Commercial Metals (CMC) makes the steel rebar inside concrete, recycles the scrap it is made from, and since December 2025 has bought two precast-concrete businesses. At the Oct 9 close it trades at 12.0 times trailing earnings (13.6 times once a one-time tax credit is taken out) and 1.56 times book value, about 25% below its 52-week high — against 19.8 and 21.4 times earnings for Nucor and Steel Dynamics. It has paid a dividend for 247 consecutive quarters and been profitable in all 11 fiscal years from 2015 to 2025. It also borrowed $2.0 billion to fund the change, carries a roughly $370 million antitrust judgment under appeal that is not counted in its net debt, and reports its fiscal year on October 15.
Cheap against its peers, fair against Graham, and honest about the debt: all three are true at once.A cyclical steel company with an unusually long dividend record, priced below its peers and 3% below its Graham Number once its tax credit is taken out — in the middle of a debt-funded move into higher-margin precast concrete. Not a bargain and not a trap: a fair price for a business whose next few quarters will say a lot about the move, with the first test four days after this file.
Commercial Metals Company, founded in 1915 and based in Irving, Texas, employs about 12,700 people and reports in three segments. North America Steel recycles scrap into steel in electric-arc mini-mills and sells rebar, merchant bar and fabricated downstream products; in the third quarter of fiscal 2026 (ended May 31) it shipped 750,000 tons of steel products at an average price of $989 a ton and earned $253.5 million of adjusted EBITDA. Europe Steel is its Central European operation, including Poland. Construction Solutions, the newest, makes precast concrete, concrete pipe and soil-stabilization products (Tensar).
The change in character is recent. CP&P (Concrete Pipe and Precast, Mid-Atlantic and South Atlantic) was bought on December 1, 2025 for about $675 million in cash from the balance sheet. Foley Products, a precast and pipe supplier in the Southeast, followed on December 15, 2025 for $1.84 billion in cash, funded with $2.0 billion of new senior notes: $1.0 billion at 5.750% due November 2033 and $1.0 billion at 6.000% due December 2035. In the third quarter the two together added $175.7 million of sales and $52.9 million of EBITDA. Construction Solutions as a whole sold $394.6 million, double a year earlier, at a 24.7% adjusted EBITDA margin — against 14.2% for the company overall.
The stock is 25% below its 52-week high of $84.87 and 20% above its low of $53.08. The path between them, using weekly closes plus the days numbered below:
The sharper contrast is with the steelmakers it is usually compared to. Year to date is price only, since December 31, 2025; the tax rate is what each paid over the last twelve months (22% in the restated row is the assumption):
| Company | Year to date | P/E | Debt / equity | Tax rate |
|---|---|---|---|---|
| CMC, as reported | −7.7% | 12.0× | 0.75 | 11.8% |
| CMC, at 22% tax | — | 13.6× | 0.75 | 22% |
| Nucor | +53.5% | 19.8× | 0.32 | 20.7% |
| Steel Dynamics | +40.7% | 21.4× | 0.44 | 26.3% |
The discount has reasons: Commercial Metals carries about twice the debt relative to equity, earns less on its capital, and is the one of the three that has just spent a quarter of its market value on acquisitions. Nucor and Steel Dynamics pay a normal tax rate, so the fair comparison for CMC's multiple is the normalized one — about 13.6 times against 19.8 and 21.4.
Reported earnings per share ran from $0.67 in fiscal 2015 to a peak of $9.95 in fiscal 2022 and back to $0.74 in fiscal 2025. That is what a steel company looks like across a cycle — and it is why Graham's rule of averaging earnings over several years exists.
The legal charge. In November 2024 a jury found for a competitor, Pacific Steel Group, in an antitrust case and awarded $110 million, which the court trebled by law. CMC recorded a $350.0 million provision in the first quarter of fiscal 2025 (about $265 million after tax), turning that quarter into a $175.7 million loss and holding the year to $84.7 million of net income. CMC lost its post-trial motion on September 29, 2025 and filed its notice of appeal on October 24, 2025.
The tax credit. The last twelve months look far better: net income of $595.1 million and diluted EPS of $5.31. But the effective tax rate was 7.9% for the nine months to May 31, 2026, against 21.7% a year earlier, “primarily due to the recognition of a federal investment tax credit related to the ongoing construction of the West Virginia micro mill, with commissioning currently expected during 2026.” Over the four quarters CMC paid 11.8% on $674.7 million of pre-tax income. At 22% instead, net income would be $526.3 million and EPS $4.70 — about 61 cents a share lower. The filing does not say how long the credit continues. (The third quarter also included a $20.4 million CO₂ credit in Europe and a $20 million mill-outage cost in North America, which roughly offset each other.)
About 61 cents a share of the last year's earnings is a tax credit on a mill that has not opened yet.Net debt was $0.3 billion at August 31, 2025 and $2.8 billion at May 31, 2026, after $2.5 billion of acquisitions and $2.0 billion of new notes.
| Measure | Reading | Note |
|---|---|---|
| Net debt / EBITDA, trailing | 2.5× | On trailing reported EBITDA of $1.15 billion |
| The company's own “net leverage adjusted for acquisitions” | 2.1× | Replaces the precast businesses' actual results with $245 million of expected annualized EBITDA and removes about $36.5 million of one-time costs. Target: under 2× “well ahead of” its mid-2027 goal |
| Net debt plus the recorded lawsuit amount | 2.8× | My arithmetic: ($2.84 billion + $0.37 billion) / $1.15 billion |
| Interest expense | $40.2M a quarter | $10.9 million a year earlier. The new notes alone cost $115 million a year |
| Interest coverage | 6.9× | Trailing |
| Goodwill and intangibles | $2.60 billion | 27% of total assets; tangible book value is $17.41 a share, so the price is about 3.8× tangible book |
| Capital spending, last 12 months | $513M | 56% of operating cash flow; trailing free cash flow $405 million, about 5.6% of the market value |
| Share buybacks | $18.9M in Q3 | $600 million added on August 5; about $717 million now authorized, roughly 10% of the market value |
The lawsuit deserves its own sentence. If the appeal fails, CMC pays about the $370 million recorded so far, plus interest that keeps accruing; its own filing says that payment “could have a significant impact on our liquidity.” If the verdict is overturned, the expense and liability reverse. A second suit by the same plaintiff, in California state law, seeks about $29 million of lost profits (part subject to trebling); discovery is substantially complete and no trial is scheduled.
Figures are from company statements read for this file, plus arithmetic on them — not figures travisvaluation.ca carries; the app runs on SEC filings and a typed price. Price ratios are at the Oct 9 close; earnings ratios are trailing twelve months through the quarter ended May 31, 2026 (diluted); balance-sheet items are from May 31, 2026.
| Method | Reading | What it says |
|---|---|---|
| Price to earnings (trailing, diluted) | 12.0× reported 13.6× at 22% tax | Under Graham's ceiling of 15 either way |
| Price to book | 1.56× | Narrowly over Graham's 1.5 |
| P/E × P/B | 18.8 reported 21.3 at 22% tax | Inside Graham's alternative limit of 22.5 on both; narrowly on the second |
| Graham Number | $69.86 reported $65.72 at 22% tax | Price is 9% below the first and 3% below the second — a thin margin of safety, not a wide one |
| Earnings power value (7-year average EPS $4.21 ÷ AAA yield 6.03%) | $69.75 | 8% above the price; averages the peak year and the legal-charge year together |
| EV / EBITDA (trailing) | 8.6× | Nucor 10.9×, Steel Dynamics 12.7× (market data, October 6 prices) |
| Return on equity / on invested capital | 13.6% / 8.5% | Company's FY2029 mid-cycle target for ROIC: 13.0% to 14.5% |
| Piotroski F-Score | 6 / 9 | Decent |
| Altman Z-Score | 2.91 | Grey zone (1.81–2.99) — not distress, not pristine |
| Dividend | $0.80 a year 1.3% yield | $0.20 a quarter; payout about 14% of earnings. $0.12 until 2021, then $0.14 (Oct 2021), $0.16 (Oct 2022), $0.18 (Mar 2024), $0.20 (Mar 2026) |
| Analyst consensus | 12 buy / 13 hold / 2 sell | Target $82 (range $75 to $89) |
The app's own research page for CMC currently reads its latest filed annual report — fiscal 2025, ended August 31, 2025, filed October 16, 2025 — and says so on the page. On that basis it shows a price-to-earnings of 86.3, a Graham Number of $25.12 and Graham 1962 and 1974 formula values of $19.96 and $14.57, because the earnings behind them are the $0.74 of the legal-charge year; its current ratio (2.78) and its long-term-debt test (debt $1.31 billion against net current assets $2.24 billion, a pass) describe the company before the December acquisitions. Those figures are computed correctly from that filing and are not repeated as valuation here. They will move to fiscal 2026 once CMC files its annual report (last year it filed the day it reported results). The page's seven-year-average earnings power value, $69.75, is the one measure built to smooth a year like that, and it agrees with the figure above.
| Test | Result | Detail |
|---|---|---|
| Adequate size | Pass | About $7.1 billion market capitalization |
| Current ratio ≥ 2 | Pass | 2.33 (May 31, 2026) |
| Long-term debt < net current assets | Fail | Long-term debt $3.31 billion against net current assets of $1.97 billion. It passed before the acquisitions ($1.31 billion against $2.24 billion) |
| Positive earnings, last 10 years | Pass | All 11 fiscal years, 2015–2025; the smallest profit was $46.3 million in 2017 |
| Twenty years of dividends | Pass | 247 consecutive quarterly payments, per the company's June 25 release (the app can only confirm 17 years from structured filings) |
| Earnings growth over the window | Pass | Graham's three-year averages: $0.51 (fiscal 2015–17) to $4.04 (fiscal 2023–25) — though the path is erratic |
| P/E ≤ 15 | Pass | 12.0 reported; 13.6 at 22% tax |
| P/B ≤ 1.5 | Fail | 1.56 — but the alternative P/E × P/B ≤ 22.5 passes at 18.8 (21.3 at 22% tax) |
Six of eight — better than EQT's three, and the two fails are real. The debt test fails because the company spent $2.5 billion it had not spent a year ago; the book test fails by a hair and is rescued only by Graham's own alternative. Neither is a quirk of how the test reads this kind of business.
Over the twelve months to the date of value, insiders bought about $854,000 of stock on the open market and sold about $2.38 million. Every trade below was read directly from its SEC Form 4.
| Date | Insider | Trade | Value |
|---|---|---|---|
| Aug 13, 2026 | John McPherson, director | Bought 1,390 at $71.92 | $99,969 |
| Jul 10, 2026 | Peter Matt, President and CEO | Bought 8,230 at $61.30 | $504,499 |
| Jan 20, 2026 | Dennis Arriola, director | Bought 2,000 at $74.69 | $149,380 |
| Oct 20, 2025 | John McPherson, director | Bought 1,722 at about $58.09 | $100,037 |
| Feb 3, 2026 | Jennifer Durbin, former Chief HR and Communications Officer | Sold 25,050 at $79.97 | $2,003,249 |
| Oct 31, 2025 | Brian Halloran, SVP, North America Steel | Sold 6,232 at $59.87 | $373,110 |
Two things in the history change how to read the buying. These are habitual buyers. Peter Matt has now bought three times in three years — 6,200 shares at $40.30 (October 2023), 6,100 at $48.30 (March 2025) and 8,230 at $61.30 (July 2026) — and John McPherson has bought in April 2022, April 2025, October 2025 and August 2026 (the purchases before this twelve-month window are from an insider-trading data feed). A cluster of strangers buying is a stronger signal than directors who buy every time the stock dips. The sales were discretionary. The Rule 10b5-1 box on both sale filings is unchecked and neither carries a footnote, so neither was made under a pre-set trading plan — the opposite of the EQT finding in Workfile No. 8. Durbin sold about a third of her holding at $79.97 near the February high; Halloran about a tenth, at $59.87.
At its August 5, 2026 Investor Day, CMC set fiscal 2029 “mid-cycle” targets: core EBITDA of $1,650 million to $1,800 million, a core EBITDA margin of 15.0% to 16.0%, return on invested capital of 13.0% to 14.5%, and “free cash flow” of $1,375 million to $1,525 million. For scale, the third quarter's core EBITDA of $353.6 million is a pace of about $1.4 billion a year and its trailing ROIC is 8.5%.
CMC defines that free cash flow as core EBITDA less capital expenditures — before interest, taxes and working capital. It is not comparable to free cash flow on the cash-flow statement, which was $405 million over the last twelve months. The targets are management's, for a mid-cycle year, and not a forecast: the same company reported $0.39 of earnings per share in fiscal 2017 and $9.95 in fiscal 2022.
Set October 6, 2026; reviewed at the October 9, 2026 date of value. The first test arrives October 15.
1. Fiscal fourth-quarter core EBITDA of about $390 million or more. The third quarter was $353.6 million and CMC said the fourth would be higher: North America Steel adds about $20 million from the absence of the outage and a similar amount from volume and margin, Construction Solutions grows at a mid-teens rate (about $15 million on $97.4 million), and Europe is “modestly higher” excluding CO₂ credits (the $20.4 million credit does not repeat). That sums to roughly $388 million, my arithmetic. A figure below $354 million, a decline from the third quarter, is a clear fail.
2. Steel margins hold. North America Steel's metal margin was $610 a ton in the third quarter, down $13 from the second. Below $610 means scrap is still outrunning prices, the squeeze behind June's fall.
3. Debt starts coming down. Net debt below the May 31 level of $2.84 billion, and the company repeating its goal of under 2× (its own measure) by mid-2027.
4. The earnings are not just a tax credit. An effective tax rate near 8% again in the fourth quarter means the credit is still flowing; a rate nearer 22% means the 22%-tax figures in section 3 are the better guide. Watch whether the West Virginia mill commissions in 2026, as the filing says it expects.
5. The lawsuit. Any ruling in the Pacific Steel Group appeal. An adverse result puts about $370 million of cash out the door, against $0.56 billion of cash on hand and about $1.8 billion of total liquidity.
6. Construction demand. The 10-year Treasury yield is above 5.3%, near a 25-year high. CMC calls its backlogs “robust”; softer language on bookings or backlog on the October 15 call would be the early sign that rates are reaching its customers.
Re-checked on October 6 against market data, with price ratios at the October 5 close. The Radian and Textron insider purchases are from an insider-trading data feed, not the Form 4s themselves.
| Considered | Rejected because |
|---|---|
| Cognizant (CTSH) | An IT-services name down 33% from its high amid AI-disruption fears, at 12.5× earnings. Set aside this week on Curtis's call: on reported numbers it narrowly fails Graham's combined test (about 23), and whether AI shrinks its work is unresolved. It stays on the watch list. |
| Nucor (NUE) and Steel Dynamics (STLD) | Less debt per dollar of equity, but 19.8× and 21.4× earnings and 2.5× and 3.7× book — both fail Graham's price tests. |
| Gentex (GNTX) | About 11.5× earnings, 1.8× book, essentially no debt, a current ratio near 3. But earnings per share have been flat to lower for three years ($1.84, $1.76, $1.74 for 2023–2025) and no insider has bought since October 2022. |
| Radian (RDN) | 8.5× earnings and 0.94× book, with the incoming CEO buying about $5.8 million in June — but a housing-credit bet immediately after Lennar. |
| Linamar (LNR.TO) | 8.4× earnings and 0.88× book (C$97.46), but already a worked example in the book. |
| Textron (TXT) | A director bought about $1 million on May 1 and the stock is within 1% of its 52-week low, but the dividend is $0.08 a year and the return on equity is about 11%. |
At the date of value the author owns no shares of Commercial Metals and will not buy any until after the companion letter has gone out.
Cheap against its peers, fair against Graham, and honest about the debt: all three are true at once.