トピー工業株式会社（Topy Industries, Limited）

研究原文：Topy Industries, Limited (7231, TSE Prime and Nagoya Premier) is an integrated steel-to-parts group: electric-furnace steelmaking from scrap at Toyohashi, steel products, and downstream steel and aluminium wheels and construction-machinery undercarriage components, plus a small tail of synthetic mica, construction, property and sports operations. Founded 1921, formed in present shape by the 1964 merger, capital JPY 20,983m, 24,077,510 shares issued, 32 consolidated subsidiaries and 4 affiliates, 16 overseas production bases in seven countries. The parent alone earns only JPY 1,880m of operating profit, so consolidated statements are the only meaningful economic perimeter. Legal English name on the securities-report cover is Topy Industries, Limited.

内部研究类型：company_identity_perimeter

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Ordinary shareholders own a cyclical integrated manufacturer whose profit is earned in subsidiaries; analysis must start from the steel-to-parts chain, not from a pure parts-maker frame.

不确定性（原始记录）：Minority-interest and subsidiary-level attribution were not extracted; group boundary is read from filings only.

研究原文：What is durable is the frame, not the earnings level. Integration from scrap to qualified components, the parts segment's demonstrated ability to reprice, a 51-53% equity ratio, a sellable securities portfolio, a DOE-anchored return formula and an alliance anchor on the register have been observed in both bad steel years (FY3/22, H1 FY3/27) and good ones (FY3/24, FY3/26) without depending on cycle direction. What is not durable is any particular year's profit: the five-year operating range of JPY -1,706m to +10,440m, reported net income's dependence on portfolio sales, and a temporary element inside the latest parts figure all say earnings will keep moving with the scrap spread and the machinery cycle. The structural judgment is two-level: own the frame for balance-sheet compounding and parts repricing ability, but price earnings at through-cycle levels (roughly JPY 5-6bn operating profit) until steel holds at least breakeven through a full scrap cycle and the parts margin survives a negotiation round without accounting assistance.

内部研究类型：durable_business_portrait

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Separates what compounds (book value, dividend formula, repricing franchise) from what oscillates (spread-driven profit), which is the basis for valuing the equity at through-cycle rather than peak or trough earnings.

不确定性（原始记录）：Parts margin durability has been observed for only about two years and has not been tested through a full OEM negotiation round or a full scrap cycle under the new pricing framework.

研究原文：FY3/26 operating profit rose 46.8% to JPY 7,783m on sales of JPY 297,750m (down 1.0%; a 2.6% consolidated margin), but the entire increase came from the parts segment (profit up 145.5% to JPY 10,917m at a 5.4% margin) while steel profit fell 61.1% to JPY 2,470m (2.8% margin) on scrap costs and repair expense. The headline growth is therefore a composition story: a downstream margin programme replaced steel as the group's engine.

内部研究类型：earnings_composition_finding

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Group profit capacity now depends on one franchise (parts) carrying an unmoated, spread-exposed one (steel); every point of segment margin moves a thin consolidated margin materially.

不确定性（原始记录）：Repair expense inside steel and the split between structural reform and repricing inside parts are not separately quantified.

研究原文：Reported FY3/26 net income of JPY 10,151m includes a JPY 7,143m one-off gain on sales of policy shareholdings (partly offset by impairment of JPY 754m and disposal losses of JPY 635m); ordinary profit of JPY 8,618m is the cleaner recurring anchor, including about JPY 835m of recurring dividend and equity-method income. The revised FY3/27 ordinary-profit forecast is JPY 8,600m, essentially flat, while net income is guided down because no shareholding sale is assumed. Shareholding-sale gains have now supplemented reported profit in three consecutive reporting periods (a further JPY 1,371m in H1 FY3/27), so portfolio sales are becoming a habit but remain finite in supply. Three separate mechanisms flatter reported profit - portfolio sales, impairment offsets and a pass-through accounting change - and each must be stripped to see the operating trend.

内部研究类型：profit_quality_structure

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Valuation and return judgments should be anchored to ordinary profit and segment run rates, not reported net income or the reported ROE.

不确定性（原始记录）：The size of recurring versus finite portfolio gains going forward depends on the remaining saleable holdings and market prices.

研究原文：Reported FY3/26 ROE of 7.3% clears the plan's FY2027 target of at least 6.0% only because JPY 7,143m of securities-sale gains lifted it; management itself says the ex-gain level still needs to rise. Ordinary profit of JPY 8,618m on about JPY 145bn of equity is roughly a 6% pre-tax return on equity, so the target is a floor still to be defended, not a milestone passed. No ROIC series was available and the company's March 2026 cost-of-capital response document was identified but not extracted.

内部研究类型：capital_efficiency_judgment

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Until recurring ROE clears the target without asset sales, the discount to book has a rational core; sustained sub-cost-of-capital returns justify a sub-1x multiple.

不确定性（原始记录）：Pre-tax/after-tax and period-basis approximations are the review's own; ROIC and the quantified cost-of-capital response remain unknown.

研究原文：H1 FY3/27 flipped the mix again: sales grew 9.7% and operating profit rose 10.9% to JPY 3,611m, more than twice the revised half-year guide of JPY 1,500m, because parts reached a record JPY 7,411m (6.7% margin; external sales up 16.2%) while steel lost JPY 1,283m. JPY 2,918m of the half's profit arrived in Q2 alone as parts price adjustments landed, after a weak Q1 (JPY 693m) in which steel spreads were hit by a scrap surge and parts strength led the company to raise the half-year guide in August. Parts profit includes a temporary pass-through-method gain the company flags as one-off.

内部研究类型：current_change_finding

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：The half-year beat is real but its quality attaches to the parts figure rather than to demand, so it supports the forecast raise without proving run-rate margins.

不确定性（原始记录）：Composition of the Q2 parts profit between volume, currency and the temporary adjustment is undisclosed.

研究原文：Steel profitability is a metal-spread problem, not a volume problem. The electric-furnace segment buys scrap at market prices and sells into negotiated prices, so a high scrap plateau compresses the margin until selling-price revisions catch up. FY3/26 steel profit fell 61.1% as scrap rose in the second half; in H1 FY3/27 steel external sales fell only 3.4% while profit swung JPY 2,779m (from +1,496m to -1,283m) year on year, a spread event on stable construction demand.

内部研究类型：causal_mechanism

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Steel has no moat and no supplier-concentration rescue; only price pass-through protects margin, so steel is an option on spread normalisation rather than a foundation.

不确定性（原始记录）：The scrap level at which steel returns to profit and the company's price-increase assumptions are undisclosed.

研究原文：The parts segment is the earnings engine: JPY 10,917m of FY3/26 operating profit (about 76% of JPY 14,269m segment profit, after steel JPY 2,470m and other JPY 882m) and JPY 7,411m in H1 FY3/27, more than the whole consolidated H1 profit, against a steel segment that swung to loss. The margin reset results from structural reform and a 'sustainable selling price' policy rather than volume growth, so durability depends on holding the new price formation. Parts is about two-thirds of reported segment revenue but a larger share of profit, so mix shift toward downstream is a quiet structural trend inside the cycle.

内部研究类型：profit_engine

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Any failure of the new price formation would hit group profit immediately because steel cannot currently offset it.

不确定性（原始记录）：Underlying run-rate margin below the reported 6.7% is unquantified; durability through an OEM negotiation round is untested.

研究原文：A change in the method of reflecting raw-material prices in selling prices to OEM customers produced a temporary adjustment gain inside H1 FY3/27 parts profit of JPY 7,411m. The company names three drivers (volume, currency and this adjustment) and flags the adjustment as one-off, but does not quantify it, so the segment's underlying half-year margin is somewhere below the reported 6.7%. This is the key measurement gap in the current record.

内部研究类型：accounting_timing_effect

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Segment margins for H1 FY3/27 should be read as overstated by an unknown amount until the annual report resolves it; it bears directly on whether the parts reset is structural.

不确定性（原始记录）：Magnitude unknown; may be resolved by the annual securities-report segment note or briefing materials.

研究原文：The FY3/27 plan was deliberately second-half-loaded from the start (opening H1 guide JPY 1,000m of a JPY 8,000m full-year operating profit; 87.5% in H2). The November 2026 revision to JPY 8,500m operating profit arithmetically requires JPY 4,889m in H2, only 8.0% above H2 FY3/26's JPY 4,527m, with ordinary profit roughly flat at JPY 8,600m. Management frames the means as price increases and structural reform with parts volumes holding, and flags energy costs as the offsetting risk; steel must move from the JPY 1,283m H1 loss toward breakeven or better. A weak Q1 followed by strong Q2 is the plan's designed rhythm, so evidential weight sits on the full-year bridge and its steel-price assumptions rather than any single quarter.

内部研究类型：forecast_bridge

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：The market has just been taught to trust a guide whose delivery depends on the scrap-versus-price race, so H2 delivery is the next proof point.

不确定性（原始记录）：No segment-level bridge or scrap/selling-price assumptions are disclosed.

研究原文：Roughly JPY 6-6.5bn a year disappears between segment profit (JPY 14,269m) and consolidated operating profit (JPY 7,783m): the corporate centre and adjustments consume about 45% of what segments earn, a drag of about 2.2% of sales. H1 FY3/27 annualises to the same magnitude, confirming it is structural, not a one-year event. Even a good 5.4% best-segment margin leaves only a 2.6% consolidated margin.

内部研究类型：structural_margin_drag

内部来源标记：COVERAGE35

普通股东含义（原始记录）：Operating leverage works both ways on a thin consolidated margin; the equity story is about spread and mix, not scale.

不确定性（原始记录）：Composition of the adjustment line (central costs versus eliminations) is not decomposed.

研究原文：The 'other' portfolio (synthetic mica, construction, property, sports) earned JPY 882m at a 12.8% margin on JPY 6,889m of sales, about 2% of sales: the highest-margin and least cyclical part of the group, too small to move the equity case but evidence that the group can run non-cycle businesses profitably.

内部研究类型：minor_profit_pool

内部来源标记：COVERAGE35

普通股东含义（原始记录）：A small stabiliser, not a thesis driver.

不确定性（原始记录）：Business-level economics not extracted.

研究原文：The earnings record is a cycle, not a trend: sales peaked at JPY 334,496m with FY3/23 steel-price inflation; operating profit swung from a JPY 1,706m loss (FY3/22) to a JPY 10,440m peak (FY3/24), a JPY 12.1bn range around a mean of roughly JPY 5.9bn. A segment-level normalisation (steel between zero and JPY 2,470m, parts JPY 7-11bn on the reset margin, other about JPY 0.9bn, less about JPY 6.5bn of corporate costs) brackets normalised operating profit at roughly JPY 5-7bn; averaging consolidated history would import the FY3/23-24 steel-price inflation as if it were repeatable. The bracket is wide because steel is binary around spread conditions; the parts reset narrows it only if it holds.

内部研究类型：normalized_earnings_judgment

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Peak-year multiples are meaningless; at roughly JPY 66-69bn of market value the equity costs about 10-13 times normalised operating profit - cheap if the parts reset is structural, full if not.

不确定性（原始记录）：The normalised build is this review's construction from segment history, not a company figure; segment capital data are not disclosed.

研究原文：Working capital absorbs cash when scrap prices rise. Inventories climbed from JPY 57,509m at the FY3/26 year-end to JPY 64,941m at H1 FY3/27 (up JPY 7,432m, mostly raw-material value), with receivables of JPY 46,544m against payables of JPY 29,305m; total assets grew JPY 13,694m and the equity ratio eased to 51.0%. Borrowings rose about JPY 19.1bn while JPY 7,000m of bonds were redeemed, refinanced into bank debt. FY3/26 operating cash flow of JPY 13,129m was strong, but the H1 build shows the same mechanism in reverse. The scrap plateau therefore taxes the balance sheet as well as the margin. Half-year operating cash flow was not disclosed, so the cash cost is inferred from balance-sheet movement; inventory build here reflects raw-material price inflation rather than evidence of demand.

内部研究类型：cash_transmission_mechanism

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Cash conversion, not demand, is the constraint to watch in a rising scrap market; reported profit and cash diverge when scrap moves.

不确定性（原始记录）：H1 cash-flow statement unavailable; inventory costing basis and any hedging not disclosed; ownership/realisability of the inventory not examined.

研究原文：Equity ratio 53.0% at FY3/26 (51.0% at H1 FY3/27), rising from 39.9% over five years; net assets JPY 146,843m; gross interest-bearing debt about JPY 58.7bn against cash of about JPY 29.4bn at FY3/26 (net debt near JPY 29bn), rising to about JPY 39.4bn net at 30 September 2026 after the inventory build and bond refinancing - still modest against equity. Investment securities of JPY 26,713m (JPY 28,680m at H1) and land at JPY 15,244m book give non-operating backing, and the market-value equity ratio was 23.1% at FY3/26. Book value per share compounded from JPY 6,660.64 to JPY 6,863.81 in six months despite a loss-making steel half.

内部研究类型：financial_resilience_structure

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：A thick equity layer absorbs the cycle rather than transmitting it to solvency, which underwrites the dividend formula and buyback.

不确定性（原始记录）：Maturity ladder of the remaining JPY 18,000m of bonds and covenant detail not extracted; consolidated cash availability to the parent not analysed.

研究原文：Land is carried at JPY 15,244m book value (head office and plant sites at historical cost); its market value is not disclosed in the sources used, so it is an unquantified option that historical-cost accounting leaves off the valuation table. Investment securities, by contrast, are fully observable at fair value, with a JPY 13,239m unrealised-gain reserve inside net assets.

内部研究类型：unquantified_asset_option

内部来源标记：COVERAGE35

普通股东含义（原始记录）：At about 0.46 times book the market pays for neither block at full value; the securities alone approximate a large fraction of market capitalisation, while land realisability and tax leakage on any monetisation are unknown.

不确定性（原始记录）：No appraisal, encumbrance or tax-basis information; realisability and management willingness to monetise are unproven.

研究原文：Policy shareholdings stood at 11.3% of consolidated net assets at March 2026 (14.8% earlier); the sub-10% goal was missed on market prices and its deadline extended to March 2027. Disposals are real (JPY 7,143m gain in FY3/26, JPY 1,371m in H1 FY3/27, JPY 12,446m of FY3/26 proceeds recycled into debt repayment, dividends and buybacks) but the ratio lags because remaining holdings appreciated; investment-securities book value rose again to JPY 28,680m at September 2026. The programme is measurable in sales but not yet in the ratio.

内部研究类型：balance_sheet_conversion_programme

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：A source of both reported profit and balance-sheet optionality, but a second extension would damage the credibility of the conversion pattern.

不确定性（原始记录）：Whether the March 2027 target is met through disposals rather than price moves is open.

研究原文：Shareholder returns follow a stated formula: a progressive dividend with a DOE guide of about 2.5% during the plan, paid at JPY 85, 103, 130 and guided to JPY 135 for FY3/27 (a 37.9% payout on the revised EPS forecast) even though the initial FY3/27 profit forecast fell. Because it is anchored to book value, the dividend ratchets with the balance sheet rather than the steel cycle and has been executed through a loss-making steel half; behaviour in a consolidated loss year is untested since adoption. Financial policy is formulaic and defensive: 53.0% equity ratio, DOE-anchored dividend and an annual August buyback.

内部研究类型：capital_return_policy

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：The dividend is the most reliable number in the equity story and yields about 4.3-4.5% on the guide, paying holders to wait while earnings quality resolves.

不确定性（原始记录）：The successor plan's return framework after FY2027 is unknown; no fixed buyback amount is committed.

研究原文：Buybacks are a standing August instrument, not a signal: treasury stock rose from 8.4% to 12.0% of issued shares (2,900,013 shares) in eighteen months, shrinking the share count on which per-share metrics are computed and compounding book value per share independently of the steel cycle. Share-count denominators therefore matter in per-share valuation (issued 24,077,510 versus ex-treasury).

内部研究类型：per_share_compounding_mechanism

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：The share count only falls; dilution is not part of this story.

不确定性（原始记录）：Buyback prices versus book and funding source (partly securities proceeds) not analysed.

研究原文：Nippon Steel holds 23.0% of voting rights (22.1% direct plus 0.9% aggregated; 22.01% of shares excluding treasury) as an equity-method alliance partner. It is simultaneously a register stabiliser, a supplier of part of the raw material and a customer for toll rolling, giving it commercial ties in both directions with the steel business, yet the filing states it holds no approval right over management. Ownership, governance and the steel segment's commercial position therefore cannot be analysed apart from each other.

内部研究类型：anchor_shareholder_relationship

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Aligns the steel business commercially with Japan's largest steelmaker without conferring control; the anchor's own portfolio policy is the one slow variable.

不确定性（原始记录）：Alliance economics (volumes, pricing terms) and any change in the holding policy are not disclosed.

研究原文：The 23.0% commercial-alliance holding stabilises the register and caps the probability of register-driven events; hostile or abrupt register change is unlikely, so the governance debate is operational rather than about control. The rising treasury share quietly concentrates remaining holders.

内部研究类型：register_stability_judgment

内部来源标记：COVERAGE35

普通股东含义（原始记录）：Register-driven catalysts are unlikely; re-rating must come from earnings and capital policy.

不确定性（原始记录）：Large-shareholding reports and the full register were not examined beyond the anchor.

研究原文：Governance is moving in the direction the market asks of low price-to-book companies: transition to an audit-and-supervisory-committee structure announced March 2026, a published and updated cost-of-capital response, takeover defence abolished May 2025, active subsidiary portfolio changes and the annual buyback. It supports the re-rating case at about 0.46 times book but does not change near-term earnings.

内部研究类型：governance_direction

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：A mild positive for the multiple, not a thesis driver.

不确定性（原始记录）：Post-transition board independence and skill composition not extracted.

研究原文：Tokyo Stock Exchange pressure on companies trading below book value addresses a population this stock sits squarely in at about 0.46 times book; the updated cost-of-capital response, the buyback cadence and the cross-shareholding programme are the company's documented answers.

内部研究类型：policy_context

内部来源标记：OPEN_DISCOVERY

普通股东含义（原始记录）：Provides external pressure that can convert balance-sheet assets and capital discipline into a narrower discount.

不确定性（原始记录）：Content of the March 2026 response not extracted; policy pressure is a slow, non-dated catalyst.

研究原文：The market prices the balance sheet at about half its book value and mid-cycle earnings at a full multiple: at JPY 3,130 after the November revision the equity was valued at roughly JPY 66-69bn (66bn on ex-treasury shares), about 0.46 times H1 book value per share of JPY 6,863.81 and about 8.8 times the revised FY3/27 EPS forecast of JPY 356.04 (6.7 times FY3/26 EPS, which contains one-off gains), with a dividend yield of about 4.3%. The +5.03% one-day response credited the revision without re-rating the structure. Embedded expectations are low and specific (deliver the revised forecast, keep the dividend formula, no steel surprise); the market believes the balance sheet and doubts the earnings, which matches the evidence. The price pays for balance-sheet repair and parts durability but not for a steel recovery or a failure of the H2 plan, and the gap closes only if through-cycle earnings rise or balance-sheet conversion accelerates.

内部研究类型：valuation_expectations_judgment

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：The priced-in future is the dull one; both the steel option and balance-sheet conversion sit outside it, so shareholders are paid to wait while the earnings-quality question resolves.

不确定性（原始记录）：Market expectations are inferred from price and multiples, not analyst consensus; peer multiples were not pulled; the FY3/26 P/E is shown for reference only.

研究原文：Valuation snapshot at the research cut-off (close of 6 November 2026; cut-off 9 November 2026): share price JPY 2,980; market capitalisation JPY 71.75bn (on 24,077,510 shares issued, including treasury shares); forecast PER 8.4x on revised FY3/27 EPS of JPY 356.04; actual PBR 0.45x on FY3/26-end net assets per share of JPY 6,660.64; FY3/26 reported ROE 7.3% (includes policy-shareholding sale gains); forecast dividend yield 4.53% on the FY3/27 JPY 135 forecast; EV/EBITDA 5.7x (EV JPY 111.18bn = market cap plus net interest-bearing debt of JPY 39,427m at 30 September 2026, being short-term borrowings JPY 32,235m, bonds JPY 18,000m and long-term borrowings JPY 20,645m less cash JPY 31,453m; EBITDA JPY 19.56bn = FY3/26 operating profit JPY 7,783m plus depreciation JPY 11,777m). Basis differences versus the post-reaction view (JPY 3,130 on 9 November; 0.46x on H1 book value; ex-treasury market value about JPY 66bn) must be kept distinct: different price date, share-count basis and book-value date. The EV/EBITDA mixes a September 2026 net-debt figure, inflated by the scrap-driven inventory build, with FY3/26 EBITDA.

内部研究类型：valuation_snapshot_with_basis

内部来源标记：OPEN_DISCOVERY

普通股东含义（原始记录）：Reference multiples for the point-in-time price; the discount to book and mid-single-digit EV/EBITDA frame the margin of safety but rest on mixed period bases.

不确定性（原始记录）：Multiples mix price dates, share counts (issued versus ex-treasury), book-value dates and a period-mismatched EBITDA; they are derived inferences, not primary facts.

研究原文：The named growth fields - aluminium wheels in India and Indonesia (lightweighting plus new geographies), forged aluminium wheels for commercial vehicles and recycling - carry no disclosed capex, capacity or return figures, so they are options on the story rather than part of the priced earnings base. Technology is cumulative process metallurgy (electric-furnace steelmaking, integrated rolling and forming, aluminium light-weighting), without R&D spending figures in the record; it supports the mix-shift story but not a growth multiple.

内部研究类型：unpriced_growth_option

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：The equity case does not require a technology breakthrough; any success is upside outside the price.

不确定性（原始记录）：Volumes, capacity, returns, alliance terms and R&D spending are unknown.

研究原文：Macro reaches this P&L through prices, not volumes. Scrap prices (steel spread and inventory values), Middle East-linked energy prices (electric-furnace costs, flagged by the company for H2 FY3/27), exchange rates (a named contributor to H1 parts profit) and US trade policy are the four documented channels. Scrap is bought at market with some raw-material purchases from Nippon Steel; selling prices catch up only through negotiation, and that lag is the transmission mechanism. US tariff exposure on parts was negotiated to limited impact in FY3/26, removing one feared drag.

内部研究类型：external_input_price_exposure

内部来源标记：OPEN_DISCOVERY

普通股东含义（原始记录）：Macro monitoring for this stock means watching the scrap price, energy and the yen, not GDP.

不确定性（原始记录）：Quantified sensitivities, scrap procurement terms, hedging and the FY3/27 tariff outcome are undisclosed.

研究原文：Construction-machinery production drives demand for undercarriage components, the parts segment's largest current growth source; it drove the H1 FY3/27 beat and carries the second-half plan, while passenger-car wheel volumes had declined in FY3/26. Demand is derived from OEM production schedules, cyclical with machinery and construction cycles, and the demand that carries the current forecast is the same demand that carried the last peak.

内部研究类型：end_market_driver

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Concentration in one cycle is the exposure behind the parts engine.

不确定性（原始记录）：No customer concentration, order backlog or OEM production plans for H2 are disclosed.

研究原文：The one genuine moat sits in components: wheels and undercarriage parts are qualified into OEM programmes, switching costs sit with the customer, and incumbency persists, but the rent is collected only through periodic price negotiation, so the pricing framework matters as much as qualification. Steel is undifferentiated and priced against the scrap spread; the moat does not extend there. Valuation should not average a moated parts business with an unmoated steel business.

内部研究类型：competitive_position_judgment

内部来源标记：COVERAGE35

普通股东含义（原始记录）：Pricing discipline once established with OEMs is a defensible asset; the steel side has no such protection.

不确定性（原始记录）：No competitor financials or third-party market-share data were pulled; position is stated qualitatively from company disclosure, and the framework has not survived a full negotiation round.

研究原文：FY3/26 revenue fell 1.0% on price and mix (steel down, parts up 5.7% on repricing); H1 FY3/27 revenue rose 9.7% led by parts volumes (external +16.2%) while steel external sales fell 3.4% despite price increases. Segment totals include intersegment sales (about JPY 10.7bn in steel), so shares are indicative. Each yen shifted from steel products to qualified components carries roughly double the margin, so flat group sales can still mean rising profit capacity.

内部研究类型：revenue_quality_finding

内部来源标记：COVERAGE35

普通股东含义（原始记录）：Judge revenue by segment composition, not growth rate.

不确定性（原始记录）：External-only segment revenue for the full year and sub-segment volume/price splits are not disclosed.

研究原文：FY3/26 operating cash flow of JPY 13,129m covered capital expenditure of JPY 9,911m (below depreciation of JPY 11,777m, a harvesting posture), but investing cash flow was positive only because of JPY 12,446m of securities-sale proceeds, and financing absorbed JPY 13,678m of debt repayment, dividends and buybacks. Free cash flow after capex is real but modest; the generous FY3/26 returns were co-funded by asset sales. The H1 FY3/27 release announced a new Sanwa Parts crawler-works plant to optimise undercarriage production - capital following the parts margin.

内部研究类型：cash_and_capital_deployment

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Operations fund the factory and the portfolio funds deleveraging; whether FY3/27 operating cash covers capex and the raised dividend without further asset sales is unknown.

不确定性（原始记录）：Plant cost, capacity and savings undisclosed; H1 FY3/27 cash flow not in the interim release.

研究原文：Three regimes: 2021-22, pandemic and steel downturn with consecutive operating losses and a price low of JPY 990 in April 2022; FY3/23-24, steel-price inflation carrying sales to JPY 334,496m and operating profit to the JPY 10,440m peak; FY3/26, the parts margin reset (profit +145.5%) replacing steel as the engine while portfolio sales began funding returns. The equity ratio's climb from 39.9% to 53.0% is the financial signature. The parts-led configuration is the newest and least tested.

内部研究类型：historical_turning_points

内部来源标记：COVERAGE35

普通股东含义（原始记录）：History argues for judging the current regime on its own evidence (about two years) rather than the longer steel-led average.

不确定性（原始记录）：Intra-year timing follows release calendar; regime durability unproven.

研究原文：Downside layers: a 51-53% equity ratio; cash of about JPY 28-31bn plus a sellable securities portfolio of JPY 28,680m; a dividend anchored to book value rather than earnings; a parts business that stayed profitable through the steel downturn; an anchor shareholder with commercial reasons to hold. Counterweight: the register produced operating losses in FY3/21 and FY3/22, so protection is financial, not operational. A bad year costs earnings and multiple, not solvency or the dividend formula, provided the securities portfolio remains saleable near book (untested in a disorderly market).

内部研究类型：downside_protection_layers

内部来源标记：COVERAGE35

普通股东含义（原始记录）：Asymmetry supports holding through the cycle rather than trading it.

不确定性（原始记录）：DOE dividend in a consolidated loss year is untested; securities saleability under stress and cash availability at the parent unknown.

研究原文：Each catalyst works on a different term of the valuation: Q3 FY3/27 and H2 delivery against the JPY 4,889m bridge (earnings proof); the successor to 'Active & Challenge 2027' (framework, including whether ROE ambition exceeds 6%); completion of the sub-10% cross-shareholding target by March 2027 (balance-sheet conversion); any quantification of aluminium-wheel expansion; continued buybacks (per-share compounding).

内部研究类型：catalyst_calendar

内部来源标记：COVERAGE35

普通股东含义（原始记录）：The investor is not waiting on a single event; proof points arrive quarterly.

不确定性（原始记录）：Timing beyond the disclosure calendar is uncertain.

研究原文：Scenario value is concentrated in one interaction: if parts pricing durability (a business-model fact) holds while the steel spread merely stops worsening (a market fact), consolidated operating profit approaches the FY3/24 peak with a far stronger balance sheet than the one that produced it. The re-rating case is the product of the two domains, not of either alone; if the parts reset proves accounting and cycle rather than structure, the stock returns to pure spread-cycle pricing.

内部研究类型：scenario_value_interaction

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Identifies the single interaction on which upside and downside scenarios turn.

不确定性（原始记录）：The scenario is conditional and unquantified; neither leg has been observed together through a full cycle.

研究原文：Thesis: Topy has quietly changed which business carries it - a repriced, qualified parts franchise now funds a DOE-anchored, book-compounding return stream on a 51-53% equity-ratio balance sheet, while steel and the securities portfolio are options rather than foundations - and at about 0.46 times book the market still prices the old steel-led company. Evidential weight: one demonstrated fact (parts repricing power), one demonstrated programme (balance-sheet conversion), one unproven option (steel spread normalisation). Falsifiers: (1) the H2 bridge misses with steel still loss-making at Q3; (2) parts margin falls back below about 4% at the next OEM negotiation round or as the pass-through effect annualises out; (3) the cross-shareholding ratio is not below 10% by March 2027 through disposals; (4) on price, a monthly close below JPY 2,615 and decisively below the 60-month average near JPY 2,234. Confirmation: H2 at or above the revised guide with steel at least breakeven, a successor plan raising recurring-ROE ambition, and a monthly close above JPY 3,710. If falsifiers stay silent, expected path is book-value compounding plus a 4.3% yield, with re-rating as upside.

内部研究类型：thesis_and_falsification

内部来源标记：COVERAGE35 / OPEN_DISCOVERY

普通股东含义（原始记录）：Defines what would confirm or break the holding case; price falsifiers are retained as stated in the Technical-aligned research record.

不确定性（原始记录）：Each falsifier is observable within about two quarters except the negotiation-round test, which resolves in the successor plan's first year.

研究原文：Material evidence boundaries carried as Research Debt across domains: no competitor financials or margin benchmarks (the thinnest domain); no third-party market-share data; no R&D expenditure figures; no land appraisal; the March 2026 cost-of-capital response not extracted; no analyst consensus or peer multiples; internal transfer economics between steel and parts and alliance terms with Nippon Steel undisclosed; bond maturity profile for the remaining JPY 18,000m and board composition not extracted; scrap procurement terms and any hedging undisclosed; FY3/27 tariff outcome not yet observable; segment capital data for returns-based normalisation absent; ROIC series absent.

内部研究类型：research_debt_evidence_boundaries

内部来源标记：COVERAGE35

普通股东含义（原始记录）：These gaps bound confidence on competitive position, relative valuation and capital returns without being negative evidence.

不确定性（原始记录）：Several are resolvable through the securities report, integrated report or a dedicated peer pull; none has a reasonable current public path in the supplied record.

研究原文：How large is the temporary adjustment gain inside H1 FY3/27 parts profit of JPY 7,411m from the change in raw-material pass-through method, and what is the segment's underlying half-year run rate without it? If the gain is large, the 6.7% margin overstates the reset and falsifier (2) moves closer.

内部研究类型：research_debt

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Decides whether parts durability, the core thesis leg, is structural.

不确定性（原始记录）：Unresolved; next evidence: FY3/27 annual segment note and any briefing materials.

研究原文：What scrap-price and selling-price assumptions underpin the revised forecast's implied H2 steel result, and at what scrap level does steel return to profit? Within the JPY 4,889m H2 operating profit, how much comes from steel spread recovery versus parts volumes, and is a segment-level bridge available?

内部研究类型：research_debt

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Determines whether the H2 guide is conservative or dependent on steel recovery, i.e. falsifier (1).

不确定性（原始记录）：Unresolved; next evidence: Q3 FY3/27 print and briefing materials.

研究原文：What does the successor plan after 'Active & Challenge 2027' target for operating profit and ROE, and does the progressive-dividend DOE framework survive into it?

内部研究类型：research_debt

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Sets whether recurring ROE ambition rises above 6% and whether the dividend formula persists.

不确定性（原始记录）：Future disclosure; no current public path.

研究原文：Will policy shareholdings fall below 10% of consolidated net assets by March 2027, given investment-securities book value rose again to JPY 28,680m by September 2026 on market prices?

内部研究类型：research_debt

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：A further missed target would damage the credibility of the balance-sheet conversion pattern (falsifier 3).

不确定性（原始记录）：Depends on disposal pace and market prices through March 2027.

研究原文：What volumes, capacity and returns are expected from aluminium-wheel expansion in India and Indonesia and forged aluminium wheels for commercial vehicles, and what does the new Sanwa Parts crawler-works plant cost, what production does it consolidate, and what saving or capacity gain does it deliver? Without these, capital deployment returns remain unassessable.

内部研究类型：research_debt

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Determines whether incremental capital earns above cost of capital.

不确定性（原始记录）：No disclosed figures; next evidence: integrated report, annual report and plant announcements.

研究原文：Is any change expected in Nippon Steel's 23.0% holding or the business alliance, in a sector where anchor shareholders have been reducing cross-holdings?

内部研究类型：research_debt

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：A change would alter register stability and steel-segment commercial flows.

不确定性（原始记录）：No current public path; monitor large-shareholding reports.

研究原文：After borrowings rose roughly JPY 19bn in H1 FY3/27 while a JPY 7,000m bond was redeemed, what leverage level does management regard as appropriate for a company with a 53% equity-ratio zone?

内部研究类型：research_debt

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Bears on how far scrap-driven working capital can push gross debt before capital policy is constrained.

不确定性（原始记录）：Management leverage tolerance not disclosed.

研究原文：What does management regard as through-cycle operating profit of each segment at mid-cycle scrap and construction-demand levels, against the five-year consolidated range of JPY -1,706m to +10,440m?

内部研究类型：research_debt

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Would replace this review's constructed normalisation bracket with a company view.

不确定性（原始记录）：Management is unlikely to disclose segment through-cycle levels; the review's JPY 5-7bn bracket remains its own construction.

研究原文：Nippon Steel holds 23.0% of voting rights as an 'other related company' and Topy is its equity-method affiliate under a business alliance; the relationship stabilises and commercially aligns the steel business without conferring control.

内部研究类型：anchor_holder_equity_method_affiliate

内部来源标记：OPEN_DISCOVERY

研究原文：The steel segment buys scrap at market while its selling prices are negotiated, so scrap plateaus compress the spread until price revisions catch up - the single largest swing factor in consolidated profit.

内部研究类型：input_price_supplier_exposure

内部来源标记：OPEN_DISCOVERY

研究原文：Construction-machinery production drives undercarriage component volumes, which drove the H1 FY3/27 beat and carry the second-half plan.

内部研究类型：end_market_demand_driver

内部来源标记：OPEN_DISCOVERY

研究原文：The change in the raw-material pass-through method produced a temporary H1 gain, meaning part of the record parts profit is an accounting-timing effect rather than run-rate margin.

内部研究类型：qualifies_run_rate

内部来源标记：OPEN_DISCOVERY

研究原文：Policy shareholdings, held under a public commitment to fall below 10% of net assets, supply both reported profit (JPY 7,143m FY3/26 gain) and balance-sheet optionality, lifting net income and ROE above operating reality.

内部研究类型：inflates_reported_profit

内部来源标记：OPEN_DISCOVERY

研究原文：A DOE-anchored progressive dividend converts book-value growth into dividend growth and, with the recurring buyback, mechanically supports per-share metrics even in flat-earnings years.

内部研究类型：formula_supports_per_share

内部来源标记：OPEN_DISCOVERY

研究原文：At about 0.46 times book the company sits in the population the exchange's capital-cost reforms address; its cost-of-capital response and buyback cadence are its documented answers and a possible route to a narrower discount.

内部研究类型：external_policy_pressure_on_discount

内部来源标记：OPEN_DISCOVERY

研究原文：Parts profit, driven by vehicle and construction-machinery end demand, now exceeds the consolidated operating profit it feeds; its margin reset carries the group result and steel cannot currently offset a failure.

内部研究类型：drives_consolidated_profit

内部来源标记：OPEN_DISCOVERY

研究原文：The revised FY3/27 forecast requires steel to move from a JPY 1,283m H1 loss toward breakeven or better in H2 through price increases; the bridge is arithmetically modest but fully exposed to the scrap-versus-price race.

内部研究类型：forecast_dependence

内部来源标记：OPEN_DISCOVERY

研究原文：The 51-53% equity ratio and growing book value per share make the DOE dividend formula and annual buyback self-funding, so balance-sheet strength converts directly into shareholder return.

内部研究类型：capital_support_funds_returns

内部来源标记：OPEN_DISCOVERY

研究原文：The 23.0% alliance holding with commercial ties in both directions makes hostile or abrupt register change unlikely and anchors governance debate in operations rather than control.

内部研究类型：stabilises

内部来源标记：OPEN_DISCOVERY

研究原文：The same high scrap plateau that compresses the steel spread inflates inventory values and pushes borrowings up, so one market price moves the P&L, balance sheet and cash flow in the same direction at once.

内部研究类型：same_driver_transmits_to_cash_and_debt

内部来源标记：COVERAGE35 / OPEN_DISCOVERY

研究原文：Inventory and receivable repricing grew assets faster than retained profit, easing the equity ratio from 53.0% to 51.0% and raising net debt, though the base remains strong.

内部研究类型：erodes_equity_ratio

内部来源标记：COVERAGE35

研究原文：A structural JPY 6-6.5bn gap between segment and consolidated profit compresses even a strong parts year to a 2.6% consolidated margin.

内部研究类型：structural_deduction

内部来源标记：COVERAGE35

研究原文：The undisclosed size of the temporary adjustment is carried as the highest-priority Research Debt.

内部研究类型：unresolved_measurement_gap

内部来源标记：COVERAGE35

研究原文：The one-off securities gain is what lifts reported ROE to 7.3%, above the 6.0% target; ex-gain returns are near or below the target.

内部研究类型：flatters

内部来源标记：OPEN_DISCOVERY

研究原文：Securities-sale proceeds of JPY 12,446m went out as debt repayment, dividends and buybacks, so portfolio conversion co-funded FY3/26 returns while operations covered capex.

内部研究类型：proceeds_fund_deleveraging_and_returns

内部来源标记：OPEN_DISCOVERY / COVERAGE35

研究原文：Through-cycle operating profit of roughly JPY 5-7bn determines whether the equity at JPY 66-69bn is cheap or full; peak-year multiples are not usable.

内部研究类型：normalisation_basis_for_multiple

内部来源标记：COVERAGE35

研究原文：Book value per share of JPY 6,863.81 backed by an equity ratio above 50% gives a floor logic to the 0.46x multiple, but asset value should be read with tax leakage, realisability and parent-availability caveats unresolved.

内部研究类型：asset_backing_against_discount

内部来源标记：COVERAGE35

研究原文：OEM qualification and switching costs are the mechanism through which negotiated repricing is collected; the moat supports the parts reset only if the pricing framework survives negotiation rounds.

内部研究类型：supports_pricing_durability

内部来源标记：COVERAGE35

研究原文：Parts pricing durability is one of the two legs whose product defines the re-rating scenario.

内部研究类型：scenario_leg

内部来源标记：COVERAGE35

研究原文：Steel spread merely ceasing to worsen is the second leg of the re-rating scenario.

内部研究类型：scenario_leg

内部来源标记：COVERAGE35

研究原文：A missed H2 bridge with steel still loss-making at Q3 is falsifier (1) of the thesis.

内部研究类型：falsifier_source

内部来源标记：COVERAGE35

研究原文：Parts margin falling back below about 4% when the pass-through effect annualises out would show the reset was accounting and cycle, falsifier (2).

内部研究类型：falsifier_source

内部来源标记：COVERAGE35

研究原文：Cross-lane observation only: the fundamental lane's unresolved earnings-quality question (one-off gains in FY3/26, a flagged temporary gain in H1 FY3/27) is the same question the protected Technical State leaves open as 'continuation unconfirmed'. The fundamental positivity of the balance sheet and dividend does not rewrite that independent Technical State, and the Technical State does not alter the fundamental adjudication.

内部研究类型：cross_lane_question_alignment

内部来源标记：OPEN_DISCOVERY / TECHNICAL

研究原文：Cross-lane: the fundamental discount-to-book and re-rating optionality is a separate lane from the pure price state; the post-revision +5.03% response to JPY 3,130 is consumed as price-response evidence within the Technical lane, and fundamental cheapness may not upgrade a non-confirming Technical State. The Technical falsifiers (monthly close below JPY 2,615, the 60-month average near JPY 2,234) remain price-defined and are mirrored, not redefined, in the thesis.

内部研究类型：cross_lane_non_rewrite_boundary

内部来源标记：TECHNICAL

研究原文：Durable Business Portrait: integrated chain, one moated engine, balance-sheet compounding｜The company is an integrated scrap-to-component chain whose profit has moved downstream to a qualified, repriced parts franchise, carried by a conservative balance sheet and a book-value-anchored return formula; steel is a spread-exposed option and the earnings level is cyclical.

内部研究类型：durable_business_portrait_structure

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Own the frame for balance-sheet compounding and demonstrated repricing; value earnings at through-cycle levels.

不确定性（原始记录）：Parts durability tested for about two years only; steel viability through a full scrap cycle unproven.

研究原文：Two engines, two cycles: scrap spread squeezes steel while parts repricing carries the group｜Scrap costs hit steel immediately while selling prices lag; parts repricing and construction-machinery volumes more than offset it at the segment level; a structural corporate drag then converts a strong parts year into a thin consolidated margin.

内部研究类型：main_causal_chain

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Segment mix matters more than group scale; consolidated profit hinges on the parts margin surviving negotiation and steel stopping its losses.

不确定性（原始记录）：Parts underlying run rate unknown because of the unquantified temporary gain.

研究原文：Earnings-quality ladder: operating, ordinary, net profit and ROE｜Profit quality is the hinge between domains: the same one-off items that flatter reported earnings (shareholding-sale gains in FY3/26 and H1 FY3/27, the temporary pass-through gain in H1 parts profit) must be stripped to reach ordinary profit of about JPY 8.6bn and a recurring ROE near the 6% target.

内部研究类型：profit_quality_hinge

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Anchor valuation to ordinary profit and segment run rates, not reported net income or 7.3% ROE.

不确定性（原始记录）：Size of the temporary gain and of remaining saleable holdings.

研究原文：One scrap price moves P&L, inventory, debt and cash together｜A high scrap plateau compresses the steel margin, inflates inventory values by billions of yen, and lifts borrowings in the same period, so industry, working-capital and leverage domains are a single mechanism.

内部研究类型：feedback_transmission_chain

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Watch cash conversion and net debt alongside the spread; reported profit and cash diverge when scrap moves.

不确定性（原始记录）：No H1 cash flow statement; hedging and inventory costing basis unknown.

研究原文：Balance-sheet conversion loop: portfolio sales, deleveraging, DOE dividend and buyback｜Portfolio sales fund deleveraging and returns, the equity ratio rises, bank access stays unconditional, and a DOE-anchored dividend plus buyback compound book value per share independently of the steel cycle (BVPS JPY 6,660.64 to 6,863.81 in six months).

内部研究类型：capital_conversion_loop

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：The income stream has risen every year while operating profit oscillated; dependence on finite portfolio proceeds is the limit.

不确定性（原始记录）：Portfolio supply is finite and the March 2027 ratio target is pending; FY3/27 cash cover of capex plus dividend without asset sales unknown.

研究原文：FY3/27 second-half bridge and plan rhythm｜A designed back-loaded plan whose revised H2 requirement is only 8.0% above last year's H2 but depends on steel price follow-through and unquantified parts run rate.

内部研究类型：forecast_test_structure

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Q3 and full-year prints are the near-term proof or falsification of the thesis.

不确定性（原始记录）：No segment bridge or price assumptions disclosed.

研究原文：Valuation bridge: priced balance sheet, doubted earnings, normalised profit｜At about 0.46x book and 8.8x revised EPS the market believes the balance sheet and doubts the earnings; normalised operating profit of JPY 5-7bn makes the equity roughly 10-13x mid-cycle operating profit, cheap only if the parts reset is structural. Growth options and land are outside the price.

内部研究类型：valuation_bridge

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Holders earn about a 4.3% yield while waiting; re-rating requires rising through-cycle earnings or faster balance-sheet conversion.

不确定性（原始记录）：Mixed price dates, share counts and book dates across the multiples; no consensus or peer multiples.

研究原文：Thesis, scenario interaction and falsification conditions｜One demonstrated fact, one demonstrated programme and one unproven option, with four numeric or observable falsifiers.

内部研究类型：thesis_falsification_structure

内部来源标记：COVERAGE35 / OPEN_DISCOVERY

普通股东含义（原始记录）：Book-value compounding plus yield if falsifiers stay silent; spread-cycle repricing if falsifier (2) triggers.

不确定性（原始记录）：Negotiation-round test resolves only in the successor plan's first year.

研究原文：Research Debt carried forward｜The highest-value unresolved items are the parts run rate and the H2 composition; the rest bound confidence on competition, capital deployment and relative valuation.

内部研究类型：research_debt_cluster

内部来源标记：OPEN_DISCOVERY / COVERAGE35

普通股东含义（原始记录）：Resolution of the first two can move the thesis; the others mostly calibrate confidence.

不确定性（原始记录）：Most items depend on future disclosure rather than a reasonable current public path.

研究原文：Protected Technical Synthesis (RESEARCH_TRUTH.technical_state)

内部研究类型：protected_technical_synthesis

内部来源标记：TECHNICAL

不确定性（原始记录）：受保护的 Technical Research 结构；公开边界以其正文路由为准。

公开边界：受保护 Technical Research 仅保留可追踪身份，本索引不重写其内部语义。