A five-stage ladder for reading where a circular value chain actually stands,
what binds it at that point, and which financing pattern can see it. The model
is SINE-authored ⟨Inference, 2026-11-06⟩, built at Tell Münzing’s framing —
“Die Stimmen in den Reports zeigen die unterschiedlichen Bewertungsmuster bzw.
Reife des Systems wie die Stufen und Akteure ineinandergreifen. Da entsteht ein
temporales to-do set” (Slack, 2026-11-06) — out of the three POLYPROBLEM
reports read as one set:
06 Der Circularity Code,
07 Die Kluft im Kopf (sources/roechling/2025-07-01-polyproblem-07-die-kluft-im-kopf.md,
no page yet) and
08 Wechsel-Geld.
Each stage is anchored in quoted voices; the ladder itself is the inference. The point of the model is not classification. It is that the three reports disagree about the first step because each speaker is standing on a different rung, and that the financing gap everybody describes turns out to be a mismatch between stages and valuation patterns rather than a shortage of capital.
The five stages
| Stage | Binding constraint | What gets valued | Representative voices |
|---|---|---|---|
| S0 Unvalued | follow-on costs sit outside every balance sheet | nothing — the circular asset is invisible | 08 p. 4; Rembe (07, p. 27); 06’s cost-bearer ≠ value-captor |
| S1 Single-actor, isolated | no counterparty on the other side of the loop | project NPV, and it fails | Betancourt Munoz (08, p. 34); GIZ (06); Richter (07, p. 22); PreZero (08, p. 23) |
| S2 Bilaterally underwritten | the guarantee is a one-off, not an institution | the guarantor’s covenant, not the model | Traceless / Land Hamburg (08, p. 15); Stremlau’s first-loss tranches (08, p. 9); offtake guarantees (06, p. 13) |
| S3 Multilaterally synchronised | who finances the shared layer and how the value is split | the chain, not the actor | 08 p. 23; Maier (07, p. 15); DIN SPEC 91446 and the DPP (06); Betancourt Munoz (08, p. 35) |
| S4a Disclosure demanded | a regime asks for numbers the layer below cannot supply | compliance, not value | PPWR, EPR and recyclate quotas (08, p. 23); Pöppelmann’s “erfolgreich gescheitert” (08) |
| S4b Bookable | none specific — it prices like any other asset class | cashflow and collateral, conventionally | Gruber (08, p. 33); Stremlau’s negative list (08, pp. 9–10); Griestop (07, pp. 25–26); Seo (07, p. 28) |
S0 — unvalued
The defect all three reports state and none attributes to the others. 08, p. 4: the linear model pays “nur, weil ihre Folgekosten nicht in den Bilanzen auftauchen und weitgehend von der Allgemeinheit getragen werden”. 07, p. 27, Gina Rembe, from the municipal side: “Die Kosten tragen Städte und Kommunen bzw. wir als Steuerzahlerinnen und Steuerzahler. Das muss sich ändern. Einweg muss stärker eingepreist werden und von den finanziell profitierenden Unternehmen getragen werden.” 06 states it in the data layer: those who bear the measurement cost are not those who capture the resulting value (see ecosystem financing § cost-bearer ≠ value-captor).
S1 — single-actor, isolated
The stage most pilots never leave. Betancourt Munoz (IDB Invest), 08, p. 34: “Man findet Projekte, die den Kreislauf intern schließen. Aber dann haben sie keine Kunden für ihr Produkt.” The same shape in each report: 06’s GIZ finding of more than twenty separate collection apps in Indonesia alone (giz.md); 07’s Unverpackt sector falling from roughly 500 to ~230 shops with Katharina Richter’s “das wirtschaftliche Umfeld ist ziemlich feindselig” (p. 22), and REWE’s eleven-store packaging-free pilot extended from six to twelve months and still barely used; 08’s PreZero at −€386m in 2024 despite owning the whole loop.
S2 — bilaterally underwritten
Something does move here, but only because a third party absorbs the part nobody can value. Traceless obtained a Hamburg state guarantee; the report’s own comment, via Johanna Thoma: “Wenn das eine systemische Lücke ist, dann braucht es vielleicht auch eine systemische Lösung” (08, p. 15). Stremlau’s blended-finance instruments — guarantees, Bürgschaften, first-loss tranches (p. 9) — and 06’s long-term industry offtake guarantees (p. 13) are the same move. ⟨Inference: S2 is where the published success stories live, which is why the sector reads as further along than it is. A one-off guarantee is evidence that the valuation problem was contracted around, not solved.⟩
S3 — multilaterally synchronised
The stage at which value is created by the chain rather than by a firm. 08’s Schwalbe case (p. 22): specialist retail, logistics, a technology partner and in-house R&D are all required, so “Der Wert entsteht somit nicht in einem einzelnen Unternehmen. Er entsteht in der Verbindung verschiedener Akteure.” Its own open question, same page: “wie eine gemeinsame Infrastruktur finanziert und der daraus entstehende Wert verteilt wird”. 07 states the identical requirement from the demand side — Judith Maier (REWE), p. 15: “Insellösungen behindern eine breite Mehrweg-Akzeptanz”, with standardised cross-sector containers as the remedy and “Niemand kann dieses Ziel allein erreichen”. 06 supplies the instruments — quality grading under DIN SPEC 91446, the digital product passport — and the condition that they are worthless if only one stage of the chain supplies data.
S4 — bookable
⟨Revision 2026-11-06: the stress test below splits this rung into S4a (a disclosure or quota regime demands numbers) and S4b (the numbers can be priced). The circular reports conflate them; the biodiversity case separates them cleanly. The text below describes S4b.⟩
The end state, described in all three reports only by analogy to renewables. Gruber (DKB), 08, p. 33: “Durch die Einspeisevergütung hat der Gesetzgeber das Risiko beim Cashflow rausgenommen, und damit war das Thema für die Banken leicht zu beherrschen.” 07 names the demand-side equivalent — Griestop (WWF), pp. 25–26: “Anstatt Mehrweg günstiger anzubieten, würde ein deutlich teureres Einweg-Angebot eher zu einer Verhaltensänderung führen” — and Seo (UBA) the level playing field (p. 28). Stremlau’s inversion belongs here too: she would “eher definiert …, was nicht nachhaltig ist, und das regulatorisch stärker gewichtet” (08, pp. 9–10).
The stages interlock; they do not sequence
⟨Inference — the core claim of this page. The ladder is not a path an individual firm climbs. In one chain, actors occupy different stages at the same time, and the chain performs at the level of its least mature link. Schwalbe sits at S2–S3; its collection, logistics and reprocessing partners sit at S1; the recyclate market sits at S0 — which is why a technically working loop still reads as a cost centre. This is the mechanism behind 08’s own closing correction (p. 23) that the transition must be financed “synchronisiert … unternehmensübergreifend und entlang der Wertschöpfungskette”, and behind the behavioural measurement in 07: Lukas Schuck reports the same intervention producing between 0 % and 80 % effect across stores of the same chain (pp. 26–27). Identical instrument, different local maturity, two orders of magnitude in outcome.⟩
The corollary matters for sequencing advice: an intervention aimed at a stage above where the weakest link stands is wasted. 07 proves this against itself — nudges at 1–2 %, reusables at 1.6 % of the market in 2023, single-use rising from 13.6bn (2022) to 14.6bn (2023) while reusables rose only from 101m to 232m, which the report calls “homöopathische Wirkung” (p. 25). Behaviour work is an S3/S4 instrument applied to an S1 system. Nathalie Bujold states the limit directly (pp. 12–13): “letztlich stehen uns nur die Optionen zur Verfügung, die das bestehende System erlaubt.”
Valuation patterns map to stages — the gap is a mismatch, not a shortage
⟨Inference: each class of financier can only underwrite at particular stages, because its valuation pattern presupposes them.⟩
| Financier class | Reads value as | Can underwrite |
|---|---|---|
| Venture capital | growth on an asset-light model (08, p. 28: not “das Hockey-Stick-Wachstum”) | S1 |
| Impact funds | return threshold — Anthropia’s €10m vehicle failed at ~12 % | S1–S2 |
| Public guarantor / blended finance | the guarantee, not the model (Stremlau, p. 9; 90-10 rule, p. 10) | S2 |
| Multilateral development banks | development impact plus “eine andere Art von Risiko” (Betancourt Munoz, p. 34) | S2–S3 |
| Commercial banks | cashflow and physical collateral (Gruber, p. 32) | S4 only |
| Proprietary circular desks | own methodology — Rabobank €270m; Intesa Sanpaolo €11.7bn in 2023 on a CE methodology built with the Ellen MacArthur Foundation (08, p. 6) | S2–S3, privately |
No pattern covers the S2→S3 transition in public. That is the precise location of the valley of death described in 08, p. 27 — “zu risikoreich für klassische Kreditgeber, zu kapitalintensiv für Venture Capital und zugleich noch nicht reif genug für großvolumige institutionelle Finanzierung” — and the reason only 13 % of surveyed circular firms use bank credit against 46 % of European SMEs (08, p. 6).
⟨Inference: this reconciles the report’s unreconciled voices. Hafenmayer (“Nicht Nachhaltigkeit macht ein Unternehmen investierbar, sondern ein funktionierendes Geschäftsmodell”, pp. 30–31) and Betancourt Munoz (the obstacle is “nicht unbedingt … ein Mangel an Kapital oder Investoreninteresse”) are right that capital is not scarce; Nusseck is right that nothing flows. Both hold if the defect is that no valuation pattern applies at the stage where the chains currently sit. It also explains the Rabobank/Intesa anomaly recorded on the valuation gap page: those two banks did not wait for a public method, they built a private one — so the S2–S3 method demonstrably exists and is not shared. That is a disclosure problem, which is 06’s subject, and 08 does not recognise it as such.⟩
The temporal to-do set
⟨Inference: derived from the stage transitions, with SINE coverage marked honestly — including where it is absent.⟩
| # | Transition | Task | SINE capability |
|---|---|---|---|
| T1 | S0→S1 | make follow-on costs measurable per chain segment — measurement, not yet pricing | partial: method yes, pricing is politics |
| T2 | S1→S2 | turn one-off guarantees into a repeatable instrument; the guarantor must be able to verify without the firm disclosing | yes — governance design plus confidentiality-preserving verification |
| T3 | S2→S3 | the anchor-tenant question: who operates the shared layer, on what revenue — 08’s open question verbatim | core — see bootstrapping problem |
| T4 | S3 (parallel) | claims made credible without disclosure: grading (DIN SPEC 91446), DPP, residual-value grading, and the audit of data quality that no report addresses | core |
| T5 | S3→S4 | ex-ante value-distribution rule for the shared loop | core, conditional on competition-law clearance — unanswered in all three reports |
| T6 | S4 | price regime: negative-list taxonomy, single-use pricing, level playing field | no — SINE supplies the evidence base, not the rule |
| T7 | ongoing, lagging | defaults instead of campaigns | no — but the model says when a behaviour budget is premature |
The capability window
⟨Inference: SINE’s governance and coordination offer is not stage-neutral. At S0–S1 there is nothing to coordinate; at S4 the market coordinates itself. The offer is worth most across S2→S3, which is exactly where the three POLYPROBLEM reports, read together, place the German plastics and packaging system. This makes the timing argument sharper than the capability argument: the question to put to a partner is not whether coordination helps, but whether they are at the rung where it is the binding constraint.⟩
Stress test: the biodiversity analogue ⟨2026-11-06⟩
Requested as a durability check ⟨Tell Münzing, Slack, 2026-11-06⟩: does the ladder hold in a domain with different actors, a different substance and different institutions? Run against the wiki’s biodiversity material — essential variables, nature-related reporting frameworks, GEO BON, Kim et al. (2026).
The mapping
| Stage | Biodiversity instance | Evidence |
|---|---|---|
| S0 | nature’s follow-on costs outside every balance sheet; the Aichi Targets missed by 2020, attributed in part to “the dearth of reporting measures” | Kim et al. §1, via SCBD 2020, Xu et al. 2021 |
| S1 | voluntary pilots, no counterparty: EuropaBON, FrenchBON, BHI instances in San Martín, the Pilbara and Korea; EESVs have no formal variable list at all | essential variables § Status |
| S2 | no instance found — see below | — |
| S3 | the essential-variable layer itself: 6 EBV classes / 22 subclasses, 84 prioritised specific variables from EuropaBON, two published crosswalks, BON in a Box, GBiOS; SEEA EA endorsed by the UN Statistical Commission as a global statistical standard | essential variables; nature-related reporting frameworks |
| S4a | TNFD core disclosure metrics, NPI state-of-nature metrics, NGFS central-bank scenario frameworks — all arriving before the layer that would make their numbers comparable | nature-related reporting frameworks § corporate and financial side |
| S4b | not reached | — |
What holds
The stage definitions transfer intact. S0 and S1 have the same signature in both domains, down to the fragmentation measurement: twenty-plus collection apps in Indonesia (06) and a proliferation of national and thematic BONs with no formal variable list are the same pathology in different substances.
The interlocking claim survives its strongest test. Biodiversity holds S3-grade structure (crosswalks, an endorsed statistical standard) on top of S1-grade supply (voluntary pilots) with S4a-grade demand (central banks) above both. The model predicts that the chain performs at its least mature link — and the wiki recorded exactly that outcome independently, before this page existed: a disclosure regime arriving ahead of the data layer that would make its numbers believable, “where carbon was before PACT”.
The valuation-pattern mismatch sharpens. NGFS is attempting to underwrite nature-related financial risk at S4 against a supply sitting at S1. That is the same mismatch as a commercial bank being asked to lend against an S2 circular chain, two rungs further apart.
What breaks — two amendments the test forces
1. The ladder is not monotone in demand, so S4 must split. In the circular case the demand regime and bookability look like one rung, because neither has arrived. Biodiversity separates them: TNFD, NPI and NGFS demand numbers (S4a) while nothing is priced (S4b). ⟨Inference: this retro-illuminates the circular case. PPWR, recyclate quotas and EPR are S4a instruments too, and all three reports treat them as though they were S4b — as investment signals rather than as compliance demands landing on an S1 supply. Pöppelmann’s “Wir sind erfolgreich gescheitert” is precisely the damage of S4a without S4b: the early mover meets the quota and competes against latecomers who waited more cheaply. The model did not say this before the stress test.⟩
2. S2 exists only where there is something to bargain over. Bilateral underwriting needs a party with a private return a guarantor can secure. The state-side biodiversity actors are states reporting to a convention and national statistical offices — on the wiki’s own reading this is a near-pure coordination failure, with commercial confidentiality never named and cryptography never reached for (coordination vs. bargaining failure). There is no rent to divide, so there is nothing to underwrite, and the domain moves S1 → S3 directly. ⟨Inference: the presence or absence of an S2 rung is therefore diagnostic of the failure shape, not a measure of progress. A chain stuck at S2 has a bargaining component; a chain that skipped it does not.⟩
What this does to the capability window
Combining the two amendments changes the claim rather than weakening it. Biodiversity’s shared layer is being built without SINE’s instrument set and does not need it — the operator exists (GEO BON), the funding is public appropriation (the European Commission’s commitment to regional EBV coordination after EuropaBON), and no participant has anything to lose by disclosing. ⟨Inference: so the window is not a stage. It is a stage plus a bargaining component: SINE binds where a chain sits at S2→S3 and the reporting entity is exposed by disclosure. In biodiversity that condition is met in exactly one place — the corporate boundary, where TNFD makes a firm the reporting entity — which this wiki already flagged as the sharpest unclaimed space in the domain and marked unverified. The stress test and the earlier open task converge on the same unread document.⟩
A third finding, pointing the other way and worth carrying to the camp: the biodiversity S3 layer is financed because it was claimed as a public statistical function — SEEA EA sits under the UN Statistical Commission and is explicitly the ecosystem sibling of the System of National Accounts. The circular S3 layer has never been claimed that way; it is expected to find a private business model, which is why its financing question stays open. ⟨Inference: “who pays for the shared layer” may be the wrong question. The biodiversity precedent suggests the prior question is whether the layer is a market good at all — and the one circular instrument in the wiki that got funded privately (cirplus financing DIN SPEC 91446, because its business model is transparency) is the exception that shows how narrow the private route is.⟩
Open questions
- The biodiversity stress test below is run against one source community (Kim et al. 2026, written by the people who would operate the layer they propose) plus a stub page on GEO BON. No TNFD, NPI or NGFS primary source has been read, and the wiki holds no source at all on biodiversity credits, offsets or debt-for-nature swaps. The missing S2 rung is therefore partly an absence of evidence; it is argued structurally below, not demonstrated.
- Is there any documented instance of an actor moving S2→S3 without public money? None appears in the three reports.
- Does the stage-mismatch account predict anything the capital-shortage account does not? Proposed test: a chain whose weakest link is lifted to S3 should attract conventional credit without additional subsidy.
Cross-references
- Ecosystem financing — the S3 financing question in its own terms, with the instrument inventory.
- Circularity valuation gap — why S4 is not reachable by information alone; Hafenmayer’s counter-position.
- Bootstrapping problem — T3, the anchor-tenant test and the ABtC pattern.
- Coordination vs. bargaining failure — the diagnostic that places a chain at S2 rather than S3.
- Data-space business models — the same “who operates and pays for the shared layer” question in the data layer.
- Digital product passport, residual value grading — T4 instruments.
- Circularity definition problem — why the renewables analogy for S4 needs a measured unit first.
- Ostrom commons governance — T5, the distribution rule.
Sources
sources/roechling/2026-08-01-polyproblem-08-wechsel-geld.md— p. 4 (externalised follow-on costs), p. 6 (PREVENT quote; Rabobank, Intesa; 13 % vs 46 %), pp. 9–10 (Stremlau), p. 15 (Traceless, Thoma), pp. 22–23 (Schwalbe, PreZero, synchronisation, the two open questions), p. 24 (Stammer, TecPart), pp. 27–28 (valley of death, Nusseck, Anthropia), pp. 30–31 (Hafenmayer), pp. 32–33 (Gruber), pp. 34–35 (Betancourt Munoz). Summary: Wechsel-Geld.sources/roechling/2025-07-01-polyproblem-07-die-kluft-im-kopf.md— p. 7 and p. 10 (Jackson), pp. 12–13 (Bujold), p. 15 (Maier, REWE), pp. 20–21 (Wermter), p. 22 (Richter), p. 25 (reuse figures, Rembe), pp. 25–26 (Griestop), pp. 26–27 (Schuck), p. 28 (Bethmann, Seo), pp. 32–34 (Miserocchi). No wiki page yet — ingest pending.sources/roechling/2023-12-01-polyproblem-06-circularity-code.md— p. 12 (“Die Technik ist vorhanden, das Geschäftsmodell nicht”), p. 13 (Helftewes, offtake guarantees; Schiller, DIN SPEC 91446), p. 25 (single-use levy), p. 53 (who pays for the infrastructure). Summary: Der Circularity Code.