ABC-Mart, Inc.2670.T
Strip the cash out of the balance sheet and a different company appears. The reported 12.1% return on equity reads like a tired retailer ; the operating business underneath earns roughly 24% on capital, on a 50.7% gross margin that held through COVID and a ¥160 yen. The gap between the two is ¥213.9bn of dormant net cash — 31% of the market capitalisation — sitting under 62.4% family control. The dislocation is real and the operating business is cheap. What governs the upside is not whether the cash exists, but whether it ever moves : the family has announced two buybacks in six years and executed neither.
The Japan retail business, at ¥56.4bn of normalised segment operating profit on the 10x multiple its 20.7% margin and verified oligopoly pricing earn, is worth roughly ¥564bn — about 94% of the operating value.
Korea (¥4.7bn at 5x, deleveraging) and the rest of overseas (¥2.3bn, Danner wholesale plus Taiwan) add ~¥38bn together. The operating sum is ¥601bn ; at the spot enterprise value of ¥468bn, the market pays 7.4x operating profit for it.
Net cash of ¥213.9bn, credited at 75% for the dormancy the family imposes on it, adds ¥160bn and brings equity to ¥759bn in the base case.
The market capitalises ABC-Mart at ¥681.7bn. The operating business is cheap and the cash is discounted below even a 25% haircut. What the price does not hold is any value for the cash actually being returned.
The number that defines ABC-Mart is the one that is wrong. A 12.1% return on equity files the company as a mature, middling retailer, and the market has filed it there — the share is roughly flat over ten years while the TOPIX has roughly tripled. But that 12.1% is not what the business earns. It is what the business earns after ¥213.9bn of idle net cash — 53% of equity — is dropped into the denominator. Take the cash out and the operating engine returns about 24% on capital. The reported figure is half of the real one, and the missing half is sitting in a deposit account.
That gap is the whole case, so it is worth being precise about what created it. Over the decade operating profit rose from ¥39.7bn to ¥63.3bn and earnings per share from ¥107.9 to ¥187.2 — +60% and +73%, with no financial engineering behind it. Buybacks over those ten years were zero and the share count is flat. This is the mirror image of the usual Japanese compounder that flatters a thin operating record with accretive repurchases : ABC-Mart earned real per-share growth and then buried it under retained cash. Net cash roughly quadrupled to ¥214bn and has sat above half of the balance sheet for nine straight years, the payout pinned near 40%. The company has produced the value ; it has declined to hand it over.
So the question is not whether the business is good — it is — but whether the discount that quality trades at is corrigible or permanent : a governance false-negative the TSE efficiency push can force open, at which point the return on equity re-rates and the multiple follows, or a value trap a founding family with no internal reason to disgorge lets justify itself indefinitely. The two readings sit on the same operating facts and resolve in opposite directions.
The cellular work pushes the honest answer toward the harder side. The family controls 62.4% of the votes, so any return of capital needs its consent — and the base rate is not encouraging. Two buyback programmes have been announced, ¥5.0bn in 2020 and ¥7.5bn in 2022, and both were executed at zero ; the 2022 authorisation lapsed in six weeks without a single share bought. There is no activist on the register to force the issue, and at a 1.71x price-to-book the stock sits above the threshold that would put it on the TSE's named-and-shamed list. The most direct catalyst, in other words, does not apply, and the agent who would otherwise apply it is absent.
What that leaves is a long that earns its keep from the floor. The operating business is cheap at 7.4x ex-cash, the net cash is ¥864 a share — close to a third of the price — and the bear case is a timing disappointment of about 15%, reversible, with the domestic oligopoly rent intact underneath. The weighted fair value is ¥2,996, modestly above the ¥2,753 spot ; the asymmetry is 2.93x because the bull case is large even at a low probability. Conviction is moderate, and it rests on the asymmetry and the floor — the restitution that would deliver the bull case is a free option on a null base rate, not a dated event. The first observable test is the Q1 print on 8 July 2026.
The decade reads in three regimes, and the useful thing is that they isolate what is structural from what is cyclical. Through to FY2020 the company was a steady domestic compounder — revenue grinding higher, an operating margin in the high teens, net cash already doubling in the background as a comfortable cushion. COVID then tested the model in FY2021 : revenue fell −19% and the operating margin compressed to 8.9%, but it stayed positive, where the apparel names in the bucket fell into loss. Replacement demand for shoes absorbs a discretionary shock. From FY2023 the inbound-tourism super-cycle drove revenue from ¥290bn to ¥379bn and operating profit to a record ¥63.3bn — and the share de-rated through all of it.
| Inflection | FY 2017Pre-COVID | FY 2020Peak margin | FY 2021COVID trough | FY 2024Inbound boom | FY 2026Record |
|---|---|---|---|---|---|
| Revenue (¥bn) | 239.0 | 272.4 | 220.3 | 344.2 | 378.6 |
| EBIT (¥bn) | 41.9 | 43.4 | 19.5 | 55.7 | 63.3 |
| EBIT margin | 17.5% | 15.9% | 8.9% | 16.2% | 16.7% |
| Gross margin | 53.9% | 52.6% | 50.1% | 51.0% | 50.7% |
| Return on equity | 13.4% | 11.2% | 7.0% | 12.3% | 12.1% |
| Net cash (¥bn) | 115.6 | 147.1 | 145.4 | 175.0 | 213.9 |
| Net Income (¥bn) | 28.4 | 29.7 | 19.2 | 40.0 | 46.3 |
| EPS (¥) | 114.6 | 120.0 | 77.6 | 161.6 | 187.2 |
Source: Data pack 10 June 2026, closing-month convention (year ended February ; 1:3 split August 2023 retro-adjusted). EBIT = reported operating income ≈ managerial AOP. Gross margin holds in a 50.1–53.9% band across COVID and a ¥160 yen ; the swing in EBIT margin is SG&A leverage on volume, not a break in gross margin. FY2021 net income of ¥19.2bn is partly non-operational — operating profit was ¥19.5bn, with ~¥7.8bn of one-off COVID subsidies below the line ; FY2022 carries a ¥3.9bn impairment.
Three allocation decisions explain the flat share. The cardinal one, still running, is the thesaurisation itself : a payout frozen near 40% and zero buybacks let net cash quadruple to ¥214bn, which on the company's own sensitivity diluted the return on equity by roughly four points. The second is the overseas retail build at sub-domestic returns — a Korean and Asian network at ~6.6% margin, a quarter of the per-store profit of Japan, now contracting. The third is the failure to convert the inbound windfall into capital return : through a +31% revenue boom the cash grew by ~¥40bn in FY2026 alone, exactly when the TSE pressure and the P/B discount called for an act, and the answer was to retain more.
The engine only resolves once the consolidated 16.7% margin is pulled apart by geography, because the two pieces are not the same business. Japan earns 20.7% on ~72% of revenue and ~89% of segment profit. The total overseas line earns 6.4%, with Korea inside it at 7.3% and falling. A single group number is a weighted average of an oligopoly rent and a sub-scale satellite — which is why a consolidated multiple is the wrong instrument and the sum of the parts is the right one. At the store level the spread is stark : a Japanese store turns ~¥51mn of profit, a Korean store ~¥13mn, so every yen of capital reallocated abroad moves the blended return down.
Where the value is actually made is the conversion, not the price. Gross margin — about 50.7%, and roughly what any footwear retailer earns — is not the edge. The edge is an SG&A ratio near 34%, against ~47% for a high-service apparel format, which is what the scale and self-service of a dense store network buys. That is also why the margin is a two-edged variable : it is operating leverage, and it runs both ways. Korea proved the downside cellularly in FY2026, shedding ¥1.7bn of operating profit on only ¥3.3bn of lost revenue — an incremental fall-through above 50%, the signature of semi-fixed costs deleveraging when volume retreats. The gross margin held ; the SG&A line did the damage.
The cash conversion is structurally sound and was conjuncturally poor in the last print : free cash flow normalises near ¥42bn on a low-capex, self-service model, but FY2026 fell to ¥28.2bn on a ¥14bn inventory build over +1.7% revenue, most plausibly pre-positioning. The pathology is downstream of generation. The free cash flow yield is ~6.2%, ~9.0% ex-cash ; the shareholder yield is 2.9%, dividend only. That ~3.3-point gap is the governance defect, quantified : roughly half the cash the business throws off is neither returned nor productively reinvested — it accumulates dormant.
This pillar carries the thesis because it is where the locked value lives. Three of its four parts are excellent : return on capital ex-cash near 24%, an oligopoly gross margin of 50.7% held through COVID and a ¥160 yen, and a low-capex, cash-generative model. The fourth — capital allocation — drags the score down : net cash quadrupled with no productive use, reported return on equity diluted from ~14.6% to 12.1% by the cash alone, FY2026 conversion at ~45% on the inventory build. Genuine quality, held back by one thing the income statement does not show.
This is the second cardinal and the lowest score, because it is the only pillar whose correction would re-rate the whole dossier and the only one whose entrenchment would perpetuate the discount regardless of operating quality. Zero buybacks in a decade, a payout pinned near 40%, net cash above half of equity for nine years. The structural cause is that the founding family holds 62.4% of the votes and has no internal reason to disgorge. Two announced buyback programmes were executed at zero. Correction depends on external pressure, which is what makes the thesis conditional rather than owned outright. The dispersion between this pillar and the operating ones is the thesis.
A defensive replacement staple — beta 0.29, operating profit positive through the COVID trough — overlaid with a cyclical, FX-dependent inbound stream that is not disaggregated, plus a Korean line in contraction (−4.5%).
A real domestic distribution oligopoly — network density, buying power, pricing proven under a ¥160 yen — but a moat of channel, not of intellectual property. The hero brands belong to Nike and adidas, leaving it exposed to their direct-to-consumer shift.
Domestic execution is excellent — margin lifted 19.2% to 20.7% in two years. The allocation judgement is the weak point : doubling down on a deleveraging Korea through the FOLDER acquisition, exactly as the organic Korean line declines.
A good operating business carrying one concentrated, corrigible governance defect. Without the allocation drag this scores 19–20/25 — a defensive compounder. The score sits below a quality compounder such as Food & Life and well above a value trap. The spread from 2.0 to 3.5 is the whole point : it measures the distance between the operational quality, which is high, and the translation into shareholder value, which is not.
A corrigible governance false-negative, or a value trap of dormant cash ?
Inbound : a structural lift, or a borrowed peak ?
Domestic growth blends a replacement staple — structural, low-single-digit — with an inbound overlay that the weak yen carries and that the company does not isolate. Strip the inbound and underlying growth is probably ~2–3%. A yen normalising toward ¥130 removes the tailwind and reveals whether the staple base stands on its own. The −0.8% beta and the operating resilience through COVID say the base is defensive ; the unanswered question is how much of the recent print it actually accounts for.
Korea / FOLDER : accretive expansion, or destructive cross-subsidy ?
Korea's margin fell from 9.1% to 7.3% on −8.1% revenue — a near-total deleverage — and management is adding to it through FOLDER, a Korean select-shop and New Balance distributor, exactly as the organic line declines. The question is whether that is a scale play or domestic high-return capital subsidising a market that destroys it. In value terms Korea is small — ~¥24bn of enterprise value against the Domestic ¥564bn — so it is the allocation signal that matters more than the line itself.
At ¥2,753 the market pays 7.4x operating profit for the business ex-cash, 14.7x on reported earnings, 1.71x book. Three pessimisms sit stacked in that price : restitution at zero in perpetuity, the cash valued below par with no premium for the option ; a cyclical inbound peak, a P/E that de-rated from ~18–20x to 14.7x during a revenue boom on the read that the record earnings are tourist-inflated ; and a return on equity capped near 12% indefinitely. The method follows the structure — a sum of the parts, because the consolidated P/E values an artificially capped denominator and the bare P/B ratifies the thesaurisation as fate. The Domestic pole carries 94% of operating value at 10x, anchored on Nitori's ~8.2x EV/EBITDA plus a quality premium, and the cash is credited by scenario, never at par.
The yen normalises and the inbound overlay refluxes, compressing domestic same-store sales toward the staple base alone ; Korea's deleverage continues ; and the thesaurisation entrenches under TSE pressure, so the market concludes the cash is structurally trapped and discounts it harder (dormancy to 35%). The floor holds at ¥2,344 because the net cash (¥864/share) and the defensive domestic earnings power underpin it. This is a timing disappointment, reversible — not a permanent impairment.
The domestic plan executes without surprise — margin held near 20.7%, mid-single-digit organic growth with inbound decelerating but the staple stable — and the cash keeps accumulating, no restitution and no destruction. The cellular sum of the parts delivers ¥3,065 on ~¥63bn of normalised operating profit, crediting the cash at 75%. A consolidated re-rating may or may not come ; the fair value does not need it. In this case the net cash compounds from ¥214bn toward ¥306bn over five years — the value trap, drawn cellularly.
The capital event fires : under TSE pressure the family announces a substantial buyback or a capital-structure target, the cash is credited at par, the return on equity re-rates mechanically and the multiple follows. In parallel the inbound persists on a durably weak yen and FOLDER stabilises Korea. The path needs the allocation decision above all, and on a 62.4%-controlled register with two reneged programmes behind it, that decision is not signalled today — which is why the probability is 15% and the case is a free option rather than a base.
| KPI | Latest value | Status | What it tells us |
|---|---|---|---|
| Capital-return execution | 0 of 2 programmes | Cardinal | The only event that unlocks the bull. Two buybacks announced (¥5.0bn 2020, ¥7.5bn 2022) and executed at zero ; the base rate is nil. A third announced-then-reneged programme plus a rising family stake confirms the value trap and forces a Neutral re-underwrite. |
| Korea segment operating margin | 7.3% FY2026 | Cardinal | The bear trigger. Down from 9.1% on −8.1% revenue — near-total deleverage. Sustained below 6.5% pulls the Korea pole from 5x to 4x and confirms the cross-subsidy. Q1 reads ~¥4.73bn, −26% YoY. |
| Domestic same-store sales ex-inbound | Not disclosed | Watch | Separates the structural staple from the borrowed inbound peak. Below zero across two prints would expose a stagnant base and confirm the bear's demand leg. |
| Family-controlled stake | 62.4% | Holding | EM Planning 49.89% + Miki Masahiro 9.54% + Miki Michiko 3.00%. A supermajority that vetoes any restitution. No activist on the register (Orbis 2.17%, patient value). |
| P/B ratio | 1.71x | Reference | Above the TSE "PBR<1" threshold, so the most direct catalyst does not apply. Below 1.40x approaches the de-rating trough and compresses the Domestic multiple. |
| USD/JPY | ¥160.52 | Watch | Near a ten-year high. A weak yen is net positive (inbound dominates the COGS pass-through) ; appreciation toward ¥130 thins inbound demand even as it relieves the cost of goods. |
| Net cash accumulation | ¥213.9bn FY2026 | Trigger | ¥864/share, 31% of cap, including ~¥20.1bn of long-term securities. Accumulation beyond ¥250bn with no return is the value trap deepening ; ¥197.6bn is core distributable cash. |
| Free cash flow conversion | ~45% FY2026 | Watch | Depressed by a ¥14bn inventory build on +1.7% revenue — most likely pre-positioning. Renormalisation above ~70% confirms it was conjunctural ; persistence flags a structural drag. |
The case turns decisively positive if the cash moves. A substantial buyback, a published capital-structure target, or a payout lifted toward 60% at the FY2027 result or AGM would convert the dormant balance sheet from a discount into a re-rating, double the return on equity mechanically and open the bull path to ¥3,953. It is observable on a dated calendar ; it is simply not signalled by a family that has reneged twice.
The case turns negative if the narrow engine stalls. A Korean operating margin held below 6.5% together with domestic same-store sales ex-inbound below zero across two consecutive prints resets fair value toward ¥2,344. A domestic gross margin slipping below ~49% over two quarters would be more serious — it would signal that the oligopoly pass-through has hit a ceiling under the weak yen, and convert a margin we treat as a structural floor into a variable.
The allocation risk is the one to watch most carefully, because the company has shown the instinct already. A third announced-then-reneged buyback alongside a rising family stake would confirm that restitution is dead and force a re-underwrite to Neutral — a pure value trap. The other permanent-loss path is structural rather than behavioural : measurable Nike or adidas direct-to-consumer share gains eroding the domestic distribution channel would depreciate the oligopoly rent itself, which is the one asset the floor rests on. Neither is signalled today.
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.