A different kind of Scope 3
Most companies doing Scope 3 accounting spend their effort on Category 1 (purchased goods and services) or Category 11 (use of sold products), because that's where the mass of a manufacturer's or retailer's footprint sits. Banks, asset managers, and insurers don't have that problem. Their own operations — office energy, business travel, employee commuting — are small. What dominates their inventory is Category 15: Investments.
Category 15 covers the emissions associated with an organization's investment and lending portfolio: equity holdings, corporate bonds, business loans, project finance, and similar instruments. For a financial institution, this category isn't a niche add-on. It routinely accounts for the overwhelming majority of a reported footprint, because the emissions of the companies and projects a bank finances are, in aggregate, vastly larger than the emissions of running the bank itself. A regional lender's own buildings and travel might amount to a rounding error next to the emissions of the industrial and energy borrowers on its loan book.
This is why financed emissions get treated as almost a separate discipline within Scope 3 accounting rather than just another category to tick off. The GHG Protocol's Corporate Value Chain Standard defines Category 15 in general terms, but it doesn't specify exactly how to allocate a borrower's or investee's emissions back to the lender or shareholder. That gap is what the Partnership for Carbon Accounting Financials, known as PCAF, was built to close.
What PCAF actually standardizes
PCAF extends GHG Protocol logic to financial instruments by defining, asset class by asset class, how to attribute a portion of a company's or project's emissions to each investor or lender. The core idea is an attribution factor: the share of a borrower's or investee's emissions assigned to a given financial institution based on the size of that institution's financial stake relative to the total capital structure of the entity being financed.
In practical terms, this means a bank doesn't just add up the total emissions of every company it lends to. It calculates, for each loan or investment, what fraction of that company's emissions corresponds to its share of the company's financing — typically outstanding debt plus equity, or an equivalent value measure depending on the asset class. Sum those fractions across the whole portfolio and you get the institution's financed emissions.
PCAF sets out methods across several asset classes, including listed equity and corporate bonds, business loans and unlisted equity, project finance, commercial real estate, mortgages, motor vehicle loans, and sovereign debt. Each asset class has its own attribution logic because the underlying financial structures differ — a mortgage is not a corporate bond, and neither behaves like project finance for a wind farm.
Data quality: same problem, different vocabulary
Anyone who has worked through a standard Scope 3 inventory will recognize the underlying issue here. The GHG Protocol's data quality hierarchy for Scope 3 runs from spend-based estimates at the weak end, through average-data and hybrid methods, up to supplier-specific primary data at the strong end. PCAF applies the same basic logic to financed emissions, scoring data quality from best (verified, company-reported emissions figures) to weakest (emissions estimated from proxies like sector averages or revenue when no reported data exists at all).
The practical consequence is that a bank's financed emissions number is only as reliable as the emissions disclosure of the companies in its portfolio. A loan to a company that publishes third-party-assured Scope 1 and 2 data produces a defensible attribution. A loan to a private company with no emissions disclosure gets modeled using sector-average intensity figures applied to revenue or asset value — which is directionally useful but not something to build a net-zero claim on without caveats. This is why financed emissions inventories tend to show a wide spread of data quality scores within the same portfolio, and why year-on-year changes can reflect improved borrower disclosure as much as actual emissions changes.
Why this is showing up on more compliance checklists
Financed emissions used to be a voluntary, investor-relations-driven exercise for institutions with public climate commitments. That's changing as mainstream climate disclosure regimes catch up. Under frameworks built on double materiality, such as the EU's ESRS E1 under CSRD, a financial institution's material Scope 3 emissions are expected to include Category 15, and PCAF is the most established methodology available to produce that number in a comparable, auditable way. California's SB 253 similarly requires Scope 3 disclosure from large companies doing business in the state, phased in with third-party assurance requirements — and a covered financial institution's Scope 3 profile will be dominated by the same category.
This creates a specific operational challenge: assurance providers increasingly expect not just a financed emissions total, but a defensible data quality score and a documented attribution methodology behind it. Institutions that treated Category 15 as a spreadsheet exercise a few years ago are now being asked to show their work.
What this means in practice
For a financial institution building or upgrading a financed emissions inventory, the priorities are usually: map the portfolio to PCAF asset classes correctly before doing anything else, prioritize obtaining reported emissions data from higher-emitting or higher-exposure borrowers first, and document attribution factor calculations in a way that survives external assurance. None of this is exotic Scope 3 work — it's the same measurement discipline applied to a category that happens to carry almost the entire footprint. Institutions building this capability from scratch, or preparing financed emissions figures for CSRD or SEC-adjacent disclosure, are the ones most likely to find the data quality gaps before an auditor does.