend or buyback 8-Ks (Hasbro declares by press release); zero credit-rating-action 8-Ks (ratings are not disclosed in the FY2025 10-K — an open question, and material given the 0.25%–2.00% coupon step-up); zero standalone impairment 8-Ks; zero occurrences of “Boston” across all 56 filings (the HQ relocation is disclosed only in the FY2025 10-K Item 2); and only five 8-Ks in all of 2025, all routine — a quiet year between the 2022–24 repair and the 2026 cyber shock.

1.6 Insider filings (Forms 3, 4 and 4/A)

Corpus: 281 Form 4 + 5 Form 4/A + 16 Form 3, covering 393 transactions across 35 filers, 2021-03-23 → 2026-06-30 (all public on SEC EDGAR). The transactions were machine-parsed from the XML filings and aggregated by transaction code, not eyeballed.

Every open-market purchase in five years — all three, all within a five-day window during the Alta Fox proxy fight, six weeks before the vote, and ~10 weeks after Alta Fox publicly stated no director had bought a share in a decade:

Further facts relied on: only 2 of 35 insiders ever bought a share; exactly one of 32 sales cites a Rule 10b5-1 plan (Sibley, 2022-01-04, 2,000 shares, $210,000) — i.e. ~99.7% of sale dollars were discretionary, and the aff10b5One checkbox (present in 166 filings from 2023) is always 0; 2023 — the year containing the five-year low of $38.76 — contains zero code-P and zero code-S transactions (nobody bought the bottom); and the Feb-2026 cluster in which Cocks exercised 377,512 options (strikes $55.78–$96.79) and sold 377,992 shares at $100.32–$106.48 for ~$38.6M at roughly the 99th percentile of the five-year price range — with Goetter, Sibley, Kilpin, Bunge and Barbacovi selling alongside. Fair nuance carried into this analysis: Cocks’ total holdings rose 249,861 → 303,310, so this was option monetization, not an exit. Goldner’s $21.3M July-2021 sale at ~$100–102 preceded his October-2021 leave and death and is treated as context, not signal.

1.7 SEC data APIs

• SEC EDGAR XBRL company-facts API — https://data.sec.gov/api/xbrl/companyconcept/CIK0000046080/us-gaap/{GoodwillImpairmentLoss, AssetImpairmentCharges, CommonStockDividendsPerShareDeclared}.json — used to build the six-year impairment series and the dividend-per-share series directly from the filer’s structured data, independently of any aggregator. Accessed 2026-07-17.

2. Earnings-call transcripts

Transcripts were sourced from public earnings-call transcript providers and the company’s own investor-relations releases, accessed 2026-07-17. Coverage is earnings-call-centric; the February-2025 “Playing to Win” investor day deck is not part of the transcript set, so this analysis relies on the proxy’s description of the five Playing to Win building blocks (a primary source) rather than on the investor-day deck itself.

Management commentary is a hypothesis, not evidence. All management commentary above is a HYPOTHESIS, not evidence. Where this analysis uses it, it is either (a) validated against a filing or external data, (b) labelled as management’s claim, or © explicitly rejected. Three specific instances carried: Cocks’s “80–90% player or player-collector” mix claim is not accepted as evidence (§7); the “juggernaut” characterization of Monopoly GO! is carried as a contradiction against third-party bookings data (§7); and the WPN “11,000+ stores / +20%” figures are load-bearing to the moat verdict and independently unverifiable (§7).

3. Analytical framework context

The valuation and competitive analysis draw on two published frameworks and on publicly-observable comparisons to peer IP, gaming and toy businesses:

• Bruce Greenwald & Judd Kahn, Competition Demystified — the three genuine advantage types (supply/cost, demand/captivity, economies of scale + captivity); the market-share-stability test; the 15–25% ROIC bar as a moat pass/fail check; “think local”; “brand alone is not a barrier” (the Mercedes-Benz case); and expansion only inside the franchise. Applied per-asset in §4, to the industry in §3, and to the video-game spend in §7 of this analysis.
• Edward Chancellor (ed.), Capital Returns (Marathon Asset Management) — supply-side capital-cycle analysis; the asset-growth anomaly; high returns attract capital and mean-revert. Applied to the TCG capital cycle in §3.
• Peer reference points (all public): a mature franchise-IP publisher such as EA has historically traded around ~15–18x P/E / ~14–16x EBITDA absent a deal premium — the benchmark used to strike the implied-Wizards multiple in §10; the ~30% app-store platform tax that first-party publishers (EA, Take-Two, Roblox) cite as the largest external claim on their profit pool, which frames why Hasbro’s above-the-toll licensor position is underrated; the goodwill-drag ROIC decomposition seen at Disney (reported blended ROIC understating segment returns), reproduced in §6.4 as the 23.5%-reported vs 13.7%-on-capital-deployed split; and the own-history percentile methodology (comparing a recovery name to its own history, not a cross-section of peers at different cycle points), applied in §10.2 to reject the “richest-ever multiple” framing for this name.

This is fresh coverage of Hasbro; the analysis was built from the primary and public sources catalogued in Sections 1, 2, 5 and 6.

5. Quantitative data feeds — with authority caveats

All four feeds below are third-party. None is primary. For a US filer the filing governs; where a feed and a filing disagree on a material number, the filing wins and the discrepancy is recorded. Two of the four returned material defects on this ticker, documented below. No figure from any of them is a price target or a valuation conclusion.

5.1 Third-party fundamentals aggregator — NOT PRIMARY

A third-party financial-data aggregator was used, accessed 2026-07-17, for HAS (annual FY2018–FY2025 + quarterly), MAT and GAW.L: the three statements, profitability/valuation ratios, enterprise value, per-share data, and the earnings-call transcripts.

What was used: revenue, gross profit, cash-flow lines and balance-sheet lines (all reconcile to the filings); the 10-year own-history EV/EBITDA and P/S series; MAT and GAW.L comparatives; and the transcripts (§2).

Three material defects found on this ticker — all three matter, and all three are recorded:

(a) HAS operating income / margin / EBITDA / ROIC / ROE are a pseudo-adjusted series, not GAAP — the defect runs through every year. The aggregator reports FY2025 operating income $1,058.0M / 22.5% margin. The filing reports $11.1M / 0.2%. The feed reclassifies the $1,021.9M goodwill impairment and the $25.0M eOne disposal loss below the operating line — which is not the filing’s presentation: the 10-K places the impairment inside the Consumer Products segment’s operating result. Arithmetic check: 1,058.0 − 11.1 = 1,046.9 = 1,021.9 + 25.0.

Consequence: the 22.5% figure is never quoted as a reported operating margin. GAAP FY2025 operating margin is 0.2%. Coincidentally the feed’s $1,058.0M equals the independently-derived normalized (ex-impairment, ex-disposal) operating profit — so the level is usable as a normalized input, but the label is wrong. Relatedly, the aggregator’s ROE (-15.9% FY2025 / 16.5% FY2024) and ROIC (12.2% FY2024; not reported at all for FY2023/FY2025) are computed on a book equity of $538.5M supported by -$1,174.9M of tangible common equity — the negative-tangible-equity problem. All ROIC in §6.4 is hand-computed from filing line items. The feed’s cash-conversion cycle (FY2025 60.5d / FY2024 73.1d) reconciles for FY2025 (hand-computed 61.0d) but not FY2024 (hand 60.4d, a 12.7-day gap) — hand-computed figures govern.

(b) The aggregator’s enterprise value is a fiscal-quarter-END snapshot — stale by construction. It printed market cap $13.18B / EV $15.47B, struck at the 2026-03-31 quarter close (~$93.6/share). The stock closed $82.09 on 2026-07-16 (-12%). EV was therefore re-derived at the live price off the Q1 FY2026 balance sheet (debt $3,623.3M, cash $857.1M, minority $25.9M → net debt $2,766.2M; 143.2M diluted shares): EV = $14.55B, which is what §10 uses. The stale snapshot would have overstated EV by ~$0.9B and every multiple struck on it.

© The aggregator returns GAW.L’s market cap in PENCE while the statements are in GBP — a 100× unit mismatch. The printed EV/EBITDA of 1,933.5x and EV/sales of 818.2x are garbage (market cap returned as 505,322,790,000). Corrected: market cap ≈ GBP 5.05B, EV ≈ GBP 5.05B (net cash), i.e. EV/EBIT ~19.3x and EV/sales ~8.2x — the figure used as the Games Workshop comp. Always sanity-check aggregated market cap for LSE-listed (pence-quoted) issuers.

5.2 Price history — the price source of record

Full split-and-dividend-adjusted daily price history (11,677 rows from 1980-03-17; last close $82.09 on 2026-07-16) was used. Every price, range and return statistic was recomputed programmatically over both the adjusted close and the unadjusted_close series — not read by eye. This is what caught the all-time-high error recorded in §7.

5.3 Own-history valuation percentiles — own-history context ONLY

An own-history valuation-percentile series, at 2026-07-16, price $82.09: composite percentile 90.175 (based on 2 usable components); P/B 18.06x / 90.8th percentile; P/S 2.42x / 89.5th percentile; P/E null (GAAP TTM EPS -$1.65).

Caveats, all applied rather than merely noted. (a) Own-history context ONLY — never cross-sectional. (b) The P/B percentile is meaningless for this name: book equity of $538.5M is entirely goodwill and intangibles against negative ~$1.17B tangible common equity, hollowed by buybacks and serial impairments; price/tangible-book prints -10.0x and tangible book per share is -$8.19. © The P/E percentile is null because GAAP EPS is impairment-distorted. (d) With P/E null and P/B meaningless, the composite rests on effectively one usable component (P/S) — and §10.2 argues that the 89.5th-percentile P/S is a mix artifact, not evidence of excess: consolidated operating margin went from 3.8% (FY2023) to ~22% normalized (FY2025), and a company whose margin structure has quintupled should trade at a record price/sales. This analysis therefore rejects the “richest-ever multiple” framing for this name and leads with EV/EBITDA (11.3x TTM vs a ~12.2x ten-year average) and forward P/E (~14x).

5.4 News-sentiment feed — uninformative for this name; recorded as a finding

A third-party news-sentiment feed returned only 11 articles, and they are generic market-wide items (e.g. “Stocks Finish Sharply Higher on Iran Hopes”), not HAS-specific news — uninformative for Hasbro. Consequence: the §8 recent-events timeline and the five-year event map were built instead from the 8-K corpus, the 10-K, the 10-Q, the transcripts and trade press — recorded so the absence is not mistaken for a quiet tape. It is not a quiet tape; it is an absent feed.

5.5 Factor / risk model — third-party statistical estimates, not primary

A third-party quantitative risk/factor model (loadings dated 2026-07-16, 756-day window), accessed 2026-07-17.

Relied on for: the risk-adjusted track record (y10 +3.28%/yr, Sharpe 0.04; y5 +0.46%/yr, Sharpe -0.05; y3 +12.86%; y1 +11.75%; vol ~33% at every long horizon; lifetime max drawdown -63.8%, y5 max DD -55.1%); factor loadings (Base R²=0.260: Market +0.876, DividendYield +0.382, Momentum -0.312, SmallSize +0.247, Quality +0.049, Value +0.034); R² of only 0.26–0.32 ⇒ 68–74% of variance is idiosyncratic, with specific vol 24.85% vs ~28.5% y1 total vol (~87% of realized vol stock-specific); the 252-day factor regime (DividendYield +17.13%, z=1.69 — the best-performing style factor); and factor-similar peers (COWZ 0.914, APTV 0.906, LEA 0.902, JBHT 0.889, SCHD 0.886 — no toy, games or IP peer anywhere in the list).

Two binds. (a) The annualization convention: every horizon’s return is annualized, including the short windows. The reported quarter-return of -0.382 is not a -38% quarter — de-annualized, (1-0.382)^(1/4)-1 = -11.3%, which matches the computed price-history -11.3% exactly. This analysis reports the raw moves. (b) Loadings and returns are reportable facts; “it will continue / mean-revert” is interpretation, labelled and regime-caveated. The negative Momentum loading coexisting with a +115% run off the Apr-2025 low is a construction artifact of the 756-day window and the 12-1m momentum definition — flagged so it is not read as “the stock has not gone up.”

6. Industry, competitor and market sources

Organized by topic. Grade column: P = primary (company/regulator/filing); T = trade press or industry data with a named methodology; W = hobby-press / wiki / community — direction reliable, magnitude approximate; L = low-quality syndicated. See §7 for what each grade is permitted to support.

6.1 Toy industry sizing and category decomposition

6.2 Mattel — the direct comp

Why it matters: the best-run pure-play toy company alive — with Barbie, Hot Wheels and a live Hollywood flywheel — did -1% revenue with margins down 200bp in a +6% US / +7% global year. That is what makes §3.2’s verdict an industry verdict rather than an execution verdict — and it is the benchmark that shows Hasbro’s CP (3.3% ex-impairment margin, loss-making in Q1) is a materially worse version of an already-mediocre business.

6.3 Games Workshop — the cleanest listed analogue for Wizards

6.4 Pokémon — the #1 competitor, and the capital cycle in plain sight

6.5 Disney Lorcana — the leading indicator, and the most valuable comp in the file

Why this is the most valuable comp: Lorcana is the same product cycle as Magic, started ten years later, and its speculative tranche has already rolled over while player demand held — the exact decomposition the market is not applying to Magic, and the decomposition that Magic’s own collector-vs-play-booster pricing (§6.7) then confirms on the target itself.

6.6 Competitive set, TCG marketplace and the graveyard

6.7 Magic — product, pricing, cadence and the secondary market

6.8 Dungeons & Dragons — the OGL natural experiment

Not used, and recorded so it is not picked up later: a ~$460M FY2025 D&D revenue figure circulates in trade commentary. It is not company-sourced and does not appear in this analysis as fact. Hasbro has never disclosed D&D revenue, D&D Beyond subscribers or D&D profitability in any filing in the five-year corpus.

6.9 Video games, personnel, and Hasbro corporate

6.10 eOne, Power Rangers and the 2022 proxy contest — press corroboration

6.11 Monopoly GO! / Scopely — the decaying royalty

6.12 Tariffs and regulation

6.13 Collectible grading — the speculative superstructure

Disclosure limit recorded: PSA’s parent Collectors Universe is private (Nat Turner / Steve Cohen-backed), so grading disclosure is partial. Grading volume is nonetheless the best available real-time leading indicator for the TCG speculative cycle.

6.14 TCG market sizing — LOW QUALITY, NOT A VALUATION INPUT

These three vendors disperse ~55% on the same year ($8.4B vs $13.3B). Used directionally only (“a ~$9–13B market growing high-single-digit”), never as a valuation input. See §7.

6.15 Analytical frameworks

• Analytical frameworks: Greenwald & Kahn, Competition Demystified (the three genuine advantage types; the market-share-stability test; the 15–25% ROIC bar; “think local”; scale advantages must be defended move-for-move; brand alone is not a barrier — the Mercedes-Benz case; “in the long run everything is a toaster”; expansion only inside the franchise — “if you don’t bring anything to the dance, don’t expect to take anything home”) and Marathon, Capital Returns (supply-side capital-cycle analysis; the asset-growth anomaly; high returns attract capital and mean-revert). Applied per-asset in §4, to the industry in §3, and to the video-game spend in §7 of this analysis.

7. Source quality and evidence grading

This section states what this analysis’s conclusions actually rest on. It is deliberately unflattering where the evidence is thin.

7.1 Grading table

7.2 Unresolved contradictions this analysis carries (rather than resolves)

These are surfaced in the Fact-vs-Interpretation table and Open Questions, not smoothed over.

7.3 Corrections made before publication

Recorded as evidence process — each of these would have reached this analysis uncorrected.

7.4 Known unreconciled items and disclosure gaps

Every material claim in this analysis traces to a source listed above. No BUY/SELL recommendation and no price target appears in this appendix.