The demand side. Three regimes that attach financial consequences to a product-level embedded-emissions figure — the first regulation that cannot be satisfied by an entity inventory. Characterised in §3 of The Carbon Accounting Landscape Review (ICC, 2026).

Framing principle: “Any government regime that determines a tariff, a credit, a compliance obligation, or a penalty requires a robust system of measurement, reporting, and verification (MRV). […] the underlying metric must be calculated using a defined method, cover a defined scope, and be subject to independent verification.”

The three regimes compared (Table 2, p. 18)

DimensionEU CBAMCalifornia LCFSChina CFMS
Status & purposeMandatory; price carbon embedded in imports; prevent leakageMandatory; credit/deficit market on fuel carbon intensityEmerging national PCF infrastructure; trade compatibility
LevelProduct (per imported good)Product (per fuel pathway)Product (per good)
CalculationSpecific embedded emissions (tCO₂e/t) by standardised process, allocated following ETS-consistent process attribution rules, reported per CN codeWell-to-wheel life-cycle carbon intensity (gCO₂e/MJ) via the mandated CA-GREET4.0 modelLife-cycle PCF per GB/T 24067-2024 (ISO 14067-based); national emission-factor database
CoverageIron and steel, aluminium, cement, fertilisers (incl. ammonia), electricity, hydrogen; expanding to further ETS-relevant sectorsTransportation fuels (gasoline, diesel, ethanol, RD, RNG, hydrogen, electricity, alt-jet)Cross-sector; 100 key products by 2027, 200 by 2030; priority high-trade-exposure sectors
BoundaryGate-level direct emissions plus indirect from electricity where applicable, plus emissions attributed from certain precursor inputsCradle-to-grave (well-to-wheel), including indirect land-use changePer study; often cradle-to-gate or broader
Data basisVerified primary preferred; punitive tightening defaults where unavailableOperating inputs to a fixed model; Tier 1 secondary defaults + limited primary / Tier 2 high primary share (customised modelling)Primary encouraged; national secondary database for gaps
VerificationThird-party, accredited, reasonable assurance (definitive phase)Mandatory initial validation + annual third-party verificationTraceable, auditable data; assurance regime maturing
Removals / biogenicRemovals not currently recognised; biogenic per RED criteriaNet life-cycle treatment; biogenic in pathway CIPer ISO 14067 life-cycle treatment

Legal bases cited: Regulation (EU) 2023/956 and Commission acts; Directive 2003/87/EC; California Title 17 CCR §§ 95480–95503 with CA-GREET4.0 effective 1 July 2025, CARB LCFS Verification Program and Guidance 22-02, 2024 amendments effective 1 July 2025; China CFMS / GB/T 24067-2024.

The contrast case: Australia

Australia’s Safeguard Mechanism (Clean Energy Regulator) is a facility-level baseline-and-credit scheme: declining emissions-intensity baselines, ACCUs or SMCs surrendered on exceedance, tradable SMCs generated by beating the baseline, measured under NGERS as the mandatory MRV backbone — the way CA-GREET4.0 underpins LCFS reporting. It operates at facility level and produces no product-level embedded-emissions data.

The Review uses it to make its own point sharper: “no existing regime yet reaches true product-level accounting.” (Footnote 14, p. 19.)

What regulators are converging on (§3.5)

Four themes, read across the three regimes:

  1. Product-level boundaries — tight and prescribed, not left to the reporting entity: gate-level for CBAM, well-to-wheel for the LCFS.
  2. Standardised calculation — “The LCFS mandates a single model. CBAM mandates a single methodology. Both offer a narrower range of methodologies than voluntary LCA-based standards — a restriction designed to promote comparability.”
  3. Primary data, enforced by consequence — verified primary data as the default; secondary data made “progressively more costly or more strongly penalised”.
  4. Mandatory third-party verification — reasonable assurance by accredited verifiers “becoming the norm rather than the exception”.

The strategic move: harmonise the data layer, not the regimes

The most transferable idea in this section, and one the wiki should keep (§3.5):

“It must do so, however, without presuming that future regimes will converge on identical designs. Jurisdictions face different political constraints, and their boundary, coverage, and instrument choices will continue to differ […] The practical implication is to harmonise the data layer rather than the regulatory designs that sit on it: a single underlying set of verified, transferable embedded-emissions records from which each regime draws what its rules require.”

⟨Inference: this is the same architectural bet as the Data Commons — infrastructure below the policy layer, neutral to which policy sits on top — arrived at from the regulatory side rather than the governance side. See Data Commons.⟩

The gap the Review reports

“For those regulations at the product-level, no standard in existence today has been consistently referenced to underpin compliance and in many cases, policymakers are defining bespoke approaches to meet individual needs.” The consequence it names: cost and complexity of compliance rising for corporations “without optimising the opportunity for emissions reduction at scale.”

Also worth recording: most emissions regulation today still operates at the entity or facility level. Product-level regulation is the leading edge, not the mainstream.

Cross-references

Sources

  • sources/whitepapers/2026-08-17-icc-carbon-accounting-landscape-review/2026-08-17-icc-carbon-accounting-landscape-review.md (§3.1–3.5 incl. footnotes 11 and 14, Executive summary, “Regulatory regimes” reference list)
  • sources/whitepapers/2026-08-17-icc-carbon-accounting-landscape-review/tables-recovered-from-pdf.md (Table 2)