Koei Tecmo Holdings3635.T
Koei Tecmo prints a 42.0% operating margin at FY March 2026 on ¥88.4bn of revenue, with an operating ROIC inferred at ~50% once the ¥217bn financial portfolio is isolated. That portfolio sits at 79.6% of equity and ~44% of market cap, behind Koyu Holdings — the Erikawa family vehicle controlling 51.92% of votes. SOTP fair value reconstructs to ~¥1,490 at a 30% control discount on the portfolio ; spot is ¥1,481. The single variable governing the asymmetry is the control discount applied to the portfolio.
Operating core ex-portfolio EV is approximately ¥281bn on FY March 2026 EBIT of ¥37.2bn — 7.6x reported, ~9x on AOP normalised ex-FX.
The financial portfolio of ~¥217bn sits at 79.6% of equity and ~44% of market cap.
Consolidated market cap at spot is ¥497.8bn.
At a 30% control discount on the ~¥209bn of excess financial assets, SOTP lands at ~¥1,490 against spot of ¥1,481. The discount carries the entire gap.
The dossier rests on one question. The de-rating — P/B from 5.0x to 1.8x, −38% over one year — either sanctions a post-peak mirage of earnings inflated by a non-repeatable financial portfolio and a weak yen, or it has under-paid a catalogue publisher of first-rank operating economics once the portfolio is isolated. Consolidated reading values one object. The dossier holds two of opposite nature: an asset-light operating core at 42.0% margin and ~50% ROIC ex-portfolio, and a ~¥217bn financial portfolio managed as a deliberate group activity.
The consensus reads the consolidated shell as a fairly de-rated false positive. The forward P/E of 12.6x reads on a net income line that the portfolio inflates in up-years and that fell below EBIT in FY March 2023.
The variant locates the value in the dissociation. Stripped of the financial assets, the core changes hands at ~9x EV/EBIT on normalised earnings — a multiple inconsistent with a 42% reported margin and ~50% ROIC. The hidden value of the variant is hidden only if the control discount is refused. Koyu Holdings, the family parent, controls 51.92% of votes and runs the portfolio as a claimed investment activity, which the governance does not justify refusing.
The asymmetry resolves on the control discount, not the operating margin. At no discount, fair value clears ¥1,600 ; at a 50% discount it falls toward ¥1,360. The core is settled. The minority access to the portfolio is the open variable. The single diagnostic is the financial-assets-to-equity ratio and any signal of portfolio reduction or exceptional return over the next 12–18 months.
Position framing is monitoring at current levels, with a long bias conditional on a governance or allocation trigger. Sizing is zero absent that trigger. The weighted asymmetry is slightly negative, and the downside is floored by book and portfolio. Conviction is moderate.
The decade from FY March 2016 to FY March 2026 reads as three regimes. A Japanese niche plateau through FY March 2019, revenue flat at ~¥37–39bn, EBIT margin 24–33%, net income already carried above EBIT by the portfolio. A musou-plus-COVID-plus-yen breakout from FY March 2020 to FY March 2023, revenue rising to ¥78.4bn and EBIT margin to the 49.9% peak, with the market re-rating P/B to 5.0x in FY March 2021. A post-peak normalisation and recapitalisation from FY March 2024, margin troughing at 33.7% then rebuilding to 42.0%, P/B collapsing to 1.8x, and a FY March 2026 balance-sheet event — ¥24.0bn of property capex, treasury cut from ¥37.7bn to ¥3.7bn, ~18m shares re-issued.
| Inflection | FY Mar 2016Niche plateau | FY Mar 2021Musou breakout | FY Mar 2023Cycle peak | FY Mar 2024Margin trough | FY Mar 2026Recapitalisation |
|---|---|---|---|---|---|
| Revenue (¥bn) | 38.3 | 60.4 | 78.4 | 84.6 | 88.4 |
| EBIT (¥bn) | 11.1 | 24.4 | 39.1 | 28.5 | 37.2 |
| EBIT margin | 28.9% | 40.4% | 49.9% | 33.7% | 42.0% |
| Net income (¥bn) | 10.9 | 29.6 | 30.9 | 33.8 | 42.8 |
| Net income − EBIT (¥bn) | −0.2 | +5.2 | −8.2 | +5.3 | +5.7 |
| Financial assets (¥bn) | 75 | 130 | 133 | 177 | 217 |
| FCF (¥bn) | 10.4 | 28.2 | 28.9 | 34.8 | 9.0 |
| DPS (¥) | 16.7 | 45.0 | 50.0 | 54.0 | 66.0 |
Source: Data pack 31 May 2026 · koei_tecmo_3635.xlsm (Income Statement, Balance Sheet, Cash Flow). Financial assets = cash + short-term securities + long-term investments. FY March 2026 = year ended 31 March 2026. The Net income − EBIT row is the portfolio signature: positive in up-markets, −¥8.2bn at the FY March 2023 operating peak through the 2022 equity drawdown.
The value created on the decade is operating and real — EBIT ×3.4, margin structurally higher, overseas from ~25% to 44.5% of revenue. It was amplified by a non-repeatable peak, masked by a portfolio that inflates net income in up-markets and destroys it in down-markets, and sub-optimised by an allocation that accumulated ~¥217bn of financial assets rather than returning or redeploying them at the ~50% operating ROIC.
Demand is a triptych. Own premium releases set the cadence — Dynasty Warriors Origins, Atelier Yumia, Nioh 3 — correlated to slate timing and decoupled from the hardware cycle through simultaneous PC launches. Collaboration royalties are the capital-light layer, the musou-as-a-service model on third-party IP (Pokémon Pokopia, Hyrule Warriors, Fire Emblem Warriors) that management named as the explicit driver of the FY March 2026 revision. The digital and mobile back-catalogue is the annuity at 171.4m annual downloads. The recurring layer is real ; the release cadence remains the metronome.
Monetisation gains by cost and mix. Musou is mid-price, and there is no pricing power of expansion. The margin expansion from 33.7% to 42.0% over two years rests on combat-engine reuse, a rising digital-catalogue mix, live-service cost cuts, and a non-operating layer — the weak yen translating ~44.5% overseas revenue. The critical cost is the timing of upstream development investment, which moves the margin by ~16 points across the cycle ; management guides FY March 2027 lower for upstream investment on the Switch 2 slate. The second cost is FX.
The unit that matters is the spread between the two capitals. Operating NOPAT of ~¥27.9bn on operating capital of ~¥55.7bn implies a ~50% operating ROIC against a reconstructed WACC of ~7.2%. The ~¥217bn portfolio returns ~7–10%, with non-operating income of ~¥12–18bn. Each yen held in the portfolio rather than returned or redeployed into content gives up ~40 points of spread. That is the exact cost of the allocation.
The operating economics are exceptional: gross margin 64.6%, operating margin 42.0%, ROIC ex-portfolio inferred at ~50% against a ~7.2% WACC, FCF/EBIT ~80–90% ex-events. The model is scalable on combat-engine reuse. Capital allocation pulls the grade down: ~¥217bn of financial assets at 79.6% of equity sit idle by the standard of the operating ROIC, the reported ROE is crushed to 18.5%, and the payout is pinned at ~50% with no structured buyback. An operating model that would score ~4.5 in isolation, dragged to 3.5 by the allocation. Movement to 4.0–4.5 requires a material reduction of the portfolio toward return or content redeployment.
The weakest pillar and the brake on the thesis. Koyu Holdings — the Erikawa family vehicle in rental property and securities holding — controls 51.92% of votes at 31 March 2026, with family officers in concurrent roles at parent and at Koei Tecmo. Asset management is a claimed family activity. The minority is structurally unprotected, and the stock is immunised against activism and TSE-reform pressure, which neutralises the allocation optionality central to the sectoral thesis. The supports are a maintained dividend at ~50% payout and a floored downside. Movement to 3.0 requires a credible governance commitment on capital return or a dilution of control, both unlikely. This pillar is why the dislocation exists and why it may not close.
Triptych demand with a real recurring layer (mobile 171.4m downloads, digital catalogue, royalties), decoupled from the hardware cycle. Cadence of hits is the metronome ; Asia concentration at 23.1% incl. China.
Proprietary musou IP, a reusable mass-combat engine, and reference-partner status for third-party adaptations (Pokémon, Zelda, Fire Emblem). No pricing power, a mature genre, and a collaboration model that depends on partners’ willingness to license.
First-rank operating execution — record FY March 2026, simultaneous PC launches, disciplined live-service cost cuts. The capital event is the weak point: treasury 39→3.7bn with +18m shares, ¥24bn property capex, payout frozen. Management excels at games and at investing, less at minority return.
Below Capcom (19/25) and Konami (18/25), comparable to a two-speed asset: an operating core that would score 4+ in isolation, bridled by the allocation (Economic model) and the control (Governance). The grade does not justify a premium consolidated multiple. It justifies a premium on the operating core only if the control discount on the portfolio is ever allowed to close. The decisive pillar for the thesis is Governance, because it determines whether the hidden value can be crystallised.
Will the ~¥217bn portfolio ever be returned, or is it a permanent family activity ?
Is the 42% margin a catalogue annuity or a peak of musou plus FX ?
Consensus prices a post-peak summit aided by the yen, normalising toward ~38% (FY March 2027e EBIT ~¥33.4bn). The engine carries a real annuity — digital catalogue, mobile, capital-light royalties — more durable than the consensus prices, but masked by FX and the cyclicality of upstream investment. The open question is how much of the 42% survives a USD/JPY ~130 normalisation, where the framework removes ~6 points.
Is the collaboration model a capital-light moat or a partner dependence ?
Consensus reads collaboration royalties as an opportunistic add-on tied to one-off hits. The reference-partner status is a recurring, low-IP-cost asset, confirmed as an explicit driver of the FY March 2026 revision. Koei Tecmo does not own the partner IP and depends on the willingness to renew. It is a borrowed moat.
At ¥1,481 spot and 1.8x P/B against a 5-year average of 4.1x, the market prices three things. A margin normalisation toward ~38% (consensus FY March 2027e EBIT ~¥33.4bn). A deep control discount on the portfolio — book of ¥816 per share is ~80% financial assets, and the price credits only part of that book. And the absence of an allocation catalyst. The stock did not re-rate on record FY March 2026 results. The P/E of 12.6x and the consolidated EV/EBITDA of 12.9x are rejected: net income is portfolio-distorted and has been below EBIT, and consolidated EV includes a portfolio at ~44% of market cap. The SOTP on the operating core plus the portfolio at a control discount is the method retained.
Margin ex-FX proves structurally low at ~30% — the 42% was FX plus cadence. The market concludes the portfolio is never returned and widens the control discount to 50%. The operating multiple compresses to 8x on a quality de-rating. The mechanism is a timing disappointment on margin doubled by a re-rating of the discount — reversible. The floor sits at ~¥900–950 on book (¥816, of which ~¥649 financial assets) and dividend.
Koei Tecmo executes the catalogue, the margin normalises ex-FX toward ~36% on AOP of ~¥30.6bn at an 11x core multiple, the portfolio remains a stable family activity, and the market keeps a ~30% control discount. No allocation catalyst. The SOTP clears the operating core at ~¥337bn plus real-estate NAV ~¥15bn, and the portfolio at a 30% discount on ~¥209bn excess. Roughly fair.
The invisible catalysts materialise together. The margin holds at ~40% on royalties and the Switch 2 pipeline. A governance or TSE signal — portfolio reduction, exceptional return, an evolution at Koyu Holdings — tightens the control discount to ~15%, and the core re-rates to 13x. Improbable without a governance signal, high payoff.
| KPI | Latest value | Status | What it tells us |
|---|---|---|---|
| Financial assets / equity | 79.6% | Cardinal | The single ratio governing the dossier. A reduction greater than 10 points, or any exceptional return, re-opens the file with a long bias and lets the control discount close. Status quo holds the discount permanent. |
| Control — Koyu Holdings votes | 51.92% | Trigger | Any evolution of the control structure re-opens the file. The current level immunises the stock against activism and TSE pressure. |
| Entertainment margin (ex-FX, est.) | ~36% est. | Cardinal | Sustained ≥35% across FY March 2027–2028 refutes the FX mirage and re-rates core value up. Below ~30% confirms the peak and cuts core value ~15–20%. |
| Operating margin (reported) | 42.0% FY26 | Holding | Entertainment segment alone at 44.6%. Consensus FY March 2027e implies ~38%. Not a floor — the FX layer is ~6 points. |
| Net income vs EBIT | +¥5.7bn FY26 | Asymmetric | The portfolio amplifier. Below EBIT by ¥8.2bn at the FY March 2023 peak. The P/E is inoperative ; value the EBIT ex-portfolio. |
| FCF Yield (reported / normalised) | 1.8% / ~6.0% | Watch | Reported collapsed by the ¥24bn property capex. Normalised ex-event ~¥30bn FCF supports the floor. |
| P/B vs 5-year average | 1.8x vs 4.1x | Reference | ~45% below the 5-year average, at the decade floor. Encodes the control discount and the margin normalisation already. |
| Asia mix (incl. China) | 23.1% | Watch | The decade growth driver and the regulatory / competitive exposure. Overseas at 44.5% carries the FX-amplified margin. |
| Beta 2Y vs TOPIX | 0.26 | Reference | Among the lowest in the sub-industry. Confirms the masked-financial-asset profile and the floored downside. |
A credible governance or allocation signal — an evolution of the Koyu Holdings control structure, a portfolio reduction greater than 10 points of the financial-assets-to-equity ratio, an exceptional return, or a structured buyback at the discount — re-opens the file in R1 with a long bias. The control discount becomes a closable gap, and the Bull at +34% activates.
An Entertainment operating margin ex-FX confirmed at or above 35% across FY March 2027–2028 refutes the FX mirage and re-rates the core value up on its own. The diagnostic publishes at the FY March 2027 annual results around April–May 2027.
Two consecutive readings of margin ex-FX below ~30%, or confirmation that the portfolio keeps growing faster than shareholder return over 12–18 months, confirms the permanent control discount and invalidates the long bias. A new low-return capex event or a portfolio expansion in place of return signals a return to the pre-event allocation pattern. 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.