A circular-economy metric proposed by the Hitachi-AIST CECRL and being driven toward international standardization. It expresses the economic efficiency of circular activity as a productivity ratio computed from ordinary financial-statement items.
⟨Page substantially rewritten 2026-10-16 after the CE Association white paper arrived; the first version, written from Hitachi’s one-paragraph summary, got the input data and the system boundary wrong. The corrections are flagged inline, because both errors are instructive.⟩
Definition
CVP = Y_CE / E_CE (1)
CVP = (S_CE − I_CE) / (C_CE + G_CE) (2)
CVP_VN = Σ(S_CE,i − I_CE,i) / ( Σ(C_CE,i + G_CE,i) − Σ C_D,i ) (3)
Y_CE value added from circular actions · E_CE expenditure on circular actions · S_CE sales · I_CE intermediate goods costs · C_CE cost of goods sold · G_CE SG&A · C_D purchase costs of resources exchanged within the value network, netted out of (3) to prevent double counting. Value added by the subtraction method. Primary source: the H-AIST research report (April 2026); identically stated in the CE Association white paper, pp. 2–4.
Two discretionary inputs, not one. Besides the CE ratio, I_CE (intermediate goods costs) “are not separately identified in financial statements and must therefore be estimated by reclassifying financial statement items, depending on company-specific accounting practices and data availability”. The worked example maps them to Japan’s SMEA survey categories and omits energy costs while noting they could belong there. ⟨Inference: the published ratio makes the reclassification invisible. An auditor checking a CVP figure is checking a mapping, not a number.⟩
Worked values, for calibration (dummy data, from the research report): a single waste-management company scores 0.64; a three-company recycling network scores 1.27 with intra-network purchase costs netted out and 0.49 without, against standalone values of 0.47, 0.50 and 0.49.
The inputs are accounting data, not material data. ⟨Correction: this page previously claimed CVP’s numerator requires material composition, origin, prior use and remaining lifetime, and built an argument about digital product passports on it. Eq. (2) refutes that. The only non-financial judgement CVP needs is the segmentation — which revenues and costs count as circular.⟩
Where the formula came from. The earliest public form, in the June 2025 overview, is
CVP = (Sales Revenue – Intermediate Input Costs) / Total CE Operating Costs
— structurally Eq. (2) already, with the denominator undivided and “including labour, logistics, depreciation”. The 2026 research report splits it into COGS + SG&A and names the subtraction method; it adds no new cost category. ⟨Inference: between the two documents the work went downward into financial-statement line items, not outward in scope. The 2025 paper also claims cross-sectoral benchmarking outright, which 2026 retracts into a precondition (“normalization is necessary”). Development here narrows claims rather than widening them — a good sign about the authors, and a reason to read the 2025 text as positioning rather than method.⟩
No investment term, in either generation. The 2025 paper faults CTI for incorporating neither “operating costs or investment needs”; Hitachi’s own summary below speaks of “the economic efficiency of circular investments”. Neither Eq. (2) nor its 2025 ancestor contains a capital term — investment enters only as depreciation inside operating cost. ⟨Inference: CVP is an operating-efficiency ratio wearing investment-appraisal language. A firm asking “should we fund this circular line?” gets told how efficiently the existing line converts period cost into value added. That is a real and useful thing to know and not the thing the marketing sentence promises.⟩
One real case, asserted ⟨2026-10-16⟩. Every published calculation uses
dummy data except one: at the lab’s
Third Open Forum (July 2026)
Murasato reported a case study of “dismantling, shredding and sorting
operators involved in a magnet-remanufacturing business”, where
“improvements in efficiency and output quality achieved through the
introduction of advanced sorting equipment were reflected in higher CVP
values … using real operational data”. No figure, no operator, no period is
given. ⟨Inference: this does not contradict the missing investment term —
an operating ratio can register the consequences of capital spending
(throughput, output quality) without measuring the spend. Hold both: the
denominator has no capital term, and the authors claim the effects of
capital show up anyway. The claim is asserted, not evidenced; nothing in
sources/ lets anyone check it.⟩
And a second output stream, for the record. CVP is one of two “forms of value visualization” in the lab’s rule-making theme. The other is residual-value grading, which operates on goods rather than accounts — and is the layer SINE’s work actually touches. A conversation about CVP alone is a conversation about the half that needs us least.
Hitachi’s own one-paragraph statement of purpose, from the Sustainability Report 2026, p. 36:
“Existing CE metrics have generally lacked a cost-efficiency perspective, limiting their ability to incentivize investment. Hitachi works to address this gap by standardizing a new CE metric called Circular Value-added Productivity (CVP). This metric captures the economic efficiency of circular investments by making improvements in investment efficiency visible.”
The diagnosis is correct, and evidenced
A circularity indicator that does not move capital does not move anything. The white paper does not leave this as an assertion: it surveys ~1,400 CE indicators from 34 types of resource material, condenses them to 45, maps them onto six assessment scopes × eleven assessment items, and finds no indicator in the “focus on cost and value added” cell except CVP itself — not from WBCSD’s CTI, not from ISO 59020, not from ISO 59040. “There is also a shortage of indicators from this perspective internationally.”
The authors are also explicit that CVP is complementary, not a replacement: in a remanufacturing case CVP tracked operating profit while WBCSD’s Circular Material Productivity did not, and combining indicators is “essential not only for comprehensive assessment but also for preventing greenwashing”.
⟨Inference: concede the diagnosis without argument. Inside an industrial group nothing gets funded that cannot be stated as value creation — see Hitachi’s Corporate Value Enhancement Formula, which organises the whole value-creation cycle “from an economic value perspective” (Integrated Report 2026, p. 4).⟩
How the network figure is built — a consolidation without a group
The research report names three candidate methods and adopts the third: each company computes its own CVP (rejected — “does not represent the CVP of the VN as a whole”); treat the VN as one company (rejected — internal circulation “cannot be evaluated properly”); or aggregate value added and expenditure network-wide while netting out intra-VN purchase costs (adopted, “regarding the VN as a single production system”). The authors state the model for the netting outright:
“This approach is based on the same concept as eliminating intra-group transactions when preparing consolidated financial statements.”
⟨Inference: CVP_VN is a consolidation without a group — consolidated income-statement arithmetic applied to firms sharing no parent, no accounting policy, no auditor and no obligation to one another. And the factor of 2.6 between 1.27 and 0.49 shows how much rides on the netting: get the elimination wrong and Method 3 collapses to the Method 1 the authors rejected. The metric’s entire value-network claim therefore rests on a cross-company reconciliation that no participant can perform alone, and the report does not say who performs it.⟩
Where it actually breaks
Not at the choice of boundary — the boundary is a network (see value networks and the ISO 59000 series) — but at what can be computed and believed across it. The study group’s own obstacle list, verbatim:
| Obstacle recorded | Remedies offered |
|---|---|
| ”The disclosure of data used in CVP calculation is difficult.” | internal use without disclosure; third-party certification of the final value only; year-on-year comparisons instead of levels |
| ”The disclosure of numerator and denominator data within VNs is difficult." | "Data management by financial and other institutions as third parties”; disclosure would help “maintain[] equity among the stakeholders that comprise the VN" |
| "The acquisition of data separating CE business and existing business is difficult.” | a “CE ratio”, “left up to the companies”; a “simple certification system and guidelines” |
The research report goes further than the table and names the mechanism:
“Calculating VN-level CVP requires determining value added and costs for each participating company, but it is not necessarily required to disclose these figures to all participants. A method in which a trusted, impartial third-party organization collects data from individual companies and calculates CVP can enable evaluation without public disclosure to the entire VN. … A third-party body can exist outside the VN, or it can exist within and become part of the VN’s formation.”
⟨Inference: every remedy is a trust arrangement — an auditor, a trustee, a coarsening, or self-declaration. Not one makes a figure checkable without disclosing it. A standard whose load-bearing inputs are self-segmented and whose cross-party aggregation requires a data trustee will be auditable against the reporter’s own books and unauditable against its counterparties. That is the confidentiality gap, stated by the proposers.⟩
Four limitations the authors state about their own indicator
From §4 of the research report, not from us: the CE scope “can easily become arbitrary — a result that may impair the comparability and explainability of CVP”; CVP “does not directly evaluate resource circularity rates or the reduction of environmental impact” and must not be used alone (greenwashing); financial structures differ enough by industry that absolute values should not be compared across companies; and in platform networks capturing all participants “is difficult in practice”, so the calculation boundary is drawn narrower than the network.
⟨Inference: this is a frank statement that CVP is a management-analysis instrument, not a disclosure instrument — its authors advise against cross-company comparison of levels and against standalone use. Putting it into an ISO annex or the WBCSD GCP invites exactly the disclosed, comparable use §4 warns against. That tension between the research and the standardization route is worth tracking; it is not yet a contradiction in the sources, because neither document addresses the other on this point.⟩
⟨Inference, and a limit on our own pitch: arbitrary CE segmentation is not a confidentiality problem, and neither a data trustee nor a multi-party protocol fixes it. A trusted collector computes an exact ratio from self-classified inputs. Signed data speaks to whether a figure is faithfully reported, not to whether the segmentation behind it was drawn honestly. SINE should keep the two apart; claiming the second would be overselling. See open tasks, item 18.⟩
⟨Correction: the earlier version of this page argued that “a metric designed to incentivise investment tends to be defined over data the investor already holds, so the system boundary gets drawn where the data stops — the company boundary.” The first clause holds; the conclusion does not. CVP_VN is defined over the network. The failure is not a small boundary, it is a boundary with no instrument for crossing it.⟩
The ratio form, and Tell Münzing’s thesis
The observation that Hitachi’s efficiency focus displaces systemic change still stands, and this source neither confirms nor refutes it — it narrows where it applies. Hitachi converts governed environmental quantities into productivity ratios, repeatedly:
| Quantity | Metric form | Stated reason |
|---|---|---|
| Circularity | CVP — value added per unit of circular expenditure | existing metrics fail to “incentivize investment” (SR 2026, p. 36) |
| Scope 3 Cat. 1 + 11 | reduction per unit of added value (−52 % by FY2030 vs FY2022) | absolute emissions “fluctuate greatly … difficult to see” (pp. 51–52, 64) |
| Cat. 11 near-term | GHG per unit of gross profit, −40 % by FY2027 | same |
⟨Inference: each is defensible on its own — ratios separate management action from portfolio noise. The aggregate is that no absolute quantity remains under a target. CVP is the weakest instance of the pattern, though, because unlike the Scope 3 intensities it is not a substitute for an absolute measure: it is a profitability test added alongside circularity measures its own authors insist must be used with it. The sharper version of the critique is therefore not “ratios instead of quantities” in general, but the specific one above: the ratio is cheap to compute inside a firm and impossible to verify between firms, so adoption will concentrate where verification is not required.⟩
What this means for SINE
Complementary, concretely. CVP answers is the circular activity economically viable? The open question it hands to someone else is can the parties in a value network compute Eq. (3) together without any of them revealing its sales, COGS and SG&A?
⟨Inference: that is a textbook secure-multi-party-computation problem, and unusually well-shaped for one — Eq. (3) is additively decomposable, with the single cross-party term C_D being an amount both sides of each intra-network transaction already know and can reconcile bilaterally. Confirmed by the primary source in the strongest available form: the authors specify a “trusted, impartial third-party organization” that collects the figures and computes the ratio “without public disclosure to the entire VN”, and leave its placement open. The role is specified; the instrument is not. A trusted collector who sees everything is the trust-maximal implementation of the computation a multi-party protocol performs trust-minimally — the authors wrote the interface and left the implementation open. This is a slot they drew, not a gap we argue them into. See cryptographic trust infrastructure.⟩
A second, smaller opening: the segmentation. The CE ratio is self-declared. Certification of a declaration is not verification of it. ⟨Inference: anything that binds the circular-revenue claim to evidence a counterparty can check — the Man & Kleppmann move of coupling the claim to the financial record so that misreporting it requires misreporting the accounts — applies here almost directly, and CVP is already financially anchored, which is the precondition that architecture needs. See Emission Impossible.⟩
Enter as a contributor. Hitachi and AIST have an apparatus — a standardization committee under the CTO, a Rule-making Promotion Group, an IEC SMB seat, a 40-member association organised around ISO/TC 323, and ministries in the room. SINE cannot build that and should not flank it. The live work item is the ISO 59011 annex and TC 323 WG3; the live gap is the one the study group wrote down itself.
Open questions worth putting to them, none rhetorical:
- In Eq. (3), who computes the sum, and what do the participants have to reveal to that party? Is the “trusted, impartial third-party organization” of §3.3 meant to be an existing institution (a bank, an industry body) or a new one — and inside or outside the network?
- How is ΣC_D reconciled? Each intra-network cost is a sale on the other side; matching them is where a protocol would act.
- Is third-party certification of a CVP value intended to cover its inputs, or only its arithmetic?
- The H-AIST proposal has a second half — “data model standards for information on goods including quality information” — that this white paper never describes. Where is it, and does it carry provenance?
Cross-references
- H-AIST CVP research report (April 2026) — the primary definition
- Driving Circularity Transitions (H-AIST, June 2025) — the earliest public statement of the metric
- Third Open Forum report (July 2026) — the magnet-remanufacturing case study
- Residual-value grading — the other half of the same proposal
- CE Association white paper (2026) — the implementation document
- Hitachi Sustainability Report 2026 — the one-paragraph version
- Value networks and the ISO 59000 series — the boundary question
- Hitachi-AIST CECRL · Circular Economy Association · Hitachi, Ltd.
- Confidentiality gap in carbon accounting
- Coordination failure vs. bargaining failure
- Cryptographic trust infrastructure (CCTI stack)
- WBCSD — CTI, CMP and the GCP
- Verifiable carbon accounting · Decision-grade carbon data criteria
Sources
sources/hitachi/2026-04-13-h-aist-cvp-research-report.pdf(definition, worked calculations, Methods 1–3, §3.3 third party, §4 limitations)sources/hitachi/2025-06-19-h-aist-driving-circularity-transitions-cvp.pdf, §§3, 5–6 (first published form of the formula, cross-sectoral benchmarking claim)sources/hitachi/2026-07-30-h-aist-ce-lab-third-open-forum.md(Murasato’s magnet-remanufacturing case study; CVP as one of two value-visualization forms)sources/standards/2026-10-16-ce-association-social-implementation-cvp.pdf, printed pp. 1–8, 12 ⟨page numbers corrected 2026-10-29; printed = PDF − 3⟩sources/standards/2026-10-16-ce-association-social-implementation-cvp/tables-recovered-from-pdf.mdsources/hitachi/2026-10-15-hitachi-sustainability-report-2026.pdf, pp. 36, 51–52, 56, 64sources/hitachi/2026-10-15-hitachi-integrated-annual-report-2026.md, p. 4