Tsuruha Holdings3391.T
Tsuruha is the largest drugstore operator in Japan and, on the number that matters, the only one in its peer group that earns less on its capital than that capital costs. The Welcia merger that built the scale also built ¥448bn of goodwill — 51% of equity — that the return on capital does not service. The inherited reading was an optically cheap name: a 1.2x book multiple, a 7.2x forward EV/EBITDA in line with the sector's best compounders. Read against the certified data, both are traps. The book is 2.6x tangible ; the multiple is the price of quality applied to a business that destroys value. What is left to decide is not whether the name is expensive, but by how much — and that answer turns entirely on how the goodwill is treated.
On EV/EBITDA — the metric that adds back the ¥24.7bn of annual goodwill amortisation and makes the group comparable to its IFRS peers — a base 6.3x on forward EBITDA of ¥148bn puts enterprise value at ¥932bn and equity at ¥930bn, or ¥2,051 per share.
On normalised earnings — the mirror metric, carrying the full drag of goodwill amortisation, interest and minorities — earnings power of ~¥33bn at a value-destroyer's 15–17x reconstructs to ¥1,100–1,240 per share.
The ¥800-plus gap between the two readings is the goodwill question in one number: ¥448bn of it sits on the balance sheet, 51% of equity, amortised through operating profit for two decades. Whether it holds value, or its amortisation is the accounting record that it does not, is what the modelling cycle has to resolve.
Spot is ¥2,353. Both readings sit below it. The dispute is the magnitude, not the direction.
The question is not whether Tsuruha is a good business — the sector work settled that — but why the market holds the weakest dossier in its group at the middle of its own valuation corridor while it prices the four compounders at the floor. At 9.1x LTM EV/EBITDA, the 46th percentile of a decade of trading, Tsuruha sits where a quality operator sits. On the certified numbers it is the bucket's only value-destroyer. The dossier turns on whether that preserved multiple is a repricing lag that compresses as the Welcia dilution lands, or whether purchasing synergies restore the spread and justify the multiple after the fact.
The destruction is worth stating plainly. Return on invested capital, reconstructed ex-cash for FY2/2026, is 4.8% against a 6.01% cost of capital ; lease-adjusted for the ¥570bn of capitalised rent, the spread runs −1.6 to −3.1 points. Every other name in the bucket earns a positive spread of +1.9 to +6.4 points. The merger did not cause this — return on capital had already halved, from 12.4% in May 2020 to 7.1% in May 2024, while capex tripled — but it institutionalised it, adding ¥448bn of goodwill the return does not service.
The variant view lives in the double charge the market has not capitalised. Consensus prices the FY2/2027 guidance as given, treating the ~90bp margin dilution as a settled fact and the mid-corridor multiple as an equilibrium. What that reading misses is that the full dilution only appears from FY2/2027, that its recovery is conditional on a 90bp synergy capture four structural bottlenecks make improbable, and that the reported earnings the multiple rests on are themselves doped — a ¥10.58bn step-revaluation gain lifts reported net income ~30% above the ~¥33bn the business actually earns.
The optical cheapness is the trap, twice over. The 1.2x book multiple that screens as the sector's cheapest is 2.6x on tangible book, because 51% of equity is intangible. The 7.2x forward EV/EBITDA that reads like a discount sits on an EBITDA inflated in absolute yen — ¥65bn to ¥148bn — by a low-margin revenue base : a lower-quality earnings stream priced, correctly, at a lower multiple. The two most attractive metrics in the dossier are its two most dangerous.
The framing is avoidance, with a short prefiguration to confirm after the modelling cycle. Every valuation method converges on overvaluation — from −14% on EV/EBITDA to −54% on normalised earnings — so the direction is not in question, only the magnitude. The one caution is timing : under Aeon's 50.3% absolute majority no activist catalyst is possible, which makes the case well-founded but weakly triggered. Conviction is moderate-to-strong ; sizing is contained.
The decade reads as a company that grew without compounding, then bought scale it could not make pay. Across the autonomous years to May 2024, revenue compounded +8.7% a year and net income +1.5% — eight years of double-digit top-line growth for near-flat earnings, because the operating margin eroded structurally, from 5.94% to 4.59%, and return on capital halved. The merger did not reverse the trajectory ; it took it to industrial scale, doubling the capital base with goodwill while pulling the consolidated margin down toward 3.89% on guidance.
| Inflection | May 2016Net-cash compounder | May 2020Pre-COVID peak | May 2024ROC halved | FY2/2026Welcia · 3mo | FY2/2027Full dilution · g |
|---|---|---|---|---|---|
| Revenue (¥bn) | 527.5 | 841.0 | 1,027.5 | 1,450.6 | 2,555.0 |
| EBIT (¥bn) | 31.3 | 45.0 | 47.2 | 63.0 | 99.4 |
| EBIT margin | 5.94% | 5.35% | 4.59% | 4.35% | 3.89% |
| Return on capital | 13.3% | 12.4% | 7.1% | ~4.8% | n.c. |
| FCF (¥bn) | 42.2 | 26.5 | 20.0 | 59.7 | n.c. |
| Net debt (¥bn) | −60.6 | −45.5 | −8.6 | +6.4 | — |
| Net income (¥bn) | 19.3 | 27.9 | 21.7 | 42.7 | 41.5 |
Source: pack 14 July 2026 · Tanshin 31 May 2026. Pre-transition years close mid-May ; year-end moved to end-February from the period closed February 2025. EBIT = reported operating income. FY2/2026 net income is doped ~30% by a ¥10.58bn step-revaluation gain ; FY2/2027 guidance net income is −2.7% despite +76% revenue. Return on capital FY2/2026 reconstructed ex-cash (NOPAT ¥43.1bn / invested capital ¥902.1bn).
Three management decisions explain the shape. Return on capital was allowed to halve across 2020–2024, capex tripling from ¥7bn to ¥32bn for a falling marginal return, with no visible pivot toward distribution or store closure — growth for its own sake, where size outranks return. Welcia was then overpaid into ¥448bn of goodwill, 51% of equity, on a business earning a structurally lower margin (~2.83% against ~4.5%), which tipped the spread below the cost of capital. And the ¥78bn one-off buyback booked alongside the merger, judged "not probative" in the sector work, produced no re-rating while consuming cash as the balance sheet tightened — an opportunistic signal in place of a disciplined, quality-backed return policy.
The engine is a traffic model monetised by mix, not by price — and the mix does not earn the capital back. Food is 32.6% of merchandise revenue at 17.2% gross margin : not a margin business but a frequency generator, there to bring the customer into the store. The margin sits in the cross-sell — OTC at 40.7%, cosmetics at 33.0%, dispensing at ~36% — which together carry roughly 34% of gross profit on 27% of revenue. Tsuruha composes a basket ; it does not set a price.
The distinction between mix power and pricing power is the whole of it. The decade's gross-margin expansion, 28.3% to 30.6%, is entirely a mix shift toward dispensing and private label, with no autonomous price effect visible. Food is a competitive loss-leader ; dispensing margin is administered by the reimbursement schedule ; OTC and cosmetics face cross-industry competition. The scale the merger created is meant to convert into purchasing margin, but in a sector the company itself describes as one where competition "continues to intensify," the buying gain is more likely recycled into price to defend traffic than captured in margin.
The consolidated line hides two opposite populations. The legacy Tsuruha store is healthy — same-store sales are accelerating, +4.3% against +2.6% a year earlier, a two-year stack of +7.0% — while the Welcia contribution dilutes at a ~2.83% margin. Free cash flow of ¥59.7bn in FY2/2026 looks robust but is doubly flattered : a ¥20.2bn working-capital release, and the add-back of ¥24.7bn of non-cash goodwill amortisation. Normalised, and after rising interest, the figure is nearer ¥35–45bn. The destruction sits in the marginal capital, not in the operation ; a healthy legacy store cannot repair an acquisition that was overpaid.
This pillar carries the thesis because it is the destruction itself. Return on invested capital, lease-adjusted, runs −1.6 to −3.1 points below the 6.01% cost of capital — the only negative spread in the bucket, against +1.9 to +6.4 for the five peers. Cash conversion looks strong but is volatile across the decade (−50% to +130%) and currently flattered by working capital and the goodwill add-back. The theoretical scalability of scale is neutralised by the Welcia dilution. Moving this to 3.0 needs the spread back above zero — a ~90bp synergy capture that is not demonstrated.
The lowest pillar, and the one that turns a mediocre dossier into an avoid. Aeon holds 50.3% — an absolute majority — so the minorities (BlackRock 3.6%, JPMorgan 2.9%, Nomura 2.5%) hold no counterweight. Capital allocation serves Aeon's sector-consolidation agenda rather than the outside shareholder : the ¥78bn one-off buyback, timed with the merger and judged not probative, is the symptom, not a return policy. A dividend at 2.0% yield is the only alignment signal, and it is modest. Lifting this to 2.5 would need a formal minority-protection commitment — improbable under majority control.
Genuinely defensive and domestic. A demographically-supported dispensing base (16.4% of revenue) and legacy same-store sales at +4.3% keep the score off the floor. The frequency engine — food, 32.6% of revenue at 17.2% margin — is the most commoditised link, exposed to cross-industry competition and sector overcapacity.
The No.1 scale (¥2,555bn, 5,665 stores) is real but shallow. In an oversupplied sector it confers neither pricing power nor switching cost ; it is a size advantage, not a return advantage. It has none of Matsukiyo's data/brand moat or Sundrug's cost moat, and pays only if the purchasing synergy converts to margin — unproven.
Legacy operating execution is competent — same-store sales, gross margin — but it is eclipsed by the allocation record : return on capital halved before the merger, Welcia overpaid into ¥448bn of goodwill, and an opportunistic one-off buyback in place of a disciplined return.
A trap / illusion profile — the margin exists but does not pay for the capital. Below every compounder in the bucket (15.5–17.5 / 25) and the lowest grade in the group. Governance is the decisive pillar precisely because it is the lowest : it is what converts a weak economic case into one to avoid rather than merely to skip, and it caps any re-rating for as long as Aeon holds the majority.
Do the purchasing synergies recover the ~90bp of dilution, or is the merger permanent destruction ?
Does the mid-corridor multiple survive the normalisation of earnings power ?
The reported ¥42.7bn is doped ~30% by the step-revaluation gain ; on the normalised ~¥33bn the P/E is ~32x, above the historic average, and the guidance itself puts net income −2.7% and EPS −29%. The mid-corridor multiple rests on a denominator that is eroding. This is the divergence at the centre of the dossier : add the goodwill amortisation back (EV/EBITDA) and the name is 14% rich ; carry it (normalised earnings) and it is 47–54% rich.
Is the Aeon control discount priced, or still ignored ?
The 50.3% absolute majority subordinates the minority structurally. Treated as a neutral, ancient fact, it is a latent discount ; treated as already priced, it caps any re-rating durably. Under majority control the capital allocation — a dilutive merger, an opportunistic buyback — serves the consolidation agenda more than the outside return, and the minority has no lever to change it.
At ¥2,353 and 7.2x forward EV/EBITDA, the market prices Tsuruha as a compounder of its own bucket — exactly the Cosmos (7.4x) and Sundrug (6.6x) range — while it carries the only negative ROIC−WACC spread and a 1.5 / 5 governance score. It applies a quality multiple to a destruction dossier. The tell is the metric divergence : add the goodwill amortisation back and the name is 14% rich ; carry it and it is 47–54% rich. The 1.2x book that reads cheap is 2.6x tangible. A rigorous sum of the parts cannot be built — one reported segment, no operating profit by category, no legacy/Welcia split — which is itself why the goodwill question stays open, and why the case routes to the full modelling cycle rather than resolving here. The probability-weighted fair value (55 / 30 / 15) is ~¥2,026, −14% on the EV/EBITDA base and ¥1,100–1,240 on normalised earnings : the direction is common to every method, the magnitude method-dependent.
Synergies do not materialise — recycled into price to defend traffic in overcapacity — margin stalls or slips below 3.9%, and an exogenous shock lands : a −2 to −3% dispensing-schedule revision on the 16.4% officine, or a BoJ hike on the 2.2x-EBITDA debt. The market re-discriminates and the multiple compresses toward the corridor's absolute floor (5.3x). Reversible as a timing disappointment while the legacy core holds ; it turns into permanent loss only if the legacy same-store base collapses and the goodwill is impaired.
Tsuruha executes the FY2/2027 guidance without material synergy — EBIT margin near 3.9%, Welcia dilution booked — and the mid-corridor multiple compresses slowly toward the compounder range as the market re-discriminates quality. The EV→equity bridge on ¥148bn EBITDA at 6.3x lands at ¥2,051. No shock, no redemption ; the fair value simply sits below the spot.
The synergies land faster than planned (EBIT margin toward 4.3–4.5%), consolidated same-store sales confirm Welcia converging on the healthy legacy, and the spread crosses back above zero at 24 months. The market re-rates toward a stabilised compounder multiple (7.5x) — but no higher, because Aeon governance and the leverage cap it. The path needs both the operational lift and the allocation restraint, neither signalled today.
| KPI | Latest value | Status | What it tells us |
|---|---|---|---|
| Consolidated EBIT margin (FY2/2027) | 3.89% guidance | Cardinal | The synergy test. Above 4.3% across 2–3 quarters confirms capture and forces a revision away from the short ; at or below 3.9% confirms permanent dilution. |
| ROIC−WACC spread (lease-adj) | −1.6 to −3.1 pt | Cardinal | The only negative spread in the bucket, against +1.9 to +6.4 for peers. Back above 0 on >¥30bn of synergy at 24 months invalidates the short. |
| Net income vs guidance (FY2/2027) | ¥41.5bn g / ¥44.2bn cons. | Priced | Consensus prices net income above guidance on a doped base ; a print at or below guidance without synergy triggers the multiple compression. |
| Consolidated SSS ex-M&A | legacy +4.3% | Watch | The legacy core is healthy and accelerating. Whether Welcia converges on it or dilutes it, once the ex-M&A line is published, separates the floor from the compression. |
| Aeon control / related-party | 50.3% | Trigger | Absolute majority ; minorities hold no counterweight. A distribution cut to fund Aeon-piloted M&A confirms subordination and deepens the destruction. |
| EBIT interest cover | 25x FY2/2026 | Watch | From 132x in 2016. Debt at 2.2x EBITDA ; the variable/fixed split is undisclosed and BoJ-sensitive. The net-cash immunity is gone. |
| EV/EBITDA (LTM / forward) | 9.1x / 7.2x | Reference | Compounders at the 6.6–7.4x floor ; corridor 5.0–15.7x. Forward 7.2x is a low multiple on an inflated EBITDA, not a discount. Compression toward 5.3x is the Bear. |
| Goodwill amortisation | ~¥24.7bn/yr | Reference | J-GAAP, inside operating profit, ~20-year schedule ; ¥454.6bn → ¥448.4bn over Q1 cross-validates the run-rate. The EBIT/EBITA reading it drives moves fair value by tours. |
The case turns neutral if the scale finally pays. A consolidated EBIT margin held above 4.3% across 2–3 FY2/2027 quarters, or a ROIC−WACC spread back above zero at 24 months on more than ¥30bn of captured synergy, would show the merger recovering the dilution it created and force a revision away from the short. Either is observable on the published FY2/2027 calendar ; neither is signalled today.
The case hardens if the narrow lever fails. A margin confirmed at or below 3.9% without synergy across the FY2/2027 prints validates the mid-corridor compression toward ¥2,051 and below. The permanent-loss path is narrower but modelable : if the purchasing synergies fail and the legacy same-store base collapses under overcapacity together, a goodwill impairment on the ¥448bn becomes a central scenario and the floor falls toward ¥1,100–1,240.
The allocation risk is the one to watch most, because the control structure makes it likely. A further Aeon-piloted acquisition funded on the stretched balance sheet — a repeat of the Welcia overpayment — would burn the free cash flow at the minority's expense and deepen the destruction irreversibly. A −2 to −3% dispensing-schedule revision in 2026 would do the same from the regulatory side. Neither is signalled ; both are un-priced.
The information provided on this website is for informational and educational purposes only and should not be construed as financial, investment, legal, or tax advice. All content reflects the personal opinions, interpretations, and analyses of the author at the time of writing and is subject to change without notice. Nothing contained herein constitutes, or should be interpreted as, a recommendation, solicitation, or offer to buy or sell any securities, financial instruments, or other investment products. The author is not a licensed financial advisor, broker, or investment professional. Any references to specific assets, markets, or strategies are illustrative in nature and do not constitute personalized investment advice. Investing in financial markets involves risk, including the potential loss of capital. Past performance is not indicative of future results. Readers are solely responsible for their own investment decisions and should conduct their own independent research and due diligence before making any financial commitments. You are strongly encouraged to consult with a qualified financial advisor, legal professional, or other relevant specialist before making any investment or financial decisions. By accessing and using this blog, you agree that the author shall not be held liable for any direct or indirect losses, damages, or consequences arising from the use of, or reliance on, the information presented herein. All content is provided "as is" without any warranties of completeness, accuracy, or reliability.