Consolidating emissions across entities without double counting

A single-entity emissions figure is an arithmetic problem. A group figure is a policy problem wearing an arithmetic costume, and the four things that break it are all decisions rather than calculations.

The four things that break consolidation

Before the specifics, the shape of the problem. Financial consolidation has decades of machinery behind it: an agreed reporting entity, intercompany elimination, a group chart of accounts, a close calendar. Emissions consolidation typically has a spreadsheet with one tab per subsidiary and a SUM at the bottom.

The four failure modes are intercompany double counting, factor differences between jurisdictions, non-aligned reporting calendars, and entities that joined or left part way through the period. Each is manageable alone. Together, in a group of any size, they produce a number nobody can reconstruct.

Intercompany double counting

The clearest case: one group entity generates electricity or heat and supplies it to another. The supplying entity records it in Scope 1. The receiving entity records purchased energy in Scope 2. Both are correct at entity level. Added together at group level, the same emissions appear twice.

Shared sites do the same thing more quietly. Two subsidiaries occupy floors of the same building, both receive an apportionment from the same landlord, and neither knows the other did.

Intercompany purchases are the third: the buying entity records the purchase in Scope 3 Category 1, while the selling entity has already reported the underlying emissions in its own Scope 1 or 2. At group level, you have counted the same tonne as a direct emission and as a value chain emission.

And joint arrangements are the fourth, where two parents may both have a claim on the same asset depending on the measurement approach each applies.

The fix is procedural rather than clever: eliminate at the same point where finance eliminates intercompany transactions, using the same intercompany mapping, and keep the eliminations visible as adjustments rather than netting them off silently. An assurance provider will want to see the gross entity figures, the elimination, and the group total — not just the last one.

Different emission factors by jurisdiction

Here the instinct to standardise is actively wrong. A UK entity and a German entity buying identical quantities of electricity should not use the same factor, because they are drawing from different grids with different generation mixes. Forcing one factor across the group produces a tidier process and a less accurate number.

What consistency actually requires is that the rule for choosing a factor is the same everywhere: for example, use the national publisher's factor for the reporting year where one exists, and a stated fallback where it does not. Then record, against each figure, which source and which publication year was applied.

That last part is the one that gets skipped, and it is the one that breaks reproducibility, because publishers revise factors and a factor table updated in place quietly changes last year's answers. The mechanism is covered in could you reproduce this figure in eighteen months.

Publication lag is worth planning for. Some national factors for a reporting year are not published until well into the following one, so your fallback rule needs to say what you do in the meantime and whether you restate when the final factor lands.

Different reporting calendars

Subsidiaries with year ends that do not match the parent are common in acquisitive groups, and there is no single correct treatment.

You can align the subsidiary to the group reporting period, which is more work and more comparable. Or you can use the subsidiary's own period and disclose that you did, which is less work and requires the disclosure to be explicit. Both are defensible positions.

What is not defensible is the third option, which is what most groups actually have: a mix, decided per subsidiary by whoever was collecting the data, undocumented, and impossible to explain a year later. Pick one, write it down, and note the exceptions with reasons.

Entities acquired or disposed mid-period

An acquisition in August raises a question that sounds trivial and is not: does the group figure include that entity's emissions from August, or for the whole year?

Both approaches are used. What matters is that you choose one, apply it to disposals as well as acquisitions, and record the effective date against the figure. The organisational boundary you selected under the GHG Protocol Corporate Standard — equity share or control — drives how the entity enters the group measure in the first place; the ISSB's educational material on greenhouse gas disclosure sets out how that selection determines the organisational boundary.

The reason the date has to be recorded, rather than just remembered, is the following year. Your comparative has to be built on the same basis, and by then the person who ran the acquisition integration will be doing something else.

What consistency actually requires

Pulling those together, consistency in a group context means four things, none of which is uniformity.

The same rule for selecting factors, producing different factors in different places. The same boundary approach across entities, applied to whatever each entity turns out to be. The same period policy, with documented exceptions. And the same elimination logic as the financial consolidation, so that two people asking where the intercompany adjustment went get the same answer.

Uniformity of inputs would be wrong. Uniformity of method is the target.

Documenting a methodology decision

Every one of the choices above is a judgement, and judgements are what assurance providers test when the arithmetic is straightforward.

A methodology decision record is short. What you decided. What else you considered and why you rejected it. What you relied on. Which entities, periods and figures it governs. Who decided, and with what authority. When, and from which period it applies.

Then the field almost everyone omits: what would cause you to revisit it. A new factor release, an acquisition in a new jurisdiction, a change in the organisational boundary. Without a revisit trigger, methodologies do not get reviewed — they drift, and the drift is discovered by someone else.

Write these when you make the decision. A decision reconstructed in March for something agreed the previous June is a reconstruction, and it reads like one.

The position: a group emissions figure is only as good as the four policies underneath it, and those policies are usually undocumented rather than wrong. Most consolidation problems I have seen were not calculation errors. They were reasonable decisions, made once, by someone who has since moved on, that nobody wrote down — which is a records problem, not an emissions problem, and it is the one a spreadsheet is worst at.

Common questions

How do I avoid double counting emissions across group entities?

Eliminate at the same point in the process where you eliminate intercompany transactions in the financial consolidation, and for the same reason. The four common points are intercompany energy supply, shared sites where two entities allocate from one landlord apportionment, intercompany purchases recorded by the buyer in Scope 3 Category 1, and joint arrangements where the boundary overlaps.

Should every entity in a group use the same emission factor?

Usually not. A UK entity and a German entity buying electricity draw on different grids, and using one factor for both would misstate them. Consistency means applying the same rule for choosing a factor everywhere, and recording which factor and publication year was applied to each figure, rather than forcing one number across jurisdictions.

What if a subsidiary has a different financial year end?

You need a stated policy and you need to apply it consistently: either align the subsidiary to the group reporting period, or use its own period and disclose that you did. Both are defensible. What is not defensible is a group figure where nobody can say which period each entity contributed.

How do we handle an entity acquired part way through the year?

Decide whether you are including emissions from the acquisition date or for the full period, apply that consistently across acquisitions and disposals, and record the decision. The organisational boundary you selected under the GHG Protocol drives the answer, and the date needs to be recorded against the figure so the following year’s comparative can be built.