The five names sit inside the same TSE bucket but no longer trade on the same logic. Food & Life Companies (3563.T), Toridoll Holdings (3397.T), Zensho Holdings (7550.T), Saizeriya (7581.T) and Skylark Holdings (3197.T) span four economic archetypes — domestic vertical integrator, multi-banner mature, specialty international expander, mono-format value with a contested overseas leg. The forward P/E spread between them runs from 18.6x to 40.4x. The dispersion is itself the first read.

The post-COVID decade settled one thing for this bucket. The Japanese institutional bid no longer pays the defensive scarcity premium of the negative-rate era. The five-year P/E averages — Skylark 82.5x, Toridoll 86.2x, Food & Life 68.4x — are not benchmarks; they are residuals of a regime that ended. The new regime pays one thing durably: the legibility of consolidated margin trajectory toward FY+3/FY+5 mid-cycle, demonstrated quarter by quarter at the segment level. Everything else — vertical integration without overseas execution, mature FCF without a growth leg, international optionality without proven unit economics — is paid partially or transitorily.

What follows below sits in three layers. The economic engine and the cross-operator inputs describe what is shared across the five. The archetype map and the names section sort them. The mispriced variables and the structural watchlist track what each consensus is reading wrong and what would force a reframing.

The single most decisive structural variable across the bucket is vertical sourcing integration. Zensho's MMD platform delivers a consolidated gross margin of 54.2% in FY2025 against a typical mature foodservice 30 to 40% band — a 1,400 to 2,400 basis-point operating moat that is genuinely difficult to replicate at scale. Food & Life sources fish through a direct partnership network rather than owning the upstream, which is more capital-efficient but transfers cyclical risk to suppliers. Toridoll's Marugame Seimen moat sits in the store rather than upstream — the Menshokunin noodle craftsman, two to three years of training, demonstrably non-replicable internationally as the UK impairment proved. Saizeriya runs a partial vertical play through its Brisbane farm. Skylark's central procurement and industrial kitchen mutualisation is real but commoditised — it matches peers of similar scale without producing margin leadership.

Unit economics tell the second story. A mature Sushiro store overseas produces roughly 95m JPY of annual profit against 40m JPY for the same store domestically — a 2.4x inversion that mechanically pulls consolidated margin higher as the overseas mix shifts toward 35% by FY2027–FY2028. A mature Marugame Seimen domestic store ran a segment business profit margin of 17.8% in 1H FY March 2026 on 71.4bn JPY of half-year sales — the highest segment margin observed in the bucket. Sukiya pre-incident sat at 9.1% and is now in recovery at 3.1%. Saizeriya's Japan stores hold a 7 to 9% segment margin while the Asia leg posted operating profit declines of −82.7% in Beijing, −92.5% in Shanghai and −96% in Guangzhou in Q1 calendar 2025. Skylark's Gusto suburban anchor sits in a 4 to 6% band that has not expanded structurally over the cycle.

Cash conversion completes the picture: consolidated EBITDA-to-FCF ratios diverge by a factor of two — roughly 17% for Zensho in FY2025 against 54% for Food & Life, with Skylark higher again at mid-cycle. Most of the dispersion is the growth-capex cycle. Zensho is mid-build on the MMD international platform, Food & Life is scaling overseas, Skylark is in harvest. Mid-cycle conversion converges tighter than reported — somewhere in the 50 to 70% band — but the timing differential determines current headline FCF yield.

The thread that ties this together is that consolidated margin becomes an unreliable tool when segment dispersion inside an operator exceeds roughly 500 basis points. Toridoll is the cleanest case — Marugame domestic at 17.8% against an overseas leg in restructuring is a gap above 1,500 basis points, and the consolidated P/E of 30.5x is reading the average of a compounder and a workout. A correct read requires a sum-of-the-parts the market has not yet priced.

The first cross-operator input is the yen. The depreciation from roughly 108 USD/JPY in early 2022 to 152–160 across 2024 explains, by the factor decomposition embedded in the historical performance work, between 25 and 40% of the overseas-leg outperformance of the past five years. For Food & Life the contribution sits at the upper end; for Zensho's international platform somewhere in the middle; for Toridoll's overseas business at the lower end. The right modelling discipline takes 130–135 USD/JPY as mid-cycle base and runs a sensitivity around 110/130/150. At spot, headline overseas EBITDA is mechanically inflated by 15 to 25% relative to a normalised yen. The operators that have built a real overseas business will absorb the reversion; the ones that have not will see it taken out of the multiple.

The second input is the joint pressure of Shunto wages and the structural labour shortage in Japanese hospitality. The prefectural minimum-wage trajectory adds 3 to 5% per year to entry-level cost on a compounded basis, and foodservice labour participation continues to tighten. The operators most exposed are the multi-banner family-dining models with the lowest pricing power — Skylark in particular, where pass-through has run at a reconstructed 60 to 75% against more than 100% at Sukiya pre-incident. Automation is the partial answer: Sushiro's IC-chip, auto-waiter and Digiro layer is the strongest digital moat in the bucket; Marugame's seasonal item engineering and Sukiya's MMD throughput are operational rather than digital answers. The dispersion in pass-through capacity sorts the bucket more cleanly than any other macro input.

The third input runs through food inputs, inbound tourism and suburban demographics together. Wheat flour, dominant in Marugame's COGS, ran +25 to +35% cumulative through the post-Ukraine cycle and was absorbed by seasonal pricing without volume destruction — the most demonstrable pricing-power case in the bucket. Sukiya absorbed +25% cumulative on beef bowl input pre-incident. Inbound tourism, structurally elevated post-2023 in the Tokyo–Osaka corridor, has lifted urban Sushiro, Gusto and Sukiya station-side covers; the contribution is concentrated, not distributed. Both the input softening and the inbound tailwind are partly a yen story — the same reversion that takes out overseas EBITDA softens both. Underneath sits the structural suburban demographic drag of 0.5 to 1.0% of real traffic per year, hardest on Skylark, softer for the names with international scaling or a domestic compounder banner. There is no aggregate macro tailwind to lean on — outperformance has to come from somewhere specific.