Honda Motor Co., Ltd.7267.T
Sixty-four percent of the revenue is cars, and at a normalised yen the car business earns approximately nothing. The profit lives in two places the consolidated line hides: a motorcycle franchise that held an 18.2% margin through the worst loss in the group's history, and a captive bank whose debt makes the balance sheet unreadable. Valued part by part, the sum runs well ahead of the price. What has to be decided is narrower: whether the electric-vehicle drain is bounded, and whether a discount ten years old can be collected at all.
Motorcycles, at ¥730bn of normalised operating profit on the 9.0x that an 18.2% margin and a 27.0% return on assets earn, are worth roughly ¥6,570bn.
The automobile core, which produces somewhere between −¥177bn and +¥24bn once the yen is normalised to 130, is carried at zero. Power products, at zero.
The captive bank, on ¥2,800bn of segment equity at 0.80x book, adds ¥2,240bn; an industrial net cash position of ¥1,700bn and ¥331bn of minorities bring equity to ¥10,179bn, or ¥2,615 per share.
The market capitalises Honda at ¥6,068.5bn. Strip out the captive and the cash and it is paying between 3.4x and 6.6x the operating profit of the motorcycle franchise, whose listed comparables trade between 10x and 14x.
Honda files as a car company and is traded as one, benchmarked against Toyota and Nissan, screened on automotive margins. The segment note tells a different story. In the last clean year, motorcycles and the captive finance arm produced 80.7% of group operating profit between them; the automobile business, on 64% of external revenue, produced roughly a fifth. Normalise the yen to 130 and the car business produces approximately nothing at all — two independent reconstructions put it between −¥177bn and +¥24bn. So the entity the market is pricing and the entity that actually earns money are not the same entity, and the gap between them is what the whole dossier turns on.
The year that made this visible was brutal. In the year to March 2026 Honda booked its first operating loss of the decade, −¥414.3bn, on a ¥1,577.8bn electric-vehicle charge — cancelled North American models, written-off development, provisions against alliance and supply commitments. The decomposition matters more than the headline. Impairments and write-offs of ¥852.8bn sit on assets whose recoverable value was assessed at zero, and cannot be impaired twice. The remaining ¥667.4bn is provisions, of which ¥561.1bn covers undertakings to third parties on unpublished terms. The debate is confined to that second component.
What makes the case interesting is what the price implies once you take the pieces apart. Value the captive on its equity at 0.80x book, credit the industrial cash position at the middle of its range, and the residual the market is paying for the motorcycle business works out at 3.37x its operating profit. Take the most punitive assumptions available — captive at 0.60x book, industrial cash credited at zero — and it is still only 6.63x. Hero MotoCorp, the closest comparable on quality with a 25.9% return on equity, trades at roughly 14.3x EBIT; Yamaha, diversified and running a 2.05% return on invested capital, at roughly 10.3x. Honda's motorcycle business held its margin to within ten basis points through the worst year in the group's history and returned 27.0% on the assets it employs. It is being valued below the least impressive comparable available.
There is a hard limit on how far that observation can be pushed. The sum-of-the-parts base case implies a price-to-book of 0.86x, and Honda has never crossed its book value in ten years — the fiscal-year closes run from 0.8279x down to 0.4142x, the low set in March 2026. The bull implies 1.40x, a level the stock has not approached. So the arithmetic can be right and still uncollectable, because collecting it requires a re-rating the sector's own behaviour says does not happen: ¥1.4tn of buybacks across two years retired 28% of the share count and the multiple halved anyway.
The position framing is observation with a documented long bias, sizing at zero. Direction is robust — no cell on the sensitivity grid produces a fair value below the price, and the lowest tested still prints ¥1,616. Magnitude is unestablished, because the industrial net cash position ranges from zero to ¥3,348bn and Honda publishes neither liabilities nor cash flow by segment. Weighted fair value is ¥2,549. Conviction is moderate; the diagnostic arrives with the half-year to March 2027, around November 2026.
Read across the decade and the striking thing is how little the operating economics moved. Revenue grew about 49%, from ¥14.6tn to ¥21.8tn, but the operating margin never held above 7% and spent most of the period between 3% and 6%. The peak, 6.8% in the year to March 2024, coincided with the weakest yen of the period. Return on equity oscillated between 6% and 9% outside a single spike in FY March 2018 that came from a one-off US tax benefit. The group grew, absorbed a pandemic, reflated, and arrived at a structurally identical level of profitability.
| Inflection | FY 2016Pre-cycle | FY 2021COVID trough | FY 2024Reflation peak | FY 2025Last normal year | FY 2026EV reset |
|---|---|---|---|---|---|
| Revenue (¥bn) | 14,601.2 | 13,170.5 | 20,428.8 | 21,688.8 | 21,796.6 |
| Operating profit (¥bn) | 503.4 | 660.2 | 1,382.0 | 1,213.5 | −414.3 |
| Operating margin | 3.4% | 5.0% | 6.8% | 5.6% | −1.9% |
| Return on equity | 5.0% | 7.7% | 9.3% | 6.7% | −3.5% |
| Return on capital | 3.0% | 4.2% | 5.7% | 3.9% | −1.2% |
| Free cash flow (¥bn) | 755.8 | 754.0 | 398.6 | −218.7 | 523.2 |
| Net debt (¥bn) | 4,838.0 | 5,280.4 | 5,541.4 | 7,245.4 | 8,740.0 |
| Shares net of treasury (m) | 5,406.9 | 5,180.0 | 4,828.9 | 4,346.5 | 3,892.6 |
| Book value per share (¥) | 1,250.5 | 1,753.4 | 2,629.4 | 2,836.0 | 3,035.9 |
Source: workbook and data pack, pull 22 July 2026, close-year convention (FY 2026 = year ended 31 March 2026). Operating profit is IFRS reported. Free cash flow and net debt are consolidated and therefore distorted by the captive: the swing from −¥218.7bn to +¥523.2bn is driven by movement in finance receivables, not by industrial performance. Book value per share is computed on the net-of-treasury divisor and reconciles to the Tanshin to the yen.
Three decisions account for the shape. The electric-vehicle over-extension is the expensive one: capacity committed on a North American demand curve that reversed, unwound in a single ¥1,577.8bn charge on assets that could be neither sold nor repurposed. The second is quieter and cost more over time — a decade of holding an automobile core at a normalised margin of about nothing while it absorbed 39% of the capital employed and ¥1,143bn of research spending outside that charge. The third is that ¥1.4tn was returned without any structural work to make the parts legible. The capital went back. The discount stayed.
The unit that explains Honda is capital employed by segment, not margin. The Tanshin publishes assets, depreciation and capital expenditure per segment, which makes the calculation possible, and the result is a spread of roughly fifteen to one. In the last clean year motorcycles returned 29.50% on segment assets while holding 7.4% of them. Automobiles returned 2.05% on 39.0% of assets. Financial Services returned 2.01% on 51.7% of assets and consumed 77.4% of group capital expenditure. Power products returned −1.64%. Against an industrial cost of capital of 6–7%, one of four businesses clears the bar, and it is the one holding the least capital.
That single table dissolves most of what the consolidated ratios appear to say. A group return on equity of 6–9% is the weighted average of an asset-light rent and a balance sheet in which more than 90% of the capital employed earns about 2%. Motorcycles run at 2.6–3.4% capital expenditure to sales against 5.5–6.2% for automobiles, which is what makes them a rent rather than a peak. The margin moved from 18.3% to 18.2% in the year the group margin went from +5.6% to −1.9%. A ten-basis-point move through a shock of that size is a brand price being held, and it is the only pricing power here — the guidance bridge puts the rest at −¥313.0bn on the price-and-cost line, costs running ahead of prices everywhere else.
The critical cost is neither the tariff nor commodities. On the eight lines of the guidance bridge from FY March 2026 to FY March 2027, the two electric-vehicle items are worth +¥953.6bn net against +¥147.0bn for tariffs and −¥142.0bn for currency — six times the two combined. Honda's critical cost is therefore endogenous and discretionary, which separates it from the exporters in the bucket, where the binding constraint is trade policy and the yen. The tariff exposure is in fact the least severe in the sub-industry, at roughly 90% local US production against 57.7% of revenue earned in North America. And the guidance bridge carries a second, less comfortable reading: strip the electric-vehicle arithmetic out and the operating base degrades by ¥39.3bn. The FY March 2027 rebound is an accounting event.
Which leaves the cash. No consolidated conversion ratio is meaningful — movement in the captive's finance receivables dominates operating cash flow, and ¥612bn of profit-and-loss capital expenditure sits against ¥3,805bn of segment capital expenditure, ¥2,766bn of it the captive leasing fleet. What can be computed is industrial: ¥1,038.6bn of capital expenditure against ¥691.7bn of depreciation, a ratio of 1.50x consistent with a technology transition. And the ¥943.8bn distributed in FY March 2026 — ¥272.9bn of dividend, ¥670.9bn of buyback — amounts to 111% of the ¥850bn of normative operating profit, paid in a loss year out of the balance sheet and the captive's capacity to remit cash upward. The 15.6% shareholder yield is a balance-sheet floor rather than an earnings floor, and its durability is itself a debate.
The lowest grade in the set, and the one that carries the thesis, because it is the only pillar whose movement could reconcile intrinsic value with the price. The vulnerability is arithmetic rather than operational: more than 90% of the capital employed earns about 2%, far below any plausible industrial cost of capital, and the 7.4% of assets sitting in motorcycles at 29.50% cannot lift that weighted average. Layered on top is ¥1,577.8bn destroyed on assets assessed at zero recoverable value. Reduce or make legible the capital held in the captive and the group's return on capital moves mechanically toward the motorcycle number, at which point the complexity discount loses its analytical basis. No improvement in the other four pillars produces that effect.
The second cardinal because it is the value anchor and the floor under the downside. The motorcycle moat has just been given the best empirical test available, and it held to within ten basis points on a 27.0% return on assets and 3.4% capital intensity. That is a rent, verified cellularly rather than narratively, and it is what makes the bear case a timing disappointment. The limit is reach. The moat covers 18% of revenue and stops. Automobiles are a commodity product in an oversupplied market where Honda holds no cost or brand advantage; the captive is a refinancing spread, reproducible by any manufacturer with a comparable credit rating. Authentic, narrow, and the only thing here worth a franchise multiple.
Motorcycle revenue grew 10.8% in the loss year, driven by emerging-market motorisation rather than the developed discretionary cycle; captive revenue held at +0.6%. Automobiles fell 2.2% with share eroding in China and ASEAN. No aftersales recurrence of the kind that shelters Isuzu.
The reset was taken in one movement on an identified date, published with unusual granularity, and the dividend was held at ¥70 through a loss. Against that: ¥1,577.8bn destroyed by pro-cyclical over-extension, and a decade holding a zero-margin core without arbitration. Above Nissan, below Isuzu.
The strongest pillar on facts that are hard to argue with: 15.6% shareholder yield, 14.1% of issued capital cancelled in one year, no dilution and no options outstanding across ten years, disclosure quality above the sector. What caps it is the absence of any structural work to make the parts visible.
A real asset lodged in an adverse capital structure. That places Honda third of seven in the sub-industry, above Nissan at 8.5 and the currency-illusion exporters, below Toyota at 17.5 and Isuzu at 16.0. The grade explains the price: at 0.514x book the market is pricing the structure of the balance sheet, not the quality of the assets inside it. Honda ranks sixth of seven on the moat and first on asymmetry — the inverse ordering is the fact worth carrying out of this section.
Is the electric-vehicle charge a calibrated reset, or the first instalment of an unbounded drain ?
Is the capital return sustainable, or is it the balance sheet being drawn down ?
One camp treats the 15.6% yield as an acquired valuation floor; the other notes it was paid in a loss year and reads a short-dated defence of the share price. The arithmetic favours caution: ¥943.8bn distributed against ¥850bn of normative operating profit is 111%, funded from the balance sheet and the captive's capacity to remit cash upward. That capacity is bounded, since the captive must hold its capitalisation ratios to protect its funding cost — and the group's cost of debt has already moved from 0.44% to 2.61% in two years.
Is the captive's profit an autonomous rent or a transfer price from the car business ?
The conventional treatment values a captive on its equity as a financial institution, and the sub-industry framework carries it as a genuine pocket of value. The competing reading is that part of the 7.8% margin is automobile margin reclassified: the captive quotes the end customer a rate that subsidises the vehicle sale, in which case valuing both separately counts the same profit twice, and the residual-value writedowns already taken on the electric leasing fleet are the deferred cost of it. Honda states that intersegment transactions approximate arm's length without demonstrating it.
At ¥1,559 and 0.514x book, three things are embedded. The complexity discount is priced as permanent — ten years without crossing book, a multiple drifting from 0.82x to 0.51x, two years of record buybacks fully neutralised. The electric-vehicle drain is priced as a series rather than a term: the stock did not react to the +¥1,453.6bn accounting tailwind written into FY March 2027 guidance, so the announced non-repetition is not being credited. And the motorcycle business is priced at no standalone value, the residual paid for it working out at 3.4x to 6.6x operating profit against 10x to 14x for its comparables. What the price does not hold is the industrial cash position, which no aggregator reconstructs and for which the desk confirms no standardised field exists. What it over-prices, in the other direction, is the group's ability to have that value recognised without structural action. Method is sum-of-the-parts with price-to-book as control; price-to-earnings is rejected on a negative trailing base and an incoherent forward, EV/EBITDA on a captive that inflates both assets and depreciation.
The mechanism is sequential and is not the mirror of the bull. Provisions prove understated, further charges land on alliance and supply contracts, and automobiles move from zero value to negative — a ¥150bn annual drain capitalised over five years at roughly −¥650bn. Financing that drain then consumes the captive's upward remittance and the industrial cash, so the cash assumption collapses to zero at the same moment. Buybacks are suspended, the dividend cut, the captive de-rated to 0.60x book. Motorcycles at 7.0x. Implied price-to-book 0.42x — approximately where the stock closed in March 2026, four months ago. Reversible, and the floor rests on neither inferred variable.
Execution in line, without recognition. The electric-vehicle residual stabilises at the guided ¥500bn and burns down; motorcycles grow at a high single digit in emerging markets at an unchanged margin; automobiles sit at break-even at a normalised yen; the captive holds its spread. The dividend stays at ¥70 and the buyback is trimmed to ¥400–500bn. Motorcycles at 9.0x, captive at 0.80x book, industrial cash credited at ¥1,700bn. Implied price-to-book 0.86x, the exact top of a decade corridor never crossed — the constraint on this scenario, and the reason no structural action is assumed inside it.
The invisible catalysts fire together. The residual burns down faster than guided, into the ¥335–345bn the consensus already carries; automobiles recover a normalised +¥150bn on tariff insulation and a lower break-even, capitalised at 6x. And structural action arrives — separate industrial disclosure, securitisation, or a partial listing of the captive — which makes the parts legible and forces recognition. Motorcycles at 12.0x, captive at 0.95x book, industrial cash at ¥3,300bn. Implied price-to-book 1.40x, a level Honda has never approached; the scenario assumes a break in the valuation regime itself, which is why probability is held at 18%.
| KPI | Latest value | Status | What it tells us |
|---|---|---|---|
| Semi-annual electric-vehicle losses | ¥500bn guided FY Mar 2027 | Cardinal | The diagnostic, first observable at the half-year to March 2027, around November 2026. Above ¥250bn for the half, or any top-up to the onerous-contract provisions, invalidates boundedness and moves the dossier to a justified trap. |
| Motorcycle segment margin | 18.2% FY Mar 2026 | Holding | The value anchor and the floor, held within ten basis points through the group's first operating loss of the decade, on a 27.0% return on assets. Degradation here destroys the only quality asset in the group. |
| Industrial net cash ex-captive | ¥0 to ¥3,348bn | Unestablished | Governs 55% of the market capitalisation and swings base fair value from ¥2,178 to ¥3,026 on its own. Resolution runs through Item 18 of the 20-F filed 18 June 2026 — the precondition on the Temps 2b cycle. |
| Buyback run-rate and dividend | ¥670.9bn · ¥70 DPS | Trigger | Distribution ran at 111% of normative operating profit. Below ¥400bn annualised, or a dividend under ¥70, removes the floor the yield-fund holder base is buying, with no other buyer category to take over. |
| Automobile quarterly result ex-EV | −10.0% margin FY Mar 2026 | Watch | Read against the effective quarterly rate. Below ¥50bn a quarter above 140 confirms the normalised zero. Above ¥100bn on the same terms restores positive value to the core. |
| Financial Services margin | 7.8% FY Mar 2026 | Watch | Holding above 7.5% while North American automobile volume declines confirms an autonomous rent. Degrading in proportion, or a further residual-value writedown on the electric fleet, confirms cross-subsidy and deepens the discount on the segment. |
| Price-to-book | 0.514x | Reference | Decade corridor 0.41–0.85x on fiscal-year closes, low set at March 2026. Base fair value implies 0.86x, the top of that range; the bull implies 1.40x, never approached. The re-rating channel is the unmodelled variable of the case. |
| Consensus versus guidance | ¥372.3bn vs ¥260.0bn | Reference | Seventeen contributors, revised 14–15 July 2026, so the +43.2% gap is a live divergence rather than stale data. It sits entirely on the size of the electric-vehicle residual. |
The case turns actionable on two conditions, and only the first is dated. A half-year to March 2027 printing electric-vehicle losses at or below ¥250bn with no addition to the onerous-contract provisions would establish that the reset was calibrated, and would remove the reason the discount exists. The second is the one that converts value into price: separate financial disclosure for the industrial perimeter, a securitisation, or a partial listing of the captive. Either would make the sum of the parts legible; without one of them the arithmetic can stay right and uncollected for another decade, as it has for the last one.
The case turns negative on the conjunction rather than on either leg alone. Two consecutive publications showing both a top-up to the electric-vehicle provisions and a reduction in the capital return would mean the drain has stopped being ring-fenced and has started consuming the motorcycle business's operating cash and the captive's distribution capacity. In that configuration the floor falls below ¥1,000, motorcycles are valued as a constrained asset rather than a rent, and the hidden-asset thesis is void. Neither condition is observed today.
Two things would require a complete re-underwriting. A new major capacity commitment — electrification, batteries, or an alliance — before the residual of the last one has burned down would repeat the error that cost ¥1,577.8bn; the signal is industrial capital expenditure accelerating beyond ¥1,038.6bn. And a competitive electric two-wheeler offer landing in India or ASEAN would attack the whole of the normative earnings power. That is the only event capable of altering the economics of the group irreversibly, and it is the one risk this framework cannot price.
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.