The primary source for CVP: twelve pages published by the Hitachi-AIST Circular Economy Cooperative Research Laboratory in April 2026. Everything the wiki previously said about CVP came from secondary statements of it — one paragraph in the Sustainability Report and the CE Association white paper, which is an implementation proposal rather than a definition.
Supplied as an attachment by Tell Münzing, 2026-10-16, closing the fetch
half of open task 15. Source: sources/hitachi/2026-04-13-h-aist-cvp-research-report.pdf
and its Markdown twin.
No correction forced. Eqs. (1)–(3) match exactly what had been recorded from the white paper. What the report adds is the arithmetic, the method choice behind Eq. (3), and — the substantive part — four limitations the authors state about their own indicator.
1. The worked single-company calculation
A waste-management company with resource recovery, dummy figures (M$): sales 5; purchased goods and raw materials 1.5; outsourcing 0.5; freight and packing 0.1; plus labour, rent, depreciation, advertising, entertainment, taxes.
5 − (1.5 + 0.5 + 0.1)
CVP = ───────────────────────────────────────────────────────────────── ≒ 0.64
(1.5+0.5+0.1+0.5+0.25) + (1+0.1+0.1+0.1+0.05+0.25+0.05)
(numerator: sales less intermediate goods costs; denominator: COGS plus SG&A, as itemised in the report’s Table 1.)
The input that is not in the accounts. 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.” For the trial, the authors align definitions with Japan’s SMEA 2023 Basic Survey of Small and Medium Enterprises and count purchased goods/raw materials, outsourcing, and freight and packing as intermediate — while noting that energy costs “can also” belong there and are omitted here.
⟨Inference: so CVP has two discretionary inputs, not one. The CE ratio decides which share of the business is circular; the I_CE reclassification decides where the line between value added and intermediate consumption falls. Both are reconstructions by the reporting company, and the second is invisible in the published ratio. An auditor checking a CVP figure is checking a mapping, not a number.⟩
2. Value networks: three methods, and why Method 3
The report names and rejects two obvious approaches before adopting the third:
| Approach | Why rejected / adopted | |
|---|---|---|
| Method 1 | each company computes its own CVP; average or weighted-average them | ”remains, in essence, a set of company-level CVP values and therefore does not represent the CVP of the VN as a whole” |
| Method 2 | treat the VN as a single company | resources circulating inside the boundary “cannot be handled appropriately” and the internal circulation structure “cannot be evaluated properly” |
| Method 3 | aggregate value added and costs across the network, netting out intra-VN purchase costs | adopted — “from the perspective of regarding the VN as a single production system” |
The stated model for the netting is explicit and is the most useful single sentence in the report:
“This approach is based on the same concept as eliminating intra-group transactions when preparing consolidated financial statements.”
(The cited authority is an EY Japan commentary on consolidation, not a standard.)
⟨Inference: CVP_VN is therefore a consolidation without a group — the arithmetic of a consolidated income statement applied to firms that share no parent, no auditor and no obligation to each other. Everything that makes consolidation tractable inside a group is exactly what is absent between independent firms: a single accounting policy, a controlling entity entitled to the underlying books, and a statutory audit over the result. This is the cleanest way to state what the metric needs and does not have.⟩
3. The number that shows how much rides on the netting
Trial for a three-company recycling VN (collection → recycling → material), dummy data:
- Value added: ΣS_CE − ΣI_CE = 14.0 − 8.4 = 5.6 M$
- Expenditure: Σ(C_CE − C_D) = 0.7 + 1.2 + 2.5 = 4.4 M$
- CVP_VN = 5.6 / 4.4 = 1.27
Without eliminating the intra-VN purchase costs the denominator is 11.4 M$ and CVP_VN falls to 0.49. The three companies’ standalone CVPs are 0.47, 0.50 and 0.49 — which the authors read as benefits being “distributed relatively evenly”.
⟨Inference: the entire claim of the value-network view is that factor of 2.6. If the elimination is wrong, CVP_VN collapses to roughly the mean of the standalone values and Method 3 buys nothing over the Method 1 the authors rejected. The metric’s whole value therefore rests on a cross-company reconciliation — matching each intra-VN sale against the corresponding purchase — that no participant can perform alone and that the report does not say who performs. That is a narrower and far more tractable problem than “share your financials”: C_D is the one quantity both sides of each transaction already hold, and they need to agree on it rather than reveal everything around it.⟩
Also recorded: external resources bought by the first company in the chain “can, in principle, be included in the denominator”, but were excluded in the trial because including them “could make Company A’s cost tend to appear relatively large when evaluating the distribution of costs and value added”. An admitted ad-hoc choice, “to be examined further using real data”.
4. The third-party collector — stated as the mechanism
§3.3, and this is the passage that matters most for SINE:
“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. Such efforts are expected to help prevent large imbalances in benefits and burdens within the VN. A third-party body can exist outside the VN, or it can exist within and become part of the VN’s formation.”
In the white paper the data trustee was one of several remedies in an obstacle table. Here it is the recommended mechanism, with its placement (“outside the VN, or … part of the VN’s formation”) treated as a design choice.
⟨Inference: the role is specified and the instrument is not. A trusted collector who sees everything is the trust-maximal implementation of exactly the computation a multi-party protocol performs trust-minimally; the authors have written the interface and left the implementation open. This is the most direct opening the Hitachi material has produced so far — not a gap we argue them into, a slot they drew. See confidentiality gap and cryptographic trust infrastructure.⟩
5. Five value-network topologies (Fig. 1)
| Type | Shape |
|---|---|
| 1 | resource form unchanged along the chain — e.g. iron scrap collected, sorted, refined by A, B, C |
| 2 | the focus resource is closed inside the boundary (per ISO 59010, possibly only part of the loop) |
| 3 | the resource changes form inside the VN — e.g. combustible residue converted to energy, used to process another resource |
| 4 | the resource is combined with an external resource and converted |
| 5 | platform VN — A and B operate platform P with an unspecified number of members C₁…Cₓ; between each member and P only information and money flow, while goods circulate directly among the members |
In practice these combine. ISO/TR 59032:2024 (Review of existing value networks) is cited alongside ISO 59010 as the source of the typology — a standard not previously in the wiki.
⟨Inference: Type 5 is the topology SINE’s work sits in — a coordinating layer that sees information and money but never the goods. It is also the one the authors admit CVP handles worst (see below).⟩
6. Four limitations the authors state themselves
This section is the report’s most valuable content, because it is a self-assessment rather than a critique from outside.
- Arbitrary CE scope (§4.1.1). Which activities count as CE “has a major influence on the results”; for multi-business companies the definition “can easily become arbitrary — a result that may impair the comparability and explainability of CVP”. Remedy offered: define it in advance and apply it consistently over time.
- Greenwashing (§4.1.2). CVP “does not directly evaluate resource circularity rates or the reduction of environmental impact”, so judging CE progress on CVP alone “may not adequately reflect the actual state of circularity”. Must be combined with environmental indicators.
- Industry structure (§4.1.3). Financial structures differ enough by industry that absolute CVP values should not be compared across companies; use industry averages, own history, or normalised metrics.
- Resource transformation and open participation (§4.2). If externally introduced resources come to dominate, the VN’s CVP “will approach the CVP of that company” and stops describing the network. In platform VNs (Type 5) capturing all participants “is difficult in practice”; the proposed workaround is to draw the calculation boundary narrower than the VN and exclude the unspecified participants — with the boundary and method disclosed.
⟨Inference: taken together these are a frank statement that CVP is a management-analysis instrument, not a disclosure instrument. Its own authors advise against cross-company comparison of levels, against standalone use, and against reading a platform VN’s figure as covering the platform. Any later use of CVP as a disclosed, comparable number — which is what putting it into an ISO annex or the WBCSD GCP would invite — contradicts this section. Worth tracking as the gap between what the research says and what the standardization route implies.⟩
⟨Inference, separate and important: limitation 1 is not a confidentiality problem and no trustee or cryptographic protocol fixes it. A trusted collector computes an exact ratio from self-classified inputs. Signed data addresses whether a figure is faithfully reported; it does not address whether the segmentation behind it was drawn honestly. SINE should keep these two apart in its own argument — claiming the second would be overselling.⟩
What it does not contain
No mention of digital product passports, product passports, declarations, or ISO 59040 / PCDS — checked across the full text. Nor any reference to the other half of the H-AIST proposal (the data model for goods information). This report is about the CVP half only. See open tasks, item 17.
Cross-references
- Circular Value-added Productivity (CVP) — the concept page this source defines
- Value networks and the ISO 59000 series — the typology and the standards cited
- CE Association white paper (2026) — the implementation proposal built on this report
- Hitachi-AIST CECRL — the authors
- Confidentiality gap in carbon accounting · Coordination failure vs. bargaining failure
- Cryptographic trust infrastructure (CCTI stack)
- WBCSD — CMP, cited here as an indicator within CTI v4.0
Sources
sources/hitachi/2026-04-13-h-aist-cvp-research-report.pdfsources/hitachi/2026-04-13-h-aist-cvp-research-report/2026-04-13-h-aist-cvp-research-report.md(Markdown twin). ⟨Note: Figs. 1–3 are images in both PDF and twin and were not read as images; their content here is taken from the report’s own descriptions of them in §§3.1.1, 3.1.2 and 3.2, which are complete enough to stand alone. Table 2 survives the conversion only in garbled form — the per-company figures quoted above are the ones the report states in its Step 1–3 arithmetic, not read off the table.⟩