Coca-Cola Bottlers Japan2579.T
Business income doubled in a single year, from ¥12.0bn to ¥24.5bn, and the share is up 74% over twelve months. Pull that recovery apart and roughly half of the margin gain is accounting: a depreciation reset and a useful-life extension that lower the charge without touching the cash. The operational half is real — FY2025 is the highest the certified series has ever printed — but it sits on a franchise that is smaller than it was in 2018, earning a tangible return below its cost of capital. The market is paying for a turnaround it can already see. What is left to decide is whether the margin holds once the accounting tailwind runs out.
Clean EBITDA — business income plus depreciation, before impairment — runs to roughly ¥70bn in FY2026. On the disciplined 9.0x the franchise earns mid-cycle, enterprise value is ¥630bn ; net of ¥57.5bn of debt and across 164.12m shares net of treasury, that is a base-case fair value of ¥3,488.
Weighting the bear at ¥2,574 and the bull at ¥4,402 against that base brings the probability-weighted fair value to ¥3,424.
The market prices the share at ¥4,307 — a net capitalisation of ¥706.9bn, an enterprise value of ¥764.4bn, and a clean EV/EBITDA of 10.9x, above the 10.3x ceiling of the entire decade. Even held at that ceiling, the implied fair value reaches only ¥4,043. The doubling everyone extrapolated does not survive the multiple it is already paid at.
Coca-Cola Bottlers Japan is the one pure converter in its bucket: an industrial and logistics operator with no brand, no concentrate it can price, and no business outside Japan. It runs about 90% of the Coca-Cola system's volume in the country, buys its concentrate at a price The Coca-Cola Company sets, and earns a network-execution margin — never a brand margin. So when business income doubled from ¥12.0bn to ¥24.5bn between FY2024 and FY2025, the question was never whether the recovery was real. It was what kind of recovery it is, and whether a converter can earn its way out of a survival model.
The honest answer is that the doubling is two things at once. A proxy decomposition puts roughly 46% of the FY2024-to-FY2025 gain in the depreciation line: the charge fell from ¥45.5bn to ¥39.7bn after the vending impairment, and a useful-life extension confirmed in Q1 FY2026 lowers it further. That is a non-cash tailwind, and it flatters the margin without improving the cash. The other 54% — pricing and channel mix — is operational and genuine. The certified series settles one part of the debate: FY2025 business income of ¥24.5bn is the highest the company has ever printed, 5.4% above the post-merger FY2018 peak and at a higher margin despite lower revenue. A structural component exists. The size of it is the whole argument.
What the market has done is pay for the success and not the risk. The clean EV/EBITDA of 10.9x sits above a decade ceiling of 10.3x, the price-to-book has re-rated from 0.48x at the 2021 trough to 1.78x, and the share is up 74% over the year and 7.5% in the three days into this read. Embedded in that multiple is a business income margin reaching the ≥5% Vision 2030 target, a tangible return finally crossing the cost of capital, and a PBR re-rating that holds. None of those is proven. The owner earnings — clean EBITDA less maintenance capital expenditure — come to about ¥40bn, a 4.4% margin that falls short of the 5% target the price assumes.
The second feature of the case is that the negative read does not translate into a position. The weighted fair value is ¥3,424, some 20.5% below spot, and the directional bias is short. But the short is inexecutable. A 4.3pp buyback yield, an active TSE/PBR bid, and momentum of +74% over the year make shorting a self-help name a squeeze trap. The bias is short ; the trade is not. That gap is the practical heart of the file.
The position framing is observation, not ownership and not a short. There is no actionable entry at this price: the asymmetry is too negative to be long and the technical floor is too firm to be short. Conviction is moderate. Two things are worth watching on the published calendar — the business income margin once the depreciation tailwind exhausts, and the trajectory of vending revenue.
The decade reads as one long round-trip rather than a compounding. The April 2017 merger of Coca-Cola West and Coca-Cola East Japan built the present entity and lifted nominal revenue to ¥927bn at the FY2018 full-year — a perimeter artefact that makes any comparison across 2017 meaningless. Then the integration consumed it. A ¥61.9bn goodwill impairment in FY2019 took reported profit deeply negative, COVID cut FY2020 revenue by 11% and hollowed out vending, the dividend was halved from ¥50 to ¥25, and net debt peaked at ¥155.2bn. Gross margin slid from 50.8% in 2015 to a 43.6% trough in 2022. Business income fell from a ¥23.3bn post-merger peak to a ¥14.7bn loss. The recovery since FY2023 is genuine, but it returns the company to roughly where it started: revenue at ¥893.8bn is below the FY2018 figure, and the franchise has shrunk organically rather than grown.
| Inflection | FY 2019GW impairment | FY 2021COVID trough | FY 2022Margin trough | FY 2024Recovery base | FY 2025Re-rating |
|---|---|---|---|---|---|
| Revenue (¥bn) | 890.0 | 785.8 | 807.4 | 892.7 | 893.8 |
| Business income (¥bn) | 11.4 | −14.7 | −14.4 | 12.0 | 24.5 |
| EBIT reported (¥bn) | −58.9 | −20.8 | −11.6 | 13.4 | −72.5 |
| Gross margin | 47.3% | 44.6% | 43.6% | 45.1% | 44.7% |
| FCF (¥bn) | −35.6 | −3.3 | 10.0 | 20.7 | 31.3 |
| Return on capital (reported) | −8.0% | −0.2% | −1.1% | +1.2% | −8.9% |
| Net debt (¥bn) | 76.4 | 103.6 | 95.9 | 54.2 | 57.5 |
Source: Data pack 22 June 2026, certified business income series (IS_ADJUSTED_OPER_INC), tanshin filings. Reported EBIT and return on capital are distorted by non-cash impairments (¥61.9bn goodwill FY2019, ~¥88bn operational vending FY2025) and carry no economic signal ; the business income line is the only reliable read of recurring profit. Dividend rebuilt ¥25 (FY2020) → ¥50 (FY2021) → ¥60 (FY2025) → ¥72 guidance (FY2026).
Three management choices explain the U. The 2017 merger was over-capitalised — the ¥61.9bn goodwill impairment two years later was the admission that part of the acquisition price was never going to earn its return, and it dropped return on capital to −8%. The vending fleet was kept and invested in long after its economics turned dilutive ; the ~¥88bn operational impairment in FY2025 is the late crystallisation of a decline that was visible for years. And the 2020 dividend cut, from ¥50 to ¥25, showed how thin the cash cushion really was under a shock. The discipline since — ¥30.2bn of buybacks in FY2025 against ¥4.6bn the year before, a dividend rebuilt to a ¥72 guidance, net debt down to ¥57.5bn — is real and corrective. It is also funded in part by letting the asset base age, and the same dormant balance sheet that funds today's buyback funded yesterday's vending over-build.
The engine only makes sense at the channel level, because the consolidated line averages businesses with very different economics. In Q1 FY2026 the OTC channel — supermarkets, drugstores, convenience stores, e-commerce — earned an income margin near 9.2%, food service near 8.3%, and vending near 1.8%. That 740bp spread between OTC and vending is the cardinal fact of the file. A single ~2.7% consolidated business-income margin is the weighted average of a healthy channel and a dying one, which is why a single consolidated multiple is the wrong instrument and why the mix shift, rather than any uniform improvement, drives the margin.
Where the pricing actually lives matters, because the word does a lot of quiet work here. CCBJI has run eight price increases since 2022 and added ¥15.1bn of price contribution in FY2025, and Q1 FY2026 carried +4.0% volume alongside a higher wholesale revenue per case — price passed without volume destruction, so far. But this is execution pricing, not brand pricing : the company passes price through its channels, while The Coca-Cola Company owns the intellectual property and sets the concentrate cost. The lever is real and bounded — bounded by elasticity, estimated around ¥200 per PET bottle, and by the dictated input. It is the channel mix, OTC over vending, that is the genuine margin vector.
The cost base is something the converter carries rather than controls. The concentrate price is fixed by The Coca-Cola Company and the logistics fleet is fixed cost ; the only endogenous lever is efficiency, and the company has used it — headcount is down 26% since 2017, from 17,197 to 12,667, and capital expenditure has normalised to about ¥30bn. The arabica reflux of −41% from its 2025 peak feeds a margin tailwind for Georgia coffee into FY2026-2027, lagged by two to three quarters, but it is transitory and should not be read as structural quality.
The cash bridge looks excellent and is partly an illusion. Free cash flow of ¥31.3bn converted at 128% of business income in FY2025, but capital expenditure of ¥29.8bn ran below depreciation of ¥39.7bn — a ¥9.9bn run-down. As the useful-life extension pulls depreciation further below maintenance capital expenditure, business income will increasingly overstate the sustainable cash earnings. Normalised to a ¥30bn maintenance charge, the owner earnings are about ¥40bn, a 4.4% margin below the 5% target. The cash machine is real — CFO held at ¥61.1bn through a loss-making reported year — but the recovery is being read at a quality the underinvestment does not support. The consolidated margin is the average of a channel that creates value and one that destroys it, and the price reads only the average.
This is the pillar that decides the thesis, and it is the weakest. Return on capital ex-cash is 3.9% and ex-goodwill 4.4%, both below a cost of capital near 6%, and the gap between the two is almost nil. That last point is the diagnosis : the problem is not an obese balance sheet hiding good operations, as it is at Asahi, where the ex-goodwill return is far higher. It is the tangible capital itself earning below its cost — a root pathology, not an allocation one. CFO of ¥61.1bn and FCF of ¥31.3bn are real but do not cover the ~8% cost of equity, and the FCF is flattered by capex running under depreciation. No amount of deleveraging or buyback corrects capital that earns below its cost ; only a durable margin lift that pushes the tangible return through the cost of capital would, and that is the proof the file does not yet have.
The moat is the second cardinal because it explains why the converter cannot capture the value its volume implies. Territorial exclusivity over the Coca-Cola system in Japan — about 90% of system volume — is a genuine execution barrier that a third party could not replicate quickly, and the switching cost for The Coca-Cola Company to replace its bottler is high. But it is a barrier of volume, not of margin. It has never produced a return on capital above the cost of capital, and gross margin eroded six points, from 50.8% to 44.7%, despite the barrier holding the whole time. The converter remains a price-taker on concentrate. The exclusivity secures the share ; it has never secured the price.
Defensive in aggregate — beverage is a staple — but structurally shrinking : volume −0.7% a year on demographics, vending −2.7%, 100% Japan with no geographic smoothing. The OTC pivot carries the growth and the margin on a declining base.
Recent execution is genuinely good — eight price increases, −26% headcount, buybacks up to ¥30.2bn, and a certified margin above the FY2018 peak. The decade record is heavy : two impairments and the 2020 dividend cut. Good management in a poor model.
The TSE/PBR pressure aligns the company with minorities for now — dividend up 2.9x, large buybacks, deleveraging, share count down 180m to 164m. But the re-rating depends on an exogenous PBR bid, The Coca-Cola Company's influence over concentrate is a structural conflict, and the 2020 cut dented the payout's credibility.
A survival model in real but capped recovery — neither a pure trap nor proven value. The dispersion between execution, scored at 3.0 on management and alignment, and the model itself at 2.0 is exactly the file: good management in a poor model. Above a value sink, well below a quality compounder such as Food & Life at 19–20/25, and consistent with the valuation — the multiple is already paid above the decade ceiling, and the per-channel sum, if it could be built, would read lower rather than higher.
Is the business-income recovery structural, or mechanical and already in the price ?
Does vending stabilise, or keep destroying value ?
Consensus reads the OTC pivot as offsetting the vending decline and the 2025 impairment as having cleaned the problem. Vending is still 45% of revenue at a ~1.8% income margin, contracting −2.7%, and the operational impairment — goodwill certified at zero — confirms a uniformly uneconomic asset base rather than a one-off clean-up. If the decline accelerates, the 45% erodes faster than OTC at 47% can carry the consolidated margin, and the mix shift becomes a race the company loses.
Does the PBR re-rating survive without a fundamental relay ?
Consensus reads the 0.48x-to-1.78x price-to-book re-rating as justified by self-help and capital return, with a ~5.8% shareholder yield underpinning it. But the re-rating is a regime, not a fundamental : it pays a low optical book with self-help, not quality — the decade EPS is negative and return on capital is below the cost of capital. The shareholder yield is carried 4.3pp by the buyback, the price-to-book sits above the decade corridor on a destroyed book, and the 0.53 beta, the highest in the bucket, makes the name a regime-amplitude proxy.
At ¥4,307 and a clean EV/EBITDA of 10.9x, the market is pricing a turnaround that mostly succeeded and stays succeeded — a business-income margin reaching ≥5%, a tangible return crossing the cost of capital, and a sustained capital return. The multiple sits above the decade ceiling of 10.3x ; even held at that ceiling, the implied fair value is ¥4,043, still below spot. The tell is in the regime: the share is up 74% over the year and 7.5% in three days, the price-to-book has re-rated from 0.48x to 1.78x on a destroyed book, and the shareholder yield is carried by buyback, not earnings. P/E is dead here — the decade EPS is negative — so the discipline is clean EV/EBITDA, which has the further merit of being immune to the depreciation reset. Valued on its own honest metric, the share is paid for the success and not its risk.
Vending accelerates its contraction, the business-income margin caps below 4% — revealing the FY2025 doubling as mechanical — and the PBR bid fades as value funds exit. Clean EBITDA stays at ~¥64bn and the multiple compresses to 7.5x, the low end of the decade corridor. The downside is mostly reversible — a regime relapse rather than permanent loss — with a floor near ¥2,900–3,000 held by the resilient CFO and the FCF yield. It tips below only if the vending network becomes structurally sub-critical.
Pricing and channel mix hold, the margin edges toward ~4%, but the tangible return stays below the cost of capital and the depreciation tailwind runs out. Clean EBITDA of ~¥70bn on the FY2026 guidance, FCF near ¥30bn, capital return sustained. The multiple settles at 9.0x, the upper-middle of the corridor, for an enterprise value of ¥630bn and a fair value of ¥3,488. Execution is good ; the converter cap is intact.
The ≥5% business-income margin is held by pricing and mix after the depreciation tailwind is spent — maintenance capex stable, FCF/AOP not degraded — the tangible return approaches the cost of capital, and the re-rating perpetuates outside the corridor. Clean EBITDA reaches ~¥78bn by FY2027 with an owner-earnings margin above 5%, on a multiple held at 10.0x. The market already prices this path, which is why the residual upside is barely positive.
| KPI | Latest value | Status | What it tells us |
|---|---|---|---|
| Business-income margin | 2.74% FY2025 | Cardinal | The swing variable. A margin held ≥5% across FY2026-27 with stable maintenance capex and undegraded FCF/AOP confirms a structural recovery ; capping below 4% confirms the value trap and pulls fair value toward ¥2,574. |
| Vending revenue YoY | −2.7% FY2025 | Watch | 45% of revenue at a ~1.8% margin. Below −3% over two consecutive quarters confirms continued destruction ; stabilisation toward 0% validates the OTC pivot. |
| OTC vs vending margin spread | 740bp Q1 FY2026 | Watch | OTC ~9.2% vs vending ~1.8%. The consolidated margin is set by how fast OTC growth outruns the vending decline — the dispersion, not the average, is the read. |
| Return on capital ex-cash vs WACC | 3.9% vs ~6% | Priced | The root pathology. Ex-cash 3.9% ≈ ex-goodwill 4.4%, gap near nil — tangible capital earning below its cost. A durable crossing of the cost of capital is the proof the multiple already assumes. |
| Maintenance capex vs D&A | 29.8 vs 39.7 FY2025 | Watch | A ¥9.9bn run-down flatters FCF. As the useful-life extension pulls depreciation below maintenance capex, business income overstates sustainable cash earnings ; owner earnings normalise to ~¥40bn (4.4% margin). |
| Clean EV/EBITDA (TTM / forward) | 11.9x / 10.9x | Reference | Against a 5.9–10.3x decade corridor, average 8.5x. Above the ceiling reads expensive ; compression toward 9.0x is the base case. Even at the 10.3x ceiling, implied fair value ¥4,043 is below spot. |
| Shareholder yield / buyback pace | 5.7% · ¥30.2bn FY2025 | Trigger | Carried 4.3pp by buyback, not earnings. It is what makes the short inexecutable. Deceleration below ~¥15bn a year removes the technical floor and feeds the regime-relapse case. |
| Café (arabica) spread | −41% from peak | Reference | A 2-3 quarter lagged margin tailwind for Georgia into FY2026-27. Transitory — treated as a temporary uplift, not capitalised as structural quality. |
The case turns positive if the margin proves structural rather than accounting. A business-income margin held at or above 5% across two consecutive prints to FY2027, with maintenance capex stable near ¥30bn and FCF/AOP not deteriorating, would show the recovery is carried by pricing and mix and not by the depreciation reset — and would make the tangible return crossing the cost of capital underwritable. The technical route is simpler : a price reflux below ~¥3,600 opens a positive asymmetry toward the ¥3,424 fair value and would move the file from watchlist to a long entry. Either is observable ; neither is signalled today.
The case is confirmed negative if the narrow engine stalls. Vending revenue falling more than −3% YoY over two consecutive quarters, alongside a business-income margin capping below 4%, would confirm the FY2025 doubling was mechanical and the value trap real — holding the watchlist with the short bias reinforced, and still inexecutable. A regime relapse, the PBR bid fading without a return-on-capital crossing, compresses the multiple from 10.9x toward the corridor and the floor near ¥2,900–3,000.
The allocation risk is the one to watch most carefully, because the company has made it before. The buyback is being run at 10.9x EV/EBITDA, above the decade corridor and above the ¥3,424 intrinsic fair value — a capital return that quietly destroys value if sustained above the worth. Worse would be growth capex reinvested into the declining vending channel, a repeat of the over-build that produced the FY2025 impairment and an irreversible destruction. Neither is signalled now, but both sit inside the dormant balance sheet that funds today's discipline.
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.