1. Scorecard

2. Argument

The short version

Tianli runs fee-paying senior-high schools across mainland China. At HK$1.02, the shares trade at 0.559x tangible book, against RMB1.5608 of filing tangible book per share, and at 4.4x normalized earnings. Revenue and profit were at multi-year highs in the latest half, and the F-Score is 8/9. The bargain is real.

It is also a contract-controlled operating network whose compulsory-education perimeter was broken by regulation once already. The board suspended the interim dividend even as profit rose, and the group still carries a large interest-free receivable and guarantees for the deconsolidated schools. The shares qualify for BUY - BIG on the numbers, but mainland-China exposure caps fresh-money entry at Standard.

Why it qualifies

• At HK$1.02 the shares trade at 0.559x tangible book, versus RMB1.5608 per share, a 44.1% discount.

• The latest half produced RMB2,142.8m of revenue and RMB471.4m of profit, up 14.2% and 21.0%.

• FY2025 operating cash flow was RMB852.6m against RMB648.1m of profit and RMB403.6m of capex; the F-Score is 8/9.

• The debt load is manageable on the reported numbers: RMB2,535.8m of bank borrowings, 10.1x EBITDA interest coverage, and RMB1,432.7m of undrawn committed facilities.

• The chairman's vehicle bought 37,579,000 shares during the prior six months while the price was falling.

Why it is not a clean bargain

• The group controls its PRC operating schools through structured contracts rather than equity, and regulation forced the deconsolidation of 30 compulsory-education schools in 2021, with a RMB1,085.2m impairment.

• The deconsolidated Affected Business owed the group RMB873.3m interest-free and unsecured, while the group had provided RMB254.0m of guarantees for its bank facilities without consideration.

• The interim dividend was nil despite profit rising 21.0%; cash was RMB449.3m and first-half capex rose to RMB540.2m, leaving a current ratio of 0.416.

• The chairman is also chief executive, and his construction company builds schools on a cost-plus premium of 9% to 11%; first-half construction fees were RMB50.1m against RMB7.3m in the comparable period.

• The history includes RMB100.0m of impairment reversals in FY2025 and RMB81.9m in the latest half, so reported profit and tangible book still need to be read alongside the regulatory perimeter and asset quality.

How a minority holder gets paid

A minority holder gets paid through cash distributions and a re-rating, if the business continues to operate and the market stops treating the contract structure as terminal. The dividend is the first route, although the nil interim payment means it must be re-established rather than assumed. The larger payoff is mean reversion toward the RMB4.00 - RMB6.77 sell range, with the lower end the more credible exit because the upper end reflects a valuation regime the cohort may no longer receive. The controlling shareholder's 46.13% stake and prior buying provide alignment, not a promise.

The RMB4.00 - RMB6.77 (HK$4.67 - HK$7.91) sell range comes from 29 observations of Tianli's own P/TBV history, not from a forecast or takeover assumption. The lower end is credible if the business keeps operating and the multiple returns to its historical median. The top is less credible on current earnings power because it requires a return to the stronger valuation regime seen earlier in the cohort, so it is an optimistic endpoint rather than the base exit case.

What would change the view

• FY2026 results and the final dividend show whether the interim suspension was temporary or a change in distribution policy.

• The Affected Business receivable stops growing, is repaid, or receives an appropriate allowance, while the related guarantees continue to fall.

• Senior-high qualification and classification rules provide evidence that the structured contracts can continue to support the operating network.

• The next filing shows debt maturities being refinanced without exhausting cash or committed facilities, and the current ratio recovers.

• Related-party construction fees move back toward the cap trajectory and product mix, margins, and impairment reversals normalize.

Sources and gaps

The FY2025 annual report and Interim Report 2026 were read in full. The June to September 2026 exchange-announcement window and the post-February insider register were not reached directly, so those items remain identified as third-party or unverified where relevant.

3. Backup and sources

This last layer is the full detail: the sources, calculations, and open questions behind the argument above.

• Field: Company; Value: Tianli International Holdings Limited

• Field: Ticker; Value: SEHK:1773 (1773.HK), The Stock Exchange of Hong Kong Limited, Main Board

• Field: Analysis Date; Value: 2026-09-07

• Field: Current Price; Value: HK$1.02 (7 September 2026) = RMB0.873; market capitalization approximately HK$2.10bn (RMB1.79bn; US$269m)

• Field: P / TBV; Value: 0.559× (derived: price ÷ filing TBV/share of RMB1.5608 at 28 February 2026)

• Field: Verdict; Value: BUY, BIG

• Field: Position disclosure; Value: Long 2,000 shares, purchased at HK$0.915 per share; trade date is not shown in the supplied screenshot

• Field: Buy-Below; Value: RMB2.10 (HK$2.45), 1.35× TBV, 7.2-year own-history, qualified: above the 1.00× absolute ceiling, so the operative buy price is RMB1.56 (HK$1.82) at 1.00× TBV

• Field: Sell Range; Value: RMB4.00 - RMB6.77 (HK$4.67 - HK$7.91), 2.56× - 4.34× TBV, 29 observations over 7.2 years

• Field: Moat; Value: Narrow

• Field: Governance Risk; Value: High (2 capital-allocation warnings fired)

Primary Sources Reviewed

• Source: Annual Report; Detail: Annual Report 2025 (HKEX), incl. Independent Auditor's Report, Report of the Directors, Corporate Governance Report and audited consolidated financial statements, FY ended 31 August 2025, filed/published 2025-12-30

• Source: Latest Quarterly; Detail: Interim Report 2026 (HKEX), incl. Independent Review Report and interim condensed consolidated financial statements, Six months ended 28 February 2026, filed/published 2026-05-15

• Source: Proxy / MIC; Detail: Poll Results of Annual General Meeting held on 28 January 2026, AGM 2026, filed/published 2026-01-28

• Source: Earnings Release; Detail: Not separately obtained, the FY2025 results announcement and the H1 FY2026 results announcement were superseded by the full Annual Report 2025 and Interim Report 2026, both read in full

• Source: Insider Activity; Detail: Hong Kong SFO Part XV disclosure-of-interests record as reproduced in the Annual Report 2025 (as at 31 August 2025) and Interim Report 2026 (as at 28 February 2026). The standalone exchange disclosure-of-interests register for the period after 28 February 2026 was not reached

• Source: Other Key Sources; Detail: SEHK:1773 last traded price HK$1.02 (7 September 2026, 16:08 HKT) and market capitalization HK$2.13bn; SEHK:1773 quote HK$1.025, YTD -56.93%, 5-day -7.24%, market cap HK$2.14bn, analyst consensus 2 covering brokers; HKD/CNY mid-market rate 1 HKD = 0.8559 CNY (7 September 2026); Discloseable transaction: subsidiary to acquire 10% of an art-education group for RMB104m plus RMB140.8m capital injection, taking the holding to 51% (announced 9-10 June 2026); Share retirement of 13,227,000 repurchased shares (16 June 2026); resignation of Ms Zhang Xiao as joint company secretary (31 August 2026); Broker note maintaining an Outperform rating with FY2026-28 revenue estimates of RMB3,915m/4,225m/4,499m and adjusted net profit of RMB741m/821m/901m; Sector context: Hong Kong-listed private-education issuers, 16 of 21 down more than 70% from peak and 11 trading below HK$1; Peer valuation context: a Hong Kong-listed private-education comparable trading on a 2.95x trailing P/E and an 11.3% dividend yield (April 2026)

• Source: Data Quality; Detail: GOOD

Tianli International Holdings Limited, Value Investing Analysis

1. Executive Summary

Tianli International Holdings runs a network of fee-paying senior high schools in mainland China, 58 schools across 36 cities at the last annual reporting date, teaching roughly 54,000 high-school students during the year and about 60,000 at the start of the autumn 2025 term, and monetizes that network four ways: tuition and comprehensive education fees, sales of products into its own campuses, canteen and boarding logistics, and management and franchise fees from schools it operates for third parties. The verdict is BUY, BIG, with Governance Risk High and a structural-risk caveat a reader should weigh before the tier: the shares trade at HK$1.02 (7 September 2026) against a filing-derived tangible book value of RMB1.5608 per share (HK$1.82), a P/TBV of 0.559×, while the business behind that book grew revenue 14.2% and profit 21.0% in the most recent half year.

The gap between price and fundamentals is the whole story. In the six months to 28 February 2026 the group earned RMB471.4 million on RMB2,142.8 million of revenue; over the twelve months to that date it earned roughly RMB730.0 million, against a market capitalization of about HK$2.10 billion (RMB1.79 billion), a trailing multiple under 2.5×. The Piotroski F-Score is 8/9, the only failed signal being leverage, which rose as the group termed out short-dated bank debt. Operating cash flow was RMB852.6 million in FY2025 against RMB403.6 million of capital expenditure, and bank borrowings of RMB2,535.8 million are 76.3% of equity, comfortably inside the absolute debt filter. The margin of safety against tangible book is 44.1%. Right-of-use assets are 74.2% of tangible book value, so this is cheap against tangible book value rather than cheap against a liquidation floor.

The reasons to be careful are specific. The group does not own equity in its PRC operating schools; it controls them through structured contracts, and in 2021 that structure was legislated out from under the compulsory-education half of the business, forcing deconsolidation of thirty schools and a RMB1,085.2 million impairment. Those schools, the "Affected Business", remain entangled: the group is owed RMB873.3 million by them, interest-free and unsecured, has guaranteed RMB254.0 million of their bank borrowings for no consideration, and owes them RMB1,399.8 million. Two capital-allocation warnings fired: that continuing flow of capital and credit support to entities the group no longer consolidates, and the board's decision on 10 April 2026 to pay no interim dividend at all for the half year (against RMB5.78 cents a share a year earlier) in a period when profit rose 21%. The chairman, who is also chief executive, controls 46.13% of the shares and his own construction company builds the schools on a cost-plus-9-to-11% basis.

Data quality is GOOD. Both primary filings, the FY2025 annual report and the interim report to 28 February 2026, were read in full, and the litigation, debt, related-party and subsequent-events notes were reached in the primary documents. The gap is the six months since: exchange announcements from June to September 2026 and the substantial-shareholder register could not be opened directly, so that window rests on third-party summaries and is labeled as such throughout.

2. Business and Market Overview

The revenue base splits four ways. For FY2025 (year ended 31 August 2025), comprehensive educational services, tuition and the core teaching contract, were RMB1,868.0 million, up 7.0%; sales of products into the school network were RMB992.3 million, up 8.0%; comprehensive logistical services (canteen, boarding, campus services) were RMB613.0 million, up 2.6%; and management and franchise fees from 18 entrusted schools were RMB115.6 million, up 93.9% on the addition of eight schools under management (Annual Report 2025, Financial Review). Total revenue was RMB3,588.9 million, up 8.1%. In the half year to 28 February 2026 the mix shifted sharply: product sales rose 45.6% to RMB691.3 million while comprehensive educational services grew only 3.4% to RMB1,046.6 million and logistics fell 1.7%. Management and franchise fees rose 38.7% to RMB78.2 million. That mix shift matters, because product sales carry a visibly thinner margin, gross margin fell 2.4 points to 35.2% in the half, which management attributes directly to product revenue growing 45.6% against procurement cost growing 49.6% (Interim Report 2026, Management Discussion and Analysis).

The customer is the fee-paying family, so there is no customer concentration in the industrial sense; the concentration risk sits on the other side, in geography and regulation. The group is centered on Sichuan and spans 18 provinces and municipalities. There is one reported operating segment. Capital intensity is high and rising: owned property, plant and equipment was RMB5,046.4 million and right-of-use assets RMB2,379.5 million at 28 February 2026, together 74.1% of the RMB10,020.1 million balance sheet, and capital expenditure in the half year was RMB540.2 million against RMB218.1 million a year earlier. Revenue is seasonal, the company lists seasonality among its principal risks, and the fiscal year ends 31 August, aligning to the academic cycle.

Regulatory exposure is the defining feature, not a footnote. Senior high school is post-compulsory education and may lawfully be run for profit; compulsory education (grades 1 to 9) may not, and since 1 September 2021 private compulsory schools have also been barred from transacting with related parties. That single change removed thirty schools from the consolidation. The group's own regulatory commentary is candid that there are substantial uncertainties regarding the interpretation and application of the Implementation Regulations, and that national and local classification-management rules have still not been issued (Annual Report 2025, Regulatory Updates).

As of 9 to 10 June 2026: a subsidiary agreed to acquire 10% of an art-education group for RMB104 million and to inject a further RMB140.8 million of capital, taking the holding to 51%. The announcement itself was not opened; this is a third-party account and is not independently verified against the filing.* Alongside it, management is pushing an AI-education strategy, a self-developed large model that passed the national generative-AI service filing in June 2025, sold internally to cut teacher cost and externally through business, government and community channels.

Understandable? Yes, it is a school operator. Predictable? At the operating level, unusually so: enrolment is contracted a year at a time and prepaid, which is why contract liabilities of RMB952.7 million sit on the balance sheet. Cyclical? Not economically, but acutely policy-cyclical. Structurally advantaged? Only inside a regulatory perimeter the state has already moved once.

* Sourced from third-party reporting; not independently verified against source filings.

3. Moat, Competitive Position, and Industry Cycle

Against the five moat sources the evidence is mixed but not empty. Intangibles and brand are real and locally specific: the group has operated since 2002, and in the 2025 Gaokao roughly 90% of graduates from established schools cleared the undergraduate admission line and 58% cleared the first-tier line, with 399 offers from QS top-50 universities against 272 the prior year (Annual Report 2025, Management Discussion and Analysis). In a market where the product is measured once, publicly, by a national examination, verifiable results are the brand. Switching costs are high in the way that matters: a family that enrols a child in the first year of a three-year senior-high program is locked in by the examination cycle, not by contract. Cost advantage is asserted through a centralised, standardized management system that lets new campuses be opened to a template; the evidence for it is a gross margin that has held between 33.7% and 33.8% while revenue quadrupled from RMB884.4 million in FY2022 to RMB3,588.9 million in FY2025. Efficient scale operates city by city, a single high-quality private senior high school in a prefecture-level city faces limited direct local competition. Network effects are absent.

The classification is Narrow. It is a genuine moat producing genuine returns, return on equity was 21.8% in FY2025 and return on capital employed roughly 12.8% after tax, but it is licensed rather than owned. The group operates under approvals, permits and filings the state grants, through structured contracts rather than equity, and it does not own the underlying school sponsorship rights the way an owner of a factory owns the factory. A moat a regulator can drain with a single administrative instrument is narrow by construction, whatever its width in normal weather. The company itself lists eleven of its thirteen principal risks as some form of regulatory, licensing, capacity or admissions-standard exposure.

On the cycle, the operating and valuation cycles have separated, and this is the most important observation in the report. Operationally the group is not near a trough: high-school enrolment rose about 11% to roughly 60,000 at the start of the autumn 2025 term, revenue rose 14.2% in the most recent half, profit for the period rose 21.0% to RMB471.4 million, and full-time teacher headcount at self-owned schools rose from 2,446 to 2,610. The valuation cycle is a different matter. Across the Hong Kong-listed private-education cohort the de-rating has been near-universal, 16 of 21 listed names down more than 70% from their peaks and 11 trading below HK$1, and a listed peer was quoted on a 2.95× trailing earnings multiple and an 11.3% dividend yield in April 2026.* Tianli's own shares have fallen roughly 57% year to date and about 80% from the 12-month high of HK$5.05. The classification is Below mid-cycle: the business is at or near its own operating high while the multiple assigned to it sits at the bottom of its recorded range.

* Sourced from third-party market data and press coverage; not independently verified against source filings.

4. Management, Governance, and Capital Allocation

Mr Luo Shi, 52, founded the group in 2002, has been a director since January 2017 and has been both chairman of the board and chief executive officer since June 2018. That combination is an explicit, disclosed departure from code provision C.2.1 of the Corporate Governance Code, the only such deviation the company reports for FY2025. He also chairs the nomination committee. Mr Wang Rui, 44, has been chief financial officer, executive director and joint company secretary since January 2018, arriving from finance roles at Longfor, New Hope and China Vanke. Three independent non-executive directors sit on a board of eight, Mr Liu Kai Yu Kenneth (audit committee chairman, a chartered accountant and former exchange listing officer), Mr Yang Dong (an education academic) and Mr Cheng Yiqun (a PRC lawyer). Audit and remuneration are majority-independent and independently chaired; nomination is not.

As of 31 August 2026: Ms Zhang Xiao resigned as joint company secretary. The underlying announcement was not opened; this is a third-party account.* No reason has been verified from a primary source, and it is recorded here as an unexplained departure rather than assumed routine.

Insider ownership is high and, unusually, has been rising into the collapse. Mr Luo Shi's aggregate long position was 973,034,569 shares, 46.13%, at 28 February 2026, comprising 934,556,316 shares held through his wholly owned vehicle Sky Elite Limited, 30,000,000 held beneficially, 6,521,733 vested award shares and 1,956,520 held by his spouse (Interim Report 2026, Directors' and Chief Executive's Interests). Six months earlier, at 31 August 2025, Sky Elite held 896,977,316 shares (42.53%) and the aggregate family interest was 44.35% (Annual Report 2025, Substantial Shareholders' Interests). That is an increase of 37,579,000 shares, roughly 1.8% of the company, acquired by the controlling shareholder's vehicle between 1 September 2025 and 28 February 2026, while the shares fell from roughly HK$3.60 to under HK$3.00. Mr Wang Rui held 8,956,520 shares (0.42%), Mr Zhang Wenzao 1,702,000 (0.08%) and Mr Pan Ping 13,043,289 through his spouse (0.62%). The regime searched was the Hong Kong Securities and Futures Ordinance Part XV disclosure-of-interests record as reproduced in the interim and annual reports; the standalone exchange disclosure-of-interests register for the six months after 28 February 2026 could not be opened, so post-February insider transactions are not confirmed.

Pay is not the problem here. Total directors' and chief executive's remuneration was RMB12.0 million in FY2025, down from RMB27.1 million in FY2024, against profit of RMB648.1 million. Mr Luo Shi received RMB8.6 million, of which RMB7.3 million was the non-cash share-option charge and only RMB1.35 million cash pay; Mr Wang Rui received RMB2.5 million and the three independent directors RMB600,000 between them, unchanged. Running the share-based-compensation test explicitly: total equity-settled award and option expense was RMB25.5 million in FY2025 against net income of RMB648.1 million, a ratio of 3.9%, down from RMB41.4 million and 7.4% in FY2024. The ratio is below 10% and fell rather than doubled, so the numeric test does not fire.

Related-party dealing is where the scrutiny belongs, and the note was read in full in both primaries. Checking each relationship type rather than only the obvious one. Service and construction fees: Sichuan Nanyuan Construction, wholly owned by a holding company approximately 75.80% indirectly owned by Mr Luo Shi, builds and refurbishes the group's schools under a framework agreement running to 31 August 2027, priced at actual costs plus a premium in the range of 9% to 11%. The amounts were RMB223.0 million in FY2024, RMB60.7 million in FY2025 and RMB50.1 million in the six months to 28 February 2026 against RMB7.3 million in the comparable half, a 6.9 times increase, against annual caps that step down from RMB310 million (FY2025) to RMB210 million (FY2026) to RMB170 million (FY2027). Financial relationships: the group is owed RMB873.3 million by the deconsolidated Affected Business, unsecured and interest-free, and owes it RMB1,399.8 million on the same terms, repayable within one to three years; it also owes RMB54.3 million to Nanyuan Construction and RMB20.3 million to a subsidiary's minority holder. Guarantees: the group has guaranteed RMB254.0 million of the Affected Business's bank facilities across four schools and has no consideration received from those guarantees (Interim Report 2026, note 20(c)(5)). No nominee or back-to-back arrangement is disclosed. There is one contingent purchase obligation: the 20% minority in an acquired subsidiary may be put to the group on 7 September 2026 at not less than the acquisition-date valuation of its identifiable net assets.

Capital allocation over the lookback, consolidated into single totals rather than described program by program. Buybacks: the company repurchased 12,061,000 shares in FY2025 for HK$43.1 million (approximately RMB38.8 million) at prices between HK$3.13 and HK$4.10, a blended average of roughly HK$3.57, and a further 18,055,000 shares in the six months to 28 February 2026 for approximately RMB50.4 million. Across the two periods that is 30,116,000 shares for roughly RMB89.2 million (HK$99.0 million), a blended average of about HK$3.29, now marked at HK$1.02, a mark-to-market loss of roughly HK$68.3 million, or 69% of the capital deployed. Separately the award-scheme trustee sold 13,000,000 shares for HK$51.5 million during FY2025. Dividends: RMB9.68 cents per share declared for FY2025 (RMB5.78 interim plus RMB3.90 final), a 30% payout ratio held flat with FY2024, with RMB207.8 million of cash paid during FY2025. Then the break, on 10 April 2026 the board resolved to pay no interim dividend for the half year to 28 February 2026, against RMB5.78 cents a share a year earlier, in a period when profit for the period rose 21.0%. Mergers and acquisitions: no material acquisition or disposal in FY2025 or the interim period; RMB36.4 million of cash went to acquiring subsidiaries in FY2025, and a RMB8.8 million goodwill impairment was taken against a Gaokao-repetition school whose autumn 2025 enrolment fell after Sichuan changed its examination model. Capital programs reversed or impaired: the FY2021 impairment of RMB1,085.2 million on property and right-of-use assets occupied by the deconsolidated schools is still partly on the books, with RMB100.0 million reversed in FY2025 and RMB81.9 million in the interim half.

Working the five capital-allocation warnings one at a time:

1. Control without minority protection, NOT FIRED, borderline. The 46.13% stake is above the threshold, so the affirmative test applies rather than the percentage. The annual general meeting held on 28 January 2026 did put three director re-elections to a poll, one non-executive and two independent non-executive directors, each carried with between 99.74% and 99.97% support, alongside the auditor's re-appointment, the dividend, the general mandates and a special resolution amending the articles. A genuine annual director election is in the record, which is the mitigant the test looks for. Against it: chairman and chief executive are the same person, the nomination committee is chaired by that person, and control over the operating schools runs through contracts rather than equity. Borderline, not fired.

2. Capital routed to parent-mandated, policy-driven or low-return projects, FIRED. The interest-free receivable from the deconsolidated Affected Business has risen from RMB686.9 million to RMB873.1 million over two years, with RMB183.9 million of fresh advances in the most recent half against RMB80.2 million a year earlier, alongside RMB254.0 million of guarantees given for nothing. This is shareholder capital and shareholder credit supporting entities from which shareholders take no economic benefit, and the flow is increasing rather than running off.

3. Dividend cut or suspension within roughly the last three years, FIRED. The interim dividend went from RMB5.78 cents a share to nil in a half year when profit rose 21.0%, with no explanation in the interim report beyond the bare resolution of 10 April 2026.

4. Non-answers to direct shareholder questions, UNASSESSABLE. No earnings-call transcript or shareholder question-and-answer record was reached in this run. Absence of evidence here is not evidence of absence.

5. Outsized pay or off-market related-party deals, BORDERLINE. Pay fails the numeric test in the company's favor, as set out above, and cash compensation is modest against earnings. The related-party construction arrangement is disclosed, capped, approved by independent shareholders at an extraordinary general meeting and reviewed by the auditor under Chapter 14A of the listing rules, which is what a well-governed cost-plus arrangement looks like, but cost-plus pricing with a controlling shareholder's own contractor gives that contractor no incentive to control cost, and the 6.9 times half-on-half jump against a declining cap schedule is the specific thing to watch. Borderline.

Two warnings fired. Governance Risk: High (2).

5. Corporate Ownership, Subsidiaries, and Joint Ventures

Ownership, taken from the issuer's own Part XV disclosure-of-interests record as at 28 February 2026:

• Holder: Sky Elite Limited (BVI, 100% owned by Mr Luo Shi); Shares: 934,556,316; % of issued: 44.31%

• Holder: Mr Luo Shi, aggregate incl. spouse, trust and beneficial; Shares: 973,034,569; % of issued: 46.13%

• Holder: First Beijing Investment Limited (investment manager); Shares: 344,079,000; % of issued: 16.31%

• Holder: Norges Bank (beneficial owner); Shares: 105,994,000; % of issued: 5.03%

• Holder: Mr Pan Ping (through spouse's vehicle Shang Long Limited); Shares: 13,043,289; % of issued: 0.62%

• Holder: Mr Wang Rui; Shares: 8,956,520; % of issued: 0.42%

• Holder: Mr Zhang Wenzao; Shares: 1,702,000; % of issued: 0.08%

• Holder: The Core Trust Company Limited (award-scheme trustee, at 28 January 2026); Shares: 71,065,817; % of issued: 3.37%

There is a single class of ordinary shares of HK$0.1, one vote each, no dual-class structure, and no acting-in-concert or voting agreement disclosed as in force at the latest filing. Free float outside the controlling shareholder, the two disclosed institutions, the directors and the trustee is roughly 33%. First Beijing Investment Limited went from 12.00% to 16.31% over the same six months, a second large buyer alongside the chairman during the decline.

The structure below the listed company is not trivial, and it is the single most important thing in this report. The Cayman company owns Tibet Yongsi Technology Co., Ltd., a wholly foreign-owned enterprise, which controls the PRC operating entities through structured contracts, an exclusive business cooperation agreement, exclusive call options, school sponsors' and directors' rights entrustments, powers of attorney, spouse undertakings, equity pledges and a loan agreement, because PRC law prohibits foreign ownership of primary and middle schools and restricts foreign participation in senior high schools to Sino-foreign cooperation subject to a qualification requirement the group has not yet met and is still working towards. Economic benefits reach the group through those contracts, not through equity. The group states it is not aware of any non-performance of the structured contracts.

• Structure: PRC operating entities (self-owned schools); Consolidation: Consolidated via structured contracts, no equity held; Economics: All revenue and profit

• Structure: Affected Business (30 compulsory-education schools); Consolidation: Deconsolidated 31 August 2021; Economics: RMB873.3m receivable, RMB1,399.8m payable, RMB254.0m guarantees, nil profit

• Structure: Associates (incl. Luzhou Tianli Kindergarten); Consolidation: Equity method; Economics: RMB13.1m carrying value; RMB0.6m loss in the half

• Structure: Non-controlling interests; Consolidation: Consolidated; Economics: Negative RMB14.1m

Ring-fencing risk is concrete rather than theoretical: bank loans of RMB1,647.2 million are secured on the equity interests in certain subsidiaries and on the rights to educational service fees of certain schools, which places lenders ahead of the listed holding company on precisely the cash streams the structured contracts are designed to deliver upward. Minority leakage is immaterial, non-controlling interests are negative and small, but the group is exposed to the put option over the 20% minority of one acquired subsidiary described in the relevant discussion.

6. Historical Financial Quality and Normalized Owner Earnings

Fiscal years ended August 31; FY2025 refers to the year ended August 31, 2025.

Two things about this table need saying before any number is used. First, FY2021 is not a comparable year: it is an eight-month transition period ending 31 August 2021 whose result is dominated by the deconsolidation of the compulsory-education schools, a RMB1,085.2 million impairment plus RMB222.6 million of discontinued-operations losses, on a perimeter that no longer exists. Second, the growth from FY2022 to FY2025 is therefore the rebuild of a differently-shaped company, not organic growth from a small base. Revenue rose from RMB884.4 million to RMB3,588.9 million and net income from RMB96.2 million to RMB648.1 million over three years, while gross margin settled in a tight 33.7% to 33.8% band from FY2023 onward. That stability through a fourfold revenue increase is the strongest single piece of evidence for the operating model.

The cash record is good where it can be seen. Operating cash flow was RMB852.6 million in FY2025 against reported profit of RMB648.1 million, and RMB827.1 million in FY2024 against RMB556.2 million, conversion above 130% in both years, which is what a prepaid-tuition model should look like. Capital expenditure of RMB403.6 million and RMB406.8 million left free cash flow of RMB449.0 million and RMB420.3 million. In the six months to 28 February 2026 operating cash flow was RMB520.3 million against RMB52.8 million a year earlier, a swing the interim narrative does not explain; capital expenditure rose to RMB540.2 million from RMB218.1 million, so free cash flow in the half was negative and cash fell from RMB965.2 million to RMB449.3 million.

Normalized owner earnings. Working from the real cash-flow statement rather than a proxy: FY2025 operating cash flow of RMB852.6 million, less maintenance capital expenditure proxied at RMB195.3 million (depreciation of owned property, plant and equipment RMB187.4 million plus amortisation of intangibles RMB7.9 million, deliberately excluding the RMB96.9 million of right-of-use depreciation whose cash cost appears in lease payments), less RMB98.2 million of interest paid and RMB12.4 million of lease interest and RMB21.6 million of lease principal, all three of which the group reports inside financing activities. That gives normalized owner earnings of approximately RMB525.1 million for FY2025, or RMB0.256 per share. The reported-to-normalized ladder for FY2025 runs: reported profit RMB648.1 million; the company's own adjusted profit RMB617.6 million (which strips the RMB100.0 million impairment reversal and adds back share-based payments, goodwill and financial-asset impairments and the associated deferred tax); owner earnings RMB525.1 million on the definition above. The adjustment is symmetric, the RMB100.0 million impairment reversal is a one-off gain and is removed, and no one-off charge is added back without it.

Normalized earnings basis for the valuation. The house default is a ten-year window; only five reporting periods exist since the IPO and one of them is the eight-month stub. The basis used is the four-year average of complete post-deconsolidation fiscal years FY2022 to FY2025, RMB408.7 million, or RMB0.1989 per share, which is also the lower of that average and the trailing-twelve-month figure of RMB730.0 million, so the conservative branch of the trajectory rule and the plain average agree. The sensitivity is material and is stated in full: including FY2021 as though it were a comparable year drops the five-year average to RMB60.6 million and lifts the normalized multiple from 4.4 times to 29.6 times, which would move the base cheapness cell from the cheapest earnings column to the most expensive one. That is the single largest judgment in this report and the relevant discussion carries it through both ways.

6.5 Piotroski F-Score, 8/9 (eligible)

Periods compared: FY2025 vs FY2024.

• #: 1; Signal: Positive net income; Passes (1) if: Net income > 0; Result: 1; Underlying figures: NI 648,130.0

• #: 2; Signal: Positive operating cash flow; Passes (1) if: CFO > 0; Result: 1; Underlying figures: CFO 852,592.0

• #: 3; Signal: ROA improved; Passes (1) if: ROA(t) > ROA(t-1); Result: 1; Underlying figures: ROA 6.4% vs 5.8%

• #: 4; Signal: Earnings quality (accruals); Passes (1) if: CFO > Net income; Result: 1; Underlying figures: CFO 852,592.0 vs NI 648,130.0

• #: 5; Signal: Leverage did not increase; Passes (1) if: Total debt / avg assets (t) <= (t-1); Result: 0; Underlying figures: debt/avg assets 23.6% vs 20.8%

• #: 6; Signal: Liquidity improved; Passes (1) if: Current ratio(t) > (t-1); Result: 1; Underlying figures: current ratio 0.49 vs 0.48

• #: 7; Signal: No dilutive share issuance; Passes (1) if: Net buyback >= 0 (cash flow); fallback diluted shares(t) <= (t-1); Result: 1; Underlying figures: net buyback 31,329.0 (repurchases 38,786.0 − issuance 7,457.0)

• #: 8; Signal: Gross margin improved; Passes (1) if: Gross margin(t) > (t-1); Result: 1; Underlying figures: GM 33.8% vs 33.7%

• #: 9; Signal: Asset turnover improved; Passes (1) if: Revenue / assets (t) > (t-1); Result: 1; Underlying figures: turnover 0.352 vs 0.347

Score: 8/9. Failed: leverage_decreased. Not computable: none. Result: eligible. Policy: 0-2 reject, 3 probation (needs a stated extraordinary offsetting factor), 4-9 eligible (no ranking within the band); fewer than 6 computable signals = data-limited, say so rather than veto. Conventions: ROA/turnover scaled by beginning of year assets; leverage = total debt / AVERAGE total assets; missing debt line with assets known = zero debt; unchanged ratio passes (<=); signal 7 method = net buyback cash flow.

The score is 8 out of 9, and the single failed signal is named: leverage did not decrease. Total borrowings rose from RMB2,053.0 million to RMB2,379.0 million while average total assets barely moved, taking debt over average assets from 20.8% to 23.6%. The cause is visible in the debt note rather than in distress, the group termed out short-dated borrowing, with non-current bank loans rising from RMB1,312.1 million to RMB1,623.7 million, and management attributes the reduction in net current liabilities over FY2025 directly to that increase in long-term bank loans. Lengthening maturities is the opposite of the deterioration this signal is designed to catch, but the signal is mechanical and it failed, so it is reported as failed.

The floor rule did not fire: it requires leverage to increase and liquidity to deteriorate in the same year, and liquidity improved, the current ratio rose from 0.476 to 0.487 across FY2025. All nine signals were computable, so this is not a data-limited score. Result is eligible (band 4 to 9, with no ranking inside the band; an 8 does not outrank a 5 at a deeper discount). No cyclical or commodity override was applied. One qualification a reader should carry forward: the score is measured on FY2025 against FY2024, and the interim balance sheet has moved the wrong way since, the current ratio fell to 0.416 at 28 February 2026 and borrowings rose again to RMB2,535.8 million. On the interim data the liquidity signal would also fail, which would fire the floor rule. That is a flag for the next annual filing, not a fact about the score as computed.

7. Balance Sheet, Debt, Covenants, and Refinancing Risk

The capital structure at 28 February 2026, disaggregated before classification:

• Instrument: Bank loans, secured, current; Balance (RMB'000): 145,000; Rate: 2.90-3.60%; Maturity: 2026; Security: Educational service fee rights / subsidiary equity; Availability: ,

• Instrument: Bank loans, unsecured, current; Balance (RMB'000): 299,808; Rate: 2.60-3.60%; Maturity: 2026-2027; Security: None; Availability: ,

• Instrument: Long-term bank loans, secured, current portion; Balance (RMB'000): 321,460; Rate: 3.40-6.90%; Maturity: 2026-2027; Security: Educational service fee rights / subsidiary equity; Availability: ,

• Instrument: Long-term bank loans, unsecured, current portion; Balance (RMB'000): 111,870; Rate: 3.20-4.40%; Maturity: 2026-2027; Security: None; Availability: ,

• Instrument: Bank loans, secured, non-current; Balance (RMB'000): 1,180,740; Rate: 3.40-6.66%; Maturity: 2027-2032; Security: Educational service fee rights / subsidiary equity; Availability: ,

• Instrument: Bank loans, unsecured, non-current; Balance (RMB'000): 476,960; Rate: 3.20-4.40%; Maturity: 2027-2029; Security: None; Availability: ,

• Instrument: Total financial borrowings; Balance (RMB'000): 2,535,838; Security: of which RMB1,647,200 secured; Availability: RMB1,432,710 undrawn committed facilities, drawable over two to three years

• Instrument: Lease liabilities, current; Balance (RMB'000): 46,453; Rate: , ; Maturity: within 1 year; Security: , ; Availability: ,

• Instrument: Lease liabilities, non-current; Balance (RMB'000): 266,491; Rate: , ; Maturity: beyond 1 year; Security: , ; Availability: ,

• Instrument: Total lease liabilities; Balance (RMB'000): 312,944

• Instrument: Amounts due to related parties (unsecured, interest-free); Balance (RMB'000): 1,475,108; Rate: nil; Maturity: 1-3 years; Security: None; Availability: ,

• Instrument: Cash and cash equivalents; Balance (RMB'000): 449,264

• Instrument: Time and restricted deposits; Balance (RMB'000): 9,694

All borrowings are denominated in RMB, which removes currency mismatch, the operating cash flow is RMB too. Financial borrowings are RMB2,535.8 million and lease liabilities RMB312.9 million; they are kept separate and the classification leads on financial borrowings. Against total equity of RMB3,323.0 million, financial borrowings are 76.3% of equity and borrowings plus leases 85.7%; against total assets of RMB10,020.1 million, financial borrowings are 25.3% (the group's own stated gearing measure) and 28.4% including leases. On total liabilities of RMB6,697.0 million the ratio to equity is 201.5%, but that measure is dominated by RMB952.7 million of prepaid tuition, RMB628.6 million of deferred income and RMB1,475.1 million of interest-free related-party balances, none of which is funded borrowing, so the funded-debt measure is the one used for the absolute filter and is stated as such.

The maturity wall is the near-term item: RMB878.1 million of bank debt falls due within twelve months of 28 February 2026 against RMB449.3 million of cash. Set against that are RMB1,432.7 million of undrawn committed facilities available over two to three years, and the FY2025 profile shows the group refinancing routinely, RMB1,516.0 million drawn and RMB1,192.5 million repaid in FY2025 alone. Beyond that, the FY2025 note shows RMB828.3 million due in the second year, RMB765.3 million in years three to five and RMB30.0 million beyond five years. Interest coverage on FY2025 figures is 10.1 times on an EBITDA basis (EBITDA of RMB1,258.6 million against total finance costs of RMB124.8 million, a blended figure that includes RMB12.4 million of lease interest and is therefore conservative), and pre-tax profit covers finance costs 6.7 times. Effective rates are falling, the interim range is 2.60% to 6.90% against 3.30% to 7.00% a year earlier. No financial covenants are disclosed in either primary filing, and no covenant test, waiver or breach is reported; the security is asset-and-cash-flow pledge rather than maintenance covenant, which is the normal PRC bank structure.

Classification: Manageable. On the grid alone, debt to assets of 25.3% and coverage above 10 times, the name would read Conservative, and that is not the honest answer. The group has run net current liabilities in every one of the last five years, RMB2,313.3 million at 28 February 2026 against RMB2,126.2 million at 31 August 2025, and the directors give an explicit going-concern assessment in both filings resting on the undrawn facilities and a twelve-month cash-flow forecast. The auditor issued a clean opinion with no going-concern paragraph and no material-uncertainty emphasis, and the structural reason for the negative working capital is benign, a prepaid business model books a year of tuition as a current liability. But a company that needs a stated going-concern rationale in every filing is not Conservative. Refinancing risk is real but well covered; the group is not dependent on the equity market.

8. Real Estate, Leases, and Hidden Assets

This is a property-heavy business. At 28 February 2026 owned property, plant and equipment was RMB5,046.4 million, 50.4% of total assets and 157.3% of tangible book value, and right-of-use assets were RMB2,379.5 million, 23.7% of total assets and 74.2% of tangible book value. Right-of-use assets split RMB2,107.9 million of land use rights and RMB271.6 million of buildings and other premises, and the land-use-rights balance grew by RMB189.0 million in the half as the group prepaid for new sites, a further RMB463.5 million of prepayments for land use rights sits in non-current other receivables, up from RMB200.0 million at the year end. Matching lease liabilities are only RMB312.9 million, because most land use rights in China are acquired by a single upfront premium for a fifty-year term rather than by periodic rent, so the asset is largely paid for rather than owed.

Right-of-use assets are a capitalized right to occupy, not a liquidation floor, and they are disclosed here rather than stripped out of the tangible book value used everywhere in this report, removing an asset without removing its matching liability is the wrong operation, and in this case the matching liability is small precisely because the right was prepaid. The genuine downside support in the tangible book is the RMB5,046.4 million of owned school buildings and equipment, of which RMB651.6 million of buildings do not yet have property ownership certificates and therefore cannot be sold, transferred or mortgaged until those certificates are obtained (Interim Report 2026, note 8(a)). That is 12.9% of the owned property balance and a real, disclosed constraint on realisability, improved from RMB695.8 million at the year end.

Hidden asset value is plausible but unquantified. The land was acquired progressively from 2002 onward in second- and third-tier Chinese cities and is carried at amortised cost; no revaluation, appraisal or fair-value disclosure for the property portfolio appears in either filing, so any claim that it is worth more than book is inference, not fact. Working in the other direction, RMB1,647.2 million of bank loans are secured on subsidiary equity and school fee rights, and RMB985.2 million of the 2021 impairment against assets occupied by the deconsolidated schools was still carried net at 1 September 2025, those particular buildings are occupied rent-free by entities the group does not consolidate, which is why they were impaired in the first place and why RMB81.9 million of that impairment was written back in the half on the strength of newly obtained art-training operating licences. There is no sale-leaseback program and no disclosed intention to monetize property.

9. Capital Markets Access, Dilution, and Financing Flexibility

The share count is shrinking, and that is the finding. Shares in issue went 2,154,000,000 (31 August 2023) to 2,115,654,000 (31 August 2024) to 2,109,180,000 (31 August 2025) to 2,109,255,000 (28 February 2026), the last small rise being 75,000 shares from option exercises; over the same period treasury shares rose to 20,727,000 and a further 33,711,100 sit with the award-scheme trustee, so shares genuinely outstanding fell to 2,054,816,900. Weighted-average diluted shares fell from 2,114,964,000 (FY2024) to 2,084,739,000 (FY2025) to 2,065,649,000 in the interim half. Net repurchases were RMB31.3 million of cash in FY2025 (RMB38.8 million spent, RMB7.5 million received on option exercises) and RMB50.4 million in the interim half, and 13,227,000 repurchased shares were canceled in June 2026 on a third-party account.*

The dilution overhang is nominally 67,825,000 outstanding share options, 3.2% of issued shares, plus 4,515,000 unvested award shares. Every option is deeply out of the money: exercise prices are HK$2.38, HK$2.48 and HK$4.72 against a HK$1.02 share price, so at present none represents realistic dilution, but they also represent management incentives that are entirely worthless at today's price, which is its own governance observation. The general mandate to issue new shares was renewed at the 2026 annual general meeting with 98.46% support, so the capacity to issue up to 20% exists; it has not been used since the December 2020 placing and subscription, whose HK$695.0 million of net proceeds were fully applied by 31 August 2025. The company confirms it conducted no equity fund-raising in FY2025. With RMB852.6 million of annual operating cash flow, RMB449.3 million of cash and RMB1,432.7 million of undrawn committed bank lines, the group does not depend on the equity market to fund operations. Dilution risk is judged low.

10. Litigation, Regulatory, and Contingent Liability Risk

The notes reached in the primary documents are the annual report's Report of the Directors section headed "Litigation", the "Contingent Liabilities" and "Capital Commitments" disclosures in both the annual Financial Review and the interim Management Discussion and Analysis, note 32 / note 19 Commitments, note 37 Financial Guarantee Contracts, and note 33 / note 20 Related Party Transactions and Balances. On litigation the annual report states plainly that the group did not have any material litigation outstanding as at 31 August 2025; there is no separate legal-proceedings note in the financial statements because none is required, and the interim report repeats no litigation disclosure. On contingent liabilities both filings state that the group did not have any material contingent liabilities (31 August 2025: nil; 28 February 2026: nil), which sits awkwardly beside note 37 and is worth stating precisely: the group carries RMB254.0 million of financial guarantees for the deconsolidated Affected Business at 28 February 2026 (RMB374.0 million at 31 August 2025), recognises no expected-credit-loss allowance against them, classifies the exposure as Stage 1 and reports no transfers between stages. Management's position is that the initial fair value of those guarantees was not significant and no loss is expected; a reader should treat "no material contingent liabilities" as a statement about expected loss, not about gross exposure.

Capital commitments contracted but not provided for were RMB37.8 million at 28 February 2026, immaterial against a RMB10.0 billion balance sheet.

The prior-disclosure carry-forward produces three live items and no surprises. First, the 2021 Implementation Regulations: open since May 2021, still open, with the company stating that as at the date of the FY2025 report national and local classification-management regulations have still not been issued and it will make further announcements as appropriate. Second, the social insurance and housing provident fund non-compliance disclosed in the 2018 prospectus: the company states that as at 31 August 2025 it has established sufficient provision, which is a status update rather than a resolution, and the provision amount is not separately disclosed. Third, the qualification requirement under the Sino-foreign school-operation regulations, which the group has never satisfied: its stated remedy is to acquire or partner with overseas schools, a plan still described in the present tense in both filings. Subsequent events were nil in both, the annual reports none after 31 August 2025 up to 28 November 2025, and interim note 22 none after 28 February 2026 as at 10 April 2026. Two latent tax items round this out: RMB892.0 million of mainland tax losses carrying no deferred tax asset, and RMB2,259.2 million of unremitted subsidiary earnings on which no deferred tax has been provided, a RMB113.0 million liability at the 5% withholding rate if the group ever repatriated in full.

The single largest contingent exposure is not in any of these notes: it is the structured-contract architecture itself, discussed in the relevant discussion and carried as the top row of the risk matrix.

11. Accounting Quality and Disclosure Review

The auditor is Ernst & Young, unchanged for at least three years, and the opinion on the FY2025 statements is unqualified with no going-concern paragraph, no material-uncertainty emphasis and no reported material weakness. The interim was subject to an independent review, also by Ernst & Young, with the standard negative-assurance conclusion. Two key audit matters were reported. The first is revenue recognition, flagged for size and transaction volume rather than for any specific concern, with the usual controls, cut-off and analytical procedures. The second is the one that matters: the impairment assessment of property, plant and equipment and right-of-use assets occupied by the deconsolidated Affected Business, where the auditor engaged internal valuation specialists to test the discount rates and cash-flow projections behind a RMB100.0 million reversal of previously recognized impairment.

That reversal is the accounting item to watch. It flows straight into reported profit, RMB100.0 million in FY2025 and a further RMB81.9 million in the interim half, and it rests on directors' value-in-use calculations for assets the group does not consolidate and does not charge rent for, discounted at pre-tax rates of 17.0% and 17.4%, triggered by the schools obtaining art-training operating licences. The disclosure is good: the interim gives recoverable amount, carrying amount and reversal for each of two cash-generating units, and the recoverable amount of RMB190.7 million exceeds the carrying amount of RMB86.7 million. To the company's credit, its own adjusted-profit reconciliation strips the reversal out, so management is not presenting it as operating performance. But RMB181.9 million of pre-tax profit across eighteen months is estimate-driven, non-cash and reversible, and RMB985.2 million of the original impairment remained available to write back at 1 September 2025.

Elsewhere the picture is clean. Goodwill is small, RMB97.6 million, 3.0% of tangible book, and RMB8.8 million of it was impaired at FY2025 when a repetition school's enrolment fell; management explicitly states it performed no goodwill impairment test at 28 February 2026, which is permitted (IAS 36 requires an annual test) but is a disclosed choice worth noting the year after an impairment. Net deferred tax assets of RMB276.1 million rest largely on the impairment losses; RMB892.0 million of tax losses carry no deferred tax asset at all, which is conservative. There is one operating segment, so segment transparency is limited by design rather than by choice. Related-party disclosure is detailed and quantified, including pricing methodology, annual caps and guarantee-by-guarantee tables. Non-IFRS adjustments are reconciled line by line. Classification: Adequate, an unqualified opinion, full primary-source verification of every material note, and one genuine estimation-risk area where the auditor's pushback is described and appears proportionate.

12. Valuation and Margin of Safety

Tangible Book Value, Derivation (anchor 2026-02-28, Interim Report 2026 (six months ended 28 February 2026), Interim Condensed Consolidated Statement of Financial Position, pp.27-28; RMB thousands)

Table 1, equity bridge

• Line item: Common shareholders' equity; Amount: 3,337,133.0; Source (filing page / tag / note): Equity attributable to owners of the Company, 28 Feb 2026; Period: 2026-02-28; State: present

• Line item: − Goodwill; Amount: -97,624.0; Source (filing page / tag / note): note 10; Period: 2026-02-28; State: present

• Line item: − Other intangibles; Amount: -32,257.0; Source (filing page / tag / note): Other intangible assets, 28 Feb 2026; Period: 2026-02-28; State: present

• Line item: = Tangible book value; Amount: 3,207,252.0; Source (filing page / tag / note): derived: common equity − goodwill − other intangibles; Period: 2026-02-28; State: derived

• Line item: ÷ Shares (point in time); Amount: 2,054.817; Source (filing page / tag / note): Interim Report 2026 note 16: 2,109,255,000 shares issued and fully paid, less 20,727,000 treasury shares, less 33,711,100 shares held by the trustee under the Restricted Share Award Scheme (note 17(a)); both blocks are carried as contra-equity in the statement of changes in equity; Period: 2026-02-28; State: present

• Line item: = TBV / share; Amount: 1.5608; Source (filing page / tag / note): derived; Period: 2026-02-28; State: derived

• Line item: (Minority interest, excluded by line choice, not subtracted); Amount: -14,086.0; Source (filing page / tag / note): , ; Period: 2026-02-28; State: present

P/TBV (derived) = 0.87 ÷ 1.5608 = 0.559× (price as of 2026-09-07)

Table 2, asset composition at the anchor period

• Asset: Cash & short-term investments; Amount: 458,958.0; % of total assets: 4.6%; % of TBV: 14.3%

• Asset: Accounts receivable; Amount: 94,776.0; % of total assets: 0.9%; % of TBV: 3.0%

• Asset: Inventory; Amount: 21,195.0; % of total assets: 0.2%; % of TBV: 0.7%

• Asset: Property, plant & equipment (owned); Amount: 5,046,416.0; % of total assets: 50.4%; % of TBV: 157.3%

• Asset: Right-of-use lease assets (disclosed, never stripped from TBV); Amount: 2,379,477.0; % of total assets: 23.7%; % of TBV: 74.2%

• Asset: Long-term investments; Amount: 13,131.0; % of total assets: 0.1%; % of TBV: 0.4%

• Asset: Goodwill; Amount: 97,624.0; % of total assets: 1.0%; % of TBV: 3.0%

• Asset: Other intangibles; Amount: 32,257.0; % of total assets: 0.3%; % of TBV: 1.0%

• Asset: Other current / noncurrent assets; Amount: 1,876,227.0; % of total assets: 18.7%; % of TBV: 58.5%

• Asset: Residual / unclassified; Amount: 0.0; % of total assets: 0.0%; % of TBV: 0.0%

• Asset: Total assets; Amount: 10,020,061.0; % of total assets: 100%

Right-of-use note: ROU assets 2,379,477.0 = 74.2% of TBV. ROU is a capitalized right to occupy, not a liquidation floor, disclosed, never stripped out of the TBV/share used everywhere in this report.

No analyst adjustments are made to the bridge, so adjusted tangible book value per share equals tangible book value per share at RMB1.5608 (HK$1.82). The candidate haircut, writing off the RMB873.3 million interest-free receivable from the deconsolidated schools, is run in the scenario table below rather than folded into the headline figure, so the derivation stays a pure filing bridge. Currency: all analysis is in the reporting currency, RMB. One rate is used, HKD1.00 = RMB0.8559 (7 September 2026); the Hong Kong dollar's US peg mid-point of HKD7.80 = USD1.00 is used for the US dollar parentheticals on the decision figures only. History tables are never converted at today's rate.

Historical relative multiple valuation. A P/TBV history of 29 observations spanning August 2018 to February 2026, 7.2 years at roughly quarterly cadence, is available from a third-party ratio series; it was not rebuilt point by point from filings and is disclosed as such.* Its distribution is: minimum 0.66×, lower quartile 1.86×, median 2.56×, upper quartile 3.58×, ninetieth percentile 4.39×, top-quartile mean 4.34×, maximum 5.17×. Read in the required order: the typical band is roughly 1.9× to 3.6× tangible book, where the multiple has spent most of its listed life; the floor and ceiling are 0.66× and 5.17×, the floor set only in the most recent observation; and today the derived multiple is 0.559×, below the entire recorded series. The current point is computed from the derivation above, not read off a chart, and it is lower than the series minimum because the price has fallen a further 65% since the last observation. There is no comparable P/TBV series before the 2018 listing, and no sourced normalized-P/E history exists for this issuer, so the earnings leg of this comparison is UNAVAILABLE on the stock's own history and is anchored on peers instead.

• Metric: P / TBV (own history, 29 obs, Aug 2018 - Feb 2026); Current: 0.559×; Window floor: 0.66×; Typical band: 1.86× - 3.58×; Window ceiling: 5.17×; Implied price (band): RMB2.90 - RMB5.59 (HK$3.39 - HK$6.53)

• Metric: Normalized P/E (peer-anchored, no own-history series); Current: 4.4×; Window floor: , ; Typical band: 3.0× - 8.0×; Window ceiling: , ; Implied price (band): RMB0.60 - RMB1.59 (HK$0.70 - HK$1.86)

The two legs disagree, and the disagreement is the analysis. The book leg says the stock is priced below anything it has ever traded at, implying 3.3 to 6.4 times upside on reversion to its own typical band. The earnings leg, anchored on the 2.95× trailing multiple at which a listed Hong Kong private-education peer traded in April 2026,* says the current 4.4× normalized multiple is roughly where the whole cohort now sits, cheap in absolute terms but not cheap relative to peers. Reconciling rather than averaging: the book leg is measuring how far this specific stock has fallen from its own history; the earnings leg is measuring a sector-wide regime change that has already happened. Both are true. The reconciled demonstrated range for the scenario work below therefore anchors on tangible book multiples, with the peer earnings multiple used as the bear-case check.

* Sourced from a third-party ratio history and third-party market data; not independently verified against source filings.

Sell range (framework rule). [median P/TBV, top-quartile-mean P/TBV] × current adjusted TBV/share over the stock's own history gives RMB4.00 - RMB6.77 (HK$4.67 - HK$7.91; US$0.60 - US$1.01), equivalent to 2.56× - 4.34× tangible book, from 29 observations covering 7.2 years (August 2018 to February 2026) at roughly quarterly cadence, sorted ascending with linear interpolation at the percentile. The distribution behind it is minimum 0.66× · lower quartile 1.86× · median 2.56× · upper quartile 3.58× · ninetieth percentile 4.39× · top-quartile mean 4.34× · maximum 5.17×. The flat 0.80× reference price is RMB1.25, fallback only, not the rule for this name, since the history comfortably exceeds the five-year floor. Neither coherence flag fires: the range starts far above today's price, and the price is nowhere near the top-quartile mean. Today's derived multiple sits at the zeroth percentile of the series.

Where in the range to exit, and why: at or just above the low end, RMB4.00 (HK$4.67), not the top. The range reflects the multiple the market paid when it treated this as a growth compounder inside an expanding regulatory perimeter. The earnings power underlying it has not deteriorated, return on equity was 21.8% in FY2025 against a 16.4% average across the four post-deconsolidation years, so today's returns are above, not below, the period average, and the credibility test the house applies to a sell range (current returns 30% or more below the multi-year average) does not fire. What has changed is not earnings but the multiple regime the whole cohort is granted, and there is no evidence in the filings that the regime reverts. Treating the median as the exit and the upper quartile as an unexpected gift is the honest reading.

Scenario table. The lens is price to tangible book, the way asset-heavy school operators and the rest of this cohort actually trade, cross-checked against normalized earnings.

• Scenario: Severe downside; Key assumption: Structured contracts legislated away for the senior-high perimeter as they were for compulsory education in 2021; Method: Residual tangible book: RMB3,207.3m less the RMB873.3m Affected-Business receivable written to nil, less RMB254.0m of guarantees called in full, less a repeat of the FY2021 impairment at its actual absolute size of RMB1,085.2m, then divided by 2,054.8m shares; Value / share: RMB0.484 (HK$0.57), 0.31× today's unimpaired tangible book, not 0.31× applied as a multiple; vs current (RMB0.873): −45%

• Scenario: Bear; Key assumption: No regulatory event; earnings flat at the normalized level and the cohort multiple holds where it is; Method: Normalized EPS RMB0.1989 × the 2.95× trailing multiple observed on a listed peer, rounded to 3.0×; Value / share: RMB0.597 (HK$0.70); vs current (RMB0.873): −32%

• Scenario: Base; Key assumption: Business continues as filed; the market pays tangible book, the house's own absolute ceiling and a level this stock exceeded for its entire listed history until 2026; Method: 1.00× TBV/share of RMB1.5608. Cross-check: normalized EPS RMB0.1989 × 8.0× = RMB1.591, agreeing within 1.9%; Value / share: RMB1.561 (HK$1.82); vs current (RMB0.873): +79%

• Scenario: Bull; Key assumption: Partial re-rating to the bottom of the stock's own historical band, without recovering the median; Method: 1.86× TBV/share (own-history lower quartile) = RMB2.903. Cross-check: normalized EPS RMB0.1989 × 15.0× = RMB2.984, agreeing within 2.8%; Value / share: RMB2.903 (HK$3.39); vs current (RMB0.873): +233%

The decision figures, stated explicitly. Current price HK$1.02 / RMB0.873 (7 September 2026), market capitalization approximately HK$2.10 billion (RMB1.79 billion; US$269 million). Intrinsic value range, base case: RMB1.56 - RMB2.90 per share (HK$1.82 - HK$3.39). Margin of safety at the current price against tangible book value per share: 44.1%. Buy-Below: RMB2.10 (HK$2.45; US$0.31), 1.35× TBV, read off 7.2 years of the stock's own history, the house construction, computed from the same distribution and window as the sell range. It carries an explicit qualification. RMB2.10 is 1.35× tangible book, which is above the 1.00× absolute ceiling that governs whether a name belongs in this universe at all, so the operative buy price is the lower of the two: RMB1.56 (HK$1.82; US$0.23) at 1.00× TBV. The universe-grid readout is RMB1.01 (0.65× TBV). All three are stated because the pair is internally inconsistent for a stock whose entire recorded valuation history sits above the universe ceiling, and naming the binding constraint matters more than publishing the highest number. Buy More Below: not reachable, the base cell is already the largest size, so there is no next size up. Sell range: RMB4.00 - RMB6.77 (HK$4.67 - HK$7.91), as derived above. Cheapness type: statistically cheap. The stock is cheap on assets, on earnings, on cash flow and against its own history simultaneously, which is the classic statistical profile; the honest qualifier is that a statistically cheap PRC education asset held through contracts rather than equity is exactly the shape a value trap takes, and nothing in the filings excludes that reading.

13. Risk Matrix

• Risk: Structured contracts legislated away for senior-high schools, as they were for compulsory education in 2021; Evidence: Annual Report 2025 note 1 and Regulatory Updates; RMB1,085.2m impairment and 30 schools deconsolidated on 31 August 2021; classification-management rules still not issued; Severity: Critical; Probability: Low-Medium; Financial Impact: Loss of consolidation of the entire operating base; Valuation Impact: Tangible book largely unrealisable; the severe-downside case of RMB0.484 or worse; Mitigant: Senior high is post-compulsory and explicitly permitted to be for-profit; the group is pursuing the Sino-foreign qualification route; Monitor: Any State Council or provincial classification-management rule; any announcement on the qualification requirement

• Risk: Affected-Business entanglement: receivable and guarantees to entities outside the group; Evidence: Receivable RMB873.3m (up from RMB686.9m two years earlier); guarantees RMB254.0m with no consideration received; no expected-credit-loss allowance recognized (note 20(c)(5), note 37); Severity: High; Probability: Medium; Financial Impact: Up to RMB1,127.3m, 35% of tangible book; Valuation Impact: Minus RMB0.549 per share if fully written off; Mitigant: Net position is a RMB526.5m payable to the same counterparties; guarantee exposure fell from RMB374.0m in twelve months; Monitor: The receivable balance at each reporting date; any allowance being raised; the guarantee schedule in note 37

• Risk: Dividend suspension signals a change in payout policy or a cash constraint; Evidence: Board resolution 10 April 2026: nil interim dividend against RMB5.78 cents; cash fell from RMB965.2m to RMB449.3m in the half; capex RMB540.2m against RMB218.1m; Severity: High; Probability: Confirmed (already occurred); Financial Impact: Removes roughly half the distribution run rate; Valuation Impact: Yield support for the shares halves; Mitigant: Operating cash flow rose to RMB520.3m in the half; RMB1,432.7m undrawn facilities; Monitor: The FY2026 final dividend declaration, expected with the annual results in November 2026

• Risk: Near-term maturity wall against reduced cash; Evidence: RMB878.1m of bank debt due within 12 months of 28 February 2026 against RMB449.3m of cash; current ratio fell from 0.487 to 0.416; Severity: Medium-High; Probability: Medium; Financial Impact: Refinancing at higher cost, or capex deferral; Valuation Impact: Multiple compression on liquidity fear; Mitigant: RMB1,432.7m undrawn committed facilities; RMB1,516.0m drawn and RMB1,192.5m repaid in FY2025 without incident; Monitor: Interim and annual cash balance, undrawn facility disclosure, effective interest rate range

• Risk: Impairment reversals flatter reported profit; Evidence: RMB100.0m reversed in FY2025 and RMB81.9m in H1 FY2026, on directors' value-in-use estimates at 17.0 to 17.4% discount rates for assets occupied by non-consolidated schools; RMB985.2m of the original charge still carried net; Severity: Medium; Probability: High (recurring); Financial Impact: Up to 28% of FY2025 pre-tax profit is estimate-driven and non-cash; Valuation Impact: Overstates the earnings base a multiple is applied to; Mitigant: The company's own adjusted-profit measure strips it out; the auditor used valuation specialists on it as a key audit matter; Monitor: The impairment note in each filing; whether reversals continue and on what trigger

• Risk: Governance: 46.13% control, combined chairman and chief executive, cost-plus related-party construction; Evidence: CG Code C.2.1 deviation; Nanyuan Construction at cost plus 9 to 11%, RMB50.1m in H1 FY2026 against RMB7.3m a year earlier; nomination committee chaired by the controlling shareholder; Severity: Medium-High; Probability: High (ongoing); Financial Impact: Value leakage of unknown size through construction pricing; Valuation Impact: Persistent governance discount; Mitigant: Annual director elections held and polled; caps stepping down to RMB170m by FY2027; auditor and independent-director review under Chapter 14A; Monitor: Half-yearly related-party note; the FY2026 transaction total against the RMB210m cap

• Risk: Mix shift into low-margin product sales erodes group margin; Evidence: Product sales up 45.6% against procurement cost up 49.6% in H1 FY2026; group gross margin down 2.4 points to 35.2%; Severity: Medium; Probability: High (already occurring); Financial Impact: Each further point of margin is roughly RMB43m of gross profit annualised; Valuation Impact: Compresses the normalized earnings anchor; Mitigant: Education-services margin itself is stable; management fees are high-margin and growing 38.7%; Monitor: Half-yearly revenue mix and gross margin by line

• Risk: Buildings without ownership certificates limit realisability of the asset floor; Evidence: RMB651.6m of buildings, 12.9% of owned property, cannot be sold, transferred or mortgaged until certificates are obtained (note 8(a)); Severity: Medium; Probability: Medium; Financial Impact: Reduces collateral and liquidation value; Valuation Impact: Weakens the asset-backed argument; Mitigant: Balance fell from RMB695.8m in six months, so certification is progressing; Monitor: The note 8 disclosure at each reporting date

• Risk: Sector-wide de-rating persists regardless of results; Evidence: 16 of 21 Hong Kong-listed private-education issuers down more than 70% from peak; a peer at 2.95 times trailing earnings; the shares down about 57% year to date on results that improved; Severity: Medium; Probability: High; Financial Impact: None on the business; Valuation Impact: The base case simply never arrives; Mitigant: The chairman's vehicle bought 37,579,000 shares and a disclosed institution added 90,989,000 during the fall; Monitor: Cohort multiples; index and connect-eligibility changes; the FY2026 results reaction

14. Red Flags, Yellow Flags, and Green Flags

🟢 Green Flags

• The controlling shareholder's vehicle bought 37,579,000 shares (about 1.8% of the company) in the six months to 28 February 2026, lifting his interest from 44.35% to 46.13% as the shares fell from roughly HK$3.60 to under HK$3.00; a disclosed institution added 90,989,000 shares over the same period, going from 12.00% to 16.31%.

• Piotroski F-Score 8/9, with only the leverage signal failing, and that failure caused by terming out short-dated debt rather than by distress.

• Operating cash flow of RMB852.6 million in FY2025 on reported profit of RMB648.1 million, 132% conversion, with free cash flow of RMB449.0 million after RMB403.6 million of growth capital expenditure.

• Share count falling every year since 2023 (2,154.0m to 2,109.3m issued, and 2,054.8m genuinely outstanding), with all 67,825,000 outstanding options struck at HK$2.38 or above and therefore valueless at HK$1.02.

• Unqualified Ernst and Young opinion, and related-party disclosure that quantifies pricing method, annual caps and guarantee-by-guarantee exposure.

🟡 Yellow Flags

• Cash fell from RMB965.2 million to RMB449.3 million in the six months to 28 February 2026 while capital expenditure rose from RMB218.1 million to RMB540.2 million; the current ratio fell from 0.487 to 0.416.

• Gross margin fell 2.4 points to 35.2% in the half as low-margin product sales grew 45.6% and became almost a third of revenue.

• RMB181.9 million of impairment reversals across FY2025 and H1 FY2026 flow into reported profit on directors' value-in-use estimates for assets occupied by schools the group does not consolidate.

• RMB651.6 million of buildings cannot be sold, transferred or mortgaged pending ownership certificates.

• No goodwill impairment test was performed at 28 February 2026, the reporting date immediately after a RMB8.8 million goodwill write-down.

• As of 31 August 2026: the joint company secretary resigned, with no reason verified from a primary source.*

🔴 Red Flags

• The interim dividend was cut to nil on 10 April 2026, from RMB5.78 cents a share, in a half year when profit rose 21.0%, with no explanation offered in the interim report.

• Interest-free, unsecured lending to the deconsolidated Affected Business keeps growing: RMB686.9m to RMB873.1m over two years, plus RMB254.0 million of guarantees for which the group receives nothing.

• The entire operating business is controlled by contract, not equity, and the identical structure was overridden by regulation in 2021 at a cost of RMB1,085.2 million and thirty schools.

• Cost-plus-9-to-11% construction with the chairman's own company jumped 6.9 times half on half, to RMB50.1 million from RMB7.3 million, against annual caps that step down each year.

• Buybacks destroyed capital: 30,116,000 shares bought for roughly HK$99.0 million at a blended HK$3.29, now worth about HK$30.7 million, a 69% loss on the capital deployed.

⚡ Must-Watch Catalysts

• FY2026 annual results and the final dividend decision, expected late November 2026, the single most informative event, since it reveals whether the interim suspension was a policy change or a timing decision.

• First consolidation of the art-education acquisition announced 9 to 10 June 2026 (RMB104 million for 10% plus RMB140.8 million of capital, taking the stake to 51%), which should appear in the FY2026 accounts.*

• 7 September 2026: the earliest date on which the 20% minority holder of an acquired subsidiary may require the group to buy out its stake at not less than acquisition-date valuation.

• The autumn 2026 enrolment number, against roughly 60,000 high-school students a year earlier.

• Any State Council or provincial classification-management rule issued under the 2021 Implementation Regulations.

FINAL VALUE-INVESTING RECOMMENDATION

Verdict: BUY, BIG

• Dimension: Current Price; Assessment: HK$1.02 (7 September 2026) = RMB0.873 at HKD1.00 = RMB0.8559; market capitalization approximately HK$2.10bn (RMB1.79bn; US$269m) on 2,054,816,900 shares outstanding

• Dimension: Valuation and balance-sheet screens (the valuation framework); Assessment: PASS. P/TBV 0.559× against the 1.00× ceiling. Funded debt to equity 76.3% (financial bank borrowings RMB2,535.8m ÷ total equity RMB3,323.0m), not net of cash; 85.7% including RMB312.9m of lease liabilities, which are stated separately per IFRS 16; total liabilities to equity 201.5% but dominated by prepaid tuition, deferred income and interest-free related-party balances rather than funded borrowing

• Dimension: Piotroski F-Score (the relevant discussion); Assessment: 8/9, Eligible. Failed signal: leverage did not decrease (debt/average assets 23.6% vs 20.8%), caused by terming out short-dated debt. Floor rule did NOT fire: liquidity improved in the same year (current ratio 0.487 vs 0.476). On the interim balance sheet the liquidity signal would also fail, which would fire the floor rule at the next annual, a flag, not a fact about this score

• Dimension: P / TBV (derived: price ÷ filing TBV/share, date); Assessment: 0.559× (filing TBV/share RMB1.5608 at 28 February 2026; price 7 September 2026)

• Dimension: Normalized P/E; Assessment: 4.4× on normalized net income of RMB408.7m, being the four-year average of complete post-deconsolidation years FY2022-FY2025 and also the lower of that average and trailing-twelve-month RMB730.0m. Graham-flavoured, not literally Schloss. Sensitivity: 29.6× if the eight-month FY2021 transition stub is included as a comparable year

• Dimension: Cash Return (dividend + buyback yield); Assessment: 4.5% forward run rate on the last declared distribution (final FY2025 dividend RMB3.90 cents ÷ RMB0.873) plus roughly 2.8% of market capitalization of net buybacks in H1 FY2026 (RMB50.4m). On the FY2025 declared basis it was 11.1% dividend yield plus 1.7% net buyback = 12.8%; the interim FY2026 dividend was suspended, so the forward figure is the operative one

• Dimension: Cheapness Tier, Numbers Only (the valuation framework); Assessment: Big, the base cell only, unmodified: P/TBV row 0.50×-0.65×, normalized P/E column ≤ 12×. On the FY2021-inclusive normalization sensitivity the same P/TBV row and the > 20× column read Small

• Dimension: Capital-Allocation Warnings; Assessment: 2 fired. (2) Capital and credit routed to policy-driven entities outside the group, RMB873.3m interest-free receivable from the deconsolidated Affected Business, growing, plus RMB254.0m of guarantees for no consideration. (3) Dividend suspension, nil interim for H1 FY2026 against RMB5.78 cents, in a half when profit rose 21.0%. Warnings (1) control and (5) pay/related-party pricing are BORDERLINE, not fired; warning (4) is UNASSESSABLE

• Dimension: governance-warning total Downgrade Applied?; Assessment: No. Neither flag is applied to the tier (framework rule, 2026-08-28). Shareholder-yield flag +1 notch-equivalent (yield ≥ 3.0%, free cash flow explicitly not weak); the valuation framework 2 warnings fired, −1 notch-equivalent. Both are reported for a human to weigh after seeing them, and neither moved the tier above

• Dimension: Intrinsic Value Range; Assessment: RMB1.56 - RMB2.90 per share (HK$1.82 - HK$3.39), base to bull, from the tangible-book anchor cross-checked against normalized earnings

• Dimension: Margin of Safety; Assessment: 44.1% against tangible book value per share of RMB1.5608 at the current price of RMB0.873

• Dimension: Buy-Below Price (framework price, absolute valuation ceiling); Assessment: RMB2.10 (HK$2.45; US$0.31), 1.35× TBV, 7.2-year own-history lower-quartile mean. Qualified: that price is above the 1.00× absolute universe ceiling, so the operative buy price is the lower RMB1.56 (HK$1.82) at 1.00× TBV; the universe-grid readout is RMB1.01 (0.65× TBV)

• Dimension: Buy More Below; Assessment: Not reachable, the base cell is already the largest size, so there is no next size up

• Dimension: Sell Range (the relevant discussion); Assessment: RMB4.00 - RMB6.77 (HK$4.67 - HK$7.91; US$0.60 - US$1.01) = 2.56× - 4.34× TBV, from 29 observations over 7.2 years; percentiles sorted ascending with linear interpolation; flat 0.80× reference RMB1.25 is fallback only, not the rule for this name; no coherence flag fires; expected exit at or just above the low end, since the upper quartile reflects a multiple regime the cohort no longer receives

• Dimension: Mainland-China Sizing Overlay (the sizing overlay ); Assessment: Triggered. Numbers-only tier Big is unchanged by the overlay, but this is a mainland-core business (operations, revenue and PRC-domiciled operating entities are entirely mainland Chinese, controlled through structured contracts) merely listed in Hong Kong, the overlay's own worked example of what does NOT get the "HK listing alone" exemption. Economic-risk bucket: mainland_china. Default max fresh-money entry tier: Standard, not Big. No dated human override is on file, so Standard is the operative cap pending one

• Dimension: Position Sizing Guidance; Assessment: Numbers-only tier Big, capped to Standard by the Mainland-China entry-sizing overlay above, under a hypothetical general value mandate applied the same way to every company, with Governance Risk High and the structured-contract risk disclosed alongside it. General commentary for a publication, never sized to any individual reader

• Dimension: Expected Holding Period; Assessment: Three to five years, long enough for two full annual reporting cycles and one regulatory review window

• Dimension: Downside Risk; Assessment: RMB0.484 per share (HK$0.57), −45%, on the severe-downside construction in the relevant discussion: the RMB873.3m Affected-Business receivable written to nil, RMB254.0m of guarantees called, and the FY2021 impairment repeated at its actual absolute size of RMB1,085.2m

• Dimension: Balance-Sheet Risk; Assessment: Manageable. Debt/assets 25.3% and EBITDA interest cover 10.1× would read Conservative on the grid, but net current liabilities of RMB2,313.3m and a stated going-concern rationale in every filing argue one notch lower

• Dimension: Creditworthiness; Assessment: Adequate, all borrowing in RMB at 2.60%-6.90%, no disclosed financial covenants, RMB1,432.7m of undrawn committed facilities, RMB1,516.0m drawn and RMB1,192.5m repaid in FY2025 without incident

• Dimension: Governance Risk; Assessment: High (2 warnings fired)

• Dimension: Accounting Quality Risk; Assessment: Adequate, unqualified Ernst and Young opinion, two key audit matters with the auditor's procedures described, full primary-source verification of every material note, one genuine estimation-risk area in the impairment reversals

• Dimension: Refinancing Risk; Assessment: Moderate, RMB878.1m of bank debt due within twelve months of 28 February 2026 against RMB449.3m of cash, covered by RMB1,432.7m of undrawn committed facilities drawable over two to three years

• Dimension: Key Catalysts; Assessment: FY2026 annual results and the final dividend decision (expected late November 2026); autumn 2026 enrolment against roughly 60,000; first consolidation of the June 2026 art-education acquisition; any State Council or provincial classification-management rule under the 2021 Implementation Regulations

• Dimension: Primary Thesis Killers; Assessment: (1) Structured contracts overridden for the senior-high perimeter as they were for compulsory education in 2021; (2) the RMB873.3m Affected-Business receivable and RMB254.0m of guarantees proving unrecoverable; (3) a permanent cohort de-rating in which improving results never re-rate the shares

VALUE INVESTOR MUST-WATCH LIST

• FY2026 annual results, expected late November 2026, whether the interim dividend suspension of 10 April 2026 was a timing decision or a policy change, and whether a final dividend is declared at all.

• The Affected-Business receivable at 31 August 2026, which has risen RMB686.9m to RMB873.1m over two years. A further increase without a repayment schedule, or a first expected-credit-loss allowance against it, changes the tangible-book arithmetic directly.

• The RMB254.0m of financial guarantees in the guarantee note: whether the balance keeps falling (it fell from RMB374.0m over twelve months) and whether any Stage 1 classification changes.

• RMB878.1m of bank debt maturing within twelve months of 28 February 2026 against RMB449.3m of cash, watch the refinancing and the undrawn facility balance, which was RMB1,432.7m.

• Related-party construction with the chairman's company against the FY2026 cap of RMB210m. The half-year figure was RMB50.1m against RMB7.3m a year earlier; the cap steps down to RMB170m in FY2027.

• Gross margin and revenue mix. Product sales were 32.3% of half-year revenue and grew 45.6% while group margin fell 2.4 points to 35.2%. Each further point of margin is roughly RMB43m of annualised gross profit.

• Impairment reversals, RMB181.9m written back across eighteen months, with RMB985.2m of the original charge still available at 1 September 2025.

• Autumn 2026 enrolment, against approximately 60,000 high-school students at the start of the autumn 2025 term and roughly 54,000 the year before.

• Insider dealing after 28 February 2026, the chairman's vehicle added 37,579,000 shares and a disclosed institution 90,989,000 in the preceding six months; the record since then was not reached.

• 7 September 2026, the earliest date on which the 20% minority holder of an acquired subsidiary may put its stake to the group at not less than acquisition-date valuation.

• Any classification-management rule issued by the State Council or a province under the 2021 Implementation Regulations, and any movement on the Sino-foreign qualification requirement the group has never satisfied.

• RMB1.56 (HK$1.82), the 1.00× tangible-book line, the operative buy-below constraint and the level at which the absolute universe filter stops binding.

• RMB4.00 (HK$4.67), the low end of the sell range and the price at which the sell question goes live.

• The certification of RMB651.6m of buildings that currently cannot be sold, transferred or mortgaged; the balance fell RMB44.2m over the half.

• Cohort valuation, with 16 of 21 listed Hong Kong private-education names down more than 70% from peak, a re-rating here most likely arrives as a sector event, not a company one.

SOURCES AND DILIGENCE GAP LOG

Successfully Accessed:

• Annual Report 2025, incl. Independent Auditor's Report, Report of the Directors, Corporate Governance Report and audited consolidated financial statements | annual report | FY ended 31 August 2025 | published 2025-12-30 | HKEXnews, Tianli International Holdings Limited, stock code 1773, read in full as the primary document

• Interim Report 2026, incl. Independent Review Report and interim condensed consolidated financial statements | interim report | six months ended 28 February 2026 | published 2026-05-15 | HKEXnews, stock code 1773, read in full as the primary document

• Poll Results of Annual General Meeting held on 28 January 2026 | exchange announcement | AGM 2026 | published 2026-01-28 | HKEXnews, stock code 1773, resolution-by-resolution poll results, issued-versus-voting share reconciliation, board composition

• Last traded price HK$1.02 and market capitalization HK$2.13bn | market quote | 7 September 2026, 16:08 HKT | https://www.google.com/finance/quote/1773:HKG | SECONDARY

• Quote HK$1.025, year-to-date minus 56.93%, five-day minus 7.24%, market capitalization HK$2.14bn, two covering brokers | market data | 7 September 2026 | https://in.marketscreener.com/quote/stock/TIANLI-INTERNATIONAL-HOLD-45344000/ | SECONDARY, used as the second independent price check

• HKD/CNY mid-market rate, 1 HKD = 0.8559 CNY | FX reference | 7 September 2026 | https://wise.com/us/currency-converter/hkd-to-cny-rate/history | SECONDARY

• Discloseable transaction: a subsidiary to acquire 10% of an art-education group for RMB104m plus a RMB140.8m capital injection, taking the holding to 51% | press account of an exchange announcement | announced 9 to 10 June 2026 | https://cj.sina.com.cn/articles/view/5115326071/130e5ae7702002wd82 | SECONDARY

• Share retirement of 13,227,000 repurchased shares (16 June 2026); resignation of the joint company secretary (31 August 2026) | corporate-actions summary | to 31 August 2026 | https://in.marketscreener.com/quote/stock/TIANLI-INTERNATIONAL-HOLD-45344000/ | SECONDARY

• Broker note maintaining an Outperform rating, FY2026 to FY2028 revenue estimates RMB3,915m / 4,225m / 4,499m and adjusted net profit RMB741m / 821m / 901m | sell-side summary | 12 February 2026 | https://finance.sina.com.cn/stock/hkstock/hkgg/2026-02-12/doc-inhmpyya2953138.shtml | SECONDARY, context only, no estimate from it is used in any derivation

• Sector context: 16 of 21 Hong Kong-listed private-education issuers down more than 70% from peak, 11 trading below HK$1 | press coverage | 17 February 2026 | https://www.finet.com.cn/news/67b31e7b5a771241dd7752ca.html | SECONDARY

• Peer valuation: a Hong Kong-listed private-education comparable at a 2.95 times trailing P/E and an 11.3% dividend yield | market data | April 2026 | https://www.investing.com/equities/china-kepei-education-group | SECONDARY, the peer anchor for the earnings leg in the relevant discussion

• Third-party P/TBV ratio history, 29 observations, August 2018 to February 2026 | ratio series | to 28 February 2026 | SECONDARY, the source for the sell-range distribution; not rebuilt point by point from filings, and disclosed as such wherever it is used

Diligence Gaps, Data Not Found or Estimated:

• Litigation note, checked in the Annual Report 2025 (Report of the Directors, section headed "Litigation"): the group states it had no material litigation outstanding as at 31 August 2025, and no separate legal-proceedings note exists in the financial statements. A finding, not a gap.

• [UNAVAILABLE] Exchange announcement archive, June to September 2026. The exchange's announcement search and several mirror hosts returned 403, 429 or empty responses. Why it matters: this is the six-month window in which the shares fell roughly 45%, and it contains the June 2026 art-education acquisition, the June 2026 share share retirement and the August 2026 company-secretary resignation. Effect: all three are carried as third-party summaries with an explicit footnote, and no figure from any of them enters the tangible-book bridge, the F-Score or the valuation. No primary-source explanation for the share-price collapse was located; the analysis attributes it to the cohort-wide de-rating on peer evidence rather than to any company-specific event, and a reader should treat that attribution as inference.

• [UNAVAILABLE] Disclosure-of-interests register after 28 February 2026. Not reached. Why it matters: the controlling shareholder's vehicle bought 37,579,000 shares and a disclosed institution 90,989,000 in the six months to that date, and whether that continued is a live signal. Effect: all insider and substantial-shareholder positions in this report are stated as at 28 February 2026 and are not updated.

• [UNAVAILABLE] Earnings-call transcript or shareholder question-and-answer record. Not located for either the FY2025 or the H1 FY2026 results. Effect: capital-allocation warning 4 is recorded as UNASSESSABLE rather than as not fired, and the absence of an explanation for the interim dividend suspension could not be tested against management's own words.

• [UNAVAILABLE] Detailed FY2021 to FY2023 cash-flow and debt disclosure. The earlier annual reports were not retrieved, so operating cash flow, capital expenditure, total debt, goodwill and intangibles are populated only for FY2024 and FY2025; earlier years in the fundamentals table come from the five-year Financial Summary in the FY2025 annual report. Effect: the F-Score is computed on FY2025 against FY2024 with FY2023 supplying beginning-of-year assets, which is the standard construction; the normalization window is four years rather than the house default of ten.

• [INFERRED] Normalized earnings basis. FY2021 is an eight-month transition period whose result is dominated by the 2021 deconsolidation, and it is excluded from the normalization on comparability grounds. Why it matters: including it moves normalized net income from RMB408.7m to RMB60.6m and the normalized multiple from 4.4 times to 29.6 times, which moves the base cheapness cell from Big to Small. Effect: both readings are stated in the relevant discussion and in the Final Recommendation; this is the single largest judgment in the report.

• [ESTIMATED] Maintenance capital expenditure. Not disclosed separately; proxied at RMB195.3m for FY2025 as depreciation of owned property, plant and equipment plus amortisation of intangibles. Effect: normalized owner earnings of approximately RMB525.1m is an estimate, labeled as such, and no valuation figure in the relevant discussion depends on it.

• [UNAVAILABLE] Property valuation. No appraisal, revaluation or fair-value disclosure exists for the owned school property portfolio. Effect: any claim that the land and buildings are worth more than the RMB5,046.4m carrying value would be speculation, and none is made.

• [UNAVAILABLE] Debt covenants. Neither primary filing discloses financial covenants, maintenance tests or headroom. Effect: the report states that none is disclosed rather than that none exists, and covenant risk cannot be sized.

• [UNAVAILABLE] Social insurance and housing provident fund shortfall. The company states sufficient provision has been established as at 31 August 2025 but does not quantify it. Effect: a legacy compliance item is carried forward as open-but-provided, unquantified.

Overall Data Quality Rating: GOOD

Both primary filings were retrieved and read in full as primary documents, and the litigation, debt, related-party, subsequent-events, leases, segment, going-concern and auditor-opinion notes were all reached in those primaries, the standard for the top rating. The rating is set one notch below it for one specific reason: the most recent six months of the record, June to September 2026, containing the sharpest part of the share-price decline and three corporate actions, could not be opened at the exchange and rests on third-party summaries, and the disclosure-of-interests register for the same period was not reached. Every figure used in the tangible-book bridge, the F-Score, the debt tests and the valuation comes from a primary filing; nothing from the third-party window enters a derivation.

1. Scorecard

2. Argument

The short version

VICI owns the real estate under Caesars Palace, MGM Grand and 101 other casinos and resorts, collecting rent under leases that run 39.6 years on average and have never missed a payment since the company formed in 2017. The stock has never traded below its own tangible book value in eight years on the New York Stock Exchange. It does now.

That discount is thin, about three percent, and it exists mostly because Caesars, VICI’s biggest tenant at 38% of rent, is being taken private and management will not say what happens to the lease once the deal closes. The balance sheet is fine: leverage sits well inside house limits, the credit rating carries one notch of cushion, and the dividend has been raised every year since the company went public. What has to go right from here is Caesars staying a paying tenant on roughly the same terms once its ownership changes hands.

Why it qualifies

• Trades at 0.968x tangible book value, the cheapest the stock has been since it started trading in 2018.

• 100% of its properties are leased, and rent has been collected in full every year since 2017, including through the pandemic.

• Leverage of 58.0% of equity sits well inside the house ceiling, and net debt runs under five times cash earnings.

• The dividend has been raised every year since the IPO and now yields 7.2%, with no cut or suspension on record.

• Passes the quality screen five of seven measurable ways, with both failures explained by ordinary lease accounting rather than a real problem.

Why it is not a clean bargain

• Caesars, its single largest tenant at 38% of rent, agreed in May 2026 to be taken private, and management would not give a timeline on what that means for the lease.

• Caesars’ own results are softening: its operating profit fell year over year and it posted a net loss last quarter.

• The company cannot grow by reinvesting its own cash. Almost everything it earns gets paid out as dividends, so new acquisitions depend on selling stock or borrowing, and it has mostly stopped selling stock below book value.

• No insider bought a single share on the open market in the past year, even as the stock hit a 52-week low.

• The margin of safety is thin, about three percent below tangible book, not a deep discount.

• It just recorded its first-ever troubled loan: a small golf-resort loan now on non-accrual.

What would change the view

• Any amendment to the Caesars master lease, or a rent cut tied to the change in Caesars’ ownership.

• A ratings downgrade below investment grade at Moody’s, S&P or Fitch.

• A second loan going onto non-accrual, which would suggest the first wasn’t a one-off.

• The stock rising back above roughly $35.00 a share, which would put it in sell-range territory instead of buy territory.

• Resumption of stock issuance below tangible book value, which would signal management is prioritizing growth over protecting per-share value.

Sources and gaps

Some detail on executive pay and related-party dealings comes from a large annual filing that wasn’t read start to finish; the parts that were reached showed nothing unusual. The historical valuation range used to set a target sell price comes from an outside data series, not independently rebuilt figure by figure. See the full detail for the source ledger, calculations, and complete diligence-gap log.

3. Backup and sources

This last layer is the full working file: every source, calculation, and gap, kept so the argument above can be checked. The scorecard and the argument are the synthesis. Open this only if you want to check the work.

VICI Properties Inc.: Full Analysis

1. Executive Summary

VICI Properties is a Maryland REIT that owns the land and buildings under 103 gaming, hospitality and leisure destinations across 26 US states, one US territory and one Canadian province: Caesars Palace Las Vegas, MGM Grand, Mandalay Bay and the Venetian Resort among them: and makes essentially all of its money collecting contractual rent from sixteen operating tenants under long-term triple-net leases in which the tenant pays every property cost, including taxes, insurance, maintenance and capital expenditure (Form 10-K FY2025, Item 1: Business; Q2 2026 financial supplement, portfolio summary). The verdict is BUY — SMALL. At the 3 September 2026 close of $25.65 the stock trades at 0.968× the tangible book value derived from the 30 June 2026 balance sheet ($26.4930 per share), the first time in the eight years of quarter-end history reviewed here that it has priced below its own tangible book, and it clears both of my absolute screens: P/TBV at or under 1.00×, and financial borrowings of $16,931.2 million against $29,170.7 million of common equity, or 58.0%: comfortably inside the 100% ceiling I use (Form 10-Q Q2 FY2026, consolidated balance sheets).

What the buyer gets is an unusually clean asset: no goodwill, no intangibles, no inventory, 0.2% of tangible book in right-of-use assets, and a book value already stated net of a $1,925.8 million expected-credit-loss allowance equal to 3.84% of the $49.5 billion amortised cost of the lease and loan portfolio (Form 10-Q Q2 FY2026, Note 5: Allowance for Credit Losses). Cash generation is heavy and predictable: $2,510.0 million of operating cash flow in FY2025 against $1.3 million of capital expenditure, because the tenants fund the buildings (Form 10-K FY2025, consolidated statements of cash flows). Adjusted funds from operations were $2.38 per share in FY2025 and are guided to $2.45 to $2.47 for FY2026, so the shares change hands at roughly 10.8 times trailing and 10.4 times forward cash earnings, against a $1.84 annualised dividend: a 7.2% yield after the 2.2% increase declared for the September 2026 quarter.

The margin of safety is real but thin: 3.2% against tangible book. The case rests more on the gap between a 7.8% implied capitalisation rate on the rent roll and the 8.9% weighted average initial yield VICI achieved on FY2025 commitments than on an asset discount. Three things argue against sizing this larger. Concentration: Caesars and MGM together provide 70% of annualised rent ($1,246.2 million and $1,065.0 million of $3,312.2 million), and the Las Vegas Strip produced 49% of FY2025 lease revenue. Change of control: Caesars agreed on 28 May 2026 to be taken private by Fertitta Entertainment, and management has declined to give a timeline on the associated lease discussions. And dilution: the share count has risen 75% since FY2021. Balance-sheet risk is Manageable but Baa3/BBB−/BBB− leaves one notch of cushion; accounting quality is Adequate, with one large judgemental estimate: the credit-loss allowance, also the auditor’s sole critical audit matter; governance risk is Low, with none of the five capital-allocation flags I check firing.

2. Business and Market Overview

VICI revenue is rent. FY2025 total revenues of $4,006.1 million comprised $2,125.4 million of income from sales-type leases, $1,763.5 million of income from lease financing receivables, loans and securities, $77.5 million of other income that is an exactly offsetting gross-up of ground-lease costs passed through to tenants, and $39.8 million of golf revenue from four owned courses (Form 10-K FY2025, consolidated statements of operations). Operating costs are trivial by design: general and administrative expense was $65.1 million, or 1.6% of revenue, and the company employed 28 people at 31 December 2025 (Form 10-K FY2025, Item 1: Human Capital Management). Everything else on the income statement is interest, non-cash credit-loss provisioning, and depreciation of $3.6 million.

The lease structure is the business. Seventeen leases, fifteen of which carry CPI-linked escalation for some part of their life, initial terms of 15 to 32 years with tenant renewal options of a further 5 to 30 years, and a weighted average lease term including options of 39.6 years at 31 December 2025. Forty-two per cent of FY2025 rent and roughly 90% of rent over the long term escalates with CPI subject to caps; the balance escalates at a fixed 1% to 2%. Occupancy is 100% and management states rent collection has been 100% since formation in October 2017, including through the pandemic (Form 10-K FY2025, Item 1).

Customer concentration is the defining exposure and is disclosed plainly. MGM lease agreements produced 38% of FY2025 lease revenues and 36% of contractual rent; Caesars 36% and 37% (Form 10-K FY2025, Note 2: Concentrations of Credit Risk). On the July 2026 rent roll the ranking inverts slightly: Caesars $1,246.2 million (38%), MGM $1,065.0 million (32%), the Venetian tenant $308.7 million (9%): with no other tenant above 4.4% of the $3,312.2 million total. Both master-lease families are guaranteed at parent level by Caesars Entertainment, Inc. and MGM Resorts International, so the credit runs to the public operator rather than to the property.

Diversification is the stated strategy and is visibly progressing. As of 30 June 2026 the tenant count reached sixteen: the $1,148.0 million Golden Entertainment sale-leaseback of seven Nevada casinos closed 30 April 2026 at $87.0 million of initial annual rent on a 30-year master lease; a severance lease for MGM Northfield Park with a Clairvest affiliate began 21 April 2026 at $53.0 million; the C$200.6 million Gamehost acquisition in Alberta closed 24 June 2026, adding C$16.1 million of rent to the PURE master lease; and Club Med became the sixteenth tenant on 15 June 2026 through a $20.3 million purchase of the Carambola Beach Resort plus a $55.2 million redevelopment commitment (Form 10-Q Q2 FY2026, Note 3). A parallel $2.9 billion credit book at a 9.1% weighted average rate is originated as a route into future ownership, most conspicuously a $1.5 billion mezzanine loan on the One Beverly Hills development advanced 23 March 2026.

The business is understandable, highly predictable and structurally low-capital-intensity for the landlord, with heavy regulatory exposure only indirectly, through the gaming licences its tenants hold. It is not a melting ice cube; the honest risk is not obsolescence but tenant credit and the terminal value of leases whose contractual escalators may lag replacement cost over a forty-year horizon.

3. Moat, Competitive Position, and Industry Cycle

Against the five moat sources, VICI scores clearly on two and partially on one. Efficient scale is genuine: approximately 130 million square feet, 66,000 hotel rooms and roughly 33 acres of undeveloped Strip-adjacent land, assets that are expensive to replicate and, on the Las Vegas Strip, effectively irreplaceable, with the gaming licensing regime restricting who may operate on them. Switching costs are extreme in the only sense that matters to a landlord: a tenant that has spent decades building a branded destination on a site cannot move it, and the master-lease structure bundles strong and weak assets so it cannot cherry-pick which to renew. Cost advantage is partial and cyclical: investment-grade ratings lower the cost of capital relative to private buyers, but that is a function of the credit cycle rather than anything proprietary. Brand and network effects are absent; the tenants own the brands and the customers.

The evidence sits in the numbers rather than the narrative: 100% occupancy, 100% collection through COVID, a 39.6-year weighted average lease term, and revenue compounding from $1,509.6 million in FY2021 to $4,006.1 million in FY2025 without a single lease default. Against that stands a hard qualification: VICI cannot compound capital internally: a REIT distribution requirement and a 73% AFFO payout leave almost nothing retained: so incremental return depends entirely on issuing securities above the economic cost of the assets bought. When the equity trades below tangible book, as today, that engine stalls. Management conceded the point on the Q2 2026 call, arguing that with the loan book yielding roughly 9.5% capital is better deployed into credit than into repurchasing its own shares.

Moat: Moderate. Real and durable in the assets and the lease documents, but dependent on external capital markets rather than on internal reinvestment economics, and capped by the fact that a landlord captures the escalator, not the operating upside.

Cycle position: Below mid-cycle. Las Vegas and regional gaming fundamentals have softened: Caesars reported Q2 2026 consolidated Adjusted EBITDA of $920 million against $955 million a year earlier and a $62 million net loss: and the REIT sector is absorbing a higher-for-longer rate environment; VICI shares are down roughly 18.5% over twelve months and set a fresh 52-week low of $25.34 in the week of this analysis. Rent itself has not declined: escalators are contractual and the FY2026 AFFO guide is up about 3% per share on FY2025. The cycle pressure sits in the discount rate applied to that rent and in tenant coverage, not yet in the cash flow.

4. Management, Governance, and Capital Allocation

Edward B. Pitoniak has been Chief Executive Officer since the company was formed in 2017 and is the only non-independent director; John W. R. Payne is President and Chief Operating Officer, David A. Kieske Executive Vice President, Chief Financial Officer and Treasurer, and Samantha S. Gallagher Executive Vice President, General Counsel and Secretary: the same four throughout the period reviewed. James R. Abrahamson chairs an otherwise fully independent board of seven (Definitive Proxy Statement filed 16 March 2026). On 24 February 2026 the board appointed Jeremy L. Waxman, a seven-year employee and former Ernst and Young senior manager, as Vice President, Chief Accounting Officer effective 1 March 2026; the outgoing principal accounting officer, Gabriel F. Wasserman, moved to an expanded business-development role rather than departing, and the filing states there are no arrangements, no family relationships and no related transactions associated with the appointment (Form 10-K FY2025, Item 9B): the benign version of this event, not the unexplained-departure version.

On 25 February 2026 the company entered amended and restated employment agreements with all four named executives, removing the fixed terms and associated non-renewal severance: a shareholder-favourable change: applying uniform 12-month non-compete covenants, and setting 2026 base salaries of $1,000,000 (Pitoniak), $1,200,000 (Payne), $670,000 (Kieske) and $648,000 (Gallagher) with target bonuses of 225%, 135%, 150% and 150% of salary. Total pay is large relative to the payroll it sits on: the 2025 chief-executive Summary Compensation Table total was $14,007,585 and the average for the other three $5,691,972, so roughly $31 million of the $65.1 million FY2025 general and administrative line is four people at a 28-employee company. Against $2,510.0 million of operating cash flow that is 0.56%, in line with large-cap REIT practice, and the April 2026 say-on-pay vote passed with 94.9% support.

Insider activity over the last twelve months, from every Form 4 filed between January and July 2026, shows no open-market purchase and no open-market sale by any officer or director. The filings cluster on 2 January, 24 February, 2 April and 1 July 2026 and consist of quarterly director retainer grants, the annual executive equity award, and shares surrendered for tax withholding on vesting restricted stock. The two read in full are representative: Ms Gallagher surrendered 3,554 shares at $29.87 and 4,671 at $30.09 on 20 and 23 February 2026, then received a 30,309-share award on 24 February, ending at 368,018 shares; Mr Pitoniak gifted 20,000 shares to a non-profit educational institution on 1 June 2026, leaving 1,291,210. Absence of insider buying with the stock at a 52-week low is a non-confirmation rather than a red flag; directors face a five-times-retainer ownership guideline and the company maintains anti-hedging, anti-short-sale and anti-pledging policies.

Capital allocation has been consistent and disciplined. There have been no share repurchases in any period reviewed: the only amounts so labelled are $7.2 million (FY2025), $5.3 million (FY2024) and $5.0 million (FY2023) of shares taken back for tax withholding: and management said on the Q2 2026 call that buybacks make limited sense for a capital-dependent REIT with a 9.5%-yielding loan pipeline. The dividend has been raised every year of the company’s public life, from $1.610 per share declared in FY2023 to $1.695 in FY2024 to $1.765 in FY2025 and, as of the September 2026 declaration, to $0.46 per quarter: a 2.2% increase and a $1.84 annual rate payable 8 October 2026 to holders of record on 17 September 2026. There has never been a cut or a suspension. Equity issuance has tracked the share price: $2,385.8 million, $3,219.1 million and $2,480.1 million in FY2021 to FY2023, falling to $378.7 million and $375.3 million in FY2024 and FY2025. FY2025 commitments of approximately $2.1 billion were struck at a weighted average initial yield of 8.9%, above the 4.454% weighted average cost of debt: which is why a 75% rise in share count since FY2021 has coincided with tangible book per share rising from $19.25 to $26.49.

Related-party review found nothing to report, and the search was independent rather than inherited from an outside narrative. The 2026 proxy carries a written related-party transaction policy and states there are no material related-party transactions and no family relationships among directors or executive officers; the FY2025 annual report contains no related-party note and discloses no intercompany or affiliate lending, guarantee, deposit or nominee arrangement. On the financial-relationship limb specifically: deposits with a related-party-affiliated bank, facilities either way, guarantees either way: the filings disclose none, and the guarantees that exist run towards VICI from Caesars and MGM. The only affiliate-adjacent arrangements are ordinary-course: the golf-course management agreement with Cabot-Managed Properties, an affiliate of Cabot to which VICI is also a lender and the counterparty on an agreement converting part of a $120.0 million Cabot Citrus Farms development loan into owned real estate, and the Golf Course Use Agreement with Caesars at minimum fees of $17.6 million a year. The Cabot overlap is worth watching because VICI is landlord, lender and prospective buyer to one counterparty; nothing suggests off-market pricing, but the concentration of roles is worth stating rather than leaving implicit.

None of the five governance red flags I check fired here. No holder controls the company: the largest position is Vanguard’s 7.86%, and all seven directors were elected individually under a majority-voting standard at the 28 April 2026 annual meeting, with the lowest-supported nominee still winning 96.4% of votes cast. There is no parent to route capital toward, and FY2025’s acquisitions were struck at an 8.9% initial yield. The dividend has been raised in every year of the company’s public life and yields 7.2% today. Management did decline, on the Q2 2026 call, to give a timeline for the Caesars lease discussions or to pre-commit on regional acquisitions, but both read as ordinary confidentiality around live, sensitive negotiations rather than evasiveness: the capital-allocation question that actually mattered, why the credit allowance rose, got a specific, checkable answer. Stock-based compensation ran 0.58% of net income in FY2025 and 0.65% in FY2024, an order of magnitude below what would concern me, and falling rather than rising. Governance risk here is Low.

5. Corporate Ownership, Subsidiaries, and Joint Ventures

The ownership register is institutional and unconcentrated. Vanguard’s fund-management arms hold the two largest positions: 7.86% and 6.91%: after a January 2026 internal realignment moved The Vanguard Group’s own directly-reported stake to zero. Capital Research Global Investors holds 4.8% and Capital International Investors 1.8%. Chief executive Edward Pitoniak holds 1,291,210 shares, about 0.12% of the class, for scale against the institutional positions above. No holder controls the company, no dual-class structure exists, and the largest single reported position is under 8%.

There is one class of common stock carrying one vote per share, 1,101,074,906 shares issued and outstanding at 30 June 2026, no preferred stock outstanding against 50,000,000 authorised, and 350 holders of record at 24 February 2026. No control block, no shareholder agreement and no acting-in-concert arrangement is disclosed anywhere in the record.

The structure below the REIT is simple but not trivial. Substantially all real property is held through VICI Properties L.P., a consolidated operating partnership that is a co-registrant and publishes its own audited statements inside the same annual report; golf sits in a taxable REIT subsidiary, VICI Golf LLC. Non-controlling interests of $431,962 thousand at 30 June 2026: 1.5% of total equity: are third-party operating-partnership units plus the 20% minority in Harrah’s Joliet Landco LLC, in which a VICI LP subsidiary is the 80% owner and managing member. The remaining 49.9% of the MGM Grand/Mandalay Bay joint venture was acquired in January 2023, so that vehicle is now wholly owned; its $3.0 billion 3.558% CMBS financing runs to March 2032 and is the only secured debt in the capital structure: the one genuine ring-fencing feature, with 17.4% of borrowings sitting against two named assets ahead of the unsecured creditors. Minority leakage is small: $43.1 million of FY2025 net income and $32.2 million of distributions. The MGM tax protection agreement is a real constraint on deleveraging, covered in the debt discussion below.

6. Historical Financial Quality and Normalized Owner Earnings

Five years of audited history show a company that roughly tripled in size and then stabilised. Revenue rose from $1,509.6 million in FY2021 to $4,006.1 million in FY2025, almost all of it acquired: the April 2022 acquisition of MGM Growth Properties and the February 2022 Venetian purchase took total assets from $17,597.4 million to $37,575.8 million and then $44,059.8 million at FY2023. Since then the balance sheet has grown 6% in two years while revenue grew 11%. The FY2022 earnings dip, diluted EPS of $1.27 against $1.76 and $2.47 either side, is not an operating event: it is the $834.5 million initial credit-loss allowance ASC 326 requires on day one of a large acquisition

.

Gross margin is shown as unavailable rather than as zero: the income statement carries no cost of revenue, and the concept does not exist for a triple-net lessor. Operating income above is total revenues less total operating expenses as presented. FY2025 and FY2024 were spot-checked line by line against the consolidated statements of operations in the FY2025 annual report, and the FY2022 and FY2021 comparatives against the FY2023 annual report.

Normalized owner earnings. The cash-flow statement is used rather than a proxy. FY2025 operating cash flow was $2,510.0 million against $1.3 million of capital expenditure: tenants fund maintenance and improvement under every lease: so free cash flow was $2,508.7 million, or $2,476.5 million after $32.2 million of distributions to non-controlling interests. On the FY2025 weighted diluted count of 1,062.7 million shares that is $2.33 per share of owner earnings, against the company’s own adjusted funds from operations of $2.38; the gap is $46.8 million of debt-issuance-cost amortisation and $16.2 million of stock compensation that AFFO adds back, less the minority distribution this measure deducts. Reported diluted EPS of $2.61 is the least useful: GAAP income includes $523.9 million of non-cash lease accretion and deducts a $177.9 million non-cash provision. A symmetric one-off check finds nothing material either way.

Normalized EPS and its basis. The trajectory is growing, not cyclical, so a ten-year average is inappropriate: there are only eight years of public history and the weighted diluted count has risen 84% since FY2021, making any multi-year per-share average an average of different companies. The figure used in the valuation discussion below is therefore $2.38, FY2025 adjusted funds from operations per diluted share: a full audited year, on the current share base, on the cash measure the business distributes from. Alternatives, so a reader can substitute: five-year average GAAP diluted EPS $2.13, trailing GAAP diluted EPS $2.58, FY2026 guided AFFO $2.45 to $2.47. At the current price these give normalized multiples of 10.8, 12.0, 9.9 and 10.4 times: all landing in the same cheap-and-modest-multiple corner, so the verdict does not turn on the choice.

The credit-loss allowance roll-forward. VICI carries a large loan-loss-equivalent account, so it is worth reconstructing from the note rather than inferring from the balance-sheet movement. The reserve began FY2023 at $1,368.8 million and has only ever grown: to $1,472.4 million at the end of FY2023, $1,594.9 million at FY2024, $1,775.8 million at FY2025 and $1,925.8 million at 30 June 2026. FY2023’s $293.0 million initial allowance on new investments reflects the day-one CECL charge on the MGM and Venetian acquisitions; the ongoing quarterly change in the allowance ($172.5 million in FY2025, $37.6 million in the first half of 2026) accounts for the rest of the growth.

The most important fact in that roll-forward is what never appears in it: there have been no charge-offs and no recoveries in any period since adoption, so the provision-to-payout ratio is not merely above 1.0: it is undefined, the denominator being zero. The reserve has only ever been built. That is the opposite of the pattern this kind of check exists to catch: the $1,925.8 million deducted from tangible book is an unused cushion, not a depleted one. It was 3.72% of $47,537.4 million of amortised cost at 31 December 2025 and 3.84% of roughly $49.5 billion at 30 June 2026. There is no off-balance-sheet guarantee reserve to compare against; the unfunded-commitment allowance, $22.9 million against $6.4 million at year-end, moved the same way, so the asymmetric-reserving check is clean.

Two qualifications. The allowance is a model output, not an observation: a discounted-cash-flow estimate driven by probability-of-default and loss-given-default assumptions from a third-party provider and keyed to guarantor credit ratings, and it is the auditor’s sole critical audit matter. Its volatility is real: a $271.1 million charge in Q2 2026 alone, driven, management explained, by a private tenant issuing senior secured debt at a lower rating than the proxy company previously used in the model, with the property described as performing. Second, VICI recorded its first credit event in Q4 2025: a fully funded senior secured loan of $82.8 million on a luxury golf-resort development was placed on non-accrual, the borrower in recapitalisation discussions: 0.17% of amortised cost, immaterial to book value, material as a first data point.

Earnings quality overall. Conversion is exceptional: $2,510.0 million of operating cash flow on $4,006.1 million of revenue, essentially all free: and receivables of $34.2 million leave no working-capital trap, because rent is paid in advance. The one structural concern is that GAAP net income exceeds cash by $265.5 million, which is what fails the cash-flow quality check below; retained earnings of $3,189.4 million are correspondingly the accumulated non-cash accretion, $1,848.9 million of it since FY2023 alone. Book-value growth from retained earnings is accounting accretion on assets already owned, not reinvested cash.

Piotroski F-Score: 5 of 7 computable signals (FY2025 versus FY2024), Eligible. Five passed: positive net income ($2,775.5 million), positive operating cash flow ($2,510.0 million), an improved return on assets (6.1% versus 6.1%, unchanged and counted as a pass on this method’s convention), debt that did not increase (36.4% of average assets versus 37.4%), and improved asset turnover (0.088 versus 0.087). Two failed: cash flow came in below net income ($2,510.0 million against $2,775.5 million), and net share issuance was positive rather than a buyback (a net $368.1 million issued, $7.2 million of tax-withholding repurchases against $375.3 million issued). Two of the nine standard signals are not computable at all for this business: VICI presents an unclassified balance sheet with no current-asset or current-liability subtotals, so no current ratio exists, and there is no cost of revenue on the income statement, so no gross margin exists. Marking either a failure rather than not-computable would be a fabricated result, and seven computable signals is enough for a real score.

The two failures are both structural, not warning signs. Cash flow trails net income because sales-type lease accounting recognizes rent as non-cash interest accretion on part of the portfolio; that will fail every year the accounting stays the same, on contracts collected in full and on time. Net issuance is negative because VICI funds acquisitions by selling shares above book, not because of dilution for its own sake, and tangible book per share rose from $19.25 to $26.01 over the same five years. The eligible rating held: leverage on this method’s definition actually decreased, from 37.4% to 36.4% of average assets, so the rule that would reject a name outright on rising leverage never engaged.

7. Balance Sheet, Debt, Covenants, and Refinancing Risk

The capital structure is unusually simple for $17 billion of borrowings: one secured mortgage financing, one revolver, and eleven series of senior unsecured notes. The schedule below runs instrument by instrument at 30 June 2026 unless stated otherwise

:

Disaggregation before classification. Financial borrowings are $16,931.2 million of carrying value and $17,218.4 million of principal. Lease liabilities of $912.5 million are a separate bucket and are not promoted into the debt test: $862.5 million is finance sub-lease liability on ground leases whose rent VICI’s own tenants pay directly to the primary landlord, offset by an $835.4 million sub-lease asset. On the governing measure, financial borrowings are 58.0% of the $29,170.7 million of common equity; including lease liabilities, 62.2%; total liabilities of $18,668.5 million are 64.0%. All three are inside the 100% ceiling I use, so the debt screen passes on any reading. No current ratio can be computed: the balance sheet is unclassified: so the liquidity check runs on $2.5 billion of total liquidity instead.

Maturity wall and refinancing. The 2026 wall of $1,750.0 million has already been dealt with. On 5 August 2026 VICI agreed to sell $900.0 million of 5.400% notes due 15 October 2031 at 99.966 and $850.0 million of 5.750% notes due 15 October 2036 at 98.375; on 14 August 2026 the offering closed under a fifth supplemental indenture, roughly $1.72 billion of net proceeds earmarked to retire $480.5 million of 4.500% notes, $19.5 million of assumed 4.500% MGP notes and the $1,250.0 million of 4.250% notes due December 2026. Exact but not free: a 5.570% blended coupon against the retired 4.321% adds roughly $21.9 million of annual interest, about a cent and a half of AFFO. The next maturity is $1,500.0 million in February 2027, then $2,000.0 million in 2028, $1,892.5 million in 2029 and $2,000.0 million in 2030: laddered, no year above 12% of borrowings.

Coverage and covenants are comfortable. Net debt to annualised second-quarter adjusted EBITDA was 4.9 times at 30 June 2026, below the low end of the company’s own 5.0-to-5.5-times target; interest coverage was 4.1 times against a 1.5-times covenant floor; unencumbered assets to unsecured debt stood at 319% against a 150% minimum. The company states it was in compliance with all financial covenants at 30 June 2026, and the investment-grade tranches have suspended most restrictive covenants, leaving the unencumbered-assets maintenance test as the binding one. Ratings are Baa3 from Moody’s (upgraded from Ba1 in November 2024), BBB− from S&P and BBB− from Fitch, all stable: investment grade at all three, with exactly one notch of cushion.

Balance-sheet risk: Manageable. Financial borrowings are 35.1% of total assets, squarely in the middle of what I’d call an ordinary range, and 98.4% of debt is fixed-rate with a 5.5-year weighted average maturity. Interest coverage of 4.1 times would formally read as Elevated on an industrial-company scale, and I want to state that caveat rather than wave it away: a triple-net REIT collecting contractual rent from investment-grade-guaranteed tenants on 39.6-year leases carries more debt at the same underlying risk than a manufacturer would, which is why three agencies rate this structure investment grade at 4.9-times net leverage. The genuine constraints are the MGM tax protection agreement, which puts a floor rather than a ceiling under leverage, and the single notch of ratings headroom.

8. Real Estate, Leases, and Hidden Assets

Right-of-use assets are $50.3 million, or 0.2% of tangible book value: operating ground and use sub-leases plus the New York head office and the Cascata golf land. Matching liabilities are $50.1 million operating and $862.5 million finance sub-lease, the latter at a 50.3-year weighted average term and 5.6% to 8.3% discount rates, offset by an $835.4 million sales-type sub-lease asset presented gross because VICI is primary obligor on ground leases its tenants pay directly. Undiscounted commitments total $30.3 million, $37.6 million and $2,956.4 million respectively; the fixed-charge burden is $65.2 million a year against $3.3 billion of rent received, about 2%. Right-of-use assets are a capitalised right to occupy, not a liquidation floor, and I’m disclosing that here rather than stripping the asset out of the tangible book value used throughout: removing the asset without its matching liability would be the wrong operation. At 0.2% of tangible book the question is immaterial, so calling this cheap against tangible book value below needs no further qualification.

What the tangible book is made of matters here. Ninety-nine per cent of assets: $47,610.1 million of $48,271.2 million: are the real estate portfolio, but almost none is presented as property: sales-type leases $24,577.4 million, lease financing receivables $19,280.1 million, loans and securities $2,917.3 million, sales-type sub-leases $835.4 million, against only $238.1 million of land, development and operating equipment. Three consequences follow. There is no accumulated depreciation reducing book value, which is why tangible book per share rises rather than erodes. The carrying amounts are amortised cost, so book is a cost anchor, neither a market mark nor a depreciated one. And it is stated net of a $1,925.8 million credit-loss allowance never drawn against.

Hidden value exists but is modest and mostly optional. The most concrete item is approximately 33 acres of undeveloped or underdeveloped land on and adjacent to the Las Vegas Strip, leased to Caesars and carried inside a $148.0 million total land line, which management describes as monetisable as appropriate. Four championship golf courses sit inside $66.9 million of operating property and equipment and generate $39.8 million of revenue at a 33% margin. The embedded growth pipeline is optionality with disclosed pricing rather than hidden value: a call right on the Caesars Forum Convention Center at a 13.0-times multiple, a 7.7% cap rate, exercisable to December 2028, plus rights of first refusal over Flamingo, Paris, Planet Hollywood, Bally’s, The LINQ, Horseshoe Baltimore and Caesars Virginia, and rights of first offer with Canyon Ranch, Lucky Strike, Homefield and Indigenous Gaming Partners. The counterweight is that a 39.6-year weighted average lease term puts reversionary value very far away.

9. Capital Markets Access, Dilution, and Financing Flexibility

Dilution is the central structural cost of this business and it is large. Weighted average diluted shares went 577.1 million (FY2021) to 879.7 million (FY2022) to 1,015.8 million (FY2023) to 1,047.7 million (FY2024) to 1,062.7 million (FY2025), and the point-in-time count reached 1,101,074,906 at 30 June 2026: an 84% increase on the weighted diluted measure in five years. Net issuance was negative in every year: $2,385.8 million, $3,219.1 million, $2,480.1 million, $378.7 million and $375.3 million of gross issuance against tax-withholding repurchases of $1.7 million, $6.2 million, $5.0 million, $5.3 million and $7.2 million. There is no buyback programme and management has said it does not intend to run one while it can lend at 9.5%.

The mechanics are disciplined even if the volume is not. Equity is raised almost entirely through forward sale agreements under a $2.0 billion at-the-market programme established 6 May 2024: 7,835,973 shares were priced in FY2025 at a $32.43 weighted average, and 12,101,372 forward shares settled during FY2025 for $375.7 million. As of 30 June 2026 there were no forward shares outstanding: the last 7,750,000 settled on 29 April 2026 for $242.1 million: and no at-the-market activity occurred in the first half, consistent with management declining to issue below book. The 24.3 million shares issued on 30 April 2026 as Golden Entertainment consideration were acquisition currency at a fixed 0.902 exchange ratio, equally dilutive, and explain most of the 32.3 million-share increase in the half. Overhang is negligible: the Q2 2026 diluted count exceeded basic by 39,531 shares.

The company does not need the capital markets to survive, only to grow. FY2025 operating cash flow of $2,510.0 million covered $1,853.5 million of dividends with $656.5 million to spare and essentially no maintenance capital call; $2.5 billion of total liquidity at 30 June 2026 comprises $288.1 million of cash and $2.2 billion of undrawn revolver to February 2029. The August 2026 bond issue demonstrates unimpaired debt-market access at investment-grade spreads. Dilution risk as a judgment: moderate and self-limiting: the shares can only be issued accretively above book, and below book management has visibly stopped, which protects book value per share but also caps growth precisely when the stock is cheapest.

10. Litigation, Regulatory, and Contingent Liability Risk

Two primary sources were reached in full and both are named. Item 3, Legal Proceedings of the FY2025 annual report states that “As of December 31, 2025, we are not subject to any litigation that we believe could have, individually or in the aggregate, a material adverse effect on our business, financial condition or results of operations, liquidity or cash flows.” Note 10, Commitments and Contingent Liabilities repeats that language verbatim with no accrued loss reserve of any kind, and the corresponding note in the Q2 2026 quarterly report repeats it again as of 30 June 2026. That is boilerplate in form but boilerplate with nothing behind it: no named defendant, no class action, no regulatory proceeding, no tax dispute, no environmental matter anywhere in the annual report.

Negative confirmation (Form 10-K FY2025, Item 3 and Note 10: Commitments and Contingent Liabilities): ordinary-course claims only; no material litigation, investigations or regulatory actions disclosed; no accrued loss reserves.

The prior-disclosure carry-forward check produced nothing to carry. The FY2025 annual report discloses no subpoena, investigation, class action, covenant waiver, going-concern language or material weakness; the auditor issued unqualified opinions on both the financial statements and internal control over financial reporting. The Q2 2026 quarterly report contains no subsequent-events note and no going-concern language: every report page in that filing was checked to confirm the absence rather than infer it. The honest caveat is that a clean two-period record is weaker confirmation than a clean five-period one: the FY2023 and FY2024 annual reports were read for their financial statements, not their legal-proceedings items.

What can take a bite out of tangible book is contractual rather than litigious. Future funding commitments on the debt book were $623.5 million at 31 December 2025, subject to borrower covenant compliance, plus $55.2 million for the Club Med redevelopment. The MGM tax protection agreement indemnifies MGM against tax liabilities triggered by disposal of a protected property, by a transaction requiring exchange of MGM’s partnership interests, or by failure to maintain approximately $8.5 billion of non-recourse indebtedness allocable to MGM, for fifteen years from April 2022: uncapped and unquantified, constraining both asset sales and deleveraging, with a separate agreement to mid-2029 covering built-in gain on MGM Grand and Mandalay Bay. Ground and use lease obligations total $2,956.4 million undiscounted, for which VICI is primary obligor even though tenants pay the landlord directly. The $82.8 million non-accrual loan is the only disclosed impaired asset.

Regulatory exposure is real but second-order. VICI is not a gaming operator, but it holds licences or findings of suitability where its tenants operate, so a licensing problem at a tenant is a rent problem here. The larger dependency is tax: at least 90% of REIT taxable income must be distributed and the asset and income tests satisfied, and failure would expose the company to corporate tax at regular rates: managed by paying out 73% of AFFO and housing golf in a taxable REIT subsidiary. Item 1 flags that prediction markets and similar platforms currently operate under federal rather than state regulation, giving them an advantage over licensed tenants; that is a slow-acting threat to tenant economics rather than a contingent liability, and it is carried into the risk list below as such.

11. Accounting Quality and Disclosure Review

The auditor is Deloitte & Touche LLP, in place since 2016, and the FY2025 opinions are unqualified on both the consolidated financial statements and on internal control over financial reporting, each dated 25 February 2026. There is no going-concern paragraph, no material weakness, no restatement in the period reviewed, and management’s own conclusion agrees. Separate audited statements are issued for VICI Properties L.P. inside the same annual report, with the same opinions.

There is exactly one critical audit matter, and it is the right one: the allowance for credit losses. The auditor describes a discounted-cash-flow model projecting expected losses from probability-of-default and loss-given-default inputs drawn from a third-party provider, keyed over a two-year window to the current condition of tenants and their parent guarantors and thereafter to historical default and loss rates of comparable public companies. The procedures disclosed are substantive: controls over the model and its data were tested, credit specialists evaluated the methodology and assumptions, the credit rating and equity value of each guarantor were agreed to independent data, and cash-flow inputs were reconciled to the contracts. That last step matters, because the Q2 2026 charge arose from exactly such a data change.

Revenue recognition is the second area of judgment and is structural rather than discretionary. Whether a lease is classified as sales-type or as a financing receivable at inception determines whether cash rent arrives as interest accretion or as rent, and it is why $523.9 million of FY2025 revenue never became cash. The classification is made once, on the economics of the contract, and is disclosed, consistent and reconciled in the cash-flow statement. Non-GAAP presentation is disciplined: funds from operations and adjusted funds from operations, with add-backs named and traceable: non-cash lease accretion (a deduction, the conservative direction), the credit allowance, debt-issuance-cost amortisation, stock compensation and transaction costs. Only the $7.7 million of transaction expenses is genuinely non-recurring, and the rest are labelled as recurring.

Impairment risk to book value is nil from the usual source. There is no goodwill and no intangible asset on the balance sheet in any period, so no part of tangible book is exposed to a write-down; the analogous exposure is the credit allowance, already deducted. Deferred tax assets of $10.5 million are immaterial for a REIT that paid $7.3 million of cash tax in FY2025, and no pension obligation is disclosed. Segment reporting is thin but adequate: real property and golf, the latter 1% of revenue: supplemented by tenant-level rent disclosure. Related-party disclosure is where the annual report is quietest: no related-party note at all, with the proxy’s substance being a policy and a negative statement rather than a schedule.

Disclosure quality: Adequate. Unqualified opinions with no material weakness and near-full primary verification would support a stronger rating but for two things: the sole critical audit matter is a genuine and large estimation-risk area whose output moved reported earnings by $271.1 million in a single quarter, and the proxy’s compensation tables and audit-fee schedule were not reached in full. Neither is a concern about integrity; both are reasons not to claim the top rating.

12. Valuation and Margin of Safety

Tangible book value, derived from the 30 June 2026 balance sheet. Common shareholders’ equity was $29,170.7 million (Form 10-Q Q2 FY2026). VICI carries no goodwill and no other intangible assets in any period since formation, so tangible book value equals common equity exactly: $29,170.7 million. Dividing by the 1,101,074,906 shares issued and outstanding at 30 June 2026 (a point-in-time count, not a weighted average: no filing-supported dilutive increment exists at that date, and the Q2 2026 diluted count exceeded basic by only 39,531 shares) gives $26.4930 per share. The only non-tangible deferred charges anywhere are $14.5 million of unamortised revolver financing costs and $6.8 million of deferred acquisition costs, together $0.02 per share, which would move the multiple from 0.968× to 0.969× if backed out: immaterial, and not applied. The $431,962 thousand of non-controlling interest is excluded by construction: the valuation uses equity attributable to VICI’s own shareholders, not the consolidated total.

P/TBV (derived) = $25.65 ÷ $26.4930 = 0.968× (price as of 2026-09-03).

At the 30 June 2026 anchor, cash and short-term investments were $288.1 million (1.0% of tangible book), accounts receivable $34.2 million (0.1%), owned property, plant and equipment $238.1 million (0.8%), right-of-use lease assets $50.3 million (0.2%, disclosed and never stripped from the figure above), and long-term investments: the bulk of the real estate portfolio, held as leases and loans rather than owned property: $47,610.1 million, or 163.2% of tangible book against 98.6% of total assets. There is no goodwill and no intangible balance in the mix. Total assets were $48,271.2 million.

Historical relative multiple valuation. The book-value leg is the primary anchor and rests on 33 quarter-end P/TBV observations from 2018-03-31 to 2026-03-31, an 8.25-year window

.

One further observation, dated 2017-12-31, is excluded from that window: the common stock did not begin trading on the New York Stock Exchange until 1 February 2018, so no traded price stands behind it, and the 74.17× multiple carried there is arithmetically impossible against the equity then reported. Leaving it in would have dragged the top-quartile mean to 9.73× and produced a sell range topping $257 per share; the exclusion is disclosed here rather than buried and is the only edit made to the series. On the book-value leg, the stock sits at 0.968× against a window whose floor is 0.979× (March 2020) and whose ceiling is 1.862× (June 2022), with a typical band of 1.243× to 1.473× implying $32.94 to $39.02 a share. On trailing GAAP earnings the stock trades at 9.9 times, below the five-year floor of 10.8× (FY2025) and well under the 10.8×-to-13.4× typical band excluding the CECL-distorted FY2022 print of 25.8×, implying $27.86 to $34.57 a share. On funds from operations the multiple is 10.8× FY2025 adjusted funds from operations of $2.38 per share, with no multi-year window sourced for comparison. The implied capitalisation rate on annualised contractual rent is 7.76% at today’s price.

The two legs agree and need no averaging: on book value the stock sits below the floor of its own eight-year window, and on earnings below the lowest annual observation in five years. Today’s 0.968× is at the 0th percentile of the 33-observation distribution: cheaper than at any quarter-end since listing, including the March 2020 pandemic low of 0.979×. Where they diverge is magnitude: the book leg implies $32.94 to $39.02 on a return to the typical band, the earnings leg $27.86 to $34.57, the difference being dilution. The reconciled demonstrated range is $28 to $39. The 8.25-year P/TBV history behind this window is a market-data series that has not been independently rebuilt point by point against filings; the current-period P/TBV used throughout this report is derived directly from the balance-sheet bridge above.

Sell range. This method sets the sell range from the median and the top-quartile mean of the stock’s own P/TBV distribution, applied to current tangible book per share, giving $35.00 to $43.70 (1.321× to 1.649× tangible book) on the 33-observation, 8.25-year quarterly window. The distribution: minimum 0.979×, p25 1.243×, median 1.321×, p75 1.473×, p90 1.722×, maximum 1.862×. The flat 0.80× reference of $21.19 is fallback only, not the rule for this name. The range begins 36% above today’s price.

An exit at the low end, around $35, is the reasonable expectation, for a returns reason rather than a preference. The top-quartile multiples in this series were printed between mid-2021 and mid-2022, when VICI was adding rent at more than 70% a year through the MGP and Venetian acquisitions and funding it at a sub-4% marginal cost of debt. Today rent grows at roughly 3% from escalators plus incremental deals, and the marginal cost of debt is the 5.570% blended coupon struck in August 2026. The multiple a 3%-growth business earns is the middle of its own distribution, not the top. Return on common equity was 10.0% in FY2025 against an 8.7% five-year average, so nothing in the credibility of the return argues for a lower band: but the growth that earned the ceiling is absent.

Scenario table. The lens is net asset value on capitalised rent, cross-checked against the tangible-book apparatus. Every case uses one formula: NAV per share = [ annualised contractual rent ÷ capitalisation rate + loans and securities, net + cash − total borrowings at principal − non-controlling interests ] ÷ 1,101.075 million shares, where the constant deduction is $17,218.4m − $2,917.3m − $288.1m + $432.0m = $14,445.0 million. Land, development property and operating equipment of $238.1 million are excluded entirely, and the $862.5 million finance sub-lease liability is netted against its matching $835.4 million asset and dropped; both omissions are conservative.

• Severe downside: Caesars rent reset −20% to $3,063.0 million, cap rate widens to 8.75% → $18.67, −27.2%.

• Bear: Caesars rent reset −10% to $3,187.6 million, cap rate 8.25% → $21.97, −14.3%.

• Base: rent as reported, $3,312.2 million, cap rate 7.75% → $25.70, +0.2%.

• Bull: rent +2.5% to $3,395.0 million from escalators and new deals, cap rate 7.00% → $30.93, +20.6%.

• Anchor, tangible book: the derived filing bridge at 30 June 2026 → $26.49, +3.3%.

• Anchor, own-history p25 multiple: 1.243× tangible book on the 8.25-year window → $32.94, +28.4%.

Two cross-checks. The cap-rate anchors are not invented: VICI’s own call right on the Caesars Forum Convention Center is struck at a 13.0-times multiple, a 7.7% cap rate, and FY2025 commitments were made at an 8.9% weighted average initial yield, so 7.00% to 8.75% brackets both the trophy and the regional end of the portfolio. Separately, a book-value stress rather than a value estimate: were the credit allowance to double to $3,851.6 million: no precedent, given zero charge-offs since inception: tangible book would fall to $24.74 per share and today’s price would be 1.037× tangible book rather than 0.968×.

Stated explicitly. Current price $25.65 at the 3 September 2026 close, a market capitalisation of approximately $28.2 billion on 1,101,074,906 shares; the stock traded at $25.42 intraday on 4 September 2026, within $0.08 of a 52-week low of $25.34. The reporting currency is the US dollar throughout. The debt exhibit translates the CAD and sterling revolver balances into USD at 30 June 2026 rates of C$1 = US$0.7037 from the Bank of Canada and £1 = US$1.3273 from the Bank of England; the original balances remain in the exhibit notes. Intrinsic value range: $25.70 to $29.85 per share, the base-case net asset value at capitalisation rates of 7.75% down to 7.00%, with derived tangible book of $26.49 inside it. Margin of safety: 3.2% against tangible book per share and 7.7% against the $27.78 midpoint: thin on both, and the discount that exists is to the stock’s own history, not to its assets.

Buy-Below is $26.49, 1.00× tangible book: and it’s worth publishing why, rather than picking whichever number looks better. This method normally sets Buy-Below from the stock’s own trading history: the average of every quarter at or below the 25th percentile of its P/TBV range, which for VICI is 1.158× tangible book, or $30.68. That figure ordinarily sits safely below the sell range, and it does here too ($30.68 against a $35.00 floor). But VICI has never traded below tangible book in eight years until now, so its own lower quartile sits above 1.00×: and publishing $30.68 as the buy price would invite a purchase at 1.16× tangible book, a level at which this method’s own cheapness screen fails and the call would be Avoid, not Buy. The lower number governs instead: $26.49, the price at which the stock stops qualifying as cheap against its own book value at all. Buy More Below, where I’d move the position from Small to a larger size, is $21.19, 0.80× tangible book.

Cheapness type: asset-value cheap, with a statistical overlay. The assets behind the book value carry no accumulated depreciation and are already stated net of a $1,925.8 million credit reserve never drawn. Paying below book for a portfolio 100% leased on a 39.6-year weighted average term with parent guarantees from two S&P 500 operators is an asset-value proposition first; the overlay is that the portfolio has never been available at this multiple before. What it is not is a high-quality-compounder discount: the company cannot compound internally: nor a deep Schloss-style asset discount, since the margin against a real-world net asset value is single-digit. The risk of value-trap cheapness turns on one question: whether the Caesars rent funding 38% of the roll survives a change of control intact.

13. Risk Matrix

• Caesars rent reset on the Fertitta change of control: High severity, Medium probability. Caesars agreed on 28 May 2026 to a take-private by Fertitta Entertainment; it provides 38% of annualised rent ($1,246.2 million), and management gave no timeline for the lease discussions on the Q2 2026 call. A 10% reset costs about $124.6 million of rent, roughly $0.11 of AFFO per share; the bear case above is −14%. Mitigants: a parent guarantee, master-lease cross-default, and a 100% collection record since 2017. Watch any 8-K amending a Caesars lease and the Fertitta deal’s regulatory timetable.

• Tenant concentration: High severity, Low probability. Caesars and MGM are 38% and 32% of the $3,312.2 million rent roll, and the Las Vegas Strip is 49% of FY2025 lease revenue. A single master-lease default would remove a third of revenue; the severe-downside case above is −27%. Mitigants: parent guarantees from two S&P 500 operators, and sixteen tenants now versus twelve two years ago. Watch tenant rent-coverage disclosure and both companies’ credit ratings.

• Credit-loss allowance volatility: Medium severity, High probability. The allowance moved $271.1 million in a single quarter (Q2 2026) and stands at $1,925.8 million, 3.84% of amortised cost, the auditor’s sole critical audit matter. It is non-cash but drives reported earnings and book value directly; doubling the reserve would take tangible book per share to $24.74. Mitigants: zero charge-offs and zero recoveries since adoption, and inputs agreed to independent data by the auditor. Watch the quarterly roll-forward and guarantor rating changes.

• Rising cost of debt on refinancing: Medium severity, High probability. The August 2026 issue priced at a 5.570% blended coupon against 4.321% retired, roughly $21.9 million of extra annual interest; $1,500.0 million matures in February 2027 and $2,000.0 million in 2028. This compresses AFFO growth toward the low single digits. Mitigants: 98.4% fixed-rate debt, a 5.5-year weighted average maturity, and $2.2 billion of undrawn revolver capacity to 2029. Watch the February 2027 maturity and credit spreads at each new issue.

• Single-notch investment-grade rating: Medium severity, Low probability. Ratings are Baa3/BBB−/BBB−, all stable, with net leverage at 4.9× against a 5.0-to-5.5× target. A downgrade would widen spreads across $17.2 billion of borrowings and raise the capitalisation rate the market applies to the rent. Mitigants: leverage already below target, and unencumbered assets at 319% of unsecured debt against a 150% covenant. Watch agency actions after any Caesars lease amendment.

• Equity-funded growth stalling below book: Medium severity, High probability. Shares outstanding are up 75% since FY2021 and there was no at-the-market issuance in the first half of 2026, with the stock at 0.968× tangible book. External growth pauses, and AFFO growth falls back to the contractual escalator, removing the growth premium in the historical median multiple. Mitigants: escalators alone deliver about 3% a year without new capital, and the $2.9 billion credit book yields 9.1%. Watch at-the-market usage each quarter and new investment yields.

• Non-accrual and future funding exposure in the credit book: Low severity, Medium probability. An $82.8 million golf-resort loan has been on non-accrual since Q4 2025, and $623.5 million of future funding commitments existed at 31 December 2025 plus $55.2 million for Club Med. The loan is 0.17% of amortised cost and the committed funding is 3.6% of borrowings, immaterial unless it becomes a pattern. Mitigant: the loan is fully funded and senior secured. Watch for any further non-accrual designation and the recapitalisation outcome.

• MGM tax protection agreement: Low severity, Low probability. The agreement requires roughly $8.5 billion of non-recourse debt allocable to MGM to be maintained for fifteen years from April 2022, an uncapped and unquantified indemnity if breached, constraining deleveraging and asset sales. Mitigants: the required balance declines over the term and no breach is disclosed. Watch any disposal of a protected property.

• Long-dated escalators against replacement cost: Low severity, Medium probability. The weighted average lease term is 39.6 years including options, with escalators of 1% to 2% fixed or CPI-linked with caps. Persistent inflation above the caps would erode real rent and lower the terminal value in any net asset value estimate. Mitigant: 42% of FY2025 rent, and roughly 90% over the long term, is CPI-linked. Watch the realised escalator against CPI each year.

14. Red Flags, Yellow Flags, and Green Flags

🟢 Green Flags

• Tangible book value per share has risen every year: $19.25 (FY2021), $22.77, $24.22, $25.12, $26.01 (FY2025) and $26.49 at 30 June 2026: despite a 75% increase in share count, because equity was issued above book and invested at an 8.9% weighted average initial yield in FY2025.

• Zero charge-offs and zero recoveries against the credit-loss allowance in FY2023, FY2024, FY2025 and H1 2026, on a reserve grown from $1,368.8m to $1,925.8m.

• Unqualified audit opinions on the financial statements and on internal control over financial reporting for FY2025, from an auditor in place since 2016, with no material weakness and no restatement.

• The 2026 maturity wall of $1,750.0m was refinanced in full on 14 August 2026 through $900.0m of 5.400% notes due 2031 and $850.0m of 5.750% notes due 2036.

• Dividend raised in every year of the company’s public life, most recently to $0.46 a quarter, on a 73% AFFO payout.

• Zero governance red flags fired, with all seven directors elected annually by majority vote on 28 April 2026 and 94.9% say-on-pay support.

🟡 Yellow Flags

• GAAP net income has exceeded operating cash flow in every year reviewed: by $265.5m in FY2025: and $1,848.9m of non-cash lease accretion since FY2023 is the largest component of the $3,189.4m retained-earnings balance.

• No insider bought a share on the open market in the twelve months to July 2026, with the stock at a 52-week low; every Form 4 is a grant or a tax-withholding surrender.

• Four named executives account for roughly $31m of a $65.1m general and administrative line at a 28-employee company.

• The credit-loss allowance is a third-party model output whose inputs changed enough in one quarter to move reported earnings by $271.1m, and it is the auditor’s sole critical audit matter.

• VICI is simultaneously landlord, lender and prospective buyer to Cabot, which also manages all four of its golf courses.

• The proxy’s compensation tables and audit-fee schedule were not reached in full.

🔴 Red Flags

• Caesars, source of 38% of annualised rent, agreed on 28 May 2026 to be taken private by Fertitta Entertainment in a transaction assuming approximately $11.9bn of Caesars debt, and management declined on 30 July 2026 to give any timeline for the associated lease discussions: the one disclosed negotiation that can permanently reset a third of the rent roll.

• Caesars reported a Q2 2026 GAAP net loss of $62m and consolidated Adjusted EBITDA of $920m against $955m a year earlier, so the largest tenant is deteriorating operationally while its ownership changes.

• The first credit event in the company’s history occurred in Q4 2025: an $82.8m senior secured loan on a luxury golf-resort development placed on non-accrual, the borrower seeking recapitalisation.

• Investment-grade ratings sit exactly one notch above high yield at all three agencies, with no second notch of cushion, against $17.2bn of borrowings.

⚡ Must-Watch Catalysts

• Q3 2026 results, expected late October 2026, with the next credit-allowance roll-forward.

• Any 8-K amending the Caesars Las Vegas or Caesars Regional master leases, and the regulatory timetable for the Fertitta acquisition of Caesars.

• The $1,500.0m of senior notes maturing February 2027 and the terms on which they are refinanced.

• Dividend payment on 8 October 2026 to holders of record 17 September 2026, and the December 2026 declaration.

• Resumption or continued absence of at-the-market equity issuance in each quarterly equity note.

Recommendation

Verdict: BUY — SMALL

At $25.65, VICI trades at 0.968× tangible book value of $26.4930 per share (30 June 2026) and clears both of my absolute screens: price-to-tangible-book at or under 1.00×, and financial borrowings at 58.0% of common equity, well inside the 100% ceiling I use (62.2% including lease liabilities, 64.0% on total liabilities). The Piotroski F-Score is 5 of 7 computable signals, Eligible, with both failures explained by ordinary lease accounting rather than a real problem. Normalized P/E is 10.8× on FY2025 adjusted funds from operations of $2.38 per share. Cash return is 7.2%, all dividend, with no buyback programme.

• Intrinsic value range: $25.70 to $29.85 a share, net asset value on capitalised contractual rent at cap rates of 7.75% down to 7.00%, with derived tangible book of $26.49 inside that range.

• Margin of safety: 3.2% below tangible book per share, 7.7% below the $27.78 midpoint of the intrinsic range: real, but thin.

• Buy-Below: $26.49, 1.00× tangible book (see the note above on why the own-history figure of $30.68 isn’t the published number here).

• Buy More Below: $21.19, 0.80× tangible book, where I’d move the position from Small to Standard.

• Sell range: $35.00 to $43.70, 1.321× to 1.649× tangible book, from 33 quarterly observations over 8.25 years. The flat 0.80× fallback of $21.19 is not the rule for this name. I’d expect an exit nearer the low end, around $35, because the top-quartile multiples in this series were earned on rent growing more than 70% a year during the MGP and Venetian acquisitions, a pace nothing in the current portfolio replicates.

• Position size: Small.

• Expected holding period: Three to five years, long enough for the Caesars change of control to resolve and for the multiple to revert toward its own median.

• Downside risk: −14.3% ($21.97) on a 10% Caesars rent reset; −27.2% ($18.67) in the severe case.

• Balance-sheet risk: Manageable. Borrowings are 35.1% of total assets, 98.4% fixed-rate, with a 5.5-year weighted average maturity. Interest coverage of 4.1× would read as Elevated on an industrial scale, but a triple-net REIT collecting contractual rent from investment-grade-guaranteed tenants on 39.6-year leases can carry more debt at the same underlying risk, which is why three rating agencies call this structure investment grade at 4.9× net leverage.

• Creditworthiness: Baa3 (Moody’s), BBB− (S&P), BBB− (Fitch), all stable, investment grade at all three but with exactly one notch of cushion. Net leverage is 4.9× against a 5.0-to-5.5× target, and unencumbered assets cover unsecured debt 319% against a 150% covenant.

• Governance risk: Low, none of the five red flags I check fired.

• Accounting quality: Adequate. Unqualified audit opinions on both the statements and internal controls, no material weakness, one genuine estimation-risk item (the credit-loss allowance).

• Refinancing risk: Low near term. The 2026 maturity wall was retired in full with the 14 August 2026 bond issue; the next maturity is $1,500.0 million in February 2027 against $2.2 billion of undrawn revolver capacity to 2029.

• Key catalysts: resolution of the Caesars lease discussions and the Fertitta change of control; Q3 2026 results in late October; the February 2027 maturity; resumption of accretive equity issuance if the shares recover above tangible book.

• What would break the thesis: a permanent cut to Caesars rent on the change of control; a downgrade below investment grade; a genuine credit loss in the master-lease portfolio, which has never happened; or sustained equity issuance below tangible book.

What to watch

• Any 8-K amending the Caesars Las Vegas or Caesars Regional master leases, and the closing timetable for the Fertitta take-private of Caesars agreed 28 May 2026.

• Caesars’ own quarterly results and rent coverage; consolidated Adjusted EBITDA fell to $920m in Q2 2026 from $955m.

• The quarterly credit-loss allowance roll-forward: the first charge-off in the company’s history would be a thesis event, not a modelling event.

• The $82.8m golf-resort loan on non-accrual since Q4 2025 and whether its recapitalisation closes.

• The $1,500.0m of senior notes maturing February 2027 and the coupon at which they are refinanced.

• Ratings actions at Moody’s, S&P and Fitch: one notch separates this balance sheet from high yield.

• At-the-market equity issuance in each quarterly stockholders-equity note: resumption below tangible book would be dilutive rather than accretive.

• Q3 2026 results, expected late October 2026, and the December 2026 dividend declaration.

• $35.00: the low end of the sell range, where the sell question goes live.

• $21.19: the Buy More Below price at which I’d move the sizing tier from Small to Standard.

Sources and diligence gaps. This analysis draws on VICI’s FY2025 Form 10-K, the Q2 2026 Form 10-Q, the 2026 definitive proxy statement, the Q4/FY2025 and Q2 2026 earnings releases, the August 2026 bond-offering and closing 8-Ks, the 2026 annual-meeting voting results, every Schedule 13G and Form 4 filed on VICI in 2025 and 2026, SEC XBRL company-concept data back to FY2019, Caesars’ own Q2 2026 earnings release, and the Q2 2026 earnings-call transcript. The 10-K and 10-Q were read in full; both are available from the SEC’s EDGAR system. The debt exhibit’s currency translations use the Bank of Canada’s historical currency converter and the Bank of England exchange-rate database.

Real gaps remain. The 2026 proxy’s Summary Compensation Table and audit-fee schedule were not reached in full, though the pay figures used above come from the proxy’s separately-disclosed pay-versus-performance table, which reports the same totals. The related-party section was reached only through its policy statement and negative disclosure, not a line-by-line schedule. The FY2023 and FY2024 annual reports were read for their financial statements but not their legal-proceedings items, so the litigation carry-forward in this report runs only from FY2025 forward. And the 8.25-year quarterly P/TBV history behind the sell range is a market-data series, not independently rebuilt point by point against filings; the current-period P/TBV used everywhere else in this report is derived directly from the balance-sheet bridge above, and it agrees with that series’ own latest observation within the tolerance this method requires. None of these gaps touches the balance sheet, the tangible-book bridge, the F-Score or the debt schedule, every figure in which came from primary filings and was cross-footed.

Research only, not investment advice. Position disclosure: Long, ~0.26% of portfolio.

1. Scorecard

2. Argument

The short version

Indofood Agri Resources is a Singapore-listed holding company wrapped around a large Indonesian palm oil and edible oils group. The operating recovery is real: FY2025 revenue reached Rp21.06 trillion, operating cash flow was Rp3.38 trillion, and 1H2026 attributable profit rose 32% year on year. Cash of Rp9.05 trillion almost covers borrowings and the FY2025 dividend rose for a third consecutive year.

At S$0.350 on 4 September 2026, the shares trade at 0.523x derived tangible book and about 4.9x trailing earnings. That is below the stock's own ten-year quarterly median, but it is not an extreme bargain. The verdict is BUY, SMALL. The small size is a governance decision, not a valuation decision: the controlling family is deemed interested in 86.04%, the public float is 13.92%, the group tried to take the company private in 2019, and capital has been routed into policy-driven uses such as plasma provisions and a Rp2.49 trillion forestry charge paid into escrow.

Why it qualifies

• The shares trade at 0.523x tangible book, around the 43rd percentile of the stock's real ten-year quarterly P/TBV history, below its 0.54x median.

• The balance sheet is close to net cash: Rp9.05 trillion of cash against Rp9.08 trillion of borrowings at 30 June 2026, with cash covering current borrowings 1.09x.

• Owner profit improved in 1H2026, while FY2025 operating cash flow of Rp3.38 trillion exceeded capital additions of Rp1.44 trillion.

• The Piotroski score is 6/9, the dividend reached S$0.012 for FY2025, and the current shareholder yield is 3.43% with no buybacks.

Why it is not a clean bargain

• An 86.04% deemed interest leaves minority holders with little practical ability to block ordinary resolutions, even though the board now has three independent directors.

• The Rp2.49 trillion forestry charge is paid into escrow and carried as an asset while under appeal, equal to 19.3% of tangible book with no provision.

• The government-mandated plasma programme has produced a Rp1.08 trillion write-off, continuing charges, and up to Rp523.1 billion of guarantees.

• Related parties bought Rp8.26 trillion of FY2025 revenue, while 94% of 1H2026 CPO volume moved internally, limiting the arm's-length quality of the reported segment economics.

• The operating margin is already rolling over: gross margin fell from 29.8% in FY2024 to 25.6% in FY2025 and 22.4% in 1H2026.

What would change the view

• A full or materially larger forestry loss, or a provision that shows the escrowed asset is not recoverable.

• A second consecutive period of leverage rising and liquidity deteriorating, or a failed renewal of the short-dated bank facilities.

• A call on the unreserved plasma guarantees or another large credit-loss charge.

• A CPO price reversal combined with further own-estate yield decline, or evidence that minority leakage and related-party terms are worsening.

Sources and gaps

This snapshot uses the 2025 annual report, the condensed interim statements for the six months ended 30 June 2026, the 1H2026 results materials, the 2026 AGM materials, the 31 July 2026 board-change announcements, SGX filings, and public price history. FY2021 and FY2022 audited statements were not retrievable for the supplementary annual cross-check; bank covenant headroom and inter-segment transfer prices are not disclosed; and the exchange announcement index after 21 January 2026 could not be checked in full for later director or substantial-shareholder dealings. The latest board changes were checked separately.

3. Backup and sources

This last layer is the full working file: every source, calculation, and gap, kept so the argument above can be checked. The scorecard and the argument are the synthesis. Open this only if you want to check the work.

1. Executive Summary

Indofood Agri Resources ("IndoAgri") is a Singapore-incorporated holding company for a vertically integrated Indonesian agribusiness: 278,099 hectares of planted nucleus estates and 27 palm oil mills feeding five refineries that turn crude palm oil into branded cooking oil, margarine and shortening (Bimoli, Happy, Palmia, Amanda, Simas), with smaller rubber, sugar, cocoa, tea and timber operations and a Brazilian sugar-and-ethanol joint venture. Roughly 97% of revenue is earned inside Indonesia, and the money is made twice, once on the plantation spread between fruit cost and CPO price, again on the refining and branding margin downstream. The verdict is BUY — SMALL. At S$0.350 (4 September 2026) the shares trade at 0.523× the tangible book value derived in the valuation discussion and about 4.8× trailing earnings. Measured against the stock's own ten-year quarterly P/TBV history, 0.523× sits at the 43rd percentile, cheaper than average, comfortably inside the range between the own-history Buy-Below (0.42×) and the Sell Range floor (0.54×), not at an extreme in either direction. The Piotroski screen scores a clean 6/9: two of the three failed signals are real and worth naming, leverage rose and liquidity deteriorated in the same fiscal year (debt to average assets 22.3% against 20.4%; current ratio 1.27 against 1.50), but that combination is a flag to weigh, not an automatic disqualifier, and it is stated plainly here rather than folded into a bigger score. What actually caps the position size is not valuation or the F-Score: it is governance. The controlling family holds 86.04% of the shares, tried once already to take the company private, and has a demonstrated pattern of routing capital into policy-mandated, low-return uses. That risk is real and structural, and it is why this is sized small rather than passed on entirely. The business and the balance sheet do not need a small size. The ownership structure does.

The business itself is doing better than it has in years. FY2025 revenue rose 32% to Rp21.06 trillion, profit from operations 14% to Rp3.70 trillion, and operating cash flow 52% to Rp3.38 trillion; 1H2026 net profit rose 32% to Rp1.01 trillion. The balance sheet is close to debt-free on a net basis, Rp9.08 trillion of borrowings against Rp9.05 trillion of cash at 30 June 2026, and the FY2025 dividend was raised for a third consecutive year, to S$0.012, a 3.5% yield. Tangible book is genuinely tangible: bearer plants, mills and refineries, with right-of-use land rights at only 18.1% of tangible book, so this name can fairly be described as cheap against tangible book value rather than against a capitalised lease.

What undermines it is the quality of ownership and of disclosure. Anthoni Salim is deemed interested in 86.04% of the shares; the public float is 13.92%, barely above the exchange's minimum; and the controlling group already attempted to take the company private at S$0.28 in 2019, failing only on the 90% acceptance threshold. Roughly 39% of FY2025 revenue was sold to related parties. In December 2025 the Ministry of Forestry imposed Rp2.34 trillion of forestry administrative charges, paid into a government escrow within 30 days and then carried as a non-current asset pending appeal, with a further Rp0.15 trillion in May 2026, Rp2.49 trillion, 19.3% of tangible book, capitalised rather than provided against and not identified as a key audit matter. Plasma receivables under a government-mandated smallholder programme carry a 61% allowance after Rp1.08 trillion was written off outright in FY2024. The operating recovery is real; the terms on which a minority holder participates in it are not.

2. Business and Market Overview

The Group reports two segments. Plantations (FY2025 total sales before elimination Rp14.46 trillion) develops and harvests 237,437 hectares of nucleus oil palm plus 16,203 hectares of rubber, 13,719 hectares of sugar cane and 13,616 hectares of other crops in Indonesia, processing fresh fruit bunches at 27 mills, rubber at five facilities, and cane at two mills and refineries (8,000 and 4,000 tonnes of cane per day). Edible Oils and Fats ("EOF", FY2025 sales Rp15.00 trillion) runs five refineries with 1.9 million tonnes of annual CPO processing capacity, selling consumer cooking oil under Bimoli, Bimoli Spesial and Happy and margarine under Amanda, Palmia and Royal Palmia, with industrial oils, margarine and shortening under Delima, Malinda, Palmia and Simas. Around 93% of EOF output is sold domestically; distribution runs through the parent group's own distribution network.

The economics are those of a price-taker with a branded tail; management says so directly: "As a price taker, we will remain focused on improving operational performance, strengthening cost discipline" (1H2026 announcement, the other-information note). Domestic CPO reference prices rose 8% in FY2025 to Rp14,234/kg and 5% again in 1H2026 to Rp14,931/kg, and Rotterdam CIF rose 15% to USD1,283/tonne then 13% to USD1,446/tonne, both supported by Indonesia's escalating B50 biodiesel mandate. Sugar moved the other way, raw sugar fell from roughly 24 US cents per pound in 2024 to about 14 cents by 1H2026, which is why the Brazilian sugar joint ventures cost the Group Rp178 billion in FY2025 and Rp223 billion in 1H2026 alone.

Customer concentration is extreme, and it is a related-party fact rather than an arm's-length one: FY2025 sales of goods to companies under common control were Rp6.17 trillion and to other Salim Group members Rp2.09 trillion, Rp8.26 trillion, 39% of consolidated revenue (Note 33). Internally, 94% of 1H2026 CPO volume moved from Plantations to EOF rather than to third parties, up from 84%, so the reported Plantations external revenue line is a residual rather than a market test. Capital intensity is high (FY2025 additions of Rp1.44 trillion against Rp3.38 trillion of operating cash flow) and regulatory exposure is heavy: export levies, domestic market obligations, a retail price ceiling on cooking oil, the plasma mandate and, new in this cycle, forest-area enforcement. The business is understandable and the products are staples; it is not predictable, it is decidedly cyclical, and it is not structurally advantaged.

3. Moat, Competitive Position, and Industry Cycle

Against the five classic moat sources the picture is thin. Intangibles and brand: Bimoli is a genuinely well-known Indonesian cooking-oil brand, but branded cooking oil sits under a government retail price ceiling and the EOF segment earned a 6% EBITDA margin in both FY2025 and 1H2026, a brand that cannot price is not a moat. Switching costs: none in a commodity oil sold by the litre. Network effects: none. Cost advantage: partial and eroding, the oil extraction rate improved to 21.5% in 1H2026 from 20.4%, but nucleus FFB production fell 4% and the Group is buying a rising share of its feedstock externally (493,000 tonnes in 1H2026, 30% of throughput), which converts an integrated cost advantage into a merchant margin. Efficient scale: the Indonesian palm sector has many large integrated players and no scale barrier at 237,000 hectares. The moat is Weak. The proof is in the returns: FY2025 return on equity of 8.1% (profit attributable to owners over shareholders' equity), up from 7.8% and 4.5%, is below any reasonable Indonesian cost of equity. This company does not compound capital at attractive rates; it converts commodity prices into cash and passes a large share of the profit to minority holders of its own listed subsidiaries.

On the cycle the evidence points to Above mid-cycle on price, with margins already rolling over. CPO prices are near multi-year highs and policy-supported; palm kernel prices rose 52% in FY2025 and 11% again in 1H2026 on a coconut-oil squeeze. But gross margin has peaked and turned: 29.8% in FY2024, 25.6% in FY2025, 22.4% in 1H2026, as heavier fertiliser application and external FFB purchases pushed cost of sales up 4% against 2% revenue growth. The Group is capturing a high-price environment with a deteriorating unit cost structure and a falling own-estate yield, the classic late-cycle plantation profile. Rubber is worse: management fully impaired Lonsum's rubber bearer plants and related fixed assets by Rp296.2 billion in FY2024 on yield and outlook grounds, which is a structural rather than cyclical judgement about that crop.

4. Management, Governance, and Capital Allocation

The board still has nine directors, but its composition changed on 31 July 2026. Philip Yeo Liat Kok remains Lead Independent Chairman; Mark Julian Wakeford remains CEO and Executive Director; Moleonoto Tjang and Suaimi Suriady remain Executive Directors; Tjhie Tje Fie and Axton Salim remain Non-Executive Directors; David Sungkoro and Andreas Tan remain Independent Directors; and Amelia Setiawan joined as a third Independent Director. Goh Kian Chee ceased to be a Non-Executive Director and Audit and Risk Management Committee member on the same date. Three of nine are independent, exactly the one-third minimum the listing rules require. The three non-independent non-executives are paid not by IndoAgri but by the parent, PT Indofood Sukses Makmur ("PT ISM"), or other PT ISM companies; Axton Salim is a member of the controlling family. The board refresh improves the audit committee's formal independence: it is now chaired by David Sungkoro with Andreas Tan and Amelia Setiawan, all three independent. It does not change the controlling family's ability to dominate ordinary resolutions.

Insider ownership and transactions. The register of directors' shareholdings kept under the Singapore Companies Act register, reproduced in the FY2025 Directors' Statement, shows exactly one director with any interest: Mark Julian Wakeford, 300,000 shares direct and 200,000 deemed, unchanged through the financial year and again between year-end and 21 January 2026. Every other director, including the Chairman and all three independents, holds no shares. There were no purchases and no sales by any director in the twelve months to 21 January 2026, and no options or share awards exist: "No option to take up unissued shares of the Company or its subsidiaries were granted during the year… There were no unissued shares… under option as at the end of the year." The regimes searched were that Companies Act register and the exchange's substantial-shareholder and director-dealing notification stream; the announcement index after 21 January 2026 could not be fully opened, so any insider dealing between then and 4 September 2026 remains a diligence gap. The 31 July board-change announcements were checked separately. Insiders do not eat their own cooking here: the controlling shareholder owns the company outright and the professional board owns essentially nothing.

Compensation. Wakeford received S$1,475,599 in FY2025 (84% fixed, 16% variable, no other benefits); directors' fees put to shareholders were S$375,000 against S$346,666 for FY2024; the top five key management personnel below the CEO were paid S$2,484,776 in aggregate, three of them by the subsidiary PT Salim Ivomas Pratama ("PT SIMP") rather than by IndoAgri. Group-wide key-management compensation under FY2025 Note 33 was Rp204.99 billion (FY2024: Rp191.09 billion), up 7.3% and equal to 16.1% of profit attributable to owners, a much wider definition covering the whole Indonesian group, but a number worth tracking against a Rp1.27 trillion attributable profit. There is no share-based compensation at all, so the share-based-compensation-to-net-income test computes 0.0% in both years: neither the ratio threshold nor the doubling condition is met, and warning 5 does not fire on that test.

Related-party transactions. FY2025 Note 33 and interim Note 18 were both read in full, and the related-party note covers more than the obvious trade flows. Beyond the Rp8.26 trillion of FY2025 sales already noted, purchases spanned packaging materials Rp151.0 billion, services, transport equipment and spare parts Rp171.6 billion, rental Rp48.0 billion, insurance Rp22.7 billion and royalty fees to a shareholder of Rp10.8 billion. Financial relationships are disclosed and material: Rp56.4 billion of financial income and Rp15.1 billion of financial expenses with other related parties, and Rp298.3 billion of loans to related parties repaid during the year. The Rule 920 disclosure for 1H2026 puts Salim Group shareholder loans at Rp621 billion at their peak (Rp299 billion at period end) and Salim Group corporate guarantees at Rp1,182 billion. No nominee, back-to-back or related-party bank-deposit arrangement is disclosed anywhere in the record.

Capital allocation. Dividends have been raised three years running, S$0.008 for FY2023, S$0.010 for FY2024 and S$0.012 for FY2025, the last paid on 28 May 2026, on a payout of roughly 17%. Cash to IndoAgri's own shareholders was Rp132.8 billion, Rp177.3 billion and Rp230.4 billion across FY2024, FY2025 and 1H2026, against Rp148.5 billion, Rp267.4 billion and Rp335.5 billion paid by subsidiaries to their non-controlling holders: in every one of those years more cash left the group to minorities of the subsidiaries than reached IndoAgri's own shareholders. There have been no buybacks and no issuance; the share count has been static throughout. The capital that has been destroyed went elsewhere: Rp1,084.1 billion of plasma receivables written off in FY2024 as unrecoverable, Rp296.2 billion of rubber impairment the same year, Rp160.9 billion and Rp401.4 billion of property, plant and equipment written off in FY2025 and FY2024, and a Brazilian sugar venture carried down from Rp955.8 billion to Rp884.3 billion in six months with Rp83.7 billion of Bússola losses and Rp442.6 billion of FP Natural Resources losses no longer recognised at all.

Governance warnings. The control concern is clear: the 2026 AGM included the required director re-elections, but with 86.04% deemed interest and a 13.92% float, no minority vote can carry or block an ordinary resolution. The controlling group has moved from 74.34% at the 2019 exit offer to 85.90% today and has already tried once to take the company private. That is control with a ballot but without a practical remedy.

Capital has also been routed into policy-driven or low-return uses. The plasma scheme produced a Rp1,084.1bn write-off in FY2024, further charges in FY2025 and 1H2026, and up to Rp523.1bn of guarantees. The Rp2.49tn of forestry charges paid into escrow and capitalised while under appeal belongs in the same risk bucket.

The dividend warning is not active: distributions resumed after the FY2019-FY2020 losses and have risen three years running, although a 17% payout from Rp9.05tn of cash is thin. The pre-AGM question-and-answer document was not located, so that point remains a gap. Pay was not outsized on the available evidence, and the Attorney General's Office later said it found no evidence linking PT SIMP in the export inquiry. Overall governance risk is High.

5. Corporate Ownership, Subsidiaries, and Joint Ventures

Substantial shareholders at 13 March 2026, from the Company's own Statistics of Shareholdings (percentages on 1,395,904,530 shares excluding treasury):

Ownership at 13 March 2026.

• Indofood Singapore Holdings Pte. Ltd. held 71.51% directly.

• PT Indofood Sukses Makmur Tbk held 85.90% in total, including deemed interest.

• First Pacific Company Limited and its affiliates were deemed interested in 85.90%; First Pacific Investments Limited was at 85.98%; Anthoni Salim was deemed interested in 86.04%.

• The company's stated public float was 13.92%. Treasury shares were 51,878,300, or 3.58% of issued shares, and carry no votes or dividends.

The structure matters to valuation: the Singapore listing sits above PT SIMP, Lonsum, the Brazilian interests and a further layer of equity-accounted associates.

One vote per share, one class. The top twenty registered holders account for 94.13%, and the second-largest name on the register is a broker nominee account at 11.81%, so the genuinely dispersed float is thinner than 13.92% makes it look.

The operating structure is a three-tier listed cascade:

• Indofood Agri Resources Ltd.: SGX-ST (5JS), the security analysed here and the Singapore holding company.

• PT Salim Ivomas Pratama Tbk (PT SIMP): 73.46% held directly, with essentially the entire operating group below it.

• PT PP London Sumatra Indonesia Tbk (Lonsum): held through PT SIMP, the integrated plantation business behind the Rp2,909.8 billion Lonsum goodwill balance.

• IFAR Brazil Pte. Ltd. / IFAR Trading Pte. Ltd.: wholly owned Singapore holding and trading companies for the Brazilian sugar interests.

• CMAA and Bússola: unlisted, equity-accounted Brazilian sugar and ethanol joint ventures.

• Five associates, including FP Natural Resources and PT Indoagri Daitocacao: unlisted, equity-accounted investments.

PT SIMP's FY2025 revenue of Rp21,056,904 million is identical to the Group's, which tells you what IndoAgri is: a Singapore listing wrapper over its Indonesian subsidiary, plus two Singapore holding companies and Rp165.0 billion of shareholder credit to a subsidiary maturing in 2028 and 2029. Non-controlling interests are correspondingly enormous, Rp13,081,696 million at 30 June 2026, 45.0% of total equity, against Rp15,965,955 million attributable to IndoAgri's own owners. PT SIMP alone carries 26.54% non-controlling interest, Rp6,153,009 million accumulated, taking Rp699.8 billion of FY2025 profit and Rp185.6 billion of dividends, and in certain subsidiaries the Group holds under 50% of the equity while controlling the votes, so consolidated assets overstate economic ownership further still. Cumulative losses the Group has stopped equity-accounting because its interest is exhausted stand at Rp442.6 billion for FP Natural Resources and Rp83.7 billion for Bússola, with no obligations incurred on behalf of either. The structural risks are the familiar ones for this shape: minority leakage at two levels, upstream PT SIMP guarantees securing subsidiary bank debt, and a holding company whose claim on the operating assets sits behind two layers of operating-company creditors and a 26.54% minority.

6. Historical Financial Quality and Normalized Owner Earnings

The comparison uses FY2025 and FY2024 from the audited statements, with FY2023 from the same report's three-year performance table and opening equity balances.

Fundamentals (Rp millions; per-share figures in Rp).

Net income and diluted EPS above are the amounts attributable to owners, consistent with the equity used in tangible book value; consolidated profit after tax was Rp936,000m, Rp2,110,017m and Rp2,513,854m across the same three years, so roughly half of every rupiah of group profit belongs to somebody else. The half-year confirms the direction: 1H2026 revenue Rp9,624,696m (+2.5%), gross profit Rp2,152,393m (−3.7%), net profit Rp1,012,124m (+32.4%), attributable profit Rp444,527m (+31.6%), operating cash flow Rp1,470,015m against Rp1,893,925m.

Normalized owner earnings. Real cash-flow figures exist, so no proxy is needed. FY2025 operating cash flow of Rp3,378,017m less capital additions of Rp1,437,137m (property, plant and equipment Rp1,130,791m plus biological assets Rp306,346m) gives free cash flow of Rp1,940,880m, against Rp965,771m in FY2024. Depreciation and amortisation of Rp1,544,452m exceeded total capital additions, so the Group is reinvesting below its own depreciation charge, an owner-earnings tailwind today and a yield question later, given immature oil palm area down from 23,806 hectares in 2023 to 15,978 in 2025. On the classic construction, owner earnings are Rp2,513,854m of consolidated net income plus Rp1,544,452m of D&A less Rp1,437,137m of maintenance-equivalent additions, or Rp2,621,169m, about Rp949 per share on IndoAgri's 50.5% share. The attributable split of free cash flow is not disclosed, so this comparison allocates consolidated cash flow using the profit split and is a sense-check only. That is within 5% of reported diluted EPS of Rp910, so reported earnings are a fair proxy for owner earnings here and there is no large accrual gap to unwind.

Symmetric normalization. The Group publishes a "core profit" measure stripping foreign exchange, biological-asset fair-value movements and the plasma credit-loss charge, and it cuts both ways. In FY2025 core profit of Rp2,595bn exceeded reported profit after tax of Rp2,514bn, because a Rp135bn biological-asset fair-value loss sat inside the reported figure; in 1H2026 core profit of Rp928bn was below reported profit of Rp1,012bn, because a Rp135bn foreign-exchange gain and a Rp78bn fair-value gain flattered it. Adding back the FY2025 charge without removing the 1H2026 gain would not be a normalization. On the company's own core basis 1H2026 was a 1.9% decline, not a 32% improvement, the cleanest single statement of what the first half actually did.

Normalized EPS. The trajectory is growing (Rp440 → Rp801 → Rp910, trailing twelve months Rp986 after adding 1H2026's Rp318 and removing 1H2025's Rp242), so the normalized basis is a three-year average of reported diluted EPS, Rp717. Only three years of primary income statements were opened, and the shortened window is disclosed rather than padded. On trailing earnings the multiple would be lower still; both land in the same valuation column, so the conclusion is insensitive to the choice.

The plasma reserve roll-forward. The Group carries a large allowance against government-mandated smallholder ("plasma") receivables and separately guarantees plasma bank loans. Both are reconstructed from the notes rather than inferred from balance-sheet movement:

FY2024 is the year that matters: the reserve was drawn down by Rp1,084,149m against Rp264,453m of replenishment, a provision-to-payout ratio of 0.24×, with "no reasonable expectation of recovering the cash flows". Since then no write-offs have occurred and the reserve has been rebuilt, the 1H2026 charge of Rp122,575m already exceeding the whole of FY2025's Rp103,041m, up 86.9% year on year. Gross plasma receivables were Rp1,328.4 billion at 30 June 2026 against Rp813.5 billion of allowance: a 61% provision rate on an asset the Group is legally obliged to keep creating. The asymmetry check is the uncomfortable part. The on-balance-sheet allowance rose from Rp655.2 billion to Rp813.5 billion over eighteen months, while off-balance-sheet guarantee exposure to the same borrower population rose 21% from Rp430.7 billion to Rp523.1 billion and carries no recognised reserve at all, because Note 32(d) states "there are no probable claims against the Group that may cause material impact". The guarantee is contingent and the accounting defensible, but if plasma credit is deteriorating enough to justify an 86.9% increase in the on-book charge, the Rp523.1 billion guarantee line is the next thing to move, and nothing is set aside for it.

Earnings quality and conversion. Cash conversion is good, FY2025 free cash flow of Rp1,940,880m against Rp2,513,854m of consolidated profit, with operating cash flow above net income in both years, and capex intensity is moderate at 6.8% of revenue. The trap is elsewhere: the FY2025 cash flow statement shows a Rp2,177,775m increase in other non-current receivables, essentially the forestry escrow payment, which alone consumed 64% of operating cash flow before working capital.

6.5 Piotroski F-Score, 6/9

Periods compared: FY2025 vs FY2024.

Piotroski results (FY2025 versus FY2024). Six of nine signals passed. The passes were positive net income, positive operating cash flow, improved ROA, operating cash flow above net income, no dilutive issuance, and improved asset turnover. The failures were leverage, liquidity, and gross margin. All nine signals were computable.

The combined deterioration in leverage and liquidity is the serious part: debt to average assets rose from 20.4% to 22.3%, while the current ratio fell from 1.50 to 1.27. Gross margin fell from 29.8% to 25.6% as fertiliser and bought-in fruit costs rose against a rising selling price. The half-year partly repaired the balance-sheet signal, with the current ratio back to 1.38 and borrowings down to Rp9,075,310m at 30 June 2026, but a second annual reading in the same direction would matter more than this first one.

All nine signals were computable, so this is a clean 6/9, and none of the three failures is cosmetic. Gross margin fell 420 basis points on rising fertiliser and bought-in fruit costs against a rising selling price, the wrong direction at this point in a commodity cycle. Total debt rose Rp1,429,508m to Rp9,200,754m while assets rose less, taking debt to average assets from 20.4% to 22.3%, and the current ratio fell from 1.50 to 1.27 as Rp2,334,820m of long-term loans reclassified into current liabilities and Rp2,337,480m of cash left current assets for the non-current forestry escrow. Leverage rising and liquidity falling in the same year is the floor-rule flag, named here because it is real and worth watching, not because it disqualifies the name on its own: the half-year already repaired part of it, with the current ratio back to 1.38 at 30 June 2026 and borrowings down to Rp9,075,310m, and neither failure traces to a depressed prior-year comparison that would call for a cyclical override in the other direction. The next annual reading is due with the FY2026 accounts, and a second consecutive year of the same combination would be a materially stronger signal than one year taken alone.

7. Balance Sheet, Debt, Covenants, and Refinancing Risk

Every borrowing is a Rupiah-denominated Indonesian bank facility drawn at subsidiary level: no bonds, no notes, no sukuk, no offshore debt. Instrument-level detail is disclosed annually, so the FY2025 position and the 30 June 2026 totals are shown below.

Lease liabilities are disaggregated from financial borrowings throughout, and the debt test is run on the latter: Rp9,075,310m of bank debt against Rp45,455m of leases. Financial borrowings are 31.2% of total equity and 56.8% of equity attributable to owners; total liabilities of Rp14,891,341m are 51.3% of total equity, the Schloss debt filter passes comfortably on either denominator. Debt to total assets is 20.7%, and net debt of Rp68,366m is 0.2% of total equity, which is why management describes the ratio as "less than 0.01 times".

Coverage: FY2025 adjusted EBITDA of Rp5,299bn against finance expenses of Rp548,410m gives 9.7×, and operating profit alone covers interest 6.7×; the interest line includes Rp5,241m of lease interest, so the ratio is blended and slightly conservative. Balance-sheet risk is Conservative: debt to assets of 20.7% is well inside the under-30% band, net debt is effectively nil, and no covenant concern is disclosed, the EBITDA coverage of 9.7× sits marginally below the 10× the Conservative band nominally wants, noted rather than hidden. Covenants are negative covenants at subsidiary level (restrictions on new borrowing, guarantees and asset pledges, mergers, disposals and new investment above thresholds) plus unspecified financial ratios at "certain subsidiaries"; the specific ratios and headroom are not disclosed, which is a real gap. The Group states compliance with all covenants at 31 December 2025, 31 December 2024 and 30 June 2026, and that there was "no loan default or breach of a loan agreement that has not been remedied".

The refinancing profile is the one genuinely uncomfortable feature. Rp8,343,576m, 92% of all borrowings, falls due within twelve months against a rolling stack of working-capital lines from Mandiri, BCA, DBS, SMBC, BNI and Permata, and the Group asserts "unconditional rights to rollover and/or refinance the short-term loans as and when they fall due". Cash of Rp9,052,399m covers the entire current balance 1.09 times, so even a complete failure to roll would be survivable, and 1H2026 net repayments of Rp126,054m show the lines behaving normally. Refinancing risk is low in substance but high in form: a bank-by-bank annual renegotiation at floating Indonesian rates with no disclosed covenant headroom.

8. Real Estate, Leases, and Hidden Assets

Right-of-use assets were Rp2,328,292m at 30 June 2026, 18.1% of tangible book value, 5.3% of total assets, against lease liabilities of just Rp45,455m, and that gap is the point. Almost the entire balance is Indonesian land use rights (HGB, HGU, HP and HPL, four to forty years, renewable, carrying the Group's bearer plants), paid for substantially up front rather than financed. Of the Rp2,366,815m at 31 December 2025, Rp2,351,817m was land use rights, Rp5,539m buildings and Rp9,459m office equipment; FY2025 additions of Rp425,019m were almost entirely a refinery land-lease renewal. Right-of-use is a capitalised right to occupy, not a liquidation floor, and it is disclosed here rather than stripped out of the tangible book value used everywhere in this report, removing the asset without its matching liability would be the wrong operation, and in this case the matching liability is almost nil, so the ROU line behaves far more like prepaid land than like a capitalised lease. Even so, since it is under the 25% threshold, describing this company as cheap against tangible book value is legitimate without further qualification.

The genuine downside support is elsewhere and is stronger. Cash of Rp9,052,399m is 70.2% of tangible book value on its own. Owned property, plant and equipment of Rp15,761,690m, 122.3% of tangible book, is 61% bearer plants (Rp9,695,230m net at 31 December 2025), with buildings and improvements Rp3,650,582m, plant and machinery Rp1,994,286m and heavy equipment, transport and vessels Rp721,769m. Inventory adds Rp3,797,764m.

Hidden asset value is plausible but unquantified. The 280,975 hectares of planted nucleus estates sit behind land use rights carried at Rp2,351,817m, roughly Rp8.4 million per planted hectare, a historical cost from acquisitions and renewals stretching back decades, not a mark to market. No sourced comparable transaction was obtained, so no market value is asserted; what can be said is that the carrying value is a floor rather than an estimate and that bearer plants are held at depreciated cost while the estates behind them produce at a 21.5% extraction rate. The offsetting risk runs the other way: the forestry enforcement programme in the regulatory discussion is a direct challenge to the legal status of some of that land, which is why the carrying value has become a live question rather than a dormant one.

9. Capital Markets Access, Dilution, and Financing Flexibility

The share count has not moved. Issued shares have been 1,447,782,830 at every balance-sheet date examined, with 51,878,300 in treasury (3.58%) since before FY2024, leaving 1,395,904,530 outstanding at each of 31 December 2024, 31 December 2025 and 30 June 2026. Basic and diluted counts are identical because there are no dilutive potential ordinary shares: no convertibles, no options granted, none exercised, none outstanding. There has been no issuance and no buyback, and the treasury block has been neither added to, sold, transferred nor cancelled. The count is flat, and that is the finding.

The Group does not need the market. FY2025 free cash flow of Rp1,940,880m covered the Rp177,270m paid to shareholders more than ten times over, cash rose from Rp5,945,500m to Rp8,573,289m to Rp9,052,399m across the period, and the liquidity backstop is a stack of committed working-capital bank lines rather than an equity programme. What does exist is authority rather than intent: shareholders renewed the general share-issue mandate at the April 2026 AGM permitting up to 50% of issued capital, with a 20% sub-limit for issues not made pro rata, alongside a share purchase mandate capped at 4%. A controlling shareholder at 86.04% can pass both without a single minority vote, and a 20% non-pro-rata placement is the mechanism by which a minority stake in a company like this gets diluted. Nothing suggests that is planned; the authority is standard for the market. Dilution risk is judged low on the evidence and structurally unconstrained in principle.

10. Litigation, Regulatory, and Contingent Liability Risk

The relevant disclosures are FY2025 Note 32, "Commitments and contingencies" (plasma receivables, sales commitments, capital-expenditure commitments, contingent liabilities), FY2025 Note 17, "Claims for tax refund", 1H2026 interim Note 13, "Forestry administrative charges under appeal", interim Note 15 on borrowings and defaults, and interim Note 21, "Subsequent events". Note 32(d) is the boilerplate-versus-specific test and is unambiguous: "As at 31 December 2025, there are no probable claims against the Group that may cause material impact to the Group." Note 21 is equally short: "There are no known subsequent events which have led to adjustments to this set of interim financial statements." There is no litigation note, no legal-proceedings item, no accrued loss reserve and no going-concern language anywhere in either document.

The material exposure is regulatory, not judicial, and it is large. On 1 December 2025 the Ministry of Forestry imposed forestry administrative charges of Rp2,337,480m under regulations governing forest-area governance and compliance. The Group paid the entire sum in cash on 30 December 2025, within the required timeframe, into an escrow account administered by the Forest Area Enforcement Task Force (Satgas PKH), then filed objection papers; a further Rp0.15 trillion was imposed and paid in May 2026, taking the total under appeal to Rp2.49 trillion, 19.3% of tangible book value, with the decision still pending at the date of the interim. The whole amount is recognised as an asset, "Forestry administrative charges under appeal", inside other non-current assets, with no provision. This is not an isolated action: the same task force reported recovering IDR10.2 trillion and reclaiming 5.8 million hectares from the palm-oil and mining sectors between February 2025 and May 2026, IDR3.4 trillion of it administrative fines, the programme is systematic, the counterparty is the state, and the appeal is against a policy rather than a claim.

Tax is the second exposure: claims for tax refund, advance payments plus "tax assessments being appealed to the taxation authorities", rose from Rp26,755m to Rp38,644m over the half, and the Group states that "uncertainties exist with respect to the interpretation of tax regulations". Third, the Group guarantees up to Rp523.1 billion of plasma bank loans (FY2024: Rp430.7 billion) with no recognised reserve, and PT SIMP gives upstream guarantees securing Rp2,266,268m of subsidiary borrowings.

Prior-disclosure carry-forward. The open items disclosed in the FY2025 accounts remain visible in the interim. Forestry charges: open, and escalated by Rp0.15 trillion. Plasma recoverability, a key audit matter: open, with the charge at almost double the prior-year rate. Tax assessments under appeal: open, and larger. The FY2024 rubber impairment of Rp296.2 billion: no reversal and no further charge. No covenant waiver, subpoena, class action, material weakness or going-concern qualification appears in either period. As of 29 June 2026: Indonesia's Attorney General's Office, which had questioned bank witnesses about PT SIMP's export transactions on 10 June 2026 after the Finance Minister publicly named the company in an under-invoicing and transfer-pricing investigation, stated it had found no evidence linking PT SIMP, on the ground that the company sells primarily into the domestic market. The 31 July 2026 interim makes no reference to the matter; given the timing that is consistent with there being nothing to disclose, but it is recorded because a reader relying on the filings alone would not know the episode happened. Each of these remains relevant to the risk assessment.

11. Accounting Quality and Disclosure Review

Ernst & Young LLP audited the FY2025 statements and issued an unqualified opinion dated 30 March 2026, engagement partner Lim Tze Yuen; no going-concern paragraph, no material weakness, no restatement. Two key audit matters were identified, goodwill impairment and the recoverability of plasma receivables, and the described procedures on both are substantive: an independent valuer's report reviewed for the Lonsum cash-generating unit, an internal valuation specialist engaged on discount rates and terminal growth, sensitivity analysis performed, operational assumptions benchmarked against history. The 1H2026 interim is explicitly unaudited and unreviewed, normal for the market but meaning the most recent balance sheet here carries no assurance at all.

Revenue recognition is simple and low-risk: goods transferred at a point in time, none over time, disaggregated by segment, channel, geography and product line. Non-GAAP measures are defined and reconcilable, "adjusted EBITDA" excludes foreign exchange and biological-asset fair-value movements and "core profit" additionally excludes the plasma credit-loss charge, and both are stated alongside the statutory figures rather than in place of them.

Three areas carry real estimation risk. Goodwill of Rp3,078,520m is 23.9% of tangible book and has not moved since the Lonsum acquisition, Rp2,909,757m of it in the single Lonsum integrated-estates cash-generating unit. In the October 2025 test the pre-tax discount rate on that unit was lowered from 12.46% to 11.52% and terminal growth raised from 4.95% to 4.97%, two changes that both increase headroom, in a year when Indonesian policy rates were not obviously 94 basis points lower. Management states no reasonably possible change would cause an impairment, but publishes no sensitivity figures. Biological assets of Rp1,452,511m are fair-valued, Rp957,642m of it on Level 3 unobservable inputs, and swung from a Rp135,498m loss in FY2025 to a Rp78,493m gain in 1H2026. Employee benefits are an unfunded Rp1,436,862m liability. Deferred tax is conservatively stated, a net Rp397,995m liability with only Rp6,254m of recognised tax-loss carry-forwards.

The accounting weakness is elsewhere. It is the Rp2.49 trillion of forestry administrative charges paid in cash and carried as a non-current asset while under appeal, Rp2,337,480m recognised in December 2025, 15.0% of shareholders' equity at that date, plus Rp0.15 trillion in May 2026. A contested payment to a state authority is recognised as an asset only where recovery is virtually certain; the Group's disclosure says only that objection papers have been filed and a decision is pending. No provision, no probability assessment, and, the point that matters here, the auditor did not identify it as a key audit matter in the very year it was created, while devoting two key audit matters to a goodwill balance that has not moved and a plasma allowance one-third its size. Segment transparency is a second weakness: an inter-segment transfer price that determined 94% of Plantations' CPO volume in 1H2026 is disclosed only as "based on terms agreed between the parties". Related-party disclosure, by contrast, is good.

Disclosure is classified Weak. The opinion is clean and primary-source verification of FY2023–FY2025 was full, which keeps it above Opaque; but material estimation risk met thin auditor pushback on the largest new judgement of the year, and the transfer-pricing mechanism behind most of one segment's revenue is undisclosed. The risk is elevated not because anything is provably wrong, but because the three largest soft numbers on the balance sheet, goodwill, the forestry receivable and fair-valued biological assets, total Rp7,021,031m, 54.5% of tangible book value, and all three move on management judgement.

12. Valuation and Margin of Safety

Currency convention. The Group reports in Rupiah and the shares trade in Singapore dollars. Figures are stated in Rupiah, translated at Rp13,806 per S$1, the rate the Group applies to its own 30 June 2026 balance sheet, with Singapore-dollar equivalents on the decision figures only, and US$1 = S$1.275 (4 September 2026). Earlier historical figures are not converted at today's rate.

Tangible Book Value, Derivation (anchor 2026-06-30, Condensed Interim Financial Statements for the six months ended 30 June 2026; Rp millions)

Tangible-book bridge at 30 June 2026 (Rp millions). Equity attributable to owners was Rp15,965,955m. Less goodwill of Rp3,078,520m and no other intangible balance gives tangible book value of Rp12,887,435m. Dividing by 1,395.905m current diluted shares gives Rp9,232.3183 per share. The Rp13,081,696m of non-controlling interest is excluded because the valuation uses equity attributable to IndoAgri's own shareholders.

P/TBV (derived) = Rp4,832.10 ÷ Rp9,232.3183 = 0.523× at the 4 September 2026 close.

P/TBV (derived) = 4,832.10 ÷ 9,232.3183 = 0.523× (price as of 2026-09-04)

Asset composition at 30 June 2026. Cash and short-term investments were Rp9,052,399m, or 70.2% of tangible book. Accounts receivable were Rp1,405,509m, inventory Rp3,797,764m, owned property, plant and equipment Rp15,761,690m, right-of-use lease assets Rp2,328,292m, long-term investments Rp2,257,074m, goodwill Rp3,078,520m, and other current and non-current assets Rp6,257,744m. Total assets were Rp43,938,992m.

Right-of-use assets were 18.1% of tangible book value. ROU is a capitalised right to occupy, not a liquidation floor, and it remains in the tangible-book figure used here. The balance is low enough that calling this cheap against tangible book does not require a larger lease adjustment.

Tangible book value per share is Rp9,232.32 (S$0.669) against reported net asset value of Rp11,438 (S$0.828), the Rp2,206 difference being goodwill. The one P/TBV this report uses is price divided by that derived figure: 0.523×. No adjusted tangible book is asserted, the downside cases test specific items that may prove worth less rather than pre-adjusting the bridge.

Historical Relative Multiple Valuation

The book leg uses the stock's own real trading history rather than a hand-built proxy: 40 quarterly price-to-tangible-book observations spanning 2016-06-30 to 2026-03-31 (10.0 years), each period's exchange closing price divided by the company's own reported book value at that date. The distribution: minimum 0.32×, p25 0.49×, median 0.54×, p75 0.61×, p90 0.73×, top-quartile mean 0.71×, maximum 0.82×. Today's reading of 0.523× sits at the 43rd percentile of that ten-year history, below the median, on the cheaper side of where this stock has actually traded, though not at either extreme.

A shorter, supplementary annual series (fiscal year-end closing price ÷ that year-end's tangible book value per share, FY2021–FY2025) tells the same directional story on a thinner sample:

The supplementary year-end observations were: FY2021, 0.496×; FY2022, 0.414×; FY2023, 0.409×; FY2024, 0.458×; FY2025, 0.506×; and 4 September 2026, 0.523×. FY2021 and FY2022 rely on the company's reported net asset value per share and a secondary goodwill figure rather than a fresh check of those audited statements. This short series is only a cross-check. The 40-quarter history above is the valuation anchor.

The earnings leg is unaffected by any of this and stands as its own read. Fiscal-year-end price-earnings ratios of 6.32×, 6.40×, 7.96×, 4.78× and 5.28× for FY2021 to FY2025 give a typical band of about 4.8×–8.0× with a median near 6.3×; the trailing multiple today is 4.91×.*

The comparison points are:

• P/TBV: 0.523× today. The 40-quarter history runs from 0.32× to 0.82×, with a p25 to p75 band of 0.49× to 0.61×, implying Rp4,524 to Rp5,632 per share.

• P/E on normalized EPS of Rp717: 6.74×, against a 4.8× to 8.0× historical band, implying Rp3,442 to Rp5,736 per share.

• P/E on trailing EPS of Rp986: 4.90×, against the same historical band, implying Rp4,733 to Rp7,888 per share.

The book and normalized-earnings legs overlap. The trailing-earnings leg is the outlier because Rp986 reflects a cyclical high, CPO near multi-year peaks, and half-year foreign-exchange and biological-asset gains. The book and normalized legs are the ones to lean on, producing a demonstrated range of roughly Rp4,500 to Rp5,100 (S$0.326 to S$0.369).

Two cross-checks support it. Peers: Singapore-listed palm names trade around 6–11× forward earnings, the sector's best-regarded mid-cap at roughly 8× with an 8% dividend yield. IndoAgri at 4.8× trailing and 6.5× normalized is at or below the bottom of that band, but with a 3.5% yield, an 86% controlling shareholder and two layers of minority leakage, so the discount is explained rather than dissolved. Look-through: IndoAgri's market capitalisation of about Rp6,745 billion compares with roughly Rp7,000 billion for its 73.46% of PT SIMP at the 4 September 2026 Jakarta close plus Rp2,257 billion of equity-accounted investments at book, the market prices the Singapore holding company about 29% below the sum, and about 6% below the listed stake alone. A normal holding-company discount, largely explained by the governance structure discussed above.*

Sell Range

The Sell Range applies the median and top-quartile mean of the stock's own P/TBV distribution to current tangible book value per share, using the real 40-observation quarterly series above (10.0 years, a genuinely deep history to draw one from): median 0.541×, top-quartile mean 0.712×.

Sell range: Rp4,996 – Rp6,574 per share (S$0.362 – S$0.476), i.e. 0.54× – 0.71× tangible book value. The flat 0.80× reference of Rp7,386 (S$0.535) is not used here because the full ten-year quarterly history is available. Today's price of Rp4,832 sits below the entire Sell Range, the shares are below the range where trimming would start.

Buy-Below: Rp3,889 per share (S$0.282), 0.42× tangible book, the mean of every observation at or below the twenty-fifth percentile of the same ten-year distribution. Today's price sits between Buy-Below and the Sell Range floor, meaning this name is buyable now on the numbers, without being at the deepest discount its own history has shown. On where an exit would eventually sit: return on equity has averaged 6.8% across the years behind this band (4.5%, 7.8% and 8.1%), and today's trailing figure of 8.6% is above that average, so the band is fully credible at today's earnings power rather than resting on a hoped-for recovery.

Scenario outcomes

Anchored to the stock's own P/TBV distribution and to specific balance-sheet items, not a generic percentage ladder. The lens is price to tangible book, what this sector and structure actually trade on; the earnings lens sits in the table above.

The four cases are:

• Severe downside: the forestry appeal fails in full and the net plasma receivable of about Rp515,000m is written off. Residual tangible book is about Rp7,079 per share; at the series floor of 0.32×, value is Rp2,265 (S$0.164), down 51.7%.

• Bear: the forestry appeal fails in full but the plasma balance holds. Residual tangible book is about Rp7,449 per share; at the series median of 0.541×, value is Rp4,031 (S$0.292), down 14.1%.

• Base: the forestry escrow is recovered in full and tangible book stays at Rp9,232 per share. At the series median, value is Rp4,997 (S$0.362), up 3.4% from the 4 September 2026 close.

• Bull: the forestry escrow is recovered, twelve months of retained earnings at the trailing rate lift tangible book to roughly Rp10,053 per share, and the multiple reaches the top-quartile mean of 0.712×. Value is Rp7,158 (S$0.519), up 52.5%.

Each impairment case strikes residual tangible book rather than unimpaired tangible book. The base case needs no heroic recovery assumption beyond this stock's own median multiple. A renewed take-private offer is an event rather than a valuation case: applying the 2019 offer's 21.5% premium to today's S$0.34 would be about S$0.41, or Rp5,700.

The Decision Figures, Stated

• Current price: S$0.350 (Rp4,832) on 4 September 2026; market capitalisation S$488.6 million (US$383 million).

• Intrinsic value range: Rp4,031 – Rp7,158 (S$0.292 – S$0.519), bear to bull, base case Rp4,997 (S$0.362), an analytical range, not a target.

• Margin of safety: a 47.7% discount to tangible book value, and a positive 3.4% margin against the base-case intrinsic value. Both point the same direction here: this is not a name priced at the top of its own history, and the base case does not require anything better than what this stock has already, repeatedly, been paid.

• Buy-Below: Rp3,889 (S$0.282), 0.42× tangible book, the mean of every observation at or below the twenty-fifth percentile of the stock's own ten-year quarterly distribution, 19.5% below today's price.

• Sell range: Rp4,996 – Rp6,574 (S$0.362 – S$0.476), 0.54× – 0.71× tangible book, 40 quarterly observations over 10.0 years; the flat 0.80× fallback of Rp7,386 is reference only. Today's price sits below the entire range, in the hold zone above Buy-Below.

• Cheapness type: asset-value cheap. Cash alone is 70.2% of tangible book, the balance is real mills and real trees, and today's 0.523× sits below the median of this stock's own ten-year trading history, a genuine, demonstrated discount, not a value trap dressed as one. The discount is not unexplained, though: an 86%-controlled Singapore holding company two levels above the operating assets, 45% of equity belonging to minorities of listed subsidiaries, and 39% of revenue sold to related parties are real, structural reasons the market has never fully closed this gap and may not soon. That argues for sizing the position to the governance risk, not for staying out of it.

13. Risk Matrix

Risk matrix.

• Forestry charges not recovered, high severity and medium probability. Rp2,337,480m was imposed on 1 December 2025 and paid on 30 December, followed by another Rp0.15tn in May 2026. The full Rp2.49tn is carried as an asset while under appeal, with no provision. The objection and escrow are the mitigants; watch the Satgas PKH decision and any provision in the FY2026 accounts.

• Structural minority discount, high severity and high probability. Anthoni Salim is deemed interested in 86.04%, while the stated float is 13.92%. This caps re-rating more than it threatens earnings. Watch substantial-shareholder notifications and any move toward 90%.

• Commodity price reversal, high severity and medium probability. Gross margin has fallen from 29.8% to 25.6% to 22.4% while CPO prices remain near multi-year highs. A 10% CPO fall is roughly Rp700bn of gross profit. Net cash and the downstream segment provide some offset; watch domestic and Rotterdam CPO prices and B50 policy.

• Goodwill impairment, medium severity and medium probability. Goodwill is Rp3,078,520m, or 23.9% of tangible book. The last test lowered the discount rate and raised terminal growth without publishing sensitivity figures. Watch the next annual impairment test.

• Plasma credit deterioration, medium severity and high probability. The 1H2026 charge was Rp122,575m, up 86.9%, the allowance covers 61% of gross receivables, and Rp523.1bn of guarantees has no recognised reserve. Watch the next credit-loss charge and any guarantee call.

• Estate yield decline, medium severity and high probability. Nucleus FFB production fell 2% in FY2025 and 4% in 1H2026; bought-in fruit is now 30% of throughput. Watch own-estate production, replanting and maintenance capital expenditure.

• Related-party pricing, medium severity and medium probability. Rp8.26tn, or 39% of FY2025 revenue, was sold to related parties, and 94% of 1H2026 CPO volume moved internally at transfer prices disclosed only as terms agreed between the parties. Watch the interested-person-transaction disclosure and segment margins.

• Short-dated refinancing, low severity and medium probability. Rp8,343,576m, or 92% of borrowings, matures within twelve months, but Rp9.05tn of cash covers it 1.09×. Watch bank-line renewals and the still-undisclosed covenant ratios.

• Regulatory action against the Salim group, medium severity and low probability. The Finance Minister named PT SIMP in a CPO export transfer-pricing probe; the Attorney General later said no evidence linked PT SIMP. Watch for any reopening of the investigation.

14. Red Flags, Yellow Flags, and Green Flags

🟢 Green Flags

• Net debt of Rp68,366m against Rp29,047,651m of total equity, effectively debt-free, with Rp9,052,399m of cash covering the entire current borrowing balance 1.09 times.

• Free cash flow of Rp1,940,880m in FY2025 against Rp965,771m in FY2024, with no accrual gap between reported and owner earnings.

• Dividend raised for a third consecutive year (S$0.008, S$0.010, S$0.012) on a 17% payout, with zero issuance or option dilution.

• Right-of-use assets are only 18.1% of tangible book against Rp45,455m of lease liabilities, so the tangible book is genuinely tangible rather than a capitalised lease.

• Today's 0.523× P/TBV sits at the 43rd percentile of the stock's own ten-year quarterly trading history, below the median, and between Buy-Below and the Sell Range floor, not at a historical extreme.

🟡 Yellow Flags

• The Piotroski floor rule fired: leverage rose (debt to average assets 20.4% to 22.3%) and liquidity fell (current ratio 1.50 to 1.27) in the same fiscal year, a real deterioration worth watching, not an automatic disqualifier, and already partly reversed at the half-year (current ratio back to 1.38).

• Gross margin has fallen for two consecutive periods, 29.8%, 25.6%, 22.4%, while selling prices rose: unit costs are outrunning a favourable price environment.

• Nucleus fruit production fell 2% in FY2025 and 4% in 1H2026; immature planted oil palm dropped from 23,806 hectares to 15,978, and bought-in fruit is 30% of throughput.

• Ninety-two percent of borrowings (Rp8,343,576m) mature within twelve months across six Indonesian banks, with covenant ratios and headroom not disclosed.

• The goodwill test cut the Lonsum discount rate from 12.46% to 11.52% and raised terminal growth to 4.97%, with no published sensitivity, both changes increase headroom.

• Brazilian sugar joint ventures cost Rp178,475m in FY2025 and Rp222,777m in 1H2026 alone, with carrying value down to Rp884,305m in six months.

🔴 Red Flags

• Rp2.49 trillion, 19.3% of tangible book value, of forestry administrative charges paid in cash to a government escrow in December 2025 and May 2026 and carried as a non-current asset while under appeal, with no provision, no probability disclosure and no key audit matter.

• Anthoni Salim is deemed interested in 86.04% with a 13.92% public float, up from 74.34% at the failed 2019 exit offer at S$0.28; every shareholder mandate passes without a minority vote.

• Rp1,084,149m of plasma receivables written off in FY2024 with "no reasonable expectation of recovering the cash flows"; the charge now runs 86.9% above last year, the allowance covers 61% of gross, and Rp523.1 billion of plasma guarantees carries no reserve at all.

• Roughly 39% of FY2025 revenue was sold to related parties and 94% of 1H2026 CPO volume moved internally at transfer prices disclosed only as "terms agreed between the parties".

⚡ Must-Watch Catalysts

• The Satgas PKH decision on the Rp2.49 trillion objection, the largest binary item in the analysis, timing unknown.

• FY2026 full-year results, expected late February or March 2027: the next F-Score reading, and whether the 1H2026 recovery in the current ratio to 1.38 holds.

• The final FY2026 dividend, a fourth consecutive increase, or the first break in the sequence.

• Any substantial-shareholder notification moving the controlling group toward the 90% compulsory-acquisition threshold, or a renewed exit offer.

• The October 2026 goodwill impairment test on the Rp2,909,757m Lonsum cash-generating unit; Indonesian CPO prices and B50 mandate policy.

Recommendation

Verdict: BUY — SMALL

The numbers support a buy. The shares trade below the median of their own ten-year quarterly P/TBV history, the F-Score is 6/9 with all signals computable, and the balance sheet is close to net-cash neutral. The position stays Small for a different reason: an 86.04%-controlled family, a prior take-private attempt, and a demonstrated pattern of routing capital into policy-driven uses. That is structural minority risk. It argues for sizing down, not for pretending the risk is absent.

• Current price: S$0.350 (Rp4,832) at the 4 September 2026 close; market capitalisation S$488.6m (US$383m).

• Eligibility: P/TBV is 0.523×, below the 1.00× value-investing cutoff. Financial borrowings are 31.2% of total equity and 56.8% of equity attributable to owners; total liabilities are 51.3% of total equity.

• Piotroski: 6/9. The failed leverage, liquidity and gross-margin signals are real; the first two partly reversed by 30 June 2026.

• Valuation: 0.523× P/TBV, 6.74× normalized P/E on a three-year average EPS of Rp717, and 4.90× trailing P/E on EPS of Rp986.

• Cash return: 3.43%, all dividend. The FY2025 S$0.012 dividend was paid on 28 May 2026; there were no buybacks.

• Intrinsic value range: Rp4,031 to Rp7,158 per share, bear to bull, with a base case of Rp4,997 (S$0.362). The base case is 3.4% above today's price.

• Buy-Below: Rp3,889 (S$0.282, 0.42× tangible book), 19.5% below today's price.

• Sell range: Rp4,996 to Rp6,574 (S$0.362 to S$0.476), 0.54× to 0.71× tangible book. Today's price is below the range where trimming would start.

• Position size: Small, reduced for governance risk rather than for the valuation or balance sheet.

• Holding period: Three to five years, spanning a likely forestry decision and a full palm-oil cycle.

• Main downside: Rp4,031 in the bear case and Rp2,265 in the severe case. The practical balance-sheet support is Rp9.05tn of cash, equal to 70.2% of tangible book.

• Key thesis breakers: a full forestry loss, a second consecutive year of leverage up and liquidity down, a CPO reversal into a rising unit-cost base, a call on the unreserved plasma guarantees, or further erosion of minority standing.

What to watch

• The Satgas PKH decision on the Rp2.49 trillion of forestry charges under objection. Timing unknown, outcome binary: full recovery leaves tangible book at Rp9,232 per share, a full loss takes it to Rp7,449.

• Any provision raised against the forestry asset in the FY2026 accounts, or a change in how it is classified, a move from "asset" to "expense" is the fastest way this thesis resolves.

• FY2026 full-year results, expected late February to March 2027: the next Piotroski reading, and specifically whether the leverage/liquidity combination repeats for a second consecutive year, that would be a materially stronger signal than the one year seen so far.

• The FY2026 final dividend. Three consecutive rises (S$0.008, S$0.010, S$0.012) on a 17% payout; a fourth would be evidence the controlling shareholder is willing to distribute, a break would be evidence it is not.

• The half-yearly plasma credit-loss charge, running Rp122,575m in 1H2026 against Rp103,041m for all of FY2025. Any resumption of write-offs, or any call on the Rp523.1 billion of unreserved plasma guarantees.

• Renewal of the Rp8,343,576m of short-dated bank lines across Mandiri, BCA, DBS, SMBC, BNI and Permata in the second half of 2026, and any first disclosure of the financial covenant ratios attached to them.

• Substantial-shareholder notifications. The controlling group has moved from 74.34% in 2019 to 85.90% today; 90% triggers compulsory acquisition, and a float below 10% triggers a listing problem, either would be reason to trim rather than add.

• A renewed exit or delisting offer. The 2019 attempt was S$0.28 at a 21.5% premium to the one-month volume-weighted average; the same premium today would be roughly S$0.41 (Rp5,700), inside this report's own base-to-bull range, priced by the buyer rather than the market.

• The Sell Range floor at Rp4,996 (S$0.362), the price at which trimming a position would start to make sense, well above today's level.

• The Buy-Below price of Rp3,889 (S$0.282), 19.5% below today's level, the price at which the stock reaches a materially deeper own-history discount.

• The October 2026 goodwill impairment test on the Rp2,909,757m Lonsum cash-generating unit, and specifically whether the pre-tax discount rate moves back up from 11.52%.

• Domestic KPB and Rotterdam CPO prices, and any change to Indonesia's B50 biodiesel mandate or export levy regime, the biggest single driver of the earnings line.

• Quarterly nucleus fruit production and replanting capital expenditure. Own-estate output is falling and immature area has shrunk from 23,806 to 15,978 hectares; the fix is capex the company has not yet committed.

• Any reopening of the Indonesian Attorney General's palm-oil export investigation in a way that names PT Salim Ivomas Pratama again, after the 29 June 2026 statement that no evidence had been found.

Sources and diligence gaps

This analysis draws on Indofood Agri Resources' 2025 annual report, the condensed interim statements for the six months ended 30 June 2026, the 2026 AGM notice, the 1H2026 results presentation, the statutory directors' interests register, SGX announcements, and public records on the forestry enforcement programme and the 2019 exit offer. The annual report is available from the company's investor-relations archive; exchange announcements are available through the SGX company-announcement archive.

The following gaps remain material: FY2021 and FY2022 audited statements were not retrievable, so two supplementary annual valuation observations rely on reported historical net asset values and a secondary goodwill figure; the exchange announcement index after 21 January 2026 could not be checked for later insider or substantial-shareholder dealings; the specific bank covenant ratios and headroom are not disclosed; and the inter-segment transfer price for the internally moved CPO volume is not disclosed. None of those gaps changes the tangible-book bridge or the BUY, SMALL conclusion, but each matters to how confidently a minority holder should size the position.

Research only, not investment advice. Position disclosure: not held as of this review.

1. Scorecard

Position: Long, ~0.26% of portfolio (16 shares; as of 2026-09-04)

2. Argument

The short version

Huons trades at 0.798 times tangible book value, the cheapest its shares have been anywhere in a ten-year listed history. Most of that book is real: 83% sits in owned plant and equipment, not leases or intangibles. The company clears the standard debt and quality screens, with a Piotroski F-Score of 7 of 9 and borrowings at 38.9% of equity.

The discount exists because the business is genuinely disrupted. First-half 2026 operating income fell from KRW 25.9bn to KRW 2.28bn and the group posted a small net loss, after the FDA found fabricated laboratory records at the company’s Jecheon plant and halted its US injectable exports in April 2026, followed by a formal warning letter in June. The company’s own sanctions disclosure does not mention the warning letter.

Two of five capital-allocation warnings fired: an unexplained KRW 5.28bn provision that lines up with the recall, and a related-party merger, since abandoned, that briefly halved the share price. The numbers-only tier is Standard, but final sizing is Small given the combination of the fired warnings, a fabrication finding that was deliberate rather than accidental, and FDA’s own rejection of the company’s first attempt to fix it.

Why it qualifies

• Trades at 0.798x tangible book, below every one of 38 quarterly observations in its own ten-year history.

• 83% of tangible book is owned land, plant and machinery, not leased space or goodwill.

• Piotroski F-Score 7/9; financial borrowings are 38.9% of equity with KRW 60.8bn of undrawn bank credit.

• Paid a 5.2% dividend yield over the trailing twelve months, funded entirely from cash, with no equity issued for cash in five years.

Why it is not a clean bargain

• FDA warning letter (June 2026): the microbiology team leader admitted to using a knife to cut completed pages out of a logbook, replaced them with rewritten substitutes disguised to look original, and instructed an analyst to backdate and omit bioburden test results. FDA judged Huons’ first corrective-action response (staff dismissals, a new data-integrity team) insufficient. The company’s own sanctions disclosure omits the letter entirely.

• Import Alert 66-40 blocks all drugs from the Jecheon plant from US entry, not only the three recalled injectables, and the inspection also cited separate aseptic-processing gaps (airflow, glove monitoring, disinfection procedures) beyond the paperwork fraud.

• First-half 2026 operating income fell 91% year on year and the group posted a net loss; gross margin has declined for three straight years, from 51.8% to 43.1%.

• A KRW 5.28bn provision, the largest new balance-sheet item, has no explanatory note anywhere in the half-year filing.

• A board-approved merger with a related-party affiliate was rescinded in August 2026 after the share price roughly halved and shareholders objected.

• Three years of free cash flow were negative in aggregate as capital spending outran operating cash flow, most of it funding the plant now under US import restriction.

What would change the view

• FDA re-inspection of the Jecheon plant, targeted by management for the fourth quarter of 2026, and formal closure of the warning letter; treat this as optimistic, not a base case, since FDA already judged the first corrective-action response insufficient.

• Resumption of North American injectable shipments, which management has guided for early 2027.

• Third-quarter 2026 results, due in November, as the first clean read on whether the second quarter’s operating profit was a turn or a pause.

• Whether the FY2026 annual report finally explains the KRW 5.28bn provision left unexplained in the half-year filing.

• Any second attempt at a related-party restructuring involving Huons Lab or another affiliate.

• A share price move to roughly KRW 35,000-40,500 (1.3x-1.5x tangible book), the level at which the exit question realistically goes live.

Sources and gaps

The FDA warning letter’s own page could not be fetched directly; its findings come from a search-indexed copy of the same page plus a third-party regulatory summary quoting it, cross-checked against the company’s own filing for the underlying facts. Debt covenant terms and segment-level profitability are not disclosed in either filing.

3. Backup and sources

This last layer is the full working file: every source, calculation, and gap, kept so the argument above can be checked. The scorecard and the argument are the synthesis. Open this only if you want to check the work.

Huons Co., Ltd. (KOSDAQ:A243070): Cheap against tangible assets, disrupted at the operating line

1. Executive Summary

Huons Co., Ltd. is a Korean pharmaceutical manufacturer that makes and sells prescription drugs (local anaesthetics, ophthalmic solutions, cardiovascular and metabolic products), beauty-and-wellness products (high-dose vitamin injectables, obesity drugs, health functional foods) and contract-manufactured medicines for third-party pharmaceutical companies, earning roughly 90% of its revenue inside Korea from two GMP plants in Jecheon, North Chungcheong Province. The numbers-only tier is STANDARD. After the governance review, final sizing is BUY — SMALL, because the valuation is attractive while the operating recovery and quality-system remediation remain unproven: the shares change hands at 0.798× tangible book value, the cheapest they have ever been in the ten-year history of the listed entity, while the company is simultaneously working through a US Food and Drug Administration warning letter whose findings include falsified laboratory records.

The valuation case is straightforward and asset-backed rather than earnings-driven. Tangible book value at 30 June 2026 was KRW 314.6bn, or KRW 27,003.53 per share, against a price of KRW 21,550 — a 20.2% discount to tangible book, with 83.2% of that tangible book sitting in owned property, plant and equipment and only 0.6% in right-of-use lease assets, so the discount is to real land, buildings and machinery rather than to capitalised occupancy rights. Financial borrowings of KRW 157.8bn against total equity of KRW 405.1bn (38.9%) clear the debt filter with room to spare, and the Piotroski F-Score is 7/9. Cash flow is intact: operating cash flow in the first half of 2026 was KRW 34.9bn, up 78% year on year, and the company paid KRW 13.1bn of dividends over the trailing twelve months, a 5.2% cash yield on the current market capitalisation.

What the price is discounting is a genuine and unresolved operating shock. First-half 2026 operating income fell to KRW 2.28bn from KRW 25.90bn a year earlier and the group recorded a net loss of KRW 0.89bn, with profit attributable to owners of just KRW 92m. The proximate cause is regulatory: following an inspection of the Jecheon plant from 12 to 21 November 2025, the FDA cited the company in April 2026 for insufficient sterility assurance, prompting a recall of US-bound product and a complete halt to North American shipments, and then issued a warning letter dated 15 June 2026. The company’s own filing states plainly that no exports to the United States arose in the current period and that GMP remediation is under way. Underneath the regulatory event sits a slower structural problem: consolidated gross margin has fallen from 51.8% in FY2023 to 48.3%, then 45.5%, and 43.1% in the first half of 2026.

Governance Risk is High. Two of the five capital-allocation warnings fire, and the accompanying facts are not cosmetic: a KRW 5.28bn sales-warranty provision — the largest new balance-sheet line of the period and, on the evidence, the recall reserve — appears in the half-year balance sheet with no explanatory note anywhere in the document; the FDA warning letter is absent from the report’s sanctions disclosure, which states “not applicable”; a KRW 9.03bn receivable from affiliate Huons USA, Inc. is fully provided against; and a proposed absorption of group R&D affiliate Huons Lab, resolved by the board on 18 May 2026, was abandoned on 26 August 2026 after the share price roughly halved and shareholders objected to the exchange ratio. A reader should treat Standard as the numbers-only tier and Small as the final sizing after the High governance flag is applied.

2. Business and Market Overview

Huons was carved out of Huons Global in a 2016 demerger and relisted on KOSDAQ on 3 June 2016; the underlying pharmaceutical business dates to 1965. It reports three revenue lines plus consolidated subsidiaries. Prescription pharmaceuticals were KRW 280.2bn of FY2025 revenue (45%), a broad multi-therapy portfolio — lidocaine local anaesthetics, the hyaluronic-acid eye drop Kynex, cardiovascular, metabolic, gastrointestinal and antibiotic products — sold to Korean hospitals, clinics and pharmacies. Beauty and wellness contributed KRW 162.7bn (26%): vitamin injectables, obesity drugs, arthritis injections, medical devices and the Merit health-functional-food range. Contract manufacturing was KRW 78.2bn (13%), mainly eye drops and injectables produced for other Korean and overseas pharmaceutical companies. Consolidated subsidiary revenue net of eliminations was KRW 108.0bn (17%), up from KRW 61.2bn in FY2024 as Panagene and Biorojet were consolidated.

The economics are those of a mid-sized generic-plus-specialty manufacturer in a price-regulated market. The filing is explicit that government sets pharmaceutical prices directly and that reimbursement policy therefore feeds straight into revenue and margin. Revenue is highly granular — the company states its sales consist of many products in small quantities, which is why it discloses by segment rather than by product — and no single external customer reaches 10% of revenue. Capital intensity is high and rising: gross property, plant and equipment stood at KRW 445.2bn at 30 June 2026 against KRW 288.9bn of half-year revenue, following an KRW 83.3bn property-and-equipment spend in FY2024 to build out the Jecheon second plant. Research and development ran at KRW 41.8bn in FY2025, 6.73% of revenue and rising from 5.91% in FY2024, funding a dry-eye candidate in Phase 2, a glaucoma candidate in Phase 3 and an obesity candidate in Phase 1.

As of 2026-06-30, exports were KRW 27.8bn of half-year revenue against KRW 29.2bn a year earlier, a 5.0% decline that understates the damage because North American injectable sales — KRW 18.4bn in FY2025 after 52.3% growth — went to zero part-way through the period. As of 2026-06-23, wholly-owned subsidiary Huons Bioscience was absorbed into the parent, and as of 2026-07-01 grandchild company Biorojet was absorbed into Huons N. The continuous glucose monitoring business was discontinued during the period. This is an understandable business, moderately predictable in its domestic base, not obviously cyclical, and currently disrupted rather than structurally melting — but the disruption is regulatory, and regulatory timelines are not the company’s to set.

3. Moat, Competitive Position, and Industry Cycle

Tested against the five sources of durable advantage, Huons scores thinly. Intangibles and brand: real but shallow — the Merit and Innerset consumer ranges have shelf presence, and FDA abbreviated new drug approvals for 1% and 2% lidocaine and 0.75% bupivacaine injectables were genuine regulatory assets, but the June 2026 warning letter has for now converted the most valuable of those into a liability. Switching costs: modest — requalifying a sterile line is expensive, which is why the Spectra Medical supply agreement has run since 2018 with KRW 58.1bn shipped cumulatively, but the same logic means a customer that does requalify elsewhere does not come back quickly. Network effects: none. Cost advantage: none demonstrated; gross margin is falling, not rising. Efficient scale: partial, in the narrow niche of Korean sterile eye-drop and injectable contract manufacturing, where GMP-certified capacity is scarce.

The moat is Narrow, and the honest reading is that it is narrowing. A company whose principal export franchise rests on regulatory approvals, and which has just been told by that regulator that its quality systems cannot be relied upon, does not have a durable competitive position in that franchise until it is re-inspected and cleared. Return on equity of 11.4% in FY2025, 8.5% in FY2024 and 15.5% in FY2023 does not evidence a compounding machine; it evidences a decent domestic manufacturer with a volatile export and licensing overlay. Nothing in the record suggests Huons can reinvest at rates that would justify paying a premium to tangible book.

Cycle position is Trough — but a company-specific trough rather than an industry one. Korean domestic pharmaceutical demand is not in recession; the margins of this company are. First-half 2026 operating margin was 0.79% against 8.58% a year earlier and 7.35% for FY2025 as a whole. Second-quarter operating income of KRW 2.93bn against a first-quarter operating loss suggests the bottom was the first quarter, but past the worst quarter is not the same as recovering, and the FDA import restriction that removed the export line was still in force at the balance-sheet date.

4. Management, Governance, and Capital Allocation

Song Su-young has been chief executive and board chair since roughly March 2022, was re-elected on 26 March 2025 with a term to 26 March 2028, and comes from an industrial-engineering and consulting background — Samsung Electronics, SAP Japan, Deloitte Consulting Japan — rather than from pharmaceutical manufacturing or quality assurance. The board has eight members: four inside directors, one non-executive director (the group chairman) and three outside directors who together constitute the audit committee. Directors are elected by shareholders at the annual general meeting. Board attendance was 100% for every director except one at 83%.

Compensation is not the problem here. Eight registered directors received KRW 1.80bn in FY2025 against a shareholder-approved cap of KRW 3.00bn; the group chairman was the only individual above the KRW 500m disclosure threshold at KRW 661m, all salary, no bonus, no option gains. Share-based compensation is small: the FY2025 restricted-stock charge was KRW 484.9m equity-settled plus KRW 250.4m cash-linked, and no options remain at the parent. Total FY2025 share-based compensation of roughly KRW 1.0bn is 2.4% of attributable net income and 3.0% of normalised net income, well below the threshold this method flags, and has not doubled year on year.

Insider activity is where the record gets interesting. The only significant movement disclosed for FY2025 was a gift: the entire personal holding of the group chairman, 477,750 shares or 3.99%, went to zero, distributed to three family members including a serving inside director who rose to 404,672 shares (3.38%). This is succession, not conviction buying or selling, but it tells a reader that the second generation is being positioned inside the operating company. No purchases or sales by directors or executive officers were found for 2026 in searches of the Korean insider-reporting regime; the absence of such a filing in a search is not proof that none occurred, and this is recorded as a diligence gap rather than as a finding.

Related-party dealing is extensive and warrants naming in full. In the first half of 2026 the group bought KRW 23.8bn of inventory from affiliates, plus KRW 2.4bn of outsourced processing, and paid KRW 10.1bn to parent Huons Global under other expenses, a line the note says includes dividends and which also carries headquarters occupancy. Huons Global is a pure holding company whose income consists of brand fees, service fees, rent and dividends. The filing discloses the amounts but nowhere discloses the pricing basis or benchmarks any of it to arm’s-length terms. On the financial side, the parent has provided joint-and-several guarantees for subsidiary borrowings totalling roughly KRW 25.7bn. And a KRW 9,025m trade receivable from affiliate Huons USA, Inc., arising from COVID-19 quarantine goods and medical devices, is fully provided against because collection is uncertain following a decline in local sales — roughly half a year of the normalised earnings of the whole group written off inside a related-party balance and disclosed in one sentence.

Capital allocation over the lookback has been mixed and increasingly expensive. Dividends per share ran KRW 630 (FY2023), KRW 630 (FY2024) and KRW 920 (FY2025), a 25% payout of attributable earnings, and in February 2026 the board adopted a 2026-2028 policy of KRW 200 per quarter, KRW 800 per year, rising 5-30% annually — funded, the filing states, from capital reserves transferred into retained earnings as a tax-exempt reduction dividend. That transfer was KRW 50bn in FY2025, moving share premium into distributable reserves. Consolidating the repurchase activity into one figure: the company bought back KRW 3.97bn of stock in FY2025 and KRW 2.08bn in FY2024 — KRW 6.04bn in total across the two years, with nothing repurchased in the first half of 2026. Treasury stock stood at 327,904 shares (2.74%) at 30 June 2026, most of it earmarked for an exchangeable bond or restricted-stock settlement, leaving little genuinely retirable. Acquisitions have not gone well: goodwill was impaired by KRW 2.28bn in FY2024 and KRW 1.70bn in FY2025, and unrecognised tax losses at group companies, on which no deferred tax asset is carried because realisation is judged highly unlikely, rose from KRW 32.0bn to KRW 69.1bn in a single year.

Running the capital-allocation warnings one by one. Control without minority protection: not fired, borderline — the controlling group holds 44.82%, above the threshold, but genuine director-election items occur at the annual meeting and the merger episode below shows minority shareholders can in practice defeat a controller-sponsored transaction. Capital routed to parent-mandated or low-return projects: fired. The board resolved on 18 May 2026 to absorb group R&D affiliate Huons Lab; the shares roughly halved, the extraordinary general meeting was postponed, and on 26 August 2026 the board rescinded the agreement, citing the gap between the merger consideration and the market price. Set alongside the KRW 9.0bn Huons USA write-off, KRW 3.98bn of goodwill impairment in two years and the KRW 83.3bn FY2024 plant build now under an import restriction, the pattern is capital directed toward group objectives rather than returns.

Dividend cut or inadequate payout: not fired, borderline — no cut occurred, but the FY2026 policy target of KRW 800 per share is 13% below the KRW 920 actually paid for FY2025, and funding distributions from a capital-reserve transfer rather than from earnings is a return of capital dressed as a yield. Non-answers on capital allocation, related parties or loss-making subsidiaries: fired. The KRW 5,277,642,666 sales-warranty provision is the single largest new item on the half-year balance sheet and carries no note anywhere in the document; the FDA warning letter and import restriction appear nowhere in the sanctions section of the report, which reads not applicable as of the report date; and the Huons USA receivable write-off gets one sentence with no discussion of how it accumulated. Silence counts. Outsized pay or off-market related-party deals: not fired, borderline — pay is modest and the share-based-compensation ratio test is passed comfortably; the unexplained pricing basis on KRW 23.8bn of affiliate purchases is a real unanswered question but not, on this record, evidence of off-market terms.

Two warnings fired. Governance Risk: High.

5. Corporate Ownership, Subsidiaries, and Joint Ventures

Ownership is concentrated in the group holding company, with one large institutional holder and a genuine retail float.

Ownership is concentrated in Huons Global at 40.74%, followed by Fidelity Management & Research at 9.99%, an inside director at 3.38%, other related parties at 0.70%, and treasury shares at 2.74%. The reported minority-holder group represents 36.58% of issued shares, while the controlling group and related parties hold 44.82% in total. Huons Global therefore controls the company in practice, but does not hold the two-thirds needed to pass a special resolution unaided.

The rows do not sum to 100% because the disclosed small-shareholder population excludes every holder of 1% or more, and holders between 1% and 5% are not separately identified; the controlling group and its related parties hold 5,369,472 shares, 44.82%, in total. The group chairman controls 42.76% of Huons Global and therefore controls Huons at one remove; his direct personal holding in Huons itself is now nil following the FY2025 gifts. The controlling group together holds 44.82%, comfortably above the level needed to block a special resolution but short of the two-thirds needed to pass one unaided — which is precisely why the Huons Lab merger could be, and was, defeated.

The disclosed group structure includes Huons N, Panagene, Biorojet, the former Huons Bioscience subsidiary, and Beijing Huonland Pharmaceutical. Panagene is consolidated on de facto control even though the group owns 42.66%, leaving 57.34% of its economics with outside holders. Biorojet was absorbed into Huons N on 1 July 2026, while Huons Bioscience was absorbed into the parent on 23 June 2026. Beijing Huonland is an equity-method associate.

The structure carries two specific leakage risks worth naming. Panagene is consolidated on de facto control with a minority interest larger than the parent stake, so 57.34% of its economics — currently losses of KRW 452m in the half — belong to outside holders, while 100% of its KRW 9.78bn of goodwill is deducted inside the tangible-book bridge; the group non-controlling interest balance of KRW 36.8bn is not available to Huons shareholders. Second, the parent has extended joint-and-several guarantees to subsidiary lenders totalling KRW 25.7bn, and has pledged KRW 218.5bn of property, plant and equipment, 83% of net book value, against KRW 102.9bn of bank borrowings — so the tangible asset base backing the equity is substantially encumbered.

6. Historical Financial Quality and Normalized Owner Earnings

Three audited consolidated years plus one half are available directly from the primary filings; FY2022 is included for trend context from a secondary source and marked accordingly.

Fundamentals (KRW millions; per-share figures in KRW)

The three-year consolidated record is shown in the financial-quality exhibit below. It keeps the useful comparisons together without making the reader fight a wide text table.

The record does not describe a compounder. Revenue grew every year, but each step cost gross margin. Cash generation is the more important number and the more uncomfortable one: aggregate free cash flow across FY2023 to FY2025 was negative KRW 3.6bn. The separate cash-conversion exhibit shows why the earnings quality deserves scrutiny.

Net income is the amount attributable to owners of the parent. For FY2022 the reported figures were revenue KRW 492.4bn, operating income KRW 40.9bn and net income KRW 25.0bn, taken from a secondary source and used only for trend context.

The record does not describe a compounder. Revenue grew every year, but each step cost gross margin: 51.8% in FY2023, 48.3% in FY2024, 45.5% in FY2025 and 43.1% in the first half of 2026 — 870 basis points in two and a half years as the mix shifted toward lower-margin contract manufacturing and newly consolidated subsidiaries. FY2025 operating income was still below FY2023 on 12% more revenue. Cash generation is the more important number and the more uncomfortable one: aggregate free cash flow across FY2023 to FY2025 was negative KRW 3.6bn, because the group spent KRW 153.3bn on property, plant, equipment and intangibles over three years, most of it on the Jecheon second plant whose sterile injectable line is the subject of the FDA action. Three years of earnings went into an asset that is currently not permitted to serve its highest-value market.

Normalized owner earnings. FY2025 operating cash flow of KRW 45,005.5m less a maintenance-capex proxy of KRW 20,775.1m (the year’s own property-and-equipment depreciation charge) gives owner earnings of KRW 24,230.4m, or KRW 2,080 per share. That is roughly 57% of reported attributable net income, and the gap is the honest measure of how much of the reported profit is being consumed by the asset base. In the first half of 2026 operating cash flow was KRW 34,859.0m against KRW 19,562.9m a year earlier, but the improvement is not earnings — it is a KRW 11,067.7m working-capital release as trade receivables fell from KRW 114.5bn to KRW 95.0bn on lower activity, and it will not repeat.

Symmetry check on normalization. The FY2025 effective tax rate was 13.3% against 23.6% in FY2024. Taxed at the prior year’s rate, FY2025 consolidated net income would have been roughly KRW 37.2bn rather than KRW 42.2bn — about KRW 5.0bn, or 12%, of the reported figure came from a tax benefit whose driver is not separately explained in the tax note. Adding back one-off charges without removing this would overstate normalized earnings; both are therefore carried.

Normalized EPS and its basis. The earnings trajectory is erratic rather than growing or steadily declining, so the rule for an erratic earner applies: normalized EPS is the lesser of the multi-year average and the trailing twelve months. The four-year average is KRW 3,124; trailing twelve months is KRW 1,619. Normalized EPS is therefore KRW 1,619, and normalized P/E at KRW 21,550 is 13.3×. Two alternative readings are given for context and are not used in the decision: adding back the recall provision at an assumed 22% tax rate lifts trailing EPS to about KRW 1,972 and the multiple to 10.9×; the pre-disruption FY2023-FY2025 average EPS of KRW 3,466 would put the multiple at 6.2×. The tier this analysis reaches is the same on all three readings, which is a useful robustness result.

Earnings quality is otherwise reasonable but not pristine. Operating cash flow exceeded net income in each of the last three years. Inventory rose to KRW 96.6bn from KRW 94.4bn while revenue fell, consistent with the export halt leaving finished goods on the balance sheet. Selling commissions of KRW 46.8bn in the half, 38% of selling and administrative expense, are the largest cost line after cost of sales and are the mechanism through which Korean pharmaceutical distribution economics compress margin.

6.5 Piotroski F-Score — 7/9

Periods compared: FY2025 vs FY2024.

The nine-signal audit produces 7/9. The two failed signals are gross-margin improvement and asset-turnover improvement, precisely the operating signals that matter most here. The score compares FY2025 with FY2024; it is a fact about last year, not a clean read on the first half of 2026.

Score: 7/9. The failed signals are gross margin improved and asset turnover improved — precisely the two that matter most here. A company can score 7/9 while its unit economics deteriorate, because six of the nine are balance-sheet and cash-flow tests a well-capitalised manufacturer passes almost automatically. The failures name the problem: margin is being given away and the enlarged asset base is not yet producing revenue in proportion. All nine signals were computable and the floor rule did not fire: leverage fell and liquidity improved in the same year. It must be said plainly that the score compares FY2025 with FY2024, and that the first half of 2026 — a net loss, a further 240 basis-point fall in gross margin and a current ratio down to 1.41 — would fail six of the nine signals on the same test. The 7/9 is a fact about last year, not about now.

7. Balance Sheet, Debt, Covenants, and Refinancing Risk

The capital structure at 30 June 2026 is entirely Korean bank debt plus one small privately placed exchangeable bond. There is no public bond issue, no rated debt and no syndicated credit agreement.

The instrument-level borrowing detail is shown in the debt schedule exhibit below. The liquidity-coverage exhibit compares contractual near-term maturities with cash, current financial assets, and undrawn committed facilities; the current ratio is context, not a cash resource.

Financial borrowings and lease liabilities are stated separately throughout; leases are 1.3% of the borrowing total and immaterial to any leverage conclusion. Financial borrowings are 38.9% of total equity of KRW 405.1bn; total liabilities of KRW 244.9bn are 60.4% of total equity. On either basis the debt test used here — total debt not more than 100% of equity — passes comfortably, and it also passes on the stricter total-liabilities reading.

Two things temper that. The first is the maturity wall: KRW 111.4bn of borrowings including interest falls due within twelve months against KRW 58.6bn of cash and current financial assets. Almost all of it is rolling Korean bank paper that has been renewed repeatedly, and undrawn headroom is real — KRW 148.0bn of Korea Development Bank facility commitments against KRW 87.2bn drawn — but a company under an FDA warning letter renegotiating KRW 111bn of bank paper is not negotiating from strength. The second is coverage. FY2025 operating income covered finance costs 5.9 times and EBITDA covered them 9.4 times; in the first half of 2026 operating income did not cover finance costs at all, and EBITDA covered them 4.4 times. No covenant terms, financial ratios or headroom tests are disclosed anywhere in either filing, which is itself a gap.

Debt is 24.3% of total assets, which on the leverage leg alone reads Conservative; interest coverage on the current half-year run rate reads Elevated. Taking both legs together with the twelve-month maturity concentration and the absence of any disclosed covenant terms, balance-sheet risk is classified Manageable. A second consecutive half at first-half 2026 profitability would move it to Elevated.

8. Real Estate, Leases, and Hidden Assets

The tangible book is unusually literal. Owned property, plant and equipment of KRW 261.9bn is 83.2% of tangible book value and 40.3% of total assets, comprising land, buildings, machinery, tools and equipment, and construction in progress. A further KRW 22.4bn sits in investment property, the leased-out portion of the group’s buildings and land, carried at cost less depreciation.

Right-of-use assets are KRW 1,932.3m, equal to 0.6% of tangible book value, against lease liabilities of KRW 1,994.3m. Right-of-use is a capitalised right to occupy, not a liquidation floor, and it is disclosed here rather than stripped out of the tangible book value used everywhere in this analysis, since removing the asset without its matching liability would be the wrong operation. In this case the point is close to academic: at 0.6% of tangible book, the asset-backed framing survives the check intact, and the genuine downside support is owned land and buildings plus KRW 58.6bn of cash and current financial assets.

There is one identifiable hidden value and one identifiable offset. The hidden value: investment property carried at KRW 22,370.1m has a disclosed fair value of KRW 26,170.9m, an unbooked surplus of KRW 3,800.9m or KRW 326 per share, assessed by the group using published land prices and price indices rather than an independent valuer. The offset, and it is the larger of the two: KRW 218,498.3m of property, plant and equipment, 83% of the net book value of the whole fixed-asset base, is pledged as security for KRW 102,898.0m of bank borrowings. A tangible book that is 83% owned plant is a real floor only to the extent the plant is unencumbered, and here most of it is not.

9. Capital Markets Access, Dilution, and Financing Flexibility

The share count has been stable and is now shrinking slightly. Shares in issue went from 10,859,671 at end-2021 to 11,979,665 at end-2023 and have not moved since; the increases were stock dividends and a small convertible-bond conversion, and there has been no equity issuance for cash at any point in the five-year window. Weighted-average diluted shares fell from 12,089,628 in FY2024 to 11,904,631 in FY2025 as buybacks took effect, and shares outstanding net of treasury were 11,651,761 at 30 June 2026.

Remaining dilution is small and mostly out of the money. The exchangeable bond issued in November 2025 is exchangeable into 220,317 existing treasury shares at KRW 32,002, 48.5% above the current price, so it is currently anti-dilutive and in any case exchanges treasury stock rather than issuing new shares; restricted-stock awards total 81,385 units, also to be settled in treasury shares; and no parent-level share options remain outstanding. Because no defensible current-diluted denominator can be built from a filing-supported increment at a single date, the diluted sensitivity is recorded as unresolved and the point-in-time count leads throughout.

The company does not depend on equity markets to operate: it funds itself from operating cash flow and bank facilities, has KRW 60.8bn of undrawn Korea Development Bank facility commitments, and has never run a shelf or at-the-market programme. Dilution risk is judged low; the financing risk that matters is bank rollover, covered above.

10. Litigation, Regulatory, and Contingent Liability Risk

Three court matters are disclosed, and they are small in money terms. Huons is appellant and respondent in a Patent Court appeal over a compound patent, with a claim value of KRW 683.9m; the company lost in part at first instance and both sides have appealed. Huons and five other parties are defendants in a criminal proceeding under the Pharmaceutical Affairs Act, first instance ongoing, with no monetary claim stated. The group is plaintiff in an unjust-enrichment action for KRW 95.4m which it expects to win. Beyond litigation, the group has a contract-performance surety limit with Seoul Guarantee Insurance and joint-and-several subsidiary guarantees of KRW 25.7bn.

The material regulatory exposure is not in the litigation note at all. Following an FDA inspection of the Jecheon plant from 12 to 21 November 2025, the agency restricted entry of the company’s injectable products from April 2026, the company recalled US-bound product, and the FDA issued a warning letter dated 15 June 2026 covering current good manufacturing practice violations at the sterile injectable facility. The published findings are serious in kind, not only in degree: an endotoxin test failure that was neither documented nor investigated; bioburden plates showing significant growth discarded on a team leader’s instruction; camera timestamps manipulated to create backdated records; nearly 1,900 uncontrolled blank current-good-manufacturing-practice forms found in the microbiology laboratories; and logbook pages removed and replaced with newly fabricated pages. The remediation the agency requires includes an independent data-integrity investigation, a three-year retrospective review of out-of-specification results and annual third-party GMP audits. The company’s own disclosure of all this is a single sentence in a supply-contract progress table; the sanctions section states that there are no sanctions as of the report date. Recall of at least two of the three affected products was released in June 2026, but the import restriction and the warning letter were unresolved at the balance-sheet date, and management’s stated aim of re-inspection by end-2026 and resumption of exports in early 2027 is a management assertion, not a confirmed fact.

The falsification was directed and physical, not passive neglect. The microbiology team leader admitted to using a knife to cut completed pages out of the logbook, then had the pages replaced with newly written substitutes made to resemble the originals, and separately instructed an analyst to backdate and omit testing information tied to bioburden samples. Huons told FDA investigators it could produce only a portion of the requested laboratory data, citing data-integrity breaches under which failed results had already been discarded. The company has since dismissed multiple employees involved and formed a data-integrity team over the microbiology laboratory, but FDA’s own letter assessed that response as insufficient, citing continued weak oversight — a reason to treat management’s Q4 2026 re-inspection target as optimistic rather than a base case; comparable data-integrity warning letters elsewhere have taken multiple years to close. The inspection also cited aseptic-processing gaps independent of the paperwork fraud: inadequate airflow in smoke studies, operators obstructing first air in critical areas, weak barrier-system practice, and glove-monitoring and disinfection procedures the agency found insufficiently specified. Separately, the import restriction (effective 3 April 2026) covers all drugs and drug products offered for import from this facility, not only the three named injectables.

The single new balance-sheet item that has appeared is the KRW 5,277.6m sales-warranty provision, unnoted in the filing, which on the corroborating record is the recall reserve recognised in the first quarter of 2026. The financial exposures here are individually modest against a KRW 314.6bn tangible book; the exposure that matters is the loss of the North American injectable franchise and the reputational and requalification cost of a data-integrity finding, neither of which appears as a provision.

11. Accounting Quality and Disclosure Review

The auditor is Samil PwC, appointed for FY2023 to FY2025 under Korea’s periodic auditor-designation regime, with an unqualified opinion on both the consolidated and separate statements in each of FY2023, FY2024 and FY2025, no material uncertainty related to going concern and no emphasis-of-matter paragraph. Management assessed internal accounting control as effective with no material weakness, and the audit committee met the auditor six times including a session on fraud-risk controls over cash. The designation period expires with FY2025, so the FY2026 auditor may change.

Within the financial statements, three estimates carry real judgement. Goodwill of KRW 18,031.6m, 5.7% of tangible book, sits in four cash-generating units, most of it in Panagene, a loss-making listed subsidiary in which the group holds only 42.66%, and goodwill has already been impaired in each of the last two years. Deferred tax assets are KRW 5,475.9m, but the more telling number is what is not recognised: unused tax losses of KRW 69,121.1m, up from KRW 31,979.9m a year earlier, all carried at nil because realisation is judged highly unlikely. Segment disclosure is by three revenue lines only, with no segment profitability, so a reader cannot see whether the contract-manufacturing margin or the beauty-and-wellness margin is the one eroding.

Disclosure quality is classified Adequate, not High, for specific reasons. Against: the KRW 5,277,642,666 sales-warranty provision appears as a new current liability with no note in a 200-page half-year report; the FDA warning letter and import restriction, plainly the most material development of the period, are disclosed in one sentence inside a supply-contract progress table while the sanctions section reads not applicable; the fully provided related-party receivable is explained in a single clause; and related-party purchases of KRW 23.8bn are disclosed in amount but not in pricing basis. In favour: the audited statements are complete, internally consistent and cross-foot; the auditor is a major firm under mandatory designation with a clean opinion and a rising fee; the borrowing, pledge, contingency, related-party and subsequent-event notes are all present and specific. This is a company that discloses what the forms require, competently, and volunteers nothing.

12. Valuation and Margin of Safety

Tangible Book Value — Derivation (anchor 2026-06-30; KRW millions)

Table 1 — equity bridge

The tangible-book bridge is straightforward: common shareholders’ equity of KRW 368,364.0m less goodwill of KRW 18,031.6m and other intangibles of KRW 35,693.7m produces tangible book value of KRW 314,638.7m. Dividing by 11,651,761 shares gives KRW 27,003.53 per share. No unbooked property surplus is added to this bridge.

P/TBV (derived) = 21,550.00 ÷ 27,003.5290 = 0.798× (price as of 2026-09-01). No adjustments are added to the bridge: the KRW 3,800.9m unbooked investment-property surplus identified above is deliberately not capitalised here, since it rests on the group’s own index-based estimate rather than an independent valuation.

Table 2 — asset composition at the anchor period

At the 30 June 2026 anchor, cash and short-term investments were KRW 58,603.9m, receivables KRW 101,107.7m, inventory KRW 96,636.2m, owned property, plant and equipment KRW 261,891.9m, and right-of-use assets KRW 1,932.3m. Owned property and equipment equals 83.2% of tangible book; right-of-use assets equal only 0.6%. The asset floor is real, but much of the property is pledged.

Historical relative multiple valuation

A quarterly price-to-tangible-book series of 38 observations running from 31 December 2016 to 31 March 2026 is available. Read in order, the multiple has not oscillated around a stable mean, it has de-rated almost monotonically for eight years: 4.64× at end-2016, a peak of 6.04× in March 2018, 2.17× by end-2021, 1.25× by end-2022, 0.99× by end-2024 and 0.92× at 31 March 2026. Treating the whole decade as one distribution therefore describes two different companies. The typical band for the current regime, taken as the interquartile range of the thirteen quarters from March 2023, is 0.96× to 1.34×, median 1.06×; the window floor is 0.86× (March 2025), the window ceiling for the full ten years is 6.04× (March 2018). Today’s derived 0.798× sits below every one of the 38 observations, the 0th percentile of its own history.

The two legs disagree, and the disagreement is the analysis. The book leg says the shares are cheaper than at any point in the listed history of the company, against a tangible book that is 83% owned plant and 19% cash. The earnings leg says that on the earnings actually being produced today, the shares are not cheap at all, at 13.3× depressed trailing earnings they are near the market multiple, and only become cheap if profitability returns to something near the FY2023-FY2025 average. Reconciled, the range that both legs can support is roughly KRW 26,000 to KRW 29,000, the intrinsic-value range used below.

Sell range

Applying the same rule to the same 38-observation quarterly series: Sell Range KRW 61,565 – KRW 125,311 (the local-currency range), being 2.28× to 4.64× current TBV per share, the median and the top-quartile mean of the stock’s own distribution. Thirty-eight observations at four per year over 9.5 years is comfortably past the floors this method requires, so this is not a data-limited reading. The flat 0.80× reference of KRW 21,603 is fallback only, not the rule for this name.

Credibility of that range, given today’s earnings power. The range is mechanically correct and it should not be read as a forecast. Trailing-twelve-month return on equity is 5.1% against an FY2023-FY2025 average of 11.8%, 57% below the multi-year average that underlies the sell-range period, and far below the 20%-plus returns that supported the 4×-6× multiples of 2016-2018. The top of this range was earned by a business with a growing FDA-cleared export franchise and mid-teens returns on equity; it is not credible on current earnings and should not be treated as a target. The exit question realistically goes live at the bottom of the range, and in practice much earlier: a reasonable expectation is to reassess as the multiple approaches the 1.3× to 1.5× of tangible book last seen in 2022-2023, roughly KRW 35,000 to KRW 40,500, rather than to hold for a reversion to the 2018 regime.

Scenarios

The scenario lens is deliberately separate from the mechanical own-history range:

• Severe downside: the FDA re-inspection fails or slips beyond 2027, sterile-injectable assets are written down by 25% of tangible book, and value falls to KRW 14,177 per share.

• Bear case: US recovery slips past 2027, margin stays at the first-half 2026 level, and value is KRW 23,223 per share.

• Base case: the import restriction is lifted during 2027, margins rebuild, and value is KRW 28,624 per share.

• Bull case: North American shipments resume, contract manufacturing and pipeline revenue improve, and value is KRW 35,645 per share.

The Base Case cross-checks against the earnings leg: KRW 28,624 is 8.3× the pre-disruption average EPS, a multiple this stock last traded at in 2025, so the two methods agree at the base. The Bear Case sits above the current price, which is itself the finding: at KRW 21,550 the market is pricing in something worse than the bear case set out here.

Current price KRW 21,550 (2026-09-01; market capitalisation KRW 251.1bn, KRW 251.1bn). Intrinsic value range on the base case: KRW 26,000 – KRW 29,000. Margin of safety at the current price: 20.2% below tangible book value per share. Buy-Below KRW 28,424 (1.05× TBV, 9.5-year own-history), the stock is already below it. Buy More Below KRW 13,502 (0.50× TBV). The industry-appropriate cross-lens for a capital-heavy generic manufacturer is enterprise value to EBITDA: net financial debt plus market capitalisation gives an enterprise value of roughly KRW 350.3bn against FY2025 EBITDA of roughly KRW 72.8bn, 4.8×, and against annualised first-half 2026 EBITDA of roughly KRW 39.0bn, 9.0×, the same story the earnings leg tells.

Cheapness type: asset-value cheap, with an explicit value-trap watch. The discount is to tangible assets, not to a stream of earnings, and the assets are real, owned and largely productive. But 83% of those assets are pledged to lenders, three years of free cash flow have been negative, gross margin has fallen for three consecutive periods, and the highest-return use of the newest asset is currently prohibited by a foreign regulator. This is cheap because something is wrong, and the question a buyer is underwriting is whether what is wrong is fixable within two years.

13. Risk Matrix

The main risks are easier to read as a list:

• FDA warning letter and import restriction persist, keeping the North American franchise offline.

• Data-integrity findings prove systemic rather than site-specific, forcing a permanent valuation discount.

• Gross-margin erosion continues and undermines the base and bull cases.

• Controller-sponsored transactions and related-party dealing continue to disadvantage minorities.

• Bank rollover risk rises as KRW 111.4bn of borrowings including interest comes due within twelve months.

• Pledged property, goodwill, unrecognised tax losses, and a dividend funded from capital reserves weaken the asset-value support.

14. Red Flags, Yellow Flags, and Green Flags

The single most important flag below is the first red one: a documented instance of employees fabricating quality records is a different order of problem from a missed operating quarter, and it is the fact most likely to overturn the verdict on human review.

Green Flags

• Trades at 0.798× tangible book value, below all 38 quarterly observations in its own ten-year price-to-tangible-book history, with 83.2% of that tangible book in owned land, buildings and machinery and only 0.6% in right-of-use assets.

• Financial borrowings are 38.9% of total equity and 24.3% of total assets, with KRW 58.6bn of cash and current financial assets and KRW 60.8bn of undrawn committed facilities.

• Piotroski F-Score of 7/9 on FY2025 against FY2024, with all nine signals computable and the floor rule not fired.

• Cash returned: KRW 13.05bn of dividends in the trailing twelve months, a 5.2% yield, plus KRW 6.04bn of buybacks across FY2024-FY2025, under a formally adopted 2026-2028 policy.

• No equity issued for cash at any point in five years; shares outstanding have fallen since end-2023.

Yellow Flags

• Aggregate free cash flow across FY2023-FY2025 was negative KRW 3.6bn: KRW 153.3bn of capital spending absorbed three years of operating cash flow.

• FY2025 net income was flattered by roughly KRW 5.0bn from an unexplained drop in the effective tax rate, while unrecognised tax losses rose sharply, all carried at nil.

• KRW 218.5bn of property, plant and equipment, 83% of net book value, is pledged against KRW 102.9bn of bank borrowings.

• The FY2026 dividend target is 13% below the amount paid for FY2025, and the policy is explicitly funded from a capital-reserve transfer rather than from earnings.

• No covenant terms, ratios or headroom are disclosed for KRW 152.0bn of bank debt, most of which matures within twelve months.

• Segment disclosure gives revenue by three lines and no segment profitability, so the source of the margin erosion cannot be located from the filing.

Red Flags

• The FDA warning letter of 15 June 2026 records fabricated and manipulated quality records at the Jecheon plant: the microbiology team leader admitted to using a knife to cut completed pages out of a logbook and replace them with rewritten substitutes disguised to look original, and instructed an analyst to backdate and omit bioburden test results; an undocumented endotoxin failure; nearly 1,900 uncontrolled blank CGMP forms. The company fired the staff involved and formed a data-integrity team, but FDA judged that response insufficient.

• A KRW 5,277,642,666 sales-warranty provision, the largest new line on the 30 June 2026 balance sheet, appears with no explanatory note anywhere in the half-year report, and the sanctions section of the same report states that there are no sanctions.

• First-half 2026 operating income of KRW 2.28bn against KRW 25.90bn a year earlier, a net loss, and gross margin down to 43.1% from 51.8% in FY2023.

• A KRW 9,025m related-party trade receivable from an affiliate fully provided against, explained in one clause, with no account of how it accumulated.

• The board resolved a merger with a group affiliate in May 2026 and rescinded it in August 2026 after the share price roughly halved and shareholders objected to the exchange ratio.

• Goodwill impaired in each of the last two years, on acquisitions made within the same window.

Must-Watch Catalysts

• FDA re-inspection of the Jecheon plant, targeted by management for the fourth quarter of 2026, and the closure of the June 2026 warning letter.

• Resumption of North American injectable shipments, guided by management for early 2027.

• Third-quarter 2026 results, expected in November 2026, as the first clean read on whether the second-quarter operating profit was a turn or a pause.

• Whether the KRW 800 annual dividend target is maintained at the FY2026 year-end declaration.

• Any revival of group restructuring involving Huons Lab or other affiliates.

• First-instance judgment in the Pharmaceutical Affairs Act criminal case.

• Appointment of the FY2026 auditor, as the mandatory designation period covering FY2023-FY2025 has expired.

FINAL VALUE-INVESTING RECOMMENDATION

Verdict: BUY — SMALL. The numbers-only tier is Standard; final sizing is Small, given Governance Risk High and an unresolved foreign-regulator data-integrity finding.

Key decision metrics. Current price is KRW 21,550 as of 1 September 2026. P/TBV is 0.798x, normalized P/E is 13.3x, shareholder yield is 5.2%, and the Piotroski F-Score is 7/9. The numbers-only tier is Standard. Final sizing is BUY — SMALL because Governance Risk is High, the FDA data-integrity finding remains unresolved, and the earnings recovery is not yet demonstrated. The intrinsic-value range is KRW 26,000 to KRW 29,000; Buy-Below is KRW 28,424; Buy More Below is KRW 13,502; and the mechanical Sell Range is KRW 61,565 to KRW 125,311. The realistic exit question should be reassessed much earlier, around KRW 35,000 to KRW 40,500, if the operating recovery is credible.

VALUE INVESTOR MUST-WATCH LIST

• FDA re-inspection of the Jecheon plant, management targets the fourth quarter of 2026.

• The quarterly export line: North American injectable revenue was KRW 18.4bn in FY2025 and zero from April 2026.

• Gross margin, quarter by quarter: 43.1% in the first half of 2026 against 45.5% for FY2025 and 51.8% for FY2023.

• Third-quarter 2026 results, expected November 2026.

• The KRW 5,277,642,666 sales-warranty provision, whether the FY2026 annual report finally explains it.

• Bank rollovers: KRW 111.4bn of borrowings including interest matures within twelve months.

• The dividend, and whether the KRW 800 annual target survives the FY2026 year-end declaration.

• Any revival of group restructuring involving Huons Lab or another affiliate.

• Insider filings: no director or officer purchases or sales were found for 2026.

• The Pharmaceutical Affairs Act criminal case, first-instance judgment.

• Appointment of the FY2026 auditor.

• Panagene: 57.34% of it belongs to outside holders, and it carries a meaningful share of the group’s goodwill.

• KRW 35,000-40,500 (1.3×-1.5× tangible book), the level at which the sell question realistically goes live.

• KRW 13,502 (0.50× tangible book), the Buy More Below price.

SOURCES AND DILIGENCE GAP LOG

Successfully Accessed:

• Annual business report (사업보고서), 10th fiscal period | annual statutory filing | FY ended 2025-12-31 | filed 2026-03-23 | read in full in the original Korean

• Half-year report (반기보고서), 11th fiscal period | interim statutory filing | six months ended 2026-06-30 | filed 2026-08-14 | read in full in the original Korean

• US FDA Warning Letter to Huons Co., Ltd., MARCS-CMS 724650 | foreign regulator primary document | inspection 2025-11-12 to 2025-11-21 | 2026-06-15 | https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/warning-letters/huons-co-ltd-724650-06152026 — the page returned HTTP 404 on direct fetch; the violation text quoted in the regulatory discussion was partially accessed; the company’s filing and regulator materials were used for the underlying facts

• KOSDAQ market quotation, Huons 243070: price, market capitalisation, 52-week range | market data | as of 2026-09-01 | https://www.investing.com/equities/huons-co-ltd | SECONDARY — cross-checked against a second quotation source showing the same KRW 21,550 and consistent market capitalisation

• Huons board rescinds the Huons Lab merger agreement | financial press | May-August 2026 | 2026-08-26 | https://www.mt.co.kr/thebio/2026/08/26/2026082617442755604 | SECONDARY — corroborated against the half-year report’s own disclosure-progress table

• Q2 2026 results and FDA import-restriction impact analysis | financial press | Q1-Q2 2026 | 2026-08-11 | https://www.businesspost.co.kr/BP?command=article_view&num=444588 | SECONDARY

• Q1 2026 recall provision and North American injectable export halt | financial press | Q1 2026 | 2026-08-07 | https://en.sedaily.com/finance/2026/08/07/huons-injectable-business-hit-by-fda-curbs-as-q1-swings-to | SECONDARY — corroborates the recall-related sales-warranty line in the half-year balance sheet

• FDA recall termination for North American injectables | pharmaceutical trade press | June 2026 | 2026-06-04 | http://www.mdtoday.co.kr/news/articleViewRedirect.html?oldNo=1065595120555354 | SECONDARY

• CGMP-enforcement analysis of the Huons warning letter | third-party regulatory commentary | 2026 | 2026-07-01 | https://www.nexgenhealthgroup.com/post/what-huons-fda-warning-letter-reveals-about-today-s-cgmp-enforcement | SECONDARY — used only where it repeats the letter’s own language

• FY2022 consolidated revenue, operating profit and net income | pharmaceutical trade press | FY2022 | 2023-02-01 | http://www.hitnews.co.kr/news/articleView.html?idxno=52286 | SECONDARY

• Company profile and business description | company website | current | 2026-09-01 | https://www.huons.com | SECONDARY — used only for corporate identity

• Quarterly price-to-tangible-book history, 38 observations, 2016-12-31 to 2026-03-31 | third-party ratio history | 2016-2026 | SECONDARY — the most recent observation was cross-checked against an independently derived figure and agrees

Diligence Gaps — Data Not Found or Estimated:

• Going Concern — checked in the annual business report for FY2025: no such note or language present (a finding, not a gap). The auditor recorded no material uncertainty related to going concern in any of FY2023, FY2024 or FY2025.

• Debt covenants [UNAVAILABLE] — neither filing discloses financial covenants, ratio tests or headroom for KRW 152.0bn of bank borrowings, KRW 111.4bn of which matures within twelve months. This matters because it is the one leverage risk that could crystallise faster than the operating recovery.

• Insider transactions for 2026 [UNAVAILABLE] — no director or officer purchase or sale filings were located in the Korean insider-reporting regime for calendar 2026. Absence in a search is not proof of absence.

• Segment profitability [UNAVAILABLE] — revenue is disclosed for three business lines but no segment operating profit is given, so the gross-margin erosion since FY2023 cannot be attributed to a specific line from the filing.

• The KRW 5,277,642,666 sales-warranty provision [INFERRED] — the largest new balance-sheet item at 30 June 2026 carries no note in the half-year report. It is identified here as the US recall reserve on the strength of a secondary report of a similar-size Q1 2026 recall-related charge and the filing’s own disclosure of an April 2026 recall; the identification is a reasonable inference, not a confirmed fact.

• FY2022 comparatives [ESTIMATED] — taken from a secondary report of the FY2022 results release and not independently verified against the FY2022 statutory filing. Used only for trend context and for the four-year average EPS; the normalized EPS actually used is the trailing-twelve-month figure, which does not depend on them.

• Effective-tax-rate benefit in FY2025 [INFERRED] — the tax note discloses the deferred-tax movements but does not explain why the effective rate fell. The roughly KRW 5.0bn effect is quantified from the two disclosed rates; the cause is not established.

• FDA warning letter full text [INFERRED from partial access] — the letter itself returned HTTP 404 on direct fetch. The specific findings quoted are drawn from search-index extraction of the FDA page and from a third-party regulatory analysis quoting the letter. The existence, date, inspection window and subject matter are corroborated by the company’s own filing; the itemised findings rest on those two sources rather than on a full read of the letter.

• Peer multiples [UNAVAILABLE] — no reliable current price-to-book or price-to-earnings comparison set for Korean mid-cap pharmaceutical manufacturers was sourced, so the valuation rests on the stock’s own history rather than on relative peer positioning.

Overall Data Quality Rating: EXCELLENT — both primary filings were fetched and read in full in the original Korean, and the litigation, debt, related-party, subsequent-events, lease, segment and auditor notes were all located in them. The gaps above are real but none of them touches the tangible-book bridge, the balance sheet, the cash-flow statement or the F-Score inputs, all of which come directly from audited primary statements.

1. Scorecard

2. Argument

The short version

Hsing Ta Cement runs one cement plant in northeastern Taiwan and a second in Nanjing, China, plus a portfolio of Taiwanese land it has held for decades. At NT$13.50 the stock trades at 0.56 times tangible book value, and the company carries no bank debt of any kind. Cash and marketable securities alone are worth roughly what the stock costs, before counting two operating cement plants, a limestone quarry, a ready-mix subsidiary, and land the company says is worth roughly NT$5.37 billion more than it is carried on the books.

The reason it is this cheap is earnings, not the balance sheet. Revenue has fallen 41.6% over six years as Taiwanese demand has weakened and cheaper imported cement has taken share, and the Chinese plant is selling into a market that is still shrinking. The stock pays a 6.67% dividend while a buyer waits for the discount to close, but the family that controls the company has not bought back a single share at this valuation despite having the cash to do so.

Why it qualifies

• Trades at 0.56x tangible book value, the cheapest point in a real ten-year quarterly trading history

• No bank debt of any kind since February 2020; net financial assets alone are close to the full share price

• Piotroski F-Score of 5 out of 9, with every solvency-related signal passing

• 6.67% dividend yield, covered more than once over by free cash flow

• Disclosed investment property carried far below its stated fair value, an additional cushion beyond tangible book

Why it is not a clean bargain

• Revenue down 41.6% over six years and still falling, down 17.0% for the first seven months of this year

• Gross margin has compressed from 34.4% to 9.1% over the same stretch, with an operating loss in the first quarter

• The founding family controls roughly half the shares, and the company funds a stake in an affiliated construction firm that in turn holds a board seat

• The dividend has been cut twice in the past four years

• No share buybacks despite the deep discount to book value, the clearest sign management is not planning to close the gap itself

• The Chinese plant sells into a national cement market that is still contracting

What would change the view

• Any board authorization of a share buyback

• Monthly revenue returning to year-over-year growth (it was down 20.6% in July)

• The outcome of a pending land exchange with the Taiwan government involving company-owned property

• Third-quarter results, due around mid-November, and whether the second-quarter margin recovery holds

• A ruling on the company's application for a reduced carbon-fee rate, worth roughly NT$150 million a year if granted

Sources and gaps

The half-year filing's footnotes on legal claims, commitments, and related-party dealings were not available at the time of this review; the most recent confirmation of those items comes from the prior quarter's filing. The disclosed value of the company's investment property is management's own estimate, not an audited figure.

3. Backup and sources

This last layer is the full working file: every source, calculation, and gap, kept so the argument above can be checked. The scorecard and the argument are the synthesis. Open this only if you want to check the work.

Hsing Ta Cement Co., Ltd.: Cheap against assets, weak at the operating line

1. Executive Summary

Hsing Ta Cement quarries limestone and burns clinker at an integrated plant at Su-ao in north-eastern Taiwan and at a second kiln complex at Nanjing in mainland China, sells Portland and blended cements, clinker and ready-mixed concrete into the two construction markets those plants can reach by truck and coastal barge, and collects rent and waste-processing fees on a portfolio of Taiwanese land and buildings alongside the kilns. Cement and clinker were 62.71% of FY2025 revenue and ready-mixed concrete a further 12.52%. The verdict is BUY at STANDARD sizing, and it is an asset verdict, not an earnings one: the size is held down by what the company is earning, not by what it owns. At NT$13.50 the shares change hands at 0.560x a tangible book value of NT$24.1125 per share, derived below from the 30 June 2026 balance sheet, against a business that carries no bank debt of any kind, holds NT$4.93 billion of cash and marketable financial assets against NT$1.32 billion of total liabilities, and discloses an investment-property portfolio worth roughly NT$5.37 billion more than the amount it is carried at.

The earnings side of the case is genuinely poor, and it is why the discount exists. Revenue has fallen from NT$7.59 billion in FY2020 to NT$4.43 billion in FY2025 and is running a further 17.0% below prior year for the seven months to July 2026. Gross margin has compressed from 34.4% to 16.0% across the same six years and to 9.1% in the first half of FY2026. The group reported an operating loss in the first quarter of 2026 and earned NT$0.15 of basic EPS in the first half against NT$0.49 a year earlier. Taiwanese housing-credit tightening has taken demand out of the domestic market at the same time as low-priced imported cement has taken share, and the Nanjing plant sells into a Chinese cement market whose national consumption is still contracting. This is a business earning well below its own history, not one temporarily off a peak.

What makes it investable anyway is the size of the asset cushion relative to the price. Net financial assets, cash, short-term investments and non-current financial assets less lease liabilities, after deducting the NT$307.0 million dividend paid on 16 July 2026, are roughly NT$13.37 per share against a NT$13.50 share price. A buyer pays about the value of the securities portfolio and receives two cement plants, a limestone reserve, the ready-mix subsidiary and the property portfolio for close to nothing. The Piotroski F-Score is 5 out of 9, failing four momentum-type signals and passing every solvency signal. The dividend of NT$0.90 per share is a 6.67% yield covered 1.14 times by FY2025 earnings and 1.72 times by FY2025 free cash flow. Right-of-use assets are only 2.1% of tangible book, so cheap against tangible book is a fair description here, not a lease artefact dressed up as one.

The risks that would break this are not financial-distress risks; there is no refinancing risk to speak of. They are governance and terminal-value risks. Two capital-allocation warnings fire, which sets Governance Risk: High: the Yang family controls roughly half the register while the company's own balance sheet funds a 19.90% stake in an entity that in turn wholly owns a board member, and the dividend has been cut twice in four years. Accounting quality is adequate. The auditor is PwC Taiwan, and the interim review conclusion carries a scope qualification covering subsidiaries that are 7.34% of consolidated assets.

2. Business and Market Overview

The group is a two-country cement producer with a property tail. The Taiwan business runs the Nansenghu plant at Su-ao, Yilan County, feeding it from limestone workings the company holds itself, and sold 583,000 tonnes of cement and clinker in FY2025 for NT$1,657,115 thousand, 5.9% less than in FY2024 on volumes 56,000 tonnes lower. Total Taiwanese cement consumption was 13.5 million tonnes in FY2025. The China business is Jiangsu Xinning New Building Materials at Nanjing, held through the 66.67%-owned Soaring Power Corporation, which sold 1,533,081 tonnes in FY2025 into a Nanjing regional market of roughly 8.5 million tonnes a year, where the company estimates its own share at about 18%. Taiwanese revenue fell 14.99% in FY2025 while the China plant grew 8.96%.

Alongside the kilns sit two smaller, structurally different revenue streams: resource-recycling and waste-processing income of NT$81,784 thousand in FY2025 and rental income of NT$48,599 thousand. In the first quarter of FY2026 the cement segment posted an external-revenue loss while the "other" segment, mostly recycling and rent, was the only part of the group that turned a profit that quarter.

Customer and supplier concentration is not disclosed in a form that supports a top-three figure. What is disclosed is a long list of single-year Chinese supply contracts for coal, slag powder, sandstone, gypsum, fly ash and refractories, none carrying restrictive covenants. Coal is the swing input, and management attributes the modest FY2025 profit recovery across the Chinese cement industry to lower first-half coal prices. Capital intensity is high in the accounting sense, NT$2.58 billion of owned property, plant and equipment, but low in the cash sense: capex has run between NT$128 million and NT$253 million a year for six years against depreciation and amortisation of NT$314 million in FY2025, so the plants are being maintained rather than expanded. As of 17 August 2026: July revenue was NT$292,047 thousand, down 20.6% year on year, with seven-month cumulative revenue down 17.0%. The business is understandable and cyclical. It is not predictable at the earnings line, and on the Chinese side it looks less like a cycle than a melting ice cube.

3. Moat, Competitive Position, and Industry Cycle

The traditional cement moat is efficient scale plus the freight cost of a low-value, high-weight product, and Hsing Ta has the raw materials of one: a permitted limestone reserve and an integrated kiln in a jurisdiction where nobody is going to permit a new cement plant, a second kiln with roughly 18% of the Nanjing regional market, and a downstream ready-mix subsidiary that captures part of its own output. None of the other four classic moat sources is present in any degree: no brand pricing power in bulk cement, no switching costs between two truckloads of Type I cement, no network effect, and no cost advantage. The company's stated strategy is to price at the market median, neither highest nor lowest, which is the language of a price-taker.

The freight moat is being arbitraged away by sea. Management says plainly that low-priced imported cement continues to erode the domestic market, that the industry association obtained anti-dumping duties on Vietnamese cement in July 2025 at rates up to 23.2%, and that it is now preparing an application against Indonesia. Reporting outside the filings indicates the Vietnamese duty has been substantially absorbed by falling Vietnamese prices, and that Taiwanese cement imports rose 13.77% in 2025 even as domestic production rose only 1.80%. The moat is best classified as Narrow and declining: real, but shrinking, and not enough to defend margin.

The margin history of the group is the evidence for the cycle. Gross margin ran 34.4% in FY2020, then 29.9%, 19.6%, 21.4%, 21.3% and 16.0% in FY2025, then 9.1% in the first half of FY2026. Volumes are falling in both markets, pricing is falling in China, and Taiwanese demand is being suppressed by deliberate credit policy. The right classification for the group is below mid-cycle, and for the Chinese operation specifically structurally impaired: national cement consumption there is forecast at about 1.85 billion tonnes for 2026, down a further 2.1%, and management's own case for the Nanjing plant is survival through compliance rather than growth. This company cannot compound capital at attractive rates on its operating assets; FY2025 return on equity of 4.6% against a six-year average of 10.4% says so. The investment case has to rest on assets and on the price paid, not on compounding.

4. Management, Governance, and Capital Allocation

The chairman is Yang Chih-hsiung, first elected a director in May 1991 and re-elected in June 2024 for a three-year term. The president is Yang Ta-kuan, in post since mid-2019. The board seated in June 2024 has nine members, three of them independent, the other six being five Yang family members and De Bo Investment Co., represented by Yang Po-wei. All three independent directors attended every board meeting in FY2025, and an audit committee has voted on every set of financial statements since 2022.

Insider activity was mildly positive. In FY2025 the president increased his holding by 110,000 shares to 9,025,431 (2.65% of the company); the chairman added 30,000 shares in early 2026; no other director or officer bought or sold. A 6,500,000-share pledge against one director was released during FY2025, and there is no outstanding share pledge as of the report date, a mild positive. There is no share-based compensation at all: no employee share bonus, no options, no restricted shares, nothing dilutive outstanding.

Related-party dealing is small. Two related entities are named in the most recent quarterly note, together accounting for a few million New Taiwan dollars of freight and leasing services a quarter, all on stated normal commercial terms. The only endorsements or guarantees outstanding run inside the group, between the parent and its ready-mix subsidiary, both within board-set limits.

Capital allocation over six years has been conservative to a fault: no acquisitions, no share issuance, no borrowings, capex held to maintenance levels, and NT$2,391 million returned in dividends between FY2020 and FY2025. There have been no share repurchases in FY2024 or FY2025, and that is the real criticism here. A company trading at 0.56x tangible book with NT$4.9 billion of financial assets and no debt has not bought back a single share, while it has continued to fund a 19.90% stake in an affiliated construction company. Dividends per share ran NT$1.10, NT$1.50, NT$1.50, NT$0.80, NT$1.20 and NT$0.90 for the years paid FY2020 through FY2025.

Two capital-allocation warnings fire here, and both are worth stating plainly rather than just counting. The first is control without minority protection: seven of the top ten holders are declared related within the second degree of kinship and hold 46.57% of the company; including two further family members the bloc is 51.62%. The board itself was elected by genuine shareholder vote on a real three-year cycle, with two new independent directors added in 2024, which cuts the other way. But the company itself holds 19.90% of Chin Ta Construction, which in turn wholly owns De Bo Investment, the corporate entity that occupies a board seat: shareholder capital is funding a member of the controlling bloc's own board representation, which is the opposite of minority protection. The second is a dividend cut within the past several years: the payout fell from NT$1.50 to NT$0.80 for FY2022 and again from NT$1.20 to NT$0.90 for FY2024. Two more possible flags do not fire. Related-party purchases run about NT$14 million a year against NT$4.4 billion of revenue, immaterial and stated at market terms, and director pay sits inside the company's own articles-based cap. A fifth, whether management has ever ducked a shareholder question, cannot be assessed: nine investor conferences have been held since 2017 but no transcript was reachable. Two warnings fired is what sets Governance Risk: High.

5. Corporate Ownership, Subsidiaries, and Joint Ventures

The register is concentrated in one extended family. The ten largest shareholders disclosed as of April 2026 were Yang Jen-hsiung (10.58%), Yang Chung-hsiung (10.27%), chairman Yang Chih-hsiung (10.25%), Hu Mei-hung (6.06%), Hyde Bo Capital (4.51%), Kao Yang Hsueh-ching (3.51%), Yang Chen Shu-o (3.10%), Lin Hsueh-hua (2.79%), president Yang Ta-kuan (2.65%) and Yang Ta-ching (2.40%). Together they held 56.13% of the company.

The register is therefore controlled in practice, not merely concentrated on paper.

Seven of those ten are declared related within the second degree of kinship, holding 46.57% between them. There is a single class of ordinary shares, 341,158,868 outstanding, no treasury shares, one vote each; free float is therefore about 44%. Trading volume in the first two September 2026 sessions was 52,000 and 76,000 shares, a genuinely illiquid microcap by value traded.

The group's structure and cross-holdings, as of March 2026: Hsin Yi Ready-Mixed Concrete (55.20% direct, 93.67% combined with insiders) is the Taiwan ready-mix arm and is consolidated. Hsin Ni Development (98.00%) is a Taiwan development vehicle, also consolidated. Soaring Power Corporation (66.67%) is the offshore holding company for the China operation. Chin Ta Construction (19.90% direct, 23.57% combined) is an associate that wholly owns De Bo Investment, which holds a board seat. Beneath Soaring Power sit the Nanjing kiln itself, Nanjing Xinrong and Xinning Trading. Non-controlling interests carry NT$1,934,735 thousand of the group's NT$10,178,824 thousand of equity at 30 June 2026, the large majority of that being the 33.33% of the China operation the company does not own, which matters because consolidated cash is not all attributable to Taipei shareholders. As of 22 April 2026: the subsidiary Nanjing Xinrong obtained its tax-clearance certificate, the standard precursor to deregistering a Chinese entity.

6. Historical Financial Quality and Normalized Owner Earnings

The six-year consolidated record is shown in the financial-quality exhibit below. It keeps the useful comparisons together without making the reader fight a wide text table.

The shape is unambiguous: revenue down 41.6% in six years, gross margin down 18.4 points and net income down 74%, while tangible book per share still rose 12.5% because the group retains more than it pays and has no debt to erode. The current year is worse than any of these. First-half FY2026 revenue was NT$1,857,847 thousand against NT$2,251,696 thousand, gross margin 9.1% against a prior-year 16.2%, and profit attributable to owners NT$50,378 thousand against NT$167,121 thousand, basic EPS of NT$0.15 against NT$0.49. The first quarter of 2026 was the trough, an operating loss on a 6.2% gross margin; the second quarter recovered to roughly an 11.6% margin. July revenue, down 20.6% year on year, says the recovery is not yet visible in the top line.

FY2025 owner earnings, net income plus depreciation and amortisation less maintenance capex, were NT$451,253 thousand, and because capex has run below depreciation in five of the last six years, essentially all of it is maintenance rather than growth spending. Free cash flow was NT$528,314 thousand in FY2025, comfortably covering the dividend; on a trailing twelve-month basis it is NT$282,054 thousand, still covering the dividend but only once the roughly NT$245 million spent on investment property in the first half is set against it, at which point the group is running a modest cash deficit before dividends.

The trajectory here is a post-peak decline: FY2020 was the earnings peak and every year since is lower, with the trailing figure the lowest of all. The normalized EPS this report uses is therefore the trailing twelve-month figure, NT$0.687 per share, rather than a multi-year average that would flatter the stock. Cumulative free cash flow over the six years was NT$5,494 million against cumulative net income of NT$4,894 million, a 112% conversion, but working capital has turned into a headwind: inventories rose in the first quarter of 2026 even as volumes fell.

6.5 Piotroski F-Score: 5 out of 9

Comparing FY2025 to FY2024, the score passes on positive net income, positive operating cash flow, earnings quality (operating cash flow exceeds net income by a wide margin), unchanged leverage (zero debt in both years) and no dilutive share issuance. It fails on return on assets (3.4% against 4.5%), the current ratio (5.80 against 7.00, still extraordinary in absolute terms but lower than the year before), gross margin (16.0% against 21.3%) and asset turnover (0.393 against 0.403).

All four failures are the same failure seen four ways: the business is shrinking. None of the four is a solvency signal. Every signal that speaks to solvency and to shareholder treatment passes cleanly, and the accruals signal passes by a wide margin, with operating cash flow at roughly twice net income. A score of 5 out of 9 sits squarely in the eligible band and, on its own, neither supports nor obstructs the verdict.

7. Balance Sheet, Debt, Covenants, and Refinancing Risk

There is nothing to refinance. The consolidated balance sheet at 30 June 2026 shows no short-term borrowings, no long-term borrowings, no bonds and no commercial paper, and no such line has appeared in any period since FY2020; the annual report states directly that the company's US-dollar bank loans were repaid in full in February 2020 and that it currently has no bank loans of any kind. The only interest-bearing obligations are lease liabilities of NT$57,027 thousand, 0.7% of the NT$8,244,089 thousand of equity attributable to owners. On the most punitive reading, every liability of every kind against total equity, the ratio is 12.96%. Total liabilities are 11.5% of total assets. First-quarter finance costs were NT$420 thousand, entirely lease unwind, against NT$9,479 thousand of interest income, so the group is a net receiver of interest. The classification is Net Cash, and it is not a marginal call.

The maturity wall consists of lease payments and nothing else, and there are no covenants because there is no credit agreement. As of 16 July 2026: the FY2025 dividend of NT$307,043 thousand was paid, having been accrued at 30 June and already deducted from the equity and tangible book value used throughout this report, but since left the cash balance. Adjusting for it, net financial assets less lease liabilities are approximately NT$4,562,655 thousand, or NT$13.37 per share against a NT$13.50 share price. Two caveats belong with that figure: about a third of the Chinese subsidiary's net assets belong to non-controlling interests, so not all consolidated cash is attributable to Taipei shareholders, and a material part of the cash sits in mainland China, where distribution to the parent requires the usual approvals and withholding.

8. Real Estate, Leases, and Hidden Assets

Right-of-use assets are NT$170,515 thousand at 30 June 2026, 2.1% of tangible book value, against lease liabilities of NT$57,027 thousand. The gap between the two suggests most of the balance is long-prepaid land-use rights at the Chinese plant rather than ordinary operating leases. The real downside support is elsewhere: NT$4,926,725 thousand of cash and financial assets, NT$897,564 thousand of inventory, and NT$2,583,577 thousand of owned plant.

The hidden asset is the property, and it is large. The company's investment-property note discloses a fair value of NT$7,029,841 thousand at 31 March 2026 against a carrying amount of NT$1,654,986 thousand, an unrecognised surplus of NT$5,374,855 thousand, or NT$15.75 per share before any tax, on a stock trading at NT$13.50. That is management's own estimate rather than an audited figure, derived from independent appraisal or internal assessment. Two constraints belong with it. Taiwanese land disposals attract land value increment tax, so the realisable surplus is materially below the gross figure; at a 40% combined tax and cost haircut the uplift is roughly NT$9.45 per share. And the yield on the portfolio is poor while it is held, about 1.9% on carrying value and roughly 0.4% on the disclosed fair value, which is what land held for development rather than income looks like.

Three parcels drive it. The Guanxi land, carried at NT$673,413 thousand, is the subject of a decades-old community development plan; some of the farmland portions are still registered to natural persons under trust agreements pending reclassification, and the company holds mortgages over those parcels to protect its position, a real, disclosed title risk. As of 28 October 2025: a state land exchange for this parcel was approved by the National Property Administration's valuation committee; replacement land was bought and registered by February 2026, and the exchange is now with the Ministry of Finance for final review. Separately, under a 2020 joint-construction contract, the company is contributing land in Taipei's Zhongzheng District to a project where a construction partner funds the build and the two share the completed units; the building permit was obtained in 2024 and construction started in January 2025. None of this is in earnings, and none of it is in tangible book above cost.

9. Capital Markets Access, Dilution, and Financing Flexibility

The share count has gone one way, down: from 359,955,650 in 2017 through two capital reductions and a treasury-share cancellation to the present 341,158,868, unchanged since 2023. There has been no equity issuance of any kind, and nothing outstanding, options, convertibles or otherwise, could dilute the current count.

The company does not rely on capital markets to fund itself and has not since its bank loans were repaid in 2020: operations generated NT$5,494 million of cumulative free cash flow over six years, and it holds NT$4.93 billion of cash and financial assets as its own liquidity backstop. Dilution risk is therefore assessed as none. The mirror image of that strength is the missing action: with no debt, surplus liquidity and the stock at 0.56 times tangible book, the repurchase authority has been left unused, a capital-allocation criticism rather than a financing risk.

10. Litigation, Regulatory, and Contingent Liability Risk

The company's own disclosure records no contingencies, no casualty losses, and no material litigation or administrative proceedings involving the company, its directors or any subsidiary. What is disclosed instead is a set of small regulatory penalties and one large regulatory swing factor. In FY2025 the Taiwan labour authority fined the company three times under the Occupational Safety and Health Act, for an unguarded welding terminal, a dropped refractory brick that injured a worker, and a worker struck by an overhead crane during unsupervised maintenance. The fines total NT$410,000 and are immaterial in money, but three penalties in four months, two involving injuries, are a real signal about site discipline at an ageing plant.

The large item is carbon pricing. The company is in the first cohort of Taiwanese carbon-fee payers and estimates that at the general rate the annual charge would be about NT$170 million, roughly a third of FY2025 pre-tax income. A voluntary reduction plan was approved in December 2025, qualifying it for a preferential rate, and the company applied in January 2026 for high-carbon-leakage-risk status, which management says would cut the effective cost to about NT$20 per tonne. The gap between the general rate and the qualified rate is on the order of NT$150 million a year, and qualification is annual and conditional on hitting production targets for low-carbon cement. That is the single largest contingent item in this analysis. One gap should be stated plainly: the half-year contingencies and subsequent-events notes at 30 June 2026 were not reached in this review, so the most recent note-level confirmation of "no contingencies" is from 31 March 2026.

11. Accounting Quality and Disclosure Review

The auditor is PricewaterhouseCoopers Taiwan. Partner rotation took effect in late 2025, for mandatory-rotation reasons combined with an internal reorganisation at the firm; the company records no disagreement with its auditor and no audit opinion other than unqualified in the last two years. Non-audit fees are a small fraction of audit fees and almost entirely compliance work.

One qualification does need stating. The review conclusion on the Q1 FY2026 consolidated statements is qualified: the statements of non-significant subsidiaries in the consolidation, together about 7% of consolidated assets, were not reviewed. This is a common and comparatively benign Taiwanese interim scope limitation, and the annual statements are audited without qualification, but it means roughly 7% of the asset base behind the tangible-book bridge is unreviewed at the interim dates.

The accounting itself is plain. There is no goodwill, and other intangibles are 0.2% of tangible book, so intangible impairment cannot materially damage book value. Related-party disclosure is specific and complete. Two weaknesses temper the assessment: the group reports as a single segment even though the Taiwanese and Chinese cement operations have materially different economics, and the largest single estimate in the accounts, the investment-property fair value, is management's own Level 2/3 assessment supporting assets carried at cost, and it is the number this thesis most depends on. Disclosure quality is classified as Adequate.

12. Valuation and Margin of Safety

Tangible book value, anchored to the 30 June 2026 balance sheet: common shareholders' equity of NT$8,244,089 thousand, less no goodwill and less NT$17,902 thousand of other intangibles, gives tangible book of NT$8,226,187 thousand. Divided by the 341,158,868 shares outstanding, that is NT$24.1125 per share. At NT$13.50, P/TBV is 0.560x. The equity figure is already net of the NT$307,043 thousand FY2025 dividend paid in July, so tangible book per share here is a post-dividend figure.

The book-value leg of the valuation rests on a ten-year, 40-observation quarterly price-to-tangible-book history for the shares. Today's 0.560x sits below the entire ten-year series, at the 0th percentile, cheaper than any of the forty observations and 9.7% below the lowest of them. That is the central quantitative fact of this report. No sourced multi-year earnings multiple was available for this issuer, so the earnings leg of valuation is carried separately in the scenario table below rather than blended into one number; the trailing normalized P/E of 19.65x is a reminder that the cheapness here is entirely a balance-sheet phenomenon, not an earnings one.

Sell Range: NT$19.42 to NT$24.33 (0.81x to 1.01x TBV), built from the median and top-quartile mean of the stock's own ten-year P/TBV history applied to current tangible book. That band was earned across a decade in which return on equity averaged 10.4%; trailing return on equity today is 2.84%, roughly 73% below that average. Absent either an earnings recovery or a monetisation of the property surplus, the realistic exit zone is the lower half of the range, roughly NT$19.4 to NT$21.5, and only a completed Guanxi land exchange or delivered Taipei units would make the top of the range credible.

Buy-Below: NT$16.56 (0.69x TBV), the mean of every observation at or below the stock's own 25th-percentile P/TBV over the same ten-year window. The stock is currently 18.5% below it.

Asset-heavy cement is normally also checked on enterprise value: taking market capitalisation plus minority interests at book plus lease liabilities less cash and financial assets, the group trades at roughly 3.5x trailing EV/EBITDA even after crediting minorities in full at book. Excluding minorities the enterprise value is close to nil, the same fact from a different angle.

The scenario lens is deliberately separate from the mechanical own-history range:

• Severe downside: Taiwan volumes fall a further 20%, the Nanjing kiln runs at a cash loss and is impaired; the market pays for the securities portfolio and nothing else, or NT$9.65 per share.

• Bear case: trailing earnings persist, the dividend is cut to NT$0.60 and no property event occurs, or NT$12.06 per share.

• Base case: earnings stabilise around NT$1.00 to NT$1.20 of EPS and the NT$0.90 dividend is held, or NT$19.42 per share.

• Bull case: volumes recover and the valuation returns to its own top quartile, or NT$24.33 per share.

• Asset realisation lens: the Guanxi exchange completes, Taipei units are delivered and the property surplus is realised after a 40% land-tax and cost charge, or NT$33.57 per share.

Margin of safety at NT$13.50: 44.0% below tangible book value, and 30.5% below the bottom of the intrinsic value range. This is asset-value cheap, with a live value-trap risk attached. It is not a compounder at a discount, and it is not a clean cyclical-recovery case either, because the Chinese half of the asset base faces structural rather than cyclical decline. It is a balance sheet, securities, land and two paid-for kilns, available at a 44% discount to its own tangible book, where the operating business is currently earning almost nothing and the catalyst for revaluation is neither dated nor promised.

13. Risk Matrix

The main risks are easier to read as a list than as a wide matrix:

• Chinese cement: national consumption is forecast down a further 2.1% in 2026, with prior intangible impairment and mandated emissions spending.

• Taiwan imports: domestic revenue fell 14.99% in FY2025 while imports rose 13.77% and production rose 1.80%.

• Value trap: the shares are at the bottom of their ten-year P/TBV history, but management has not repurchased stock.

• Governance: the family bloc controls 46.57% to 51.62%, reinforced by the company's 19.90% stake in an affiliate whose subsidiary holds a board seat.

• Carbon fees, Guanxi title and exchange execution, and the management-estimated property fair value remain live swing factors.

14. Red Flags, Yellow Flags, and Green Flags

Green Flags

• Zero financial borrowings since February 2020; total liabilities are 11.5% of total assets.

• NT$4,926,725 thousand of cash and financial assets against a NT$4,605.6 million market capitalisation, about NT$13.37 per share net of leases and the dividend paid in July 2026.

• Disclosed investment-property fair value up from NT$6,370,199 thousand a year earlier to NT$7,029,841 thousand.

• Share count down from 359,955,650 in 2017 to 341,158,868 today, with nothing outstanding that can dilute.

• A dividend in every year examined, NT$2,391 million cumulative FY2020 to FY2025, yielding 6.67% and covered 1.72 times by FY2025 free cash flow.

Yellow Flags

• Gross margin down from 34.4% in FY2020 to 16.0% in FY2025 and 9.1% in the first half of FY2026, with an operating loss in the first quarter.

• July 2026 revenue down 20.6% year on year and seven-month revenue down 17.0%; the top line has not turned.

• Trailing return on equity of 2.84% against a six-year average of 10.4%, why the top of the sell range is not currently credible.

• Three occupational-safety penalties inside four months of FY2025, two after injuries.

• A qualified interim review conclusion, and one reportable segment for two economically different geographies.

• Inventories up even as volumes fell, and receivables at around 90 days of sales.

Red Flags

• Family control of 46.57% to 51.62% reinforced by a company-funded 19.90% reciprocal stake in an affiliate whose subsidiary holds a board seat.

• The dividend was cut from NT$1.20 to NT$0.90 for FY2024, after NT$1.50 to NT$0.80 for FY2022.

• No share repurchase in FY2024 or FY2025 while the stock traded at an all-time-low valuation on its own ten-year record, with no debt and NT$4.93 billion of financial assets.

• The Chinese operation, roughly a third of the asset base, faces structural rather than cyclical decline and has already required an intangible impairment and heavy mandated compliance capex.

The Case in Brief

At NT$13.50, this stock trades at 0.560x tangible book value, the cheapest point in a real ten-year quarterly history, against a company that carries no bank debt of any kind and holds cash and financial assets close in value to the entire market capitalisation. The Piotroski F-Score is 5 out of 9, with every solvency signal passing. The dividend yields 6.67%. None of that is in dispute; the discount is real.

What is also real is that the company is earning very little right now, revenue down 41.6% over six years and still falling, and that the family that controls just over half the shares has not bought back a single share of stock at this valuation, while continuing to fund a stake in an affiliated construction company that reinforces its own board seat. That combination, real cheapness alongside a controlling family showing no urgency to close the gap, is why this is sized Standard rather than Big, and why the position calls for patience rather than conviction that a catalyst is coming. The dividend pays while a buyer waits. Nothing currently forces the re-rating the upper half of the valuation range depends on.

Value Investor Watch List

• Any board authorization of a share buyback, the clearest test of whether the controlling family will act on the discount.

• Monthly revenue returning to year-over-year growth; it was down 20.6% in July 2026.

• The Ministry of Finance's decision on the pending Guanxi land exchange with the Taiwan government.

• Delivery of the Taipei Zhongzheng joint-construction project, permitted in 2024 and started in January 2025.

• Third-quarter results, due around mid-November 2026, and whether the second-quarter margin recovery held.

• The Ministry of Environment's ruling on the company's high-carbon-leakage-risk application, worth roughly NT$150 million a year if granted.

• Each quarterly investment-property fair-value disclosure, which has grown at each of the last three dates reported.

• Any Indonesian anti-dumping determination, following the Vietnamese cement duties imposed in July 2025.

• The FY2026 dividend declaration, expected around March 2027, against an FY2025 payout that already absorbed 87% of earnings.

Research only, not investment advice. Position disclosure: Long, ~0.43% of portfolio

1. Scorecard

2. Argument

The short version

Asia Cement quarries limestone and makes cement, clinker, concrete, and electricity. At NT$34.55, it trades at 0.696x filing tangible book of NT$49.67, the lowest point in its 10-year history. The dividend yield is 6.66%, the company has never bought back a share, and the Piotroski F-Score is 7/9. Those are the numbers that got it onto the desk.

The catch is that the operating business earns only 5.6% on equity. China is loss-making, and 62% of first-half 2026 pre-tax profit came from outside operations. The Hualien mining-right consent and environmental review are unresolved, while a reciprocal family structure carries three capital-allocation warnings.

The verdict is BUY - STANDARD. Buy-Below is NT$38.06 from the stock's own history. The Sell Range is NT$46.96 to NT$55.64, but the lower end is the more credible exit unless return on equity recovers above roughly 9%.

Why it qualifies

• Filing-derived tangible book is NT$49.67 a share against a NT$34.55 market price.

• The 6.66% shareholder yield is entirely dividend, with no buybacks.

• The Piotroski F-Score is 7/9, with all nine signals computable.

• Cash and financial assets of about NT$97bn sit against NT$94.4bn of financial borrowings.

Why it is not a clean bargain

• Return on equity fell from 12.0% in 2019 to 5.6% in 2025 while equity grew.

• The Chinese cement operation is loss-making, with gross margin around RMB13 a tonne.

• Non-operating income supplied 62% of first-half 2026 pre-tax profit.

• The Hualien mining-right extension was revoked by final judgment, and the current consent and environmental review are not complete.

• Reciprocal ownership and related-party capital use produce three capital-allocation warnings.

What would change the view

• A sustained recovery in Chinese per-tonne gross margin from RMB13 toward the RMB25 needed for the base case.

• Clear progress through the Hualien environmental review and Truku consultation.

• Operating profit recovering while the share of non-operating gains falls materially.

• The stock reaching NT$46.96, the low end of the Sell Range, where the sell question becomes live.

Sources and gaps

The analysis uses the company's FY2025 annual report, the Q2 2026 statutory filing, and exchange price history. Individual insider transactions for the last twelve months and an earnings-call transcript were unavailable. The Hualien permitting history is drawn from company materials and secondary reporting; the live Taipower matter is the smaller civil claim remanded in May 2025. See the full report for the source ledger, calculations, and complete diligence-gap log.

3. Backup and sources

This last layer is the full working file: every source, calculation, and gap, kept so the argument above can be checked. The scorecard and the argument are the synthesis. Open this only if you want to check the work.

Asia Cement Corporation (TWSE: 1102)

Published 5 September 2026. Analysis based on the 28 August 2026 close and the 30 June 2026 statutory balance sheet. Long, approximately 0.45% of portfolio.

Verdict: BUY — STANDARD

Asia Cement is cheap on tangible book, adequately covered by its dividend, and not yet earning enough on that book to deserve a large position. The investment case is an asset discount with a cyclical recovery option attached. The operating recovery is not proven.

1. Executive summary

Asia Cement quarries limestone and makes cement, clinker, ready-mixed concrete, and electricity. The group operates two Taiwanese plants, several mainland Chinese cement subsidiaries, a gas-fired independent power plant that sells electricity to Taiwan Power Company, and a substantial portfolio of listed and unlisted financial investments.

At NT$34.55, the shares trade at 0.696x filing tangible book value of NT$49.67 per share. That is the lowest point in the 41-observation, roughly ten-year P/TBV history used here. The declared FY2025 dividend of NT$2.30 gives a 6.66% yield. The company has never bought back a share. The Piotroski F-Score is 7/9.

The cheapness is entirely an asset discount, not an earnings discount. Tangible book per share has risen from NT$35.68 at the end of 2016 to NT$49.67 today, but return on equity fell from 12.0% in 2019 to 5.6% in FY2025. Normalized earnings are NT$2.89 per share, producing an 11.95x normalized P/E. That sits inside the stock's historical earnings band, so the earnings leg is fair rather than cheap.

The balance sheet is conservative in aggregate. Financial borrowings are NT$94.4bn against NT$208.4bn of total equity, while cash and financial assets total NT$97.0bn. The refinancing risk is real because NT$49.8bn, or 53% of gross borrowings, matures within twelve months. The group has continued to issue domestic bonds, and the bonds trade close to carrying value.

The reasons to keep the position at Standard size are the Chinese cement loss, the unresolved Hualien mining-right consent and environmental review, the fact that 62% of first-half 2026 pre-tax profit came from outside operations, and a reciprocal family structure with three capital-allocation warnings. The stock can work from here, but the discount will not close by itself.

2. Business and market overview

At the parent-company level, cement and clinker account for 94% of sales. At the consolidated level, the group is more diversified. FY2025 segment revenue was approximately NT$46.7bn for Cement, NT$17.2bn for Power, NT$0.8bn for Investment, NT$0.3bn for Engineering, NT$2.1bn for Transportation, NT$3.4bn for Stainless steel, and NT$0.5bn for Leasing. Consolidated revenue was NT$71.0bn and segment result was NT$8.4bn.

The geographic split matters. FY2025 revenue was 65% Taiwan, 31% mainland China, and 4% elsewhere. Non-current assets were NT$55.8bn in Taiwan and NT$35.0bn in China. Taiwan Power Company was the only customer above the 10% disclosure threshold, taking NT$17.2bn, or 24% of consolidated revenue, from the Chia Hui Power plant under long-term capacity and energy contracts.

Taiwan's domestic cement market is protected by logistics and local production, but imports are taking share. Asia Cement sold 3.33 million tonnes domestically in 2025, equal to 33.8% of domestic producers' sales and 24.0% of island consumption. Imports reached 4.06 million tonnes, up 14.2%, and represented roughly 29% of consumption. A domestic duopoly therefore competes with a growing import channel.

Input and regulatory costs are moving against the business. The Hualien mineral tax increased from NT$10 to NT$70 per tonne, and the group has accrued a national carbon fee since 2025. The group is largely self-sufficient in limestone, but it buys slag, iron sand, gypsum, coal, and other inputs under a mixture of short- and long-term arrangements.

The consolidated balance sheet is more financial than industrial. Roughly NT$217bn of the NT$337bn asset base is cash, financial investments, equity-method investments, or investment property rather than kilns and other operating assets. That makes the asset base easier to value than the operating earnings stream. It also means the buyer is purchasing a securities-and-affiliates portfolio wrapped around a cyclical cement business.

3. Competitive position and industry cycle

Asia Cement's real advantages are limestone access, port and river logistics, and the scale of its Taiwanese production base. The Hualien plant sits beside the company's own limestone reserves and a deep-water port, and the group owns transportation subsidiaries. The Chinese plants also have access to captive limestone and river logistics.

The advantages are narrow. Cement is sold to specification, switching costs are low for ready-mixed customers, and Taiwan Power buys electricity under contract rather than because of a differentiated product. There is no meaningful network effect. Taiwan has the characteristics of an efficient-scale market, but the growing import channel is eroding that protection. China is a commodity market with structural overcapacity.

The return record confirms the limitation. Return on common equity was 12.0% in 2019, 10.0% in 2020, 9.6% in 2021, 7.5% in 2022, 6.4% in 2023, 7.1% in 2024, and 5.6% in 2025. Equity increased while the return earned on it declined. The moat classification is therefore Narrow. Asia Cement can preserve capital at acceptable rates, but it has not compounded that capital at attractive rates for the last several years.

The Chinese business is near a trough, but a trough is not a recovery. Consolidated gross margin was 29.8% in 2020 and 23.6% in 2021, fell to 13.0% in 2022 and 2023, and recovered only to 15.8% in 2025. First-half 2026 gross margin was 13.6%. Management commentary reported mainland per-tonne gross margin falling from approximately RMB40 to RMB13, with the Chinese cement operation moving from a NT$518m profit to a NT$386m loss. The group said national demand fell about 10% while its own volumes fell only 1% to 2%.

Taiwanese cement operating profit also fell 19% in the first half of 2026, as the domestic market contracted under selective credit controls. The recovery case needs mainland margin to move materially above RMB13 to RMB15 per tonne and needs Taiwan's logistics advantage to offset imports. Those are possible outcomes, not current facts.

4. Management, governance, and capital allocation

Douglas Tong Hsu has chaired Asia Cement for decades and also chairs the wider Far Eastern Group. Lee Kun-Yen is the responsible manager and Yang Yu-Ling is chief accounting officer. The chairman and chief executive are separate people and are not related according to the annual report.

The board elected in May 2026 contains ten directors and five independent directors. The previous board included eight institutional-investor directorships, all held by Far Eastern entities, alongside four independent directors. The chairman was returned. A third-party report says the chairman's son joined the board for the first time, but that detail was not confirmed in a filing reviewed for this report.

Three capital-allocation warnings are material:

• Control without enough minority protection. Asia Cement owns 22.92% of Far Eastern New Century directly, or 31.24% including entities controlled by directors. Far Eastern New Century owns 19.89% of Asia Cement. The structure is reciprocal, and every non-independent seat on the prior board was held by a group entity. Five independent directors and functioning committees are meaningful mitigants, but they do not remove the conflict created by the cross-holding.

• Capital routed to related-party or low-return uses. The group subscribed approximately NT$687m to related-party capital increases in FY2025 and donated NT$425m to a group university across FY2025 and the first half of 2026. The group also carries a large portfolio of securities and equity-method investments whose blended return is below the return shareholders would want from the asset base.

• A recent dividend cut. The FY2023 dividend was NT$2.10, down from NT$2.30. Dividends have since recovered to NT$2.20 and NT$2.30, so this is not a distress pattern, but the cut falls inside a three-year review window.

The related-party note is extensive and says transactions are on terms equivalent to unrelated-party transactions. In the first half of 2026, related-party revenue was NT$2.09bn and related-party operating costs were NT$1.06bn. Related-party receivables rose from NT$1.03bn at the end of 2025 to NT$3.99bn at 30 June 2026, including NT$1.62bn owed by Far Eastern New Century. Asia Cement also held NT$5.87bn on deposit at Far Eastern International Bank and borrowed NT$720m from it.

The group has paid dividends every year examined and has never repurchased a share. The tradeoff is that retained capital has been directed into securities, related-party equity, Chinese capacity, and group guarantees. Endorsements and guarantees for subsidiaries totalled NT$49.5bn at the latest filing.

5. Ownership, subsidiaries, and joint ventures

Asia Cement had 3,546,562,881 ordinary shares outstanding at 30 June 2026. There is one class of ordinary share and one vote per share. The ten largest holders represented 40.21% of the register, while the residual free float was 59.79%.

The largest holder was Far Eastern New Century at 19.89%, followed by the Far Eastern Medical Foundation at 5.15%. Identifiable Far Eastern-affiliated holders inside the top ten represented approximately 30.9%. The ordering of several ETF and investment-company lines was inferred from the descending sequence in the annual report, but the counts reconcile to the reported top-ten total.

The share count needs care. Shares held by equity-method associates are treated as treasury shares for earnings-per-share purposes, which is why first-half 2026 basic EPS uses 3,339,502 thousand weighted shares rather than the 3,546,563 thousand issued. Dividends are declared on the full issued count. The FY2025 dividend totalled NT$8.16bn, equal to NT$2.30 multiplied by the issued shares.

The main operating entities are:

• Asia Cement China Holdings, the consolidated Chinese cement platform, in which Asia Cement owns 67.73% directly and 72.27% including director-controlled entities.

• Chia Hui Power, the gas-fired independent power plant and the borrower under the group's only disclosed financial covenant package.

• Ya Tung Ready-Mixed Concrete, the Taiwanese downstream concrete business.

• Yuan Long Stainless Steel, a wholly owned and loss-making stainless-steel business.

• Far Eastern New Century, an equity-method associate and reciprocal holder of Asia Cement.

• U-Ming Marine Transport, an equity-method associate that contributed materially to first-half 2026 earnings.

• China Shanshui Cement, a minority equity-method investment carried at NT$15.24bn at the end of 2025. Its FY2025 share of loss was NT$812m.

Non-controlling interests were NT$23.94bn, or 11.5% of total equity, and sit mainly in the Chinese platform. Assets pledged for borrowings were NT$33.12bn, including NT$15.14bn of investment property and NT$6.99bn of equity-method investments.

6. Historical financial quality and normalized earnings

The long record is mixed. Revenue rose from NT$60.9bn in 2016 to a peak of NT$90.3bn in 2022, then fell to NT$71.0bn in 2025. Operating income was NT$22.1bn in 2019 and NT$8.4bn in 2025. Gross margin fell from 29.8% in 2020 to 13.0% in 2022 and 2023 before recovering to 15.8% in 2025.

Tangible book per share moved in the opposite direction. It increased from NT$35.68 at the end of 2016 to NT$49.09 at the end of 2025 and NT$49.67 at 30 June 2026. This is the central financial tension: the asset base grew while the operating return on that base deteriorated.

Cash conversion is adequate over a cycle but volatile. Free cash flow averaged approximately NT$13.5bn a year from FY2018 through FY2025. FY2025 operating cash flow was NT$16.84bn, capital expenditure was NT$4.04bn, and free cash flow was NT$12.80bn. The large annual swings come from working capital and the securities portfolio moving through operating cash flow.

A conservative owner-earnings calculation for FY2025 is net income attributable to owners of NT$10.03bn, plus depreciation of NT$3.90bn and amortisation of NT$0.27bn, less capital expenditure of NT$4.04bn. That produces NT$10.16bn, close to reported free cash flow. Maintenance capital expenditure is not separately disclosed, so using total capital expenditure is the conservative choice.

Normalized earnings are based on the ten-year average of profit attributable to owners. That average is NT$10.25bn, or NT$2.89 per share on the issued share count. The five-year average is 1.19 times the ten-year average, which is not enough to classify the earnings series as structurally growing. The normalized P/E is therefore 11.95x.

The quality issue is the earnings mix. Non-operating income and expense contributed 33% of FY2025 pre-tax profit, 50% of FY2024 pre-tax profit, and 62% of first-half 2026 pre-tax profit. First-half 2026 net profit attributable to owners rose 66% to NT$7.11bn while revenue fell 7%. NT$4.77bn of NT$7.66bn pre-tax profit came from outside operations. The reported earnings recovery is therefore a securities and associate-income recovery, not a cement recovery.

7. Piotroski F-Score

The Piotroski F-Score is 7/9 for FY2025 compared with FY2024. The company passed on positive net income, positive operating cash flow, cash flow above net income, lower leverage, improved liquidity, no dilutive issuance, and improved gross margin. It failed on return on assets and asset turnover.

The two failed signals describe the same problem. Revenue fell 6.9% and profit fell 22% against a balance sheet that shrank only 2.3%. This is a revenue and utilization problem, not a solvency, liquidity, accrual, or dilution problem. That is a better failure pattern than a score supported by deteriorating leverage or cash quality.

The score is useful but not decisive. It says the balance sheet and cash conversion are currently sound. It does not say that the Chinese cement business has recovered or that the group is allocating new capital at attractive returns.

8. Balance sheet, debt, and refinancing

Financial borrowings were NT$94.43bn at 30 June 2026. Lease liabilities were separately disclosed at NT$1.50bn and are not included in the financial-borrowing figure used for the absolute debt test. Financial borrowings were 45.3% of total equity and 28.0% of total assets. Cash and financial assets totalled NT$96.96bn, slightly more than gross financial borrowings.

The borrowing mix was approximately NT$33.4bn of short-term bank lines, NT$7.2bn of commercial paper, NT$6.6bn of long-term bank loans, and the balance in domestic unsecured bonds. Interest rates were generally between 0.60% and 2.46%.

FY2025 operating EBITDA was NT$12.57bn. It covered gross finance costs of NT$1.51bn by 8.3 times and net interest costs by 27.6 times after interest income. Those ratios support a Conservative balance-sheet classification, although gross interest coverage by itself is closer to Manageable.

The risk is rollover rather than immediate solvency. NT$49.82bn, or 53% of financial borrowings, matures within twelve months. The group issued NT$4.3bn of domestic bonds in the first half of 2026 and NT$18.8bn during FY2025, at 0.60% to 2.10%. The bonds were marked at NT$47.18bn against NT$47.30bn of carrying value, a small discount that suggests the domestic market remains comfortable with the issuer.

The only disclosed financial covenants sit at Chia Hui Power. The NT$10.5bn syndicated facility has a 150% total-liabilities-to-net-worth limit and a 1.5x principal-and-interest coverage requirement. Chia Hui's standalone statements were not obtained, so subsidiary-level headroom cannot be confirmed. The equivalent group-level liabilities-to-equity ratio was 61.9%, well inside the limit.

9. Real estate, leases, and hidden assets

Right-of-use assets were NT$5.12bn, or 2.9% of tangible book value, against lease liabilities of NT$1.50bn. They are disclosed but not stripped from tangible book value. At this scale, tangible book is not a lease illusion.

Investment property is more important. The balance was NT$39.20bn at 30 June 2026, or 22.3% of tangible book, and it is carried at appraised fair value rather than historical cost. The auditor treated the valuation as a key audit matter. There is no large hidden revaluation waiting to appear in book value. Much of the uplift is already present.

The property portfolio includes unleased land in Kaohsiung and Taipei valued using land-development analysis, plus leased property valued using discounted rental cash flows. First-half rental income was NT$242m, approximately a 1.2% annualized yield on the carrying value, and NT$154m of fair-value gains went through the income statement.

Taiwanese rules require cumulative net fair-value increases on investment property to be appropriated to a special reserve. That reserve was NT$67.61bn, or 36.7% of equity attributable to owners, and is reversible only as properties are used or sold. The fair value is real, but the reserve makes the full amount less immediately distributable than ordinary operating cash.

The asset composition is therefore a qualification on the value case. Cash and short-term investments were NT$76.74bn, long-term investments were NT$125.41bn, and owned property, plant, and equipment was NT$41.41bn. The tangible book is liquid in parts, appraised in parts, and operational in parts. It should not be treated as a single homogeneous pile of cash.

10. Legal, regulatory, and environmental matters

The Hualien mining-right issue is the largest asset-value risk. The extension was revoked by a final judgment in 2021. A 2023 amendment to Taiwan's Mining Act requires Indigenous consent and environmental impact assessment steps. The company has continued to describe extraction as possible within the existing mining land, but the consent and environmental process remains unresolved.

The risk is not only legal. Hualien supplies limestone to the Taiwanese operating base. A failure of the process could force imported clinker or other feedstock, reduce the value of the mining rights, and impair the economics of the plant. The current value case therefore assumes the process remains workable, not that the permit issue has disappeared.

Chia Hui Power also faces a revived civil claim from Taiwan Power Company. Taiwan Power's claim is approximately NT$767m plus interest, and the Supreme Court remanded the matter for rehearing on 12 May 2025 after Asia Cement's group company had won at earlier stages. No provision was judged necessary in the reviewed financial statements. The amount is manageable against group equity, but the claim is relevant because Taiwan Power is the plant's sole customer.

The Hong Kong High Court dismissed the claims against Asia Cement in the China Shanshui conspiracy action on 12 May 2025, with costs indicated on the indemnity basis. That matter is not a current thesis problem.

The group has also accrued a carbon fee. The balance was NT$23m at 30 June 2026 after the company was designated as a high-carbon-leakage business. The current amount is small, but the cost can rise as preferential treatment changes and carbon-reduction requirements tighten.

11. Accounting quality and diligence limits

The FY2025 audit opinion from Deloitte & Touche was unmodified. The audit identified substantive key audit matters, including expected credit losses on Chinese trade receivables and the valuation of investment property. The 30 June 2026 interim review carried a qualified conclusion covering entities holding 10.4% of group assets. That is not a finding of fraud, but it lowers the confidence that a reader should attach to every consolidated number.

The company does not disclose segment assets, liabilities, and capital employed in enough detail to calculate a clean cement-only return on capital. The group-level return record is therefore more reliable than any claim about the standalone economics of the Taiwanese cement plants.

Several diligence limits remain:

• Individual dated insider trades for the last twelve months were not obtained. The annual report provides shareholding comparisons over a longer period, including a small personal sale by one director, purchases by another director, and stable or rising institutional positions.

• An earnings-call transcript was not obtained. Management's comments on per-tonne margins and demand came from secondary coverage of the August 2026 investor conference.

• Chia Hui Power's standalone financial statements were not obtained, so covenant headroom is stated at group level as a proxy.

• FY2016 and FY2017 gross profit and operating income were not used because the older presentation could not be independently reverified. Revenue, net income, equity, and tangible-book figures for those years did reconcile.

• The reported first-time appointment of the chairman's son was not confirmed from a primary filing.

These gaps affect diligence confidence, not the arithmetic of the current P/TBV, normalized EPS, dividend yield, or balance-sheet figures used in the decision.

12. Valuation and decision points

The tangible-book calculation is straightforward. Equity attributable to owners was NT$184.42bn. Goodwill of NT$2.61bn and other intangibles of NT$5.66bn were deducted, leaving tangible book value of NT$176.16bn. Dividing by 3,546.563 million issued shares gives NT$49.6706 per share. At NT$34.55, the stock trades at 0.696x P/TBV.

The P/TBV history contains 41 quarterly observations over roughly ten years. The lower-quartile mean is 0.766x, the median is 0.945x, and the upper-quartile mean is 1.120x. The current ratio is below the lower-quartile mean and below the historical median. This is a real asset discount, not a rounding artifact.

The earnings leg is less generous. Normalized EPS is NT$2.89 and normalized P/E is 11.95x. That is inside the stock's historical earnings band. Trailing P/E is lower at 9.54x, but trailing profit is flattered by the non-operating gains described above. The current earnings multiple should therefore not be treated as evidence that the operating business is unusually cheap.

The current decision figures are:

• Buy-Below: NT$38.06, equal to 0.766x tangible book. This is the lower-quartile mean of the company's own P/TBV history and is the appropriate company-relative entry reference.

• Buy More Below: NT$32.29, equal to 0.65x tangible book. This is the deeper fresh-money level for a larger position under the valuation grid.

• Sell Range: NT$46.96 to NT$55.64, equal to 0.945x to 1.120x tangible book. The lower end is the more credible exit while return on equity remains around 5.6%. The upper end becomes defensible only if return on equity recovers above roughly 9%.

The stock's operating enterprise value is less flattering. Market capitalisation of approximately NT$122.5bn, plus NT$94.4bn of borrowings, less NT$97.0bn of cash and financial assets, produces an enterprise value near NT$119.9bn. Against FY2025 operating EBITDA of NT$12.6bn, that is roughly 9.5x. A cement business at a cyclical low is not obviously cheap on this measure. This is why the book discount supports a Standard position rather than a concentrated one.

The peer comparison points in the same direction. On the exchange's conventional book-value measure, the other Taiwanese cement major traded near 0.79x book with a 3.29% dividend yield and no meaningful trailing P/E, while Asia Cement traded near 0.66x book with a 6.66% yield and a 9.54x trailing P/E. Asia Cement is cheaper and more income-covered, but the discount is partly the market's response to its lower operating returns and more complicated asset structure.

13. Risk review

Hualien mining rights. Severity is High and probability is Medium. Failure of the consent and environmental process could impair up to NT$5.28bn of mining rights and disrupt the limestone feed for the Taiwanese plants. The mitigants are the company's claim that extraction can continue within existing mining land and its stated reserve life. Monitor consent meetings, environmental milestones, and any renewed approval.

Chinese cement and receivables. Severity is High and probability is High. The Chinese cement operation was loss-making in the first half of 2026, with per-tonne margin falling to approximately RMB13. Mainland property weakness also makes trade-receivable losses a live accounting risk. Monitor per-tonne margin, volumes, receivable ageing, and new expected-credit-loss charges.

Dependence on non-operating gains. Severity is High and probability is High. Non-operating items supplied 62% of first-half 2026 pre-tax profit. A flat securities market or weaker associate income would pull reported earnings back toward the lower normalized base. Monitor the operating versus non-operating split every quarter.

Reciprocal ownership and related-party capital use. Severity is Medium and probability is High. The cross-holding and related-party transactions can direct capital toward group objectives rather than the highest-return use for Asia Cement shareholders. Five independent directors and formal committees help, but the structure remains a permanent governance discount. Monitor new share subscriptions, donations, related-party receivables, and changes in the cross-holding.

Refinancing. Severity is Medium and probability is Low under current conditions. More than half of borrowings mature within twelve months, but the group has cash and financial assets exceeding gross borrowings and has continued to issue domestic bonds. Monitor bond coupons, commercial-paper rates, short-term bank lines, and the next refinancing cycle.

Investment-property valuation. Severity is Medium and probability is Medium. A 20% reduction in the NT$39.2bn investment-property portfolio would reduce tangible book value by roughly NT$2.2 per share. The independent appraisal and special reserve are mitigants, but fair value is not the same as cash received. Monitor annual appraisal assumptions and comparable land prices.

Power-plant customer concentration and litigation. Severity is Medium and probability is Medium. Taiwan Power accounts for 24% of consolidated revenue, and the Chia Hui civil claim remains in rehearing. The claim is modest against group equity, but a contract or tariff dispute would matter more than the claim amount alone. Monitor the rehearing and power-tariff policy.

Taiwan imports and carbon costs. Severity is Medium and probability is High. Imports were roughly 29% of 2025 consumption and rose 14.2%. Carbon fees are small today but can rise. Monitor import volumes, domestic cement pricing, the carbon-fee rate, and the company's blended-cement programme.

14. Final recommendation and watch list

BUY — STANDARD. Asia Cement is a reasonable purchase at NT$34.55 because the tangible book is real, the balance sheet is sound in aggregate, the dividend is meaningful, and the shares trade below the lower end of their own P/TBV history. The stock is not a large-position candidate because the operating return is weak, China is loss-making, much of current profit is non-operating, and the ownership structure creates a permanent minority-shareholder concern.

The practical framework is simple. Buy-Below is NT$38.06. A deeper fresh-money level is NT$32.29. If the shares reach NT$46.96, the sell question becomes live. Hold for the upper end near NT$55.64 only if return on equity has recovered above roughly 9% and the Chinese operation is earning rather than consuming capital.

The most important items to watch are:

• Chinese cement margin per tonne, especially any move above the RMB13 to RMB15 floor toward RMB25.

• Huanggang Line 2, scheduled to start in the third quarter of 2026, and the remaining Chinese capital programme.

• The Hualien consent and environmental-review process.

• The Taiwan High Court rehearing of the Chia Hui and Taiwan Power claim.

• The operating versus non-operating split of third-quarter 2026 profit.

• Mainland receivable ageing and expected-credit-loss charges.

• Related-party receivables, especially the NT$1.62bn owed by Far Eastern New Century.

• Any new related-party share subscription, group donation, or change in the reciprocal cross-holding.

• The next domestic bond issue and the rollover of NT$33.4bn of short-term bank lines.

• The FY2026 dividend decision expected in March 2027.

Sources and methodology note

The primary financial sources were the Asia Cement 2025 annual report, the 30 June 2026 interim statutory filing, and the FY2025 audited statutory filing. Price history and exchange valuation data came from the Taiwan Stock Exchange daily data and per-stock valuation data. The company's mining-right FAQ was also reviewed. Management's per-tonne margin commentary and the Hualien chronology were checked against secondary coverage and are identified as such in the relevant discussion.

This report is research, not investment advice. The author is long Asia Cement, approximately 0.45% of portfolio.

1. Scorecard

2. Argument

The short version

Jiangxi Ganyue Expressway operates eight state-controlled toll roads in Jiangxi province, trading at 0.47x tangible book value and 9.1x normalized earnings. Tangible book per share has compounded upward every year for seven straight years without interruption, and the concession schedule behind the roads now runs 7 to 23 years, with only one small road expiring near-term.

The discount looks larger than it is. At the enterprise level, where net debt roughly equals the market value of the equity, the real discount to tangible capital is 33.7%, not the 52.9% headline. Two disclosed flags keep this a Standard-size position rather than the largest one the numbers alone would justify: the company's cheapest financing runs through its controlling shareholder rather than standing on its own credit, and 27.6% of tangible book was capitalized in the first half of the year on a concession expansion whose re-approval is not yet confirmed.

Why it qualifies

• Trades at 0.47x tangible book value, a 52.9% discount at the equity level and 33.7% at the enterprise level.

• Normalized earnings multiple of 9.1x on a seven-year basis.

• Tangible book per share has compounded upward every single year for seven straight years, with no interruption.

• Eight toll concessions now confirmed to run 7 to 23 years; only one small road, 2.75% of toll revenue, expires near-term.

• A 4.4% trailing dividend yield and an F-Score of 6 out of 9.

Why it is not a clean bargain

• CNY 5,278.7 million, 27.6% of tangible book, was capitalized in the first half of the year on a concession expansion whose re-approval is not yet confirmed.

• The cheapest financing runs through the controlling shareholder: cash on deposit at its finance company, a major loan intermediated through it, and fees paid to it to operate the roads. The same fact functions as both a balance-sheet strength and a governance flag.

• Related-party transaction flows run 4.5x to 7.5x the board-approved CNY 180m cap, with no confirmed explanation for the gap.

• The dividend was not covered by free cash flow in two of the last three years.

• 96.7% of short-term investments is a single related-party brokerage stock, and its mark-to-market swing drove most of this year's reported profit decline.

• The controlling shareholder holds 47.84% of the company, a fired governance concentration warning on its own.

What would change the view

• Confirmation, either way, of whether the concession behind the recently completed expansion is re-approved.

• A primary filing on the January 2026 change of control to the provincial state asset regulator.

• Resolution of the 4.5x to 7.5x gap between the board-approved related-party cap and actual related-party flows.

• The related-party shareholder loan maturing 2026-09-07.

• Price at or below CNY 4.02, or at or above CNY 5.03, the top and bottom of the stock's own valuation range.

• Confirmation of whether the second-largest shareholder, at 4.99%, is state-linked.

Sources and gaps

This draws on the company's own H1 2026 and annual filings and a national credit rating agency's report, with the tangible-book construction independently confirmed against a mirror of the exchange filing. The concession re-approval, the January 2026 control change, and the related-party cap gap remain unconfirmed by a primary source; where a fact could not be verified, the report says so rather than estimating it. See the full detail for the source ledger, calculations, and complete diligence-gap log.

3. Backup and sources

This last layer is the full working file: every source, calculation, and gap, kept so the argument above can be checked. The scorecard and the argument are the synthesis. Open this only if you want to check the work.

Company: Jiangxi Ganyue Expressway Co., Ltd. (江西赣粤高速公路股份有限公司) Ticker: 600269, Shanghai Stock Exchange As of: 2026-08-28, price CNY 3.86, market capitalisation CNY 9,014.7m (USD 1,341m at USD/CNY 6.7217). Buy-Below, Sell Range, Add-More Trigger and the tier's stated reasoning below were corrected 2026-08-29; see the note at the start of the valuation discussion. P/TBV: 0.471× (filing TBV/share CNY 8.1876 at 2026-06-30). Verdict: BUY — STANDARD. Buy-Below: CNY 4.02 (0.49× TBV, 10.2-yr own-history). Sell Range: CNY 5.03 – 6.59 (0.61× – 0.81× TBV). Add-More Trigger (held position): CNY 2.64 (0.32× TBV), a review price for adding, not a mechanical trigger. Moat: Narrow. Governance risk: High, 2 fired warnings, 3 borderline. Position: Long, ~0.5% of portfolio, initiated 2026-08-26.

Primary sources reviewed: the 2025 Annual Report (179pp, audited), for the fiscal year ended 2025-12-31, filed 2026-03-31, read in full from the exchange's designated disclosure venue; the 2026 Interim Report (144pp, unaudited), for H1 ended 2026-06-30, filed 2026-08-10; the 2026 First Quarter Report (19pp), filed 2026-04-28; the 2025 annual general meeting resolutions of 2026-06-18 (dividend, removal of a director, D&O insurance) as reported in the interim report, though the AGM circular itself was not opened; the 2025 preliminary results bulletin was identified but not opened, superseded by the audited annual report; insider activity from the 2026 Interim Report's board, senior-management, and officer shareholding sections (China has no Form 4 analogue); every figure below comes from a primary filing or a stated secondary source, not a pre-built dataset. Other sources: the 2025 Annual Report Summary, the 2023 and 2021 Annual Report Summaries, the announcement on change in accounting estimate (2026-08-11), exchange monthly closing-price history from 2016-03 to 2026-08 and the 2026-08-28 closing quote, press coverage of the H1 2026 result and the 2026-07-14 profit warning, and a USD/CNY reference rate. Data quality: EXCELLENT.

Jiangxi Ganyue Expressway Co., Ltd. — Value Investing Analysis

1. Executive Summary

Jiangxi Ganyue Expressway operates eight tolled expressways with close to 800 km of mainline in Jiangxi province and collects a toll on every vehicle that uses them, and it runs four smaller non-toll businesses — smart-transport engineering, service-area fuel retailing, property development and construction, and a financial-investment book — alongside that core. The verdict is BUY — STANDARD. At CNY 3.86 (2026-08-28) the shares carry a market capitalisation of CNY 9,014.7m against a filing-derived tangible book value of CNY 19,121.3m, or CNY 8.1876 per share — a P/TBV of 0.471×, near the bottom of the roughly ten-year, 41-quarter series this position's price targets are built from — fractionally above the series' own floor of 0.454×, not the single cheapest print, but close to it. The tangible book is genuine road, not accounting air: owned property, plant and equipment is CNY 26,400.4m, right-of-use lease assets are CNY 44.4m (0.2% of tangible book), and goodwill is CNY 3.6m on a CNY 38.0bn balance sheet. The name is cheap against tangible book value in the strict sense the phrase is meant to carry.

The reason the market is offering it at this price is a reported-earnings collapse that has almost nothing to do with the roads. Attributable profit for the first half of 2026 fell 72.81% to CNY 207.7m, but profit before non-recurring items fell only 14.27%, to CNY 536.4m, and toll service revenue actually rose 0.62% to CNY 1,807m. The gap is a CNY 466.2m negative fair-value swing on a listed-equity book — principally 78,500,053 shares of Guosheng Securities marked at CNY 11.00 on 30 June 2026 — plus a CNY 97.7m fall in the Xinda Real Estate holding taken through other comprehensive income. This is a toll-road company whose reported profit is being driven by a securities portfolio it did not need to own, and that is a governance finding before it is a valuation one.

The balance sheet supports the discount, with one qualification. Financial borrowings of CNY 12,321.7m sit against total equity of CNY 21,129.4m and total assets of CNY 37,973.5m — 58.3% of equity and 32.4% of assets — with lease liabilities a separate and trivial CNY 45.4m; net debt to adjusted EBITDA is 3.17× as reported, 3.46× excluding the CNY 811.7m held at the parent's finance company. Funding is cheap and lengthening: the most recent medium-term notes priced at 1.51% and 1.55% in late July 2026, the issuer rating is AAA, and a CNY 12.0bn unified debt-financing programme was registered in the first half. The Piotroski F-Score is 6 of 9, failing the ROA, leverage and gross-margin comparisons. Cash return is real: CNY 0.17 per share for FY2025, paid 8 July 2026, a 4.40% yield at today's price, with cumulative dividends since the 2000 listing above CNY 6.7bn at an average payout near 31%.

What holds this position to Standard rather than the Big tier the mechanical matrix reads is my own judgment, not a formula — the current method deliberately no longer folds governance flags into the tier automatically. Jiangxi Provincial Communications Investment Group holds 47.52% directly and a further 0.32% through a wholly-owned subsidiary acting in concert, under the ultimate control of the Jiangxi provincial state-asset regulator; the company pays that same controlling shareholder CNY 49.97m a year to operate its own toll, maintenance, safety and rescue centres; CNY 811.7m of the group's CNY 3,307.9m of cash sits on deposit with the parent's captive finance company; and the capital that produced the first-half loss went into a related-party securities house rather than into the road network. Two of five capital-allocation warnings fire and I weigh them as real enough to hold this one notch below what the numbers alone say, setting Governance Risk at High. The asset discount is wide enough to pay for that; it is not wide enough to ignore it.

2. Business and Market Overview

The toll business is eight separate expressways, all commercially operated concessions: Changjiu (127.187 km), Changzhang (104.978 km), Changtai (147.72 km), Jiujing (134.712 km), Wenhou (35.5 km), Penghu (63.93 km), Changfeng (39.05 km) and Fengtong (131.183 km). They sit on the north-south and east-west trunk of the Jiangxi provincial network and connect the provincial capital Nanchang to the province's main cities. Revenue is a toll per vehicle at prices, terms and charging methods set by government — the company says exactly that in its own risk disclosure — so volume is the only variable management controls, and even that is shaped by what else the province builds.

The five reported segments for the first half of 2026 give the shape of the group precisely: toll service revenue CNY 1,822.6m against CNY 888.9m of segment cost; engineering construction CNY 179.6m; refined-oil sales CNY 742.3m; property development CNY 31.2m; other CNY 30.8m; intersegment eliminations CNY -146.0m; total CNY 2,660.5m. Pre-tax profit by segment shows the same concentration — toll CNY 284.1m, refined oil CNY 110.7m, engineering CNY 8.4m, property CNY -42.9m. The toll segment carries CNY 43,166.6m of gross segment assets before eliminations against total group assets of CNY 37,973.5m; everything else is small.

Group revenue fell 13.66% year on year in the first half and none of that came from tolls. The smart-transport subsidiary Fangxing Technology saw revenue fall from CNY 468.3m to CNY 158.1m and earned CNY 1.7m of net profit; the property subsidiary lost money; refined-oil volumes were squeezed by international price moves and by electric substitution, which the company is answering by opening battery-swap stations for heavy trucks in its service areas — the first twelve went into operation in February 2026 under a strategic agreement signed in October 2025 with a CATL-controlled counterparty. Customer concentration inside the toll business is total but not credit-risky in the ordinary sense: at parent level 99.34% of receivables and contract assets are owed by the Jiangxi Provincial Traffic Monitoring and Command Centre, the state body that clears networked toll collection.

As of January 2026: the Zhangji expressway reconstruction and expansion project reached dual eight-lane running condition, and CNY 5,278.7m of construction in progress was transferred into fixed assets during the half — construction in progress fell from CNY 5,513.7m to CNY 235.0m while fixed assets rose from CNY 20,472.4m to CNY 26,400.4m. The company says it will apply to have the Zhangji toll period re-determined once the whole line is open, as it successfully did after the Changjiu and Changzhang widenings, and its own accounting-estimate announcement assumes a re-determined term of 30 years. That application is the most valuable pending item in the business and none of its value is in book equity today; see the concession schedule below, and the red flags list later in this report, for why this is treated as a live, unresolved risk rather than a formality.

Concession schedule (added after this draft's original writing — the single largest open item going into publication, retrieved from a national credit-rating agency's 2025 report after six earlier failed attempts):

Concession periods by road: Changjiu 1998.04 to 2049.09; Changzhang 1999.07 to 2044.03; Wenhou 1999.01 to 2028.12; Jiujing 2000.11 to 2030.11; Changtai 2003.06 to 2033.06; Penghu 2010.09 to 2040.09; Changfeng 2011.12 to 2041.12; Fengtong 2012.10 to 2042.10.

Only Wenhou expires near-term (2028.12), and it is 2.75% of 2024 toll revenue on the rating agency's own footnote. The agency's conclusion, in its own words: "the remaining operating periods of the other controlled toll roads are all 5+ years, with no near-term toll-collection-rights expiry pressure." Changjiu was re-approved out to 2049.09 after its own expansion completed, which is the precedent that matters for Zhangji. Hedge, stated plainly: this schedule is a secondary source, a rating agency's table dated to the end of 2024, not the concession grants themselves — it narrows the Zhangji/Changtai risk (Changtai itself runs to 2033.06, well beyond any near-term concern) but does not confirm that Zhangji's specific re-determination has actually been granted. That remains open; see the red flags list later in this report.

This is an understandable, highly predictable, mildly cyclical, capital-intensive regulated business bolted to four much less predictable ones. It is not a melting ice cube — the roads are being widened and their tolling lives re-cut — but essentially all of the group revenue growth of the last five years came from the non-toll segments, and those are exactly where the losses are.

3. Moat, Competitive Position, and Industry Cycle

Of the five moat sources only two apply here, and both apply strongly within their limits. Efficient scale is the real one: a built expressway on a trunk corridor cannot be economically duplicated, and the company's own listing-era undertakings record that the provincial government agreed in 1998 not to build, rebuild or expand any parallel road within 50 km of the Changjiu, Yinsanjiao and Changzhang alignments while average daily traffic stayed below defined thresholds. Cost advantage is the second: incremental traffic costs almost nothing to serve, which is why toll segment revenue of CNY 1,822.6m carried only CNY 888.9m of cost in the half. Brand, switching costs and network effects are irrelevant to a road.

The moat is real but it is Narrow, for three reasons the filings themselves supply. First, it is finite-lived — these are concessions depreciated by traffic volume over an estimated remaining tolling period, not perpetual assets, and every extension has to be applied for and granted. Second, the 1998 non-parallel-road undertaking has effectively lapsed: the interim report states that the promise cannot be performed because of the national policy of vigorously developing expressways, and that the actual controller and controlling shareholder have publicly explained the situation rather than compensated for it. The protection that made the corridor a monopoly is gone, and the party that removed it also controls the company. Third, the same controlling group owns the competing provincial network, so route competition is an internal allocation decision rather than a market one.

Compounding is therefore limited. Return on equity has run between 3.3% and 7.1% over the seven years to FY2025 and was 6.6% in FY2025 — a regulated, cash-generative, low-return business, not a compounder. Capital reinvested into road widening does buy extended concession life, which is the one mechanism by which this company grows intrinsic value organically, but the return on that capital is set by the same government that grants the extension.

On the cycle, the toll business is Mid-cycle and there is no depressed-earnings snapback embedded in the price. National road freight volumes rose 3.2% and road passenger movements 0.9% in the first half of 2026, while the company's toll service revenue rose 0.62% — running behind the national numbers, which is consistent with diversion onto newly built parallel capacity rather than with weak underlying demand. FY2025 toll service revenue was CNY 3,608m, up 0.77%. Toll margins are stable, volumes flat to marginally positive. What is cyclical in this share price is not the road; it is the mark-to-market on a securities book.

4. Management, Governance, and Capital Allocation

The chairman and legal representative is Han Feng, who signs the accounts as the person responsible for the company. The finance director is Li Shikun, appointed by the board on 11 February 2026; the board secretary is Fu Yan. Beyond that, the management table has turned over hard in eight months. Hu Wei resigned as independent director on 25 November 2025 for personal reasons and was replaced by Yan Qingqing, elected at the extraordinary general meeting of 23 January 2026. Xu Zhihua resigned as director, general manager and member of the strategy and investment decision committee on 27 May 2026, again for personal reasons, and left the company entirely. Nie Jianchun was removed as a director by resolution of the 18 June 2026 annual general meeting, recorded as a work transfer. Xu Yibiao resigned as deputy general manager on 21 July 2026. That is five board and senior-management changes in eight months, including the general manager, and the interim report names no successor general manager. None of the departures is explained beyond "personal reasons" or "work transfer" — an absence of explanation that is itself a fact to report rather than a reassurance.

Insider ownership is nil. The interim report's section on shareholdings of serving and departed directors and senior managers is marked not applicable, meaning no director or officer holds shares and none traded in the period. China has no Form 4 analogue; the periodic report and the exchange's shareholding-change announcement regime are the disclosure venues, and both are silent, so "none found" here is a statement about the regime searched rather than an inference from absence. There is no equity incentive plan, no employee share ownership plan and no other incentive scheme. The board adopted new remuneration measures for directors and senior managers in the first half of 2026, and the annual general meeting approved directors' and officers' liability insurance capped at CNY 150,000 a year — modest, and not evidence of outsized pay, though individual compensation is not disclosed in an interim report and was not verified here.

Ownership is concentrated. Jiangxi Provincial Communications Investment Group holds 1,109,774,225 shares, or 47.52%, and Jiangxi Highway Development Co., its wholly-owned subsidiary and a declared party acting in concert, holds a further 7,379,960 shares, or 0.32%, for 47.84% combined. The ultimate controller is the Jiangxi provincial State-owned Assets Supervision and Administration Commission, which gave an undertaking in 2025, on completion of a free transfer of the controlling stake, not to transfer the 47.85% it indirectly controls for eighteen months. The largest non-state holder is Great Wall Life Insurance with 159,985,682 shares (6.85%) of own funds plus 13,258,410 (0.57%) in a universal-life account. The register held 49,523 ordinary shareholders at 30 June 2026, down from 60,547 at the end of 2023.

Related-party dealing is heavy, and disclosed in full. Under agreements signed on 31 December 2024, the controlling shareholder operates the toll collection, maintenance, safety and rescue centres of the company's own roads for CNY 49.97m a year through to 31 December 2027 — CNY 34.50m from the parent company, CNY 10.29m from Changtai and CNY 5.18m from Changtong. Ordinary-course related-party transactions for 2026 are capped at CNY 110m. At 30 June 2026 the group owed related parties CNY 3,300.4m and was owed CNY 1,414.7m; CNY 811.7m of cash sat with the parent's finance company against a contractual daily cap of CNY 850m; and CNY 140m was lent to Guosheng Securities as subordinated debt running to 27 November 2030. Guosheng Securities is also the company's largest listed-equity holding at CNY 863.5m and appears in the related-party dealings table, which places the largest financial-asset position — and the CNY 466.2m first-half fair-value loss it drove — inside the controlling group's own orbit.

Capital allocation has two faces. The dividend record is good and improving: CNY 0.12 per share for FY2021, cut to CNY 0.10 for FY2022, then CNY 0.16 for FY2023, CNY 0.17 for FY2024 and CNY 0.17 for FY2025, the last paid on 8 July 2026 for CNY 397.0m in total and 31.03% of attributable profit. Cumulative distributions since the 2000 listing exceed CNY 6.7bn at an average payout near 31%. There has never been a buyback, and none is proposed in the "quality, efficiency and returns" action plan the company has run since April 2024. On the other side, capital went into a securities portfolio that lost CNY 466.2m of fair value in six months, into a Fangxing spin-off worked on from March 2021 and abandoned in March 2025 on changed policy and market conditions, and into the Changtong subsidiary, whose equity now stands at CNY 363.1m against registered capital of CNY 4,500m and which lost CNY 97.8m in the half after losing CNY 44.2m in the comparable period.

The capital-allocation warnings are mixed. (1) Control without offsetting minority protection — FIRED: 47.84% concert-party control by a provincial state group under the state-asset regulator, and a listing-era protective undertaking whose sister promise has already been allowed to lapse. Correction: the evidentiary basis originally cited alongside this — "no disclosed independent-director veto over related-party capital allocation" — is not sound as stated: a Board Strategy & Investment Decision Committee demonstrably exists (the departing general manager sat on it before his May 2026 resignation). The warning still fires on concentration alone; that limb of the original reasoning is withdrawn, not the conclusion. (2) Capital routed to parent-mandated, policy-driven or low-return projects — FIRED: road widening executed to the provincial "transport-strong province" plan; operation of the company's own roads outsourced to the controlling shareholder at CNY 49.97m a year; CNY 811.7m of cash placed with the parent's finance company; CNY 863.5m of equity and CNY 140m of subordinated debt in a related-party securities house; and a loss-making Changtong funded pro rata by a CNY 1,989.9m shareholder loan. (3) Dividend cut or an inadequate payout — BORDERLINE, not fired: there was a cut, from CNY 0.12 to CNY 0.10 for FY2022, but it is three years past and the payout has since been raised twice to CNY 0.17; no public shareholder complaint about the payout was located. My own judgment: a 31.03% payout is adequate as-is for a capital-intensive infrastructure operator running ongoing expansion capex (the Zhangji project alone was a CNY 12.1bn total investment) — this stays borderline, not fired. (4) Non-answers on capital allocation, related parties or loss-making subsidiaries — BORDERLINE, upgraded from "not fired": the original clean reading rested on the company having answered 21 investor questions on the exchange's e-interaction platform in the half at a 100% response rate, plus results briefings and institutional meetings. Correction: a response-rate count is a measure of activity, not of substance — nobody actually read the content of those 21 answers to confirm they addressed capital allocation, related parties or loss-making subsidiaries rather than deflecting. Downgraded to borderline until the answers themselves are read. (5) Outsized pay or off-market related-party deals — BORDERLINE, not fired: the volume of related-party dealing is very large, but every item is disclosed with a stated pricing basis at or better than market, the finance-company deposit rate is contractually floored at the central-bank benchmark and at what other group members receive, and no evidence of off-market pricing was found. Individual director and officer compensation was not independently verified, which is part of why this item stays borderline rather than clear.

Two warnings fire, three sit borderline, none is clear. Governance Risk: High (2 fired).

5. Corporate Ownership, Subsidiaries, and Joint Ventures

Ownership at 30 June 2026 — one class of A-shares, one vote per share, 2,335,407,014 shares in issue and unchanged in the period:

Top holders: Jiangxi Provincial Communications Investment Group, the state-controlled parent, 1,109,774,225 shares (47.52%); Great Wall Life Insurance own funds, 159,985,682 shares (6.85%); Jiangxi Provincial Port and Shipping Construction Investment Group, a state legal person, 116,536,810 shares (4.99%); Wang Weizhen, a domestic natural person, 27,946,225 shares (1.20%); Huang Guozhen, a domestic natural person, 27,740,000 shares (1.19%); Hong Kong Securities Clearing Company (the Stock Connect nominee), 23,253,574 shares (1.00%); Great Wall Life Insurance universal-life account, 13,258,410 shares (0.57%); Huang Guoming, 8,444,659 shares (0.36%); Chen Guixing, 8,222,052 shares (0.35%); and Jiangxi Highway Development, wholly owned by the controlling shareholder and acting in concert with it, 7,379,960 shares (0.32%). Directors and senior management hold nil as a group.

Principal subsidiaries at 30 June 2026, all acquired by incorporation: Jiangxi Changtai Expressway (Changtai and Zhangji roads), 76.67% owned, assets CNY 9,730.7m, liabilities CNY 3,540.3m, H1 net profit CNY 185.8m, minority interest CNY 1,461.4m. Jiangxi Changtong Expressway (Changfeng and Fengtong roads), 57.69% owned, assets CNY 5,242.3m, liabilities CNY 4,879.3m, H1 net loss CNY 97.8m, minority interest CNY 153.6m. Jiangxi Fangxing Technology (smart transport, systems integration), 78.59% owned, assets CNY 1,754.9m, H1 revenue CNY 158.1m against CNY 468.3m a year earlier, net profit CNY 1.7m. Jiangxi Ganyue Industrial Development (service-area fuel retailing), 100% owned, 13 service-area filling-station pairs and one urban station. Shanghai Jiarong Investment Management (financial investment), 97.22% direct and 2.78% indirect, holds the fund and private-equity positions. Jiangxi Jiayuan Construction Development (property and building construction), 100% owned, became wholly owned and was renamed in September 2025. Jiangxi Jiaheng Industrial (road maintenance, asphalt supply), 95% direct and 5% indirect.

Minority interests are material and are excluded from tangible book by line choice: CNY 1,893.9m of the CNY 21,129.4m of total equity belongs to outside holders of Changtai, Changtong and Fangxing, and the CNY 19,235.5m used in the bridge is the parent share only. The structure carries two specific leakage points. Changtai, the strongest subsidiary, is 23.33% owned by outsiders, so nearly a quarter of the newly widened Zhangji economics never reaches the listed shareholder. Changtong is the mirror image — 42.31% minority, equity of CNY 363.1m against registered capital of CNY 4,500m, and a CNY 1,989.9m shareholder loan advanced pro rata by the controlling shareholder. Security is specific rather than structural: 15% of Changjiu toll income is pledged against a CNY 1.2bn Agricultural Bank facility and Changtai has pledged the Zhangji toll right and all its proceeds against a CNY 9.7bn syndicated facility, so the listed parent's residual claim on those two roads sits behind pledged revenue streams. The interim report records no significant restriction on using group assets or settling group debts, and no joint ventures material enough to require separate financial disclosure — associates are disclosed only in aggregate at a carrying value of CNY 1,070.7m.

6. Historical Financial Quality and Normalized Owner Earnings

Seven fiscal years are sourced from the company's own annual reports and summaries. The record is that of a stable regulated utility with a volatile tail: toll revenue barely moves, operating cash flow is large and reliable, and reported net income swings on property-cycle write-downs and securities marks.

The chart covers revenue, net income, operating cash flow, free cash flow, diluted EPS, dividend per share, tangible book per share and ROE across FY2019-FY2025. Two further lines the filings disclose only from FY2022 onward: gross margin 32.1%, 32.0%, 39.3%, 37.7% (FY2022-FY2025), and operating income (CNY millions) 1,095.9, 1,814.3, 2,050.6, 1,944.6 over the same four years; capex over the same window was 1,050.8, 2,306.5, 2,426.4, 1,690.0 (CNY millions), and total dividends paid, from FY2021, were 280.2, 280.2, 233.5, 373.7, 397.0 (CNY millions).

Three things stand out. Operating cash flow has been between CNY 1.9bn and CNY 3.4bn every year for seven years and has never come close to threatening the dividend. The revenue line is not a toll line — the FY2023 peak of CNY 7,492.2m and the FY2024 fall to CNY 5,985.0m are the property and engineering segments moving, not traffic; the reported jump in gross margin from 32.0% to 39.3% between those years is the same effect in reverse, low-margin construction and property revenue leaving the mix, and it should not be read as toll pricing power. And tangible book per share has compounded steadily from CNY 6.73 to CNY 8.30 across the window, a 3.5% annual rate, which is what retaining roughly two-thirds of a mid-single-digit ROE produces.

Owner earnings. The cash-flow statement is the right starting point and it was read in the primary filings, so no proxy is needed for the top line. FY2025 operating cash flow was CNY 2,974.9m and total capex CNY 1,690.0m, giving reported free cash flow of CNY 1,284.9m. That understates owner earnings because most of the capex is expansion, not maintenance: the Zhangji and Changzhang widenings absorbed the bulk of the CNY 2,306.5m and CNY 2,426.4m spent in FY2023 and FY2024 and are now complete or near-complete. Road maintenance is expensed through cost of sales rather than capitalised under this company's policy, so maintenance capex is genuinely small; taking it at roughly CNY 300–400m a year gives normalized owner earnings of about CNY 2,575–2,675m against a market capitalisation of CNY 9,014.7m [ESTIMATED — the filings do not split maintenance from expansion capex, and this is the single largest judgement in the section]. Even on the conservative reported-FCF figure the FY2025 free cash flow yield on today's market value is 14.3%, and the CNY 397.0m dividend was covered 3.2 times.

Normalized EPS. The house basis is a ten-year average with a trajectory switch. Ten years were not sourced; seven were, from FY2019 to FY2025, and the switch tests are computed on that window. The five-year average of attributable profit is CNY 1,064.8m against a seven-year average of CNY 994.8m — a ratio of 1.07, inside the 1.25 "growing" trigger and well above the 0.80 "declining" trigger, so the trajectory is stable and the full-window average governs. Normalized EPS is therefore CNY 994.8m ÷ 2,335,407,014 shares = CNY 0.4260, and normalized P/E at CNY 3.86 is 9.06×. Two sanity checks sit either side of it: trailing twelve-month attributable profit is CNY 723.4m (FY2025 plus H1 2026 less H1 2025), or CNY 0.3098 per share for a trailing P/E of 12.5×, depressed by the securities mark; and first-half profit before non-recurring items annualises to about CNY 0.46 per share, an 8.4× multiple. The normalized figure sits between them, which is what it should do.

Earnings quality is good at the cash level and poor at the reported level. Operating cash flow has exceeded net income in every year shown. Working capital is not a trap — inventory of CNY 2,017.3m is unsold property being run down, and receivables of CNY 384.3m on CNY 2.66bn of half-year revenue are trivial. The quality problem is the opposite one: reported profit contains a large, unhedged, unrelated equity-market exposure, and in the first half of 2026 that exposure alone took CNY 466.2m off pre-tax profit.

6.5 Piotroski F-Score — 6/9

Periods compared: FY2025 vs FY2024.

Nine Piotroski signals, FY2025 against FY2024: positive net income, passes (NI CNY 1,279.5m). Positive operating cash flow, passes (CFO CNY 2,974.9m). Return on assets improved, fails (3.4% against 3.6%). Earnings quality, CFO exceeding net income, passes (CNY 2,974.9m against CNY 1,279.5m). Leverage did not increase, fails on the corrected construction (non-current liabilities over average assets, 27.06% against 23.34%; see the note below). Liquidity improved, passes (current ratio 1.32 against 1.02). No dilutive share issuance, passes (net buyback 0.0). Gross margin improved, fails (37.7% against 39.3%). Asset turnover improved, passes (turnover 0.169 against 0.167).

Conventions: ROA and asset turnover scaled by beginning-of-year assets; leverage on Piotroski's canonical construction, non-current liabilities over average total assets; signal 7 computed on net buyback cash flow.

Score: 6/9. The original leverage signal (5) was computed on a non-canonical total-debt-over-assets basis (32.6% vs 34.5%, a false pass); recomputed on Piotroski's actual construction — non-current liabilities over average assets — leverage rose 23.34% to 27.06% and the signal fails. The correction cuts both ways worth stating plainly: the same FY2025 current-ratio improvement used to pass signal 6 (1.02 to 1.32) is largely the mirror image of the same event — CNY 1,611.6m of liabilities moved from current to non-current while total liabilities fell only CNY 332.3m — so the liquidity "improvement" and the leverage "deterioration" are substantially one balance-sheet reclassification counted twice with opposite signs, not two independent facts. Both signals had already reversed by the 2026-06-30 anchor date used everywhere else in this report: the current ratio round-tripped to 1.0312 and the asset-liability ratio rose a further 1.14 points to 44.36%. The three failures are roa_improved, leverage_did_not_increase and gross_margin_improved. House status is eligible (4–9 band, no ranking inside it). The floor rule did not fire.

7. Balance Sheet, Debt, Covenants, and Refinancing Risk

The borrowings note, the bonds note, the long-term payables note and the financing-liabilities rollforward were all read in the 2026 interim report, and the instrument list below is complete to the balance-sheet date. Financial borrowings are disaggregated from lease liabilities before any classification is made: financial borrowings CNY 12,321.7m, lease liabilities CNY 45.4m (CNY 28.4m non-current plus CNY 17.0m current). The rollforward in the cash-flow note cross-foots to CNY 12,824.1m of financing liabilities including leases and the CNY 457.0m dividend payable, which reconciles to the instrument list exactly.

Debt instruments, unsecured unless noted (CNY millions): 26 Ganyue SCP001 super short-term paper, 1,003.9 at 1.48%, maturing 2026-12-21. 26 Ganyue SCP002, 801.5 at 1.39%, maturing 2027-02-08. Accrued SCP interest and other, 1,010.4, within 12 months, included in other current liabilities. 24 Ganyue MTN001, 898.7 at 2.17%, maturing 2027-04-24, annual coupon, bullet. 24 Ganyue MTN002 (rural revitalisation), 698.9 at 2.10%, maturing 2027-06-11, annual coupon, bullet. 24 Ganyue MTN003, 1,299.1 at 2.05%, maturing 2027-07-31, annual coupon, bullet. 25 Ganyue MTN001, 1,099.7 at 1.70%, maturing 2028-07-09, annual coupon, bullet. Agricultural Bank fixed-asset loan for the Changjiu widening, 1,045.0, LPR 5-year-plus less a spread, maturing 2037 (a 15-year term from 2022), with 15% of Changjiu toll income pledged. ICBC fixed-asset loan for the Zhangji widening and Changtai, 1,608.0, LPR 5-year-plus less a spread, maturing 2042 (an 18-year term from 2024), drawn from a CNY 5,000m facility. Bank of China fixed-asset loan for the Changzhang phase II, 226.4, LPR 5-year-plus less a spread, maturing 2038 (a 15-year term from 2023), drawn from a CNY 1,500m facility. Other credit long-term borrowings, 725.8, LPR-linked, various maturities, unsecured. Short-term borrowing from the group finance company, 4.7, rate not disclosed, maturing 2026-12-15, related party. Parent group-borrowing pass-through for the Changjiu widening, 554.0 at 1.20% fixed, maturing 2029 (a 14-year term from 2015/16), on-lent from national policy banks via the parent. Shareholder loan to Changtong from the controlling shareholder, 1,989.9, rate not disclosed, stated maturity 2026-09-07, advanced pro rata and carried as non-current. Accrued interest on long-term borrowings and payables, 81.5, within 12 months. Total financial borrowings: 12,321.7. Lease liabilities (current CNY 17.0m plus non-current CNY 28.4m), 45.4, implicit rate, to 2030, kept out of the debt test. Cash and equivalents, 3,307.9, including CNY 811.7m at the parent finance company and CNY 3.3m restricted. Listed and fund securities at fair value, 1,418.8 (CNY 892.5m FVTPL, CNY 198.0m FVOCI, CNY 328.3m other).

Also issued after the balance-sheet date and therefore not in the table: 26 Ganyue MTN001, CNY 300m at 1.51%, issued 27 July 2026 and maturing 27 July 2028, and 26 Ganyue MTN002, CNY 500m at 1.55%, issued 30 July 2026 and maturing 30 July 2029. Two super short-term papers, 25 SCP006 and 25 SCP007, CNY 500m each, were repaid on 3 August 2026. A CNY 12.0bn unified debt-financing instrument registration was obtained in the half — the company says it is the first local enterprise in Jiangxi and in the expressway sector to hold it.

Ratios, on the company's own disclosed basis: total liabilities to assets 44.36% at 30 June 2026 (43.22% at end-2025); current ratio 1.03 (1.32); quick ratio 0.78 (0.99); EBITDA interest cover 9.97× in the half against 16.73× a year earlier; EBIT interest cover 4.23× against 11.27×, the fall being entirely the securities mark inside pre-tax profit; cash interest cover 16.68×, essentially unchanged. Loan repayment and interest payment rates are both 100%, no interest-bearing debt is overdue, and the AAA issuer rating with stable outlook has been maintained. On financial borrowings alone the ratio is 32.4% of assets and 58.3% of total equity.

Net debt to EBITDA: on FY2025 adjusted EBITDA (CNY 2,843.4m, stripping the non-operating securities fair-value gain — discussed further in the valuation section below), gross debt to adjusted EBITDA is 4.33× and net debt to adjusted EBITDA (net of all CNY 3,307.9m of cash) is 3.17×. The adjustment that matters: CNY 811.7m of that cash sits on deposit at the controlling shareholder's captive finance company — the same fact cited as evidence for the second fired capital-allocation warning. Netting it against debt while correctly refusing to net the related-party equity stake in the securities book is inconsistent; excluding that deposit from cash, net debt to adjusted EBITDA is 3.46×, at or through the boundary this report elsewhere classifies as "Elevated." Whether the finance-company deposit should count as cash for this purpose is an economic-substance question no filing settles either way.

Classification: Manageable. Debt to assets sits at the low end of the 30–50% band and coverage at the top of the 5–10× band on the EBITDA measure, with no covenant breach, no overdue instrument and a demonstrated ability to refinance at falling rates. The capital-intensity caveat applies explicitly and is not a waiver: regulated toll-road infrastructure with contracted, government-set pricing structurally supports more leverage at the same risk than an industrial would, and a 32% debt-to-asset ratio here is conservative for the sector rather than merely adequate — but the qualification filter debt test is still run on the unadjusted number, and it passes on both the financial-borrowings measure (58.3% of equity) and the wider total-liabilities measure (79.7% of total equity, 87.6% of parent equity), so no sector allowance is needed to clear it.

The maturity wall is the item to watch and it is concentrated in 2027: CNY 2,896.7m of medium-term notes fall due between April and July 2027, on top of CNY 1,805.4m of super short-term paper maturing in December 2026 and February 2027. Rolling that at current interbank pricing is not a stretch for an AAA name that just printed 1.51%, but it is the reason cash rose to CNY 3,307.9m in the half. One classification item deserves flagging: the CNY 1,989.9m shareholder loan to Changtong is carried within non-current long-term payables while the related-party note gives its maturity as 7 September 2026 — nine weeks after the balance-sheet date. Either an extension is assumed and undisclosed, or the classification is wrong; the filings do not say which, and this is a diligence gap rather than a solvency concern given the lender is the controlling shareholder.

8. Real Estate, Leases, and Hidden Assets

Right-of-use assets are CNY 44.4m, which is 0.2% of tangible book value of CNY 19,121.3m, against lease liabilities of CNY 45.4m; total lease cash outflow in the half was CNY 11.4m and short-term and low-value lease expense CNY 6.9m. Right-of-use assets are a capitalised right to occupy, not a liquidation floor, and they are disclosed here rather than stripped out of the TBV per share used everywhere in this report — removing the asset without its matching liability would be the wrong operation. At 0.2% the question is immaterial either way, which is the point worth making: unlike an asset-light retailer whose "tangible book" is largely leasehold, this company's book is road.

What the book actually consists of is CNY 26,400.4m of owned property, plant and equipment, of which road assets carried CNY 18,402.0m at 31 December 2025 before the CNY 5,278.7m Zhangji transfer in the half. Those assets are depreciated by traffic volume over the estimated remaining tolling period rather than on a straight-line life, and they are carried at depreciated historical cost. Two sources of value sit outside that number. First, the pending re-determination of the Zhangji tolling period: the company assumes a 30-year re-determined term in its own accounting-estimate work, and an extension granted on a road already built and paid for is close to pure incremental value that book equity does not carry. Second, investment property is held at cost — CNY 165.9m net against CNY 328.0m of original cost — and the Changtai Building in Ji'an is let to the municipal engineering company on a lease running to 31 July 2030 at rents that escalate 3.56% a year, while Jiarong lets a floor of the Zhide Xingzuo tower in Beijing's Xicheng district to 28 February 2030.

The genuine downside support is not exotic. Cash and equivalents are CNY 3,307.9m with only CNY 3.3m restricted; liquid listed and fund securities add CNY 1,418.8m at market; inventory of CNY 2,017.3m is completed and near-completed property being sold down, which is the least reliable of the three and is where the CNY 29.2m of impairment charged in the half landed. Against CNY 12,321.7m of borrowings, the CNY 4,726.7m of cash and marketable securities is the first line of defence and the toll cash flow is the second.

9. Capital Markets Access, Dilution, and Financing Flexibility

The share count has not moved. There were 2,335,407,014 shares in issue at each of the seven fiscal year-ends from FY2019 to FY2025 and at 30 June 2026, and the interim report states that total shares and share structure were unchanged in the period. There are no options, no restricted stock units, no employee share plan, no convertible bonds — the interim report's convertible-bond section is marked not applicable — no warrants and no treasury shares. Basic and diluted earnings per share are identical in every period shown, which is what a capital structure with no dilutive instruments produces. There has never been a buyback either, so the shareholder yield is entirely the dividend.

The company does not need the equity market. Financing runs through bank facilities and the interbank debt-financing market, where it holds an AAA rating, a newly registered CNY 12.0bn unified programme, and undrawn committed bank capacity — CNY 3,392m of the ICBC Zhangji facility and CNY 1,273.6m of the Bank of China Changzhang facility remain undrawn on the disclosed drawdown figures, alongside a CNY 9.7bn syndicated facility for Zhangji signed in 2023 that is barely used. The one equity-adjacent event in the window was a CNY 58.6m minority injection into a subsidiary in FY2024, which dilutes the group's share of that subsidiary rather than the listed share count. Dilution risk is as close to nil as a listed company gets.

10. Litigation, Regulatory, and Contingent Liability Risk

The note reached is Note XVI, Commitments and Contingencies, of the 2026 interim report, read in the primary filing, together with Section V items 6 to 9 of the same report. Commitments are explicitly nil: "as at 30 June 2026 the company had no significant commitments requiring disclosure." Section V item 7 records that there were no material litigation or arbitration matters in the period, item 8 that neither the company nor its directors, senior managers, controlling shareholder or actual controller were subject to investigation or penalty, and item 3 that there were no irregular guarantees.

One contingency is disclosed and it is a carry-forward. In July 2022 the fuel subsidiary sued Sichuan Zhonghangyou Xinye Petroleum and a second defendant for return of CNY 143,000,000 of prepayments on four fuel purchase contracts, with a claim that the second defendant be jointly liable to the extent of CNY 55m of funds it had allegedly moved. The Nanchang Intermediate Court ruled for the company on 23 April 2023, the Jiangxi High Court dismissed the appeal on 16 December 2023, and in February 2024 CNY 55.09m of enforcement proceeds net of fees reached the subsidiary. In June 2024 the Supreme People's Court ordered the case up for retrial and suspended enforcement; in July 2024 the Nanchang court terminated enforcement; the Supreme People's Court heard the retrial in July 2025 and, as at the date the interim financial statements were signed, judgment is still awaited. The exposure runs both ways: the unrecovered balance of the CNY 143m claim, and a possible clawback of the CNY 55.09m already collected if the retrial goes against the company. Provisions carried are small and specific — CNY 21.1m in total, comprising CNY 4.3m for unresolved litigation, CNY 15.9m of construction warranty retentions and CNY 0.9m of late-delivery compensation on property.

Carry-forward checks against the FY2025 annual report found no other open items: the Fangxing spin-off, worked on from March 2021, was formally stopped by board resolution on 21 March 2025 and is closed; there is no subpoena, investigation, class action, covenant waiver, going-concern language or material weakness in either filing; and the subsequent-events note of the interim report is marked not applicable in all four of its sub-items, meaning no material non-adjusting event was disclosed even though the company issued CNY 800m of medium-term notes in late July 2026 and repaid CNY 1,000m of paper on 3 August 2026 — those appear in the bond section rather than in the subsequent-events note, which is a presentation point rather than an omission. Regulatory risk is not litigation risk here: the material regulatory exposure is the government's power to set toll rates, tolling periods and charging methods, which the company names as its second-ranked risk factor and which the Third Plenum decision to "optimise toll road policy" keeps live.

11. Accounting Quality and Disclosure Review

The FY2025 financial statements were audited by Zhongshenzhonghuan CPAs (special general partnership), report number Zhonghuan Shen Zi (2026) 1500011, signed by Cai Suhua as engagement partner, and carry a standard unqualified opinion. The audit report was read in the primary filing. There is no going-concern material-uncertainty paragraph — only the boilerplate responsibility wording — and no material weakness is reported. The half-year statements are unaudited, which the company states plainly.

There is exactly one key audit matter and it is the right one: depreciation of road assets. The auditor records road-asset carrying value of CNY 18,402.0m at 31 December 2025 and FY2025 road depreciation of CNY 953.0m, notes that the units-of-traffic method depends on management's estimate of total remaining traffic over the tolling period, and identifies significant management judgement and potential bias because the forecast depends on the history of parallel expressways and national and provincial roads and on other road-network construction in the same region. The procedures described are substantive rather than nominal: benchmarking the policy against comparable listed expressway operators, assessing the independence and competence of the third-party traffic forecaster, back-testing prior forecasts against actual traffic, agreeing actual traffic to external data and recomputing the charge.

That estimate moved in the period under review, which is why it matters now. On 7 August 2026 the board approved a change of accounting estimate: the Ministry of Transport Planning and Research Institute re-forecast traffic for all eight roads and new unit-traffic depreciation rates apply from 1 January 2026, while from 1 February 2026 the Zhangji depreciation base is reset on an assumed re-determined tolling period of 30 years. The disclosed effect on the first half is accumulated depreciation and cost of sales up CNY 46.47m, tax down CNY 8.53m, net profit down CNY 37.94m and attributable profit down CNY 29.10m; had the same estimate applied from 1 January 2023, net profit would have been lower by CNY 115.7m, CNY 120.7m and CNY 124.7m in 2023, 2024 and 2025 respectively. The direction is conservative — the company took more depreciation, not less — and the pro-forma disclosure is unusually complete. It also means reported FY2023–FY2025 earnings were, on today's traffic assumptions, roughly CNY 120m a year better than they would now be stated; the normalized EPS used in this report is not adjusted for that, which biases the normalized P/E modestly optimistic.

Elsewhere the disclosure is detailed. Non-recurring items are itemised line by line — the CNY -328.7m first-half total is dominated by CNY -468.5m of financial-asset fair-value and disposal effects, partly offset by CNY 20.2m of government grants and CNY 111.5m of tax effect — so the recurring and non-recurring halves of the result can be separated by the reader without estimation. Segment reporting gives revenue, cost, investment income, impairment, depreciation, pre-tax profit, tax, net profit, assets and liabilities for five segments plus eliminations. Related-party disclosure is exhaustive to the point of listing individual balances with fifty counterparties. Goodwill is CNY 3.6m, so impairment risk to book value from that source is nil; other intangibles are CNY 110.6m; deferred tax assets of CNY 406.5m are the larger judgement and are not separately stress-tested in the filings. Revenue recognition is unremarkable — tolls at the point of passage, construction over time, property on handover.

Classification: Adequate. Every material note was opened in a primary document and the opinion is clean, which supports the top of the scale; what holds it at Adequate rather than High is that the single most important number in the accounts — road depreciation — rests on a traffic forecast that management commissions, that has just been revised in a way that would have cut three prior years by about CNY 120m each, and whose sensitivity is not disclosed.

12. Valuation and Margin of Safety

Valuation method. This section's price-target construction (Buy-Below, Sell Range, Add-More Trigger) has been corrected in two ways since this report was first drafted. First, the self-built seven-point annual P/TBV series used below has been replaced with the same 41-quarter own-history series used everywhere else on this position (2016 to 2026, quarterly), which is simply a longer and more reliable version of the same idea. Second, Buy-Below itself now uses a different, company-relative construction, corrected 2026-08-29: the original universe-wide grid reading landed Buy-Below above the bottom of this stock's own Sell Range, an incoherent pair of numbers for a name that has traded almost entirely inside the "cheap" band across its own history. The corrected figures are below; the reasoning that produced them replaces the equivalent reasoning in the earlier analysis.

Tangible Book Value — Derivation (anchor 2026-06-30, 2026 Interim Report; CNY millions)

Table 1 — equity bridge. Common shareholders' equity: CNY 19,235.5m (2026 Interim Report, consolidated balance sheet, equity attributable to parent, 19,235,530,185.53), as of 2026-06-30. Less goodwill: CNY -3.6m (note 7(27), goodwill 3,609,824.44). Less other intangibles: CNY -110.6m (note 7(26), intangible assets 110,592,472.20). Equals tangible book value: CNY 19,121.3m (derived: common equity minus goodwill minus other intangibles). Divided by shares outstanding: 2,335.407m (2026 Interim Report, share capital 2,335,407,014; the interim report's own share-capital disclosure confirms no change in the period). Equals TBV per share: CNY 8.1876 (derived). Minority interest, excluded by line choice and not subtracted: CNY 1,893.9m (2026 Interim Report, minority interests 1,893,850,794.32).

No analyst adjustments are made, so adjusted TBV per share equals TBV per share at CNY 8.1876.

P/TBV (derived) = 3.86 ÷ 8.1876 = 0.471× (price as of 2026-08-28).

Table 2 — asset composition at the anchor period. Cash and short-term investments: CNY 4,200.4m, 11.1% of total assets, 22.0% of TBV. Accounts receivable: CNY 384.3m, 1.0%, 2.0%. Inventory: CNY 2,017.3m, 5.3%, 10.6%. Property, plant and equipment (owned): CNY 26,400.4m, 69.5%, 138.1%. Right-of-use lease assets, disclosed and never stripped from TBV: CNY 44.4m, 0.1%, 0.2%. Long-term investments: CNY 1,596.9m, 4.2%, 8.4%. Goodwill: CNY 3.6m, 0.0%, 0.0%. Other intangibles: CNY 110.6m, 0.3%, 0.6%. Other current and noncurrent assets: CNY 3,209.3m, 8.5%, 16.8%. Residual/unclassified: CNY 6.1m, 0.0%, 0.0%. Total assets: CNY 37,973.5m (100%).

Right-of-use note: ROU assets 44.4 = 0.2% of TBV. ROU is a capitalized right to occupy, not a liquidation floor — disclosed, never stripped out of the TBV/share used everywhere in this report.

Road-only asset composition. The "Property, plant & equipment (owned)" row above (26,400.4 / 69.5% of assets / 138.1% of TBV) is TOTAL owned PP&E — buildings, equipment and land alongside the roads, not a road-only figure. An earlier characterisation of this position as "roughly 70% of assets and 138% of tangible book" in road PP&E conflated the two. The only audited road-specific carrying value in this record is CNY 18,402.0m at FY2025 — 48.8% of assets and 94.9% of tangible book (no road-only figure exists at the 2026-06-30 anchor date used for the bridge above). The remaining roughly 18% of tangible book not accounted for by roads is the related-party brokerage stake, other fair-value-through-OCI equity, and property inventory.

Historical Relative Multiple Valuation

An annual seven-point P/TBV check is retained as supplementary context; the governing series is the company's own quarterly P/TBV history — 41 observations, 2016-06-30 to 2026-06-30.

Full 41-quarter distribution: min 0.454×, p25 0.519×, median 0.615×, p75 0.681×, p90 0.829×, lower-quartile mean (mean of every observation at or below p25) 0.491×, upper-quartile mean (mean of every observation at or above p75) 0.805×, max 0.902×. Today's 0.471× sits at roughly the 5th percentile of this series — still near the bottom of its own ten-year range, though not the literal minimum (that was 0.454×, in Q3 2022).

Annual check (supplementary):

Year-end price, TBV per share, and P/TBV, in CNY: 2019-12-31, 4.13, 6.6693, 0.62×. 2020-12-31, 3.42, 6.7438, 0.51×. 2021-12-31, 3.43, 6.9119, 0.50×. 2022-12-31, 3.39, 7.1435, 0.47×. 2023-12-31, 4.11, 7.5023, 0.55×. 2024-12-31, 5.61, 7.9356, 0.71×. 2025-12-31, 5.19, 8.2986, 0.63×. Today (2026-08-28): 3.86, 8.1876, 0.471×.

The annual points sit inside the fuller quarterly range and would have understated the true floor (0.454×, a quarter this coarser series never sampled). The quarterly series is the governing history.

Sell Range and Buy-Below

Sell Range is [median, upper-quartile-mean] × current TBV/share, over the same 41-quarter series: CNY 5.03 – 6.59 (0.61× – 0.81× TBV). Neither coherence flag fires: the range does not start below the current price, and the current price is not above it — today's price sits well below the whole range.

Buy-Below uses the same 41-quarter distribution as Sell Range. The lower-quartile mean is CNY 4.02 (0.49× TBV, 10.2-yr own-history). The result sits below the Sell Range floor by construction; the universe-wide matrix still governs the sizing tier and Buy More Below.

Buy More Below: not reachable — Big is already the largest size on the matrix, so no next size up exists at any price.

Add-More Trigger (new): CNY 2.64 (0.32× TBV). A single, entry-anchored review price for adding to this specific position — entry CNY 3.87, 2026-08-26, minus one 0.15-P/TBV-point step — distinct from Buy-Below/Buy More Below, which read off the universe-wide grid from today's price rather than this position's own entry point. It is a guideline for revisiting the thesis if the price gets there, never a mechanical buy signal, and never applied to sizing or the tier.

At CNY 3.86, the stock trades below its own Buy-Below (4.02) and well below the bottom of its own Sell Range (5.03), with a real gap between the two.

Scenario Table

The anchor for every row is the company's own P/TBV window and, on the earnings leg, an EV/EBITDA cross-check: enterprise value at CNY 3.86 is CNY 19,922.4m (market capitalisation CNY 9,014.7m plus financial borrowings CNY 12,321.7m plus minority interests CNY 1,893.9m less cash CNY 3,307.9m). EBITDA basis. Reported FY2025 EBITDA of CNY 3,130.1m includes a CNY 286.7m pre-tax fair-value gain on the securities book that this report treats as non-operating everywhere else; only the mark-inclusive construction prints the 6.4× multiple below. Stripping the mark gives adjusted EBITDA of CNY 2,843.4m and EV/EBITDA of 7.01× (a second, independent construction from an operating-income line lands close by, at 6.77×) — still the cheap end of the sector's normal 7×–10× range, just on the correct, comparable basis.

Severe downside: tolling periods shortened by policy, a further CNY 1.0bn of securities and property write-downs, multiple stays at the window floor — 0.42× TBV of CNY 7.75, giving CNY 3.26, -16% versus today's CNY 3.86. Bear: traffic flat, no Zhangji extension granted, securities book falls a further 20%, multiple stays at today's level — 0.47× TBV of CNY 8.05, giving CNY 3.78, -2%. Base: toll revenue grows 1-2%, capex normalises, securities book stable, multiple reverts to the seven-year median — 0.55× TBV of CNY 8.30, giving CNY 4.57, +18%. Bull: Zhangji tolling period re-determined at 30 years, securities book recovers, multiple reaches the top-quartile mean — 0.67× TBV of CNY 8.60, giving CNY 5.76, +49%.

Stated explicitly. Current price CNY 3.86 (2026-08-28), market capitalisation CNY 9,014.7m (USD 1,341m at USD/CNY 6.7217, the single reference rate used in this report and applied nowhere except the decision figures). Intrinsic value range, base case: CNY 4.57 – 5.45. Margin of safety, both levels stated: at the current price against tangible book per share of CNY 8.1876, the equity-level discount is 52.9%; at the enterprise level, where net debt per share (CNY 3.8596) is almost exactly the share price itself, EV/tangible capital is 0.663× and the real discount is 33.7%, about a third smaller than the headline number. Both belong in the same sentence; the 52.9% figure should never travel alone. Buy-Below: CNY 4.02 (0.49× tangible book, own-history basis) — see above. Sell Range: CNY 5.03 – 6.59 — see above. Buy More Below: not reachable — Big is already the largest size on the matrix, so no next size up exists at any price. Add-More Trigger (held position only): CNY 2.64 — new, see above. Cheapness type: asset-value cheap, with a statistically cheap overlay — the discount is to a real, depreciating, cash-generating road network, not to a normalised earnings stream, and the earnings leg agrees rather than carrying the case.

13. Risk Matrix

• Controlling shareholder directs capital to non-core, related-party uses. Evidence: CNY 863.5m equity and CNY 140m subordinated debt in a related-party securities house; CNY 811.7m of cash at the parent finance company; CNY 49.97m/yr paid to the parent to operate the company's own roads. Severity high, probability high (already occurring). Financial impact: CNY 466.2m fair-value loss in H1 2026 alone. Valuation impact: sustains the 0.47× discount, the main reason the tier is Standard not Big. Mitigant: disclosed, capped and priced at stated market floors, 100% investor-question response rate. Monitor: the FY2026 related-party transaction cap and any increase in the securities book.

• Toll policy change — rates, tolling periods or charging method. Evidence: the company names it as risk factor 2; the Third Plenum decision to "optimise toll road policy"; the 1998 non-parallel-road undertaking already allowed to lapse. Severity high, probability medium. Financial impact: the toll segment is CNY 1,822.6m of H1 revenue at 51% margin. Valuation impact: a shortened tolling period cuts the depreciable life and the asset value directly. Mitigant: provincial control of the operator cuts both ways; Changjiu and Changzhang extensions already granted. Monitor: the Zhangji tolling-period application and any provincial toll-rate consultation.

• Securities-book volatility swamping reported earnings. Evidence: H1 2026 attributable profit -72.81% while recurring profit -14.27%; CNY 466.2m fair-value loss; CNY 97.7m OCI loss on Xinda. Severity medium, probability high (recurring). Financial impact: up to about CNY 1.09bn of listed equity marked each period. Valuation impact: keeps the P/E leg unusable and depresses sentiment. Mitigant: non-cash; recurring profit and the dividend are unaffected. Monitor: Guosheng Securities share price and any disposal announcement.

• 2027 maturity wall. Evidence: CNY 2,896.7m of MTNs due April-July 2027 plus CNY 1,805.4m of paper due December 2026 to February 2027. Severity medium, probability low. Financial impact: refinancing cost; CNY 12.0bn programme registered. Valuation impact: minimal while AAA holds. Mitigant: AAA rating; July 2026 notes priced at 1.51%/1.55%; cash CNY 3,307.9m. Monitor: rating actions and interbank spreads at each roll.

• Loss-making Changtong subsidiary. Evidence: equity CNY 363.1m against registered capital CNY 4,500m; H1 net loss CNY 97.8m after CNY 44.2m; CNY 1,989.9m shareholder loan. Severity medium, probability high (already occurring). Financial impact: roughly CNY 100-200m a year of drag at the group level. Valuation impact: removes part of the toll cash flow the discount is priced against. Mitigant: 42.31% of the loss accrues to minorities; the roads themselves are young. Monitor: the half-yearly Changtong net result and any recapitalisation.

• Sichuan Xinye retrial outcome. Evidence: Supreme People's Court retrial heard July 2025, judgment awaited; CNY 55.09m already collected and enforcement terminated. Severity low, probability medium. Financial impact: up to a CNY 143m claim value plus a possible CNY 55.09m clawback. Valuation impact: under 1% of tangible book. Mitigant: two lower courts already found for the company. Monitor: the Supreme People's Court judgment.

• Traffic-forecast estimate revision. Evidence: the sole key audit matter; the 2026 revision cut H1 net profit CNY 37.94m and would have cut 2023-25 by about CNY 120m a year. Severity medium, probability medium (revised every three years). Financial impact: CNY 953.0m of FY2025 road depreciation is set by this estimate. Valuation impact: a downward traffic revision lowers earnings without touching cash. Mitigant: third-party forecaster; auditor back-tests prior forecasts. Monitor: the next three-yearly re-forecast and monthly toll revenue announcements.

• Governance discount is permanent. Evidence: 47.84% state concert-party control; no buyback in 26 years as a listed company; no insider ownership at all. Severity medium, probability medium. Financial impact: none directly. Valuation impact: the 0.47× could simply persist — the value trap case. Mitigant: dividend raised twice since the FY2022 cut, payout 31.03%. Monitor: any buyback authorisation and the payout ratio at the FY2026 result.

14. Red Flags, Yellow Flags, and Green Flags

🟢 Green Flags

• Tangible book of CNY 19,121.3m against a CNY 9,014.7m market capitalisation, of which CNY 26,400.4m is owned road and only CNY 44.4m (0.2% of TBV) is right-of-use.

• Operating cash flow of CNY 1.9bn–CNY 3.4bn in each of the last seven fiscal years; FY2025 free cash flow CNY 1,284.9m covered the CNY 397.0m dividend 3.2 times.

• Dividend raised from CNY 0.10 (FY2022) to CNY 0.16, then CNY 0.17 and CNY 0.17, paid 8 July 2026; a 4.40% yield with cumulative distributions above CNY 6.7bn since the 2000 listing.

• AAA issuer rating maintained, 26 Ganyue MTN001 priced at 1.51% on 27 July 2026, and FY2025 finance costs down 32.43% to CNY 211.0m.

• Share count unchanged at 2,335,407,014 for at least seven years, with no options, no convertibles and no dilutive instruments of any kind.

🟡 Yellow Flags

• Five board and senior-management changes between 25 November 2025 and 21 July 2026, including the general manager, all explained only as "personal reasons" or "work transfer", with no successor general manager named.

• Zero director and officer share ownership, and no equity incentive plan of any kind — nobody running the company owns any of it.

• Return on equity of 6.6% in FY2025 and never above 7.1% in the seven-year window: the discount is to a low-return asset, not a mispriced compounder.

• The accounting-estimate change of 7 August 2026 would have reduced FY2023–FY2025 net profit by about CNY 120m a year had it applied then, so the normalized EPS used here is modestly flattered.

• Changtong carries equity of CNY 363.1m against registered capital of CNY 4,500m and lost CNY 97.8m in the half.

🔴 Red Flags

• Added — this was named as a thesis-killer elsewhere in this report but had dropped out of the Red Flags list itself in the earlier analysis: CNY 5,278.7m (27.6% of tangible book) was capitalised in the first half on the Zhangji road expansion, whose concession re-approval is not yet confirmed by any primary or secondary source read for this report. If refused, a large part of that CNY 5.28bn depreciates over the roughly seven years remaining on the existing Changtai concession term instead of over a fresh one.

• The largest financial asset, CNY 863.5m of Guosheng Securities stock, is in a counterparty that appears in the company's own related-party dealings table, alongside a CNY 140m subordinated loan to the same counterparty running to 2030 — and it produced a CNY 466.2m fair-value loss in six months.

• CNY 811.7m of group cash, a quarter of the total, is on deposit with the controlling shareholder's captive finance company against a CNY 850m daily cap.

• The controlling shareholder is paid CNY 49.97m a year to operate the company's own toll, maintenance, safety and rescue centres under a 2025–2027 agreement.

• The 1998 undertaking not to build parallel roads within 50 km has been allowed to lapse on policy grounds, explained publicly rather than compensated — the moat was removed by the party that controls the company.

• A CNY 1,989.9m related-party borrowing is classified as non-current while the related-party note gives its maturity as 7 September 2026.

⚡ Must-Watch Catalysts

• The application to re-determine the Zhangji tolling period, expected once the full line opens; the company's own accounting work assumes 30 years.

• The FY2026 result and dividend declaration, expected late March 2027 on the pattern of the last three years — no results date has been announced.

• Monthly vehicle toll service revenue announcements, published around the 10th to 16th of each month.

• The Supreme People's Court retrial judgment in the Sichuan Xinye matter.

• The December 2026 and February 2027 super short-term paper maturities, and the CNY 2,896.7m of MTNs falling due between April and July 2027.

FINAL VALUE-INVESTING RECOMMENDATION

Verdict: BUY — STANDARD

Current Price: CNY 3.86 (2026-08-28); market capitalisation CNY 9,014.7m (USD 1,341m at USD/CNY 6.7217). Schloss Universe Filters (qualification filter): PASS. P/TBV 0.471× against the 1.00× cutoff. Debt test on financial borrowings, CNY 12,321.7m, equals 58.3% of total equity of CNY 21,129.4m (64.1% of parent equity); on the wider total-liabilities measure, CNY 16,844.1m equals 79.7% of total equity; lease liabilities of CNY 45.4m kept separate. Both below 100%, so no sector allowance is needed. Piotroski F-Score: 6/9. Eligible. Failed signals: ROA improved, leverage did not increase (corrected to Piotroski's canonical non-current-liabilities-over-assets basis, which rose 23.34% to 27.06%), and gross margin improved. Floor rule did not fire. P/TBV (derived): 0.471× (filing TBV/share CNY 8.1876 at 2026-06-30; price CNY 3.86 at 2026-08-28). Normalized P/E: 9.06× on normalized EPS of CNY 0.4260, the seven-year average of attributable profit (CNY 994.8m), trajectory tested as stable (avg5 over avg-window 1.07). Graham-flavoured, not literally Schloss. Cash Return (dividend plus buyback yield): 4.40% — dividend CNY 0.17 per share for FY2025 paid 8 July 2026; buybacks nil. Cheapness Tier — Numbers Only (tier calculation): base cell Big (P/TBV at or below the 0.50× row, normalized P/E in the 12× or lower column), unmodified — the shareholder-yield and capital-allocation review flags are disclosed for a human to weigh and are never applied to the tier itself. The mechanical readout is Big. Capital-Allocation Warnings (capital-allocation review): 2 fired, 3 borderline, 0 clear. Fired: (1) control without offsetting minority protection — 47.84% state concert-party stake under the provincial state-asset regulator (the "no independent-director veto" evidentiary limb is withdrawn; a Board Strategy and Investment Decision Committee does exist; the warning still fires on concentration alone); (2) capital routed to parent-mandated, policy-driven or low-return projects — CNY 863.5m of equity and CNY 140m of subordinated debt in a related-party securities house, CNY 811.7m of cash at the parent finance company, CNY 49.97m a year paid to the parent to operate the company's own roads. Borderline: the FY2022 dividend cut, since more than restored (my own judgment: 31.03% payout adequate as-is); the scale of related-party dealing, disclosed and priced at stated market floors; and non-answers to shareholders, upgraded from "not fired" — the original clean read relied on a response-rate count (21 questions, 100% response), not on reading the answers for substance. Tier vs. Published Verdict: the mechanical base cell is Big. The published verdict of Standard is my own disclosed judgment call on top of that reading, not a formula output — I weigh the two fired capital-allocation warning above as real enough to hold one notch below what the matrix alone says. Intrinsic Value Range: CNY 4.57-5.45 per share (base case; 0.55× to 0.67× tangible book on the stock's own annual-observation distribution, cross-checked above against the fuller 41-quarter series). Margin of Safety: 52.9% below tangible book value per share of CNY 8.1876 at CNY 3.86 (equity level); 33.7% at the enterprise level, where net debt per share is almost exactly the share price itself. Buy-Below Price: CNY 4.02 (0.49× TBV, 10.2-yr own-history).39. Now read off the same 41-quarter own-history distribution as Sell Range (the lower-quartile mean) rather than the universe-wide matrix grid, which had landed the original figure above the bottom of this stock's own Sell Range — see the valuation discussion above. Buy More Below: not reachable — Big is already the largest size on the matrix, so no next size up exists at any price. Add-More Trigger (held position): CNY 2.64 (0.32× TBV), new, did not exist in the earlier analysis. Entry-anchored review price for adding to this specific position (entry CNY 3.87, 2026-08-26); a guideline, never a mechanical signal, never applied to sizing or the tier. Sell Range: CNY 5.03-6.59 (0.61× to 0.81× TBV).49-5.45. 41 quarterly observations, 2016-06-30 to 2026-06-30 (the earlier annual sell range used only 7 annual observations); percentiles by ascending sort with linear interpolation, upper-quartile mean over observations at or above p75. No coherence flag fires. Position Sizing Guidance: Standard — the house starting point for a Standard tier under a hypothetical general mandate, not advice to any individual holder, and subject to the single-country and sector concentration limits that apply to any one name. Expected Holding Period: three to five years — long enough for the Zhangji tolling-period application, two more dividend cycles and a mean reversion in the multiple. Downside Risk: bear case CNY 3.78, severe downside CNY 3.26 — roughly 2% and 16% below the current price. The floor is a CNY 19.1bn tangible book of owned road, with CNY 4,726.7m of cash and marketable securities against CNY 12,321.7m of borrowings. Balance-Sheet Risk: manageable but explicitly fragile — financial borrowings 32.4% of assets and 58.3% of total equity, EBITDA interest cover 9.97×, no overdue debt, AAA maintained. Gross debt to adjusted EBITDA 4.33×, net 3.17× as reported, 3.46× net excluding the related-party finance-company deposit — nearer the "Elevated" boundary than the headline coverage ratios alone convey. Creditworthiness: strong. AAA issuer rating with stable outlook maintained; 100% loan repayment and interest payment rates; new money raised at 1.39% to 1.55% during 2026. Governance Risk: high (2 fired, 3 borderline — see the Capital-Allocation Warnings above for what changed). Accounting Quality Risk: adequate — unqualified opinion from Zhongshenzhonghuan CPAs, one key audit matter (road-asset depreciation), full pro-forma disclosure of the 2026 estimate change, but the largest number in the accounts rests on a management-commissioned traffic forecast whose sensitivity is not disclosed. Refinancing Risk: low. CNY 1,805.4m of super short-term paper matures December 2026 to February 2027 and CNY 2,896.7m of medium-term notes between April and July 2027, against CNY 3,307.9m of cash, a CNY 12.0bn registered programme and several undrawn multi-billion bank facilities. Key Catalysts: Zhangji tolling-period re-determination, with 30 years assumed in the company's own accounting work; the FY2026 result and dividend, expected late March 2027 on the recent pattern though no date has been announced; monthly toll revenue disclosures; any recovery or disposal of the Guosheng Securities position; the Supreme People's Court retrial judgment. Primary Thesis Killers: a toll-policy change that shortens tolling periods or cuts rates; further large fair-value losses on the related-party securities book, or an increase in its size; the governance discount proving permanent after 26 listed years with no buyback and no insider ownership; refusal of the Zhangji extension, which would strand a large part of the CNY 5,278.7m just capitalised — narrowed but not closed: a rating agency's 2025 table (discussed earlier in the business overview) now confirms seven of the eight roads run 7 to 23 more years, but does not itself confirm the Zhangji/Changtai re-determination specifically.

VALUE INVESTOR MUST-WATCH LIST

• The Zhangji tolling-period application, filed once the full line is open; the accounting-estimate announcement of 11 August 2026 already assumes a 30-year re-determined term, so a shorter grant is both an earnings and an asset-value event.

• Monthly vehicle toll service revenue announcements, published around the 10th to 16th of each month; the run rate to beat is CNY 1,807m for the half, up 0.62%.

• The Guosheng Securities share price against the CNY 11.00 mark of 30 June 2026 on 78,500,053 shares — every CNY 1.00 of movement is CNY 78.5m of pre-tax profit.

• Any change in the size of the financial-investment book, or a disposal of the Guosheng position; an increase would be a direct escalation of the second capital-allocation warning.

• The CNY 811.7m balance held at the controlling shareholder's finance company against its CNY 850m daily cap, and the annual continuing risk-assessment report on that finance company.

• Appointment of a general manager — the post has been vacant since Xu Zhihua resigned on 27 May 2026.

• The FY2026 dividend declaration expected with the annual report in late March 2027: CNY 0.17 held or raised confirms the policy, while a cut would fire the third capital-allocation warning and take the tier down again.

• The CNY 1,989.9m related-party borrowing whose stated maturity is 7 September 2026 while it is carried as non-current; the FY2026 annual report should resolve the classification.

• The December 2026 and February 2027 super short-term paper maturities and the CNY 2,896.7m of medium-term notes due April to July 2027, and the pricing achieved on each roll.

• Changtong's half-yearly result — a CNY 97.8m loss in H1 2026 after CNY 44.2m a year earlier, on equity of CNY 363.1m.

• The Supreme People's Court retrial judgment in the Sichuan Xinye matter, including any clawback of the CNY 55.09m already received.

• The next three-yearly traffic re-forecast and any interim revision to the unit-traffic depreciation rate.

• Corrected: CNY 5.03, the bottom of the corrected Sell Range where the sell question becomes live, and CNY 6.59, the top, above which the shares trade beyond their own demonstrated top-quartile multiple. CNY 4.02, the corrected Buy-Below, and CNY 2.64, the Add-More Trigger for this specific position, are the two price levels on the other side worth watching.

• Any authorisation of a share buyback, which would be the first in 26 years as a listed company and would materially change the capital-allocation read.

SOURCES AND DILIGENCE GAP LOG

Successfully Accessed:

• 2026 Interim Report (144pp, unaudited) | interim, primary | H1 ended 2026-06-30 | filed 2026-08-10 | https://pdf.dfcfw.com/pdf/H2_AN202608101827808571_1.pdf — verbatim exchange-filed PDF; balance-sheet and current-liability totals cross-footed line by line

• 2025 Annual Report (179pp, audited) | annual, primary | FY ended 2025-12-31 | filed 2026-03-31 | https://static.cninfo.com.cn/finalpage/2026-03-31/1225056936.PDF — original PDF from the exchange's designated disclosure venue; both balance-sheet years cross-footed

• Independent auditor report of Zhongshenzhonghuan CPAs, within the 2025 Annual Report | annual, primary | FY ended 2025-12-31 | 2026-03-31 | same URL — unqualified opinion; checked for going-concern material uncertainty

• 2025 Annual Report Summary (10pp) | annual, primary | FY ended 2025-12-31 | 2026-03-31 | https://static.cninfo.com.cn/finalpage/2026-03-31/1225057026.PDF

• 2023 Annual Report (170pp, audited) | annual, primary | FY ended 2023-12-31 | 2024-03-25 | https://pdf.dfcfw.com/pdf/H2_AN202403251628078575_1.pdf — FY2022 and FY2023 statements and notes

• 2023 Annual Report Summary | annual, primary | FY ended 2023-12-31 | 2024-03-26 | https://static.cninfo.com.cn/finalpage/2024-03-26/1219402252.PDF

• 2021 Annual Report Summary | annual, primary | FY ended 2021-12-31 | 2022-03-28 | https://pdf.dfcfw.com/pdf/H2_AN202203281555558373_1.pdf — FY2019 to FY2021 key figures

• 2026 First Quarter Report (19pp, unaudited) | exchange filing, primary | Q1 ended 2026-03-31 | 2026-04-28 | https://stockmc.xueqiu.com/202604/600269_20260428_NFZK.pdf

• Announcement on Change in Accounting Estimate (Lin 2026-038) | exchange filing, primary | effective 2026-01-01 and 2026-02-01 | 2026-08-11 | https://pdf.dfcfw.com/pdf/H2_AN202608101827808570_1.pdf

• 2026 Interim Report's directors/senior-management shareholding disclosure and its related-party transactions disclosure | insider activity, primary | H1 2026 | 2026-08-10 | as above

• 2025 annual general meeting resolutions as reported in the 2026 Interim Report | proxy, primary, partial | AGM 2026-06-18 | 2026-08-10 | as above

• Exchange monthly closing-price history for 600269, 2016-03 to 2026-08 | price history | SECONDARY | https://q.stock.sohu.com/hisHq?code=cn_600269&period=m

• Closing quote 2026-08-28: CNY 3.86, market capitalisation CNY 9,014,671,074.04 | price history | SECONDARY | https://push2delay.eastmoney.com/api/qt/stock/get?secid=1.600269 — market capitalisation divided by price reproduces the filed share count exactly

• Press coverage of the H1 2026 result (fair-value loss, Guosheng position, bond stock) | news | SECONDARY | 2026-08-12 | https://jx.ifeng.com/c/8vVSqYvwL4O — lead only; every figure used was re-verified in the interim report

• H1 2026 profit warning coverage (guidance CNY 180m to 260m attributable) | news | SECONDARY | 2026-07-14 | https://m.10jqka.com.cn/20260714/c678168841.shtml

• Exchange announcement index for 600269 | index | SECONDARY | 2026 | https://np-anotice-stock.eastmoney.com/api/security/ann?stock_list=600269 — used only to locate documents, each of which was then opened at the filing host

• USD/CNY reference rate 6.7217 at 2026-08-28 | FX | SECONDARY | https://tradingeconomics.com/china/currency — applied only to the decision figures

Diligence Gaps — Data Not Found or Estimated:

• Going Concern — checked in the independent auditor report within the 2025 Annual Report: no such note or language present (a finding, not a gap). Only the boilerplate responsibility paragraphs appear, with no material-uncertainty paragraph and no key audit matter on going concern.

• Maintenance capex [ESTIMATED] — the filings report total capex only and do not split maintenance from expansion. Owner earnings, discussed earlier in this report, use an assumed CNY 300–400m a year of maintenance capex. It matters because it is the difference between reported free cash flow of CNY 1,284.9m and normalized owner earnings of roughly CNY 2.6bn; the valuation does not rest on it, since the tier is set by P/TBV and the normalized P/E is computed from reported earnings, not from owner earnings.

• Goodwill and other intangibles for FY2019 to FY2021 [ESTIMATED] — not separately sourced, and held at the FY2022 filed level of CNY 147.33m when building the P/TBV history. The effect on tangible book per share in those three years is under 1% and cannot move the median or top-quartile statistics materially.

• Individual director and senior-management compensation [UNAVAILABLE] — interim reports do not disclose it, and the FY2025 annual report's remuneration detail was not extracted. It matters for capital-allocation warning 5, which is recorded as BORDERLINE rather than fired partly for that reason.

• The 2025 preliminary results bulletin and the 2026-07-14 profit-warning announcement [UNAVAILABLE as primaries] — both identified but not opened. The first is fully superseded by the audited annual report; the second is reported here only through secondary coverage and no figure from it is used in the analysis.

• The 2021 Annual Report full text [UNAVAILABLE] — the document exceeded the fetch tool's size limit, so FY2019 to FY2021 figures come from the 2021 Annual Report Summary, which carries the three-year key-figures table but not the full statements or notes. This is why total debt, capex and gross margin are blank for those three years in the fundamentals table.

• Classification of the CNY 1,989.9m shareholder loan to Changtong [INFERRED] — carried within non-current long-term payables while the related-party note gives a maturity of 7 September 2026. The filings do not reconcile the two, so an undisclosed extension is inferred; it is flagged above and on the Must-Watch list rather than resolved.

• Next results date [UNAVAILABLE] — no earnings-calendar entry was located. Late March 2027 is inferred from the 2024-03-25, 2026-03-31 and 2026-01-17 filing pattern and is labelled as an inference wherever it appears.

Overall Data Quality Rating: EXCELLENT — the audited annual report and the latest interim were both opened and read as primary documents at the exchange's designated disclosure venue or in verbatim mirrors of it, in Chinese and translated where used; the litigation, debt, related-party, subsequent-events, leases, segment and auditor-opinion notes were all located in the primary text, and going concern was checked and confirmed absent. The gaps above are real but none of them touches a figure the verdict rests on.

Action: BUY · 600269 Company: Jiangxi Ganyue Expressway Co., Ltd. · SHSE:600269 Executed: 26 August 2026, 02:00:41 (IBKR Flex reported time) Position: 800 shares · CN¥3.87 average price · CN¥3,096 execution value · 1 fill

Why I did it

This is a small China operating-company position bought at a substantial discount to filing-derived tangible book. The operating, concession, and capital-allocation risks remain part of the live thesis record.

What changes the view

The key review points are the payout ratio after the capex cycle, related-party pricing, and whether the discount proves structural rather than temporary.

Related research

Browse the Company Write-Ups → The individual 600269 write-up is forthcoming.

Back to the Trade Log →

Action: BUY · 2415.T Company: Human Holdings Co., Ltd. · TSE:2415 Executed: 24 August 2026, 01:40:15 (IBKR Flex reported time) Position: 100 shares · ¥1,655 average price · ¥165,500 execution value · 1 fill

Why I did it

This is the first non-US operating-company position under the new account. The position is deliberately small while the Japanese disclosure and ownership work stays under review.

What changes the view

I will revisit the position as the ownership and disclosure work closes, and as the company’s tangible-book economics either confirm or weaken the re-rating case.

Related research

Read the full Human Holdings write-up →

Back to the Trade Log →

Action: BUY · SHOE Company: Shoe Station Group Inc. Executed: 17 August 2026, 15:59:07–15:59:12 (IBKR Flex reported time) Position: 33 shares · $15.4913 blended price · $511.21 execution value · 4 fills

Why I did it

This was the first name to clear the full screen under this account. The stock was cheap against tangible book and eligible for the F-Score process. The balance sheet did more work in the decision than the growth story.

What changes the view

I will revisit the thesis if supplier concentration, the rebanner and capital-allocation failure, or earnings deterioration worsens materially.

Related research

Read the full Shoe Station Group write-up →

Back to the Trade Log →

Start with the full guide to the strategy, the rules, the record, and how to read this site.

Read the full Start Here guide →

1. Scorecard

2. Argument

The short version

Human Holdings runs three Japanese service businesses under one holding company: staffing, for-profit education, and elder care. At ¥1,643 the shares trade at 0.941x tangible book value, with net cash equal to roughly 116% of the market capitalization and a 4.3% dividend yield. The report's forensic checks turned up no deferred-revenue manipulation, no receivables stretch, no hidden lease liability, and no goodwill inflating the tangible-book number.

The company is controlled by the Sato family, a confirmed 45.24% bloc that could run as high as 60.45% once several unconfirmed corporate vehicles are counted. Only one of eight board seats is independent, below the exchange's own governance minimum. For roughly two years the company operated under a formal Tokyo Stock Exchange plan to widen its tradable-share ratio after it fell short of the requirement, using a self-tender and a share cancellation sourced specifically from officers and their relatives. That plan was confirmed complete as of March 2025.

Last year's profit growth was also thinner than the headline suggests. Net income fell even as revenue and operating profit rose, mostly because the effective tax rate rose 8.3 points. Most of the operating-profit gain came from cutting advertising and bonus spending, and that cut is already reversing this year.

Why it qualifies

• Trades at 0.941x tangible book value, with net cash worth about 116% of the market capitalization.

• No dilution of any kind, and capital allocation has been share-reducing: 609,201 shares, 5.5% of the register, were cancelled in December 2023.

• Revenue recognition passed every test run against it: contract liabilities, receivables, and days-sales-outstanding all behaved normally.

• Goodwill is negligible, and as a Japanese GAAP filer the company carries no capitalized lease liability inflating tangible book.

• A 4.3% dividend yield, backed by 47.3x interest coverage on a company with far more cash than debt.

Why it is not a clean bargain

• A confirmed 45.24% family ownership bloc, possibly as high as 60.45%, with only one independent director out of eight, below the exchange's own governance minimum.

• The company's float-widening moves were compelled by a Tokyo Stock Exchange listing rule, not chosen voluntarily, and the shares used in the 2023 self-tender came specifically from officers and their relatives.

• 117% of last year's operating-profit increase came from cutting advertising and bonus spending rather than the underlying business, and the bonus cut is already reversing.

• The effective tax rate rose 8.3 points last year, and next year's guidance implicitly assumes it partly comes back down.

• Officer compensation was never read, and the note that otherwise clears related-party dealings specifically excludes director pay by rule.

• Funded debt rose 14.6% in a single quarter with no explanation found, and the filing that would clarify it could not be retrieved.

What would change the view

• The next two quarterly results tested against management's full-year guidance, especially whether the tax rate and the reversing bonus cut behave as guided.

• Retrieval of the officer-compensation section of the annual filing, currently unread.

• Retrieval of the quarter's detailed results filing, which would explain the 14.6% rise in funded debt.

• Confirmation of whether the four unnamed corporate vehicles holding a combined 12.06% are affiliated with the family, which would move the real control stake toward 60.45%.

• Any tender offer, buyout proposal, or renewed listing-conformance issue involving the company.

Sources and gaps

The full report draws on the company's own annual and interim filings and two primary exchange notices. Several items remain unread or unconfirmed, including officer compensation, the newest quarterly filing, and the identity of some large shareholders. Where a fact could not be confirmed, the full report says so directly rather than estimating it.

3. Backup and sources

This last layer is the full working file: every source, calculation, and gap, kept so the argument above can be checked. The scorecard and the argument are the synthesis. Open this only if you want to check the work.

Company: Human Holdings Co., Ltd. (ヒューマンホールディングス株式会社) Ticker: 2415, Tokyo Stock Exchange, Standard Market As of: August 24, 2026, price ¥1,643 (Aug 24, 2026 close) Verdict: BUY - SMALL — a mechanical readout, not a recommendation. The base P/TBV × normalized-P/E cell alone, unmodified. The shareholder-yield and governance signals are reported as flags for a human to weigh, never applied to the tier. Tier and sizing are a human judgment call; see Position Disclosure below. Buy-Below: ¥1,400 (0.80× tangible book, 10.2-yr own-history) — see the valuation discussion below. Sell Range: ¥1,854 – 4,151 (1.06× – 2.38× tangible book). The top of this range reflects a valuation era the stock has not approached in over five years; read the annotation below before treating it as a near-term target. Add-More Trigger (held position): ¥1,393 (0.80× tangible book), a review price for adding, not a mechanical trigger. Position: Long, ~1% of portfolio (initiated 2026-08-23). Moat: None / Weak (staffing is a commodity; education is a fading legacy segment with two growing niches; nursing care is regulated utility-like). Balance-sheet risk: Net Cash (Conservative). Accounting quality: Adequate, with named gaps discussed below. Governance risk: High (family bloc 45.24%, possibly to 60.45%; one outside director of eight, below Corporate Governance Code Principle 4-8's minimum of two).

Primary sources reviewed: the FY2026/3 annual securities report (Yūhō), filed with EDINET June 24, 2026; the half-year securities report for H1 FY2026/3, filed November 14, 2025; the FY2026/3 full-year kessan tanshin results, filed May 15, 2026, including FY2027/3 guidance; the TSE listing-conformance plan (filed 2023-05-15) and the listing-conformance achieved notice (filed 2025-04-24), both retrieved and read in full; large-shareholding change reports No. 8 (2023-03-16) and No. 11 (2024-06-26), from secondary coverage, primaries not opened; the company's IR site, EDINET-sourced structured extraction, IRBank, and a second source for FY2027/3 guidance. Data quality: GOOD, with a named cap. No primary filing for the exact buyback share counts, the distribution prices, or the Q1 FY2027/3 tanshin was opened; see the sources and diligence-gap discussion at the end of this report.

Human Holdings Co., Ltd. (TSE: 2415) — Value Investing Analysis

1. Executive Summary

Human Holdings is a Japanese holding company running three businesses: staffing/HR services (59% of sales, ~89% of segment profit), for-profit education (26% of sales, the weak link), and elder-care services (13% of sales, small but improving). At ¥1,643/share it trades at 0.941x tangible book value, ~6.96x forward earnings on guidance (7.37x on last year's realized tax rate), and a 4.3% dividend yield, with net cash of ¥19,775.1mn — roughly 116% of its ¥17,051mn market capitalization. That is a genuinely rare setup for a profitably operating, dividend-paying business, and every forensic attack the adversarial panel could run against the reported numbers failed: no deferred-revenue manipulation, no receivables stretch, no capitalized-cost flattery, no hidden lease liability or goodwill inflating tangible book, no dilution. The cheapness is real.

The story around those numbers is materially different from a first read of the filing alone. This is a controlled company — a confirmed 45.24% family bloc, potentially as high as 60.45% depending on the affiliation of several corporate vehicles that could not be confirmed — with only one outside director on a board of eight, below the Corporate Governance Code's minimum of two. For roughly two years (2023–2025) the company operated under a formal Tokyo Stock Exchange continued-listing conformance plan, because its tradable-share ratio (~21%) sat below the 25% requirement for the Standard market. A self-tender in autumn 2023 acquired ~500,000 shares from officers and their relatives; a December 2023 cancellation retired 609,201 shares (5.5% of the register); and two further off-auction distributions of insider stock followed in 2024. All three actions are named in the company's own primary filings as the mechanism for curing the shortfall, and a follow-up filing confirms conformance was achieved as of 2025-03-31, at a 26.9% tradable-share ratio. The exchange-driven pressure is therefore spent, not live — but the point stands that for two years the only thing that ever moved this company's ownership structure came from the exchange, not from the company choosing on its own initiative to widen its float or reduce family control.

Earnings quality is also weaker than a surface read suggests. Reported net income fell 15% in FY2026/3 despite revenue, operating profit, and ordinary profit all rising — but the mechanical driver isn't purely one-off items. The effective tax rate rose 8.3 points (28.25% → 36.62%) on structural items (subsidiary rate mix, a larger valuation allowance, loss of a carryforward benefit) that FY2027/3 guidance implicitly assumes will partly reverse; on last year's actual tax rate, forward P/E is 7.37x, not 6.96x. Separately, 117% of last year's operating-profit improvement is accounted for by cuts to advertising and the bonus provision — underlying costs actually rose faster than gross profit — and the Q1 FY2027/3 operating-profit decline (~5.9%) is that bonus line reversing.

On balance: a genuinely cheap, balance-sheet-clean, dividend-paying business, wrapped in a governance structure that has shown exactly one mechanism for change (exchange compulsion, now apparently resolved) and an earnings trend that is more mixed than the headline suggests. The mechanical BUY - SMALL readout reflects the price and quality inputs alone (P/TBV, normalized P/E); the dividend yield and the governance risk are both stated above as separate flags, not folded into the tier — what that's worth to you, and at what size, is a human judgment call.

2. Post-Filing Events Bridge

• 2023-05-15 (TDnet, primary, retrieved 2026-08-24): filed a formal TSE continued-listing conformance plan. Tradable-share ratio was 20.85% against the 25.00% Standard-market requirement; the remedy was urging officers and relatives within two degrees of kinship to gradually sell holdings, targeting conformance by March 2025. This reframes every capital-structure move that follows as compliance-driven, not discretionary. Materiality: High.

• 2023-09-11 to 2023-10-10 (secondary aggregator; population confirmed by the 2025-04-24 primary): a self-tender acquired roughly 500,000 shares from company officers and their relatives within two degrees of kinship, at roughly 0.68x book, around ¥457mn, with no financial-statement impact until cancellation. Corroborated multiple ways, though exact terms are not primary-sourced. Materiality: High.

• 2023-12-22 (secondary aggregator, TDnet index title): cancellation of 609,201 treasury shares, the roughly 500,000 just acquired plus roughly 109,201 pre-existing, reducing issued shares from 10,987,200 to 10,377,999, a 5.5% reduction. Share-reducing capital allocation. Corroborated; the primary PDF was not opened. Materiality: High.

• 2024-06-21 (secondary aggregator): an off-auction distribution of 518,300 shares (roughly 5.0% of issued) by officers and relatives at ¥1,349, a 1.96% discount to market, a further float-widening step in the conformance plan. Corroborated; the primary PDF was not opened. Materiality: Medium.

• 2025-04-24 (TDnet, primary, retrieved 2026-08-24): filed notice confirming TSE listing-conformance achieved, with the tradable-share ratio progressing 20.8% (2023-03) to 21.2% (2024-03) to 26.9% (2025-03), clearing the 25% requirement. This closes the listing and delisting risk as of the most recent measurement date. Materiality: High.

• 2026-08-07 (scheduled, company IR calendar): Q1 FY2027/3 earnings released; operating profit reported down roughly 5.9%, confirming the bonus-provision reversal thesis discussed below. Confirmed via secondary coverage; the primary tanshin was not opened due to access failures. Materiality: High.

No litigation, credit-facility change, or large-shareholding change beyond the items above was identified between the June 2026 annual filing and August 24, 2026.

3. Business and Market Overview

Human Holdings is a pure holding company founded in 1985, listed since 2004, headquartered in Shinjuku, Tokyo, managing roughly 14 subsidiaries and 2 affiliates across four reporting lines: Staffing/HR (¥60.7bn revenue, 59% of sales, ~89% of segment profit, run mainly through Human Resocia), Education (¥26.4bn revenue, 26% of sales, weakest segment, run through Human Academy and Human Star Child), Nursing care (¥12.9bn revenue, 13% of sales, run through Human Life Care), and Other (¥2.6bn revenue, 2% of sales — a professional basketball club and a nail-salon chain).

FY2026/3 consolidated revenue was ¥102,539.3mn, up 2.2% year over year, with operating profit of ¥3,614.1mn (3.5% margin) and ordinary profit of ¥3,887.6mn. Three secular tailwinds work in the company's favor: a chronic Japanese labor shortage supporting staffing demand and pricing; rising demand for Japanese-language education from foreign nationals under loosened immigration rules; and an aging population needing more elder care. Working against it: a shrinking pool of Japanese youth is eroding the full-time vocational-college business, and low unit economics across headcount-driven staffing and regulated nursing-care reimbursement cap margins structurally. This is a stable-cash-generating conglomerate of thin-margin, labor-intensive service lines, not a growth compounder.

4. Moat, Competitive Position, and Industry Cycle

Moat classification: None/Weak, with narrow, defensible niches inside an otherwise commoditized structure.

• Staffing: No moat. Commoditized and license-gated, with many competitors (Recruit, Persol, Pasona, regional players). A modest "education → staffing" pipeline advantage exists but isn't durable. Segment profit grew 7.5% year over year on Japan's structurally tight labor market.

• Education: Narrow, eroding moat in the adult-reskilling/Japanese-teacher-training niche (recently strengthened by conversion of "Japanese-language teacher" into a national licensure). No moat in the full-time vocational-college business, directly exposed to Japan's shrinking 18-year-old cohort. Segment profit fell materially year over year.

• Nursing care: No moat beyond local licensing and facility density; heavily regulated pricing via Japan's long-term-care insurance fee schedule. Margins recovered on staffing/utilization improvements but the segment remains structurally low-margin.

• Overall: not a capital compounder. The appeal is valuation and balance-sheet strength, not franchise quality.

5. Management, Governance, and Capital Allocation

Board. Eight directors as of 2026-06-24 (seven at the 2026-03 filing date; the June addition was an internal director, so independence went from 1-in-7 to 1-in-8). One outside director — below Corporate Governance Code Principle 4-8, which requires at least two independent outside directors for Standard-market issuers. Statutory kansayaku model with three kansayaku; no nomination or compensation committee; no 独立役員 designated on the company's own governance page. The company's own CG report was not read for this run.

Ownership concentration. Confirmed top holders as of the 2026-03 record date: Yūgen-gaisha Parents 22.27%, Shingo Sato 9.64%, Tomoya Sato (President & CEO, personally) 8.31%, Koichi Sato 5.02% — a confirmed family bloc of 45.24%. Four additional corporate vehicles (Office Aguri, JPSKN, Kobee, Harvest — 12.06% combined) carry Sato-adjacent naming patterns whose affiliation could not be confirmed; if affiliated, the bloc rises to 57.3%. The top-ten register was found to be missing a line — an individual holder, 野村愛 (Ai Nomura), at 3.15% — which, if also family-affiliated, would put the ceiling at 60.45%. No trust bank, custody bank, pension fund, or foreign institution appears anywhere in the top ten.

President's personal holding does not reconcile. The president's stake is reported at 30.58% (parents 22.27% + personally 8.31%), reproducing exactly from a June 2024 large-shareholding change report. But 1,078,000 shares less the 518,300 distributed in June 2024 leaves 559,700 (5.39%), against the reported 862,412 (8.31%) — a ~302,000-share gap. Either an insider re-acquisition occurred that hasn't been disclosed, or one of the two secondary sources describing this is wrong. Unresolved; the 30.58% figure carries an at-risk flag.

Capital allocation is share-reducing, not merely non-dilutive. The autumn 2023 self-tender and December 2023 cancellation retired 609,201 shares (5.5% of the register) at roughly 0.68x book (~¥914/share, ~¥457mn), on a ~52.1% total-return payout that year. Two readings are both supported and neither is established: constructively, management acted on its own discount below book; adversely, the ~500,000 shares acquired came specifically from company officers and their relatives within two degrees of kinship — confirmed by the primary 2025-04-24 filing — not the open market, and company cash may have lifted the family bloc roughly two points (43.2% → 45.2%) in the process. The stated policy otherwise is a 30% consolidated payout ratio; FY2026/3 dividend was ¥69/share (¥65 ordinary + ¥4 one-off 40th-anniversary special), down from ¥75.50/share the prior year on a like-for-like ordinary basis. No stock options or convertible instruments are outstanding.

Insider selling is real, ongoing, and non-informational in motive. Change Report No.8 (2023-03-16) shows an estate-planning transfer of the president's personal stake into his own asset-management vehicle, not a market disposal. Change Report No.11 (2024-06-26), filed one day after an off-auction distribution announcement, shows the vehicle buying (21.59% → 22.27%) while the president personally sold (9.81% → 8.31%), at a 1.96% discount to market — disclosed motive is the listing rule, not a view on value. The supply overhang into a thin float is real; the "insiders think the shares are cheap" reading is not supported and is explicitly withdrawn.

Red flag to monitor: the March 2026 resignation of the director responsible for the underperforming Education segment (concurrently representative director of the main Education subsidiary), replaced by an internally-sourced successor at the June 2026 AGM.

6. Corporate Ownership, Subsidiaries, and Joint Ventures

Eight consolidated subsidiaries, several small non-consolidated overseas entities, and two equity-method affiliates. Human Resocia (Staffing, 100%, FY2025/3 revenue ¥58,362mn, net income ¥1,980mn), Human Academy (Education, 100%, ¥20,881mn revenue, ¥423mn net income), Human Life Care (Nursing, 100%, ¥12,336mn revenue, ¥229mn net income), plus Human Global Talent, Human Star Child, Human Academy Europe SAS, Human Planning, and Dashing Diva International. Relationship-company stock impairment losses of ¥80.4mn (FY2026/3) and ¥70.5mn (FY2025/3) were recorded against small overseas ventures (Malaysia, US) — a recurring, individually-immaterial pattern worth monitoring. No upstream guarantees or cross-default provisions identified; all named consolidated subsidiaries are wholly owned.

7. Historical Financial Quality and Normalized Owner Earnings

Three years not shown on that chart: ordinary profit was ¥3,294mn (FY2024/3), ¥3,576.2mn (FY2025/3), ¥3,887.6mn (FY2026/3); operating cash flow was ¥4,621mn, ¥1,523.1mn, ¥3,282.6mn; the equity ratio was 31.5%, 35.7%, 37.5%; EPS was ¥202.20, ¥251.21, ¥213.60; and book value per share was ¥1,554.52, ¥1,743.70, ¥1,882.08.

The net-income bridge has three legs, not one. FY2026/3's ordinary profit rose (+¥311.4mn), but net income fell, because of a net extraordinary swing of −¥447.8mn (income −¥140.1mn→¥0.6mn, loss ¥82.6mn→¥390.9mn) and a −¥253.9mn tax-charge increase from the 8.3-point effective-tax-rate rise. Both legs are structural to some degree — the tax-rate rise draws on subsidiary mix and a larger valuation allowance, not a pure one-off — and guidance implicitly assumes partial reversal. Ordinary profit, not net income, is the more reliable trend line, and it grew 8.7% year over year.

Operating-profit growth was mostly a spending cut. Advertising fell ¥174.4mn and the bonus provision fell ¥70.4mn — together ¥244.8mn, 117% of the ¥209.3mn operating-profit increase. Strip them out and underlying SG&A rose faster than gross profit. Q1 FY2027/3's ~5.9% operating-profit decline on "expanded human-capital investment" is that bonus line reversing.

Free cash flow is unstable, not merely low. FY2025/3 FCF was ¥42.2mn against dividends paid of ¥783.5mn (18.6x coverage shortfall that year); FY2026/3 FCF was ¥2,278.1mn. The two-year average FCF yield on current market cap is 6.8%, against a single-year FY2026/3 figure of 13.4% — both should be read together, not the higher number alone. Roughly ¥1.8bn of the FY2026/3 operating-cash-flow inflow is unexplained by earnings after eliminating the tuition float, receivables, and capitalization as candidate sources — genuinely open, not resolved.

Piotroski F-Score: 7–8 of 9. Signal 4 (CFO > net income) passes in FY2026/3, fails in FY2025/3 — a one-of-two-years pass. Signal 8 (gross-margin improvement) is published as indeterminate: the delta is +0.00038 percentage points, arithmetically exact but well inside the precision a management allocation between cost lines can actually support.

8. Balance Sheet, Debt, Covenants, and Refinancing Risk

As of March 31, 2026: current portion of long-term borrowings ¥2,860.4mn, long-term borrowings ¥6,496.3mn, total interest-bearing debt ¥9,356.7mn. Cash and equivalents ¥29,131.8mn, for a net cash position of ¥19,775.1mn, roughly 116% of market cap. Interest coverage 47.3x, equity ratio 37.5%.

No bonds, convertible instruments, or off-balance-sheet financing identified. Balance-sheet risk classification: Net Cash / Conservative. One open item: funded debt rose 14.6% in the June 2026 quarter (¥9,356.7mn → ¥10,722mn), against only a ¥1.0bn cash build. Netting the ¥716.1mn AGM dividend narrows the gap, and a seasonal tuition-float explanation is now unlikely because the float is flat year over year — the draw is unexplained and the Q1 tanshin that would clarify it was not read. A meaningful share of the ¥29.1bn cash balance is economically pre-funded by ¥9,893.2mn of contract liabilities (prepaid tuition), so it is not entirely "free" cash even though it carries no debt-service obligation.

9. Real Estate, Leases, and Hidden Assets

The company operates 600+ leased locations. Lease-note detail was not fully disclosed in the documents reviewed: the note discloses only ¥2,562.8mn of non-cancellable minimum lease commitments (¥353.8mn within one year), which is explicitly not the same quantity as annual rent expense — ¥8,033.1mn of the ¥22,736.2mn SG&A total is an unitemized residual with rent inside it, and annual rent expense itself is undisclosed. Three filing-derived indications the leased footprint carries real weight: guarantee and lease deposits of ¥2,269.4mn (up from ¥2,191.6mn), an implied asset-retirement obligation of ~¥1,095mn (grossed up from the related deferred tax asset), and FY2026/3 impairments falling overwhelmingly on buildings and structures at leased education and care sites. Together, these DTA and deposit items are 16.6% of the ¥18,126.1mn tangible book and are not realizable at carrying value in a wind-up scenario.

10. Capital Markets Access, Dilution, and Financing Flexibility

No shelf registration, ATM program, or rights offering identified. Share count is not flat across the period: it moved from 10,987,200 (pre-cancellation) to 10,377,999 (issued) via the December 2023 cancellation described above under governance — a share reduction, not the "flat for at least three years" picture a filing-only read produces. No dilutive instruments of any kind — positively confirmed via unchanged capital accounts, basic equalling diluted EPS on both routes, and an exhaustive net-assets decomposition leaving no room for a warrant/option component. Working capital and small bolt-on M&A continue to be funded from internally generated cash flow and modest bank borrowing.

11. Litigation, Regulatory, and Contingent Liability Risk

No specific pending litigation, regulatory investigation, fine, or contingent liability was disclosed in any section reviewed. Two independent Japanese-language sweeps — the original research and a second, separately-run adversarial-panel sweep across the parent, Human Academy, and the language-school subsidiary for administrative sanctions, lawsuits, labor-bureau action, or corrective guidance — returned nothing adverse in a licence-gated sector with a real enforcement record. This is "nothing found in the sections and searches reviewed," not an independently verified clean bill of health, and should not be re-run without new information.

12. Accounting Quality and Disclosure Review

• Auditor: Same firm 24 years, partner rotation confirmed, firm rotation not required or expected under Japanese rules — single-sourced (two independent attempts to second-source failed) but a market norm in Japan, not an anomaly. Fees ¥61.4mn (FY2026/3) vs ¥59.9mn (FY2025/3).

• Revenue recognition: passes every test run against it — contract liabilities flat relative to revenue growth, receivables flat with days-sales-outstanding falling, intangibles amortizing faster than additions.

• Impairment/goodwill: goodwill is a trivial ¥45.1mn (0.09% of assets); no capitalized lease liability under J-GAAP — the two standard tangible-book inflations both refute cleanly.

• Related-party disclosure: the FY2026/3 note reads 該当事項はありません (no applicable items) for both FY2025/3 and FY2026/3, and the transcription is not disputed — but four qualifiers now travel with that nil, replacing the single materiality caveat a first read would apply: (1) director remuneration, bonuses, and retirement allowances are categorically excluded from this note's scope under ASBJ Statement No. 11, so the nil says nothing about officer pay — the one governance section never read; (2) a share repurchase through public tender is a capital transaction and wouldn't appear here even if insiders tendered; (3) the note's window is two years and the self-tender/cancellation fall just outside it; (4) verification re-fetched the same source the original read used — one transcription checked against itself. Working the other way: the individual-related-party disclosure threshold is only ¥10mn, tight enough that the nil is genuinely informative about dealings with the Sato family personally — which threshold limb governs entities they control was not verified.

• Officer compensation: entirely unread. A previously-circulated ¥300–400mn estimate has been withdrawn as unsupported; the only figure found (~¥52.84mn per director) comes from a low-quality aggregator and is not adopted. No estimate currently on file.

Overall accounting-quality classification: Adequate, with the related-party/officer-compensation gap now the single most consequential disclosure hole in the file, given the controlled-company ownership structure.

13. Valuation

Given the net-cash position, both a simple multiple read and a cash-adjusted read matter, and the adjusted-EV construction depends on a convention choice this house has now settled.

Snapshot (2026-08-24, price ¥1,643):

• Market capitalization: ¥17,051mn

• P/TBV: 0.941x (tangible book ¥18,126.1mn; TBV/share ¥1,746.5)

• P/B on March 2026 book: 0.873x (BVPS ¥1,882.08)

• Forward P/E: 6.96x on FY2027/3 guidance (EPS ¥236.08) / 7.37x on last year's realized 36.62% tax rate applied to guided ordinary profit — the guidance-implied figure is the more optimistic of the two and should be read alongside the higher one

• Net cash: ¥19,775.1mn, ~116% of market cap

• Dividend yield: 4.3%

• FCF yield: 13.4% single-year (FY2026/3) / 6.8% two-year average — publish both, not the higher one alone

• Simple enterprise value: −¥2,724mn (negative headline, but this construction ignores the ¥9,893mn of prepaid-tuition contract liabilities entirely and should not be published alone)

Adjusted excess cash and adjusted EV depend on a deferred-revenue convention with no single right answer. Gross-deferral (subtract 100% of contract liabilities) gives adjusted excess cash of ¥9,881.9mn and EV/EBIT of 1.98x. Cost-to-complete (subtract only the ~74.30% of the float that represents the actual cost to deliver, rather than its full face value) gives ¥12,424.5mn and EV/EBIT of 1.28x. Both readings are published here since neither is clearly the single right answer, but the more conservative gross-deferral figure (1.98x) is the one this report treats as primary.

Buy-Below and Sell Range

Both price targets are computed from the same source: 41 quarterly P/TBV observations, 2016-06-30 to 2026-06-30, cross-checked against the filing-derived current P/TBV before being trusted (provider series 0.946× vs. derived 0.941× at 2026-06-30, a 0.005-point gap, well inside the acceptance test).

Buy-Below: ¥1,400 (0.80× tangible book, 10.2-yr own-history) — the mean of every quarter in that history at or below the stock's own 25th percentile.

Sell Range: ¥1,854 – 4,151 (1.06× – 2.38× tangible book) — the median-to-top-quartile-mean of the same history. Read the top of this range with real caution. The stock has not traded above roughly 1.36× tangible book at any point in the last five years; the 2.38× top-quartile-mean is driven almost entirely by 2016–2019, when the shares regularly traded at 1.5×–2.9× book. That earlier valuation regime coincided with a materially different market environment for this stock, and nothing in the current filings suggests it is close to returning. The bottom of the range (1.06×) is more credible on its own terms — the stock touched 1.36× as recently as September 2025 — but the top should be read as a historical data point, not a near-term target. The mechanical range itself is not adjusted for this; only this annotation is added.

Add-More Trigger (held position only): ¥1,393 (0.80× tangible book) — a single, entry-anchored review price for adding to this specific position (entry ¥1,655, 2026-08-24), one 0.15-P/TBV-point step below entry. A guideline for revisiting the thesis if reached, never a mechanical buy signal, and never applied to sizing or the tier.

At ¥1,643, the stock sits between its own Buy-Below (¥1,400) and the floor of its own Sell Range (¥1,854) — above the level this report would call fresh-money cheap, but below where the sell question becomes live. That middle position is consistent with a mechanical tier that is already Small, not Big, on the numbers alone.

The mechanical tier readout (front matter) is the base P/TBV × normalized-P/E cell alone — P/TBV below 1.0x on a substantial-net-cash, positive-and-growing-ordinary-profit, Piotroski 7–8/9 business. The 4.3% yield is a separate flag, not an input to that readout.

14. Risk Matrix

• Going-private on a depressed price. Evidence: a confirmed 45.24% family bloc, possibly up to 60.45%; no institutions in the top ten; company cash already used once to consolidate the position. Severity Medium. No mitigant disclosed. Monitor for any MBO or tender signal.

• FY2027/3 guidance requires two things at once. Evidence: no repeat of FY2026/3's extraordinary losses, and roughly 370bp of tax-rate recovery, against a Q1 already down roughly 5.9% in operating profit. Severity Medium-High: forward P/E is 7.37x against the 6.96x headline if guidance underdelivers. Monitor Q2/Q3 FY2027/3 prints against guidance.

• Operating momentum was a spending cut. Evidence: 117% of last year's operating-profit gain was two discretionary cost cuts, one already reversing. Severity Medium, roughly a ¥245mn/yr swing. Mitigant: underlying gross profit is still growing. Monitor the SG&A trend excluding discretionary lines.

• Funded debt rose 14.6% in one quarter, unexplained. Evidence: ¥9,356.7mn to ¥10,722mn in June 2026; a seasonal explanation is weakened by a flat tuition float. Severity Low-Medium, impact unknown. Mitigant: the net-cash position provides a cushion. Monitor by reading the Q1 tanshin when accessible.

• Related-party and officer-compensation blind spot. Evidence: this category is excluded by rule from the note this report otherwise treats as a green flag, and was never read. Severity Low-Medium, impact unknown. A threshold analysis suggests the nil is still informative for personal dealings. Monitor by retrieving the officer-compensation section before treating the related-party nil as fully clean.

• Illiquidity and thin float. Evidence: no institutional holders in the top ten; TSE Standard, not Prime. Severity Medium. No mitigant. Size positions conservatively.

15. Red Flags, Yellow Flags, and Green Flags

🔴 Red Flags:

• Controlled company: 45.24% family bloc confirmed, potentially 57.3% or 60.45%; no institutions in the top ten; one outside director on a board of eight, below CG Code Principle 4-8.

• Listing-maintenance history: tradable-share ratio ~21% against a 25% TSE Standard requirement drove a formal conformance plan filed 2023-05-15; conformance appears achieved as of 2025-03-31, but has not been re-confirmed since.

• The FY2026/3 net-income decline is only two-thirds extraordinary: ¥253.9mn of it is an 8.3-point structural rise in the effective tax rate that FY2027/3 guidance implicitly assumes partly reverses.

• 117% of the FY2026/3 operating-profit increase is an advertising and bonus-provision cut Q1 shows reversing.

• Insider selling is real and ongoing as a supply overhang, driven by the listing-conformance mechanism rather than a view on value.

• Funded debt up 14.6% in the June 2026 quarter, unexplained, with the Q1 tanshin unread.

• Roughly ¥1.8bn of FY2026/3 operating cash inflow is unexplained by earnings after eliminating the three standard candidates.

🟡 Yellow Flags:

• Free cash flow is unstable: ¥42.2mn (FY2025/3) vs. ¥2,278.1mn (FY2026/3) — several published metrics are one swing wearing different hats and reverse on the prior year.

• 16.6% of tangible book is a deferred tax asset already 55.6% written off plus landlord deposits recoverable only on lease exit against a ~¥1.1bn restoration obligation.

• Annual rent expense is undisclosed; the leased footprint (600+ locations) is materially larger than the disclosed minimum lease commitment alone suggests.

• ~¥622mn of accrued, unfunded executive retirement allowance to a family-dominated board, outside the scope of the related-party note.

• Officer compensation entirely unread — precisely the disclosure category the related-party nil excludes by rule.

• Gross margin identical to four decimal places across two fiscal years — plausibly formulaic pass-through pricing in staffing, unconfirmed.

• Auditor tenure 24 years, single firm — a market norm in Japan, single-sourced.

🟢 Green Flags:

• Net cash of ¥19,775.1mn, ~116% of market cap, against ¥9,356.7mn of funded debt; interest coverage 47.3x.

• Trading at 0.941x tangible book with the two standard inflations (right-of-use assets, goodwill) tested and refuted: J-GAAP filer with no capitalized lease liability, goodwill ¥45.1mn.

• Revenue recognition passes every test available: contract liabilities flat, receivables flat, days-sales-outstanding falling.

• Capital allocation is share-reducing: 609,201 shares (5.5%) cancelled December 2023 at roughly 0.68x book, on a ~52% total-return payout.

• No dilutive instruments, positively evidenced.

• Ordinary profit grew on revenue up 2.2% — modest growth, not a shrinking business.

• No litigation, enforcement action, or licensing sanction surfaced in two independent sweeps of a licensed sector.

• 4.3% dividend yield.

Position and Verdict

Current price: ¥1,643 (2026-08-24). Position disclosure: Long, ~1% of portfolio (initiated 2026-08-23). Mechanical tier readout: BUY - SMALL, the base P/TBV × normalized-P/E cell alone, unmodified. Flags for the human tier judgment: shareholder yield 4.3% (at or above the 3.0% threshold, a +1 notch-equivalent) and governance risk High, neither applied to the tier above. Buy-Below: ¥1,400 (0.80× tangible book, 10.2-yr own-history). Sell Range: ¥1,854 – 4,151 (1.06× – 2.38× tangible book), top annotated in the valuation discussion above. Add-More Trigger (held position): ¥1,393 (0.80× tangible book), a review price, not a mechanical trigger. Balance-sheet risk: Net Cash / Conservative. Governance risk: High. Accounting quality risk: Adequate, with a named related-party and officer-compensation gap. Refinancing risk: Low (net cash, 47.3x interest coverage). Primary thesis killers: a related-party arrangement discovered to extract value from minority shareholders; renewed listing non-conformance; loss of a core operating license; a going-private move at a price that ignores the net-cash cushion.

Still Open

Nine items the research could not close, listed so nothing here reads as more certain than it is: who specifically tendered into the 2023 buyback (population confirmed, individuals not named); the president's personal-holding discrepancy (~302,000 shares); ~¥1.8bn of unexplained operating cash inflow; the 14.6% quarterly debt rise (narrowed, not resolved); whether the four unnamed corporate vehicles (12.06%) are family-affiliated; which related-party threshold limb governs family-controlled entities; audit firm tenure/fees (single-sourced); officer compensation (never read); and a small, recurring, permanently-unexplained secondary-source discrepancy in operating profit not worth further investigation.

SOURCES AND DILIGENCE GAP LOG

Successfully Accessed:

Documents successfully accessed: the annual securities report (Yūhō, EDINET doc. S100YIDK) for FY2026/3, filed 2026-06-24; the half-year securities report (EDINET doc. S100X4BW) for H1 FY2026/3, filed 2025-11-14; the full-year kessan tanshin for FY2026/3, filed 2026-05-15, via the TDnet-hosted Yahoo! Finance Japan archive; the TSE listing-conformance plan (doc. 140120230512570763), as of 2023-03-31, filed 2023-05-15, and the listing-conformance achieved notice (doc. 140120250424522241), as of 2025-03-31, filed 2025-04-24, both via a public TDnet archive mirror as primary PDFs; the company's IR site on an ongoing basis; and IRBank and traders.co.jp for the holder table, segment page, and a second source on FY2027/3 guidance.

Overall Data Quality Rating: GOOD. The two most consequential documents this run needed — the listing-conformance plan and its achieved-conformance follow-up — were retrieved and read as primary filings during human-gate review, closing the run's largest open question. The rating is capped below EXCELLENT because several other material facts (the exact buyback share counts and price, the two 5%-rule change reports, officer compensation, the Q1 FY2027/3 tanshin, and the company's own CG report) remain sourced from secondary aggregators or entirely unread.

Diligence Gaps — Data Not Found or Estimated:

• Officer compensation section (役員の状況): never attempted. The single largest remaining gap given the related-party note's scope exclusion. [UNAVAILABLE]

• Q1 FY2027/3 kessan tanshin: the only document that would explain the 14.6% quarterly debt rise and confirm the operating-profit decline directly. [BLOCKED] — TLS/decoding failures on every access route attempted.

• The two 5%-rule large-shareholding change reports (No.8, No.11), primary: would separate genuine family selling from a co-holder leaving the joint-holding perimeter, and would resolve the president's-holding discrepancy. [UNAVAILABLE] — secondary coverage only.

• Exact buyback/cancellation/distribution share counts and prices, primary: multiply corroborated from secondary sources and internal arithmetic, but no primary TDnet PDF for the 2023-09-08 tender or 2023-12-08 cancellation notices was opened (paywalled, undecodable, or 403'd on every route attempted). [UNAVAILABLE]

• Corporate Governance Report (コーポレート・ガバナンス報告書): would show whether the company explains or disputes the CG Code Principle 4-8 shortfall. [UNAVAILABLE] — never attempted.

• Detailed lease terms (ROU assets, weighted-average term, discount rate): explicitly omitted from tanshin/interim filings as a permitted simplification. [UNAVAILABLE]

• Related-party threshold limb for director-controlled entities: which JPY threshold (individual vs. corporate) applies to entities the Sato family controls was not verified against the standard. [UNVERIFIED]

This report should be read as well-sourced but not fully exhaustive. An investor sizing a position beyond a starter/watchlist size should independently pull the officer-compensation and CG-report sections directly from EDINET (E05428) before committing further capital.

This report is for informational and research purposes only and does not constitute investment advice. It is not a substitute for independent due diligence, and the author is not a licensed financial advisor. All figures are in Japanese yen unless otherwise noted; no currency conversion to USD was performed.

1. Scorecard

2. Argument

The short version

Shoe Station Group is a 426-store family-footwear retailer trading at 0.68x filing-derived tangible book. At $15.47, the shares sit near the bottom of their own historical valuation range. The balance sheet gives the thesis room: the company has zero financial borrowings, about $116.1 million of unencumbered cash, and no refinancing wall.

The problem is that the operating business has deteriorated. Return on equity fell to 7.6% in FY2025, operating margin fell from 15.6% in FY2021 to 5.9%, and inventory became harder to clear. Management spent roughly $37.1 million converting 101 stores to the Shoe Station banner, then abandoned the program and recorded $8.3 million of impairments.

That makes this a real bargain with a real problem attached. The balance sheet can buy time. It cannot restore the old economics by itself, and tangible book is not a liquidation floor when leases and inventory make up so much of the anchor.

Why it qualifies

• The shares trade at a discount to filing-derived tangible book and at the 2.4th percentile of the company's own ten-year P/TBV history.

• The company has zero financial borrowings for a 21st consecutive year, so lender pressure and a maturity wall are not the immediate risks.

• The F-Score is 5/9 and remains eligible under the survival rules. A 5.8% shareholder yield and two capital-allocation warnings are flagged for judgment, not applied to the tier.

• The core downside question is understandable: can a debt-free retailer clear inventory and stabilize returns before the tangible-book discount stops mattering?

Why it is not a clean bargain

• The operating record is moving the wrong way. Revenue fell 5.6% in FY2025, operating margin fell to 5.9%, and the first quarter of FY2026 was a loss.

• Inventory days rose from 181.8 to 222.8 while cost of goods sold fell, and a reported dividend-cover figure of 1.59x falls to 0.11x after stripping out a payables build.

• The re-banner program is a capital-allocation failure: 101 conversions, roughly $37.1 million spent, $24.1 million of operating-income drag, and $8.3 million of impairments before the program was abandoned.

• Governance risk is High with two warnings: concentrated family control and capital routed into the failed re-banner project. The interim CEO adds uncertainty about the next capital-allocation regime, but a credible permanent CEO with better capital allocation would be a positive catalyst.

• The tangible-book anchor needs qualification. Right-of-use assets are 55.4% of tangible book and inventory is 67.9%, so this is cheap against tangible book, not an asset-backed liquidation play.

How a minority holder gets paid

For a minority holder, the payoff is likely to come from ordinary cash returns and mean reversion, not a control event. The company has raised its dividend for twelve years and has resumed buybacks while it remains below tangible book. If inventory and margins stabilize, the market can re-rate the shares toward the report's sell range without a sale of the company. If they do not, the cash balance and ongoing dividend provide time but not a floor. The holding therefore depends on a cleanup in operating returns, not on family control changing hands.

What would change the view

• Q2 diluted EPS and gross margin, especially a diluted result at or below roughly $0.3012.

• Inventory clearance at price rather than through markdowns, together with the direction of inventory days and the auditor's inventory realizable-value judgment.

• A reversal of the payables build that shows whether reported cash generation can support the dividend without working-capital help.

• A credible permanent CEO appointment followed by evidence of better capital allocation would reduce the governance discount, especially if the company does not repeat the abandoned conversion strategy.

• A dividend cut, further store impairments, a material inventory write-down, or renewed margin erosion.

Sources and gaps

The full report reviews the latest annual and quarterly filings, ownership and insider material, the historical P/TBV series, and the relevant accounting and governance disclosures. The next quarterly filing is not yet available. There is also no earnings-call transcript in the reviewed record, so management's answers on capital allocation remain unassessable rather than treated as a pass.

3. Backup and sources

This last layer is the full working file: every source, calculation, and gap, kept so the argument above can be checked. The scorecard and the argument are the synthesis. Open this only if you want to check the work.

The short layers use the 22 August market snapshot. The detailed report preserves the 17 August closing-price inputs used in its original valuation math; every price reference below is dated.

Company snapshot

• Company: Shoe Station Group, Inc. (formerly Shoe Carnival, Inc. through 2026-06-11)

• Ticker: SHOE (NASDAQ) — traded as SCVL until 2026-06-11

• As of: 19 August 2026

• Current price: $15.50 (~$417M market cap)

• Verdict: BUY — STANDARD

• Buy-below: $18.11 (0.80× TBV)

• Buy More Below: $14.71 (0.65× TBV)

• Sell range: $33–$50 (1.48×–2.20× TBV)

• Moat: Weak

• Governance risk: High (2 warnings)

Primary sources reviewed

• Annual report: Form 10-K, FY2025 (ended 2026-01-31), filed 2026-03-26 — acc. 0001193125-26-126279

• Latest quarterly: Form 10-Q, Q1 FY2026 (ended 2026-05-02), filed 2026-06-05 — acc. 0001193125-26-258362

• Proxy: DEF 14A, filed 2026-04-29 — acc. 0001193125-26-191497

• Earnings release: Form 8-K Ex-99.1, filed 2026-03-26 — acc. 0001193125-26-125001

• Insider activity: All 24 Forms 4 filed 2025-08-16 → 2026-08-16, XML parsed

• Historical filings: 10-Ks for FY2024, FY2023, FY2021, FY2019 (EPS and balance-sheet history)

• Financials and charts: Exported 2026-08-17; includes 41 P/TBV observations

• Data quality: EXCELLENT

Full detail: Shoe Station Group

1. Executive Summary

Shoe Station Group is a 426-store omnichannel footwear retailer, operating the value-focused Shoe Carnival banner and the higher-income Shoe Station banner across 35 states and Puerto Rico. It trades at 0.685× tangible book value (2026-08-17 close, $15.50) — a hair above the ten-year window floor of 0.673× recorded in April 2026, and still far below anything the stock has traded at outside the last eight months — with zero borrowed money, $129.3M of net cash equal to 31% of the share price, and a 4.44% dividend raised for the twelfth consecutive year. The verdict is BUY — STANDARD.

Two capital-allocation problems keep this from screening as an unambiguous top-tier holding despite the balance sheet. In the twelve months to August 2026 the company converted 101 stores to its Shoe Station banner at a cost of roughly $0.66 per share of earnings and a 35% increase in capex, then wrote the program down and abandoned it, taking $8.3M of impairments and announcing 18–24 store closures. It terminated the CEO who authored that strategy, paid him $4.8M in severance plus accelerated stock, and six months later still has no permanent replacement. That is not a balance-sheet problem; it is a capital-stewardship problem, and Schloss-style investing tolerates cheap assets far better than it tolerates managements that destroy value at the margin.

The central analytical question is whether the de-rating is mispricing or repricing, and the honest answer is mostly repricing. Return on equity fell from 34.2% in FY2021 to 7.6% in FY2025. A business earning 7.6% on book does not deserve the ~1.5× multiple this stock carried at its ten-year median; on a 9–10% cost of equity, something in the range of 0.75–0.85× book is defensible. At 0.685× P/TBV the stock is genuinely below that — but the margin is a third, not the three-quarters that "cheapest in a decade" implies at first glance. Anyone underwriting a return to 1.5× book is underwriting an ROE recovery for which there is currently no evidence.

What makes it investable anyway is the downside. There is no lender. Twenty-one consecutive debt-free fiscal year-ends, a $100M revolver with nothing drawn, and $129.3M of cash and securities mean no covenant can be tripped, no maturity can be missed, and no refinancing window can close. Free cash flow was positive even in FY2025's trough, and the Q1 GAAP loss of $(0.21) came alongside positive $23.1M of operating cash flow — the loss was substantially non-cash. Two facts point the same direction: the CFO bought 31,000 shares of his own money at $16.13 in April, the only open-market purchase by any insider in twelve months, and the company resumed buybacks for the first time since FY2023.

The single most important risk is not leverage or governance — it is supplier concentration. Nike alone is 24% of net sales, the top three vendors 46%, and the company states plainly that it has no long-term contracts with any supplier. A single allocation decision by Nike would impair this thesis faster than any balance-sheet strength could absorb.

2. Business and Market Overview

Shoe Station Group is an omnichannel family footwear retailer operating 426 stores across 35 states and Puerto Rico under two banners: Shoe Carnival (282 stores, value and moderate-income families) and Shoe Station (144 stores, higher-income shoppers, broader assortment). It reports as a single operating segment. The company was founded in 1978, listed in 1993, and renamed itself in June 2026 to reflect Shoe Station as the intended long-term growth vehicle.

Revenue has declined in four of the last five years: $1,330.4M (FY2021) → $1,262.2M → $1,175.9M → $1,202.9M → $1,135.3M (FY2025), a 14.7% cumulative decline from the FY2021 peak. FY2025 net sales fell 5.6% on a 5.6% comparable-store decline, with the Shoe Carnival banner down 7.7% while Shoe Station grew 2.7% organically.

As of 2026-05-21, Q1 FY2026 net sales were $270.7M, down 2.5%, with the Shoe Carnival banner's comp decline improving to −2.2% from −7.7% — the first evidence of stabilization in the core banner.

The business is understandable, moderately cyclical, and structurally challenged rather than melting. Footwear retail is a discretionary category exposed to consumer credit conditions and to tariffs on imported goods. Capital intensity is moderate: FY2025 capex of $44.7M was 3.9% of sales, inflated by the abandoned conversion program; the maintenance run-rate is closer to the $33.2M of FY2024. The dominant structural pressure is vendor disintermediation — Nike, Skechers and Crocs all continue to expand direct-to-consumer channels that compete with their own wholesale customers.

Customer concentration is negligible (individual consumers). Supplier concentration is severe and is the defining commercial risk (see the risk matrix below).

3. Moat, Competitive Position, and Industry Cycle

Moat classification: WEAK.

Against a moat, and these dominate. Supplier concentration with zero contractual protection: three vendors are 46% of sales on no long-term contract (10-K Note 16). No pricing power — gross margin has sat in a 35–37% band for five years, and FY2025's 92bp improvement is attributed by management to pre-tariff inventory timing and mix, and is guided to reverse by roughly 260bp in FY2026. Demand is not sticky, with comps down 5.6%. And the rebanner failure is direct evidence against brand transferability: 101 conversions produced "significant variability in in-store performance" and were halted inside a year. A brand with real franchise value travels; this one did not travel reliably.

For a moat, real but limited. A 47-year operating history and a store fleet concentrated in mid-size markets where management describes Shoe Carnival as "the dominant family footwear retailer" — genuine local scale that is uneconomic for a competitor to replicate in a low-density trade area. Shoe Station is a working concept, growing organically and, per management, outperforming the industry for a third consecutive year. A CRM and loyalty asset with real penetration: Shoe Perks members were approximately 78% of comparable-store net sales in FY2025.

Not "None" — the local scale and Shoe Station evidence are real. Not "Narrow" — supplier concentration and the absence of pricing power cap it. A reasonable analyst could argue Declining, given the core banner's −7.7% comp and 18–24 planned closures.

Cycle position: below mid-cycle. Operating margin of 5.9% compares to a 15.6% peak in FY2021 and a pre-COVID FY2019 that was structurally lower still. The FY2021 peak was stimulus-inflated and is not a cycle mid-point.

Can it compound capital at attractive rates? Not currently. ROE is 7.6% and ROIC 11.7% on a 5-year basis. The most recent large capital deployment destroyed value.

4. Management, Governance, and Capital Allocation

As of 2026-02-24, CEO Mark J. Worden departed and resigned from the Board the same day, treated as a termination without cause. Clifton E. Sifford — Vice Chairman, and CEO from 2012 to 2021, covering the company's strongest years — was appointed Interim President and CEO. As of this report's date no permanent CEO has been named; the search has run approximately six months.

As of 2026-08-03, Tracy Dick was appointed Chief Marketing Officer, from PetSmart, Jack in the Box and Leslie's — evidence of bench rebuilding under the interim CEO.

Ownership. J. Wayne Weaver and Delores B. Weaver together hold 8,677,328 shares, 31.5%; all executive officers and directors as a group (nine persons) hold 33.6% (DEF 14A, as of 2026-03-31).

Compensation. Former CEO Worden's FY2025 total compensation was $4,058,347 (FY2024: $5,004,173) — 7.8% of FY2025 net income — including $2,795,577 of stock awards in a year with zero non-equity incentive payout. On departure he received $4.8M in cash plus 168,184 shares through accelerated vesting, generating a $5.3M charge and a $1.6M tax-deductibility penalty. Chairman Weaver received $400,018 for FY2025 board service.

Related-party transactions: clean. The DEF 14A states that during Fiscal 2025 there were no transactions exceeding $120,000 in which a related person had a direct or indirect material interest, and none are proposed. The Audit Committee must approve all related-person transactions and may not approve any unless terms are at least as favorable as those obtainable from an unrelated party.

Insider activity. The most decisive qualitative evidence in this report — all 24 Forms 4 filed in the trailing twelve months were parsed directly.

Insider activity

Three things matter here. The CFO's purchase is the only open-market buy by any insider in twelve months, and he made it after the CEO was terminated, after the rebanner strategy went under review, and after FY2026 guidance was cut — buying into disclosed bad news with his own money. There was exactly one open-market sale all year, of 2,477 shares. And the Weavers' large dispositions are Code G gifts at $0, not sales — their combined holding fell from ~9.33M to ~8.33M shares without a single share being sold into the market, a pattern consistent with programmatic charitable or estate transfers. Any data feed that codes these as insider selling reports roughly 667,000 shares sold and inverts the true signal.

Capital allocation — the core criticism. FY2025's conversion of 101 stores cost approximately $0.66 per share of earnings and drove capex up 35%. Within roughly twelve months the Q1 FY2026 10-Q recorded that the company "is no longer pursuing a single-banner Shoe Station strategy," that "only a limited number of additional Shoe Carnival locations meet the criteria for conversion," and booked $8.3M ($0.23/share) of store-level impairments and property write-offs, with 12–14 closures in FY2026 and 6–10 in FY2027.

Against that: the dividend has been raised for twelve consecutive years (56 consecutive quarterly payments), most recently +13.3% to $0.17/quarter in March 2026 — raised into a down year — and the company repurchased 390,492 shares for $7.0M in Q1 FY2026, its first buyback since FY2023, with $43M of authorization remaining.

Governance Risk: HIGH (two warnings identified in this review). This reads harsher than the facts warrant: there is a single share class with one vote per share, no dual-class structure, a majority-independent board, an independent Audit Committee with related-party approval authority, an explicit finding of zero material related-party transactions, and a controlling family giving stock away rather than extracting from the company. The genuine concerns are narrower: a classified board with three-year terms that entrenches, a strategy reversed at shareholder expense inside a year, and a severance package that richly rewarded its architect.

5. Corporate Ownership, Subsidiaries, and Joint Ventures

Ownership structure

Single class of common stock, one vote per share. The company reports as one operating and reportable segment; no material joint ventures, minority interests, ring-fenced subsidiaries, upstream guarantees or intercompany pledges are disclosed. There is no holding-company structure separating creditors from operating assets. For a company of this size the structure is refreshingly simple, and none of the usual minority-leakage risks apply.

6. Historical Financial Quality and Normalized Owner Earnings

All figures $ thousands, from audited statements in the FY2025 10-K and the FY2024 10-K.

The ROE row is the most important line in this report. Returns fell from 34.2% to 7.6% in four years while book value grew 52%. The company retained and reinvested earnings into a business earning progressively less on capital. That is the mechanical explanation for the multiple compression discussed in the valuation section below, and it is why "cheapest P/TBV in a decade" is a weaker argument than it appears.

Normalized owner earnings. The cash-flow statements support a real, sourced free cash flow figure rather than a proxy: CFO $82.2M (10-year basis) less capex $49.8M gives normalized FCF of ~$32.4M. On the FY2023–FY2024 capex run-rate, normalized FCF would be closer to $50M. Against a $415.4M market cap that is a 7.8%–12.0% normalized FCF yield.

Normalized EPS — three bases, and a material caveat. Diluted EPS split-adjusted:

Annual diluted EPS history

FY2016 EPS is not independently verified against a primary filing — the oldest 10-K retrieved for this report reaches only FY2017; FY2016's Diluted EPS of $0.64 is a reasonable substitution for a plain historical line item like this. One sanity check before use: $0.64 sits as a plausible continuation of the adjacent-year trend (FY2017 $0.575, FY2018 $1.225, neither an outlier) and of FY2016's own revenue ($1,001.1M) against FY2017's $1,019.2M — a smooth, unremarkable step, not a break. The split-adjustment basis also checks out: the FY2017 figure ($0.57) sits within rounding of the report's independently-derived $0.575, so $0.64 has had the same adjustment applied as the other nine years, not a different one.

Normalized EPS basis

The 10-year average is used deliberately here, not the more common 5-year basis. A 5-year basis would label the earnings trajectory "Growing" simply because it compares the 5-year level against the 10-year level, and the stimulus-inflated FY2021 sits at the start of that 5-year window. Within that same window EPS actually fell from $5.42 to $1.90. A 5-year basis inflates normalized EPS by roughly 58% against the 10-year figure and is not a like-for-like read on current earning power. Every basis considered here, trailing, forward, and all three normalization methods, lands in the same cheap range regardless, so this choice does not change the verdict.

6.5 Piotroski F-Score — 5 / 9

FY2025 vs FY2024, ROA and asset turnover on beginning-of-year assets.

Total 5/9 — Eligible (4–9 band). Two disclosures: signal 7 is scored strictly at 0, though the increase is entirely treasury reissuance for vested stock compensation with issued shares constant at 41,049,190 and weighted-average diluted shares essentially flat — a lenient reading gives 6/9. And signals 3, 6 and 9 all failed for one underlying reason: the balance sheet grew (inventory +14%, capex +35%) while sales and earnings fell. That is a genuine negative, not an artifact.

Note: with zero long-term debt in both fiscal years, leverage did not increase, so the leverage signal passes.

7. Balance Sheet, Debt, Covenants, and Refinancing Risk

Classification: NET CASH — the strongest balance-sheet category this report tracks.

Capital structure and liquidity

There is no debt line of any kind on the balance sheet. The liability section runs accounts payable → accrued and other → current portion of operating lease liabilities → long-term portion of operating lease liabilities → deferred income taxes → deferred compensation → other. No notes payable, no long-term debt, no current maturities, no finance-lease liability.

10-K Note 10 confirms a $100M amended and restated revolving credit facility expiring 2027-03-23, inventory-collateralized: "No borrowings were outstanding under the Credit Agreement as of January 31, 2026 or February 1, 2025, and we did not borrow under the Credit Agreement during Fiscal 2025 or Fiscal 2024." At year-end, $1.0M of letters of credit were outstanding with $99.0M available. FY2025 was the 21st consecutive debt-free fiscal year-end. Covenants: minimum net worth $250M against actual equity of $689.7M (a 2.8× cushion), consolidated interest coverage ≥3.0×, and a $20M cap on additional debt.

The lease-versus-debt distinction, stated explicitly. The $371.4M of operating lease liabilities is not borrowed money. It is the discounted present value of contractual store rent, brought on-balance-sheet only in 2019 under ASC 842, offset by a $349.6M right-of-use asset. It carries no financial covenants, no acceleration clause tied to leverage, no refinancing wall, and no lender able to call it. Under the Schloss debt filter this report uses financial borrowings — $0, or 0.0% of equity — while also reporting the lease-inclusive construction at 53.6%. Both pass the ≤100%-of-equity test with room. Any data feed showing ~$371M of "total debt" for this company is capitalizing leases; the honest formulation is zero borrowed money and substantial contractual rent.

That said, the rent is real and non-cancellable: $371M of obligations against a $415M market cap, and closing an underperforming store does not extinguish its lease. The 18–24 planned closures will carry exit costs.

Refinancing risk: none. There is nothing to refinance. Interest coverage is 164× — a figure that only makes sense once you understand there is no interest-bearing debt.

8. Real Estate, Leases, and Hidden Assets

The company leases essentially all 426 stores; it is not a property owner. Right-of-use assets are $349.6M against $371.4M of lease liabilities — the small excess of liability over asset is normal amortization timing.

One caveat that matters to the valuation, and it cuts against the thesis. Tangible book value of $614.5M includes that $349.6M of right-of-use assets — 57% of the total. An ROU asset is not realizable in a liquidation the way inventory, cash or owned real estate is; it is the capitalized right to occupy space the company must also pay for. Tangible book value should therefore not be read as a liquidation floor. The genuine downside support is the $129.3M of net cash plus inventory, not the full TBV figure. This distinction is not made by standard screens, and it materially qualifies the asset-value case.

No material hidden assets, sale-leaseback optionality or under-marked property are identified.

9. Capital Markets Access, Dilution, and Financing Flexibility

The company does not rely on capital markets to fund operations and has not issued equity for capital. FY2025 cash proceeds from stock issuance were $172k, entirely employee stock purchase plan. Shares outstanding rose 199,871 (0.7%) via treasury reissuance for vested compensation; weighted-average diluted shares were essentially flat at 27,524k → 27,535k.

Financing flexibility is excellent and entirely unused: $99.0M available on the revolver, $129.3M of cash and securities, and $43M remaining on the buyback authorization. Dilution risk is minimal and limited to ongoing equity compensation.

10. Litigation, Regulatory, and Contingent Liability Risk

No material litigation is disclosed. 10-K Note 16, verbatim: "From time to time, we are involved in certain legal proceedings in the ordinary course of conducting our business. While the outcome of any legal proceeding is uncertain, we do not currently expect that any such proceedings will have a material adverse effect on our consolidated balance sheets, statements of income, or cash flows." There is no separate commitments-and-contingencies note carrying loss accruals and no litigation reserve on the balance sheet.

Tariffs are the live regulatory exposure and are explicitly unquantified. In February 2026 the US Supreme Court struck down certain IEEPA tariffs and the executive branch reimposed tariffs under other statutes. 10-K Note 17 states there is "considerable uncertainty regarding the scope and duration of current and potential tariffs... including availability and timing of refunds of tariffs paid under IEEPA," and that the company "cannot reasonably estimate the total financial impact." This cuts both ways: potential IEEPA refunds are an unpriced call option, while renewed tariffs are a cost headwind already embedded in the guided 260bp gross-margin decline.

11. Accounting Quality and Disclosure Review

Classification: HIGH.

Unqualified audit opinion, no going-concern language, no material weaknesses, no restatements in the lookback period. Segment reporting is simple (one segment). Related-party disclosure is explicit and clean. Revenue recognition is straightforward point-of-sale retail. There are no pension assumptions of consequence and no aggressive capitalization policies identified.

Non-GAAP usage is present but appropriately reconciled and not misleading: Q1 FY2026 adjusted SG&A excludes $13.6M of non-recurring charges from the CEO transition and the strategic review, and the company reports both GAAP EPS of $(0.21) and adjusted EPS of $0.23 side by side. The adjustments are genuinely non-recurring in character.

Two items warrant attention rather than concern. The Q1 FY2026 effective tax rate of (11.2)% versus 28.1% a year earlier is a benefit that flatters the reported loss — the pre-tax result is worse than the after-tax figure suggests. And goodwill of $18.0M plus intangibles of $40.9M arising from the Shoe Station acquisition sit on the balance sheet while the strategy built around that brand has just been curtailed; the FY2025 impairments were taken against store-level long-lived assets, not against the brand intangible. That is defensible — the Shoe Station banner is growing — but it is the asset most exposed to a further write-down if the concept stalls.

12. Valuation and Margin of Safety

Current price $15.50 (2026-08-17 close, the day the position was initiated) · 27,151,308 shares · market cap ~$417M.

Tangible book value — filing-derived bridge

TBV/share is shown and proven from the filing bridge below, not asserted:

Equity (2026-05-02, 10-Q)          $673,400k
less Goodwill                       (18,018k)
less Intangibles (ex-goodwill)      (40,911k)
= Tangible Book Value               $614,471k
TBV/share = 614,471 / 27,151,308 = $22.6314
P/TBV = 15.50 / 22.6314 = 0.6849x

Net cash of $129.3M = $4.76/share = 31.1% of the share price. Enterprise value is $286.1M, putting EV/EBIT at 4.29× on FY2025 operating income (unchanged — EBIT and net cash are filing-based, not price-dependent).

Historical Relative Multiple Valuation

From the company's ten-year trading history — 41 quarterly P/TBV observations, 2016-04 to 2026-04, and 40 P/E observations over the same window.

(Distribution columns — Window Floor/Q1/Median/Q3/Window Ceiling/Upper-quartile mean — are the same 41-observation trading-history window as at original publication, 2016-04 to 2026-04; no new quarter has printed. Only the "Current" column moves, with the 2026-08-17 close.)

Recent P/TBV trajectory: 2.13 → 1.66 → 1.28 → 0.82 → 0.94 → 0.82 → 0.84 → 0.67.

The current P/TBV of 0.685× sits just above the ten-year window floor of 0.673× (the April 2026 observation) — the stock has traded fractionally cheaper only once in the available ten-year history, and today's reading sits 54% below its ten-year median.

The two legs diverge sharply, and the reason is the whole analysis. On P/TBV the stock is near an all-time low; on P/E it is at 11.48× against a 12.43× median — only modestly cheap. The divergence is not noise: earnings collapsed while book value grew. TBV per share rose $19.86 → $21.71 → $23.04 across FY2023–FY2025 while EPS fell $2.68 → $1.90. A P/E computed on depressed earnings looks unremarkable; a P/TBV computed against a rising book looks extreme. The book-value leg is doing the work, and it is doing it because the market has repriced the return on that book, not merely the book.

This is why the historical band cannot be used as a price target. Multiplying TBV by the 1.244×–1.762× typical band implies $28–$40 per share. That band was earned by a business generating 20–34% ROE. At FY2025's 7.6% ROE and a 9–10% cost of equity, a defensible multiple is roughly 0.75×–0.85× book, implying $17–$19. Reversion to the historical band requires an ROE recovery for which there is currently no evidence, and underwriting it would be assuming the conclusion.

Intrinsic value — four anchors

Intrinsic-value anchors

Illustrative anchor range (analyst cross-check, not the Sell Range): $18 – $25, central ~$21 (≈0.94× tangible book). These four anchors are static — TBV, the 10-year normalized EPS, and FY2026 guidance are all filing/guidance-based, not price-dependent. This range is informational only. The Sell Range below, not this range, is what governs.

Scenario table

• Severe downside — $11.32/share: Nike relationship impaired, comps −10%, and ROE falls toward −4%.

• Bear — $15.50/share: FY2026 guidance missed and margin falls below 34%; no re-rating.

• Base — $18.11/share: FY2026 delivered at $1.50 and ROE stabilizes around 7–9%.

• Bull — $24.89/share: A permanent CEO is installed, ROE recovers toward 12%, and closures complete.

Sell Range

Sell Range = [median P/TBV, top-quartile-mean P/TBV] × current TBV/share, derived from the same 41-observation distribution above:

• Low = median 1.475× × $22.6314 = $33.38 → $33

• High = top-quartile mean 2.244× × $22.6314 = $50.79 → $50

• Sell Range: $33 – $50 (1.48× – 2.20× TBV), 41 observations so not data-limited.

Buy-Below: 0.80× TBV = 0.80 × $22.6314 = $18.11. This is an entry guide, not the Sell Range above: it marks the price below which the shares are cheap enough to warrant a larger position, not a price to sell at.

Buy More Below: $14.71 (0.65× TBV). Two capital-allocation red flags are noted separately — concentrated family and board control, and the capital sunk into the failed re-banner program — and are a matter for judgment alongside the price, not folded into it.

Valuation summary

• Current price: $15.50

• Illustrative anchor range: $18–$25, central ~$21

• Sell Range: $33–$50 (1.48×–2.20× TBV)

• Buy-Below: $18.11 (0.80× TBV)

• Buy More Below: $14.71 (0.65× TBV)

• Cheapness type: Asset-value cheap, with genuine value-trap risk

The cheapness is real but its character matters: this is not a high-quality compounder on sale, and it is not a pure statistical bargain either. It is an asset-value situation with a fortress balance sheet and a deteriorating earnings stream, backed by net cash and tangible assets rather than franchise quality or earnings power.

13. Risk Matrix

Risk matrix

• Critical — Supplier concentration: Nike is 24% of sales and the top three suppliers are 46%, with no long-term contracts. Monitor vendor commentary and DTC expansion.

• High — Continued earnings decline: Operating margin fell from 15.6% to 5.9% and ROE from 34.2% to 7.6%. Monitor quarterly comps and gross margin against the 34% guide.

• High — No permanent CEO: The company has been operating with an interim CEO since 2026-02-24. Monitor the appointment announcement.

• High — Gross margin guide: FY2026 guidance is roughly 34% versus 36.6% previously. Monitor the Q2 report, expected in early September.

• Medium — Tariffs: The 10-K says the impact cannot reasonably be estimated. Monitor trade rulings.

• Medium — Further store impairments: $8.3M has been taken and 18–24 closures are planned. Monitor closure cadence.

• Medium — Shoe Station intangible write-down: $18.0M of goodwill and $40.9M of intangibles remain tied to the curtailed strategy. Monitor the annual impairment test.

• Medium — Governance entrenchment: A 31.5% block and classified board limit outside influence. Monitor proxy proposals.

• Medium — ROU assets overstate liquidation value: $349.6M of $614.5M in tangible book is right-of-use assets.

14. Red Flags, Yellow Flags, and Green Flags

🟢 Green Flags

• Zero financial borrowings; 21 consecutive debt-free fiscal year-ends

• $129.3M net cash — 31% of market cap; current ratio 4.02×

• CFO bought 31,000 shares (~$500k) at $16.13 in April 2026, after the bad news

• 12th consecutive annual dividend increase, +13.3%, raised into a down year

• Buybacks resumed — 390,492 shares for $7.0M, first since FY2023, $43M remaining

• Positive operating cash flow ($23.1M) despite a Q1 GAAP loss — the loss is substantially non-cash

• Zero material related-party transactions; single share class

• Cheapest P/TBV in the ten-year record

• Shoe Perks loyalty ≈ 78% of comparable-store sales

🟡 Yellow Flags

• Revenue down four of five years; −14.7% from the FY2021 peak

• Piotroski 5/9 — the four failures are all deterioration signals

• FY2026 gross margin guided down ~260bp

• Q1 tax benefit of (11.2)% flatters the reported loss

• $18.0M goodwill + $40.9M intangibles tied to a curtailed strategy

• 57% of tangible book is right-of-use assets, not liquidation-realizable

• Classified board entrenches a 31.5% holder

• Most recent hard financials are 3.5 months old; Q2 unreported

🔴 Red Flags

• Nike = 24% of sales, top three = 46%, on no long-term contracts

• ROE collapsed 34.2% → 7.6% in four years

• Capital allocation failure: 101 conversions, ~$0.66/share cost, written down and abandoned within twelve months

• $4.8M severance plus accelerated stock to the CEO who authored the abandoned strategy

• No permanent CEO six months into an interim arrangement

• Operating margin down nearly two-thirds from FY2021

⚡ Must-Watch Catalysts

• Q2 FY2026 results, due early September 2026 — first read since guidance was reaffirmed

• Permanent CEO appointment

• Gross margin against the ~34% FY2026 guide

• Execution of the remaining $43M buyback at sub-book prices

• Store-closure count versus the guided 18–24

• Any change in Nike allocation or terms

• Tariff rulings and potential IEEPA refunds

• Further insider open-market purchases

FINAL VALUE-INVESTING RECOMMENDATION

Verdict: BUY — STANDARD

Recommendation summary

Why it may work

• The stock is cheap against filing tangible book and normalized earnings, with net cash and a meaningful shareholder return.

• The house buy-below is $18.11; the illustrative sell range is $33–$50 if the valuation normalizes.

• The upside case is mean reversion, not a heroic earnings-growth forecast.

What keeps it at STANDARD

• Two warnings fired: concentrated control/classified governance and the failed rebanner capital-allocation program.

• Operating performance has deteriorated, supplier concentration is material, and the company still lacks a permanent CEO.

• The balance sheet limits some downside, but tangible book is not the same as liquidation value because leases, inventory, and intangibles matter.

What would change the verdict

• Q2 FY2026 results and gross margin against the roughly 34% guide.

• A permanent CEO appointment and evidence of better capital allocation.

• Buyback execution, store closures, Nike terms, inventory, dividend policy, and any revolver draw.

VALUE INVESTOR MUST-WATCH LIST

• Q2 FY2026 results and gross margin against the guided ~34%.

• Permanent CEO appointment and that person's capital-allocation record.

• Buyback execution, store closures, inventory, and Shoe Carnival/Shoe Station comps.

• Nike, Skechers, and Crocs allocation, terms, and DTC expansion.

• Dividend declarations, tariff rulings, and any revolver draw.

• Goodwill and Shoe Station intangible impairment testing at year-end.

SOURCES AND DILIGENCE GAP LOG

Sources accessed

• SEC EDGAR: FY2025 10-K, Q1 FY2026 10-Q, DEF 14A, FY2025 results 8-K, 24 Forms 4, and historical 10-Ks.

• Business Wire: Name/ticker change announcement dated 2026-06-11.

• Market-data cross-check: Price, market cap, and 52-week range retrieved 2026-08-16.

• Derived calculations: TBV bridge, current P/TBV, and the 41-observation historical P/TBV distribution.

• Data quality: EXCELLENT.

Diligence gaps

• Q2 FY2026 results: Not yet filed; most recent hard financials are 3.5 months old; expected early September.

• Permanent CEO: Identity and search timeline not disclosed.

• FY2026 capex guidance: Not guided; Q1 actual was $10.4M.

• Tariff impact: Issuer says it cannot be estimated.

• FY2016 EPS: Resolved from a secondary historical line item; oldest retrieved 10-K reaches FY2017, so this figure is not independently filing-verified.

• Q1 buyback average price: Estimated from secondary commentary; aggregate $7,002k is confirmed in the cash flow statement.

• TBV/share basis: Filing-derived Q1 FY2026 10-Q bridge governs at $22.63/share; a secondary dataset used a different period/basis.

• Piotroski signal 7: Judgment disclosed; strict score 5/9, lenient reading 6/9.

• Normalized-EPS basis: Judgment disclosed; the 10-year basis is used deliberately. Overall Data Quality Rating: EXCELLENT

Both the annual (10-K, FY ended 2026-01-31) and interim (10-Q, quarter ended 2026-05-02) filings were directly fetched and read from SEC EDGAR — not summaries — along with the DEF 14A, the 8-K earnings exhibit, four earlier 10-Ks, and the complete XML of all 24 trailing-twelve-month Forms 4. The litigation, debt and related-party notes were each located and read in the actual filing. Aggregators were used only for the live share price, cross-checked across two sources and validated against an independently computed TTM EPS.

Research only, not investment advice. Long, ~0.5% of portfolio (initiated 2026-08-17, 33 shares, average fill $15.49 — a cent below the $15.50 close used throughout this report's valuation math).