The most significant editorial findings in the Sep 24 briefing.
Editorial insight
This is a catch-up cycle. It widens the usual 14-day lookback to 15 days specifically to cover September 23, the one day this month no earlier check ever reached.
Two companies share an identical placeholder vector and an identical measured delta, but only one of them has ever had its rise applied.
“Tronox and Freeport-McMoRan share an identical published 18.8 placeholder vector and an identical measured +13.7 delta, yet Freeport-McMoRan's rise was applied in July while Tronox's is withheld this cycle — does a band-wide Critical-cluster review resolve that asymmetry, or does the next matching placeholder test screening rule five again, one company at a time?”
Abu Dhabi's published score of 21.9 out of 100 had never been individually checked before this cycle. A first review proposes 28.1 — still Developing band, no tier change, but the largest upward finding in recent cycles.
Why it matters
An instrument that only ever finds harm isn't measuring anything. This is the same evidence standard, applied to money a court actually returned to workers.
A coordinator-level review of the never-checked Critical-band Fortune 500 cluster Tronox shares with Penn Virginia, Rex Energy and Sanchez Energy is the next step.
A UN treaty-body communication on the new exit-ban decree, or a documented individual case, is the next test for a score sitting half a point from the next band.
How to read this briefing— Bands, scores, and terms
Schema guide
The 5 performance bands
critical0–20
developing20–40
functional40–60
established60–80
exemplary80–100
Two scales
Each of the 8 dimensions is scored 1.0–5.0; these combine into a 0–100 composite score, mapped to the 5 bands above.
Key terms
Band crossing
A score change large enough to move an entity from one performance band into an adjacent one — the most structurally significant finding in a given cycle.
Boundary watch
An entity whose current score is within 3 points of a band threshold, flagged for priority reassessment in the next cycle.
Carry-forward
A dimensional credit retained from a prior assessment when new evidence is insufficient to revise a specific dimension; disclosed explicitly.
First baseline
An entity's inaugural composite score — no published score exists to compare against, so delta is not shown.
Floor designation
The most serious finding: all 8 dimensions resolve at the lowest behavioral anchor (1.0/5.0) across multiple cycles, yielding a composite of 0.
Forward trigger
A scheduled future reassessment event (e.g., a policy implementation date or legislative deadline) that may materially change an entity's score.
Progressive's published score was never checked. A federal jury's finding of bad faith — not an accusation, a verdict — is the first hard evidence the benchmark has on the company.
Progressive's published score of 76.3 out of 100 had never been individually measured; six of its eight published dimensions sat at an identical, unchecked 4.0. A first measurement proposes 36.9, two tiers lower.
Where this sits
Read the full signal
On September 11, 2026, a federal jury in the Northern District of Oklahoma found Progressive breached its duty of good faith handling a driver's claim for $25,000 in underinsured-motorist benefits, awarding $20 million in compensatory damages and $20 million in punitive damages. The jury heard evidence that Progressive's adjusters "lacked a basic understanding of how to handle Oklahoma underinsured motorist claims" and were paid under a company-wide profit-sharing program tied to cutting claim costs — a structural incentive, not one employee's mistake.
The judgment is not final. Progressive is seeking post-verdict relief in Tulsa federal court and an appeal is expected, which is why the measurement does not credit the company for repairing the harm yet. Genuine strength is also counted: Progressive ranked #19 on Fortune's 100 Best Companies to Work For in 2026, a ninth consecutive year, and #1 on Forbes's list for employer culture from a 217,000-worker survey.
Progressive's adjusters lacked a basic understanding of how to handle Oklahoma underinsured motorist claims
A court found a disability insurer underpaid the staff who decide disability claims. That is the sharpest kind of finding — it is the company's own workforce, adjudicated, not alleged.
Unum Group's published score of 60.9 out of 100 had never been individually checked; it sits on a starting vector identical to three other insurers (The Hartford, Travelers, MetLife). A first measurement proposes 39.4, two tiers lower.
Where this sits
Read the full signal
On April 17, 2025, a federal court in Tennessee gave final approval to a $14.8 million wage-and-hour settlement. The class: about 910 Unum disability benefits specialists — the employees who process claims and decide whether disabled people get paid — who said they were wrongly classified as exempt from overtime pay. Roughly 52 of them received more than $30,000 each.
Real strength is counted too: Unum's return-to-work and vocational rehabilitation programs, designed to restore a claimant's earning capacity, score as the single strongest finding in the assessment. And history is kept separate from today: a 2004 settlement with 48 states, which forced Unum to reassess 200,000 disability claims, is recorded as real but compelled reform, not current conduct.
This is the one genuine re-measurement of the night, not a first check — real prior evidence plus a fresh Human Rights Watch report moved the number.
Australia's published score of 62.5 out of 100 falls to a proposed 48.8, crossing from the benchmark's Established tier into its Functional tier. Unlike tonight's other three proposed changes, this one rests partly on a real earlier baseline, not purely on an unchecked placeholder.
Where this sits
Read the full signal
Human Rights Watch marked 25 years of Australia's offshore detention system on September 10, 2026: more than 5,000 people transferred to Nauru and Manus Island since 2001, at least 14 deaths since 2012, and an estimated $1.8 billion paid to Nauru to permanently accept people Australia turns away. Eight days later, Australia's Home Affairs Minister announced 100 more compliance officers and 250 new detention beds, a measure a rights group warned could deter exploited workers without papers from ever reporting abuse.
Australia is also credited for real transparency: its own Productivity Commission published data in July showing 4 of 19 Indigenous-outcome targets are getting worse — a government publishing its own bad news, counted as a genuine strength even as the overall score falls. A widely circulated figure of roughly 350 deportations to Nauru could not be confirmed by either primary source and was left out of the measurement.
to accept people whom Australia rejects and expels
The same math the benchmark applied to a different company in July reappeared tonight on Tronox. That repeat is evidence the number belongs to the benchmark's own unmeasured data, not to either company.
Tronox measured 32.5 out of 100 tonight, 13.7 points above its published 18.8 and a tier higher — but no change was proposed. The finding is recorded and set aside for a wider review instead.
Where this sits
Read the full signal
The reason: Tronox's published 18.8 is shared, digit for digit across all eight measured behaviors, with Penn Virginia and Rex Energy — two companies checked this same cycle and found to no longer exist as independent public companies. A number one live company shares with two dissolved companies cannot be a measurement of the live one.
The match runs deeper. In July 2026, the benchmark applied the identical move — 18.8 up to 32.5, a 13.7-point rise — to a different company, Freeport-McMoRan, starting from the identical published number. Two independent measurements, five months apart, landing on the exact same number from the exact same start, is strong evidence the starting number itself is the problem. Filing Tronox alone would fix that same underlying number one company at a time instead of all at once, so tonight's finding is being routed to a review of the whole group of never-checked Fortune 500 companies in the benchmark's lowest tier.
Fortune 500 — Two Insurers Get Their First Real Measurement, Both Fall
Progressive and Unum Group both had first-ever checks tonight; both landed well below their never-measured published scores.
Read the full signal
Progressive (76.3 published, 36.9 proposed): a $40 million bad-faith jury verdict, not yet final, is the first hard evidence checked against a placeholder score.
Unum Group (60.9 published, 39.4 proposed): a court-approved $14.8 million settlement found the company underpaid its own disability-claims staff.
Against already-checked insurers (State Farm 13.8, Cigna 20.3, Centene 32.8, Humana 35.2), both proposed values still sit above the pack — but the direction is consistently downward whenever an insurer placeholder gets its first real check.
·medium
Fortune 500 — A Placeholder Match Repeats a July Pattern Exactly
Tronox's finding matches a real change applied in July to a different company, down to the exact number.
Read the full signal
Tronox measured 13.7 points above its published 18.8 — the identical delta the benchmark applied to Freeport-McMoRan in July from the identical starting number.
Tronox's published vector is shared, digit for digit, with Penn Virginia and Rex Energy, both confirmed this cycle to no longer exist as independent companies.
No change was filed for Tronox; the finding is routed to a review of the whole never-checked Critical-band Fortune 500 cluster instead.
9 signals shown
Risk signals
Developments that may affect future scores. Watch items from the Sep 24 briefing.
Risk
Tronox's placeholder match with two defunct companies, and its identical delta to a real July change on a different company, points at a wider Fortune 500 Critical-band cluster that has not been individually checked.
Risk
Two more insurers (Progressive, Unum Group) have now had their starting placeholders replaced by real evidence and both measured well below published — the remaining insurers sharing the same starting numbers have not yet been checked.
Risk
China's published score sits half a point above the floor of its band; a UN treaty-body communication or a documented individual exit-ban case would be the next test for this position.
Risk
Abu Dhabi's positive labor-court finding rests entirely on UAE government and UAE-owned sourcing; independent verification would resolve whether this becomes a confirmed upgrade.
Score movements
Entities with score changes this cycle, followed by confirmed positions.
A methodology decision on how the benchmark treats a composite that lands exactly on a band line, and any consolidated workforce disclosure from Berkshire, are the next steps for this row.
Whether Google DeepMind's newly launched AGI-debate institute publishes work unflattering to Google, and its funding and independence terms, are the next tests for this boundary position.
A UN treaty-body communication on the new exit-ban decree, or a documented case of someone barred from leaving with no notice or appeal, would be the next test for this position.
Any documented change to Unum's own overtime and claims-staffing practices since the 2025 settlement would be the next test for whether this reading holds or moves further.
A coordinator-level review of the never-checked Critical-band Fortune 500 cluster Tronox shares with Penn Virginia, Rex Energy and Sanchez Energy is the next step, not further one-off measurement.
Whether Djibouti's own government contributes documented funding or capacity to the Yemeni-refugee reception, beyond the currently 15%-funded UNHCR appeal, is the next test for this position.
floor-confirmed
Evidence ledger
Primary sources reviewed in this briefing cycle. 15 sources linked.
abu-dhabiTier 2 · Journalism2026-09-21
Full payments were secured for 457 workers, while 491 others received partial payments, with remaining balances placed under structured instalment plans.
Employers retain disproportionate control under the kafala sponsorship system; workers who leave abusive jobs risk false absconding charges leading to detention and deportation; trade unions remain banned and there is no non-discriminatory minimum wage.
progressiveTier 5 · Gov/Court2026-09-11
Progressive's adjusters lacked a basic understanding of how to handle Oklahoma underinsured motorist claims
On April 17, 2025, the Eastern District of Tennessee granted final approval of a $14.8 million wage-and-hour class settlement covering about 910 Unum disability benefits specialists who alleged they were misclassified as exempt and denied overtime; participating members averaged over $10,000 each.
Of 19 Closing the Gap Indigenous-outcome targets, 1 is met, 3 are on track, 5 are improving but off track and 4 are worsening — published by the government's own agency.
Tronox's published dimensions are byte-identical to Penn Virginia's and Rex Energy's; this cycle independently established that Penn Virginia ceased to exist as a company in 2023 and Rex Energy sold substantially all its assets in 2018.
In July 2026 the benchmark applied a +13.7-point upgrade to Freeport-McMoRan, from 18.8 to 32.5 — the identical starting value and identical delta this cycle's independent Tronox measurement produced.
West Virginia's Bureau for Social Services published its FFY2025 Child Fatalities and Near Fatalities Annual Report on September 21, 2026, naming corrective actions including a new prevention-partnership model.
Djibouti's UNHCR operation is only 15 percent funded while appealing for $13.5 million to serve a surge of Yemeni arrivals.
Floor designations
·8 entities at composite 0 with documented evidence pattern
Composite scores resolving at zero — methodology disclosure
These entities consistently score the worst result across all 8 dimensions of compassionate conduct — the benchmark's most serious classification.
What “floor” means: every one of the 8 dimensions (Recognition, Response, Reduction, and 5 others) resolves at the lowest behavioral anchor (1.0/5.0) across multiple assessment cycles, yielding a composite score of 0. Full methodology.
Daily briefings surface the headline finding. Full benchmark reports include all 40 subdimension scores, complete evidence trails, certified assessments, and sector-level analysis packages — the record researchers and journalists cite.
Independence note: entities never pay for inclusion, score changes, or suppression of findings. Commercial services support access, interpretation, and institutional use only.