Methodology: How the UN80 Executive Board Readout Was Produced

A briefing built from one UN meeting's transcript. It records what each delegation put on the record, groups them by the position they took, and shows the evidence behind every grouping. It is not a UN document and carries no official standing.

The meeting ran a single agenda item: engagement with the UN80 initiative. The heads of UNDP, UNFPA and UNOPS each made an opening statement, 27 delegations responded, and the three heads replied in two rounds. Two proposed mergers carried the session, and both were argued on cost, evidence and risk.

The readout carries a section this track's earlier readouts do not: Tracked figures, which follows specific load-bearing numbers across meetings and reports whether the room comes back to them.

Sources for every number and quote on the readout:

• Transcript: transcripts/un-funding/2026-08-25_k1ecbhx56v.json
• Scoring sheet: research/un-funding-un80-position-scores-2026-08.md
• Source analysis: readouts/un-funding-un80-2026-08-source.md

The transcript came from the UN transcripts portal, which publishes ASR transcripts with speaker attribution attached. Collection is scripted:

python pipeline/fetch_un_transcripts.py fetch asset/k1e/k1ecbhx56v --silo un-funding
python pipeline/fetch_un_transcripts.py convert "transcripts/un-funding/2026-08-25_k1ecbhx56v.json"

The portal supplies each statement's speaker as a structured object — country code, full name, function and any group the delegate speaks for. Attribution therefore comes from the source rather than from reading cues in the text, which is what lets the attribution gate below check the readout mechanically.

The JSON also carries the re-cut warning that shapes the verification section:

timestamps_flagged:          true
timestamps_note:             "UN Web TV re-edited this video after the transcript
                              was created; timestamps may no longer match the
                              current video."
original_duration_seconds:   12015
current_duration_seconds:    10556

The transcript holds 78 statements from 32 distinct speakers, including the President in the chair.

Every meeting is classified before it is analysed, because the classification decides the question the readout asks. A formal meeting is about resolving contested positions, so divergence is the finding. An informal briefing is about floating ideas, so convergence is.

This meeting is recorded formal, on four grounds:

• The Executive Board is a governing body sitting in formal session.
• The President runs a speakers' list and takes delegations in groups before inviting the entities to respond.
• Delegations read prepared national and group statements.
• Contested text is on the table. The Board must decide whether to endorse terms of reference for an independent assessment, and a draft decision on the item is under negotiation.

Because the meeting is formal, the readout uses the two-rubric scoring described below. The classification is recorded in transcripts/un-funding/INDEX.md, not inferred while drafting.

A note on which track this readout belongs to. The un-funding track's own routing rule sends sessions whose agenda item is reform to the un80 track, and this session's agenda item is reform. It was nevertheless built here, as a recorded exception, because the argument in the room was financial throughout: the divide was whether the cost case for merging holds, and the session put the three entities' funding models side by side for the first time in this series. The exception and its reasoning are written into the track index.

Every position in the readout is taken verbatim from the transcript. Nothing is inferred from a delegation's bloc, its known policy, or what it said at another meeting.

The defining divide is identified first, then everything else is organised around it. Here it is whether the financial case for merging holds up well enough to act on now. That framing was chosen over a simpler "for or against the mergers" axis because the transcript does not support the simpler one: Norway calls both mergers "constructive propositions" and still lists risks, and Denmark backs exploring both while rejecting cost saving as the rationale. What separates delegations is their reading of the evidence, not their appetite for reform.

Two rules from the project's house style bind the prose:

• No ranking of delegations. The readout reports what each delegation said; it never grades whose argument was strongest, hardest or most demanding.
• Plain language. Domain terms are explained where they first appear. Quotes are exempt from any length or vocabulary constraint — trimming a delegate's words to fit a target would falsify the record.

The horizontal position of every dot comes from a scored reading of what each delegation said. The method is in pipeline/position-scoring.md; the worked sheet with every score and quote is research/un-funding-un80-position-scores-2026-08.md.

What is scored

Because this is a formal meeting, two mirrored rubrics of seven indicators are built, each summing to 100. A delegation scores on whichever side it argued, and a delegation that argued both scores on both.

• A–G — the case is good enough to act on. The evidence before the Board supports proceeding: endorse the terms of reference, move on the merger or its alternatives, treat delay as its own cost.
• A′–G′ — the case is not made. The financial evidence is incomplete, contested or speculative, and does not support a structural decision now.

The seven indicators

How the axis works

Write P for the total scored on A–G and N for the total scored on A′–G′. Then:

net        = P - N     which way the delegation argued, and how far off neutral
on record  = P + N     the total it put on the record, either way  <- the X axis

The X axis is on record, never net. Adding the two poles rather than netting them keeps a delegation that argued both ways visible. It is also why the United States, which argued the case fails at greater length than anyone argued anything, sits at the right-hand edge rather than in the middle. Direction lives on the Y axis and in the camp colour instead, which is why the United States (104 on record) and France (82) share the right-hand column while sitting on opposite rows.

Columns are fixed thirds of the 0–100 scale — 0–39, 40–66, 67+ — so a column means the same thing here as on every other readout in the series. Within a cell, dots are ordered by on record and spaced evenly by rank. Nothing is placed randomly.

What the score does not claim

The claim is ordinal. Read a score as "argued more than", never as a measurement of how right or how influential a position is. Scoring the counter-pole on its own 100-point rubric means a delegation arguing the case fails can sit in the same column as one arguing it holds. That is a statement about how much each put on the record, not that the two positions carry equal weight in the room.

A question is never counted as opposition

A delegation asking how something will work has not taken a position against it. Belgium's question to the UNFPA Executive Director — whether a new entity created by a contentious vote would be weakened by it — scores nothing on the counter-rubric. Its counter-rubric points come from a separate, stated complaint that information has often not reached the Board in time to support its oversight role.

Criticism that reform is moving too slowly is pressure to do more, and scores on G rather than on the counter-rubric. Sweden urging the agencies to "continue to engage fully and constructively in advancing these reforms" is scored that way.

Who is not scored

The three executive heads are unscored. They authored the case the axis measures reactions to, so they cannot engage with their own proposal. They appear in their own band below the dashed line, with no horizontal position, because the axis counts how much was argued and they argued nothing — they presented. The President is procedural and is named as excluded in the scoring sheet.

No observers spoke. Every delegation on the floor was a member state or a member state speaking for a group, so the separate observer band the method provides for is not used here.

Speaking length

Word counts are computed mechanically from the transcript, never estimated:

python pipeline/word_counts.py transcripts/un-funding/2026-08-25_k1ecbhx56v.json \
    --exclude "Executive Board" --exclude "Alexander de Croo" \
    --exclude "Jorge Moreira da Silva" --exclude "Diane Kitah" --md

Across the 27 delegations the median is 364 words and the brevity threshold (70% of the median) is 255. Portugal spoke 214 words and is the only delegation carrying the brief marker on this readout.

One axis was built and then removed

A third Y dimension grouping delegations by their position on the UNFPA–UN Women merger was built and dropped before publication. It put 18 of the 27 delegations in a single cell, which shows no clustering and produced overlapping dots. The finding it would have carried — that withholding support is close to the room's default on that merger — is stated in the prose instead, where it does not need a plot to be legible.

This track carries a short list of tracked figures: specific numbers or claims from an earlier meeting that reframed what a debate was about, and that a reader would want to know the answer to going into the next session. The list lives in the track's transcript index and is read before drafting each new readout in the track.

For each open figure, the new transcript is searched for a reaction at three levels, loosest last:

1. Direct citation — the figure's own numbers, or a paraphrase close enough to be unambiguous.
2. Concept engagement without the numbers — a delegation argues the same question the figure raised, without repeating it.
3. Loose topical mention — a statement touching the same ground without engaging the point.

A result is recorded every round, including a clean miss. "Nobody engaged it this round" is itself the finding, and it only means something if it is checked and stated the same way each time. On this readout, figure 1 — the UNDP budget breakdown from 24 August — produced no reaction at any of the three levels, and that is what the section reports.

Full method: pipeline/anchor-tracking.md.

The project's standard is that every position-critical quote in a published readout is checked against the meeting audio: the video is downloaded, sliced at the transcript's own word timestamps, re-transcribed with a second speech-recognition model, and compared. That check ran on this readout on 26 August 2026. All 15 quotes matched the recording. No corrections were made.

The obstacle first, and how it was solved. UN Web TV re-cut the video after the transcript was made — from 12,015 s to 10,556 s — so the word timestamps no longer pointed at the right moments. Slicing at those timestamps would compare a quote against a different part of the meeting, which is worse than not checking, because it can stamp "verified" on the wrong speaker's words. The offset therefore had to be measured rather than assumed, following quote-verification.md §4b.

Seven probes, one constant. Fixed points across the recording were transcribed and matched back to the transcript, two of them bracketing the front and tail. Every probe returned an offset between 1,374.5 s and 1,378.8 s — a spread of 4.4 s, with no drift from the start of the meeting to its end. A single uniform offset of 1,376.7 s holds across the whole recording. Two figures are worth naming because both are wrong and both are tempting: the 1,459 s difference between the two durations is not the offset, and neither is the 1,021 s constant measured for the 24 August meeting, which was a different cut of a different recording. The full derivation is recorded in the run folder.

The result. Twelve of the 15 quotes matched on a first-pass slice with three seconds of padding. Three needed a wider window: two because the quoted words sit further from the transcript timestamp than the padding allowed, and one — Spain's efficiency figure — because the first pass returned a smoothed reading that dropped a word the speaker actually repeated. A tighter re-slice returned the readout's wording exactly, so the quote stands as written. No quotation on this readout was changed as a result of the audio check.

Attribution

Every quote rendered in the readout is matched back to the transcript and checked against the speaker object the portal supplied. This catches a quote attributed to the wrong delegation, a quote that does not appear in the transcript at all, and an actor who never spoke.

python pipeline/verify_attribution.py \
    readouts/un-funding-un80-2026-08.html \
    transcripts/un-funding/2026-08-25_k1ecbhx56v.json

Result: clean. 15 attributed quotes checked against 32 speakers, all matched, no warnings.

Scoring-sheet completeness

The scoring sheet is checked for invented score ranges, deferred work, and any transcript speaker the sheet never names. An actor left out must be named with a reason rather than covered by a group sentence.

python pipeline/verify_scoring_sheet.py \
    research/un-funding-un80-position-scores-2026-08.md \
    transcripts/un-funding/2026-08-25_k1ecbhx56v.json

Result: clean. Every one of the 32 speakers is scored, banded as a presenter, or named as excluded with a reason.

Evidence quotes

Eight evidence quotes spanning both rubrics were grepped against the transcript, each expected to appear exactly once. All eight matched.

Rendering checks

Run in a headless browser against the built page:

• All 30 dots render on both Y settings, with no overlapping dots and nothing clipped outside the plot area.
• 167 camp-coloured names were checked against their computed colour; all matched their camp, none fell back to grey.
• Edit mode enters cleanly, exposes 149 editable regions and 10 section controls for 10 sections, and leaks no editable node into the plot, toolbar, legend, evidence panel, glance table or axis explainer. Removing a section and undoing it leaves all 30 plot dots live and the evidence panel still wired.
• The page opens in light theme regardless of the reader's system setting, and the dark toggle works in both directions.

• The source is ASR, not an official record. The UN publishes these transcripts as a preview tool. Wording in any quotation may differ from the official record when it appears.
• The audio check covers the 15 rendered quotes, not the whole transcript. See Stage 4. Those 15 were verified against the recording; every other quotation reproduced from the transcript elsewhere on the readout remains ASR text.
• Scores are ordinal, not measurements. They say how much a delegation put on the record, not whether it was right, and not how much influence it carries in the room.
• One meeting only. Positions reflect what was said on 25 August 2026. A delegation silent on a point here may hold a clear position on it elsewhere; silence is never read as a position.
• Figures are as stated by speakers. The savings, cost and transition numbers on the readout are what delegations and executive heads said in the room. They have not been checked against the underlying reports, and the meeting itself established that the agencies' own figures do not yet reconcile.
• Group statements are scored as one actor. Where a delegation speaks for a group, the score reflects the statement as delivered. Members of that group who also spoke nationally are scored separately on their own statements.