Extinction rate per million species-years (E/MSY) exceeds background by 1000-10000x

Five independent claude-sonnet-4-6 calls, one per persona — Skeptic, Fact-Checker, Researcher, Compliance-Guard, Falsification-Auditor. Each writes its hardest objection from its own seat, paired with the methodological resolution it would accept. Run on the static catalogue spec Jun 3, 2026; a live council for any topic is at /try.

1. Skeptic #01
   Raised

   The 1000–10,000× signal is plausibly an artifact of IUCN assessment-effort inflation rather than genuine extinction acceleration. The Red List expanded from roughly 5,000 assessed species in 1980 to over 157,000 by 2023, and newly assessed taxa are non-randomly selected: assessors systematically prioritize species already known to be imperiled, meaning a large fraction of apparent "LC→VU/EN" transitions in the dataset are first-time assessments of chronically threatened species, not documented deteriorations from a previously secure baseline. Simultaneously, the Pliocene fossil-record background rate (~0.1–1 E/MSY) is derived overwhelmingly from marine invertebrates with hard parts, a taxonomically narrow and preservationally privileged subset that will structurally underestimate true background rates when the modern denominator includes insects, nematodes, and fungi—making the computed ratio artificially large on both ends.

   Resolved

   Filter the IUCN Red List API (v4.0 `historical_assessments` endpoint) to retain only species documented at LC or NT status in an assessment dated before 1990, then recompute E/MSY exclusively from within-cohort status deteriorations 1990–2023; if the ratio drops below 100× background using this restricted longitudinal cohort versus >1000× in the full cross-sectional dataset, the signal is driven by ascertainment bias rather than velocity of loss. In parallel, replace the raw Pliocene fossil baseline with Alroy's (2008) shareholder quorum subsampling (SQS)-corrected rates from the Paleobiology Database (paleobiodb.org), stratified to taxonomic classes present in both the fossil and modern IUCN records; a corrected background rate that rises by even one order of magnitude would shift the E/MSY ratio below the 1000× support threshold and settle whether the denominator mismatch is load-bearing for the financial-risk inference.

2. Fact-Checker #02
   Raised

   The fossil-record background extinction rate (the Pliocene baseline denominator) carries an intrinsic uncertainty of roughly one order of magnitude — commonly quoted as 0.1–1 E/MSY depending on taxon, stratigraphic interval, and preservational bias — meaning the reference value itself spans the gap between the 100× and 1000× thresholds before any modern data are introduced. Compounding this, the IUCN Red List has formally assessed only ~157,000 of an estimated 2 million described species (and perhaps 8–10 million total eukaryotic species), with coverage exceeding 90% only for birds, mammals, and amphibians; using total described species as the denominator while drawing extinctions from this heavily vertebrate-biased numerator introduces a systematic undercounting bias of 10–50×. Additionally, IUCN status reclassifications (LC→VU→EN→CR) frequently reflect improved survey effort rather than genuine population collapse, and the formal EX declaration lag (typically 10–50 years post-last-sighting) further decouples the metric from actual extinction timing. The combined multiplicative uncertainty budget — background rate ×10, taxonomic coverage ×10–50, status-change signal noise ×2–5 — easily spans two to three orders of magnitude, rendering the 100× versus 1000× discrimination threshold untestable as stated.

   Resolved

   Restrict the E/MSY numerator to confirmed EX + EW (Extinct in the Wild) declarations for the three high-completeness IUCN taxa (Aves, Mammalia, Amphibia, all with >90% species assessed per IUCN 2022 summary statistics), apply the Bland & Schreiber (2017) possibly-extinct correction factor with its 95% CI, and compute the ratio against three explicit background-rate scenarios (0.1, 0.5, and 1.0 E/MSY) drawn from Raup & Sepkoski marine fossil data and Ceballos et al. vertebrate fossil calibrations, reporting a ratio range rather than a point estimate; the hypothesis is considered supported only if the lower bound of the ratio CI exceeds 1000× under the highest plausible background rate (1.0 E/MSY), and falsified only if the upper bound falls below 100× under the lowest (0.1 E/MSY), with all GBIF occurrence endpoints filtered to GBIF data-quality flag "coordinateUncertaintyInMeters < 1000" and basis-of-record = "PRESERVED_SPECIMEN" to reduce ghost-range inflation.

3. Researcher #03
   Raised

   The dominant uncontrolled confounder is IUCN assessment-effort expansion: the number of formally assessed species on the Red List grew from roughly 16,000 in the early 2000s to over 150,000 by the mid-2020s, meaning that a large fraction of apparent "status upgrades" in any 50-year window represent first-time assessments of species entering the list directly at VU, EN, or CR — not genuine documented deterioration from a previously established LC baseline. Because the methodology counts LC→VU→EN transitions without conditioning on whether an LC baseline was ever formally recorded for the same taxon, assessment intensity and E/MSY are positively correlated through an ascertainment channel entirely independent of actual extinction dynamics, biasing the ratio upward without bound as monitoring coverage expands.

   Resolved

   Restrict the numerator to species with a minimum of two formal, temporally separated IUCN assessments on record — filtering via the `assessment_history` endpoint of the IUCN Red List API v3 — so that only within-species categorical deterioration is counted, consistent with the Red List Index methodology of Butchart et al. (2004, *Science* 313:1100). This sampled-species panel absorbs the assessment-expansion confounder by conditioning on a stable, continuously monitored cohort. The resulting E/MSY should then be cross-validated against the Zoological Society of London / WWF Living Planet Index (LPI database, accessible through the IUCN–ZSL data partnership), which tracks logged abundance trajectories independently of categorical status assignments and provides an assessment-effort-orthogonal robustness check on whether the 1000× threshold survives ascertainment correction.

4. Compliance-Guard #04
   Raised

   The ESRS E4 standard under the EU Corporate Sustainability Reporting Directive (CSRD) — which requires material, verifiable biodiversity impact metrics in regulated sustainability reports — is the primary exposure channel, because if this unvalidated E/MSY ratio is cited in corporate disclosures or prospectuses for biodiversity-linked financial instruments (nature-based bonds, biodiversity credits, or catastrophe reinsurance tranches), it could constitute a material misstatement actionable under CSRD enforcement and, where securities are involved, SEC Rule 10b-5 (17 CFR § 240.10b-5). Separately, reinsurance carriers incorporating this 1,000–10,000× multiplier prematurely into Solvency Capital Requirement models under Solvency II Directive 2009/138/EC risk misstating systemic biodiversity exposure, inviting EIOPA review or NAIC equivalent scrutiny, because IUCN status upgrades include taxonomic revisions and data-deficient reclassifications that are not confirmed extinctions and therefore inflate the computed E/MSY numerator.

   Resolved

   The gating condition requires independent cross-validation of the E/MSY figure against at least two peer-reviewed extinction-rate estimates derived from non-overlapping methodologies — for example, species-area relationship models alongside IUCN transition matrices — with the Pliocene fossil-record baseline anchored to a formally published paleontological reference carrying documented confidence intervals. The combined API data pipeline must pass a documented data-provenance audit confirming that each IUCN status transition counted as an extinction event reflects a genuine extirpation rather than a taxonomic revision, assessment-effort artifact, or data-deficient-to-assessed reclassification. Until the E/MSY ratio formally crosses the SUPPORTS threshold under this validated methodology, any regulatory filing, prospectus, or reinsurance pricing model referencing this hypothesis must carry an explicit disclaimer — conforming to the SEC statutory safe harbor under 15 U.S.C. § 78u-5 and CSRD's verifiability principle — stating that the extinction rate estimate remains a falsifiable hypothesis under active peer-review validation and has not achieved the evidentiary standard required for regulatory disclosure or financial instrument pricing.

5. Falsification-Auditor #05
   Raised

   The fossil-record background rate used as the denominator is itself uncertain across published literature by roughly one order of magnitude—estimates for the Pliocene baseline range from approximately 0.1 to 1.0 E/MSY depending on taxon, stratigraphic resolution, and preservation correction—meaning the ratio's denominator alone carries a 10-fold variance. Compounding this, roughly 20–50% of IUCN status escalations in any given assessment cycle are flagged internally as driven by improved knowledge rather than genuine deterioration, inflating the numerator by an unknown but substantial factor; and the choice of denominator for "described species" (1.5 million catalogued vs. ~8.7 million estimated total) introduces an additional 4–6x swing. These stacked uncertainties together produce a natural variance band of 30–100x around any central E/MSY ratio estimate, meaning the gap between the FALSIFIES condition (≤100x) and the SUPPORTS condition (≥1000x) lies entirely within the instrument noise and could be traversed simply by selecting among defensible published parameters—rendering the FALSIFIES threshold unreachable by genuine null-model results rather than by empirical evidence.

   Resolved

   A Monte Carlo null model should be constructed by drawing background-rate values from the full published distribution (log-uniform across the 0.1–1.0 E/MSY range, citing Raup, Foote, and Alroy estimates), crossing it with bootstrapped IUCN status-change counts after excluding records flagged as knowledge-driven reassessments (using the IUCN's own "reason for change" metadata field), and sweeping the species-total denominator from described-only to Mora et al. projected totals; the resulting ratio distribution defines the empirical null envelope. If the 95th percentile of that null envelope exceeds 100x background, the FALSIFIES threshold must be moved to below the 5th percentile of the null—likely ≤20x—to constitute a genuinely falsifiable lower bound, or a direct validation arm must be added using well-sampled clades (e.g., mammals, birds) where IUCN coverage is near-complete and species totals are tightly constrained, anchoring the ratio estimate to a subset where instrument variance is demonstrably smaller than the 10x gap between thresholds.

Unlike the static stress tests above (synthesised against the frozen catalogue spec), this is what a 3-voice council found in the most recent biweekly review. Refreshed on the 1st and 15th of each month at 09:00 UTC. Each voice runs one bounded web search via Anthropic's web_search_20260209 tool, cites what it finds, and recommends a verdict. The verdict aligns with the curated catalogue status (supported).

1. supported

   May 30, 2026 · added to catalogue at status "supported"

• Swiss Re Institute sigma-series expected annual loss for pollinator-collapse and fishery-failure perils rises 2–4× by 2028, widening reinsurance rate-on-line 40–80 bps for natural capital-correlated agricultural portfolios within two renewal cycles.
• The European Banking Authority mandates Pillar 2 capital add-ons of 60–120 bps in risk-weighted-asset density for banks with material biodiversity-exposed loan books (agriculture, forestry, fisheries) by Q4 2028 under its nature-risk prudential roadmap.
• Munich Re NatCatSERVICE revises boreal and tropical ecosystem-service-correlated peril loss expectations upward by 2–3× by 2030, repricing parametric trigger thresholds in ILS contracts covering fire, flood, and pollinator-dependent crop failure.
• The World Bank IBRD structures coupon step-up provisions of 35–75 bps on biodiversity performance bonds for high-endemism sovereigns (Brazil, Indonesia, Madagascar) within 18 months of TNFD-mandatory reporting adoption by G20 financial regulators.
• Aon Securities and Gallagher Re bring the first catastrophe bond tranches explicitly covering biodiversity-collapse-triggered fishery and crop losses to market at $500 million–$1.5 billion notional per issuance by 2027, with total ILS natural capital tranche outstanding reaching $5–10 billion by 2030.