← Back to Benefits

Benefits: evidence and policy appraisals

Detailed research records, alternatives, costs and provisional assessments. These records preserve different scopes and historical reviews; they are not an additional ranked action plan.

Evidence → choices → outcomes

What could improve it?

Support in Great Britain that reaches the people entitled to it, at a level tested against the cost of essentials, paid accurately, with suitable help into work for those who can work.

These options link to their national assessment, including funding, delivery constraints and remaining uncertainties.

  • Assessed option · provisional

    Claim what people are owed

    Assisted applications for people likely to be entitled to Pension Credit.

    8/10 criteria scored

    Coverage of the appraisal, not an impact score.

    Delivery: Amber

    Whole-life costs incomplete

    Evidence and delivery constraints

    Unclaimed entitlement establishes a gap; effect of this outreach design is untested.

    Specify eligible cohort, trusted outreach, application support and comparison group; arrange benefits-processing capacity.

    Equality and opportunity listEvidence and assessment

  • Assessed option · provisional

    A child-poverty target and essentials test

    7/10 criteria scored

    Coverage of the appraisal, not an impact score.

    Delivery: Amber

    Whole-life costs incomplete

    Evidence and delivery constraints

    A target does not itself reduce poverty. Modelled effects of a strategy do not establish the effect of the benchmark.

    Define poverty/essentials measures, then compare specific benefit, housing-cost and service changes; assign accountability.

    Equality and opportunity listEvidence and assessment

  • Assessed option · provisional

    Offer evidence-based employment support alongside healthcare

    Fidelity reviews, NHS integration and an impact evaluation inside Connect to Work.

    9/10 criteria scored

    Coverage of the appraisal, not an impact score.

    Delivery: Amber

    Whole-life costs incomplete

    Evidence and delivery constraints

    Trials exist, but effects differ by setting; pooled out-of-work groups showed no employment impact. Do not infer a uniform national growth effect.

    Specify cohort and voluntary offer; secure clinical referrals, trained employment specialists, employer links and linked-data evaluation.

    Growth and prosperity listEvidence and assessment

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

How we assessed these options

What would it take to improve this?

Detailed policy appraisal

The outcome we want: Support in Great Britain that reaches the people entitled to it, at a level tested against the cost of essentials, paid accurately, with suitable help into work for those who can work.

The main constraint: Most spending goes to pensioners and disabled people, and take-up, adequacy and work support each trade off against cost and work incentives; HM Treasury decides what higher legitimate spending is affordable, while DWP’s processing capacity and data-sharing powers limit how fast take-up can rise.

Constraint type: money, workforce, permissions.

The options on the table

What is funded now, and the changes assessed against it — the current baseline and the alternatives, before the position today and the actions.

Provisional single-reviewer assessments from the 6 October 2026 review; not approved business cases, and not government policy.

  • Claim what people are owedDelivery risk: Amber

    Funded now
    Pension Credit and DWP take-up campaigns already exist; a targeted, evaluated outreach model is a UK Facts proposal.
    The change assessed
    Assisted applications for people likely to be entitled to Pension Credit.

    Range if completed: 16.0–22.0 out of 30 (equal weights; not yet assessed: expected incremental benefit (Bn), incremental affordability (Co))

  • A child-poverty target and essentials testDelivery risk: Amber

    Funded now
    Not yet stated in the record.
    The change assessed
    Not yet defined.

    Range if completed: 12.0–21.0 out of 30 (equal weights; not yet assessed: increment beyond baseline (Ad), expected incremental benefit (Bn), incremental affordability (Co))

Assessed with another page as home

These options also bear on this topic. The full assessment sits on the home page named under each.

  • Offer evidence-based employment support alongside healthcareDelivery risk: Amber

    Home: Jobs & pay.

    Funded now
    Connect to Work has taken referrals since April 2025 and aims to support about 300,000 people by 2029/30; IPS also operates in NHS mental health services. The proposal here is to hold it to fidelity and integrate it with care, not to start something new.
    The change assessed
    Fidelity reviews, NHS integration and an impact evaluation inside Connect to Work.

    Range if completed: 17.0–20.0 out of 30 (equal weights; not yet assessed: incremental affordability (Co))

  • Prevent avoidable loss of rented homesDelivery risk: Amber

    Home: Housing.

    Funded now
    Funded baseline: the 56-day prevention duty (extended to section 8 notices from 1 May 2026), the Homelessness, Rough Sleeping and Domestic Abuse Grant (£885.0m in 2026–27), the Housing Loss Prevention Advice Service and a £16m Legal Support Grant (October 2026 to March 2029). Testing a defined early-advice offer against comparison areas is a UK Facts proposal, not government policy.
    The change assessed
    Test early rent-arrears/legal advice and targeted assistance for a defined at-risk cohort before homelessness occurs.

    Range if completed: 16.0–22.0 out of 30 (equal weights; not yet assessed: expected incremental benefit (Bn), incremental affordability (Co))

Alternatives already on record: 16 options listed under the points in the position today (6 without an action record), and 4 earlier options under other options compared. They have not been scored in the review.

The position today

  • What we can build on (1)
  • What is holding us back (4)PAC reported that incorrect earnings-averaging guidance had led to an estimated 26,000 carers being wrongly recorded as overpaid. DWP began reassessing affected cases in April 2026; that start does not establish completed redress. (PAC February 2026; DWP April 2026 · Great Britain)
  • What could improve things (1)Raising take-up of existing entitlements, testing whether the basic rate covers essentials and offering suitable employment support address different failures in the same system.
  • 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

What could improve things Opportunities

What could make it worse Threats

The actions worth examining

Our shortlist for this page. Research quality, policy status, cost and funding are judged separately; none of these is government policy unless its status says so.Our shortlist. None is government policy unless its status says so.

  1. Claim what people are owed — not yet costed / funding not identified
  2. Set an explicit test for whether benefits cover essentials — not yet costed / funding not identified
  3. Offer evidence-based employment support alongside healthcare — indicative costing / partial funding

  • Action 1A25Policy status: site-editorial action

    Claim what people are owed

    Use lawful data matching and assisted applications to reach people likely to be entitled, then verify and pay their claims.

    The thread: issue, evidence, action

    1. Issue Weakness: Only 62% of people entitled to Pension Credit claimed it; up to 900,000 entitled families did not, leaving up to £2.4bn unclaimed in 2024–25 (one year’s estimate). Also: opportunity, strength
    2. Evidence Pension Credit take-up (caseload)
    3. Action A25, this action

    Assessment (draft, 6 October review)

    Who decides:
    DWP, councils and advice providers (Great Britain)
    Evidence:
    Untested proposal. The size of unclaimed entitlement is measured; what outreach would achieve has not been evaluated here.
    Expected effect:
    Direction plausible; size not established.
    Cost and funding

    up to £2.5bn a year, 2023–24, cash (nominal) prices; 1 further cost line, some not yet costed

    Funding: Not yet costed / funding not identified

    Why it is shortlisted here: The take-up gap is measured and large; people are missing support they are already entitled to. Success raises legitimate spending, which must be budgeted.

    What it takes
    The problem

    Hundreds of thousands of pensioner families entitled to Pension Credit do not claim it.

    Who is affected, and where

    Pensioner families in Great Britain entitled to Pension Credit who do not claim. Territory: Great Britain. Northern Ireland administers its own social security; not assessed.

    Policy status

    UK Facts proposal. Pension Credit and DWP take-up campaigns already exist; a targeted, evaluated outreach model is a UK Facts proposal.

    The evidence
    What the research found:
    62% of entitled people claimed Pension Credit in 2023–24; up to 910,000 families did not claim, and up to £2.5bn went unclaimed — about £2,600 a year per family.
    Evaluated on:
    Pensioner families in the Family Resources Survey; Great Britain, 2023–24
    Our recommendation (separate from the research):
    Test targeted outreach with an evaluation before claiming any take-up effect, and budget for the extra entitlement it would pay.
    Strongest counter-evidence:
    Survey-based estimates can misclassify households as entitled, so a statistical estimate is not individual eligibility.
    Effects on record
    • Income of eligible pensioners who start claiming: Direction plausible; size not established. plausible spillover, limited confidence; Great Britain. An unclaimed family misses about £2,600 a year on average, but how many more would claim after outreach has not been evaluated.
    Cost and funding

    Cost completeness: Partly costed. Funding: Not yet costed / funding not identified.

    • Extra Pension Credit paid if more entitled people claim (household transfer; paid by central government): up to £2.5bn a year, 2023–24, cash (nominal) prices. Up to £2.4bn went unclaimed in 2024–25 (up to £2.5bn in 2023–24): the ceiling if every entitled family claimed. Successful take-up raises spending; a realistic increase depends on what outreach achieves, which is not established.
    • Outreach, data matching and assisted applications (administration; paid by central government, councils): Not yet costed. Administration by DWP, councils and advice services.
    Who pays:
    Central government: extra Pension Credit payments and administration; councils and advice services for assisted applications.
    Financing proposed:
    Benefit spending is demand-led; outreach needs an administration budget.
    Gap remaining:
    Outreach is not costed; extra entitlement could be up to £2.5bn a year if every entitled family claimed.
    What delivery needs
    Skills:
    Advisers trained in Pension Credit.
    Operational capacity:
    DWP processing.
    Infrastructure:
    Data-matching systems.
    Permissions:
    Lawful data-sharing gateways.
    Law:
    Existing powers; check data-sharing basis.
    Public consent:
    Older people’s willingness to be contacted.
    First step and stages

    First step: Identify likely unmet entitlement without treating a statistical estimate as individual eligibility. (owner: DWP with councils)

    1. Test outreach in defined areas — DWP and councils (proposed stage, not an official timetable)
    2. Verify awards and measure retention — DWP (proposed stage, not an official timetable)
    Trade-offs
    Who gains and who loses:
    Gains for low-income pensioners; higher spending for the Exchequer.
    What else the money or effort could do:
    Administration and benefit spending compete with other priorities.
    Strongest counterargument:
    Successful take-up raises spending; a campaign that does not reach the right people wastes money.
    Risks:
    Privacy, digital exclusion and contacting people who are not entitled.
    How we would judge it
    Baseline:
    Pension Credit take-up (DWP), 62% by caseload in 2023–24.
    Leading indicator:
    Eligible claimants reached and successful awards.
    Outcome:
    Take-up by caseload and expenditure; income gains; processing time.
    Continue, adapt or stop:
    Continue where awards per contact beat a comparison area; stop otherwise.
    Earlier option on this page

    This card replaces the option “Raise take-up of support people are already entitled to” (operating: Campaigns operate; automatic award is an external proposal, not government policy.). Its evidence: DWP measures the gap (62% caseload take-up, up to £2.4bn unclaimed in 2024–25) but has not published an evaluation showing which outreach methods raise take-up, or by how much. Automatic award has not been tested in the UK for Pension Credit. DWP, Income-related benefits: estimates of take-up, financial year ending 2025

    Sources

    Source published 30 October 2025; evidence reviewed 2026-10-04. Reconsider if: The next take-up release; An evaluation of a Pension Credit campaign.

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

  • Action 2A16Policy status: external proposal

    Set an explicit test for whether benefits cover essentials

    An independent advisory body, on the model of the Low Pay Commission, would publish evidence each year on the cost of essentials for set household types and advise on the standard allowance and deductions; ministers and Parliament would still decide.

    The thread: issue, evidence, action

    1. Issue Weakness (Inequality): 27.4% of children (4.0 million) were in relative poverty after housing costs, and 13% were in deep material poverty, lacking at least four of thirteen essentials. Also: threat, opportunityHow acting here reaches it: A published test of whether the basic benefit rate covers essentials sets payments against the cost of the items that deep material poverty counts children as lacking.
    2. Evidence Children in deep material poverty
    3. Action A16, this action

    Assessment (draft, 6 October review)

    Who decides:
    DWP and HM Treasury (rates and remit), Parliament (any legislation) (Great Britain)
    Evidence:
    Untested proposal. A policy proposal drawing on UK and overseas advisory bodies; no evaluation shows that such a body changes hardship, and the essentials estimates are the proposers’ own modelling.
    Expected effect:
    Direction plausible; size not established.
    Cost and funding

    Not yet costed

    Funding: Not yet costed / funding not identified

    Why it is shortlisted here: There is no published test of whether the basic rate covers essentials; one would make the adequacy choice and its cost explicit.

    What it takes
    The problem

    No official benchmark says what the basic rate of Universal Credit is meant to cover; it is set by ministers at the annual uprating review and, for 2026–2030, by a fixed path in legislation, without a published comparison with the cost of essentials.

    Who is affected, and where

    Households on Universal Credit in Great Britain, and people entitled to it who do not claim. Territory: Great Britain. Universal Credit is legislated for Great Britain; Northern Ireland runs a parallel system and Scotland pays some devolved benefits on top (for example the Scottish Child Payment). Equivalent decisions there are not assessed here.

    Policy status

    Proposed by researchers or others, not government policy. A Joseph Rowntree Foundation proposal, not government policy. The Universal Credit Act 2025 (Royal Assent 3 September 2025) already sets the standard allowance on a path that ends 4.8% above inflation-only uprating by 2029–30; that is a fixed path chosen by Parliament, not a test of adequacy, and it is the baseline any option must be compared with.

    The evidence
    What the research found:
    JRF reports that the standard allowance from April 2026 (about £98 a week for a single person aged 25 or over, £154 for a couple) is below its destitution thresholds (£105 and £160) and below the JRF/Trussell estimate of essentials (£120 and £205). It proposes an independent advisory process but does not cost the body or any change in rates.
    Evaluated on:
    Not evaluated. JRF compares 2026–27 Universal Credit rates with its own and Trussell’s estimates of essential costs for a single adult and a couple.; United Kingdom, 2026–27 rates
    Transfer to this case:
    The Low Pay Commission and New Zealand’s advisory group answer different questions; their record is not evidence that a benefits body would reduce hardship.
    Our recommendation (separate from the research):
    Specify the essentials basket, household types and decision rule first, then publish a distributional costing of named options against the 2025 Act baseline. UK Facts does not set its own adequacy threshold; the JRF/Trussell figures are their estimates.
    Strongest counter-evidence:
    An essentials basket is a normative choice, and a published benchmark creates pressure for large recurring spending that must be financed by tax, cuts elsewhere or borrowing; higher basic rates also change work incentives at the margin. JRF itself does not cost the change.
    Effects on record
    • Material hardship among households on Universal Credit: Direction plausible; size not established. plausible spillover, limited confidence; Great Britain. A benchmark changes nothing unless rates change; how much hardship would fall depends on the option, take-up and housing costs, none of which has been modelled here.
    • Recurring benefit spending: A cost or harm; size not established. plausible spillover, limited confidence; Great Britain. Shown as a trade-off, not a benefit. The size depends on the option and has not been costed on the current baseline.
    Cost and funding

    Cost completeness: Not yet costed. Funding: Not yet costed / funding not identified.

    • An independent advisory body on benefit adequacy (secretariat, research, consultation) (administration; paid by central government): Not yet costed. JRF does not cost the body. Comparable bodies (the Low Pay Commission, pay review bodies) are small, but their budgets are not a price for this one.
    • Any change in Universal Credit rates or deductions following the benchmark (household transfer; paid by central government): Not yet costed. Must be modelled on the current caseload, take-up and interactions with housing support, against the Universal Credit Act 2025 baseline. No older campaign price tag is reused.
    Who pays:
    Taxpayers, through higher benefit spending (or cuts elsewhere, or borrowing); DWP for the small administrative cost of an advisory body.
    Financing proposed:
    None. Any rate change would need a recurring tax, spending or borrowing decision matched to a costing on the current caseload and take-up.
    Gap remaining:
    Unknown: it depends entirely on the option chosen. The advisory body is a small administrative cost; rate changes could be large and recurring.
    What delivery needs
    Skills:
    Analysts able to price essentials by household type.
    Operational capacity:
    DWP and HM Treasury modelling of caseloads, take-up and interactions with housing support.
    Infrastructure:
    None beyond existing payment systems.
    Permissions:
    None.
    Law:
    A statutory body or duty would need legislation; a non-statutory panel would not.
    Public consent:
    Agreement on what counts as an essential; public attitudes to benefit spending.
    First step and stages

    First step: Specify the essentials basket, household types and the exact options to cost against the 2025 Act baseline. (owner: DWP with HM Treasury)

    1. Consult on remit, basket and household types — DWP (proposed stage, not an official timetable)
    2. Set up the advisory body and publish a first benchmark — DWP; Parliament if statutory (proposed stage, not an official timetable)
    3. Annual report, with a published government response at the uprating review — Advisory body; DWP ministers (proposed stage, not an official timetable)

    Depends on: A25: higher rates help only people who claim; Housing support (Local Housing Allowance) interacts with any change

    Trade-offs
    Who gains and who loses:
    Gains for households on Universal Credit, largest for those with the least other income; costs fall on taxpayers or on whatever spending is cut. Rates for pensioners and disability benefits are outside this test unless its remit is widened.
    What else the money or effort could do:
    Money for higher rates is not available for services, employment support or debt reduction.
    Strongest counterargument:
    A benchmark turns a political judgement into an apparently technical one and builds in pressure for spending that may not be affordable; uprating already follows a legislated path to 2029–30.
    Risks:
    Basket choices contested as arbitrary; recommendations ignored, eroding trust; higher rates absorbed by rents where housing support is capped.
    How we would judge it
    Baseline:
    Deep material poverty and food insecurity among people in low-income households (DWP Households Below Average Income).
    Leading indicator:
    The published gap between each household type’s rate and the benchmark.
    Outcome:
    Material hardship, food insecurity and income after essential costs; employment and caseload effects alongside.
    Continue, adapt or stop:
    Review the remit if the benchmark is not used in uprating decisions, or if hardship does not change after rates move towards it.
    Sources

    Source published 30 January 2026 (JRF); 2025 (Universal Credit Act); evidence reviewed 2026-10-04. Reconsider if: A government response to an independent adequacy proposal; The next uprating review; A costed JRF/Trussell essentials-guarantee update.

    Also on: Inequality, Inflation — one shared record, so its cost is counted once.

  • Action 3A09Policy status: funded programme

    Offer evidence-based employment support alongside healthcare

    Individual Placement and Support puts an employment specialist alongside health services to find matched work quickly and keep supporting the person once they start.

    The thread: issue, evidence, action

    1. Issue Weakness (Jobs & pay): 2.74 million people aged 16–64 were economically inactive because of long-term sickness, against 2.08 million at the end of 2019. Also: threat, opportunityHow acting here reaches it: Many people on health-related benefits want to work; voluntary, evidence-based employment support moves some into work, which the long-term sickness inactivity count on the employment page records.
    2. Evidence Economically inactive because of long-term sickness, aged 16–64
    3. Action A09, this action

    Assessment (draft, 6 October review)

    Who decides:
    DWP and local authority delivery areas (Connect to Work), with NHS commissioners (England and Wales)
    Evidence:
    Experimental or quasi-experimental evaluation. Randomised trials of an adapted model found modest effects that varied by place; evidence is strongest for severe mental illness, a different group.
    Expected effect:
    +4 percentage points (22% vs 18%): working 13 or more weeks in the year after joining (out-of-work participants) (West Midlands Combined Authority). No effect for out-of-work participants in Sheffield City Region, and no effect on earnings at 12 months. 2 more results in “What it takes”.
    Cost and funding

    £1.03bn in total, 2025–26 to 2028–29, cash (nominal) prices; 2 further cost lines, some not yet costed

    Funding: Indicative costing / partial funding

    Why it is shortlisted here: Many people on health-related benefits want to work; evidence-based, voluntary support is already funded through Connect to Work and should be judged on sustained work.

    Builds on the option Employment support for disabled people and people with health conditions compared below.

    What it takes
    The problem

    People with health conditions who want to work often get employment help that is separate from their care and does not stay with them in work.

    Who is affected, and where

    Adults with health conditions or disabilities who want to work, in England and Wales. Territory: England and Wales. Connect to Work covers England and Wales; Scotland and Northern Ireland run their own employment support.

    Policy status

    A funded programme already exists. Connect to Work has taken referrals since April 2025 and aims to support about 300,000 people by 2029/30; IPS also operates in NHS mental health services. The proposal here is to hold it to fidelity and integrate it with care, not to start something new.

    The evidence
    What the research found:
    Working 13+ weeks in a year: +4 percentage points for out-of-work participants in the West Midlands; +3 points for people in work in Sheffield City Region (weaker evidence); no effect for out-of-work participants there; no earnings effect at 12 months; small health gains in Sheffield City Region only.
    Evaluated on:
    9,785 adults with mild to moderate health conditions, in or out of work; Sheffield City Region and West Midlands, 2018–2020
    Transfer to this case:
    Results for severe mental illness or for one area should not be assumed for all economically inactive people.
    Our recommendation (separate from the research):
    Keep delivery at fidelity, integrate it with NHS care, and judge it by sustained work and health, not referrals or starts.
    Strongest counter-evidence:
    One of three trial groups saw no employment effect and the Exchequer return within a year was 1–2p per £1; benefits were mainly in health and wellbeing, and only in one area.
    Effects on record
    • Working 13 or more weeks in the year after joining (out-of-work participants): +4 percentage points (22% vs 18%) (randomised control group receiving usual support; 12 months after randomisation, 2018–2020; significant at the 99% level; no interval published). evaluated, moderate confidence; West Midlands Combined Authority. No effect for out-of-work participants in Sheffield City Region, and no effect on earnings at 12 months.
    • Working 13 or more weeks in the year after joining (people in work but struggling): +3 percentage points (randomised control group; 12 months after randomisation, 2018–2020; significant only at the 90% level; no interval published). evaluated, limited confidence; Sheffield City Region. Weaker evidence than the West Midlands result.
    • Self-reported health (EQ-5D-5L) at 12 months: +0.10 standard deviations (randomised control group; 12 months after randomisation; significant at the 90% level (99% when Sheffield groups are pooled); no effect in the West Midlands). evaluated, limited confidence; Sheffield City Region. Small; not found in the West Midlands. Not evidence of reduced NHS demand.
    Cost and funding

    Cost completeness: Partly costed. Funding: Indicative costing / partial funding.

    • Connect to Work grant and programme costs, 2025–26 to 2028–29 (recurring spending; paid by central government): £1.03bn in total, 2025–26 to 2028–29, cash (nominal) prices. Sum of the four years in the revised business case. Excludes £15.5m spent in 2023–24 and 2024–25.
    • Connect to Work costs, 2029–30 to 2031–32 (recurring spending; paid by central government): £134.4m in total, 2029–30 to 2031–32, cash (nominal) prices. Sum of the last three years in the revised business case.
    • NHS clinical integration, referral and supervision for IPS (recurring spending; paid by NHS): Not yet costed. Not in the DWP business case. An older national IPS price is not a current rollout budget.
    • Autumn Budget 2024 and Spending Review 2025 settlements for Connect to Work: £1.09bn, 2025–26 to 2028–29 — committed (allocated to this scope). £114m for 2025–26 plus £978m for 2026–27 to 2028–29. The business case says funding and costs differ because volume forecasts changed.
    • Connect to Work funding to be requested at a future Spending Review: £138m, 2029–30 to 2031–32 — not committed (no allocation). “Will be requested in a future Spending Review.”
    Who pays:
    DWP for Connect to Work grants; the NHS for clinical integration and IPS in mental health services.
    Financing proposed:
    £138m for 2029–30 onwards is to be requested at a future Spending Review.
    Gap remaining:
    Connect to Work is funded to 2028–29; the later years are not committed and the NHS-side integration cost is not established.
    What delivery needs
    Skills:
    Trained employment specialists and supervisors.
    Operational capacity:
    NHS referral routes and fidelity reviews.
    Infrastructure:
    None beyond services.
    Permissions:
    Data sharing between health and employment services.
    Law:
    No new law.
    Public consent:
    Voluntary for participants.
    First step and stages

    First step: Select the eligible group, service standard and an existing capacity gap in one area. (owner: DWP delivery areas and NHS commissioners)

    1. Recruit and train to fidelity — Connect to Work delivery areas (proposed stage, not an official timetable)
    2. Refer and support into work — Providers with NHS teams (proposed stage, not an official timetable)
    3. Measure sustained work and health from payroll and survey data — DWP (proposed stage, not an official timetable)

    Depends on: NHS capacity to refer and integrate; Employer vacancies

    Trade-offs
    Who gains and who loses:
    Gains for participants who find suitable work; intensive support per person means fewer people reached.
    What else the money or effort could do:
    The same money could fund less intensive support for more people, or health services directly.
    Strongest counterargument:
    Trial effects were modest and inconsistent; scaling may dilute fidelity further.
    Risks:
    Unsuitable jobs, pressure on people who are not ready, and over-claiming benefit savings.
    How we would judge it
    Baseline:
    Economic inactivity due to long-term sickness (ONS Labour Force Survey).
    Leading indicator:
    Programme starts and the share in paid work within 12 months.
    Outcome:
    Employment sustained 6–12 months later (HMRC payroll data), job quality and health.
    Continue, adapt or stop:
    Continue where sustained work beats a matched comparison; adapt or stop where it does not.
    Sources

    Source published April 2023 (trials); July 2026 (business case); evidence reviewed 2026-10-04. Reconsider if: Connect to Work impact evaluation results; The next Spending Review; New trials in different populations.

    Also on: NHS, Economy, Jobs and pay — one shared record, so its cost is counted once.

How the options are assessed

One shared record per option — the same one the national plan reads. Delivery risk, gate and list membership are separate judgements.

  • Claim what people are owed — Amber delivery RAG; gate pending
  • A child-poverty target and essentials test — Amber delivery RAG; gate pending

Provisional single-reviewer assessments from the 6 October 2026 review; not approved business cases, and not government policy.

  • A25Delivery risk: AmberGate: pending

    Claim what people are owed

    Existing action A25, assessed at the same scope.

    Provisional single-reviewer assessment; not an approved business case.

    Who and where
    Pensioner families in Great Britain entitled to Pension Credit who do not claim. (from the action record) Jurisdiction: Great Britain.
    Result so far
    Range if completed: 16.0–22.0 out of 30 (equal weights; not yet assessed: expected incremental benefit (Bn), incremental affordability (Co))
    Funded now (baseline)
    Pension Credit and DWP take-up campaigns already exist; a targeted, evaluated outreach model is a UK Facts proposal. (from the action record)
    The change assessed (marginal)
    Assisted applications for people likely to be entitled to Pension Credit. (from the superseded 4 October selection table)
    What the evidence supports—and the next decision

    Estimated non-take-up identifies an access problem, but cannot predict how many additional awards a new campaign will secure.

    AI-assisted analytical follow-up · 8 October 2026. Proposed appraisal design; the attributed scores below are unchanged. No funding or delivery agreement is implied.

    Define the change
    Test assisted Pension Credit applications for an identified likely-entitled group against existing outreach.
    Test whether it works
    Additional correct awards, processing time, household income and sustained receipt; distinguish applications from entitlement and awards.
    Establish the cost
    Outreach, advice and processing plus additional benefit expenditure. Report fiscal transfers separately from social resource costs.
    Protect against harm
    Use consent and secure data matching; minimise distress, false eligibility messages and digital exclusion.

    Read with the outcome evidence and sources in the full assessment. Monitoring proposals are not existing baseline measurements. A numerical benefit or cost score needs evidence for the defined increment, not just the scale of the problem.

    Full assessment of Claim what people are owed: ten criteria, scenarios, delivery, costs and monitoring
    The ten criteria (0–3)

    Each criterion is scored 0 to 3 by the reviewer. “Not yet assessed” means unknown: it is never counted as zero.

    1. Scale and severity (Sc)2 of 3 — Serious harm to a defined group or widespread moderate harm

      Reviewer’s reason: Provisional problem-importance judgement from linked site pages; not the scale of achievable benefit.

    2. Increment beyond baseline (Ad)2 of 3 — Specific material change to existing delivery

      Reviewer’s reason: Specify eligible cohort, trusted outreach, application support and comparison group; arrange benefits-processing capacity. A missing rating requires a defined marginal change.

    3. Confidence in the proposed action (Ev)0 of 3 — No effect evaluation

      Reviewer’s reason: Unclaimed entitlement establishes a gap; effect of this outreach design is untested.

    4. Expected incremental benefit (Bn)Not yet assessed

      Reviewer’s reason: Unknown unless a reasonably defined target-population benefit can be judged; no bonus merely for spanning domains.

    5. Residual harms (Hm)2 of 3 — Manageable harm with specified safeguards

      Reviewer’s reason: Privacy, mistaken advice and processing delays; fiscal cost is not equivalent to social loss.

    6. Incremental affordability (Co)Not yet assessed

      Reviewer’s reason: Outreach and administration plus additional entitlement paid; show fiscal transfers separately from social resource costs.

    7. Delivery capacity (Ca)2 of 3 — Stretched but plausibly obtainable

      Reviewer’s reason: Specify eligible cohort, trusted outreach, application support and comparison group; arrange benefits-processing capacity.

    8. Distribution (Fa)3 of 3 — Benefits concentrated on disadvantaged people with assessed burdens

      Reviewer’s reason: Provisional distribution judgement for the stated target; verify who pays, who gains and excluded groups.

    9. Meaningful first outcome (Ti)3 of 3 — Within one year

      Reviewer’s reason: First meaningful outcome, not announcement: Successful sustained take-up, disposable income, errors and claimant experience.

    10. Readiness (Rd)2 of 3 — Named commissioning/allocation/coordination step

      Reviewer’s reason: DWP, councils and advice providers; Assisted-application trial

    Four weightings

    Normalised as 10 × weighted sum ÷ sum of weights, on a 0–30 scale. Where a criterion is not yet assessed, the range is the lowest and highest result if it were later scored 0 or 3 — not a score, a confidence interval or a rank.

    • Equal weights: Range if completed: 16.0–22.0 out of 30. Not yet assessed: expected incremental benefit (Bn), incremental affordability (Co). Funding gate: not passed.
    • Evidence doubled: Range if completed: 14.5–20.0 out of 30. Not yet assessed: expected incremental benefit (Bn), incremental affordability (Co). Funding gate: not passed.
    • Scale doubled: Range if completed: 16.4–21.8 out of 30. Not yet assessed: expected incremental benefit (Bn), incremental affordability (Co). Funding gate: not passed.
    • Cost ignored (the funding gate still applies): Range if completed: 17.8–21.1 out of 30. Not yet assessed: expected incremental benefit (Bn), incremental affordability (Co). Funding gate: not passed.

    Funding gate: Not passed — the cost is not yet assessed or no allocation is committed for this scope. Ignoring cost in a weighting never removes the gate.

    Delivery risk and gate

    Delivery risk shown: AmberReview’s claim: Amber. Best the recorded evidence allows: Amber.

    Review’s reason: Outreach and administration plus additional entitlement paid; show fiscal transfers separately from social resource costs.

    Why it is not Green yet

    • Scope (population, place and increment) not verified against the delivery baseline.
    • No verified owner with the powers in place.
    • Delivery capacity not verified.
    • Lifetime incremental costs not complete.
    • No credible funding (a committed allocation for the scope).
    • Safeguards not specified.
    • Monitoring lacks a metric, a data source or a baseline.

    Green needs verified scope, a verified owner with the powers, verified capacity, complete lifetime costs with credible funding, specified safeguards and monitoring with a metric, source and baseline. Amber means appraisal gaps remain; Red means a material delivery obstacle. Red does not mean unimportant.

    Gate: Pending — appraisal incomplete (review label: “Pending: incomplete appraisal”).

    Why: Scores incomplete; Funding gate not passed (cost unknown or no credible funding).

    The gate is separate from national list membership: being listed does not make an option eligible, and an eligible option need not be listed.

    Cost and funding

    Lifetime costing: Partial — some lines costed, no whole-life total. Review: “Full incremental costing not established in reviewed material”.

    • Extra Pension Credit paid if more entitled people claim (partial; C-A25-ENTITLEMENT)
    • Outreach, data matching and assisted applications (not costed; C-A25-OUTREACH)

    Still to cost: Outreach and administration plus additional entitlement paid; show fiscal transfers separately from social resource costs.

    Funding status: not identified.

    Owner, powers and capacity

    Proposed owner: DWP, councils and advice providers (proposed, unverified).

    Powers: needs a decision or allocation. Capacity: unverified.

    Powers and capacity read from the review’s Rd and Ca anchors; no proposed owner has agreed to act (the review: “Proposed owners have not agreed to act”).

    England:
    DWP, councils and advice providers (action record)
    Scotland:
    DWP, councils and advice providers (action record)
    Wales:
    DWP, councils and advice providers (action record)
    Northern Ireland:
    not assessed

    Northern Ireland administers its own social security; not assessed.

    Dependencies, safeguards and risks

    Depends on: None recorded.

    Safeguards

    None specified yet.

    Risks and residual harms: Privacy, mistaken advice and processing delays; fiscal cost is not equivalent to social loss.

    Monitoring and measured results

    Metric: Successful sustained take-up, disposable income, errors and claimant experience.

    Data source: Not yet named. Baseline: Not yet established.

    Measures already on the domain page

    A page measure is not necessarily the review’s metric; linking the two is an editorial decision still to be made.

    Evidence for the outcome

    Outcome: Successful sustained take-up, disposable income, errors and claimant experience.

    Unclaimed entitlement establishes a gap; effect of this outreach design is untested.

    Effect records for this action

    • Income of eligible pensioners who start claiming: size not established (plausible spillover, limited confidence; Great Britain)
    Review record

    Assessment status: Provisional single-reviewer assessment; not an approved business case.

    Reviewer: Original assessment author (name not supplied). Reviewed: 6 October 2026.

    Scored in the working review of 6 October 2026 on the shared ten-criterion scale (one reviewer); not independently calibrated.

    First 90 days (review): Produce a signed scope and current-delivery baseline; Specify eligible cohort, trusted outreach, application support and comparison group; arrange benefits-processing capacity. Obtain incremental costs and an accountable decision before expansion.

    Decision route: Assisted-application trial

    An earlier assessment exists on the 4 October 2026 selection table (nine H/M/L bands; superseded 7 October 2026). Its bands are kept, not converted into these scores.

    An editorial screening framework by one reviewer, not a formal appraisal or business case. Proposed owners have not agreed to act.

  • A16v2Delivery risk: AmberGate: pending

    A child-poverty target and essentials test

    A new scope of action A16 (A16v2); A16 itself is unchanged and nothing is inherited from it.

    Provisional single-reviewer assessment; not an approved business case.

    Who and where
    Population not yet defined. Jurisdiction: not defined.
    Result so far
    Range if completed: 12.0–21.0 out of 30 (equal weights; not yet assessed: increment beyond baseline (Ad), expected incremental benefit (Bn), incremental affordability (Co))
    Funded now (baseline)
    Funded baseline not yet stated
    The change assessed (marginal)
    Marginal change not yet defined
    What the evidence supports—and the next decision

    A poverty target and an essentials test are measurement choices; benefits depend on the policies adopted in response.

    AI-assisted analytical follow-up · 8 October 2026. Proposed appraisal design; the attributed scores below are unchanged. No funding or delivery agreement is implied.

    Define the change
    Separate the monitoring framework from costed tax, benefit, housing and service options beyond the current Child Poverty Strategy.
    Test whether it works
    Relative and absolute low income after housing costs, material deprivation and subgroup outcomes; document definition changes and income-model uncertainty.
    Establish the cost
    Cost administration separately from recurring household support; model interactions and take-up rather than summing headline policy estimates.
    Protect against harm
    Protect families just outside thresholds and report who loses as well as who gains.

    Read with the outcome evidence and sources in the full assessment. Monitoring proposals are not existing baseline measurements. A numerical benefit or cost score needs evidence for the defined increment, not just the scale of the problem.

    Full assessment of A child-poverty target and essentials test: ten criteria, scenarios, delivery, costs and monitoring
    The ten criteria (0–3)

    Each criterion is scored 0 to 3 by the reviewer. “Not yet assessed” means unknown: it is never counted as zero.

    1. Scale and severity (Sc)3 of 3 — Severe harm across a broad population

      Reviewer’s reason: Provisional problem-importance judgement from linked site pages; not the scale of achievable benefit.

    2. Increment beyond baseline (Ad)Not yet assessed

      Reviewer’s reason: Define poverty/essentials measures, then compare specific benefit, housing-cost and service changes; assign accountability. A missing rating requires a defined marginal change.

    3. Confidence in the proposed action (Ev)0 of 3 — No effect evaluation

      Reviewer’s reason: A target does not itself reduce poverty. Modelled effects of a strategy do not establish the effect of the benchmark.

    4. Expected incremental benefit (Bn)Not yet assessed

      Reviewer’s reason: Unknown unless a reasonably defined target-population benefit can be judged; no bonus merely for spanning domains.

    5. Residual harms (Hm)2 of 3 — Manageable harm with specified safeguards

      Reviewer’s reason: Unfunded targets; narrow metrics; inadequate treatment of disability and housing costs.

    6. Incremental affordability (Co)Not yet assessed

      Reviewer’s reason: Separate low-cost monitoring from recurring policy measures; model household distribution and behavioural effects.

    7. Delivery capacity (Ca)2 of 3 — Stretched but plausibly obtainable

      Reviewer’s reason: Define poverty/essentials measures, then compare specific benefit, housing-cost and service changes; assign accountability.

    8. Distribution (Fa)3 of 3 — Benefits concentrated on disadvantaged people with assessed burdens

      Reviewer’s reason: Provisional distribution judgement for the stated target; verify who pays, who gains and excluded groups.

    9. Meaningful first outcome (Ti)1 of 3 — More than three but less than ten years

      Reviewer’s reason: First meaningful outcome, not announcement: Poverty depth, material deprivation and child outcomes after housing costs.

    10. Readiness (Rd)1 of 3 — Legislation/external agreement or major enabling work

      Reviewer’s reason: Treasury, DWP and relevant service departments/devolved governments; Replace benchmark-only action with costed options

    Four weightings

    Normalised as 10 × weighted sum ÷ sum of weights, on a 0–30 scale. Where a criterion is not yet assessed, the range is the lowest and highest result if it were later scored 0 or 3 — not a score, a confidence interval or a rank.

    • Equal weights: Range if completed: 12.0–21.0 out of 30. Not yet assessed: increment beyond baseline (Ad), expected incremental benefit (Bn), incremental affordability (Co). Funding gate: not passed.
    • Evidence doubled: Range if completed: 10.9–19.1 out of 30. Not yet assessed: increment beyond baseline (Ad), expected incremental benefit (Bn), incremental affordability (Co). Funding gate: not passed.
    • Scale doubled: Range if completed: 13.6–21.8 out of 30. Not yet assessed: increment beyond baseline (Ad), expected incremental benefit (Bn), incremental affordability (Co). Funding gate: not passed.
    • Cost ignored (the funding gate still applies): Range if completed: 13.3–20.0 out of 30. Not yet assessed: increment beyond baseline (Ad), expected incremental benefit (Bn), incremental affordability (Co). Funding gate: not passed.

    Funding gate: Not passed — the cost is not yet assessed or no allocation is committed for this scope. Ignoring cost in a weighting never removes the gate.

    Delivery risk and gate

    Delivery risk shown: AmberReview’s claim: Amber. Best the recorded evidence allows: Amber.

    Review’s reason: Separate low-cost monitoring from recurring policy measures; model household distribution and behavioural effects.

    Why it is not Green yet

    • Scope (population, place and increment) not verified against the delivery baseline.
    • No verified owner with the powers in place.
    • Delivery capacity not verified.
    • Lifetime incremental costs not complete.
    • No credible funding (a committed allocation for the scope).
    • Safeguards not specified.
    • Monitoring lacks a metric, a data source or a baseline.

    Green needs verified scope, a verified owner with the powers, verified capacity, complete lifetime costs with credible funding, specified safeguards and monitoring with a metric, source and baseline. Amber means appraisal gaps remain; Red means a material delivery obstacle. Red does not mean unimportant.

    Gate: Pending — appraisal incomplete (review label: “Pending: incomplete appraisal”).

    Why: Scores incomplete; Funding gate not passed (cost unknown or no credible funding).

    The gate is separate from national list membership: being listed does not make an option eligible, and an eligible option need not be listed.

    Cost and funding

    Lifetime costing: Not established. Review: “Full incremental costing not established in reviewed material”.

    Still to cost: Separate low-cost monitoring from recurring policy measures; model household distribution and behavioural effects.

    Funding status: not established in review.

    Owner, powers and capacity

    Proposed owner: Treasury, DWP and relevant service departments/devolved governments (proposed, unverified).

    Powers: needs legislation or external agreement. Capacity: unverified.

    Powers and capacity read from the review’s Rd and Ca anchors; no proposed owner has agreed to act (the review: “Proposed owners have not agreed to act”).

    England:
    not assessed
    Scotland:
    not assessed
    Wales:
    not assessed
    Northern Ireland:
    not assessed

    Not defined in the review.

    Dependencies, safeguards and risks

    Depends on: None recorded.

    Safeguards

    None specified yet.

    Risks and residual harms: Unfunded targets; narrow metrics; inadequate treatment of disability and housing costs.

    Monitoring and measured results

    Metric: Poverty depth, material deprivation and child outcomes after housing costs.

    Data source: Not yet named. Baseline: Not yet established.

    Measures already on the domain page

    A page measure is not necessarily the review’s metric; linking the two is an editorial decision still to be made.

    Evidence for the outcome

    Outcome: Poverty depth, material deprivation and child outcomes after housing costs.

    A target does not itself reduce poverty. Modelled effects of a strategy do not establish the effect of the benchmark.

    Review record

    Assessment status: Provisional single-reviewer assessment; not an approved business case.

    Reviewer: Original assessment author (name not supplied). Reviewed: 6 October 2026.

    Scored in the working review of 6 October 2026 on the shared ten-criterion scale (one reviewer); not independently calibrated.

    First 90 days (review): Produce a signed scope and current-delivery baseline; Define poverty/essentials measures, then compare specific benefit, housing-cost and service changes; assign accountability. Obtain incremental costs and an accountable decision before expansion.

    Decision route: Replace benchmark-only action with costed options

    An editorial screening framework by one reviewer, not a formal appraisal or business case. Proposed owners have not agreed to act.

Decision brief

  • Write to pensioners on Housing Benefit who are missing Pension Credit, help them claim, and count the claims that succeed

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 briefPublic-policy proposalReviewed 2026-10-04

Write to pensioners on Housing Benefit who are missing Pension Credit, help them claim, and count the claims that succeed

The thread: issue, evidence, action

  1. Issue Weakness: Only 62% of people entitled to Pension Credit claimed it; up to 900,000 entitled families did not, leaving up to £2.4bn unclaimed in 2024–25 (one year’s estimate). Also: opportunity, strength
  2. Evidence Pension Credit take-up (caseload)
  3. Action Help eligible people claim support they already qualify for
Decision required
DWP writes to every pensioner household that its Housing Benefit records suggest is entitled to Pension Credit but not claiming, offers a named route to help complete the claim, and contacts households in a random order so the extra successful claims and income can be measured.
Who and where
Pensioner households in Great Britain receiving Housing Benefit but not Pension Credit whose Housing Benefit records suggest entitlement. Applies in: Great Britain.
Accountable body
Department for Work and Pensions (Pension Credit policy and the Pension, Disability and Carers Service)
Intervention
Send invitation-to-claim letters and a reminder to pensioners whose Housing Benefit records suggest Pension Credit entitlement, with a named person who can help complete the claim.
Best evidence, and where it applies
In DWP’s 2023 trial, 27.5% of households written to under its refined criteria made a successful claim within about four months, against 4.6% of similar households not written to; the comparison was not randomised. A US randomised trial found help with the application roughly tripled enrolment compared with no contact.Studied in: About 144,500 pensioner households on Housing Benefit but not Pension Credit; 2,409 written to (708 under refined criteria); refined comparison group 45,400 (Letters July 2023, reminder September 2023; claims followed for almost four months).How closely the comparison transfers to this decision: Pennsylvania, United States — partial; London, England (17 boroughs) — high.
Extra cost of this change
Capital: none identified in the brief.Recurring: Extra Pension Credit paid when more entitled people claim — up to £2.5bn a year, 2023–24, cash (nominal) prices; Average award per successful claim in the trial — £50.57 per household per week; Outreach, data matching and assisted applications — Not yet costed; Help to complete claims (phone and home visits) — not yet established.One-off or mixed: Evaluation using the randomised contact order — not yet established.
Funding
Not funded: no funding route identified yet.Committed to this change: none.Possible source, not an award: none identified.Unfunded: The administration cost is not established and is unfunded; the extra benefit spending depends on how many claims succeed.
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
DWP confirms the legal basis, refreshes the refined Housing Benefit screening, and publishes the evaluation plan with the randomised contact order before the first letters go out.The proposed first milestone; a 90-day date is not yet specified. Owner: DWP. Proposed by UK Facts, not an official deadline.
Twelve-month milestone
Not yet specified.Time to a first useful result: About four months after the first letters, the period over which the 2023 trial counted claims; awards and income need around a year.
Outcome, and when to scale or stop
Measure: Calls and claims started within eight weeks of the letter, and the share of claims abandoned before a decision.Scale if: Successful claims and income gained are clearly higher among households written to than in the comparison group, at an administration cost per extra successful claim DWP judges acceptable.Stop if: Letters produce no more successful claims than the comparison group, or refusals and complaints show the screening is wrong.
Main trade-off
Every successful claim adds to benefit spending, and letters still miss entitled pensioners who claim no benefit at all.
Source and review date
Pension Credit ‘Invitation to Claim’ Trial (DWP research report RR 96) (7 October 2024); Take-up and Targeting: Experimental Evidence from SNAP (NBER Working Paper 24652) (May 2018 (Quarterly Journal of Economics, 2019)). Reviewed 2026-10-04.

An editorial proposal by UK Facts for discussion, not government policy. Records: A25 · effects E-A25-INCOME · costs C-A25-ENTITLEMENT, C-A25-OUTREACH

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

1.The change, in plain words

Write to pensioners on Housing Benefit who are missing Pension Credit, help them claim, and count the claims that succeed

2.The decision requested

Decision
Commit to writing, with a reminder, to all pensioner households on Housing Benefit identified by DWP’s refined Pension Credit criteria; fund a help-to-claim route by phone or home visit for those who respond; contact households in a randomised order over several months; and publish successful claims, awards and income gained.
Accountable institution
Department for Work and Pensions (Pension Credit policy and the Pension, Disability and Carers Service)
Geography
Great Britain
Powers
Existing social security administration powers and DWP’s administration budget. DWP used Housing Benefit data for this purpose in its 2023 trial, so no new legislation is expected; the legal basis and data protection impact assessment must be confirmed before scaling.
Delivery partners
Councils (Housing Benefit and Council Tax Reduction administration; Crisis and Resilience Fund advice services); Independent advice providers such as Citizens Advice and Age UK local organisations, under contract; DWP Pension Credit processing teams

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

Target population
Pensioner households in Great Britain receiving Housing Benefit but not Pension Credit whose Housing Benefit records suggest entitlement.
Eligibility
Over State Pension age, on Housing Benefit, not on Pension Credit, and passing DWP’s refined screening on income and capital. A letter is an invitation to claim, not a decision: entitlement is still assessed on the claim.
Scale
Set by DWP’s own records: its refined criteria identified just under 121,000 Housing Benefit claimants who may be entitled (2023). Housing Benefit claimants are only part of the gap; DWP notes that around one in six entitled non-claimants claim Housing Benefit.
Baseline
Pension Credit take-up was 62% by caseload in 2023–24; up to 910,000 entitled families did not claim and up to £2.5bn went unclaimed, about £2,600 a year per family (DWP, financial year ending 2024). In DWP’s 2023 trial, 27.5% of households written to (refined criteria) made a successful claim within almost four months, against 4.6% in the comparison group.

Source: Department for Work and Pensions, Income-related benefits: estimates of take-up, financial year ending 2024 (Sections 1 and 2); Department for Work and Pensions, Pension Credit ‘Invitation to Claim’ Trial (DWP research report RR 96) (Executive summary; section 4)

Change beyond existing law or funding
DWP has run a national awareness campaign since April 2022 and trialled letters in 10 areas in 2023. Help to Claim covers Universal Credit, not Pension Credit. The added change is a complete, published letter round to the identified group with a reminder, a funded route for help with the claim, a randomised contact order that creates a comparison group, and published results. Whether DWP has already written to the wider group since the trial was not found in a source we could verify; if it has, the increment is the help-to-claim route, a second round for non-responders and the published evaluation.

4.How it would work

Do: Send invitation-to-claim letters and a reminder to pensioners whose Housing Benefit records suggest Pension Credit entitlement, with a named person who can help complete the claim Changing: more entitled pensioners making a claim, and fewer abandoned or incomplete claims

  • More successful Pension Credit claims among the households written to

    Hypothesis to test

    DWP’s 2023 trial found 27.5% against 4.6% (refined criteria) over almost four months, but its treatment areas were not randomly chosen. No validated effect record exists, so this is untested for a national round.

  • Higher income for pensioners who start claiming

    Hypothesis to testEffect record E-A25-INCOME

    The mean award in the trial’s refined group was just under £51 a week (2023). Income gained should be measured from awards, not projected from the national average unclaimed amount.

  • Access to help that Pension Credit unlocks, such as help with council tax, energy bills and health costs

    Hypothesis to test

    Pension Credit is a gateway to other support; whether these are then claimed is not measured in the trial and should be tracked separately.

  • Wasted claims and worry among households that are not entitled

    Hypothesis to test

    A risk, not a benefit: under the original criteria only 37.4% of claims after the letter succeeded; under refined criteria 75.3% did.

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

5.Best evidence

  • Pension Credit ‘Invitation to Claim’ Trial (DWP research report RR 96)

    Author or institution:
    Department for Work and Pensions (Clara Yates, Gary Gifford, Catriona Parpworth, Joe Shuter); research report RR 96
    Published:
    7 October 2024
    Evaluation period:
    Letters July 2023, reminder September 2023; claims followed for almost four months
    Design:
    Planned test with a non-random comparison group: households in 10 local authority areas written to; identified households elsewhere in Great Britain not written to; administrative claims data and 19 interviews
    Sample or population:
    About 144,500 pensioner households on Housing Benefit but not Pension Credit; 2,409 written to (708 under refined criteria); refined comparison group 45,400
    Effect and uncertainty:
    Successful claims 27.5% of those written to against 4.6% in the comparison group (refined criteria); 11.1% against 2.3% under the original criteria. Mean award just under £51 a week. No confidence intervals reported.
    Limits:
    The 10 areas were not a random sample, so DWP says it cannot conclude the same result would follow for all households; follow-up under four months; criteria refined part-way through.

    Where: Executive summary; sections 1.3.2, 2.3, 2.4 and 4. Affiliation is provenance, not proof of quality.

  • Take-up and Targeting: Experimental Evidence from SNAP (NBER Working Paper 24652)

    Author or institution:
    Amy Finkelstein (MIT) and Matthew J. Notowidigdo, with Benefits Data Trust and the Pennsylvania Department of Human Services; NBER Working Paper 24652
    Published:
    May 2018 (Quarterly Journal of Economics, 2019)
    Evaluation period:
    2016, nine months’ follow-up
    Design:
    Three-arm randomised controlled trial (letter only; letter plus application help; no contact)
    Sample or population:
    About 30,000 people aged 60 and over in Pennsylvania, likely eligible for food assistance but not enrolled
    Effect and uncertainty:
    Enrolment after nine months 6% (no contact), 11% (letter only), 18% (letter plus help); differences significant (p < 0.001).
    Limits:
    A different benefit and administration in the United States; the extra enrollees were less needy than average enrollees; costs are rough 2016 dollar estimates.

    Where: Abstract; Introduction pp.1–3. Affiliation is provenance, not proof of quality.

  • How London used data to make pensioners £8 million a year better off

    Author or institution:
    Policy in Practice (the delivery provider), funded by the Mayor of London
    Published:
    15 September 2023
    Evaluation period:
    February to August 2023
    Design:
    Data-matched letters with signposting to advice partners; provider monitoring with no comparison group
    Sample or population:
    8,239 pensioner households in 17 London boroughs
    Effect and uncertainty:
    More than 2,000 successful claims worth just under £8.4m in the first year (gross, not net of claims that would have happened anyway).
    Limits:
    No counterfactual; reported by the company that ran it; costs not published.

    Where: Opening paragraph; “Being smart with data through meaningful collaboration”. Affiliation is provenance, not proof of quality.

6.Alternatives, counter-evidence and who loses

Closest feasible alternative
Join up Pension Credit and pension-age Housing Benefit claims so that one claim is assessed for both, instead of relying on letters. That would reach future claimants without outreach but needs systems and possibly rule changes, and has not been evaluated.
Strongest counter-evidence
Generic awareness work has not measurably raised take-up: DWP has run a campaign since April 2022, yet take-up was 62% in 2023–24 against 65% the year before (not a statistically significant change). In the US trial, outreach drew in less-needy eligible people more than the neediest. Claims rose sharply in late 2024, but alongside the Winter Fuel Payment being tied to Pension Credit, so that rise is not evidence for letters alone.

Source: Department for Work and Pensions, Income-related benefits: estimates of take-up, financial year ending 2024 (Section on Pension Credit take-up (campaign from April 2022)); National Bureau of Economic Research, Take-up and Targeting: Experimental Evidence from SNAP (NBER Working Paper 24652) (Abstract; p.3 (targeting)); Department for Work and Pensions, Tens of thousands more pensioners now in receipt of Pension Credit (Press release text)

Distributional effects
Gains go to low-income pensioners, especially renters on Housing Benefit; in the trial, letters had the largest relative effect for people entitled to small weekly amounts. Entitled pensioners who are owner-occupiers or claim nothing are not reached by this data match. Cost falls on the Exchequer.
Main downside
Success raises benefit spending with no offsetting saving guaranteed, and DWP processing capacity can become the bottleneck: it deployed 500 extra staff during the 2024 surge.

The options compared, including current policy

  • The proposal

    Letters plus help to claim for pensioners on Housing Benefit, in a randomised order (this brief)

    Who is reached
    Pensioners on Housing Benefit who seem entitled to Pension Credit but do not claim
    Evidence
    DWP trial: 27.5% successful claims against 4.6% (non-random comparison); US randomised trial supports help with claims
    Likely outcome and timing
    Extra claims within about four months; income gains over the following year
    Cost and funding
    Administration not costed and not allocated; extra benefit spending demand-led
    Main trade-off
    Higher benefit spending; misses entitled people who claim nothing

  • Current policy

    Current policy: awareness campaign, the 2023 letter trial, and council data shares

    Who is reached
    Pensioners who see adverts or are contacted by their council; Help to Claim covers Universal Credit only
    Evidence
    Take-up did not rise measurably after the 2022 campaign began
    Likely outcome and timing
    Gradual, uneven change
    Cost and funding
    Existing DWP campaign and council budgets
    Main trade-off
    No national comparison, so what works is not learned

  • Alternative

    Council-led data-matched campaigns using the Crisis and Resilience Fund (see the Hampshire decision brief)

    Who is reached
    Residents known to councils, including pensioners not on Housing Benefit but on Council Tax Reduction
    Evidence
    London: more than 2,000 successful claims from 8,239 letters, no comparison group
    Likely outcome and timing
    Depends on each council’s choice and capacity
    Cost and funding
    Within councils’ ring-fenced allocations, displacing crisis payments
    Main trade-off
    Patchy coverage and uneven evaluation

  • Alternative

    Join up Pension Credit and pension-age Housing Benefit claims

    Who is reached
    Future pensioner claimants of either benefit
    Evidence
    Not evaluated
    Likely outcome and timing
    Years: needs new systems
    Cost and funding
    Not costed here
    Main trade-off
    Does not reach today’s non-claimants without outreach

  • Alternative

    Set a test of whether benefits cover essentials

    Who is reached
    All recipients if rates change
    Evidence
    A measurement and decision rule, not an outreach intervention
    Likely outcome and timing
    A published test, then a rate decision
    Cost and funding
    Any rate rise costed separately
    Main trade-off
    Helps people already claiming, not those missing out

  • Alternative

    Employment support alongside healthcare (see the national decision brief)

    Who is reached
    Working-age people with health conditions who want to work
    Evidence
    See the national decision brief on /employment
    Likely outcome and timing
    Sustained work over one to two years
    Cost and funding
    Connect to Work allocations
    Main trade-off
    Different group; does not address pensioner take-up

7.What other countries or the devolved nations show

Country lesson · Pennsylvania, United States, Sample drawn late 2015; letters sent in the first half of 2016; enrolment followed for nine months

Writing to older people likely to be eligible for food assistance in Pennsylvania

Evidence: Experimental or quasi-experimental evaluationPartial match: some conditions differ

What was changed
State and Medicaid records were linked to find people aged 60 and over who were not enrolled in the Supplemental Nutrition Assistance Program (SNAP) but were likely to be eligible. One group received a letter from the state human services department and a reminder postcard; a second group received the same letter with a number for a non-profit that gave phone help to complete the application and gather documents; a third group received nothing.
Study design and population
Randomised controlled trial of about 30,000 people in three equal groups, registered in the AEA RCT Registry; outcomes from state administrative records. Older people already enrolled in Medicaid but not in SNAP, identified from linked administrative data as likely eligible.
Measured outcome there
After nine months, 6% of the control group had enrolled, against 11% of those sent the letter only and 18% of those also offered help to apply; the differences were statistically significant (p < 0.001). About 30% of each letter group called in, and about 75% of applications succeeded in all three groups. The authors’ rough cost per additional enrollee was about $20 for the letter and about $60 with application help.
Cost status
Historic and approximate: the authors’ own rough calculation in US dollars for 2016; no UK cost can be read from it.
What failed or remains uncertain
Most people who were written to still did not enrol. The extra enrollees received lower benefits and were less sick than people who enrolled without the letter, so the outreach reached less-needy eligible people better than the neediest. Dropping the reminder postcard cut the letter’s effect by about a fifth. A working paper finding; the authors’ welfare conclusions rest on a model.
Why the UK differs
SNAP is a food benefit with a different application process, run by a state; Pension Credit is a cash benefit run by DWP, with a backdating rule and passported help. The data match here used Medicaid enrolment; the UK equivalent is Housing Benefit and Council Tax Reduction records, which miss entitled pensioners who claim neither.
First change required here
Randomise the order in which identified households are written to, and test letter only against letter plus a named person who helps complete the claim, so the UK learns the cost per extra successful claim and who is reached.

Not a UK forecast: the effect above was measured in that place and population. Source: National Bureau of Economic Research, Take-up and Targeting: Experimental Evidence from SNAP (NBER Working Paper 24652) (Abstract; Introduction pp.1–3). Reviewed 2026-10-04.

Country lesson · London, England (17 boroughs), February to August 2023 (first phase)

Targeted Pension Credit letters across 17 London boroughs

Evidence: Observational analysisClose match to the UK decision

What was changed
Funded by the Mayor of London, Policy in Practice analysed the boroughs’ own records (with its Low Income Family Tracker software) to identify 8,239 pensioner households likely to be entitled to Pension Credit, wrote to them, and pointed them to partners such as Age UK to help them claim.
Study design and population
Campaign monitoring by the delivery provider: claims and awards tracked in council data after the letters; no comparison group. Pensioner households known to the boroughs (for example through Housing Benefit or Council Tax Reduction) whose records suggested Pension Credit entitlement.
Measured outcome there
More than 2,000 successful claims, worth just under £8.4 million in their first year (average £3,879 per household); about £700,000 of Attendance Allowance was also claimed by 186 people. DWP’s own trial report describes the campaign’s claim success rate as around 28%.
Cost status
Not published in the source read: the cost of the data work, letters and advice is not stated.
What failed or remains uncertain
With no comparison group, some of these households would have claimed anyway, so the figures are gross results, not the campaign’s effect. The provider ran and reported the campaign. Most households written to did not make a successful claim. Results for the later phase were not checked.
Why the UK differs
An English case, so the benefit rules match. London boroughs are unitary councils that hold both benefit records and the local welfare budget; in Hampshire the county holds the crisis fund while the 11 districts hold Housing Benefit and Council Tax Reduction records, so a data-sharing agreement is needed first.
First change required here
Use the same cohort definition across every participating council and stagger the order of contact, so a comparison group exists from the start.

Not a UK forecast: the effect above was measured in that place and population. Source: Policy in Practice, How London used data to make pensioners £8 million a year better off (Blog post, sections “Being smart with data through meaningful collaboration” and opening paragraph); Department for Work and Pensions, Pension Credit ‘Invitation to Claim’ Trial (DWP research report RR 96) (Section 2.3 (comparison with the Mayor of London initiative, footnote 8)). Reviewed 2026-10-04.

8.What delivery needs first

  • law — A confirmed legal basis and data protection impact assessment for using Housing Benefit records to invite Pension Credit claims at national scale. (Genuine veto: nothing proceeds without it; owner: DWP data protection officer and legal advisers)
  • staff — Pension Credit processing staff to decide the extra claims quickly, so backdated awards are not delayed. (Scarce resource shared with other priorities; owner: DWP Pension, Disability and Carers Service)
  • staff — Trained advisers who can help older people complete claims by phone or at home, including people with limited digital access. (Scarce resource shared with other priorities; owner: DWP with councils and contracted advice providers)
  • data — Up-to-date Housing Benefit income and capital data, screened with the refined criteria, and a way to link claims back to letters for the evaluation. (Requirement; owner: DWP analysts with council Housing Benefit teams)
  • public consent — Clear letters that explain why the person was contacted and that a letter is not a decision, so people are not alarmed or misled. (Requirement; owner: DWP communications, tested with older people’s organisations)

9.What it would cost

How the cost is built up: Letters and reminders + screening + assisted claims × households helped + extra processing + evaluation = administration cost; separately, successful claims × measured average award = extra benefit spending. The two are reported side by side, never netted against each other.

  • Extra Pension Credit paid when more entitled people claim

    Pensioner families entitled to Pension Credit who start claiming

    up to £2.5bn a year, 2023–24, cash (nominal) prices (ledger C-A25-ENTITLEMENT)

    recurring; price year not stated; Each year the award continues; Historic cost

    Read from the ledger: the national unclaimed total is a ceiling, not a forecast of what a letter round would add.

    Source: Department for Work and Pensions, Income-related benefits: estimates of take-up, financial year ending 2024 (Section 2)

  • Average award per successful claim in the trial

    Households in the trial’s refined group who made a successful claim

    £50.57 per household per week

    recurring; 2023–24 prices; Weekly, while entitled; Historic cost

    Multiply by measured successful claims, not by the number of letters sent.

    Source: Department for Work and Pensions, Pension Credit ‘Invitation to Claim’ Trial (DWP research report RR 96) (Section 2.4 table (refined treatment group))

  • Outreach, data matching and assisted applications

    Letters, reminders, screening and help to claim

    Not yet costed (ledger C-A25-OUTREACH)

    recurring; price year not stated; While the programme runs; Not yet established

    Read from the ledger, which has no amount.

  • Help to complete claims (phone and home visits)

    Households who respond and want help

    Not yet established per assisted claim

    recurring; price year not stated; While the programme runs; Not yet established

  • Evaluation using the randomised contact order

    Linking letters to claims, awards and passported help

    Not yet established per evaluation

    one-off revenue; price year not stated; First 12 to 18 months; 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
The administration cost is not established and is unfunded; the extra benefit spending depends on how many claims succeed.
Proposed financing
DWP’s administration budget for letters, processing and evaluation; extra Pension Credit is demand-led benefit spending. Councils may choose to fund help with claims from the resilience part of their Crisis and Resilience Fund allocations. Nothing is allocated to this proposal.
Responsible payer
central government, councils
Budget displaced
Other DWP administration and processing work; for councils, other uses of the Crisis and Resilience Fund such as crisis payments.

11.First milestone and time to a useful result

DWP confirms the legal basis, refreshes the refined Housing Benefit screening, and publishes the evaluation plan with the randomised contact order before the first letters go out. (Owner: DWP. Proposed by UK Facts, not an official deadline.)

Time to first useful result: About four months after the first letters, the period over which the 2023 trial counted claims; awards and income need around a year.

What could delay it: Processing capacity keeps pace with extra claims Housing Benefit data are current enough to screen accurately Advice partners can be contracted in time No competing policy change (such as a new passported payment) distorts the comparison

12.How we would know

Baseline
Pension Credit status of every identified household at the start; take-up of 62% (2023–24) as the national context.
Comparison
Households randomly assigned to be written to later act as the comparison for those written to first; within the early group, letter only against letter plus a named helper.
Leading indicator
Calls and claims started within eight weeks of the letter, and the share of claims abandoned before a decision.
Follow-up
At four and 12 months: successful claims, average award, refusals and reasons, passported help claimed, and who was reached by age, household type and area.
Continue if
Successful claims and income gained are clearly higher among households written to than in the comparison group, at an administration cost per extra successful claim DWP judges acceptable.
Adapt if
Many claims are refused or abandoned: tighten the screening, change the letter or add help before writing to more households.
Stop if
Letters produce no more successful claims than the comparison group, or refusals and complaints show the screening is wrong.
Thresholds
Not yet set: Numerical levels need the cost per letter and per assisted claim, the refreshed cohort size and a power calculation; none is published.

What is not yet known, and how to find out

  • What do letters, screening and help with claims cost per extra successful claim? (cost)Next step: Price the letter round and the assisted-claim service, then divide by extra successful claims measured against the randomised comparison.Scope: One national letter round to the identified Housing Benefit group, with help to claim, for 12 months · Data or quotes needed: DWP unit costs for letters and processing; quotes from advice providers per assisted claim; evaluation tender · Who can estimate it: DWP finance and analysts, with advice providers
  • Does a national letter round repeat the trial’s result, and does help with claims add to it? (effect)Next step: Randomise the contact order and the offer of help in the first round.Defensible route: bounded pilot
  • Has DWP already written to the wider Housing Benefit group since the trial, and with what result? (data)Next step: Ask DWP to publish any roll-out dates, numbers written to and claim outcomes.
  • How can entitled pensioners who claim no benefit at all be identified? (data)Next step: Test other lawful data sources, such as council tax records, in council-led campaigns and compare hit rates.
  • Can processing keep pace without delaying other claims? (delivery)Next step: Phase letters to match processing capacity and publish clearance times.

Sources last checked 4 October 2026 · Owner: Evidence editor (James Harris) · Next check: DWP income-related benefits: estimates of take-up, financial year ending 2025 (22 October 2026) or the first trigger: A change to Pension Credit or Housing Benefit rules.

More decision briefs on other pages

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

    Identify likely unmet Pension Credit entitlement without treating a survey estimate as individual eligibility

    Owner: DWP with councils

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

  2. Then, the decision

    Specify the essentials basket and options to cost; budget for extra entitlement from take-up

    Owner: DWP with HM Treasury

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

  3. Then, delivery

    Test outreach in defined areas; keep employment support voluntary and at fidelity

    Owner: DWP, councils, advice services and Connect to Work providers

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

  4. Finally, the measured outcome

    Higher take-up, smaller payment errors both ways, more people in sustained work

    Owner: DWP statistics

    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. No total is shown because these costs are on different bases or periods and cannot be added.

Cost and funding ledger

  • Extra Pension Credit paid if more entitled people claim

    A25 · household transfer · partial

    Amount
    up to £2.5bn a year, 2023–24, cash (nominal) prices
    Who pays
    central government
    Committed
    None identified
    Proposed, not committed
    —
    Gap
    Not established

  • Outreach, data matching and assisted applications

    A25 · administration · not costed

    Amount
    Not yet costed
    Who pays
    central government, councils
    Committed
    None identified
    Proposed, not committed
    —
    Gap
    Not established

  • An independent advisory body on benefit adequacy (secretariat, research, consultation)

    A16 · administration · not costed

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

  • Any change in Universal Credit rates or deductions following the benchmark

    A16 · household transfer · not costed

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

  • Connect to Work grant and programme costs, 2025–26 to 2028–29

    A09 · recurring spending · complete for scope

    Amount
    £1.03bn in total, 2025–26 to 2028–29, cash (nominal) prices
    Who pays
    central government
    Committed
    £1.09bn
    Proposed, not committed
    —
    Gap
    None for this scope

  • Connect to Work costs, 2029–30 to 2031–32

    A09 · recurring spending · complete for scope

    Amount
    £134.4m in total, 2029–30 to 2031–32, cash (nominal) prices
    Who pays
    central government
    Committed
    None identified
    Proposed, not committed
    £138m (Connect to Work funding to be requested at a future Spending Review)
    Gap
    £134.4m

  • NHS clinical integration, referral and supervision for IPS

    A09 · recurring spending · not costed

    Amount
    Not yet costed
    Who pays
    NHS
    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.

Effects on record, with evidence ratings (6 records)

What the evidence says each action changes

Effects on record — these measure different outcomes, so they are not drawn on one scale
ActionOutcome and areaEffectCompared withEvidence
Offer evidence-based employment support alongside healthcareWorking 13 or more weeks in the year after joining (out-of-work participants) Work and pay+4 percentage points (22% vs 18%)significant at the 99% level; no interval publishedrandomised control group receiving usual support; 12 months after randomisation, 2018–2020evaluated; moderate confidence; West Midlands Combined AuthorityNo effect for out-of-work participants in Sheffield City Region, and no effect on earnings at 12 months.
Offer evidence-based employment support alongside healthcareWorking 13 or more weeks in the year after joining (people in work but struggling) Work and pay+3 percentage pointssignificant only at the 90% level; no interval publishedrandomised control group; 12 months after randomisation, 2018–2020evaluated; limited confidence; Sheffield City RegionWeaker evidence than the West Midlands result.
Offer evidence-based employment support alongside healthcareSelf-reported health (EQ-5D-5L) at 12 months Health and care+0.10 standard deviationssignificant at the 90% level (99% when Sheffield groups are pooled); no effect in the West Midlandsrandomised control group; 12 months after randomisationevaluated; limited confidence; Sheffield City RegionSmall; not found in the West Midlands. Not evidence of reduced NHS demand.
Help eligible people claim support they already qualify forIncome of eligible pensioners who start claiming Pensions and retirementDirection plausible; size not established—plausible spillover; limited confidence; Great BritainAn unclaimed family misses about £2,600 a year on average, but how many more would claim after outreach has not been evaluated.
Set an explicit test for whether benefits cover essentialsMaterial hardship among households on Universal Credit Poverty and inequalityDirection plausible; size not established—plausible spillover; limited confidence; Great BritainA benchmark changes nothing unless rates change; how much hardship would fall depends on the option, take-up and housing costs, none of which has been modelled here.
Set an explicit test for whether benefits cover essentialsRecurring benefit spending Public finances and tax · harmSize not established—plausible spillover; limited confidence; Great BritainShown as a trade-off, not a benefit. The size depends on the option and has not been costed on the current baseline.

How we would know

Pension Credit take-up (caseload); Benefit spending overpaid (fraud and error); Benefit spending underpaid (error).

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

Measured results

What each option would be judged on, and the latest value of the measures already on this page. A measure moving is not proof that an option caused it.

Where the ideas come from

Research, counter-evidence and methods (14 sources; 2 claims held back)
  • Universal Credit quarterly statistics, 29 April 2013 to 14 May 2026 — Department for Work and Pensions, 18 August 2026.Official statistics; Great Britain. Relevant passages checked, 2026-10-04.
  • Tackling fraud and error in benefit expenditure 2024–25 — Committee of Public Accounts, 2026-02-11.Committee report; Great Britain. Relevant passages checked, 2026-10-11. Limits: Dated estimates and stated objective, not a current count of completed refunds.
  • Review of Carer’s Allowance Overpayments — Department for Work and Pensions, 2026-04-13 (update).Government update; Great Britain. Relevant passages checked, 2026-10-11. Limits: Start and scope of an exercise, not proof all affected claimants have received redress.
  • Tackling benefit overpayments due to fraud and error — National Audit Office, 2025-10-22.Audit report; Great Britain. Relevant passages checked, 2026-10-11. Limits: Dated audit findings; not a claim that later implementation has failed, or that all overpayments are fraud.
  • Income-related benefits: estimates of take-up, financial year ending 2025 — Department for Work and Pensions, 2026-10-22.Official statistics; Great Britain. Relevant passages checked, 2026-10-23.
  • Fraud and error in the benefit system, financial year ending 2026 estimates — Department for Work and Pensions, 14 May 2026.Official statistics; Great Britain. Relevant passages checked, 2026-10-04.
  • Income-related benefits: estimates of take-up, financial year ending 2024 — Department for Work and Pensions, 2025-10-30.Official statistics; Great Britain. Relevant passages checked, 2026-10-04.
  • Independent process to ensure Universal Credit covers essential costs (briefing) — Joseph Rowntree Foundation (Iain Porter, Isobel Richardson), 2026-01-30.Policy briefing with UK and international precedents; United Kingdom. Full text checked, 2026-10-04. Limits: Does not cost an advisory body or any rate change; essentials estimates are the proposers’ modelling. Makes no claim about savings targets and error.
  • Health-led Employment Trials: evaluation synthesis report — Department for Work and Pensions, 2023-04.Randomised controlled trials of an adapted IPS service with economic and process evaluations; 9,785 people randomised in Sheffield City Region and the West Midlands, May 2018 to October 2019; people with mild to moderate physical or mental health conditions, in or out of work; Two English combined authorities. Relevant passages checked, 2026-10-04. Limits: Adapted IPS for mild to moderate conditions; results differ sharply by site and group. Not evidence about severe mental illness or all economically inactive people.
  • Universal Credit Bill: Universal Credit rebalancing — impact assessment — Department for Work and Pensions, 2025-07-07.Impact assessment; Great Britain. Relevant passages checked (read from an Internet Archive copy (Parliament’s site blocked automated requests); the gov.uk copy is marked withdrawn and points here), 2026-10-04. Limits: Estimates are modelled; DWP describes the behavioural effects as the most uncertain part.
  • Universal Credit Act 2025 (c. 22) — legislation.gov.uk, 2025-09-03 (Royal Assent).Legislation; Great Britain. Relevant passages checked, 2026-10-04.
  • Connect to Work: Programme Business Case 2 summary — Department for Work and Pensions, 2026-07-30.Programme business case; England and Wales. Full text checked, 2026-10-04.
  • Investment priorities for mental health 2025 — Centre for Mental Health, 2025-05-07.Evidence review and policy proposals. Full text checked, 2026-10-04. Limits: Supports examining Individual Placement and Support for appropriate populations. Check original trials, service fidelity and the vintage of any national costing before reuse.
  • The Pensions Review: final recommendations — Institute for Fiscal Studies / abrdn Financial Fairness Trust (Jonathan Cribb, Carl Emmerson, Paul Johnson, Heidi Karjalainen, Laurence O’Brien), 2025-07-02.Policy review with microsimulation of automatic-enrolment reforms; Private-sector employees aged 16–74 (ASHE 2021, earnings uprated to 2025 prices); United Kingdom (automatic enrolment: Great Britain; Northern Ireland has parallel legislation). Relevant passages checked (read from an Internet Archive copy), 2026-10-04. Limits: Not a 12% package (12% appears only as a comparison). Modelled costs, not an approved settlement; transition and administration not included.

Counter-evidence and criticism

  • Claim what people are owed: Survey-based estimates can misclassify households as entitled, so a statistical estimate is not individual eligibility. (Department for Work and Pensions)
  • Set an explicit test for whether benefits cover essentials: An essentials basket is a normative choice, and a published benchmark creates pressure for large recurring spending that must be financed by tax, cuts elsewhere or borrowing; higher basic rates also change work incentives at the margin. JRF itself does not cost the change. (Joseph Rowntree Foundation)
  • Offer evidence-based employment support alongside healthcare: One of three trial groups saw no employment effect and the Exchequer return within a year was 1–2p per £1; benefits were mainly in health and wellbeing, and only in one area. (Department for Work and Pensions)

Claims held back

  • Error, exclusion and hardship rise when savings targets dominate service design (brief threat, cited to S37) — The JRF briefing makes the case for an adequacy process; it does not present evidence on savings targets and error. Needed: An audit or evaluation linking savings measures to error and hardship; the threat is re-sourced to DWP’s own impact assessment.
  • Any adequacy threshold for benefits — UK Facts sets no threshold of its own; the JRF/Trussell essentials estimates are shown as their modelling. Needed: An independent, published benchmark with a stated basket and household types.

“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.

Other options compared — not on the shortlist (4)

Other options compared

Options from earlier work on this page that are not on the shortlist. They remain valid choices; their evidence is shown as before.

  • structuralStatus: legislated

    Rebalance Universal Credit: a higher standard allowance, a lower health element for new claims

    Aims to:
    A higher basic rate for all Universal Credit households and, the government intends, a weaker incentive to be assessed as unable to work.
    How:
    The standard allowance rises above inflation each year to 2029–30; the health (LCWRA) element for most new claims from April 2026 is about half the previous rate, with protection for the most severe conditions and people at the end of life.
    Who decides:
    DWP and HM Treasury.
    Main constraint:
    Behavioural effects are uncertain; people with health conditions who claim after April 2026 receive less for the same assessed need.
    Evidence:
    Not evaluated — a mechanism to test, not an evidenced fix
    See the trade-offs and evidence
    What the evidence shows
    The effects are modelled, not observed. DWP’s impact assessment expects 750,000 people on the lower health rate by 2029–30 and says the behavioural effects are “the most uncertain aspects”. No evaluation of the change has been published yet.
    Policy status
    Universal Credit Act 2025. From April 2026 the standard allowance for a single person aged 25+ is £424.90 a month; the new-claim health element is £217.26 a month (was £423.27).
    Resources
    Benefit spending; Work Capability Assessment capacity; employment support for people with health conditions.
    Consent and decisions needed
    Enacted by Parliament in 2025; further change would need new legislation.
    Time
    Rates changed April 2026; full caseload effect by 2029–30.
    Cost and yield
    DWP impact assessment (July 2025, cash): the standard allowance increase costs about £1.85bn in 2029–30; the health-element change saves about £2.1bn net in 2029–30 after protections.
    Trade-offs and who pays
    Gains for all Universal Credit households are partly paid for by new claimants with health conditions.
    How we would know it worked
    Leading indicator: Number of new claims on the lower health rate, and Work Capability Assessment referrals. Outcome: Employment and household income of people with health conditions on Universal Credit, compared with earlier cohorts. It would not be working if: Employment among new health claimants does not rise while their incomes fall.
  • capacityStatus: funded

    Employment support for disabled people and people with health conditions

    Aims to:
    More people who want to work and who have health conditions in sustained jobs.
    How:
    Voluntary, tailored support — employment specialists, more work-coach time and links with health services — offered to people on health-related benefits.
    Who decides:
    DWP; local authorities (Connect to Work); NHS; employers.
    Main constraint:
    Recruiting specialists; NHS capacity; employers offering adjusted jobs. Money does not create jobs or health care on its own.
    Evidence:
    Moderate evidence — evaluated, with limits on scope or transfer
    See the trade-offs and evidence
    What the evidence shows
    DWP cites evaluations of earlier programmes: Work Choice participants were 11–14 percentage points more likely to be in work eight years later; extra work-coach time for people in the UC health group raised employment by 3 points after a year. Effects are real but modest. Detailed evidence on Individual Placement and Support is on the Jobs and pay page.
    Policy status
    £2.2bn of new employment support from 2026–27 to 2029–30 (per the July 2025 impact assessment).
    Resources
    Work coaches and employment specialists; NHS referral routes.
    Consent and decisions needed
    Funded; voluntary for participants.
    Time
    Job outcomes within 12 months of starting; caseload effects over several years.
    Cost and yield
    £2.2bn over 2026–27 to 2029–30 (cash). DWP estimates fiscal savings of about £180m a year for every 10,000 extra people in full-time work (modelled).
    Trade-offs and who pays
    Intensive support reaches fewer people; the money could raise benefit rates instead.
    How we would know it worked
    Leading indicator: Starts and job entries within a year. Outcome: Employment sustained at 12 months against a comparison group. It would not be working if: No difference in sustained employment from similar people not offered support.
  • preventionStatus: legislated

    Pay the right amount: bank-data checks and targeted case reviews

    Aims to:
    Lower overpayments (and earlier detection, so claimants build up smaller debts).
    How:
    Banks flag accounts receiving specified benefits that break eligibility rules (for example savings above £16,000 on Universal Credit); DWP reviews them before deciding anything.
    Who decides:
    DWP; banks and building societies; Information Commissioner (oversight).
    Main constraint:
    Data protection and proportionality; false positives; banks’ IT; capital and abroad errors are hard to detect from other data.
    Evidence:
    Limited evidence — plausible mechanism, little direct evaluation
    See the trade-offs and evidence
    What the evidence shows
    DWP estimates the eligibility verification measure will find 50,000–100,000 overpayments a year and save about £500m a year when fully rolled out; the OBR validated savings of up to £940m over five years. These are estimates, not evaluated results; the measure is being rolled out on a test-and-learn basis.
    Policy status
    Public Authorities (Fraud, Error and Recovery) Act 2025 (Royal Assent December 2025); gradual rollout.
    Resources
    DWP investigators and data analysts (up to 3,000 more staff across fraud, error and debt work); bank systems.
    Consent and decisions needed
    Legislated; privacy concerns raised in Parliament and by the Information Commissioner shape the codes of practice.
    Time
    Gradual; DWP expects £1.5bn savings from the Act by 2029–30.
    Cost and yield
    Savings estimate about £500m a year at full rollout (DWP), up to £940m over five years (OBR-validated); implementation costs not separately published here.
    Trade-offs and who pays
    Wrongly flagged claimants face checks and stress; data access raises privacy concerns. Recovering more overpayments can deepen hardship if deductions are high.
    How we would know it worked
    Leading indicator: Accounts flagged and the share confirmed as overpaid. Outcome: The fraud and claimant-error overpayment rates in DWP’s annual estimates (2.2% and 0.6% in 2025–26). It would not be working if: Overpayment rates for capital and savings errors do not fall in the benefits covered.
  • structuralStatus: announced

    Reform how PIP is assessed (Timms Review)

    Aims to:
    A PIP assessment that reflects the extra costs of disability fairly and is trusted, within the spending forecast.
    How:
    A review co-chaired by Sir Stephen Timms (Minister for Social Security and Disability), Sharon Brennan and Dr Clenton Farquharson, with a 12-member steering group, recommends changes to the assessment.
    Who decides:
    DWP (Minister for Social Security and Disability); Parliament for any legislation.
    Main constraint:
    The spending forecast; assessment quality (67% of PIP appeals decided at a tribunal hearing in January–March 2026 went the claimant’s way); legal challenge.
    Evidence:
    Not evaluated — a mechanism to test, not an evidenced fix
    See the trade-offs and evidence
    What the evidence shows
    No recommendations yet: the interim report (July 2026) sets out evidence and says final recommendations will come in the autumn. Its terms of reference require recommendations to sit within forecast PIP spending (over £41bn by 2030–31, DWP spring 2026 forecast, 2026–27 prices).
    Policy status
    Review under way; final report due autumn 2026.
    Resources
    Assessment capacity; DWP decision-makers; tribunal capacity if appeals rise.
    Consent and decisions needed
    Parliament directed that the review be co-produced with disabled people; any change to eligibility needs regulations or primary legislation.
    Time
    Recommendations autumn 2026; any new assessment would take years to apply to existing claims.
    Cost and yield
    No official costing until recommendations are published.
    Trade-offs and who pays
    Tighter criteria reduce spending but remove support from some disabled people; looser criteria raise cost.
    How we would know it worked
    Leading indicator: Share of initial decisions changed on reconsideration or appeal. Outcome: Material deprivation and extra-cost hardship among disabled people. It would not be working if: Appeal overturn rates and disabled people’s hardship do not improve after reform.
The options side by side
OptionResourcesTimeEvidenceOutcome to watch
Rebalance Universal Credit: a higher standard allowance, a lower health element for new claimsBenefit spending; Work Capability Assessment capacity; employment support for people with health conditions.Rates changed April 2026; full caseload effect by 2029–30.not evaluatedEmployment and household income of people with health conditions on Universal Credit, compared with earlier cohorts.
Employment support for disabled people and people with health conditionsWork coaches and employment specialists; NHS referral routes.Job outcomes within 12 months of starting; caseload effects over several years.moderateEmployment sustained at 12 months against a comparison group.
Pay the right amount: bank-data checks and targeted case reviewsDWP investigators and data analysts (up to 3,000 more staff across fraud, error and debt work); bank systems.Gradual; DWP expects £1.5bn savings from the Act by 2029–30.limitedThe fraud and claimant-error overpayment rates in DWP’s annual estimates (2.2% and 0.6% in 2025–26).
Reform how PIP is assessed (Timms Review)Assessment capacity; DWP decision-makers; tribunal capacity if appeals rise.Recommendations autumn 2026; any new assessment would take years to apply to existing claims.not evaluatedMaterial deprivation and extra-cost hardship among disabled people.

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