Two first baselines read about 14 to 18 points lower, but only one has the source quality the band-crossing rule requires.
“Madrid's band crossing cleared the benchmark's source-quality rule on a government source and a treaty-body decision; would a court, regulator or government record of University of Glasgow's reported BAE Systems protest coordination clear the same rule for its crossing?”
Valencia is ranked 39th among global cities with a published score of 60.9 out of 100. That number was never checked against evidence.
Why it matters
Valencia's published 60.9 came from a template number, not research. A first evidence-based reading proposes 41.9. The city's flood response earned credit. Its homelessness record did not.
A court, regulator or government record on Glasgow's reported coordination with BAE Systems over a student protest would raise the evidence quality behind its proposed crossing.
TBD
Cornell UniversityMEDIUM
The outcome of the New York Attorney General's review and the reopened Tompkins County criminal case.
Findings of the governor's third-party review of lethal-injection procedures.
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.
Madrid's city hall helped an evicted 87-year-old woman return home within a week and counts its homeless every year. Its published 65.6 still outran the evidence. A first check proposes 48.1.
Madrid's published score is 65.6 out of 100, in the Established band. It had never been individually assessed. A first full assessment proposes 48.1, in the Functional band. Confidence is low: 20 of the 40 sub-measures have no evidence and sit at the default midpoint of 3 out of 5.
Where this sits
Read the full signal
The published score is unchanged, and nothing has been applied.
The city's recent conduct was good. On 23 September 2026 Mayor José Luis Martínez-Almeida offered an evicted 87-year-old woman immediate housing. The city's housing company then mediated an agreement that returned her home by 29 September. Madrid also counts homelessness every year: 1,015 people were sleeping rough in 2025, down 9%.
The weight on the other side is Cañada Real, a settlement on Madrid's edge. About 4,500 people, at least 1,800 of them children, have lived without electricity since October 2020. In 2025 the European Committee of Social Rights found Spain in violation of the European Social Charter. Responsibility is shared among the national government, the region and the city, so the assessment scores the city's share with low confidence.
Ayer entendieron que el Ayuntamiento era el instrumento adecuado para poder llegar a un acuerdo
The index's top-ranked university leads on repairing its slavery past, and overrode its own staff and students on arms. A first check proposes 48.7, but the adverse sources are news reports, not court or government records.
The University of Glasgow is ranked first among 100 universities at 62.5 out of 100, in the Established band. That score was never assessed against evidence. A first assessment proposes 48.7, in the Functional band, a 13.8-point gap. Confidence is low.
Where this sits
Read the full signal
The published score is unchanged, and the proposal has not been applied.
Glasgow's record on repairing historical harm is the strongest the benchmark has seen in a university. It published a report on its own profits from slavery in 2018 and runs a joint reparative-justice master's degree with the University of the West Indies. Its Top-Up programme targets students from deprived areas, care-experienced students, refugees and young carers.
On current conduct the record is weaker. In November 2024 the University Court voted 14 to 7 against arms divestment, after its own consultation found 81% of staff and 86% of students in favour. Emails reported by Al Jazeera on 24 September 2026 show Glasgow told BAE Systems, a funder, that it had informed Police Scotland of a planned student protest at a 2023 careers fair. In June 2026 it opened a voluntary severance scheme over a forecast shortfall of about GBP 10 million.
The proposed band crossing rests on news reports and student press. No court, government or international-body source was located, so the assessment holds this finding at low confidence.
Meanwhile, emails show the University of Glasgow apparently reassured BAE it had informed Police Scotland of a planned protest targeting an October 2023 careers fair.
Cornell's score had never been checked. The proposed 41.9 rests on a state review, the university's own statement and a faculty finding. Lawsuit allegations are not scored as fact.
Cornell University is ranked 12th among 100 universities at 54.7 out of 100, in the Functional band. A first assessment proposes 41.9, also Functional, so no band change would result. The published score is unchanged, and the proposal has not been applied.
Where this sits
Read the full signal
The New York State Attorney General announced a review of Cornell's initial response to a 2024 sexual-assault allegation, and the Tompkins County District Attorney reopened the criminal case (Boston Globe, 30 September 2026). Cornell's 21 September 2026 statement calls the allegations deeply disturbing but acknowledges no failure and routes the response to the legal process. The assessment scores those actions. It does not treat the lawsuit's allegations as established fact.
Separately, Cornell's Faculty Senate found in November 2025 that temporary suspensions were used to deter, retaliate or compel compliance, and that about 85% of suspended students had taken part in protest. On the credit side, Cornell's need-based aid covers the full cost of attendance for families earning up to $75,000, and a March 2026 task force adopted evidence-based prevention programmes.
Vietnam's published score was never checked. State media confirmed two Viet Tan activists were detained, and the group says they were abducted from Cambodia. The assessment scores the confirmed detention, not the abduction claim.
Vietnam is ranked 106th among countries at 34.4 out of 100, in the Developing band. It had never been individually assessed. A first assessment proposes 23.8, still Developing. The published score is unchanged, and the proposal has not been applied.
Where this sits
Read the full signal
On 29 September 2026 ABC News reported that Vietnamese state media confirmed the detention of Australian Hiep Tran and Norwegian-Vietnamese Duc Thuan Nguyen. Their group, Viet Tan, alleges both were abducted in Cambodia and taken across the border. The abduction is an allegation and is not scored as fact. The detention is.
The assessment also weighs Human Rights Watch's 2026 World Report: more than 160 political prisoners in 2025 and at least 32 convictions under Article 331 of the penal code. On the credit side, Vietnam abolished the death penalty for eight offences in June 2025 and set 2026 and 2030 targets in a health-care roadmap. Fourteen of the 40 sub-measures rest on default scores, so confidence is medium, not high.
Unlike the first-baseline cases, Bangladesh already had a real measurement. Three September 2026 laws strip the human rights commission of power over enforced disappearances, and a proposed 33.1 reflects them.
Bangladesh's published score is 39.8 out of 100, in the Developing band. It was set on 16 June 2026 after labour reforms. A re-check proposes 33.1, still Developing. The published score is unchanged, and the proposal has not been applied.
Where this sits
Read the full signal
Human Rights Watch reported on 22 September 2026 that three laws passed in September remove the National Human Rights Commission's power to investigate enforced disappearances, and move the Rapid Action Battalion's personnel and powers to a new unit with fewer safeguards. HRW says the laws break the governing party's reform commitments.
Of the 6.7-point gap, 4.8 points were measured on 21 September 2026 and were too small to file on their own. The other 1.9 points come from the new laws. The June labour-reform credit is kept. Confidence is medium: the evidence on the new laws is a single watchdog source, but the laws themselves are public acts of parliament.
Global Cities — Two Spanish Cities Meet Their First Real Evidence
Valencia and Madrid both read about 18 points below never-assessed starting values on city-government evidence alone.
Read the full signal
Valencia: published 60.9, proposed 41.9. The city's flood protocols are credited; the regional government's flood alert is excluded.
Madrid: published 65.6, proposed 48.1, at low confidence. The city's housing help for an evicted 87-year-old is credited; Cañada Real's power cuts are scored with shared responsibility.
Spain's own country score is separate: 51.2 measured this cycle, within 0.6 points of a proposal already pending at 50.6.
New York City appears in two indexes with different published scores: 48.4 in global cities (never assessed) and 56.3 in US cities (assessed 10 June 2026).
10 signals shown
Risk signals
Developments that may affect future scores. Watch items from the Oct 1 briefing.
Risk
Published scores that were never measured may sit well above what a first reading of the evidence supports.
Risk
The top of the universities index rests on scores that were never assessed against evidence.
Risk
Spain's two emergency housing decrees need a parliamentary vote within 30 days, and that vote bears on Spain's housing record.
Risk
Florida's attorney general has asked a court to pause new OpenAI model development and bar minors from ChatGPT; no court has ruled.
Risk
Egypt, Iran and Hong Kong each recorded new press-freedom arrests or sentences within four days, in measures already at or near the lowest score.
Score movements
Entities with score changes this cycle, followed by confirmed positions.
A court, regulator or government record on the redundancy round's withheld equality impact assessment would raise the evidence quality behind the crossing.
Evidence that Barrick has carried out the Loulo-Gounkoto agreement on overtime pay, expense reimbursement and earlier labour agreements would bear on the Critical line.
The city activated orange-alert protocols before the episode began, suspended activity at 26 centres in flood-prone zones, and handled 250 fire incidents and 211 police calls.
valencia2026-10-01
el Ayuntamiento se ha negado a hacer un censo oficial de las personas sin techo
Opposition parties state the city refused an official homeless census and that shelter capacity (57, 20, 10 places) is insufficient; a city count at one settlement rose from 21 (2022) to 114 (2025).
After a regional court annulled the 2024 moratorium, the city enacted a 2% per-neighbourhood cap on tourist flats in May 2026, described as Spain's most restrictive.
madrid2026-09-29
Ayer entendieron que el Ayuntamiento era el instrumento adecuado para poder llegar a un acuerdo
The ECSR found Spain violated the European Social Charter over Cañada Real power cuts affecting about 4,500 people.
university-of-glasgow2026-09-24
Meanwhile, emails show the University of Glasgow apparently reassured BAE it had informed Police Scotland of a planned protest targeting an October 2023 careers fair.
The NY Attorney General is reviewing Cornell's handling.
cornell-university2026-09-21
Cornell University is aware of a lawsuit concerning the deeply disturbing allegations of sexual assault involving members of the student community that occurred in 2024.
OHCHR welcomed Vietnam's abolition of the death penalty for eight offences (countervailing).
bangladesh2026-09-22
Three laws approved in September 2026 by the Bangladeshi parliament have reversed human rights protections and broken the ruling Bangladesh Nationalist Party's (BNP) commitments to reform.
Egyptian security forces raided homes and arrested six investigative journalists on 28-29 September 2026; no reasons were disclosed and phones and laptops were seized.
A federal judge tentatively approved a $78.7 million class settlement over the January 2020 Delta fuel dump on 28 September 2026; about $51.2 million would reach the class after fees.
After the failed execution of Christa Pike on 30 September 2026, Governor Bill Lee halted executions for the rest of his term and ordered a third-party review.
On 29 September 2026 the Council of Ministers approved two Royal Decree-laws extending protection against evictions of vulnerable households to 31 December 2030.
South Africa2026-09-28
Police said we can't open cases because we are foreigners.
Refugees in Durban told Al Jazeera that police refused to open cases ahead of the 30 September 2026 anti-migrant deadline.
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.
Copy-ready citation string for journalism, research, or academic use.
Compassion Benchmark. "Daily Briefing — Oct 1." compassionbenchmark.com/updates/2026-10-01. Accessed [Month Year]. Independent — entities never pay for inclusion, score changes, or suppression of findings.
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.