UK Spend-Based GHG Intensity by SIC Section — ONS Industry Factors
When you know what a supplier charged you but not what they emitted, this is the fallback. Spend-based estimation multiplies an economic figure by an industry-average intensity. It is the least precise method in the hierarchy and every framework says so — but it is the one that lets you put a number against a whole ledger rather than leaving it blank.
This page publishes the UK industry intensities held in MasterBrain: 18 SIC 2007 sections from the ONS Atmospheric Emissions framework, in kg CO₂e per pound of gross value added.
These figures are per pound of gross value added — an industry’s output minus what it bought in to produce that output. They are not per pound of invoice. GVA is typically a minority of turnover, so applying a GVA-basis intensity directly to a supplier’s invoice understates the result, and by a different amount in every sector. This matters especially when combining sources: the US EPA rows in the same dataset family are published per dollar of purchaser price, which is the invoice basis. The two are not interchangeable, and a portfolio that mixes them without converting is comparing different quantities.
What spend-based intensity is
An environmentally-extended input-output figure: total greenhouse gas emissions attributed to an industry, divided by that industry’s economic output. ONS produces it by pairing the UK Environmental Accounts atmospheric-emissions series with national-accounts gross value added, at SIC 2007 section grain, on a UK residence basis.
It answers a question no supplier-specific factor can: what is a reasonable emissions figure for a purchase where I know only the sector and the amount? That makes it the last resort in every data hierarchy — the GHG Protocol Scope 3 spend-based method, and PCAF data quality score 4 — and it should be replaced with primary data wherever primary data exists.
The spread across sectors is wide: agriculture at 2.71 kg CO₂e per pound of GVA is 271 times information and communication at 0.01. Sector selection therefore dominates the answer far more than the precision of the spend figure does. Getting the sector right matters more than getting the pounds exactly right.
The 18 industry sections
All 21 SIC 2007 sections are listed so the gaps are visible. Eighteen carry a published number.
| SIC | Industry section | kg CO₂e / £ GVA | Year |
|---|---|---|---|
| A | Agriculture, forestry and fishing | 2.71 | 2024 |
| B | Mining and quarrying | 0.54 | 2024 |
| C | Manufacturing | 0.28 | 2024 |
| D | Electricity, gas, steam and air conditioning supply | 1.49 | 2024 |
| E | Water supply; sewerage, waste management and remediation | 0.91 | 2024 |
| F | Construction | 0.08 | 2024 |
| G | Wholesale and retail trade; repair of motor vehicles | 0.05 | 2024 |
| H | Transport and storage | 0.88 | 2024 |
| I | Accommodation and food services | 0.06 | 2024 |
| J | Information and communication | 0.01 | 2023 |
| K | Financial and insurance activities | [low] | — |
| L | Real estate activities | [low] | — |
| M | Professional, scientific and technical activities | 0.01 | 2024 |
| N | Administrative and support service activities | 0.02 | 2024 |
| O | Public administration and defence | 0.03 | 2024 |
| P | Education | 0.02 | 2024 |
| Q | Human health and social work activities | 0.03 | 2024 |
| R | Arts, entertainment and recreation | 0.05 | 2024 |
| S | Other service activities | 0.03 | 2024 |
| T | Activities of households as employers | 0.03 | 2024 |
| U | Activities of extraterritorial organisations and bodies | — | negligible |
[low] is the ONS marker, defined in the source as “a low figure but not a real zero” — these sections are published without a number, not as zero. Section U is noted by ONS as “assumed to be negligible”. Section J is carried at its 2023 value because ONS marks it [low] for 2024.
The sections with no number
Three sections are absent from the dataset, and that absence is the source’s, not an omission here.
- K — Financial and insurance activities and L — Real estate activities are marked
[low]by ONS in every published year, including 2024. The source defines[low]as “a low figure but not a real zero”, so there is no value to carry and substituting zero would be wrong. - U — Extraterritorial organisations is described by ONS as “assumed to be negligible”.
- J — Information and communication became
[low]in 2024 but has a published 2023 figure, so it is carried at 2023 and labelled with its own year rather than dropped.
If you need a figure for finance or real estate, this dataset cannot give you one. Use a supplier-specific footprint, or an input-output model that publishes those sectors, and record which you used.
Application — worked example
The arithmetic is one multiplication. The care is all in the denominator.
emissions (kg CO2e) = value added attributable to the purchase (GBP) x sector intensityA manufacturer’s activity generating £2,000,000 of gross value added, SIC section C:
2,000,000 x 0.28 = 560,000 kg CO2e = 560 t CO2eMove the same £2m of value added into agriculture (section A, 2.71) and it becomes 5,420 t — nearly ten times more. That ratio, not the pound figure, is what a spend-based estimate is really telling you.
If what you hold is an invoice rather than a value-added figure, do not apply these intensities to it directly. Either convert to a value-added basis using the sector’s ratio of GVA to output, or use a factor set published on a purchaser-price basis instead — and say which in your audit trail.
Calculators that use these factors
The worked examples above are automated in these GreenCalculus tools, which read this dataset live from the MasterBrain so results always reflect the current vintage:
- Scope 3 Cat 1 spend-based calculator — purchased goods and services from procurement spend.
- Scope 3 Cat 2 capital goods calculator — capital expenditure on a spend basis.
- Small business carbon footprint calculator — a quick spend-based estimate for SMEs.
Common reporting errors
- Applying a GVA intensity to invoice totals. The most consequential error on this dataset. Gross value added excludes bought-in inputs; an invoice does not. Applying these figures to spend understates the result by a sector-specific margin.
- Mixing GVA-basis and purchaser-price-basis factors in one portfolio. The UK and EU rows in this dataset family are value-added basis; the US EPA rows are purchaser price. Combining them without conversion compares different quantities.
- Reading a missing sector as zero. Finance and real estate are
[low], not nil. An estimate that silently assigns them zero will understate a services-heavy portfolio. - Treating a section average as a company figure. These are whole-industry averages. Within any section the spread between individual firms is far wider than the gap between neighbouring sections.
- Leaving spend-based estimates in place once primary data arrives. Every framework treats this as the weakest tier. It is a placeholder, and the data-quality score should improve over time.
- Assuming AR6. These are AR5 GWP-100 per the DEFRA and Ricardo convention. If the rest of your inventory is AR6, note the mismatch rather than silently blending bases.
Methodology, boundaries & uncertainty
Unit. ONS publishes this table in thousand tonnes CO₂e per £ million of GVA. That is exactly kilograms per pound — the millions cancel — so the published number is carried through unchanged. This is worth stating plainly because misreading it as tonnes per £ million is an easy and consequential mistake.
Geography and basis. United Kingdom, residence basis — emissions attributed to UK-resident economic units wherever they occur, which is the national-accounts convention and differs from a territorial inventory.
GWP. AR5 GWP-100, per the DEFRA and Ricardo convention. ONS migration to AR6 is anticipated but has not shipped.
Grain. SIC 2007 section — the single-letter level, the coarsest published grain. ONS publishes a finer industry breakdown elsewhere, but the intensity series is section-level only. A section covers wildly heterogeneous activity, and that heterogeneity is the dominant uncertainty in any spend-based estimate.
Vintage. Reference year 2024 for 17 sections, 2023 for section J. ONS publish annually in June, with interim corrections; values move between editions, so apply the edition contemporaneous with your reporting year.
Implementation & provenance chain
Every row traces to the published ONS table. Nothing is hand-transcribed: the table is generated from the same MasterBrain rows the REST endpoint serves, and the generator asserts that all 18 source keys render exactly once before the page is built.
| Stage | What it is |
|---|---|
| Primary source | ONS, Atmospheric emissions: greenhouse gas emissions intensity by industry, 5 June 2026 release, prepared by Ricardo Energy and Environment |
| Cell reference | Each row carries its sheet, table and year column (e.g. Table 1a, year column 2024, row C) in the source_ref field |
| Data layer | GreenCalculus MasterBrain — canonical keyspace spend_based.uk.sic.<letter>.<slug>, version-stamped |
| Filter | The feed is pinned to the ONS source id. The wider spend_based section holds 2,226 rows across five publishers covering three continents — none of the others appear here |
| Licence | Open Government Licence v3.0 — free to reuse with attribution |
Data access — REST API & CSV
The full 18-row dataset is available as a machine-readable REST endpoint and as a flat CSV download, versioned by MasterBrain edition and citable.
Cache-Control: max-age=3600; X-GC-Version signals updates./wp-json/greencalculus/v1/ons-sic-factors
Click to generate ↓
Citation guidance
If you use this dataset in a published tool, report or academic work, cite the ONS primary source.
Office for National Statistics (2026). UK Environmental Accounts — Atmospheric emissions: greenhouse gas emissions intensity by industry. Newport: ONS, prepared by Ricardo Energy and Environment. Contains public sector information licensed under the Open Government Licence v3.0.
ONS primary citation
Cite this dataset (GreenCalculus compilation). A versioned, machine-readable snapshot of these factors is archived on Zenodo with a citable DOI:
Say, Jeremiah (2026). UK spend-based GHG intensity by SIC 2007 industry section (ONS) (machine-readable) (v2026.107). GreenCalculus. Zenodo. https://doi.org/10.5281/zenodo.21831227
GreenCalculus dataset DOI
Frequently asked questions
Not directly. These intensities are per pound of gross value added, which is an industry’s output minus the inputs it bought in to produce that output. An invoice is not value added — it includes everything the supplier bought too. Applying a value-added intensity to invoice totals understates the answer, by a different margin in every sector. Either convert your spend to a value-added basis using the sector’s own ratio, or use a factor set published on a purchaser-price basis, and record which you did.
ONS marks both as [low] in every published year, and defines that marker in the source as “a low figure but not a real zero”. There is genuinely no number to carry, so those sections are absent rather than being filled with a zero that would be wrong. If you need an intensity for a financial or property counterparty, take it from a supplier-specific footprint or from an input-output model that publishes those sectors, and say which you used.
Because ONS moved that section to [low] for 2024, so there is no 2024 number for it. Rather than dropping the sector entirely, the row keeps its published 2023 figure and carries its own reference year, which every consumer of this dataset reads and can disclose. The dataset is therefore deliberately mixed-vintage — seventeen sections at 2024, one at 2023.
Weak, and deliberately so. These are whole-industry averages at the coarsest published grain, so two companies in the same SIC section can differ by an order of magnitude while sharing a factor. Every framework places spend-based estimation at the bottom of its data hierarchy — it is PCAF data quality score 4 — because it tells you about a sector, not about a counterparty. Use it to get complete coverage, then replace it wherever primary data becomes available.
Not without care. The EPA factors are published per dollar of purchaser price — the invoice basis — while these are per pound of gross value added. They also use different industry classifications, different GWP sets and different currencies. A portfolio spanning both needs each side computed on its own basis and the difference disclosed, not a single blended multiplier applied across the lot.
AR5 GWP-100, following the DEFRA and Ricardo convention that ONS applies to the underlying emissions series. Migration to AR6 is anticipated but has not shipped. If the rest of your inventory is on AR6, that is a real basis mismatch: note it rather than blending the two silently, because the difference is material for methane-heavy sectors such as agriculture and waste.
Emissions are attributed to UK-resident economic units wherever in the world they occur — a British airline’s flights abroad count, a foreign firm’s UK operations do not. That is the national-accounts convention, and it is what makes these figures comparable with GVA, which is measured the same way. It differs from a territorial inventory, which counts emissions inside UK borders regardless of who caused them, so the two are not interchangeable.
Two reasons compound. Agriculture emits large volumes of methane and nitrous oxide from livestock and soils, which carry high warming potentials, and it generates relatively little value added per unit of physical output. A high numerator over a small denominator gives 2.71 kg CO₂e per pound — 271 times the figure for information and communication. It is a reminder that these intensities measure emissions per unit of economic value, not per unit of anything physical.
Version history
| Version | Date | MasterBrain | Summary |
|---|---|---|---|
| 1.0 | 2026-08-06 | v2026.107 | Initial publication. Complete ONS SIC-section intensity set (18 sections, 5 June 2026 release): value-added denominator warning, suppressed-section explanation, worked example, 8-item FAQ, REST + CSV access. |