Skip to content

ADR-0018 — SFDR / EU-Taxonomy signal: assessed and declined for Phase 1; not built

Field Value
Status Accepted
Date 2026-05-23
Supersedes —
Superseded by —
Related ADR-0017 (SBTi source-taxonomy drift — moving-vocabulary precedent); quarterly source-instrument continuity review

Context

Cycle 15d ingested ORIT (INST-0076, Octopus Renewables Infrastructure Trust, sector 55 Renewable Electricity). ORIT is an Article 9 SFDR fund that publishes an EU Taxonomy alignment percentage in its annual reports. During the Phase-1 close pass (cycle 17) this surfaced as a rule gap: no SFDR or EU-Taxonomy scoring rule exists in the catalogue.

The gap renders correctly — no phantom or broken rule appears on ORIT's institution detail page; the absent rule simply produces no signal. Criterion (a) of the Phase-1 close pass was declared MET with this gap documented pending a scoping decision. This ADR is that decision.

The question is whether to build a scoring rule (or rule set) that credits SFDR product classification (Art. 6 / 8 / 9) or EU Taxonomy alignment-% as a positive signal for relevant institutions.

Decision

Do not build an SFDR or EU-Taxonomy scoring rule.

ORIT continues to render with no applicable SFDR / EU-Taxonomy rule. The criterion (a) gap is closed as assessed and declined, not deferred. No scraper, no migration, no engine change.

Rationale

Three independent grounds, each sufficient on its own:

1. n=1 and structurally so. SFDR product classification (Art. 6 / 8 / 9) applies to funds and fund-management products (UCITS, AIFs, etc.). ORIT is the only fund-type entity in the 76-institution universe; every other institution is a corporate. For financial corporates the analogous Taxonomy metric is the Green Asset Ratio (GAR); for non-financial corporates it is Taxonomy-aligned turnover, capex, and opex. A rule built on ORIT's Art. 9 signal does not generalise to the remainder of the universe without creating three structurally distinct sub-rules that each apply to a different (and currently empty or near-empty) cohort. The engineering cost is real; the incremental coverage is zero beyond the one institution that prompted it.

2. Self-declared signal in tension with neutral-prior. SFDR and the EU Taxonomy Regulation are disclosure regimes, not investment standards. There is no requirement to invest in taxonomy-aligned activities; the Art. 9 label is a manager's product-level declaration, and the taxonomy-alignment percentage is the manager's own portfolio computation, calculated under the manager's chosen methodology. Crediting either as a positive ESG signal sits in direct tension with ADR-0012's neutral-prior principle, which is specifically designed not to reward unverified self-description. The substantive signal — the alignment-% — is more meaningful than the label, but the same tension applies: it is unaudited self-reporting with no third-party verification layer comparable to SBTi target validation or PRB signatory accountability.

3. Mid-transition vocabulary. The European Commission's SFDR 2.0 proposal (November 2025) replaces the Art. 6 / 8 / 9 framework with three new categories (Sustainable / Transition / ESG Collection), each on a 70% portfolio threshold. The final framework is not expected to apply before 2028. Encoding Art. 8 / 9 vocabulary now means onboarding a known-moving target for one entity — the same taxonomy-drift class of problem documented in ADR-0017 (SBTi "Commitment Removed") and the quarterly source-continuity review. When the vocabulary changes, any rule built against it becomes a maintenance liability.

Clarification — label vs. alignment percentage: SFDR classification (a product label: Art. 6 / 8 / 9) is distinct from EU Taxonomy alignment (a 0–100% activity-based metric). The substantive ESG signal is the alignment-%; the label alone is descriptive of the product wrapper, not the investment activity. Even granting that the alignment-% is the more meaningful input, all three grounds above still hold.

Consequences

  • Criterion (a) of the Phase-1 close pass is closed as assessed-and-declined. It is no longer an open gap or a Phase-1.5 carry item.
  • ORIT's Art. 9 classification and EU Taxonomy alignment-% are uncredited by the scoring engine. ORIT's ESG score reflects the NF rules that do apply (NF-G4, NF-S3, NF-S4, UN-E1, CHRB not_found, BHRRC clean, etc.).
  • If the universe later gains a material cohort of EU-fund / Art. 9 entities, this ADR should be revisited as a fresh scoping decision against the SFDR 2.0 category framework then in force. Do not attempt to reverse-engineer a rule from ORIT's current Art. 9 data.
  • The quarterly source-instrument continuity review should note SFDR 2.0's expected application date (2028) as a standing watch item.

Alternatives considered

Build a narrow ORIT-specific rule — rejected. A rule that applies to a single institution is not a rule; it is a one-off adjustment. It would also require a new applicable_sectors pattern not yet needed anywhere else and would set a precedent for bespoke rules on n=1 entities.

Build a generalised EU Taxonomy alignment-% rule for all institutions — rejected. No other institution in the 76-institution universe publishes a comparable metric in a consistent, machine-readable form. The data coverage is effectively zero beyond ORIT.

Defer to Phase 1.5 pending SFDR 2.0 finalisation — rejected. Given the 2028 application timeline, deferral without a declining decision merely leaves the item open indefinitely. Assessed-and-declined is cleaner than an open carry with no expected resolution date.

References

  • Cycle 15d scoring run (run_id=39) — SFDR/EU-Taxonomy gap first confirmed; ORIT detail page renders cleanly with no phantom rule
  • Cycle 17 Phase-1 close pass — criterion (a) declared MET with gap documented
  • docs/design/orit-physical-asset-infrastructure-carve-out.md — ORIT methodology note
  • European Commission SFDR 2.0 proposal (November 2025)
  • ADR-0017 — SBTi source-taxonomy drift (moving-vocabulary precedent)
  • ADR-0012 — neutral-prior scoring methodology (self-declared signal tension)
  • Quarterly source-instrument continuity review (parked queue, backlog)