The CBAM factor everyone inverts
In 2026 the CBAM factor is 0.975. A great many spreadsheets, briefing notes and consultancy decks use 0.025. It is the same number read the wrong way round, and it overstates the free allocation adjustment thirty-nine fold.
The part that surprises people: the more efficient your plant, the more the mistake costs you. Worked numbers below, with the instruments cited.
Where the confusion comes from
CBAM is usually explained like this: "free allocation is phased out at 2.5% in 2026, rising each year until 2034." That sentence is correct. The trouble starts when the 2.5% is carried into the arithmetic, because the regulation's "CBAM factor" is not the share being withdrawn — it is the share that remains.
Two things are being counted, and they are complements:
- 2.5% — the free allocation withdrawn in 2026. The headline number in most summaries.
- 97.5% — the free allocation still in place. This is the CBAM factor, and this is what goes in the formula.
The formula
An obligation is a subtraction. Embedded emissions, minus an adjustment representing the free allocation an EU producer would have received:
Guidance document 4, equations (1), (2) and (6)
SEFA(g,y) = CBAM(y) × CSCF(y) × BM(g)
FAA(g) = SEFA(g,y) × M(g)
obligation = embedded emissions − FAA(g)
where CBAM(y) is the share of free allocation remaining in year y, CSCF(y) is the EU ETS cross-sectoral correction factor, BM(g) is the applicable benchmark and M(g) is the mass of the goods.
Note what the formula does not say. The obligation is not embedded emissions multiplied by the CBAM factor. That is a second, separate error we have seen in the wild, and it fails in the opposite direction — understating a steel obligation by roughly elevenfold.
The two fixed inputs for 2026
What inverting it does
Take reinforcing bar, CN 7214 20 00, production route C, no precursors declared — so the whole-chain benchmark of 1.364 tCO₂e/t applies.
Correct: 0.975 × 1.0 × 1.364 = 1.3299 tCO₂e/t of adjustment.
Inverted: 0.025 × 1.0 × 1.364 = 0.0341 tCO₂e/t.
The adjustment term is wrong by 39×. But the adjustment is subtracted, so what it does to your obligation depends on how far your plant sits from the benchmark:
| Embedded | Correct | Inverted | Overstated |
|---|---|---|---|
| 1.40 t/t | 0.0701 t/t | 1.3659 t/t | 19.5× |
| 1.45 t/t | 0.1201 t/t | 1.4159 t/t | 11.8× |
| 1.60 t/t | 0.2701 t/t | 1.5659 t/t | 5.8× |
| 1.90 t/t | 0.5701 t/t | 1.8659 t/t | 3.3× |
| 2.30 t/t | 0.9701 t/t | 2.2659 t/t | 2.3× |
So "about forty times" — which is how this error usually gets described, including by us before we sat down and worked it out — is the wrong headline. 39× is the error in the adjustment term. The error in the number you would actually budget is between roughly 2× and 20×, and it is worst for the cleanest producers, whose real obligation is small enough that the inflated figure dwarfs it.
Which is the perverse part. Invest in an efficient plant, get close to the benchmark, and a careless spreadsheet will tell you that you owe twenty times what you do.
Why it matters in both directions
An overstated obligation is not a safe error. We have seen it lead to:
- Quoted prices padded with a cost that does not exist, losing tenders to competitors who did the arithmetic correctly.
- Budget provisions an order of magnitude too large, and the capital decisions that follow from them.
- Abandoning EU customers entirely, on the belief that the market has become uneconomic.
The understated version — multiplying rather than subtracting — is worse still, because nobody notices until the declaration is filed.
What the phase-out actually does to you
Same plant, same emissions, same 10,000 tonnes, embedded at 1.45 tCO₂e/t. Only the CBAM factor changes:
| Year | CBAM factor | Obligation | At €80/t |
|---|---|---|---|
| 2026 | 0.975 | 0.1201 t/t | €96,080 |
| 2027 | 0.95 | 0.1542 t/t | €123,360 |
| 2028 | 0.90 | 0.2224 t/t | €177,920 |
| 2029 | 0.775 | 0.3929 t/t | €314,320 |
| 2030 | 0.515 | 0.7475 t/t | €598,032 |
| 2032 | 0.265 | 1.0885 t/t | €870,832 |
| 2034 | 0 | 1.4500 t/t | €1,160,000 |
Twelve times the 2026 figure, for identical production. This is the number worth putting in front of a board, and it is the reason the obligation is a strategic variable rather than a compliance cost.
Check your own spreadsheet in two minutes
- Find where the CBAM factor or free allocation enters the calculation.
- If the cell holds 0.025, or 2.5%, or anything that rises over time toward 1 — it is inverted. The factor falls toward zero.
- Check it is multiplied by a benchmark and then subtracted, not multiplied against embedded emissions.
- Check the benchmark matches your production route. A single steel CN code can carry benchmark candidates that span fourfold, and picking the wrong one swamps everything above.
- Check your CSCF is 1.0 for 2026–2030, and flagged as undetermined after that.
Sources
Every figure here is traceable. Where the Commission has determined a value we use it; where it has not, we say so rather than carrying a preliminary value as though it were settled.
- CBAM factor schedule — Guidance document 4 on CBAM implementation, Table 2-1 (free allocation adjustment); Article 10a(1a) of the EU ETS Directive; Commission Implementing Regulation (EU) 2025/2620.
- CSCF — Commission Implementing Decision (EU) 2026/1862 of 23 July 2026, Article 1. Determines 100% for the 2026–2030 allocation period. Values from 2031 follow the next period's notification.
- Benchmarks — Annex to the Free Allocation Adjustment Act, point 5. The Commission's published workbook is explicitly informational; the binding values are in Implementing Regulation (EU) 2025/2620.
- Formula — Guidance document 4, section 2.2.1.1 and equations (1), (2), (6).
Want your own number?
Send your CN codes, tonnage and production route and we will return a one-page calculation with the method written out and every assumption labelled. No charge, and you do not need our software to use the result.
This guide is an explanation of a published method, not legal or assurance advice. It describes the regulation as it stood on 10 October 2026; CBAM has been amended repeatedly, so check the instruments above before relying on a figure. If you think something here is wrong, write to hello@ecometricsbih.com — we would rather correct it than defend it.