Scope 3 Spend-Based Emissions: Why the Emission Factor Is Only the Starting Point

The first greenhouse gas calculations most teams learn are the reassuring ones. A company knows how many litres of diesel it burned, finds a published factor in kilograms of CO₂e per litre, multiplies one by the other and has a Scope 1 figure it can defend. The activity data and the emission factor describe the same physical thing in the same physical unit, so the multiplication carries almost all of the work.

Spend-based Scope 3 looks like the same exercise in different clothing. Where better information is not available, the GHG Protocol accepts expenditure as a proxy for activity, and the arithmetic reads identically: expenditure multiplied by an emission factor expressed in kilograms of CO₂e per unit of currency. The difficulty is that the two sides of that multiplication rarely arrive compatible with one another. A litre is a litre in any year and in any country. A ringgit is not.

Money carries a currency, a date and a price level. A monetary emission factor carries its own version of each of those, inherited from the economic data it was built from. Before the multiplication means anything, the expenditure has to be expressed on the same basis as the factor about to be applied to it. That is where most of the work in a spend-based calculation actually sits, and none of it is visible in the formula.

A spend-based factor is not a physical constant

Spend-based factors are generally environmentally extended input-output factors, built by combining a country’s economic accounts with its environmental accounts to give emissions per unit of economic output for each sector of that economy. That derivation is exactly what makes them useful. They cover the whole of an economy, including the parts of a supply chain a company will never see and could not survey if it tried. It is also what gives every one of them a set of attributes that a factor for a litre of diesel simply does not have.

None of this is hidden. Publicly documented input-output models state the currency and the price year their published values are expressed in, because the people who build them know it matters. The information is simply easy to walk past on the way to a number, particularly when the number is one cell in a table of several hundred.

The Scope 3 Standard frames the consequence as a question of data quality rather than of arithmetic. Among the indicators it sets out for assessing the data behind an inventory are technological, temporal and geographical representativeness: whether the data used genuinely represents the activity, the time period and the location being reported. Currency and price year are the temporal question in monetary form. Geography is a larger subject and has the third article in this series to itself.

Put concretely, MYR 4 million of expenditure recorded in a 2026 financial year and a factor expressed per USD of output in an earlier price year have two gaps between them rather than one. Neither gap is difficult to close. Both are decisions, and both should be written down at the time rather than reconstructed from memory when somebody asks.

The two totals differ by around five percent, and nothing whatever about the company’s activity differs between them. The same goods were bought on the same dates at the same prices. Only the monetary basis handed to the factor has changed, and since the factor is expressed per unit of currency, that difference reaches the reported emissions in full.

This is not an argument that a single closing rate is wrong. In a year when the currency barely moves, the methods converge and the choice is immaterial. It is an argument that the choice should be made deliberately and recorded, because in the years when it does matter it can move a category total by more than most of the genuine improvements a sustainability team will spend that year working on.

Finance teams already solve a version of this problem, which is worth remembering before inventing a new answer. MFRS 121, Malaysia’s adoption of IAS 21, translates a foreign currency transaction at the spot rate on the date of the transaction and permits an average rate for a period where it approximates the actual rates and exchange rates have not fluctuated significantly. A GHG inventory is not bound by that standard. The reasoning behind it transfers cleanly, and a company whose finance function already applies a documented translation policy has an obvious and defensible place to start.

The price year behind the factor

The second gap is less visible than the first, because nothing in the data announces it. An expenditure figure does not carry a label saying which price level it belongs to, and neither does the value in the factor table.

A factor derived from economic activity in a given year is tied to the purchasing power of that year. It says, in effect, that a unit of currency spent in that sector in that year corresponded to a certain quantity of emissions. Apply it unchanged to expenditure from a later year and the calculation quietly assumes that a unit of currency still buys what it bought then.

This is not a refinement companies are left to discover for themselves. Among the activity data the GHG Protocol’s calculation guidance lists for a spend-based calculation is inflation data to convert market values between the year of the EEIO emission factors and the year of the activity data. The adjustment belongs to the method rather than sitting on top of it as an optional embellishment.

Usually it does not. Suppose a company spends USD 1 million on a category of goods in 2026, and the best available factor is expressed per USD of output in 2021. If prices in that sector rose appreciably over those five years, the 2026 million buys materially less physical output than the 2021 million would have. Applying the factor to the unadjusted amount credits the purchase with output it never received, and the estimate drifts upwards for reasons that have nothing to do with the company or its suppliers. Where prices have fallen, the drift runs the other way, which is why this is a price-year question rather than an inflation question.

Choosing the adjustment is a methodological exercise in its own right, and a more interesting one than it first appears. General consumer price inflation and the price movement of the particular category purchased are not the same thing, and the gap between them can be wide. Electronics, freight, construction materials and professional services have all behaved quite differently from any headline index over the past decade. The index applied should bear some relationship to what was actually bought, the reasoning should be documented, and the same reasoning should hold across categories and across reporting years.

A more general habit sits underneath both of these sections. IFRS S2 asks an entity to disclose the measurement approach, inputs and assumptions behind its reported greenhouse gas emissions, and exchange rate treatment and price-year adjustment are inputs and assumptions in precisely that sense. They are also among the easiest things in a Scope 3 calculation to lose track of, because neither leaves a visible trace in the result. Two companies can publish the same figure for the same purchasing having made opposite choices about both, and nothing in either disclosure would reveal it unless somebody wrote it down. Building that habit early is a large part of what IFRS S1 and S2 readiness involves in practice.

Where the method becomes a data problem

Everything described so far applies to a single invoice, and for a single invoice it is an afternoon’s work. Groups do not have a single invoice.

A listed group preparing a Scope 3 inventory may be working from tens or hundreds of thousands of transaction lines, drawn from several subsidiaries, denominated in a handful of currencies, posted against ledger structures that differ between entities, spread across every date in the reporting year and occasionally sitting behind financial year ends that do not align. Each individual calculation stays simple. The inventory is the sum of all of them, made the same way, and demonstrably so.

Excel is not incapable of any of this, and the point is not that spreadsheets are the problem. It is that at sufficient scale a workbook stops being a spreadsheet and starts being a transaction-processing system, with everything that implies: version control, reproducibility, reconciliation, change management and an audit trail. Most workbooks acquire those responsibilities long before anyone decides to give them the controls that go with them.

The arithmetic is not what gets harder. These are:

• Applying the same treatment to comparable transactions across every entity in the group
• Reproducing a published figure months later, from the same inputs, and getting the same answer
• Knowing which version of which factor set produced which result
• Changing a methodology and understanding what it does to every line it touches
• Tracing a category total in the report back to the transactions that produced it
• Handling exceptions without handling them differently each time
• Recalculating when a factor is revised, which happens more often than people expect

The last item on that list deserves a note of its own. Published factors get revised, methodologies improve and better activity data arrives to replace spend-based estimates, all of which raises the question of whether the historical figures should move as well. That is the subject of restating GHG emissions, and a calculation that cannot be re-run on an earlier year’s data has quietly answered the question by default.

Assurance sharpens the whole of this. As assurance expectations tighten through the phase-in of Malaysia’s National Sustainability Reporting Framework, the questions asked about a spend-based figure are rarely about the multiplication. They are about whether comparable transactions received the same treatment, whether the assumptions were applied consistently across entities and periods, and whether a category total in the report can be traced back to the transactions underneath it. Those are properties of a process rather than of a formula, and they are difficult to retrofit once a number has been published.

Not every transaction belongs in the same calculation

There is a further question that sits before all of the above, and no amount of care with factors will answer it.

An expenditure ledger is organised for financial accounting. It was never designed as a list of things to apply emission factors to. Run a spend-based calculation across the whole of it and the result will draw in items that are measured properly somewhere else:

• Freight, calculated from weights, distances and modes in its own category
• Waste, calculated from tonnages and treatment routes
• Business travel, which has its own activity data
• Fuel and electricity purchases already measured in Scope 1 and Scope 2
• Capital expenditure, which belongs to the capital goods category
• Intercompany transactions, which appear twice in a group ledger and represent one flow of goods

Applying a well-chosen factor correctly to a transaction that should never have reached it still produces a wrong inventory, and it does so silently. The calculation succeeds. The number looks plausible. Nothing in the output indicates that the same activity has been counted twice, which is what makes this failure mode so durable: it survives review precisely because there is nothing to see.

Selecting the correct factor is therefore only part of the exercise. Determining which transactions should reach that factor can be equally important, and it is the subject of the next article in this series.

Where the money was spent is not where the goods were made

One further assumption travels with a spend-based factor, quietly, and it concerns geography.

A Malaysian company buying electronic equipment from a Malaysian supplier has a Malaysian invoice. The equipment may have been assembled in one economy from components manufactured in several others, using materials and energy drawn from several more. The invoice records where the company bought the goods. It records nothing at all about where the emissions from producing them arose, in which economies, or at what emissions intensity.

Currency and price-year adjustments make the financial value compatible with the factor. They do not make the factor’s geographical assumptions representative of the supply chain behind the purchase. That is a separate problem with a separate treatment, and it is the subject of the third article in this series.

What the formula leaves out

Spend-based Scope 3 is routinely described as expenditure multiplied by an emission factor. The formula is not wrong. It is an incomplete description of the work involved in producing a figure anybody can defend, and what it leaves out is not a refinement at the margin.

The emission factor matters. So does what the expenditure represents, when it was incurred, which currency recorded it, which monetary year the factor belongs to, how the transaction was classified and whether it should have been in that calculation at all. Get those right and a reasonable factor does its job well. Get them wrong and an excellent factor will produce a figure that describes nothing in particular, with no indication on the face of it that anything has gone astray.

That is the more useful way to read the formula. Mathematically, multiplication is the final step. Methodologically, it is usually the easiest one.

Ace CSR supports Malaysian companies with Scope 3 assessment, GHG accounting, IFRS S1 and S2 advisory and sustainability reporting, including spend-based and activity-based calculation methodologies, emission factor selection, and the methodology notes that explain a Scope 3 figure to the people who have to rely on it.