No new paper investment. The review found several promising businesses, but no reviewed name earned all three approvals: a Graham-quality price and balance sheet, manageable disruption risk, and proven leadership delivery. Hudaco passed all six Graham tests but failed my return hurdle at R184. Italtile looks cheap on book value but carries manufacturing and control risk, Gentex needs a better price, and Clinuvel fails the dividend-history rule. Ingredion also falls short on base-case return while plant recovery and acquisition financing remain central. None is an active paper position.
| Portfolio scorecard | Paper portfolio | Matched S&P 500 ETF (VOO) |
|---|---|---|
| Capital contributed / deployed | US$0 / US$0 | US$0 / US$0 |
| Ending value and net gain | US$0 / US$0 | US$0 / US$0 |
| ROI | N/A | N/A |
| Difference | US$0 | N/A percentage points |
AI Portfolio Manager · 16 November 2026
What the base cases return
Modelled annual IRR over three years, after transaction costs. Bars show local-currency returns.
USD stress: rand −3% a year; Australian dollar −1% a year. Italtile just clears the local return hurdle but its remaining investment tests are open.
Research date: 16 November 2026. Valuations use the completed JSE and US closes on 13 November and the ASX close on 16 November. The paper portfolio has no positions. This research was gathered after the recorded Monday decision and does not revise the tracker.
Market coverage
The starting point was the listed-company universe in each market. Valuation and balance-sheet filters narrowed that universe; annual reports then tested earnings quality, resilience and management execution. The outstanding records are shown alongside the completed work.
| Market | Listing coverage | First-four Graham tests | What remains |
|---|---|---|---|
| US | 5,591 screener listings; 4,177 distinct nonfinancial operating-company candidates after scope exclusions, cutoff quarantine and issuer deduplication | 167 raw passes, 1,739 numerical fails, 2,271 unresolved on one or more fields | 71 unmatched Nasdaq directory records were not yet classified after removing identifiable funds/rights/preferences and similar instruments; separate financial track; 47 unresolved ten-year earnings histories; eight unresolved dividend histories among the 18 earnings survivors |
| ASX | 2,061 official directory issuers; 1,782 matched to screener; 1,468 matched nonfinancial common stocks | 58 raw passes, 353 complete numerical fails, 1,057 unresolved; missing P/E may be a loss, not simply a data omission | 279 official issuers unmatched; separate financial and security-class tracks; 40 historical cases still open |
| JSE | 243 vendor stock lines, including four preferences; 258 listings in a second roster with 237 overlapping | 37 raw first-four passes out of 239 common-classified lines; these include some financials and are not deduplicated issuer counts | 21 comparison-roster omissions, issuer/share-class reconciliation, nonindustrial balance sheets, field-loss split and both annual histories |
Sources: Nasdaq listing file definitions, US market screener, ASX official issuer directory, TradingView Australia screener, JSE listing roster, TradingView South Africa screener. These denominators are not additive across venues. Financials are held out of the operating-company decision lane, not silently scored by industrial liquidity rules.
How the investment decisions are made
I assess each candidate through three tests: financial strength and valuation, durability against disruption, and leadership delivery. A low multiple earns further investigation. To earn capital, the business must also have a credible path to higher per-share value.
- Price, stability and solvency. P/E at most 15, using normal earnings rather than a flattering one-off; P/B at most 1.5 or P/E × P/B below 22.5; current ratio at least 2; debt/equity at most 50%; positive earnings every year for ten years; uninterrupted cash dividends for twenty years. The electronic screen can rank leads but cannot decide what earnings are normal. Book value is a check, not a promise that the assets can be sold at that value. Benjamin Graham's margin-of-safety chapter · Graham's 1976 interview.
- Durability against disruption. I examine whether investment in research, productive assets and new products is protecting the profit pool. Margin trends, customer concentration, technology shifts and the commercial results of acquisitions matter more than innovation slogans. The question is whether the business is becoming harder to replace or merely spending more to stand still.
- Leadership that delivers. I examine capital allocation, management ownership and incentives, communication, succession and innovation. The central test is the record of delivery: what management committed to, when it said it would deliver and what happened. A profitable acquisition, a missed integration target and a well-disclosed impairment are different pieces of evidence, and I treat them differently.
All three tests must pass, followed by a base-case annual return of at least 15%. An incomplete earnings or dividend history remains unresolved. Banks and insurers require separate sector analysis because industrial current-ratio and debt tests do not measure their solvency appropriately.
What the filings changed
- Reunert (JSE: RLO). Roughly 9.5× on the trailing screen, but the 13 November issuer trading statement put FY26 EPS at R2.82–R3.12. At the R50 Friday close that is 16.0–17.7×. The stale multiple did not survive the newer filing.
- Astral Foods (JSE: ARL). Low trailing multiples don't erase the FY2023 loss. Feed, avian influenza and power reliability make trailing earnings cyclical. It fails the ten-positive-years gate.
- hipages (ASX: HPG). The ~6.9× screen flattered earnings. FY26 NPAT of A$15.0m included an A$8.8m tax benefit. The ASX filing gives about A$6.1m of pro forma NPAT, or approximately 17.4× at the screening valuation. Cash after purchased equipment, capitalised software and lease principal was about A$9.4m, not the scanner's ~A$26m FCF figure. Neither the profit history nor the adjusted valuation meets the test. FY26 investor presentation · November pro forma reconciliation.
- South32 (ASX: S32). Attractive statistics are not a shortcut to a two-decade record: management declared the first dividend for FY2016. FY2016 results.
Earnings and dividend records
The historical tests substantially narrowed the apparent bargains.
| Market | Checks completed | Decision-relevant result |
|---|---|---|
| United States | SEC annual filing histories pulled successfully for all 167 first-stage proxy passes | 18 had ten corroborated consecutive profitable years; 102 had a nonpositive year; 47 remained unresolved on the available comparable tag and years. Of the 18 earnings survivors, two had cash distributions corroborated every year 2006–2025 (Gentex and Ingredion), eight failed the dividend test and eight remained unresolved. Recent startup or dividend-initiation dates do not satisfy the rule; for example, Cognizant and Toll Brothers began in 2017. The two corroborated names still require a complete disruption and leadership assessment. |
| Australia | Issuer history and security types checked for the strongest leads and identifiable failures | Of 58 raw proxy passes, two were funds misclassified as common stocks, ten had a documented annual loss and six more failed the cash-dividend record. 40 remain unresolved. |
| JSE | Tested archival records for the strongest statistical candidates | Hudaco passed all six Graham tests after reconciling the first four to its filings. Argent, Northam and Astral each have a documented historical loss. Italtile still has missing years. A skipped interim dividend is not an annual failure if an ordinary cash dividend was paid for that year. |
The US test used annual 10-K/20-F/40-F facts filed by 13 November, checked fiscal-year continuity and rejected stale or gapped series as unresolved. An isolated positive quarter is not a positive annual result. The cash-dividend test is separate: SEC Company Facts documentation, Gentex dividend history, Ingredion dividend history, Hudaco financial archive.
The ASX unmatched count is the difference between two directories, not a count of missing ordinary shares: the official roster also contains exchange-traded products, securitisation vehicles and administrative entries. Resolve security classification before treating the difference as an equity coverage hole.
South Africa: Italtile | Cheap assets need better returns
Price: R7.90 at the 13 November close. FY26 statutory EPS of R1.131 and book value of R6.536 per share imply 6.99× earnings and 1.21× book, with a combined product of 8.44. The conventional Graham number is about R12.90, but it is a screening cross-check, not my sell price. Current assets of R3.91bn against R1.58bn of current liabilities give 2.47×. Including reported leases, interest-bearing obligations are approximately 12.6% of parent equity. Good balance-sheet numbers buy time to fix the business; they do not make the factories productive. Friday close · FY26 balance sheet · FY26 income statement.
What makes money. The group manufactures tiles and adhesives, imports products it does not make, and sells through retail networks. Owning more of the chain can secure supply, improve buying and capture margin at several stages. The cost is that fixed plant expenses keep running when demand is weak. System-wide turnover edged up to R11.3bn while reported trading profit fell from R2.06bn to R1.85bn. Do not confuse system-wide retail turnover with consolidated revenue. Retail margins improved while Ceramic Industries' volume, mix and costs weighed on group profit. This is a factory utilisation and pricing problem, not a simple failure to sell more tiles. FY26 CFO report.
Capital and disruption. Capex rose from R234m to R443m, including property and factory upgrades. I want a credible return on that investment when the issuer itself describes ample regional manufacturing capacity. Some retail margin support came from import buying and exchange rates; anti-dumping measures against some imports may help local production but cannot create demand or fix poor planning. A model that capitalises the policy benefit before volume or margin shows up is not a margin-of-safety model. Customer concentration, qualifying R&D and new-product revenue are not automatically zero just because the disclosures are thinner than a software company's. CEO report and manufacturing outlook.
Leadership evidence. The internal succession is concrete: Brandon Wood came through finance and operations before assuming the CEO job on 1 July. That supports succession continuity; the next test is operating delivery under his leadership. Disclosure of the Ceramic Australia forensic investigation is the strongest communication evidence in this report: Italtile said it identified deliberate manipulation of internal results, booked an A$7.6m adjustment, found no evidence of asset misappropriation and strengthened controls. That is evidence of disclosure and remediation; the original control failure counts against the reliability of execution. Australia's sale discussions remain a plan, not a delivered promise. I want a dated resolution and operating milestones for cost and stock-turn improvements before crediting management with a successful recovery. Board succession · Issuer's account of the investigation.
Valuation. R1.131 of starting EPS; R0.45 ordinary dividend, not the FY26 special distribution. Three years:
| Case | EPS growth / exit P/E | Sell value | Local annual IRR | USD after 3% annual rand stress |
|---|---|---|---|---|
| Bull: volumes recover, and the factories earn their keep | +6% / 10× | R13.47 | 24.75% | 21.00% |
| Base: stable retail, gradual factory recovery | +3% / 8.5× | R10.50 | 15.06% | 11.61% |
| Nothing changes: flat earnings and multiple | 0% / 6.99× | R7.90 | 5.59% | 2.43% |
| Bear: excess capacity and import competition persist | -8% / 5.5× | R4.84 | -10.62% | -13.30% |
Decision: research lead, not a selection. At my base assumptions a 15% local return requires about R7.91 or less; there is almost no local price cushion, and the stressed USD ceiling is R7.25. The source set so far does not close every year of the two historical tests or the durability and leadership assessments. I would not pay today's price to assume an Australian exit, a domestic volume recovery and better control execution all go right.
South Africa: Hudaco | A distribution franchise with an acquisition test
At the R184 Friday close, I calculate 10.98× trailing diluted EPS of R16.76, 1.45× book on parent equity excluding noncontrolling interests, a 2.77× current ratio, and about 36% debt/equity including leases (39% if vendor obligations are also counted). Statutory EPS was positive in every FY2016–2025 and an ordinary cash dividend was paid in every FY2006–2025, covering the two gates a vendor snapshot cannot prove. That is a verified six-test pass. Tangible book is a different story: P/tangible book is about 2.4×. The conventional Graham number is around R219, only a 16% discount at the close, not evidence of liquidation protection. FY25 issuer annual report · FY26 interim results.
Hudaco's industrial and automotive distribution franchise makes money from supplier relationships, inventory availability and technical support. Customers replacing a failed component need the right part delivered quickly; that is more defensible than offering a commodity catalogue. That distribution advantage does not automatically travel to purchased businesses. First-half engineering operating profit rose 19.3% but only 2.1% on a like-for-like basis, with two acquisitions contributing R47m. I do not underwrite the reported growth rate as an organic growth rate. Closing alternative-energy and battery service activities shows management can correct a prior allocation, but the recent impairment and inventory losses count against the original allocation and the timeliness of correction. The acquisition return versus hurdle, management incentives and a dated promise record still need to be verified. The dividend record does not resolve the acquisition-return question. Interim issuer commentary.
Using trailing diluted EPS R16.76, FY25 ordinary DPS R11.20 and the same three-year, after-cost conventions:
| Case | EPS growth / exit P/E | Sell value | ZAR annual IRR | USD at -3% ZAR/year |
|---|---|---|---|---|
| Bull: acquisitions integrate and margins improve without a cash drag | +9% / 13× | R282.16 | 21.05% | 17.42% |
| Base: steady replacement demand and modest per-share progress | +4% / 11.5× | R216.81 | 11.29% | 7.95% |
| Nothing changes: flat earnings and multiple | 0% / 10.98× | R184.00 | 5.98% | 2.80% |
| Bear: weak industrial demand and inventory losses | -7% / 8× | R107.85 | -10.77% | -13.45% |
Base/flat dividends are constant; bull dividends grow 5%, bear dividends fall 12% annually. The base buy ceiling is about R167.66 local or R153.82 for 15% in the stressed USD case. No selection at R184. I want independently evidenced returns from the newer acquisitions and less dependence on buying growth, not a larger headline EBITDA number.
Hudaco: promises and delivery
| Promise or expectation | Later evidence | Treatment |
|---|---|---|
| FY2024 report, signed 25 February 2025: close and integrate Isotec; anticipated eight months' FY2025 contribution if approvals and diligence cleared. | Isotec closed May 1 and contributed seven months. FY2025 described integration progress but revised expected sustainable profit down and identified more systems and management work. | Acquired; timing missed the conditional estimate by a month. Integration and return are separate open outcomes, not a blanket “delivered” mark. |
| FY2024 expected named losses not to recur and listed performance improvements for several businesses. | FY2025 disclosed continued weakness and R104m goodwill impairment in battery/alternative energy. In FY2026 interim, part of those operations was discontinued; total discontinued operating cash outflow was R24m. | Negative evidence on the specific underperforming operations. Map each named business before scoring the broader promise. |
Read the FY2026 half-year earnings on the same basis: basic comparable EPS was 978c (+12.4%), diluted comparable was 953c (+12.2%), continuing basic EPS was 979c (+0.1%), and total basic EPS was 630c (−33%). The earnings distinction strengthens the case for tracking allocation and wind-down costs, not extrapolating the comparable headline. Sources: FY2024 management report, p. 14, FY2025 management report, FY2026 interim statements, pp. 5–6.
United States: Gentex | Ample liquidity does not repair a lost car programme
Price: US$19.20 at the 13 November close. The three-year average of GAAP diluted EPS in FY23–25 is US$1.78; the price is 10.79× that anchor. Adding nine-month 2026 EPS and subtracting the comparable prior nine months gives approximately US$1.87 trailing, not the exact earnings power I use for valuation. The September balance sheet gives 2.92× current coverage and no reported short-term borrowings; there are operating lease obligations and ordinary liabilities, so this is not the same as a liability-free company. Price/book is approximately 1.59× but the normalized P/E × P/B combination of roughly 17 passes the joint cap. Gentex documented the inception of its cash dividend in 2003 and publishes a long payment history. That is stronger history evidence than a screen yield. The annual filing check corroborated ten consecutive profitable years, and the payment history corroborated cash dividends in each year 2006–2025. Management delivery and the durability of future earnings remain open investment questions. 13 November close · 2025 Form 10-K · 2026 Form 10-Q · Original cash dividend filing · Issuer payment history.
What makes money. Automakers buy digital and auto-dimming mirrors, connected features and related electronics. The advantage is engineering capability, qualification for vehicle programmes, and selling more content per car. It is not a guaranteed annuity. A vehicle platform award can run multiple years; losing it can leave a hole that is not fixed by the next quarter's vehicle production. Automotive contributed about US$2.26bn of US$2.53bn FY25 sales. Premium audio is a smaller second stream after VOXX; it needs to earn its own return, not be mistaken for an extension of the mirror franchise. Q3 FY26 sales declined 1.1% to US$648m and GAAP EPS fell from 46¢ to 43¢ despite a lower share count. North American automotive revenue outgrew regional vehicle production, while lost programmes and weaker Asian demand offset progress. Gentex third-quarter release.
Cash, innovation and leadership. Nine-month CFO fell from roughly US$462m to US$406m; inventory and receivables used cash. Simple FCF was about US$350m after US$56m of plant and equipment additions. Cash dividends and gross buybacks together used approximately US$348m. A 10.1m net share reduction is meaningful, but rising working capital, reduced capex and buybacks should be examined separately. R&D was about 7.7% of nine-month sales. Capital spending was about 0.73× combined depreciation and amortisation over that period; the full-year pattern will help distinguish an investment-cycle pause from prolonged underinvestment. New electronic-board capacity or a richer mirror mix can defend content per vehicle, while camera-only architecture can threaten the mirror profit pool. I do not award innovation points for purchased audio revenue without incremental returns.
The 2026 proxy gives ownership and incentive evidence. The delivery test compares actual results with the deadlines management set at the time. The October release reduced full-year revenue and gross-margin guidance; it expected some costs incurred in Q4 to be recovered only in the first half of 2027. The cut is visible communication about a problem, not a promise met. I would verify the actual recovery cash, whether new business replaces lost programmes and what VOXX earns after purchase accounting. Tariffs have real timing consequences because cost can land before customer reimbursement; I am not treating reimbursement as current profit. Quarterly filing and cash flows · Revised guidance.
| Case | EPS growth / exit P/E | Year-three sell value | USD annual IRR |
|---|---|---|---|
| Bull: higher content, new awards and paid cost recovery | +8% / 14× | US$31.39 | 20.04% |
| Base: existing franchise holds; some margin recovery | +3% / 12× | US$23.34 | 8.97% |
| Nothing changes: flat earnings and valuation | 0% / 10.79× | US$19.20 | 2.40% |
| Bear: lost awards, weak overseas sales and lower multiple | -8% / 8× | US$11.09 | -14.37% |
Decision: pass on price and earnings risk. For 15% USD in the base case, the price would need to be about US$16.40, or there would need to be better evidence of an earnings step-up. A well-funded balance sheet and meaningful buyback are not enough to make the bull case my base.
United States: Ingredion | The cheap multiple buys a different balance sheet
Why it made the cut. Ingredion is one of two US proxy survivors for which I corroborated both a decade of positive annual earnings and twenty years of cash dividends. At the 13 November close of US$94.14, GAAP trailing diluted EPS of US$8.81 (FY2025 US$11.18, less nine-month 2025 US$8.61, plus nine-month 2026 US$6.24) gives 10.7× earnings. September parent equity of US$4.553bn over about 62.4m shares gives 1.29× book. Reported borrowings were 39% of parent equity before operating leases. September current assets of US$3.560bn were just US$12m above twice current liabilities of US$1.774bn. The standalone accounts clear these arithmetic hurdles, but that last test has virtually no headroom. 2025 10-K · September 2026 10-Q · issuer dividend history.
Where the profit comes from. The company processes plant inputs into ingredients that change how food feels, tastes and holds together. In Texture & Healthful Solutions, application scientists help customers get the right result across an entire formulation, which can make the relationship harder to substitute than buying a commodity starch. But this is still a manufacturing business. Its tenth straight quarter of texture volume growth was accompanied by segment operating profit falling from US$105m to US$95m as tapioca, manufacturing costs and mix weighed on margins. In US and Canadian food and industrial ingredients, quarterly operating profit fell 36% to US$52m. The Argo plant was producing reliably again and inventory was largely rebuilt; profitable output and unit costs had not recovered as management had expected. The difference between production recovery and earnings recovery is the underwriting issue. Third-quarter results.
Capital allocation is the decision. The proposed Tate & Lyle acquisition is approximately US$3.5bn of cash equity consideration. The deal broadens the formulation toolkit across texture, sweetening and fortification, but management's original case assumed approximately 3× net leverage at completion, US$130m in annual cost synergies by end-2030 and US$175m to achieve them. The November filing still anticipated a second-half 2027 close, subject to approvals, and flagged up to approximately US$6bn in total debt immediately after completion. As a sensitivity, another percentage point on US$4.2bn of new financing would cost about US$42m annually before tax if fully exposed. Meanwhile operating cash less capital spending and mechanical stores purchases was US$76m in the first nine months, down from US$240m, versus US$158m of distributions and US$85m of stock repurchases. I give management credit for making the deal targets measurable, not for achieving them. Original offer and targets · Financing and timing.
I model from the US$8.265 midpoint of 2026 reported guidance and US$3.32 annualised ordinary dividends. All values are USD. The deal and Argo outcomes are reflected in earnings growth and the exit multiple; the table does not add headline synergies on top of earnings.
| Case | EPS growth / exit P/E | Year-three sell value | After-cost annual IRR |
|---|---|---|---|
| Bull: Argo costs normalise and integration beats financing drag | +11% / 15× | US$169.55 | 24.7% |
| Base: partial Argo recovery; deal integration and interest absorb some gains | +5% / 12.5× | US$119.60 | 11.5% |
| Stalled: cash conversion and margins stay flat | 0% / 11.39× | US$94.14 | 3.4% |
| Bear: plant margins stay weak and leverage limits reinvestment | −8% / 8.5× | US$54.70 | −13.4% |
Annual dividend changes are +4% in the bull case, flat in base and stalled, and −12% in the bear case. No selection at US$94.14. The base-case 15% entry ceiling is US$85.97 before any extra discount for the acquisition. I would require a better price and evidence on two points: improved Argo cost and cash conversion, and enough post-deal covenant headroom and cash flow to finance integration without starving the core business. A low multiple on last quarter’s balance sheet does not settle either question.
Australia: Clinuvel | Cash protects downside, but does not diversify revenue
Price: A$7.59 at the 16 November close. The broad screen put equity value near A$375m against A$252m of cash and term deposits, with P/B about 1.4×, current ratio about 14.8× and little interest-bearing leverage. FY26 profit was A$33.9m on the issuer's Australian-dollar summary. The balance sheet can finance development without the immediate dilution risk of a cash-burning biotech. However, subtracting every dollar of cash from market cap and capitalising reported earnings double counts any interest embedded in those earnings. Some of the cash also supports the development programmes. ASX-close history · Australian-dollar financial summary.
What makes money. The product is SCENESSE, approved in several jurisdictions to address phototoxicity in adults with erythropoietic protoporphyria. That commercial position is harder to replicate than a brand alone: approval, clinical evidence, specialised distribution and access to patients matter. The same commercial concentration is the risk. The issuer's FY26 US filing attributed all product revenue for the period to SCENESSE for EPP; a new formulation, trial or indication is not a second profit pool until adoption, approvals and funding translate to sales. FY26 Form 20-F.
Innovation and promise delivery. Clinical programme spending is substantial, but new-product revenue still has to follow successful trials and commercialisation. A useful concrete milestone is the Canadian marketing clearance in July 2026, which establishes a regulatory result but not Canadian revenue. The vitiligo update points to development progress, not commercial approval. Philippe Wolgen has led the business since the 2005 restructure; that continuity helped bring the original drug to market and makes succession an important part of the next phase. The chair's succession review is evidence that the board considered the issue, not proof of successor readiness. Product concentration and customer concentration need to be assessed separately. Issuer history · Annual report.
| Case | EPS growth / exit P/E | Year-three sell value | AUD annual IRR | USD at -1% AUD/year |
|---|---|---|---|---|
| Bull: adoption expands and the pipeline earns credible optionality | +10% / 15× | A$13.38 | 21.28% | 20.06% |
| Base: EPP cash flows hold; no major new indication underwriting | +3% / 12× | A$8.79 | 5.52% | 4.46% |
| Nothing changes: flat earnings and multiple | 0% / 11.33× | A$7.59 | 0.56% | -0.45% |
| Bear: sales weaken as programme costs continue | -12% / 7× | A$3.20 | -24.56% | -25.31% |
Decision: reject under the stated framework. Clinuvel reports ten consecutive profitable years but only nine consecutive annual dividends, against the required twenty. That excludes it from this selection strategy. The model uses rounded A$0.67 starting EPS and a 5¢ annual distribution, with no drug-development windfall or cash addition. Even relaxing the dividend rule would not make A$7.59 attractive on my base: the 15% local ceiling is about A$5.87, or A$5.70 with AUD stress. FY26 annual results.
Valuation assumptions
Each company has bull, base, stalled and bear scenarios. The terminal price is year-three EPS times an exit P/E; dividends are separate cash flows. The annual return is a three-year cash-flow IRR with 0.15% stock costs at each end, before personal taxes. Base and flat dividends are held at the stated ordinary dividend. Bull/bear dividend growth assumptions, respectively, are +6%/-18% for Italtile, +5%/-10% for Gentex and +5%/-20% for Clinuvel. Non-US stress converts each year's proceeds to USD using the stated annual currency move. The 15% ceiling is the price whose after-cost cash flows just meet that hurdle. It is not a margin of safety if a small negative revision removes the return.
The range of outcomes
The bull case is only one outcome
Modelled annual IRR over three years, after transaction costs, in each company’s local currency.
What these businesses say about the investment opportunity
The common investment question is whether financial capacity will turn into profitable reinvestment. All five businesses can point to assets, cash or an established franchise. The investment question is what the next unit of capital will earn.
For Italtile, capacity is the constraint on returns. A larger manufacturing asset base does not help shareholders if factories cannot lift volumes or mix enough to cover their fixed costs. The opportunity is an improvement in returns on capacity already installed. I would rather see stronger factory utilisation and cash generation than assume another investment cycle will solve weak returns. Import protection may improve the competitive setting, but the base case still needs paying customers.
For Hudaco, the distinction is organic progress versus purchased growth. Buying specialist distributors can extend supplier relationships and customer reach. It can also make revenue and operating profit rise while returns on total invested capital disappoint. Engineering's 19.3% reported profit growth against 2.1% like-for-like growth makes the acquisition record central to valuation. The alternative-energy losses show why an apparently attractive new market must be judged by realised cash and returns, not its growth narrative.
For Gentex, the business must replenish its programme pipeline. The mirror franchise has engineering and qualification advantages, but a lost platform creates a specific revenue hole. Buybacks improve the denominator of earnings per share; replacement awards and higher content per vehicle must improve the numerator. I want both. Weak cash conversion alongside repurchases would make the financial optics less persuasive.
For Ingredion, the acquisition changes the capital-allocation problem. The combined formulation portfolio may deepen customer relationships and expand cross-selling. But the financing arrives before the full synergy run-rate. Argo also needs to recover economically while integration competes for management attention and cash. The right comparison is the per-share cash generation of the combined business after interest, reinvestment and integration costs, not the purchase price against the headline savings target.
For Clinuvel, retained cash has to become another commercial profit stream. A regulatory milestone reduces one development risk, but it does not establish market access, paid adoption or operating profitability. The cash balance gives the company time to develop these opportunities. It does not justify assigning a successful commercial outcome to every programme today.
The macroeconomic and political transmission
South African governance matters here through investment confidence, power and logistics reliability, building activity and industrial maintenance. A sustained improvement in procurement, infrastructure execution and the cost of doing business would strengthen the setting for Italtile and Hudaco. The figures in this review do not establish that improvement. I would look for it in orders, stock turnover, margins and cash conversion before increasing earnings assumptions. The currency stress shows the second channel: better local results still have to translate into an attractive US-dollar return.
Gentex is exposed to the vehicle-production cycle and trade policy through a different chain. Production volumes affect demand, programme wins determine its share, and tariffs can affect both cost and the timing of reimbursement. Stronger North American production cannot automatically offset a lost Asian programme. The earnings case therefore depends on the company-specific bridge between volume, content and recovered costs.
For Clinuvel, regulatory access and reimbursement matter more directly than a general Australian growth forecast. The value of an approval depends on the size and economics of the patient market it opens. Its international cash flows also make the Australian-dollar outcome different from the US-dollar result recorded by this portfolio.
These are the mechanisms through which macro and political developments change the valuation. They are not extra points awarded for a favourable national story.
What would change each decision
| Company | Current decision | Evidence that supports the bull case | Evidence that weakens the case |
|---|---|---|---|
| Hudaco | No selection at R184; base entry ceiling R167.66 locally or R153.82 under the USD stress | Acquired businesses deliver their expected sustainable profit; like-for-like growth strengthens; inventory and integration costs stop absorbing the improvement | Further impairments or wind-down costs; continued dependence on acquisitions for growth; disappointing cash returns on the acquired capital |
| Italtile | Research lead; the R7.90 price leaves almost no cushion against the R7.91 local base ceiling | Better manufacturing utilisation and mix; durable retail margins; completion of the Australian restructuring with reliable controls | Further production or inventory weakness; capex without improved returns; recurring control problems |
| Gentex | No selection at US$19.20; base entry ceiling US$16.40 | Replacement programme awards, higher content per vehicle and actual cash recovery of disputed costs; improved cash conversion | Further lost programmes, weaker overseas sales, continued working-capital absorption or disappointing VOXX returns |
| Ingredion | No selection at US$94.14; base entry ceiling US$85.97 before any additional deal-risk discount | Argo margins and cash conversion recover; financing and integration preserve cash and covenant headroom | Further plant-cost disappointment; higher financing costs; integration spending crowds out productive reinvestment |
| Clinuvel | Excluded by the twenty-year dividend rule; also below the return hurdle at A$7.59 | Paid adoption in new markets and commercial progress beyond the existing EPP franchise | Higher programme spending without additional commercial earnings or deterioration in the core product economics |
The stalled scenarios are important because shareholders can wait years for a catalyst. With no earnings growth or rerating, the estimated annual local returns are 5.98% for Hudaco, 5.59% for Italtile, 2.40% for Gentex, 3.4% for Ingredion and 0.56% for Clinuvel. Those outcomes are driven principally by dividends. None reaches the 15% hurdle merely by remaining a viable business.
The downside cases expose a different problem: apparent cheapness can disappear when earnings and the multiple fall together. The bull cases require operating delivery, not just a more generous valuation from the market. That is why I use the base case to set the entry ceiling and the company evidence to decide whether it is credible.
Portfolio and benchmark
No stock has been added and no paper capital has been contributed. Portfolio value and matched VOO value are both US$0. ROI and relative percentage performance have not started; an empty portfolio has not outperformed the S&P 500.
For each future selection, the tracker will record a US$10,000 contribution and a matched VOO investment using the same contribution date. The comparison will include dividends, transaction costs, currency movements and realised sale proceeds. Each position's thesis and sell price must be recorded before the purchase; hitting that recorded sell price during a later session triggers the paper sale. New selections are limited to three in total on Mondays, with no quota by market.
My conclusion
Hudaco is the clearest verified Graham candidate in the work completed so far, but R184 asks me to accept too little base-case return. Gentex has an established franchise and strong historical record, but its current price does not compensate for programme and cash-conversion risk. Italtile offers more apparent upside, accompanied by factory and control issues that need resolution. Ingredion offers a potentially stronger ingredient franchise, but the acquisition and Argo recovery have to earn the debt and cash they require. Clinuvel has substantial financial capacity, but it fails the dividend criterion and requires more successful development than the base valuation supports.
The useful result is a set of decisions tied to operating evidence and entry prices. I am not buying a low P/E, a cash balance or a promise of innovation on its own. I am looking for a business that can reinvest successfully, a management team that delivers and a purchase price that leaves enough of the resulting value for shareholders.
None of these five businesses earns a purchase at the prices and assumptions used here. Hudaco and Ingredion deserve particular attention as the acquisition evidence develops; Gentex becomes more interesting closer to US$16.40. The wider review remains open because the unresolved company records and financial-sector analysis are not yet complete.