Financials

Financials

Isetan Mitsukoshi is no longer the company its income statement remembers. Operating margin has gone from -2.6% in FY2021 to 14.7% in FY2026 — a 17-point swing inside five years — while gross sales have climbed back to a 12-year high. Net income, free cash flow and ROE all set fresh decade records in the year just reported. The balance sheet is under-levered (equity ratio 50.8%), capital return has stepped up to nearly half of CFO, and the stock trades at ~17x earnings — a clear premium to listed Japanese department-store peers.

The investment question is no longer "can this business earn anything." It is whether a P/E of 17x and a ~¥1.3 trillion market cap are pricing a sustainable luxury-and-inbound margin (bull: structurally re-rated) or the peak of an FX/tourism cycle (bear: cyclical earnings at the top). Every judgement below comes back to that question.

What the headline numbers say

Two terms worth defining. Gross sales is the total ticket value of merchandise sold across IMHDS stores — the legacy top-line investors followed for decades. Revenue (net) is the figure required under Japan's 2022 revenue-recognition standard, which forces department stores to book consignment merchandise on a commission basis. The standard change is why reported revenue collapsed from ¥1,196B in FY2019 to ¥418B in FY2022 — same business, smaller line. Use gross sales for cycle comparisons; use net revenue when computing margin on currently reported figures.

The decade in one chart

The top-line decade is the boring chart: gross sales basically round-trip from ¥1.29T (FY2016) to ¥1.30T (FY2026), with a brutal COVID crater in between (-32% peak-to-trough). The interesting chart is the second. In FY2016 IMHDS turned each ¥100 of gross sales into ¥2.60 of operating profit. In FY2026 it earned ¥6.20 on the same basis — and almost ¥15 on every ¥100 of net revenue. That is not a recovery; that is a different business.

The drivers are well-rehearsed: inbound-tourism luxury wave magnified by a weak yen, deliberate mix skew toward higher-ticket goods, CRM-driven "individual customer business" lifting spend-per-visit, and operational fixed-cost discipline post-COVID. The risk: two of those four (yen, tourism) are policy-sensitive.

Year-wise statements - the standard view

Three things jump off the table:

1. Pre-COVID was mediocre, not great. FY2016-FY2020 averaged op margin of 2.0% and ROE of 1.5% — closer to a utility than a luxury retailer. Worth remembering before extrapolating today's returns.
2. FY2021 nearly broke the equity base. A ¥41B loss and -7.9% ROE compressed equity from ¥588B to ¥508B in two years.
3. The latest two years are different in kind, not degree. FY2025-26 produced cumulative net income of ¥128.9B and cumulative FCF of ¥175.9B — roughly equal to total net income earned across the prior six years combined.

Earnings quality - does the cash arrive?

CFO has run above net income in 10 of the last 11 years, including every year of margin expansion. CFO/NI conversion in FY2024-FY2026 averaged 1.39x — even excluding FY26's unusual positive investing cash flow (asset disposal lifted reported FCF to ¥112B), underlying CFO of ¥91B against ¥76B of net income is excellent for a working-capital-light retailer that books most COGS on consignment.

The year to flag is FY2021: CFO collapsed to ¥1.2B against a ¥41B GAAP loss, indicating the loss was real (not non-cash) and working capital provided no buffer. This is the data point the bear case is built on — when this business turns down, it turns down hard.

Balance sheet - flexibility, not constraint

Equity ratio is the cleanest single read: 41.9% at the FY2021 trough, 50.8% at FY2026 year-end — the strongest of the major listed Japanese department-store operators. Achieved while paying ~¥75B in dividends and buying back over ¥100B across FY2024-FY2026. Total assets are essentially flat at ~¥1.22T over a decade — growing returns on a stable asset base, not a balance sheet inflating through acquisition.

A few items the table cannot show:

• Real-estate optionality is real. The unusual +¥21.6B investing cash inflow in FY2026 confirms that disposals are being used to recycle property capital.
• No solvency stress markers. Equity ratio above 50%, total liabilities (¥597.8B) below shareholders' equity (¥620.2B), and interest cover (recurring profit ¥86.6B) all signal a balance sheet that is a tool, not a constraint.
• The Credit, Finance & Friendship Club segment carries the only meaningful interest-bearing borrowing (funding credit-card receivables). Structurally low-risk Japanese consumer credit, but the reason gross liabilities sit close to ¥600B despite the asset-light core.

Returns on capital - convergence, not compounding (yet)

For context: a globally premium softlines retailer compounds at 20-30% ROE; a typical Japanese GMS earns mid-single digits. IMHDS at 12.5% ROE / 6.6% ROA in FY2026 sits between — and the new Medium-Term Plan explicitly cites ROIC discipline. The pre-COVID decade averaged 1.7% ROE, so anything above 10% is genuinely new ground. Two structural reasons it can stay there:

• Higher net margin in the new revenue-recognition world. When the top line is a commission, the same yen of operating profit drops more cleanly through.
• Asset turn has not deteriorated. Total assets fell 5% over the decade while gross sales returned to a 12-year high.

The bear case: FY2026 embeds a FY2023-FY2026 luxury/tourism tailwind. The pre-COVID 1-2% ROE base case pressure-tests poorly.

Capital allocation - the recent step-change

Disclosed integrated-report figures only run through FY2024, but cash-flow statements show FY2025 financing outflows reached ¥94.9B and FY2026 reached ¥76.9B — far above the ¥40-50B FY2024 blended dividend-plus-buyback footprint. Dividend has scaled with earnings: ¥56 per share in FY2026 versus ¥12 in FY2019. Total return ratio ran 30-50% from FY2023; FY2025-FY2026 evidently pushed it materially higher.

Two practical reads:

1. Management is treating excess cash as shareholders' capital, not strategic reserve. A recent shift; should structurally support the multiple if it persists.
2. Buybacks at a higher multiple are less accretive. With the stock at ¥3,722 versus the ¥2,500-2,700 mid-2025 zone, every yen of buyback delivers ~40% less EPS lift than it would have eighteen months ago. Expect the form of capital return to tilt back toward dividends and disposals if the share price holds at these levels.

Quarterly trajectory - the inflection is still on

Q4 FY2026 net income of ¥24.8B beat consensus by nearly 80%, and the most recent four quarters delivered bottom-line beats even while revenue printed below estimates. That divergence — operating efficiency catching the Street off guard, top-line softer than expected — is the single best near-term proxy for whether the FY2026 op margin can stick.

The deceleration to watch: Q1-Q2 FY2026 gross sales softened sequentially (¥301B → ¥295B); foot traffic and inbound spend will be tested across Q1-Q2 FY2027 (released summer 2026).

Peer comparison - paying up, on purpose

IMHDS sits at the top of the listed peer set on three of four dimensions: largest gross sales (¥1.3T), highest market cap by 2x, strongest balance sheet (equity ratio 50.8%). On op margin (14.7%) it is below Marui's 18.1% — but Marui is structurally a finance company stitched onto a retailer, not a comparable. On ROE (12.5%) it sits behind Marui (11.6%) and H2O (9.8%) only because equity has compounded faster than earnings.

The valuation gap matters. At 17.3x P/E, IMHDS trades at a discount to J. Front (22.6x) and modestly above Marui (18.2x), with H2O (10.6x) and loss-making Takashimaya at the cheap end. On gross-sales basis, IMHDS prints ~1.0x market-cap-to-gross-sales versus J. Front's ~0.5x and H2O's ~0.5x — a clear, intentional premium. The market is paying for best-in-class equity ratio and highest absolute profit scale. If margin normalises to 10% (FY2024 level), the implied P/E on normalised earnings rises closer to 25x — at which point peers look cheap.

Valuation in context - history says cheap, peers say full

On trailing earnings (¥215 EPS, ¥3,722 price) the stock prints 17x — modestly below the Japanese multiline-retail industry average of ~18x and below J. Front's 23x. Simply Wall St's Aug 2025 model flagged it "undervalued" at 15.6x. On consensus analyst targets the picture is reversed: average 12-month target is ¥2,984 (-12% downside), dragged by CLSA's Sell at ¥2,100 and offset only modestly by JPMorgan's Buy at ¥3,200. Morgan Stanley downgraded to Equalweight, cutting target sharply to ¥2,200.

The simple multiples say cheap; the analysts who model the cycle say expensive. That gap lives entirely in whether FY2026's 14.7% op margin is the new normal or the cycle peak.

What I would do with this page

The page confirms two things and contradicts one. It confirms (1) the operating model has structurally changed, with margin, ROE and FCF at simultaneous decade highs, and (2) the balance sheet is a competitive weapon, not a constraint. It contradicts the simple "P/E below industry → undervalued" narrative: peer P/Es should be compared at peer margins, and peer margins are notably worse, so the multiple is doing real work.

The first financial metric to watch is the operating margin on net revenue in Q1-Q2 FY2027 (released August and November 2026). FY2026 printed 14.7%, FY2025 13.7%, FY2024 10.1%. Annualised above 13% over the next two quarters supports the new-business-model thesis and the 17x. A slip toward 10% on softer inbound spend or yen normalisation would make the consensus ¥2,984 target look fair. Everything else — buybacks, balance sheet, capital allocation — is a function of that one line.