Seven & i Holdings3382.T

The interesting thing about Seven & i is not that the consolidated numbers are bad — everyone can see that — but what the bad numbers hide. The group earns a 4.2% return on capital, below any sensible cost of capital, and the market has filed it at 1.33x book, the ninth percentile of its own ten-year range. Underneath, the picture is two businesses that have almost nothing to do with each other. One is a domestic convenience franchise that captures a royalty at a 24% operating margin and has held that margin across three regimes. The other is the largest convenience operator in North America, throwing off ¥332bn of cash operating profit that the reported line buries under Speedway goodwill. The question the dossier turns on is whether the self-help now under way — carve-outs delivered, a ¥2tn buyback running, IFRS arriving, a 7-Eleven IPO promised — forces the market to price those parts, or is financial engineering dressing a sub-WACC consolidated whole.

The distinction matters because the consolidated line is close to meaningless here. The reported operating margin of 4.8% is the weighted average of a franchise rent at 24% and a fuel-heavy US operation at 2.6% — two economies that do not belong in the same ratio. The reported return on capital of 4.2% is the same average, carrying ¥2,110bn of Speedway goodwill. Strip the ¥138bn of goodwill amortisation back out and consolidated EBITA is ¥561bn, not the ¥423bn the market reads ; the North American pole alone moves from ¥222bn reported to ¥332bn before goodwill. IFRS, arriving for the year ending February 2028, makes that add-back accounting-real. The market follows the reported line ; the cash line is about a third higher.

The second thing the price does not hold is the balance sheet's dormant value. Net debt of ¥3,355bn looks heavy, but ¥1,564bn of it is capitalised leases ; the hard financial net debt is ¥1,791bn. Against that sit a listed 39.9% stake in Seven Bank whose market value is directly observable, a 35.07% interest in the carved-out York vehicle, and a capital-return signal — ¥600bn of buybacks on a ¥2tn programme, the dividend lifted from ¥50 to ¥60 — that no other name in its bucket has delivered. Seven & i is the one convenience-and-discount name where founder control does not bridle the capital decision : the Ito management buyout failed, and activist pressure installed an external chief executive and a majority-independent board.

What that leaves is a sum of the parts that reconstructs about 14% above the market cap in the base case, floored by the domestic rent and the Seven Bank stake, with a wider upside the price gives no weight. That upside is a single lever : an IPO of the North American business that would crystallise the jewel at a convenience pure-play multiple. The lever is real and it is dated — but it has no calendar and no implied valuation. Couche-Tard's approach, which used to provide a take-out floor, withdrew in February 2026 ; the endogenous crystallisation replaced it, and it is more robust for depending on no external buyer.

The position framing is patient observation, not ownership at this level. The weighted asymmetry is +13.4%, below the threshold that would earn a long, and the convexity is hostage to a catalyst with no date. Conviction is moderate. The thing worth watching is the US same-store print and any firm IPO calendar ; both are observable on the published schedule.

The cleanest way to read the last decade is as a single acquisition. Seven & i entered it as a domestic retail conglomerate at a 7–8% operating margin, a net-cash balance sheet and a mid-single-digit return on capital. Then it bought Speedway at the top of the fuel cycle — roughly $21bn, closed May 2021 — doubling revenue and turning the group into the largest convenience operator in North America. The operating margin fell from 8.0% to 4.8%, return on capital slid below the cost of capital, and a ¥339bn net-cash position became ¥3,355bn of net debt. The self-help since 2025 is the attempt to unwind that decision by structure rather than operation. The shape matters because it makes any valuation anchored on the ten-year average multiple meaningless : that average spans a business that changed identity halfway through.

Three decisions explain the shape. Speedway was bought at the peak of the fuel cycle and financed with debt — the goodwill it added has depressed reported operating profit ever since. A value-destroying conglomerate, the GMS chain and the department stores, was kept a decade too long and only carved out in 2025 under activist pressure, at a cumulative special-loss cost of ¥221bn and ¥86bn across the last two years. And the capital structure was mis-set at both ends : idle net cash until 2021, then abrupt over-leverage for Speedway, with the ¥2tn buyback arriving a decade after TSE governance reform called for it. The discipline since 2025 — carve-outs, buybacks, IFRS — is corrective and externally imposed ; it is not yet evidence of a management that allocates well once the activist pressure is off.

The engine only makes sense once you split the two convenience businesses, because they are opposite economies wearing one badge. The domestic franchise captures a royalty on roughly 21,600 franchised stores at a 24.3% operating margin — ¥222.5bn on ¥914.6bn of franchisor revenue — with the franchisee carrying the store-level capital. The North American operation runs roughly 13,000 company stores at a 2.6% reported margin, on a revenue base swollen by fuel and store-level consolidation, but it is the larger absolute profit pool before goodwill : ¥332.4bn of EBITA against the domestic ¥222.5bn. A single 4.8% consolidated margin is the average of an asset-light rent and a capital-heavy scale operator, which is why one consolidated multiple is the wrong tool.

The most important thing to understand is where the pricing power actually lives, because the word flatters the US line. Domestic pricing is real : the private-label fresh-food programme — SEVEN CAFÉ, onigiri, bento — lifts margin structurally, and the current gross-margin squeeze is a rice-cost pass-through lag rather than an erosion of the franchise. US pricing is apparent. The average basket has risen +1.3% to +5.3% across the last two years, but traffic has fallen −1.8% to −8.0%, and merchandise same-store has been negative in seven of the last eight quarters. A rising basket against falling traffic is trade-up under a shrinking footfall, not pricing power — and the quality of that revenue is lower than the basket figure suggests.

The cost that drives most of the margin volatility is US operating cost — labour and rent — because the US carries ~82% of revenue and ~59% of EBITA on a thin margin, so any fixed-cost inflation is amplified. With traffic falling and SG&A rising, the largest profit pole is running negative operating leverage : the bottleneck of the whole model. The captive fuel integration through Speedway partly hedges the merchandise cost base, but it also exposes the US pole directly to the crude-retail spread and, over a five-to-ten-year horizon, to the EV transition — a two-edged feature, not a clear advantage that a Lawson or a FamilyMart does not carry.

The cash conversion looks solid on the surface — free cash flow of ¥333bn against reported EBIT of ¥423bn — but the number needs reading. The reported EBIT is depressed by ¥138bn of non-cash goodwill, so cash operating profit is higher than it looks ; against that, the year's free cash flow includes disposal proceeds and heavy US capital expenditure. The capital-return signal sits on top of this : ¥714bn returned — ¥600bn of buybacks, ¥114bn of dividends — against ¥333bn of organic free cash flow, a coverage of only 0.47x, with the balance funded by disposals and debt. Set against the leverage is a listed Seven Bank stake, a carved-out York interest and net-of-lease debt of ¥1,791bn rather than the ¥3,355bn headline. That dormant, listed capital is the real option in the name, and the price gives it almost no weight.