Every government does this, in every era, regardless of party — this platform isn't arguing otherwise, and readers of Foundation 14 will already know that's the whole point of building an accountability mechanism into government rather than trusting any single administration's good faith. What follows are six current, checkable cases — most from Labour's own record, two that span both this government and its Conservative predecessor, included because they show the pattern isn't about party. One case is included specifically because the evidence doesn't fit a simple negative story — a useful check on this article's own fairness.
A note on timing. Andy Burnham became Prime Minister on 20 July 2026. Every claim examined below was made under the same governing party, most originating under Sir Keir Starmer's premiership and continuing as live, unresolved commitments into Burnham's government. That continuity is itself the point — a manifesto pledge doesn't expire when the party changes leader.
1. "6,500 new teachers," funded by the private school VAT raid
The claim. From the 2024 manifesto onward, government has consistently said that removing the VAT exemption on private school fees would fund 6,500 additional teachers in state schools, while affecting only a small number of families — an early government estimate put pupils transferring into the state sector at around 3,000.
What actually happened. Independent data now puts the number of pupils who have left private schools since the policy took effect at over 30,000 — roughly ten times the original estimate, though the Treasury's own later "long-term steady state" projection was revised up to 35,000, an admission the initial public figure was always going to be wrong by an order of magnitude. On teacher numbers, the Department for Education points to around 4,600 additional teachers recruited toward the 6,500 target — but only once primary school teachers are excluded from the count. Overall, the total number of teachers in state-funded schools fell in 2024/25, with primary teacher numbers down by nearly 3,000. The government's justification — that falling birth rates make it "nonsensical" to count primary teachers — may well be true as a matter of demographic fact, but it is not how the original pledge was communicated to the public, and it was adopted only after the headline "more teachers" claim ran into a falling headline number.
The categorisation. This is omission on the original pupil-transfer estimate, which was published without the uncertainty that later Treasury figures admitted existed. It is misleading on the teacher pledge specifically — a plain-English promise of "6,500 more teachers" is being measured against a redefined metric adopted after the fact, in a way that lets government claim progress while the number most people would actually recognise — total teachers in state schools — went down. And it tips into outright denial in at least one instance: when reports emerged that officials were preparing contingency plans for private school closures, a government spokesperson called the suggestion "completely misleading" — even as an official source told The Times, on the record, that "we have to plan for the worst… contingency plans are necessary."
2. "We will not increase taxes on working people"
The claim. Labour's 2024 manifesto stated plainly: "Labour will not increase taxes on working people, which is why we will not increase National Insurance, the basic, higher, or additional rates of Income Tax, or VAT." Ministers have repeatedly said, since taking office, that this pledge has been kept.
What actually happened. The October 2024 Budget raised employer National Insurance from 13.8% to 15% and cut the threshold at which it becomes payable — a change government defended on the basis that the pledge covered employee NI, not employer NI, which the manifesto didn't explicitly distinguish. The independent fact-checking service Full Fact rated the pledge "Not kept," noting the OBR's own forecast that roughly 76% of the employer NI increase will ultimately be passed through to workers via lower wages. Then, in the 2025 Budget, government extended the freeze on income tax and NI thresholds to April 2031 — a change the OBR forecasts will create 780,000 more basic-rate and 920,000 more higher-rate taxpayers than previously expected by 2029/30. Pressed directly on whether this broke the manifesto pledge, the Chancellor told journalists: "I do recognise that yesterday, I have asked working people to contribute a bit more… I am not denying that this has an impact on working people" — while simultaneously maintaining the pledge had not been broken, because the headline rates themselves hadn't moved.
The categorisation. This is denial in its purest form: the Chancellor's own words acknowledge the practical effect on working people's pay, in the same breath as denying that this constitutes breaking a promise about taxes on working people. It is also misleading by omission of scale — "the rates haven't changed" is true and, on its own, gives a false impression of no impact, when the OBR's own numbers show a clear and substantial one.
3. "Smash the gangs" — the case that doesn't fit the pattern
The claim. On taking office in 2024, government pledged to "smash the gangs" behind Channel small-boat crossings, with urgent language about stopping "at nothing" to dismantle smuggling networks.
What actually happened, honestly. For roughly eighteen months, the numbers moved the wrong way. Crossings passed 30,000 in 2024, up 14% on 2023. The first half of 2025 saw the highest number of arrivals for that point in the year since records began, running 48% ahead of the same period in 2024. The full-year 2025 total reached around 41,000 — not a record (2022's 45,774 remains higher), but a clear and sustained increase under a government that had promised the opposite. Across this period, Home Office statements consistently used confident, near-identical language — "stop at nothing," "dismantle their business models" — while the trend line moved steadily against them.
Then something genuinely changed. In 2026, crossings fell substantially — provisional figures put the year on course for around 25,000, roughly 40% down on 2025. Even GB News, a broadcaster not inclined to credit this government, reported the decline with what one commentator called "grudging" acknowledgement that the strategy finally appeared to be working, crediting new cross-border enforcement agreements with France, Germany, Iraq and Serbia, and the Border Security, Asylum and Immigration Act that became law in December 2025.
The categorisation. This one is included deliberately because it resists a simple verdict. For roughly eighteen months, the confident "stop at nothing" language was misleading on timeline — it implied imminent results the policy was nowhere close to delivering, while crossings rose. But by 2026, the underlying policy appears to be producing real results, later than promised and only after sustained international agreement-building rather than any single dramatic action. The honest verdict is not "lie" or "kept" — it's that this was a pledge whose early framing overpromised speed, but whose substance may be arriving, just on a much longer timeline than the original rhetoric implied.
4. Shrinking the civil service back down after COVID — four targets, zero delivered
The claim. This is the case worth dwelling on longest, because it spans two governments of different parties, which is exactly the point — this isn't a Labour problem or a Conservative one, it's a state-capacity problem that has outlasted several changes of minister. Civil service headcount stood at 384,000 in 2016, the lowest since the Second World War. It then grew steadily through Brexit preparation and the pandemic response to around 475,000 by the end of 2021. In 2022, the Johnson government set an explicit target: cut 91,000 posts to return to 2016 levels within three years.
What actually happened, target by target. In November 2022, incoming PM Rishi Sunak scrapped the 91,000 figure entirely, calling top-down headcount targets "not the right way" to secure value, and replaced it with a vaguer instruction for departments to find their own "efficiency savings." By October 2023, that had softened further into departments being asked merely to "produce plans" on eventually reaching pre-pandemic levels — no date attached. By 2024, headcount had risen again, to around 543,000. Labour's government then set a new, much smaller target: a 10,000 reduction (roughly 2%) from late 2024. In March 2025, the Chancellor went further, announcing an intention to cut administrative running costs by 15% — which independent analysis from the Institute for Government calculated would require a headcount reduction of around 40,000, taking the total down to roughly 475,000. Instead, by March 2026, headcount had risen again, to 524,000 — an increase of 8,000 in the twelve months after that 15% commitment was made. The Institute for Government's own January 2026 analysis found that current voluntary exit schemes, even taken at face value, fall well short of the 29,000–40,000 reduction the government's own stated targets require.
The cost of not acting. This isn't a victimless target-missing exercise. A civil service workforce of roughly 524,000 — against a stated need to be tens of thousands smaller to hit the government's own published goals — represents a real, ongoing cost in a labour market where the public sector already competes directly with private employers for the same skilled workers: analysts, project managers, digital and data specialists, finance professionals. Every one of the roughly 30,000–40,000 posts the government's own analysis says are surplus to its stated targets is a person not available to a private employer trying to hire the same skills, at a moment when this platform's Labour Market Flexibility foundation documents record youth unemployment and a shrinking working-age population elsewhere in the economy. The unmet reduction also has a direct fiscal cost: on a conservative average salary estimate, an unmet gap of that size represents somewhere in the region of £1–1.5 billion a year in running costs that could otherwise reduce borrowing or fund frontline services — a small fraction of the £110 billion annual debt interest bill, but a fraction that four separate governments have now explicitly promised to deliver and none has.
The categorisation. This is denial repeated at scale: each government, on taking over the pledge, has restated a commitment to reduce headcount rather than admitting the previous target had simply failed. It is also the clearest example of no firm plan to deliver in this entire article — the Institute for Government, an independent and broadly sympathetic observer of civil service reform, has stated plainly that the mechanisms currently in place are not capable of hitting the government's own numbers, and headcount has now risen through four consecutive, published targets in a row.
The same shape, one more time: benefits into work
It's worth flagging, briefly, that this exact pattern — a confident headline pledge, followed by a target that quietly shrinks or is measured differently, followed by results moving the wrong way — shows up again in this platform's own Labour Supply and Labour Market Flexibility foundations. Government has pledged to raise the employment rate from 75.8% to 80% through its "Get Britain Working" programme, yet its own published figures show fewer than 1 in 100 people on the Universal Credit health element move into work in any given month, and around a quarter have been on that support for over a decade. The mechanism is different from the civil service case, but the shape of the gap between pledge and delivery is the same one running through every case in this article.
5. The AI investment numbers, and the scaffolding yard behind them
The claim. During President Trump's September 2025 state visit, the UK and US signed a "Tech Prosperity Deal," with government announcing £150 billion in US investment — later itemised as roughly £31–42 billion in direct AI and data centre pledges from Microsoft, Google, Nvidia, OpenAI and CoreWeave, on top of a claimed £44 billion the government said had "already flowed into" the UK's AI sector in preceding years. By November 2025, government said a further £24.25 billion had been committed "in the last month alone," putting the cumulative total above £78 billion. In January 2026, the government's own "AI Opportunities Action Plan: One Year On" report credited five AI Growth Zones with generating £28.2 billion in investment and "more than 15,000 jobs."
What actually happened. In April 2026, a Guardian investigation — reported on by the data centre trade press — found that a significant share of the pledged money "isn't necessarily real": some facilities counted toward the headline figures may not represent genuinely new investment, and job numbers were found to be unaccounted for or unverifiable. The single most concrete finding: the Essex site led by Nscale and backed by Nvidia — publicly promised as the location for the UK's largest sovereign AI data centre, under a signed contract to complete by 2026 — was, at the time of the investigation, still being used as a scaffolding yard. An AI industry chief executive used the phrase "phantom investments" to describe the pattern, warning it "could undermine the credibility of major programmes." Separately, and worth noting for balance: several other companies, including AstraZeneca, paused or scaled back UK investment plans around the same period, citing an "increasingly challenging" business environment — the headline AI figures were not matched by uniformly positive investment sentiment across the wider economy.
The categorisation. This is misleading in the specific, checkable sense that a named, contracted flagship project was represented publicly as on track while independent, on-the-ground reporting found it had not progressed past a construction staging area. It is omission at the level of the headline totals, which combine genuinely new capital with previously announced or reclassified investment without making that distinction easy for the public to verify — precisely the "phantom investment" problem industry figures themselves have now named.
6. "Maintain domestic food production" — while approving the policies that shrink it
The claim. In February 2024, the then Prime Minister stated: "Food security is a vital part of our national security. We must be more agile and responsive to at least maintain domestic food production at current levels." This is a lower bar than growth — simply a commitment not to let production fall further.
What actually happened. UK food self-sufficiency has continued falling, from around 65% toward 60% by value in the most recent year alone, as this platform's own Food Security pillar documents. Over the past 25 years the UK has lost 771,000 hectares of farmland — a 4.4% decline — contributing to a 12% fall in self-sufficiency for primary agricultural products; domestic food production per capita is now around 5% lower than in 2000, because population growth has outpaced production growth even as yields improved. A cross-party parliamentary group's November 2025 report warned that, without reform, production could fall by a further 32% by 2050 — and pointed directly at the contradiction: government's own targets for housing, solar energy, tree planting and biodiversity restoration could remove up to a quarter of remaining farmland, actively working against the "maintain production" commitment made in the same period. The report's authors explicitly called for food security to be treated "at the same strategic level as net zero" — a direct implication that, in practice, it currently isn't.
The categorisation. This is omission in its clearest form in this article: a specific, public commitment to maintain production was made without acknowledging that other active, simultaneous government policies — land-use, energy, and housing targets — directly undermine it. No single lie was told; no minister has denied the numbers. But a government can say it wants to maintain food production and mean it sincerely in the department that says it, while a different department's targets quietly work against it, and nobody in government is required to reconcile the two. That gap — between sincere departmental intent and incoherent whole-of-government delivery — is arguably harder to fix than an outright falsehood, because there's no single decision to reverse.
What this means for the platform's own standard
The common thread across all six cases isn't party politics — it's the gap between how a policy is first communicated and how its results are later measured, explained, or quietly left unreconciled with other government decisions. Redefining a metric after the fact (the teacher count), denying a plain practical effect while acknowledging it in the same breath (the NI pledge), rhetoric that outran delivery by eighteen months (small boats), repeating a headcount pledge four times without ever admitting the previous one failed (the civil service), publicising investment totals that outran what independent reporting could verify on the ground (AI data centres), and a sincere departmental commitment quietly undermined by other departments' targets (food production) are six different mechanisms, but they share a cause: government communicating in a way that manages perception rather than simply reporting what happened, and rarely being required to reconcile one department's promises against another's actions. This platform's own Government Quality, Accountability & Redress foundation exists because of exactly this pattern — and holds this platform to the same standard it applies here: state what's actually happening, plainly, even when the honest answer is "partly kept, partly not, and it's complicated."