Explore choices · Policy scenarios · UK · availability checked 3 October 2026

Change a tax or a pension rule. See who pays. Check the assumptions.

The Tax Lab shows what a rate change raises nationally, using HMRC’s published estimates, and what it does to one person’s tax at a chosen income; the Pensions Lab splits contributions between employees, employers and tax relief. Neither is a full analysis of every household, and the wider policy workbench stays withdrawn while its models are rebuilt.

Our purpose: Let users inspect the assumptions, limits and trade-offs in the models the site supports.

Evidence last reviewed 4 October 2026

Overview

The short version

Start with a working lab

  • Tax Lab · Calculated + Modelled

    Change a tax rate

    See the direct effect from HMRC’s ready reckoner, then the modelled behavioural range.

    Example question: What would a change in income tax, National Insurance or corporation tax rates raise?

    Direct effects come from HMRC’s published ready reckoner; behavioural responses are modelled within published ranges and labelled. Not a costing of any party’s plan.

    Open Tax Lab →

  • Pensions Lab · Calculated + Modelled

    Change a pension rule

    Follow contributions, tax relief and the State Pension through to their consequences for the system.

    Example question: What would higher pension contributions or a different State Pension rule mean?

    Contribution, tax-relief and State Pension levers traced to national totals and consequences, each step labelled by evidence type. Not personal financial advice.

    Open Pensions Lab →

Status: the numerical policy workbench is unavailable

Withdrawn for rebuilding; no relaunch date has been set. The labs above and the option comparisons on each topic page work now.

Why the workbench was withdrawn

The previous interactive dials mixed estimates from different years, definitions and modelling assumptions, and showed point estimates with no stated baseline, range or method. They have been withdrawn rather than repaired in place.

Choose a policy question

Beyond the two labs: where each question can be explored today, and what kind of answer you will get.

QuestionWhereType of outputScope and limits
How should a land tax treat farmers, or the UK tax recent arrivals?Policy design problemsQualitative comparisonDesign choices and their trade-offs, without a single revenue number.
How much might each revenue option actually raise?Revenue optionsPublished estimates comparedDesign-specific yield ranges from primary costings, with overlap notes. Not summed.
Which housing policies address which bottleneck?Housing remediesQualitative comparisonOptions matched to each step from permission to occupied home.
How could the nursing shortage be funded?Shortage professionsModelledFour funding scenarios with every assumption shown; none recommended as correct.
Which public investments give the biggest return?Investment returnsPublished estimates comparedPublished multipliers and benefit-cost ratios side by side; none derived or added together here.
What would cut NHS waiting times?NHS optionsQualitative comparisonOptions with evidence strength, owner, cost basis and what would show failure.
Shareable visual summary

Visual summary

Editorial explainer

An explanation of how to read the evidence. It shows no statistic.

No credible scenario without its assumptions.

The page withdrew unsupported point estimates. Every scenario needs a source, baseline, horizon, range and stated assumptions.

How to read it

  1. Step 1Source
  2. Step 2Baseline
  3. Step 3Horizon
  4. Step 4Range
  5. Step 5Assumptions

Editorial explainer: no statistic is shown.

Editorial explainer. No statistic is shown.
Download image (PNG, 1024 by 1536 pixels)

No credible scenario without its assumptions.

No credible scenario without its assumptions — editorial explainer from UK Facts.
Text and data version

No credible scenario without its assumptions. The page withdrew unsupported point estimates. Every scenario needs a source, baseline, horizon, range and stated assumptions.

Editorial explainer: no statistic is shown.

Image description: No credible scenario without its assumptions — editorial explainer from UK Facts.

All visual summaries, their source records and their publication state are listed on Shareable visuals.

More tools and comparisons

How outputs are labelled

Every number in the labs carries one of these labels, so arithmetic is never presented as a forecast.

Observed
Directly reported by an authoritative source. Not derived or estimated.
Calculated
Arithmetic performed on observed values — the calculation is shown and reproducible.
Modelled
Generated using stated assumptions (e.g. investment return, compounding, uprating).
Inferred
A plausible secondary consequence supported by evidence, but not deterministic.
Uncertain
Direction and/or magnitude cannot currently be estimated reliably.

What a scenario model needs before it returns here

  • A named owner for each model, and code and inputs that reproduce every output.
  • A source, date and unit for every parameter, or an explicit “illustrative” label.
  • The baseline, time horizon and a range — not only a central estimate.
  • Direct arithmetic kept separate from behavioural responses and wider economic effects.
  • Independent review before publication, and a dated record of changes.

Evidence → choices → outcomes

What the evidence can tell us

Policy comparisons on this site are reproducible and bounded, with baselines, assumptions, ranges and costs visible, and no tool published before it passes its own tests.

This domain in the UK SWOT →UK National Action Plan →

How we assessed these options

What would it take to improve this?

What would move the needle?

The outcome we want: Policy comparisons on this site are reproducible and bounded, with baselines, assumptions, ranges and costs visible, and no tool published before it passes its own tests.

The main constraint: No scenario model yet meets the publication standard — a named owner, reproducible inputs, sourced parameters, ranges and independent review — so the constraint is validation capacity and evidence, decided by this site’s editors, not the number of tools.

Constraint type: evidence, workforce.

The position today

  • What we can build on (2)
  • What is holding us back (1)The wider workbench is unavailable: its dials mixed estimates from different years and definitions with no stated baseline or range, so most policy questions have no interactive tool.
  • What could improve things (1)Validating the existing labs and linking every model to the same funding ledger as the action plans would let users compare options on one set of cost records.
  • What could make it worse (1)

Which quadrant a point sits in is our analytical judgement, not a statistical category. Each point says whether it is a measured fact, our synthesis or a risk.

What we can build on Strengths

What is holding us back Weaknesses

  • The wider workbench is unavailable: its dials mixed estimates from different years and definitions with no stated baseline or range, so most policy questions have no interactive tool.

    Our synthesisConfidence: High

    Source: HM Treasury, The Green Book (2026) (Chapters 6–9 (the standard the dials did not meet))

    What government has committed to

    Policy: The AQuA Book (2025) sets how government quality-assures analysis, including verification, validation and handling uncertainty; departments publish registers of business-critical models.

    Who and when:
    Starmer government (to July 2026), 30 July 2025.
    Since 20 July 2026:
    Not confirmed either way since the change of Prime Minister on 20 July 2026. No Burnham-era statement found.

    Checked 4 October 2026. A commitment is not delivery.

    Options (2)

What could improve things Opportunities

What could make it worse Threats

The actions worth examining

These actions concern transparency, accountability and informed choice — not a spending programme, and not a view on which party or system is right.

  1. Make policy models reproducible before expanding the workbench — not a direct public-spending programme
  2. Publish one funding ledger across all action plans — not a direct public-spending programme

  • Action 1A46Policy status: site-editorial action

    Make policy models reproducible before expanding the workbench

    Validate the Tax Lab and Pensions Lab against published figures, disclose baseline, scope, formulas and uncertainty, and add any new model one at a time only after it passes the same tests.

    The thread: issue, evidence, action

    1. Issue Weakness: The wider workbench is unavailable: its dials mixed estimates from different years and definitions with no stated baseline or range, so most policy questions have no interactive tool. Also: threat, strength
    2. Evidence Working policy labs; Withdrawn workbench models meeting the publication standard
    3. Action A46, this action
    Who decides:
    UK Facts, with qualified independent model reviewers (UK)
    Evidence:
    Expert or audit recommendation. HM Treasury’s appraisal and evaluation guidance sets the standard for separating cost, value, distribution and uncertainty; following it does not validate a model automatically.
    Expected effect:
    Direction plausible; size not established.
    Cost and funding

    Not yet costed

    Funding: Not a direct public-spending programme

    Why it is shortlisted here: Puts trustworthy models before more sliders: validate the two working labs, publish their limits, then add one model at a time.

    What it takes
    The problem

    The earlier policy dials mixed estimates from different years and definitions without baselines or ranges and were withdrawn; two bounded labs now work, and nothing yet meets the standard for a wider workbench.

    Who is affected, and where

    Users of the site’s policy tools. Territory: UK. The Tax Lab uses UK-wide HMRC figures; some taxes and rates are devolved (for example Scottish income tax), which each tool must state.

    Policy status

    UK Facts proposal. The Tax Lab and Pensions Lab work now and stay available. The wider workbench stays withdrawn until a model meets the publication standard on this page: a named owner, reproducible inputs, sourced parameters, ranges and independent review.

    The evidence
    What the research found:
    The Green Book (2026) keeps affordability for public bodies separate from social value and requires distributional analysis and adjustment for optimism bias. The Magenta Book (May 2026) asks for evaluation to be planned before implementation, with a counterfactual. The Tax Lab already reproduces HMRC’s published ready-reckoner figures in automated tests.
    Evaluated on:
    Not applicable (official guidance); UK central government appraisal and evaluation practice
    Our recommendation (separate from the research):
    Reproduce known scenarios, publish where each model cannot support policy claims, and only then consider one additional model at a time.
    Strongest counter-evidence:
    Matching official figures can lend false authority: ready reckoners are static, behavioural responses remain uncertain, and baselines go stale after every fiscal event.
    Cost and funding

    Cost completeness: Not yet costed. Funding: Not a direct public-spending programme.

    • Model validation, external review, data licensing and maintenance for the policy labs (private contribution; paid by UK Facts): Not yet costed. Site cost only; no public spending.
    Who pays:
    UK Facts: model development, external review, data licensing and maintenance.
    Financing proposed:
    UK Facts’ product and research budget.
    Gap remaining:
    Not costed.
    What delivery needs
    Skills:
    Tax and pensions modellers; independent reviewers.
    Operational capacity:
    A regression test suite for each tool.
    Infrastructure:
    Versioned baselines updated after each fiscal event.
    Permissions:
    Data licences where needed.
    Law:
    None.
    Public consent:
    None.
    First step and stages

    First step: Reproduce known scenarios and identify where each model cannot support policy claims. (owner: UK Facts)

    1. Validate the existing labs against published figures — UK Facts with an independent reviewer (proposed stage, not an official timetable)
    2. Publish each tool’s limits and baseline date — UK Facts (proposed stage, not an official timetable)
    3. Consider one further model, on the same standard — UK Facts (proposed stage, not an official timetable)

    Depends on: A47: one funding ledger, so model outputs and action costs use the same records

    Trade-offs
    Who gains and who loses:
    No direct distributional effect; the tools themselves report who gains and loses within their scope.
    What else the money or effort could do:
    Development time not spent on new tools.
    Strongest counterargument:
    Users want answers to many questions now; a slow, validated approach leaves most questions without an interactive tool.
    Risks:
    False precision, extrapolation beyond supported inputs, and stale tax and benefit baselines.
    How we would judge it
    Baseline:
    Two working labs; the wider workbench withdrawn (3 October 2026).
    Leading indicator:
    Share of lab outputs reproduced by automated tests against published figures.
    Outcome:
    Independent review sign-off for each tool, and users’ understanding of calculated versus modelled outputs.
    Continue, adapt or stop:
    Keep a tool withdrawn, or withdraw it, if it fails its regression tests or its baseline is out of date after a fiscal event.
    Sources

    Source published 5 February 2026 (Green Book); 14 May 2026 (Magenta Book); evidence reviewed 2026-10-04. Reconsider if: A fiscal event changes a tax or pension baseline; An independent review of either lab; A proposed new model meets the publication standard.

  • Action 2A47Policy status: site-editorial action

    Publish one funding ledger across all action plans

    Connect every action to unique cost and funding records with years, price basis and status, so totals are shown only when compatible and gaps stay visible.

    The thread: issue, evidence, action

    1. Issue Opportunity: Validating the existing labs and linking every model to the same funding ledger as the action plans would let users compare options on one set of cost records.
    2. Evidence Withdrawn workbench models meeting the publication standard
    3. Action A47, this action
    Who decides:
    UK Facts editorial and data team; policy owners remain responsible for their own commitments (UK)
    Evidence:
    Expert or audit recommendation. An accounting control grounded in official appraisal guidance; not an intervention with an economic return.
    Expected effect:
    Direction plausible; size not established.
    Cost and funding

    Not yet costed

    Funding: Not a direct public-spending programme

    Why it is shortlisted here: A scenario tool is only as honest as its cost records; one ledger stops the same money being counted twice across plans and tools.

    What it takes
    The problem

    Proposals often count the same money twice, add incompatible figures or treat tax proposals and growth effects as money in the bank.

    Who is affected, and where

    Readers of UK Facts, including journalists, officials and voters. Territory: UK. Covers every action on the site, whichever nation it applies to.

    Policy status

    UK Facts proposal. Implemented on this site in October 2026 as the policy-plan ledger, with automatic checks for duplicate allocation, invalid funding labels and incompatible totals.

    The evidence
    What the research found:
    The Green Book separates social value from affordability and requires attention to additionality and optimism bias; the NAO recommends realistic costs and published plans.
    Evaluated on:
    Not applicable; This website
    Our recommendation (separate from the research):
    Validate before showing any total; refresh with each Budget or policy update; keep unknowns unknown, never zero.
    Strongest counter-evidence:
    A neat ledger can imply more certainty than exists; it cannot judge whether a policy is worth doing.
    Cost and funding

    Cost completeness: Not yet costed. Funding: Not a direct public-spending programme.

    • UK Facts ledger implementation and maintenance (administration; paid by UK Facts): Not yet costed. Site budget; no public money. Policy totals are inherited from the records, never created from generic assumptions.
    Who pays:
    UK Facts (editorial and development time).
    Financing proposed:
    The site’s own budget; no public money.
    Gap remaining:
    Not applicable to public funding; site time is not separately costed.
    What delivery needs
    Skills:
    Data and editorial staff.
    Operational capacity:
    Refresh after each fiscal event.
    Infrastructure:
    The site’s validation suite.
    Permissions:
    None.
    Law:
    None.
    Public consent:
    None.
    First step and stages

    First step: Record periods, price bases, amounts or ranges, allocation, conditions and existing uses for every cost on the site. (owner: UK Facts data team)

    1. Validate records before publishing totals — UK Facts
    2. Refresh after each Budget or Spending Review — UK Facts (proposed stage, not an official timetable)

    Depends on: Every domain action’s cost records

    Trade-offs
    Who gains and who loses:
    Readers gain clearer information; no direct distributional effect.
    What else the money or effort could do:
    Editorial time not spent on other pages.
    Strongest counterargument:
    Official bodies (OBR, departments) already publish costings; a site ledger can duplicate them.
    Risks:
    Spurious certainty; missing data shown as zero (guarded against by validation).
    How we would judge it
    Baseline:
    Duplicate allocations and unmatched years flagged by the validator.
    Leading indicator:
    Share of actions with complete cost and funding records.
    Outcome:
    Traceable evidence for every figure; unresolved gaps visible.
    Continue, adapt or stop:
    Never publish a combined total while any material cost is unknown.
    Sources

    Source published February 2026 (Green Book); October 2024 (NAO); evidence reviewed 2026-10-04. Reconsider if: A Budget or Spending Review; A new domain plan is added.

    Also on: Public finances — one shared record, so its cost is counted once.

Decision brief

  • Open the policy scenarios page with three guided questions that anyone can reproduce

The leading option set out as a decision someone could take: who decides, who benefits, the best evidence, cost and funding, and the main trade-off. The full research record is one click away.

Decision briefSite-maintenance taskReviewed 2026-10-04

Open the policy scenarios page with three guided questions that anyone can reproduce

The thread: issue, evidence, action

  1. Issue Weakness: The wider workbench is unavailable: its dials mixed estimates from different years and definitions with no stated baseline or range, so most policy questions have no interactive tool. Also: threat, strength
  2. Evidence Working policy labs; Withdrawn workbench models meeting the publication standard
  3. Action Make policy models reproducible before expanding the workbench
Decision required
UK Facts puts three worked questions at the top of the page — one tax question and two pension questions — each with its baseline, what the answer means, what it does not mean and a link that reproduces it in the Tax Lab or Pensions Lab.
Who and where
Users of the policy scenarios page and the two labs. Applies in: UK.
Accountable body
UK Facts editors
Intervention
Give each guided question a fixed preset, a dated baseline and an output reading.
Best evidence, and where it applies
The Tax Lab’s automated tests already reproduce HMRC’s published ready-reckoner figures; the Green Book and Magenta Book set the standard for stating baselines, ranges and who gains and loses.Studied in: UK central government appraisal (Not an evaluation: official appraisal guidance).
Extra cost of this change
Capital: none identified in the brief.Recurring: Validation, independent review and baseline refresh for the two labs — Not yet costed.
Funding
Not funded yet: a possible funding route exists, but nothing is allocated to this change.Committed to this change: none.Possible source, not an award: none identified.Unfunded: Staff time and review fees are not yet costed.
Savings and wider benefits
Not yet specified: no cashable saving or avoided cost has been costed for this change. Expected outcomes are listed one by one in the research appendix and are never added together or counted as money available.
First 90-day step
The three guided questions live at the top of /policy-impact, each passing an automated reproduction testThe proposed first milestone; a 90-day date is not yet specified. Owner: UK Facts developers. Proposed by UK Facts, not an official deadline.
Twelve-month milestone
Not yet specified.Time to a first useful result: One release cycle; baselines then refreshed after the 28 October 2026 Budget.
Outcome, and when to scale or stop
Measure: All three guided outputs pass their reproduction tests at each release.Scale if: The tests pass and readers can say what each output does and does not mean.Stop if: A guided output fails its test or its baseline is out of date after a fiscal event — the question is withdrawn until fixed.
Main trade-off
Three bounded questions answer less than a full workbench, but every answer can be checked; the withdrawn workbench mixed incompatible estimates.
Source and review date
The Green Book (2026) (2026). Reviewed 2026-10-04.

An editorial proposal by UK Facts for discussion, not government policy. Records: A46 · costs C-A46-REVIEW

Research appendix: study design, effect sizes and limits, alternatives, full costing, international comparisons and how we would know

1.The change, in plain words

Open the policy scenarios page with three guided questions that anyone can reproduce

2.The decision requested

Decision
Publish three guided questions (below), each tied to a lab preset, a dated baseline and a scope warning, and move the validation record and test results to the methods section of the page.
Accountable institution
UK Facts editors
Geography
UK
Powers
An editorial and product decision for this website. No public body or budget is involved.
Delivery partners
An independent model reviewer; HMRC and OBR, as publishers of the baselines the labs reproduce

3.Who it is for, and what changes beyond today

Target population
Users of the policy scenarios page and the two labs.
Eligibility
Questions the labs can already answer within their published scope: (1) what 1p on the basic rate of income tax raises, before and after behaviour (Tax Lab, HMRC ready reckoner baseline); (2) what doubling the employer auto-enrolment minimum from 3% to 6% means for savers and employers (Pensions Lab “Employers pay more” preset); (3) how State Pension spending in 2075/76 differs between the triple lock and earnings-linked uprating (Pensions Lab, OBR long-term projection).
Scale
Three questions; no new model until one passes the publication standard.
Baseline
The wider workbench is unavailable (withdrawn for rebuilding; no relaunch date has been set); two labs work. Baselines: HMRC ready reckoner (June 2025 edition); auto-enrolment minimums of 5% employee and 3% employer on qualifying earnings; OBR long-term State Pension projections (July 2026).

Source: HM Revenue & Customs, Direct effects of illustrative tax changes (bulletin) (Ready reckoner tables); Office for Budget Responsibility, Fiscal risks and sustainability (July 2026) (Long-term projections of State Pension spending)

Change beyond existing law or funding
The page currently lists where questions can be answered. The change adds three fully worked examples with a reproducible link, a stated baseline and price year, an “outside the model’s scope” warning, and puts the test results in the methods.

4.How it would work

Do: Give each guided question a fixed preset, a dated baseline and an output reading Changing: so that a reader gets the same answer every time and can see which parts are calculated and which modelled

  • Each guided output is reproduced by an automated test

    Hypothesis to test

    A quality measure for this site, not a policy effect.

  • Readers correctly tell a calculated figure from a modelled one

    Hypothesis to test

    To be checked in reader comprehension sessions; no effect size is claimed.

Outcomes are kept separate and never added into one impact figure.

5.Best evidence

  • The Green Book (2026)

    Author or institution:
    HM Treasury
    Published:
    2026
    Evaluation period:
    Not an evaluation: official appraisal guidance
    Design:
    Guidance
    Sample or population:
    UK central government appraisal
    Effect and uncertainty:
    Requires affordability to be kept separate from social value, distributional analysis and adjustment for optimism bias.
    Limits:
    Sets a standard; does not test whether a website tool is understood.

    Where: Chapters 6–9; affordability criteria; chapter 7 (distributional analysis). Affiliation is provenance, not proof of quality.

6.Alternatives, counter-evidence and who loses

Closest feasible alternative
Restore a wider workbench of sliders now, with caveats on each output.
Strongest counter-evidence
Matching official figures can lend false authority: ready reckoners are static or use fixed behavioural assumptions, and baselines go stale after each fiscal event.

Source: HM Treasury, The Green Book (2026) (Optimism bias; uncertainty)

Distributional effects
None from the tool itself; the tax and pension questions report who pays and who gains within their scope.
Main downside
Three questions leave many reader questions unanswered until another model passes review.

The options compared, including current policy

  • The proposal

    Three guided, reproducible questions with baselines and scope warnings

    Who is reached
    Every visitor to /policy-impact
    Evidence
    Lab tests reproduce published figures; official appraisal guidance
    Likely outcome and timing
    Usable answers in one release cycle
    Cost and funding
    Site time, not yet costed
    Main trade-off
    Narrow scope

  • Current policy

    Current page: a list of where questions can be answered

    Who is reached
    Readers who click through to a lab
    Evidence
    Labs work; no worked example on the page
    Likely outcome and timing
    No change
    Cost and funding
    Existing site time
    Main trade-off
    Readers must learn the tools before getting an answer

  • Alternative

    Restore the wider workbench with caveats

    Who is reached
    All visitors
    Evidence
    The previous dials mixed incompatible years and definitions
    Likely outcome and timing
    Fast, but outputs not reproducible
    Cost and funding
    Development time
    Main trade-off
    False precision

8.What delivery needs first

  • data — Baselines refreshed after each fiscal event (the next Budget is scheduled for 28 October 2026); an out-of-date baseline withdraws the question. (Requirement; owner: UK Facts developers)
  • staff — An independent reviewer for each lab before its guided question is labelled reviewed. (Scarce resource shared with other priorities; owner: UK Facts editors)

9.What it would cost

How the cost is built up: Developer and editor time to build three presets and tests + independent review fee + refresh after each fiscal event. Site cost only.

  • Validation, independent review and baseline refresh for the two labs

    Tax Lab and Pensions Lab; three guided questions

    Not yet costed (ledger C-A46-REVIEW)

    recurring; price year not stated; After each fiscal event; Not yet established

10.Who pays, and what is missing

Programme envelope (a possible route, not an award)
None identified
Allocation committed to this change
None
Remaining gap
Staff time and review fees are not yet costed.
Proposed financing
UK Facts’ product and research budget; no public money.
Responsible payer
UK Facts
Budget displaced
Time not spent on new pages or a further model.

11.First milestone and time to a useful result

The three guided questions live at the top of /policy-impact, each passing an automated reproduction test (Owner: UK Facts developers. Proposed by UK Facts, not an official deadline.)

Time to first useful result: One release cycle; baselines then refreshed after the 28 October 2026 Budget.

What could delay it: The Budget changes the tax baseline and the ready reckoner is not yet republished An independent reviewer is available

12.How we would know

Baseline
Two working labs; no guided questions; workbench withdrawn.
Comparison
Published HMRC and OBR figures the labs reproduce.
Leading indicator
All three guided outputs pass their reproduction tests at each release.
Follow-up
After each fiscal event.
Continue if
The tests pass and readers can say what each output does and does not mean.
Adapt if
Readers misread modelled results as calculated ones; the wording or layout changes.
Stop if
A guided output fails its test or its baseline is out of date after a fiscal event — the question is withdrawn until fixed.
Thresholds
Not yet set: Pass/fail is defined by reproduction of the published figure; no numerical comprehension target is set before testing.

What is not yet known, and how to find out

  • What validation, review and refresh will cost each year (cost)Next step: Record developer time for the first three presets and obtain a reviewer quote.Scope: Two labs, three guided questions · Data or quotes needed: Hours logged; reviewer quote · Who can estimate it: UK Facts
  • Whether readers understand the calculated/modelled distinction (effect)Next step: Test the three questions in the planned reader sessions.Defensible route: bounded pilot

Sources last checked 4 October 2026 · Owner: Site methods editor (James Harris) · Next check: the first trigger: A change to the policy scenarios models or their inputs.

What it would take

The order things have to happen in, who pays, and the effects and appraisals on record.

The order things have to happen in

  1. First, the prerequisite

    Reproduce known scenarios in both labs and list what each cannot support

    Owner: UK Facts with an independent reviewer

    No date: the timing depends on the steps before it.

  2. Then, the decision

    Decide whether any withdrawn model meets the publication standard

    Owner: UK Facts editors

    No date: the timing depends on the steps before it.

  3. Then, delivery

    Link tool outputs to the shared funding ledger; add at most one further model at a time

    Owner: UK Facts

    No date: the timing depends on the steps before it.

  4. Finally, the measured outcome

    Every published output reproducible by tests; baselines dated after each fiscal event

    Owner: UK Facts

    No date: the timing depends on the steps before it.

Who pays, and what is still missing

Cost, committed funding and the remaining gap for the same scope and years. A programme envelope is not an award to a project; a proposal is not committed money. None of these costs has been established yet, so no total is shown.

Cost and funding ledger

  • Model validation, external review, data licensing and maintenance for the policy labs

    A46 · private contribution · not costed

    Amount
    Not yet costed
    Who pays
    UK Facts
    Committed
    None identified
    Proposed, not committed
    —
    Gap
    Not established

  • UK Facts ledger implementation and maintenance

    A47 · administration · not costed

    Amount
    Not yet costed
    Who pays
    UK Facts
    Committed
    None identified
    Proposed, not committed
    —
    Gap
    Not established

“Not yet costed” means no defensible cost has been found — not that the action is free. Lifetime benefits, benefit-cost ratios and forecast savings are never shown here as money available to spend.

What the evidence says each action changes

No effect has been established for these actions yet: direction plausible; size not established.

How we would know

Working policy labs; Withdrawn workbench models meeting the publication standard.

The outcome measures to watch. A baseline is where we start, not a target; where there is no official target we say so.

Where the ideas come from

Research, counter-evidence and methods (4 sources; 1 claim held back)

Counter-evidence and criticism

  • Make policy models reproducible before expanding the workbench: Matching official figures can lend false authority: ready reckoners are static, behavioural responses remain uncertain, and baselines go stale after every fiscal event.
  • Publish one funding ledger across all action plans: A neat ledger can imply more certainty than exists; it cannot judge whether a policy is worth doing. (HM Treasury)

Claims held back

  • Any revenue or cost figure from the withdrawn workbench — Its estimates mixed years, definitions and assumptions with no baseline or range. Needed: A rebuilt model that meets every point of the publication standard.

“Full text checked” means we read the report itself; “relevant passages” means the sections behind each figure; “summary or abstract only” means we could not read the full study, so no number from it is published here without another source.

Plan reviewed 2026-10-04. How we judge evidence.

How this connects

Sources and methods (2)

Method notes

  • A tool is listed here only if it works now and labels the type of each output.
  • The numerical workbench will return only when a model meets every point of the publication standard above.