State of the data
How much the open factor libraries disagree
Measured, not asserted. Every figure on this page is computed from the published comparison in the factor library, and it changes when the libraries do.
41
materials priced by two or more libraries
+45%
median gap between the highest and lowest figure, across the 34 genuine comparisons
7
apparent disagreements that are really different products under one name
The typical gap is large
Leaving aside the name collisions, the highest published figure for a material is a median 45% above the lowest, and half of all materials fall between 26% and 71%. Only 4 of 34 agree to within 10%. For 15, the highest figure is at least half as large again as the lowest. Which library an assessment is priced in is one of the largest choices in it, and it is usually made without being stated.
No comparison isolates the data from the method
In every one of the 34 genuine comparisons, at least two of the libraries characterise under different impact methods. So no gap on this page can be attributed to the underlying data alone: part of it is how each method weights the gases, and these libraries do not publish their figures under a common method that would let the two be separated. This is why an assessment should be priced in one library, with its method named, rather than assembled from whichever library has the lowest figure for each line.
A wide gap usually means two different products
Where the figures differ by a factor of three or more, the rows almost always describe different things under one name: a mix against the material itself, a raw input against a finished article. That applies to aggregates, asphalt, copper, paper, cement, leather and silicon. Reported as disagreements, they would blame careful sources for a naming problem — and it is one way a wrong number reaches a report: somebody picks the row by its name.
The widest genuine spreads
| Material | Figures, kg CO₂e/kg | Highest over lowest |
|---|---|---|
| Cardboard | 0.39 France (Base Carbone)0.56 Switzerland and Europe — BAFU (IPCC 2021)1.17 Europe — ELCD reference data (ReCiPe 2016) | +199% |
| Board | 0.439 Switzerland and Europe — BAFU (IPCC 2021)1.05 Europe — ELCD reference data (ReCiPe 2016)1.2 United Kingdom (DESNZ) | +173% |
| Plasterboard | 0.12 United Kingdom (DESNZ)0.283 Switzerland and Europe — BAFU (IPCC 2021) | +135% |
| Lead | 0.932 Switzerland and Europe — BAFU (IPCC 2021)1.81 Europe — ELCD reference data (ReCiPe 2016)2.09 France (Base Carbone) | +124% |
| Insulation mineral | 1.46 Switzerland and Europe — BAFU (IPCC 2021)1.51 Europe — construction and building products (EF 3.1)2.86 Europe — ELCD reference data (ReCiPe 2016) | +96% |
What to do about it
- Price one assessment in one library, and name it beside the result.
- Read the row label, not the material name. It says what the source measured.
- Never average across libraries. A mean of two boundaries is true of neither.
- When two suppliers quote figures from different libraries, compare the libraries before comparing the suppliers.
Every material behind these figures is on the factor library with its source rows and licence, and how PunarGen chooses between libraries is on the methodology page.
Attribution
- Contains public sector information licensed under the Open Government Licence v3.0.
- Database of the Swiss Federal Administration, BAFU:2026, Federal Office for the Environment, 2026
- Source: ADEME, Base Carbone®, Licence Ouverte / Open Licence 2.0.
Figures are reproduced for reference under the terms of their respective licences. They are not offered for resale.