What is not settled is where the underlying borrower data comes from. Most banks frame the answer as build in-house or buy a vendor platform. There is a third option, connecting to reporting infrastructure borrowers already use, and it changes a different part of the problem.
Summary
- Financed emissions sit in GHG Protocol Scope 3, Category 15. PCAF describes them as the largest portion of a financial institution's overall emissions.
- PCAF scores data quality from 1 (highest quality, verified borrower-reported data) to 5 (lowest, economic-activity proxies), with scorecards specific to each asset class.
- Supervisors have found heavy reliance on proxy data. The ECB reported that more than 80% of banks used proxies to estimate Scope 3 emissions.
- Borrower data availability is a major constraint on data quality, alongside the methodology, attribution, aggregation and governance work that sits with the bank.
- PCAF is explicit that data limitations should not stop a bank from starting, and that estimated data still identifies emission-intensive hotspots.
- Building, buying and connecting are not mutually exclusive. Shared infrastructure feeds the data layer rather than replacing a calculation stack.
What Are Financed Emissions and Why Do They Dominate a Bank's Footprint?
Financed emissions are the greenhouse gas emissions attributable to a bank's loans and investments, not its own offices or vehicle fleet. The GHG Protocol classifies them under Scope 3, Category 15 (Investments), covering equity, debt and project finance ghgprotocol.org.
PCAF, the standard-setter for this calculation, states plainly that the emissions associated with financial activities under Category 15 are the largest portion of a financial institution's overall emissions carbonaccountingfinancials.com.
The size of the gap has been measured, though the figure needs its context. CDP's first analysis of the sector, The Time to Green Finance (April 2021), found that across 332 financial institutions holding around USD 109 trillion in assets, portfolio emissions averaged over 700 times their direct emissions.
CDP later updated that to over 750 times on 2022 data cdp.net. The same report carries the caveat: only a quarter of those institutions disclosed portfolio emissions at all, so the multiple describes the banks that measured, not the sector as a whole.
What Is the PCAF Data Quality Score?
The Partnership for Carbon Accounting Financials publishes the Global GHG Accounting and Reporting Standard for the Financial Industry. Part A, now in its third edition (December 2025), covers financed emissions across asset classes including listed equity and corporate bonds, business loans, project finance, commercial real estate, mortgages, motor vehicle loans, sovereign and sub-sovereign debt carbonaccountingfinancials.com.
Alongside the calculation methodology, PCAF defines a data quality score. The direction matters and is often reported backwards: 1 is the highest quality, 5 the lowest. Scores attach to how the emissions figure was derived, and the scorecards are specific to each asset class.
PCAF's reporting requirement is transparency rather than a minimum score. Institutions should publish a data quality score weighted by outstanding amount, and where scope 3 emissions are reported, that weighted score must be reported separately from scopes 1 and 2 carbonaccountingfinancials.com.
How Much of the Sector Is Running on Estimates?
Supervisors have measured this directly, which makes it firmer ground than industry survey data.
The European Central Bank reported that its 2022 climate risk stress test found more than 80% of banks used proxies to estimate Scope 3 emissions, producing high dispersion in the figures reported.
In the same review, half the banks in the sample did not disclose financed emissions at all, and of those that did, 53% could not sufficiently substantiate their measurement approaches bankingsupervision.europa.eu.
PCAF's own emission factor database is built for this reality. It supplies factors at data quality scores 3 to 5, precisely because reported borrower data is not always available carbonaccountingfinancials.com.
Why Borrower Data Is a Major Constraint, Not the Only One
Improving from a score of 5 towards 1 depends substantially on the borrower, because scores 1 to 3 all require something the borrower supplies. A bank can license capable software and still sit at score 4 if its borrowers have nothing to report.
That is one constraint among several. Methodology selection, attribution, data mapping, portfolio aggregation, governance and systems integration all remain on the bank's side, and none of them are solved by better borrower data alone.
It is worth being precise about what PCAF asks here, because the standard is frequently read as discouraging estimates. It does not. PCAF states that options based on reported and physical activity data are preferred from a data quality perspective, and recommends using company-reported emissions where available.
In the same section it states that data limitations should not deter financial institutions from taking the first steps, and that even estimated or proxy data can help identify emission-intensive hotspots carbonaccountingfinancials.com. A score of 5 is a starting point, not a failure.
Why Smaller Borrowers Are the Hardest Segment
Larger corporate borrowers increasingly have sustainability teams and existing disclosures. Smaller borrowers usually have neither, and no reporting obligation of their own that would produce the data a bank needs.
EU law now recognises this directly. The European Commission's 2025 recommendation on voluntary SME reporting describes a trickle-down effect, whereby companies subject to reporting requirements request sustainability information from companies in their value chain that are not eur-lex.europa.eu.
Directive (EU) 2026/470 goes further, recording evidence that value chain companies including SMEs receive disproportionate requests for information, and giving smaller suppliers a right to decline requests that exceed the voluntary standard eur-lex.europa.eu.
For a bank, that is a practical signal as much as a legal one. Data collection designed as a one-off questionnaire to each borrower is running against the direction of travel.
Build, Buy or Connect: What Are the Real Options?
Most banks frame this as two options. A third exists, and the three are not alternatives to one another.
Building and buying both address how the bank processes data. Connecting addresses where the data comes from. A bank that already has internal systems or a vendor platform can still connect to shared infrastructure, which strengthens the data layer feeding those systems rather than replacing them.
What Changes When Borrowers Already Report Into Shared Infrastructure
Without shared infrastructure, the same borrower completes near-identical sustainability questionnaires for Bank A, Bank B and an enterprise customer, each in a different format, each collected separately.
The scale of that overlap has been documented. EFRAG's secretariat analysed 12 real ESG questionnaires covering roughly 26,000 SMEs, issued under initiatives spanning around 700 banks and 450 value chain companies, plus one rating agency questionnaire used by more than 100,000 companies. Energy and greenhouse gas emissions were requested in all 12; water and waste in nine each efrag.org.
Shared infrastructure separates data collection from data consumption. The borrower reports once; multiple parties, including their bank and their procurement customers, draw on the same underlying record, subject to the borrower's permissions.
Interoperability at the standards level is moving the same way. The IFRS Foundation and EFRAG jointly published interoperability guidance in May 2024 to reduce duplication for companies applying both ISSB Standards and ESRS, though the document is educational material rather than a formal statement of equivalence ifrs.org.
Frequently Asked Questions
Why do banks need financed emissions software?
It applies PCAF methodology consistently across large portfolios and produces an auditable trail. It does not generate borrower data that has not been measured, so it addresses calculation rather than collection.
Can a bank improve its PCAF data quality score without asking every borrower for data?
Only partly. Moving between the economic-activity scores is possible with better internal data, but reaching scores 1 to 3 requires borrower-specific input. That is why infrastructure borrowers already use is relevant.
Does PCAF discourage the use of estimated data?
No. PCAF prefers reported and physical activity data on quality grounds, but states that data limitations should not deter institutions from starting, and that proxy data still identifies emission-intensive hotspots. The requirement is to disclose the weighted data quality score, not to reach a particular one.
Is transition finance reporting the same as financed emissions reporting?
They are related but distinct. Financed emissions reporting quantifies the footprint of a portfolio. Transition finance reporting typically tracks the allocation and use of proceeds towards decarbonisation activities, often drawing on the same borrower emissions data.
About Gprnt
Gprnt is sustainability data infrastructure launched by the Monetary Authority of Singapore (MAS) and operated by the Global Finance & Technology Network (GFTN), a not-for-profit MAS established in 2024. It is open digital infrastructure connecting SMEs, enterprises, financial institutions and governments through one interoperable data layer.
Rather than adding another separate system, it lets institutions connect to borrowers already reporting sustainability data, through its Disclosure and Value Chain Intelligence pillars. Entry-level Scope 1 and 2 reporting is free for SMEs, and reports generated on the platform are aligned with the GHG Protocol and ISSB standards and are audit-friendly.
To see how connecting to shared infrastructure fits alongside a bank's existing financed emissions stack, visit gprnt.ai.
References
- PCAF, The Global GHG Accounting and Reporting Standard for the Financial Industry, Part A: Financed Emissions, 3rd edition (December 2025) (carbonaccountingfinancials.com)
- PCAF Emission Factor Database information sheet (2025) (carbonaccountingfinancials.com)
- GHG Protocol, Corporate Value Chain (Scope 3) Accounting and Reporting Standard (ghgprotocol.org)
- ECB Banking Supervision, The importance of being transparent (April 2023) (bankingsupervision.europa.eu)
- CDP, The Time to Green Finance: finance sector's funded emissions over 700 times greater than its own (cdp.net)
- EFRAG, VSME Exposure Draft Basis for Conclusions (January 2024) (efrag.org)
- Commission Recommendation (EU) 2025/1710 on a voluntary sustainability reporting standard for SMEs (eur-lex.europa.eu)
- Directive (EU) 2026/470 (eur-lex.europa.eu)
- IFRS Foundation and EFRAG, ESRS-ISSB Standards Interoperability Guidance (May 2024) (ifrs.org)