Separate financial-services operational emissions from financed activity

A bank’s branch inventory and a financed-emissions schedule can both be expressed in tonnes CO2e without describing the same activity. One follows the institution’s operations; the other attributes emissions associated with selected financing. Before combining them in a management slide, preserve whose emissions each number represents, which population it covers and which method produced it.

This fictional review starts with supplied calculation outputs. It helps sustainability, finance and communications teams test a proposed summary before release. It does not select an attribution method, calculate a real portfolio, assess credit risk or recommend a financing decision.

Read the result boundary before its number

GHG Protocol’s Category 15 guidance concerns investment emissions not already included in the reporting company’s Scope 1 or Scope 2. Its treatment depends on the organisational boundary. This is why an investee’s own scope label cannot simply become the bank’s operational scope label.

PCAF’s December 2025 Part A addresses financed emissions from loans and investments. Its introduction distinguishes these from capital-market facilitation and insurance-associated emissions covered by Parts B and C. A schedule prepared for one activity is not evidence that the other activities have been measured.

The invented institution, Bank B, supplies operational statement O1 for 1 April 2025–31 March 2026. Its approved boundary covers the named branches, offices and other controlled operations. O1 reports Scope 1 of 120 tonnes CO2e and two Scope 2 results: 80 location-based and 50 market-based. The technical owner has already approved those inputs and their respective methods for this exercise.

These are two views of the same Scope 2 population, not separate sets of electricity to add. The Scope 2 Guidance describes separate location-based and market-based results and says they are not a gross/net pair. Keep the method in the label; this guide does not validate any electricity instrument or factor behind O1.

Keep the financing output with its stated population

Statement F1 reports 900 tonnes CO2e attributed to Bank B for a specifically selected corporate-lending population P, for the same reporting year. Its fictional method owner confirms the calculation, the covered borrower scopes and the treatment of overlap with O1. P contains only the selected loans; it excludes other lending, investments, facilitation and insurance activities. No claim that P is a complete financed-emissions inventory is supplied.

F1 includes borrower Scope 1 and Scope 2 contributions only. Borrower Scope 3 is absent and remains an explicit limitation. The supplied 900 is a teaching input, not a result derived here, and this restricted illustration does not establish conformity with PCAF or another reporting standard.

The evidence pack also contains reference R1, totalling 9,000 tonnes CO2e for the borrowers’ own covered emissions before attribution. R1 is supporting data behind F1. It is not another 9,000 tonnes of Bank B’s direct operational emissions, nor an additional financing contribution to put beside the attributed 900.

Retain the link between F1, the borrower inputs, financial-input dates, attribution calculation, method version and reviewer. A reader should be able to find the attribution basis without reverse-engineering it from 900 divided by 9,000. That quotient would be 10%, but it is not evidence of a single portfolio-wide ownership or lending share. Different records can have different attribution factors.

Review the proposed headline before the arithmetic

The draft slide says, “Bank B’s total carbon footprint is 1,070 tonnes CO2e.” The author added 120 + 50 + 900. The arithmetic is correct, but the label is unsupported: the pack excludes other operational Scope 3 categories, other financing populations and borrower Scope 3. Correct addition cannot establish complete coverage.

The immediate repair is to show O1 and F1 separately with their boundaries. O1 supports 200 tonnes CO2e for Scope 1 plus location-based Scope 2, or 170 using market-based Scope 2. F1 supports the supplied 900-tonne attributed result for P, limited to the stated borrower scopes. Neither of O1’s subtotals should be labelled the entire institution’s footprint.

If management specifically requests a combined subtotal, retain the exclusions and method choice beside it. Under the supplied non-overlap and period assumptions, the selected combination is 120 + 80 + 900 = 1,100 tonnes CO2e on the location-based operational view, or 120 + 50 + 900 = 1,070 on the market-based view. These are alternative selected subtotals, not additive totals or a claim of complete institutional emissions.

Adding 80 and 50 together would count both operational Scope 2 views in one result. Adding R1’s 9,000 would confuse borrower totals with the bank-attributed financing result. Keep those errors distinct: the first mixes alternative methods; the second mixes reporting entities and attribution levels.

Original fictional example. The operational component uses Scope 1 plus location-based Scope 2. Q is a subset of P; its unchanged result excludes 300 tonnes from the wider selected financing population. Neither subtotal is the bank’s complete inventory.

Test a smaller financing chart without claiming progress

A revised presentation F2 displays only subpopulation Q within P. The same-year F1 result comprised 600 tonnes for Q and 300 for the omitted remainder. F2 keeps Q’s same calculations and reports 600. No borrower emissions, financial inputs, attribution factors or operating results have changed; only the chart’s selection has narrowed.

Combining unchanged O1’s location-based 200 with F2 gives 800 tonnes. Comparing that with the earlier selected 1,100 produces a 300-tonne difference, exactly the omitted remainder. It is not evidence that the bank or its borrowers reduced emissions. These are two views of the same year with different financing populations, not a year-on-year trend.

Label F2 as Q-only and keep a bridge back to P: 600 + 300 = 900. If the intended management question still concerns P, restore the omitted component or show it separately. Do not rename Q as the full portfolio. In a real comparison, changes to exposure, counterparties, methods, underlying emissions and data availability would each need their own explanation.

Make the handover answer the right question

Give each displayed figure its reporting entity, activity population, period, unit, scope treatment, method, source version and owner. For O1, preserve the organisational boundary and the two Scope 2 method labels. For F1/F2, preserve the portfolio selection, borrower scopes, attribution basis, financial-input dates and explicit exclusions. Keep R1 visibly identified as underlying borrower information.

Ask a second reader to find the evidence for four statements: why 80 and 50 are alternatives; why 9,000 is not the bank’s own Scope 1; why the 900 is limited to P; and why the 300-tonne chart difference is a selection change. If the supplied statements cannot answer one, keep that claim unresolved rather than inventing a relationship.

The financed-emissions data-request guide addresses the earlier borrower-and-facility collection task. The branch-and-ATM guide addresses operational electricity evidence. This case begins after those activities and concentrates on the meaning of their resulting summaries. For help organising a reporting handover, contact Nitium Advisory with a non-confidential outline of the figures and the intended claim.

Sources and limits

Sources checked 14 October 2026: GHG Protocol Category 15 guidance, printed page 136; PCAF Part A, third edition December 2025, printed pages 7–8; GHG Protocol Scope 2 Guidance, sections 7.4 and 7.4.2. All entities, statements, quantities and selection changes are fictional. No real inventory, lending decision, standard conformity or independent assurance is established.