Scope 3 Double Counting: When Correct Calculations Produce the Wrong Inventory

Most checks on a Scope 3 figure are checks on a calculation. Was the right factor chosen, were the units converted properly, does the multiplication hold? Those questions matter, and an inventory can pass every one of them and still be wrong.

Take three calculations any reviewer would accept on sight: expenditure multiplied by a spend-based factor, kilometres multiplied by a transport factor and tonnes of waste multiplied by a treatment factor. Each is technically correct. If two of them describe the same underlying activity, the inventory that adds them together is not. RM500,000 of logistics expenditure multiplied by an appropriate factor is a sound calculation; if the same shipments have already been calculated from their weights and distances, adding the spend-based result counts the freight twice, and nothing in either calculation will reveal it.

That is the distinction this article is about. A correct calculation is a property of a formula. A correct inventory is a property of how the data behind every formula fits together, and the two can come apart without anyone having made a mistake.

Three kinds of double counting; only one is the problem here

Double counting is a loose term, and it covers situations the GHG Protocol treats very differently. Three are worth separating before going any further.

The first is not an error. The Corporate Standard defines Scopes 1 and 2 so that two companies do not account for the same emissions in the same scope, while accepting that one company’s Scope 1 can be another’s Scope 2. Scope 3 extends the same logic along the value chain: a supplier’s direct emissions are, quite properly, part of its customers’ Scope 3. Nobody should try to engineer that overlap out of their inventory.

The second and third are different, because both happen inside a single company’s report. The Scope 3 calculation guidance states that the fifteen categories are designed to be mutually exclusive to avoid a company double counting emissions among categories, and it defines Category 1 as purchased goods and services not otherwise included in Categories 2 to 8.

The categories, in other words, are clean. The data arriving at them is not organised by category, and one economic activity can reach the inventory through a spend-based calculation, an activity-based one, a second category, a separate operational dataset or a transaction between two companies in the same group. That is the practical problem, and it is the subject of the rest of this article.

The financial ledger is not an emissions inventory

The first article in this series, on spend-based Scope 3 emissions, argued that a spend-based calculation involves far more than expenditure multiplied by a factor: currency, price year and sector classification all have to be settled before the multiplication means anything. This article starts one step earlier, with what the expenditure represents.

Financial data is attractive for Scope 3 because it is complete. Almost everything a company buys passes through the general ledger, so a spend-based calculation offers coverage that no other dataset can match. The ledger, however, was built to record financial obligations rather than physical activity, and many of the activities it records in ringgit are also recorded somewhere else, in litres, kilowatt-hours, tonnes or kilometres.

None of this means every such line must be removed from a spend-based calculation under every methodology. Where no activity data exists, the expenditure may be the best representation of an activity there is. The claim is narrower than that.

Fuel and electricity: the obvious overlaps

Fuel is the clearest case. A company running a fleet or standby generators calculates Scope 1 from litres of diesel multiplied by a combustion factor and will usually calculate the upstream emissions of producing and delivering that diesel in Category 3, which the Scope 3 Standard names fuel- and energy-related activities not included in Scope 1 or Scope 2. The same diesel appears in the ledger as fuel expenditure. Run the whole ledger through spend-based factors and that expenditure acquires a second estimate.

Where the second estimate overlaps depends on what the spend factor covers, and it is easy to miss. An input-output factor for refined petroleum products describes the emissions of producing the fuel, not of burning it, so the overlap is chiefly with the upstream emissions already in Category 3. A spend factor that includes combustion overlaps with Scope 1 directly. Either way the diesel has been represented twice, and which part is duplicated is answered by the factor’s documentation rather than by the ledger.

Electricity follows the same logic with a sharper edge. Scope 2 is kilowatt-hours multiplied by a grid emission factor, with upstream and transmission-related emissions in Category 3. An input-output factor for the electricity sector is dominated by the emissions of generating electricity, which are precisely what Scope 2 already reports. Applied to the electricity bill, it counts the same generation a second time.

The response is not to delete every fuel and electricity line from Category 1 by reflex. A utility account can include connection charges, equipment and services that sit outside the kilowatt-hours; a fuel account can hold lubricants alongside diesel. The reporting boundary and the calculation methodology have to be established first. Only then is it clear which part of a transaction is already captured and which is not.

Waste, freight and business travel

The same pattern recurs wherever an activity has data of its own. A waste contractor’s invoice is expenditure; the same collections, recorded as tonnes by waste type and treatment route, are the activity data for Category 5. Logistics expenditure sits in the ledger while the shipments behind it can be calculated from tonnes, distances and modes in Category 4. Each pair is two representations of one activity.

The GHG Protocol reports these activities in their own categories deliberately, rather than inside Category 1, to make the inventory more transparent. A spend-based run across the ledger quietly undoes that separation, because it places a second estimate of the same activity in the category that was defined to exclude it.

Business travel deserves more attention than the other two, because the activity is rarely recorded in one place. A flight booked through a corporate travel agency produces an itinerary from which passenger-kilometres can be calculated, and the agency’s invoice is posted to a travel expense account. A flight an employee books on a personal card and reclaims later appears only as an expense claim, often alongside hotels, meals and taxis.

The calculation guidance itself lists travel agency records and internal expense and reimbursement systems as sources of distance data, which says something about how fragmented the underlying records tend to be. A company that calculates distance-based emissions from agency data and spend-based emissions from its travel accounts can count every agency-booked flight twice. A company that uses the agency data alone will miss the flights employees booked themselves.

Intercompany transactions

Group structures add another layer. Suppose Company A buys materials from an external supplier and then sells or recharges them to Company B, another subsidiary in the same group. The external purchase appears in A’s ledger. The internal sale appears in B’s ledger as a purchase, usually at a higher value. Consolidate the two ledgers, run them through spend-based factors and the group has estimated emissions on both.

At group level only one of those transactions is an external economic activity; the other is a movement inside the group’s own boundary. The Corporate Standard states the underlying principle in its reporting guidance: companies should take care to identify and exclude from reporting any Scope 2 or Scope 3 emissions that are also reported as Scope 1 emissions by other facilities, business units or companies included in the inventory consolidation.

That does not make every intercompany transaction removable. An internal recharge can carry services performed within the group, whose emissions already sit in the group’s own Scopes 1 and 2, or it can pass on an external cost that the originating entity never recorded as a purchase. Some internal lines duplicate an external activity already captured; others are the only record of it. What matters is distinguishing external economic activity from internal movements, recharges and transactions, and the consolidation approach the group has chosen decides where that line falls.

Double counting across methodologies, not just categories

The usual framing of double counting is categorical: has the same category been included twice? The Scope 3 categories were designed to make that difficult, and it is rarely where the problem lies. The more useful question is whether the same underlying activity has been represented through two different datasets or calculation methods.

Spend data, activity data, operational meter readings, travel system extracts, waste manifests, logistics records, fuel cards and electricity bills are each a legitimate view of part of a company’s activity. They were collected by different functions for different purposes, and they overlap in ways that none of them records. The calculation guidance makes a related point about secondary data: care should be taken to understand the boundaries the data covers, to minimise the potential for double counting.

This is why a sophisticated inventory is not automatically a reconciled one. Replacing a spend-based estimate with activity data is a genuine improvement in quality, and it is also the moment the risk appears, because the new calculation arrives while the old expenditure is still in the ledger. Where the switch happens between reporting years it can also be a reason for restating GHG emissions, which makes it doubly important to know which representation each year’s figure was built on.

The challenge is not simply deciding which emission factor is correct. It is establishing which representation of the underlying activity belongs in the inventory.

An exclusion for GHG is not an exclusion from the dataset

Once overlaps have been found, the natural instinct is to mark the duplicated transactions and remove them. That works for one calculation and causes problems in the next.

Take electricity expenditure again. It may be excluded from a spend-based GHG calculation because the electricity is already captured as kilowatt-hours multiplied by a grid factor. That reason belongs to the greenhouse gas inventory. The same transaction may still matter to an input-output estimate of water use or land use in the supply chain, to a spend-based supplier engagement analysis or to another environmental indicator calculated on a different method. Whether it is included there depends on that metric and its methodology, not on the GHG decision.

The answer is not simply “remove the duplicates”

A transaction marked as excluded, with nothing else recorded, answers none of the questions a reviewer will reasonably ask. Excluded from which calculation, under which methodology and why? Is the activity captured elsewhere and, if so, where? A decision that cannot answer those questions cannot be reviewed, and a decision that cannot be reviewed tends to be taken again, slightly differently, every time the inventory is rebuilt.

The GHG Protocol does not prescribe how exclusions should be recorded, so what follows is a practical interpretation rather than a requirement. A robust approach keeps apart the underlying transaction, the calculation it feeds, the methodology used, whether the activity is already captured elsewhere and the reason for including or excluding it. An exclusion is defensible when it is tied to the calculation it applies to and carries its reason with it. It is fragile when it is a flag.

The calculation needs to understand the transaction

The first article in this series ended on the observation that the emission factor is only the starting point. This one adds the step before it: the transaction itself has to be understood before any factor is applied. What was bought, which activity it represents, which method suits it and whether that activity already reaches the inventory by another route are all questions about the transaction. They have to be settled before the choice of factor even arises.

One activity, counted once

A robust Scope 3 inventory is not created by calculating more transactions. Coverage matters, and the ledger provides it, but coverage without reconciliation only moves the error from what is missing to what is repeated. The underlying activities have to be understood, classified and reconciled so that the same activity is not represented through several calculation routes at once.

None of this implies that anyone has done anything wrong. Overlaps arise because the datasets a company holds were built for different purposes and because better data tends to arrive one category at a time. They are a data and methodology problem, and they respond to the same discipline as any other: decide deliberately, write the decision down, apply it consistently and make it traceable.

Ace CSR supports Malaysian companies with Scope 3 assessment, GHG accounting and sustainability reporting, including the reconciliation of spend-based and activity-based data and the methodology notes that explain a Scope 3 figure to the people who have to rely on it.