Kobayashi Pharmaceutical4967.T
Pull the recovery apart and the debate is not whether earnings come back — they already are. It is whether a capital base that tripled at the top of the cycle can be remunerated again. The fixed asset turnover fell from 8.3x to 2.5x as the beni-koji shock erased a category and left a new plant idle. Net income is climbing back ; return on capital is stuck at 1.7%. The price holds a turnaround that has mostly happened and a capital quality that has not returned.
Normalise the engine and the operating business earns about ¥30.5bn of mid-cycle EBITDA. On the 9.5–11x an out-of-scale, post-scandal, no-moat consumer name earns, that is worth roughly ¥290–320bn of enterprise value.
Add the ¥64.3bn net-cash fortress and equity reconstructs to ¥355–385bn — about ¥4,540–4,930 a share on 78.0m shares. The cellular sum of the parts lands at ¥4,470.
Probability-weight the bear, base and bull, and weighted fair value is ¥4,120.
The market capitalises Kobayashi at ¥439.2bn — ¥5,627 a share. The price holds a recovery that is mostly in it, on capital that no longer turns.
The interesting thing about Kobayashi is not whether the earnings recover, because they already are. Quarterly net income turned positive again in Q1 FY2026 (+¥1.0bn, EPS +¥13.95) after the FY2025 trough, and consensus has the diluted EPS line running from ¥49.2 in FY2025 to ¥141.8 next year and ¥228.4 by FY2028 — a near-trebling that takes earnings back within reach of the ¥268 pre-scandal peak. The question the dossier turns on is the other side of the ledger. The same earnings are now set against a capital base that tripled, and the price is paying for both the earnings coming back and the capital being worth what it was. Those are not the same bet.
The shock arrived in two distinct steps, and separating them is the precondition for any normalisation. In FY2024 the operating margin held at 15.0% while net income halved to ¥10.1bn — the beni-koji compensation provisions sat below the operating line, a one-off. In FY2025 the operating margin itself fell to 9.0% and net income dropped to ¥3.7bn ; the damage had become operational, as the supplements category was withdrawn and a new, oversized plant ran under-loaded. Domestic-segment operating profit fell −40% over that single year ; the international segment fell −36%. The FY2024 hit is expungeable from normalised earnings. The FY2025 hit is only partly so : the lost category is gone for good, and the under-absorption is reversible only if volume returns.
The capital decision is what converts a reputational shock into a durable one. Property, plant and equipment was tripled from roughly ¥22.5bn to ¥67.4bn, on capital expenditure that peaked at ¥26.1bn in FY2024 — 15.7% of sales against a ¥3–4bn historical run-rate — built at the top of a demand cycle that was already flat and delivered straight into the collapse. The fixed asset turnover that measured this idea-factory at 8.3x a decade ago now reads 2.5x. The same normalised earnings, divided by three times the capital, return a fraction of what they used to, and no recovery in the numerator changes that arithmetic.
What the price holds is a recovery that has largely happened and a capital quality that has not come back. At ¥5,627 the shares carry a P/B of 2.0x — a multiple that implies a return on equity near 10% — while the FY2025 figure is 1.7% and a normalised earnings power of ~¥200 of EPS on current book sits closer to 5–7%. The market is paying for a turnaround earnings line and a compounder's capital return at the same time, and a 2.5x fixed asset turnover cannot deliver both.
The position framing is watchlist with a documented short bias, sizing zero, conviction moderate. The asymmetry is materially negative, which removes any long ; the short is not cleanly actionable either, because the activist floor under Oasis, the net-cash fortress and the market's EV/EBITDA anchor leave no datable de-rating catalyst to lean on. The two cardinal variables — the liability bound and the re-absorption of capacity — are gated on the FY2026 accounts, published around February 2027.
The cleanest way to read the last decade is to ignore the income statement, which barely moved, and watch the balance sheet, which broke. Revenue grew from ¥128bn to ¥166bn and the operating margin held in the 14–16% band right up to FY2023 ; on those lines this is a placid, profitable niche business. The damage is one level down. Return on capital sat near 10–11% for nine straight years and then fell to 4.8% and 1.7% across FY2024 and FY2025 — and most of that fall was not the profit collapsing but the capital ballooning underneath a profit that was already going nowhere. The decade did not erode the engine. It buried the engine under three times the assets.
| Inflection | FY 2015Capital-light peak | FY 2019Peak margin | FY 2023Pre-shock plateau | FY 2024Shock below EBIT | FY 2025Stranded capital |
|---|---|---|---|---|---|
| Revenue (¥bn) | 128.3 | 158.3 | 173.5 | 165.6 | 165.7 |
| EBIT (¥bn) | 17.9 | 25.7 | 25.8 | 24.9 | 14.9 |
| EBIT margin | 14.0% | 16.2% | 14.9% | 15.0% | 9.0% |
| Gross margin | 57.4% | 57.5% | 55.6% | 52.9% | 51.1% |
| Return on capital | 9.3% | 11.2% | 10.1% | 4.8% | 1.7% |
| Fixed asset turnover | 8.29x | 7.88x | 4.21x | 2.88x | 2.46x |
| Capex (¥bn) | 3.1 | 4.4 | 12.8 | 26.1 | 13.3 |
| Net income (¥bn) | 12.4 | 19.1 | 20.3 | 10.1 | 3.7 |
| Diluted EPS (¥) | 152.7 | 244.1 | 268.2 | 135.4 | 49.2 |
Source: Data pack 4 June 2026. FYxxxx = year ended 31 December ; FY2015–FY2016 are the legacy March year-ends shown for continuity, December-end thereafter. EPS is split-adjusted (1:2, 28 June 2016). FY2024 net income carries the beni-koji compensation provisions classified below the operating line ; capex peaks at ¥26.1bn (15.7% of sales) the same year as property, plant and equipment is tripled to ¥67.4bn.
Three management decisions explain the shape. The capacity build was the cardinal one : a ×3 expansion of the asset base at the top of a demand cycle whose organic growth had already stalled near ¥155bn, on a niche model that had never needed capital intensity — delivered into the beni-koji collapse, with the FY2024 free cash flow of −¥14.8bn the cash receipt for the error. The product-safety failure was the second : up to roughly 80 deaths recorded, the supplements category withdrawn entirely, and an Oasis-led derivative action filed in Osaka — roughly ¥37bn of net income gone across two years and a category lost for good. The third is older and quieter. Through 2015–2021 the multiple was allowed to run to a 39x P/E on organic revenue that was flat, a premium the company communicated around but never converted into a growth relay ; the de-rating that followed began before the scandal, on the stagnation alone.
The engine only reads correctly at the segment level, because the consolidated 9.0% margin averages two businesses that are economically unlike. Domestic — the household and OTC niches — earned ¥14.0bn of operating profit on ¥118.1bn of revenue in FY2025, an 11.8% margin even in the worst year. International earned ¥0.8bn on ¥47.0bn, a 1.7% margin. The profitable domestic franchise is subsidising an overseas build-out that does not yet pay its way, and the consolidated line hides the cross-subsidy. The US is the only credible organic relay — revenue there grew +44% over three years to ¥23.8bn — but at a 1.7% segment margin it is a top-line story, and it gave the group no protection in FY2025, when international operating profit fell −36% alongside domestic.
Pricing power is the second thing the consolidated figures flatter. The gross margin eroded from 57.4% to 51.1% across the decade, and a meaningful part of that slide predates the scandal — a mix effect as lower-margin international grew, and a genuine loss of pricing on mature niches. The model has no subscription and no switching cost ; its premium was an innovation lead, not a protected rent, and the lead has narrowed. Extrapolating the pre-shock ~14–15% operating margin forward assumes a pricing power the gross-margin trend says is already going.
The critical cost here is not a raw material — the household and small-SKU model is not commodity-intensive — it is the under-absorption of the fixed costs of the stranded capacity, which is what turned a flat revenue year into a six-point margin fall. The cash conversion tells a cleaner story : free cash flow ran ¥12–20bn through FY2022, broke to −¥14.8bn in FY2024 on the capex peak, and was restored to ¥12.3bn in FY2025. The cash-conversion cycle has drifted from 96 days to 150 days, a working-capital drag worth watching, but the conversion itself normalises once the growth capex is digested. The free cash flow is not the problem ; the return on the capital it built is. Against that sits a balance sheet doing nothing — ¥64.3bn of net cash on a 1.7% return on capital — which is precisely the dormant capital the Oasis pressure is aimed at, and which the price gives almost no weight.
This pillar carries the thesis because the dossier is decided on the return on capital, not the earnings. The underlying model was a good one — capital-light, high-margin, ~10% return on capital for nine years — and it was degraded by an allocation decision rather than by the economics. The rating does not fall below 2.5 because the reconstructed operating return ex-stranded-capital is still convalescent at roughly 7%, and the domestic franchise stayed profitable through the worst year. It does not rise above it while the fixed asset turnover sits at 2.46x and the consolidated return on capital is 1.7%, below the ~6–7% cost of capital. Moving this to 3.5 needs a fixed asset turnover restored durably above 4x.
The second cardinal because management created the stranded capital and must now resolve it. Two errors, both costed. Capex was taken to ¥26.1bn in FY2024 — eight times the historical run-rate — at the top of the demand cycle, tripling the asset base into the collapse ; the −¥14.8bn free cash flow that year is the receipt. The beni-koji safety failure produced up to ~80 recorded deaths, the loss of a category, and a shareholder derivative action. The balance sheet was preserved and the Q1 FY2026 return to profit is being managed, which keeps the rating off the floor. But pro-cyclical allocation and a quality-control failure are structural marks against, and the dormant cash could fund a repeat as easily as a recovery.
A resilient household consumables base, low macro cyclicality — but capped organically (revenue flat near ¥155bn for years) and exposed to a confidence shock that already erased a category. The growth leg, the US, is real in top-line and unprofitable.
An innovation cadence and drugstore facings, no switching cost and no network effect. The gross-margin erosion from 57.4% to 51.1% is the moat wearing down mechanically ; each niche is replicable. A weakness, not a barrier.
Ambivalent. A net-cash fortress and a maintained dividend, with Oasis above 10% capable of forcing better capital discipline — set against a derivative action that crystallises a governance failure and ¥64.3bn dormant on a 1.7% return.
A sub-scale turnaround profile, not a compounder. Every pillar sits below 3.0, with two precise structural weaknesses — the moat and the management record — and one fragility constitutive of the archetype : a single product shock moved the return on capital from 10% to 1.7% in two years, the absence of a scale buffer that defines a niche specialist. The grade is consistent with the valuation read : there is no quality premium to claim, and the recovery is already in the price.
Is the stranded capital reversible, or permanent ?
Normative margin : ~14–15%, or structurally lower ?
Consensus extrapolates the pre-shock operating margin, reaching ~11.7% by FY2028 on the assumption of a clean recovery. The gross margin says be careful : its slide from 57.4% to 51.1% began before the scandal, part mix as international grew, part a genuine loss of niche pricing power. If that erosion is structural, the normative operating margin is capped a point or two below the pre-shock band, and a ~12.8% normalised assumption is the prudent anchor.
Is the liability bounded, and does the dormant fortress get mobilised ?
The recovery to ¥228 of EPS assumes the beni-koji liability is passed and will not repeat, but the upper bound is not certified, and a shock that ran across two fiscal years leaves room for further provisions. Against the downside sits the only real upside lever : ¥64.3bn of net cash, Oasis above 10% of the capital and a derivative action on file — the ingredients for a faster capital return that the valuation does not currently hold.
At ¥5,627 the price holds two things that do not fit together at a 2.46x fixed asset turnover : a near-complete earnings normalisation — consensus EPS trebling toward the pre-scandal peak — and a P/B of 2.0x that implies a ~10% return on equity the normalised model does not support. The P/E is the wrong tool here ; on FY2026e earnings it reads ~39.7x and on FY2027e ~30.3x, inflated by a depressed denominator. The forward EV/EBITDA of ~16.8x is itself flattered by a depressed EBITDA (¥23.7bn against a ~¥30.5bn normalised level) ; on normalised EBITDA the enterprise value implies ~12.3x, which for a sub-scale, post-scandal, no-moat name should sit nearer 9.5–11x. Valued part by part on net cash plus a decoted mid-cycle multiple, the sum lands below the market cap — the capitalisation of ¥439.2bn sits above an equity reconstruction of ~¥355–385bn.
The capacity stays idle (fixed asset turnover below 3x), the inverted operating leverage holds the margin near 7–8%, further provisions emerge on the unbounded liability, and an impairment test crystallises part of the stranded capital. This is a permanent loss, not a timing slip : normalised EBITDA of ~¥20.5bn at 8x plus net cash, less the impairment. The net-cash floor of ~¥825 a share and a distressed but still-profitable household base sit underneath it.
Earnings normalise gradually as capacity is partly re-absorbed, the operating margin re-converges toward ~12.8%, and the liability passes without a major additional provision — the supplements category stays lost. Normalised EBITDA of ~¥30.5bn at 10.5x plus the ¥64.3bn net cash reconstructs to ¥4,470 ; the SOTP and a ¥200 normalised EPS on a 16–18x control bracket ¥4,250–4,600. The recovery is real and mostly in the price already.
Both un-priced levers fire together. The liability is bounded, the capacity re-loads on new SKUs (fixed asset turnover back above 4x), the operating margin re-converges toward ~14.5%, the US segment clears a 6% margin, and Oasis forces a capital release that mobilises the dormant net cash. Normalised EBITDA of ~¥33.5bn at 13x plus net cash reaches ¥6,400. The path needs the operational lift and the allocation decision at once, and neither is signalled today.
| KPI | Latest value | Status | What it tells us |
|---|---|---|---|
| Fixed asset turnover | 2.46x FY2025 | Cardinal | The variable the whole dossier turns on, from 8.29x a decade earlier. Durably back above 4x confirms re-absorption and a credible normative return on capital ; below 4x with an impairment test confirms permanent stranded capital and pulls fair value toward ¥2,600–2,900. |
| Beni-koji liability provision | Bound not certified | Cardinal | The shock ran across FY2024 (below EBIT) and FY2025 (in EBIT). No further material provision in the FY2026 accounts validates the normalised EPS ; any additional provision invalidates it. Do not credit a normalised EPS above ~¥220 until the bound is certified. |
| Domestic-segment operating profit | ¥14.0bn FY2025 | Watch | The earnings-power floor, down −40% YoY at an 11.8% margin. A trajectory back toward ~¥20bn annualised is the leading sign that capacity is re-loading ahead of the annual fixed-asset print. |
| Gross margin | 51.1% FY2025 | Watch | Eroded from 57.4% over the decade, part of it pre-shock. Stabilising above 52% with favourable price/mix supports the ~12.8% normative operating margin ; continued erosion below 50% caps it structurally lower. |
| International-segment OP margin | 1.7% FY2025 | Watch | US revenue +44% in three years but the segment dilutes margin and gave no protection in FY2025 (−36% OP). Above 6% makes it a value relay ; durably below 3% confirms a diluter. |
| Capital mobilisation | ¥64.3bn net cash | Trigger | Net Debt/EBITDA −2.95x, Oasis above 10% with a derivative action on file. A quantified capital-return policy or a redeployment to a return above the cost of capital is the only un-priced upside. |
| EV/EBITDA (normalised / forward) | ~12.3x / 16.83x | Reference | Forward (FY2026e) sits on a depressed EBITDA. On a ~¥30.5bn normalised base the enterprise value implies ~12.3x, against the 9.5–11x a sub-scale post-scandal name should carry. |
| Net income / EPS (quarterly) | +¥1.0bn / +¥13.95 Q1 FY2026 | Reference | Turned positive after the FY2025 trough ; LTM diluted EPS ¥41.28. Confirms the numerator is rebuilding — which is the part already in the price, not the part that decides the dossier. |
The case turns off the short bias if the capital comes back to life. A fixed asset turnover durably above 4x at the FY2026 accounts (published ~February 2027), together with a stabilised beni-koji liability and no further material provision, would confirm the turnaround, lift the normative return on capital clear of the cost of capital, and reset the asymmetry from materially negative toward neutral — moving the dossier off watchlist. Either signal is observable on the published calendar ; neither is in the accounts yet.
The case turns to an actionable short if the narrow engine stalls and the market re-prices the denominator. Cosmetics has no analogue here ; the equivalent is the household segment. Domestic-segment operating profit failing to recover toward ~¥20bn annualised, alongside a fixed asset turnover stuck below 3x and an impairment test on the idle capacity, would crystallise the permanent loss and pull fair value toward ¥2,600–2,900. A multiple already anchored on a depressed EV/EBITDA would compress as the recovery narrative cracks.
The allocation risk is the one to watch most carefully, because the company has made it before. A redeployment of the dormant net cash into more capacity, or a diversification acquisition struck at a return below the cost of capital, would repeat the pro-cyclical error that created the stranded capital and burn the one lever the bull case rests on. Currently not signalled.
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.