Who would Andy Burnham appoint as Chancellor?

This pilot sits outside the Observatory’s scientific and technological record. It is published as an independent application of its claim-analysis method.

Four press reports made sufficiently definite, outcome-testable claims about Andy Burnham’s choice of Chancellor. All four named Shabana Mahmood. John Healey was appointed.

The interesting finding is not simply that the reports were wrong. They appear to have converted evidence about factional lobbying—or a provisional plan—into apparent knowledge of the final decision.

4Outcome-testable reports

0Correct

4Missed

The raw scored-report hit rate is 0/4. That figure describes only this one episode. It must not be presented as a publication or journalist accuracy rate.

The reports look like one Westminster briefing cascade. They accurately detected a factional battle, but converted evidence about who insiders wanted into apparent knowledge of what Burnham would do.

Interpretation

What the reporting actually captured

A real contest was visible

The reporting consistently identified Ed Miliband and Shabana Mahmood as the principal publicly discussed candidates. It also captured factional opposition to Miliband and a City preference for Mahmood. That part was not invented: post-appointment accounts described intense briefing by competing camps.

Advocacy was repeatedly converted into apparent intent

Early articles were relatively careful: aides wanted Mahmood; Miliband was in the running; several people were tipped. By 15 July, this had hardened into “set to become”, “expected to be named” and even “definitely happening”. The evidence visible to readers did not harden in the same way. It remained anonymous and contested.

The articles contained their own warning labels

The Guardian’s 15 July report said Burnham’s team insisted that no final decision had been made. It quoted a source describing a world in which people were being briefed into and out of jobs by people who probably did not know. The Financial Times also recorded Burnham’s spokesperson calling the reports speculation, while colleagues of Mahmood said she had not been told she was moving to the Treasury.

Those caveats matter editorially—but they do not cancel headlines and leads that presented the appointment as expected or settled.

The miss may have been caused by error or by change

The eventual choice of Healey was described as a surprise. Subsequent Telegraph reporting quoted a source saying Mahmood had been “nailed on” as recently as the preceding Friday before “something changed”. This supports two live interpretations:

the sources never knew Burnham’s settled intention; or
Mahmood genuinely was the provisional choice and Burnham changed course.

Public evidence cannot adjudicate between them. Therefore the audit scores the published outcome claim as a miss, not the underlying source as a lie.

Institutional lesson

The key failure was a collapse of claim types.

It was not simply that journalists guessed the wrong name.

01PreferenceA faction wants Mahmood.
02ProbabilityMahmood is considered the frontrunner.
03IntentBurnham plans to appoint Mahmood.
04DecisionMahmood has been selected.
05OutcomeMahmood is appointed.

The coverage moved from stages 1–2 toward stages 3–4 without publicly observable evidence of a corresponding increase in source authority. The eventual outcome then falsified the strongest public formulation.

This suggests that any larger study should measure more than hits and misses. It should test whether journalists preserve—or erase—the distance between advocacy, expectation, intention, decision and outcome.

Why there is no journalist leaderboard

This pilot cannot support one because:

the sample contains only one resolved appointment;
several stories had multiple bylines;
headlines may not be written by the reporters;
the same briefing can appear across several outlets;
reporting a faction’s lobbying can be accurate even when its candidate loses;
a real provisional decision can change after publication; and
reporters with cautious wording would be unfairly compared with reporters making categorical claims.

A defensible leaderboard would require a preregistered protocol, archived article versions and roughly 30–50 independent, resolvable claims per journalist. Scores should reward calibration, not merely count correct names.

Method and sources

Reports were scored only where their published wording made a definite or strongly expected appointment claim. Four reports met that threshold: two from the Financial Times, one from The Guardian and one from The Times. All named Shabana Mahmood. John Healey was appointed.

This was a public-web pilot, not an archive-complete study. The unit of analysis was the published claim, not the journalist. A “miss” means that a scored claim named someone other than the person appointed; it does not establish a publication or journalist accuracy rate.