The proposition that product emissions should be kept in accounts rather than compiled into inventories: each emission recorded once at its source, allocated to outputs by causal logic, transferred with legal title when the product is sold, reconciled from transaction to product to entity, and audited to the assurance level financial statements are held to.
Sourced throughout from The Carbon Accounting Landscape Review (ICC, 2026) unless marked otherwise. That document is an advocate for this architecture and says so; read the publisher page before treating its verdicts as neutral.
Counting vs. accounting
The distinction the whole argument rests on. From Table 1b (p. 14, adapted by the Review from EFI Foundation, Towards Harmonized Carbon Accounting, March 2026):
| Dimension | Carbon counting and reporting | Carbon accounting |
|---|---|---|
| Data collection | Ongoing collection with periodic consolidation, typically mixing primary and secondary sources | Transaction-based recording at each exchange, resting on data recorded at source |
| Time period | Mixes past, present and projected periods, drawn from life-cycle models; results compiled for a study period and often reused across reporting cycles | A specified period, aligned with transactions and financial records |
| System boundary | Combines attributable direct and indirect emissions; may allow consequential assessment | Attributable data within fixed boundaries (i.e. gate-to-gate) |
| Data quality | Potential for gaps, estimates, inconsistencies and overlapping emissions across reporters | Completeness and consistency framed by carbon mass balance, a rulebook, and double-entry crosschecks |
| Recording device | Single-entry inventory tabulation | A dual-sided ledger with a structured chart of accounts and double-entry recording that allows data to flow along the value chain |
| Reports generated | No widely adopted framework generates both entity- and product-level reports that reconcile | Generates fully aligned entity- and product-level reports |
The Review is emphatic that “the distinction between these two concepts is not merely terminological. Many missing elements to implement a ledger-based approach identified later in this study trace back to the difference between compiling an inventory and keeping an account.”
Note what this does not claim: accounting does not replace the quantification science. “The architecture does not substitute for the science” (§6.2). LCA stays inside the ledger.
The glossary definition (Appendix B) is the compact version: carbon accounting is “the recording of emissions as transactions within defined boundaries of time and space, using the discipline of financial accounting — double-entry recording, a chart of accounts, an audit trail — so that figures are accurate, complete, comparable, and verifiable, and reconcile from transaction to product to entity.”
Why a PCF is not a measurement
The premise underneath the premise (Executive summary, p. 5):
“Unlike sulfur content, or other product characteristics that can be tested directly, a product carbon footprint (PCF) is an informed accounting calculation built from activity data, emission factors, supplier information, boundary choices, and allocation rules. Because emissions released during production do not leave a physical imprint on the product itself, users cannot determine from the product whether one reported PCF reflects better carbon performance or simply different calculation choices.”
This is why rules, not instruments, are the binding constraint — and why the Review calls the target “the equivalent of GAAP/IFRS for carbon accounting”.
The comprehensive-ledger family
Three related but distinct approaches (§5.4). They agree on the paradigm and differ on what the two sides of the ledger represent:
| Ledger sides | Emphasis | |
|---|---|---|
| E-Ledgers / E-liability (E-Ledgers Institute; Kaplan & Ramanna) | E-liabilities = emissions, E-assets = removals | Transfer with legal title; record once at source; never remove from a ledger except by transfer to another ledger; audit-by-design; incentive-compatible primary-data transfer |
| EFI Foundation | Stocks (carbon content) and flows | Explicit carbon mass-and-energy balance; gate-to-gate; capital-goods emissions amortised into product outputs; product figures aggregate without gaps to entity totals. Built on the E-Ledgers architecture |
| CASI (Carbon Accounting Standards Initiative) | Stocks and flows | Internal bookkeeping; assumes either emission factors for supply-chain data or an E-Ledgers-style primary-data transfer; investor-grade aim |
The E-Ledgers proto-standard’s five rules, as the Review restates them (§4.1): record direct emissions at source; receive embedded emissions from suppliers on legal transfer; allocate to outputs by causal logic; verify to a reasonableness standard; and never remove emissions from a ledger except by transfer to another ledger.
Status per §5.1: “Emerging — pilots, prototypes, and conceptual development across academic, policy, and industry contexts.” Concept introduced from 2021 onward.
What the architecture is claimed to fix
From §6.3 (missing elements in accounting architecture), stated as the problems no product standard closes:
- No formal transfer. “Mainstream product standards treat a footprint as an estimate, not a quantity that moves with the product.”
- No reconciliation. Product standards “avoid double counting within a study but cannot propagate emissions as conserved, accountable quantities between entities, and so cannot reliably reconcile a product figure to an entity total without gaps or overlaps.”
- Assurance below the financial-accounting level. Assurance is “often not strictly required at the standard level […] and self-verification is permitted”, while regulation moves to mandatory reasonable assurance.
- Not transaction-level or timely. “Annual, estimation-based studies struggle to meet the timeliness and operability that compliance cycles and enterprise decisions now demand.”
Chain of custody — the unresolved choice
The Review does not pick a model, and flags the choice itself as open (§6.3). Three models, “in declining order of physical linkage” (Appendix B):
- Physical segregation — but “a decision-grade framework should not assume that physical segregation is always feasible, especially for fuels and commodities moving through shared tanks, pipelines, terminals, and trading infrastructure”.
- Mass balance — certified and non-certified inputs mixed physically but tracked in the accounts, so certified output claims cannot exceed certified inputs over a balancing period. ISCC CFC uses this for co-processing and is credited with the operational experience.
- Book-and-claim — attribute decoupled from physical flow, claimed via registry instruments. The Review will not “allow book-and-claim without safeguards”.
Open design questions it names: which custody models are acceptable by product, supply-chain segment and use case; what registry, retirement and reconciliation rules are required; how balancing periods and boundaries are set; and how claims can be made without double counting across sectoral, corporate and procurement systems.
What is settled and what is not
§6.1 credits the existing landscape with: a thriving quantification science (ISO 14040/14044 → ISO 14067, GHGP Product Standard); a shared allocation hierarchy “avoid → physical → economic”; a clear data direction (PACT’s Primary Data Share and CBAM’s tightening defaults as “two complementary mechanisms — disclosure and consequence”); working exchange infrastructure (PACT); the architecture articulated but not deployed (E-Ledgers); and regulators converging.
§6.2 lists what is missing in quantification: comparability strong enough for differentiation; a tightened causal allocation rulebook; the recognition that allocation is market design; that verification does not guarantee comparability; methodology arbitrage and transaction costs; supply-chain propagation of divergence; and consistent treatment of the “hard cases” (removals, biogenic carbon, capital goods, circular flows, market instruments, offsets).
The hard cases are explicitly left open: “This study does not take a position on those choices.”
The confidentiality gap
The Review’s Criterion B states confidentiality as part of verifiability’s design — “product-level embedded emissions can reveal energy intensity, process efficiency, and cost structure” — and then carries the requirement into none of its later instruments. ⟨Inference: the ledger architecture sharpens exactly the exposure that criterion says must be limited, because transaction-level records with legal transfer disclose far more than a period average does.⟩
This has its own page: The confidentiality gap in ledger-based carbon accounting — the verbatim passages, the instrument-by-instrument absence, the argument that the ledger worsens the problem, and the caveats against over-reading it.
What the ledger is actually for ⟨added 2026-08-17⟩
⟨Inference: mine, from reading the two verifiable-carbon-accounting papers against this architecture — see Verifiable Carbon Accounting.⟩
The Review argues for the ledger by analogy to financial accounting: dual-sided entries, a chart of accounts, an audit trail. The analogy is apt but it buries the load-bearing part. What double-entry supplies that single-entry inventory does not is a closing constraint: a rule that the reporter does not control, against which every individual record must reconcile.
That matters because it is the one thing cryptography cannot manufacture. A zero-knowledge proof can show that a set of figures reconciles to a total; it cannot establish that the total is right, nor that the total covers everything it should. Man & Kleppmann prove that a data centre’s allocation is complete relative to a signed electricity meter — and their own conclusion concedes that the single-meter assumption “allowed us to side-step the problem of ensuring that all sources of emissions have been accounted for”, while §5 concedes that an operator with several meters could simply leave some out of the accounting. Their proposed repair is the instructive part: cross-reference the meter data against the electricity supplier’s bills, so that faking consumption would require colluding with the supplier — a second, independent attestation of the same aggregate. A supply chain has neither: no meter, and no single party entitled to observe the whole of it.
⟨Inference: so the ledger’s conservation rule and the data centre’s electricity meter are the same device built out of different materials — one institutional, one physical. Read this way the architecture and the cryptography are complements rather than rivals: the ledger supplies the closing constraint across entity boundaries, the proofs let firms satisfy it without opening their books. Neither ICC nor either paper says this, and it is the strongest thing SINE has to contribute to the framework process. See open tasks, item 1.⟩
Cross-references
- The Carbon Accounting Landscape Review (ICC, 2026) — the source
- Carbon Measures — the advocate
- Decision-grade carbon-data criteria — the eight criteria, incl. Conservation and Fungibility
- The confidentiality gap — the thesis this page points to
- Verifiable Carbon Accounting — the cryptographic complement, and the closing-constraint argument in full
- Temporal resolution and the service case — two construction faults in what this architecture is scoped for
- Product-level carbon regulation — the demand side
- Pathfinder Network — the exchange layer this architecture sits above
- PACT PCF Calculation Methodology — the allocation hierarchy in PACT’s own words
- Coordination failure vs. bargaining failure
- Data Commons · Digital Product Passport
Sources
sources/whitepapers/2026-08-17-icc-carbon-accounting-landscape-review/2026-08-17-icc-carbon-accounting-landscape-review.md(Executive summary, §2, §4.1, §4.2 B, §5.1, §5.4, §6.1–6.3, Appendix B)sources/whitepapers/2026-08-17-icc-carbon-accounting-landscape-review/tables-recovered-from-pdf.md(Table 1b, Table 4)