Executive Summary

The United Kingdom spent £313 billion on welfare in 2024/25 — 24% of all government expenditure and the largest single departmental budget.¹ The public debate about this figure is almost entirely dishonest. It focuses on the least powerful recipients — disabled people, unemployed claimants, single parents — while systematically ignoring the dominant fact: nearly half of that budget flows to pensioners, through a triple lock mechanism that costs £12 billion more per year than earnings-uprating alone, in a country where 4.3 million children live in poverty.

This pillar makes four arguments. First, that the UK welfare system is structurally skewed — by electoral arithmetic rather than by need — toward the generation that has also captured the majority of the UK's £10.8 trillion in household wealth. Second, that the disability and mental health caseload explosion is being deliberately misdiagnosed as fraud and dependency when the data shows it is a public health crisis rooted in NHS waiting lists, precarious work, and inadequate housing. Third, that child poverty at 31% is not an act of god but a sustained policy choice, maintained for nine years through a two-child benefit cap that cost £3.4 billion to abolish but kept one million children in deeper poverty while it stood. Fourth, that Universal Credit, despite being the largest working-age benefit, provides support at just 12.5% of median earnings — among the least generous in comparable economies.

The international comparison is damning but clarifying. The UK does not have an over-generous welfare state. On most working-age metrics it is below the OECD average. What it has is a welfare state whose composition reflects the voting power of its beneficiaries more faithfully than it reflects the distribution of need. Until the electoral system that produced this distortion is reformed, the welfare system will continue to drift further from its stated purpose. This pillar cannot be read in isolation from the Political Reform pillar. The two are the same argument.

Key Proposals

1. The Honest Diagnosis

The UK welfare system isn't failing because it's too generous — on most working-age measures it's less generous than OECD peers. It's failing because its composition has been systematically shaped by electoral arithmetic rather than social need, and because the political conditions that would allow reform are themselves produced by the system that needs reforming.

The structural diagnosis is precise: the UK welfare system transfers money most reliably to those who vote most reliably, funded by those who vote least. Pensioners vote at rates approaching 80%. Under-35s vote at rates around 45%. The triple lock, the universal Winter Fuel Payment, the protection of pensioner council tax discounts — all of it follows the same electoral arithmetic. None of it was designed by conspiracy. It emerged from a political incentive structure that rewards responsiveness to high-turnout demographics and punishes attention to low-turnout ones.

1.1 Where the £313 Billion Actually Goes

The public argument about welfare is conducted almost entirely in the abstract, using three recurring figures who rarely appear in the spending data itself: the workless claimant who could work but won't, the woman having children she can't afford, and the immigrant drawing on a system they haven't paid into. DWP publishes exactly what the money is spent on, claimant by claimant.³ It doesn't support any of the three.

The single largest line — the State Pension — is covered in full in the Public Spending Overview and has its own dedicated pillar, the State Pension Explained; it isn't re-broken-down here to avoid stating a third, slightly different figure for the same thing (DWP's own internal accounting puts it a little below the £146.5bn used on those pages — the pattern is identical to the £313bn/£383.9bn difference already explained above, not a new error). Everything else DWP counts within the £313bn does need breaking down, because this is where the "who's actually claiming this" argument is decided:

These categories account for £159bn; the State Pension covered elsewhere accounts for most of the remainder of the £313bn, with Northern Ireland's own social security system and a long tail of smaller benefits (Winter Fuel Payment, Maternity Allowance, Bereavement Support, and others) making up what's left.³ Add Pension Credit and Attendance Allowance above — both exclusively pensioner benefits — to the State Pension itself (£146.5bn + £6.0bn + £7.8bn = £160.3bn), and pensioners account for slightly more than half of the entire £313bn before a single working-age claim is counted. That combined figure runs a little above the £150.7bn/48% headline cited earlier, which reflects DWP's own narrower pensioner-spend accounting rather than this page's own addition of the three PESA/DWP line items above — both are legitimate ways of drawing the same boundary, and neither should be read as more precise than the other.

The "workless claimant who could work" line item is £0.31bn. Jobseeker's Allowance — contributory or income-based, paid to people who are unemployed, judged capable of work, and required to actively search for it — has 68,000 claimants and cost £313 million in 2024/25.³ That is one-tenth of one percent of the £313bn total. It is not a rounding error in a spreadsheet; it is close to the entire concept the public debate is organised around. Universal Credit, which has absorbed almost everyone who would once have claimed Jobseeker's Allowance, serves an entirely different population: people already in work receiving a top-up (£22.2bn of standard allowance goes to a caseload that includes large numbers of employed claimants), single parents, carers, and £7.3bn paid specifically to claimants DWP itself has formally assessed as unable to work or unable to prepare for it. DWP does publish a finer breakdown of Universal Credit by work-search requirement — how many claimants are actively required to search for full-time work versus exempted for health, disability, or caring reasons — but it sits behind an interactive query tool (Stat-Xplore) this project could not extract in this pass; it is tracked in the Gaps Register as an open item, because it would sharpen this picture further, not soften it.

The "large family" line item is 496,000 households. The two-child benefit limit — which restricted the child element of Universal Credit and Child Tax Credit to a claimant's first two children, and was abolished from April 2026 — affected 495,990 UC households in its final year, covering 1,752,200 children.⁴ Of those households, 64% had exactly three children, not the six or eight sometimes implied in political rhetoric; 8% had five, and 4% had six or more.⁴ Measured against the roughly 4 million UC households with children, families of three-plus children were a minority of the caseload, and families the limit actually restricted were 78% of that minority — the rest qualified for an exception (multiple births, adoption, non-consensual conception).⁴ The policy's abolition cost £3.4bn a year to reverse nine years of it; the population it was ever aimed at was never the mass of claimants the debate assumes.

The immigration claim needs the legal starting point most of the debate skips. Most temporary UK visas — work, study, and family routes under Appendix FM in particular — carry a "No Recourse to Public Funds" condition by default.⁵ NRPF is not a guideline; it is a legal bar. It blocks Universal Credit, Pension Credit, PIP, Attendance Allowance, DLA, Carer's Allowance, Housing Benefit, Income Support, income-based ESA, and homelessness assistance outright.⁵ A recent migrant on a standard work or family visa is, by law, unable to claim almost everything in the table above; only settled status, British citizenship, or refugee status removes the condition. Child Benefit is a specific, narrow exception — it is not barred by NRPF.⁵ The one figure DWP does publish is nationality on Universal Credit specifically: as of June 2025, 1.3 million claimants — 16.4% of the 7.9 million caseload — were non-UK/Irish nationals, defined as anyone who has passed the Habitual Residence Test.⁶ That definition is narrower than "immigrant" in everyday usage — it excludes naturalised British citizens born abroad and includes long-settled EU nationals with British-born children — and it covers only Universal Credit, itself roughly half of working-age welfare spending; DWP has not published equivalent nationality breakdowns for PIP, Housing Benefit, Child Benefit, or legacy tax credits. So the 16.4% figure cannot be reweighted onto the £313bn total, or even the full working-age total, without inventing data that doesn't exist — any claim of the form "X% of the welfare bill goes to immigrants" that isn't explicitly scoped to Universal Credit should be treated as unsourced. The remaining gap — the rest of the system, and how the foreign-born claimant share compares to the foreign-born share of the population and workforce — is tracked in the Gaps Register.

2. What the State Pension Actually Is — A Misunderstood Transfer

Before examining the triple lock, there is a more fundamental point that almost never gets made in public debate — and without it, the entire political argument about pensions rests on a misunderstanding.

This is the fundamental structural difference between the state pension and a private or occupational pension fund. A private pension fund takes contributions, invests them in assets — equities, bonds, property — and pays out the returns on those investments as retirement income. The money exists. It has been invested and grown. When you draw it, you are drawing on accumulated capital.

The state pension has no such assets. It is a pay-as-you-go system — a direct transfer from current workers to current retirees. The phrase 'I paid in all my life and I deserve my return' — heard constantly in debates about the triple lock — describes a private pension. It does not describe the state pension. What pensioners paid in funded their parents' generation's retirement. What they receive in retirement is funded by their children's and grandchildren's earnings.

The Norway comparison is the clearest possible illustration of what a genuinely funded alternative looks like. Norway took its North Sea oil revenues — broadly comparable to the UK's — and invested them in the Government Pension Fund Global, now worth approximately $1.8 trillion. That fund generates returns that fund Norwegian public spending. The UK took comparable revenues and spent them. The state pension has no equivalent accumulated asset. This is not a peripheral observation. It is the central story of intergenerational fiscal policy in the UK over the past forty years.

3. The Triple Lock — The Mechanism in Detail

3.1 What It Does

The triple lock guarantees annual state pension increases of the highest of average earnings growth, CPI inflation, or 2.5%. In practice this means the state pension rises faster than wages in most years — the 2.5% floor guarantees real-terms growth even in years of economic stagnation. The cumulative effect is that state pension spending rises as a share of GDP year on year. The OBR projects pension spending will rise from 5% of GDP today to 8-10% over the next four decades, driven by demographics and the triple lock ratchet combined.⁷

3.2 The Steel Man

The triple lock has achieved something real. Before 2011, pensioner poverty rates in the UK were among the highest in Europe. The combination of the triple lock and the new flat-rate state pension reversed this. UK pensioner poverty fell from approximately 29% in the early 1990s to 16% today (after housing costs). The policy had a rationale grounded in genuine need — pensioners had seen their income eroded by decades of price-only uprating — and it delivered measurable outcomes. Any reform proposal that ignores this history is not serious.

3.3 The Intergenerational Arithmetic

The triple lock is funded by working-age taxpayers and distributed to pensioners regardless of wealth. The generation receiving its benefits has seen house price growth averaging 500% in real terms since 1970, defined benefit pension schemes now largely unavailable to younger workers, and stock market returns that have generated substantial real wealth. The generation funding those transfers faces stagnant real wages, a housing market that has priced homeownership out of reach in most cities, and the abolition of the defined benefit pension schemes their parents relied on.

4. The Disability and Mental Health Crisis — Misdiagnosed by Design

Working-age health-related benefit spending rose from 1.3% of GDP in 2019 to 1.7% in 2023/24⁹ — an increase of £12 billion in four years, with a further £15 billion increase forecast by 2028. One in ten working-age adults now claims a sickness or disability benefit. The political response has been consistent: this is a fraud and dependency problem requiring stricter eligibility, tighter assessments, and reduced payments. The data does not support this diagnosis.

4.1 What the Data Actually Shows

Pre-pandemic in 2019, UK spending on incapacity-related benefits at 1.3% of GDP was below the OECD average of 1.6%. The UK was not running a generous disability system before the explosion. It was running a lean one. The post-2020 rise is driven by a genuine deterioration in the health of the working-age population. Mental and behavioural conditions were the primary diagnosis for 25% of disability claimants in 2002. By 2024, that figure was 44%. PIP claims overall rose 68% between 2020 and 2024. Among 16 to 17 year olds — the group least likely to be gaming a system they have never paid into — new PIP claims rose 138%.¹⁰

4.2 The NHS Waiting List Connection

In 2024, 7.6 million people were on NHS waiting lists. People waiting years for treatment for mental health conditions, for musculoskeletal surgery, for cancer diagnosis cannot work. They claim disability benefits while they wait. The OBR and IFS have both identified NHS waiting list clearance as one of the most cost-effective long-run welfare savings available — not because it removes entitlement, but because it removes the health condition preventing employment. This is a cross-pillar dependency that has no clean solution within any single department.

4.3 The International Comparison

Countries with strong employer sick-pay obligations and structured return-to-work requirements have lower disability benefit caseloads — not because they are harsher to disabled people, but because they intervene earlier, at the employment stage, before the transition to permanent inactivity has occurred. The Netherlands requires employers to fund full pay for up to two years with mandatory return-to-work from month one. Norway requires full pay for up to 52 weeks with mandatory return-to-work from month one. Both have lower disability spending as a percentage of GDP than the UK, despite more generous individual provision. The difference is architecture, not generosity.

5. Child Poverty — A Policy Choice, Not a Natural Condition

4.3 million children — 31% of all children in the UK, after housing costs — lived in poverty in 2023/24.¹¹ 72% of those children lived in working families. Child poverty is not primarily a consequence of parental worklessness. It is a consequence of wages that are too low, housing costs that are too high, and in-work benefit support that is structurally inadequate.

The UK's child poverty rate is among the worst of any comparable wealthy economy. Germany achieves 14%. The Netherlands achieves 13%.¹¹ The Nordic countries achieve rates between 4% and 6%. These are not marginal differences. They represent millions of children in comparable economies growing up with adequate nutrition, stable housing, and access to developmental resources — while their equivalents in Britain do not.

The two-child benefit limit — introduced in 2017, maintained for nine years at a saving of £3.4 billion per year¹² — is the most precise available illustration of how the welfare system's priorities are set. The limit cost roughly one quarter of what the triple lock costs above earnings-uprating in any given year. It was maintained through nine years of child poverty evidence accumulating, on the grounds that 'taxpayers should not be expected to fund large families.' It has been abolished from April 2026. The nine years of its operation, during which the poverty rate for children in large families rose substantially, represent a documented policy choice — not an inevitability.

6. Universal Credit — The Architecture of Inadequacy

Universal Credit cost £87.8 billion in 2024/25, making it the largest single working-age welfare programme. The design logic was sound: simplify six legacy benefits into a single payment, with a taper rate designed to make work pay more reliably than the previous system. The implementation and the adequacy of the rates are not.

The standard Universal Credit allowance for a single person aged 25 or over was worth 12.5% of median earnings in April 2024 — among the least generous working-age support rates in comparable economies.¹³ The Local Housing Allowance element has been frozen, uprated below inflation, and decoupled from actual rental market rates for most of the past decade. In 2023, the LHA covered the bottom 30th percentile of local rents. In practice, in many urban areas, it covers the bottom 10th to 15th percentile. The gap is bridged by the claimant — out of an allowance already set at 12.5% of median earnings.

7. The Electoral Mechanism — Why the System Reproduces Itself

The welfare system's skew toward older voters is not maintained by malice. It is maintained by incentive. Pensioners vote at rates of 75 to 80%. Adults under 35 vote at rates of 40 to 50%. In a first-past-the-post system where marginal seats are won by thousands of votes, the electoral mathematics are precise: protecting the triple lock is worth more votes than it costs. Restricting in-work benefits, freezing Local Housing Allowance — these cost votes primarily among people who either do not vote or are concentrated in safe seats.

8. Social Care — The Bill That Hides Inside Another Bill

Social care is treated in this pillar as a component of the welfare system because its funding failure is inseparable from the welfare state's broader resource allocation failure. The political history of social care reform is the most complete available illustration of the structural problem this pillar diagnoses.

Tony Blair's government established the Royal Commission on Long Term Care in 1997. The Dilnot Commission published its recommendations in 2011. The Care Act 2014 legislated a version of those reforms. Implementation was delayed in 2015, promised again in 2021, delayed to 2025, and cancelled in July 2024. The means test threshold has been frozen at £23,250 since 2011 — a figure that would be approximately £35,000 if uprated for inflation. Average monthly care costs are now £5,000.

The political economy of this failure is precise: social care reform costs money upfront and delivers benefits dispersed across multiple departments, multiple years, and multiple electoral cycles. In a political system that allocates credit by department and accountability by parliament, that calculation consistently produces inaction.

9. Counter-Arguments

'Pensioners earned their state pension through National Insurance — it is deferred pay, not welfare'

The state pension is not a funded scheme. National Insurance contributions do not sit in a ring-fenced investment fund. Current pension payments are funded by current workers' contributions and general taxation — it is a pay-as-you-go system. The contributory framing is politically important but economically misleading. More importantly, it does not explain the triple lock. The legal minimum is earnings uprating. The triple lock's additional cost — £12 billion per year — is not a deferred entitlement. It is a political choice made every year by governments that depend on the pensioner vote.

'Cutting disability benefits will force sick people into poverty'

Correct — which is why this pillar does not propose cutting disability benefits. It proposes treating the causes rather than managing the symptoms: NHS waiting list clearance, employer return-to-work obligations, active labour market support modelled on the Netherlands and Norway, and community mental health investment. The distinction between reforming the system to help people into employment and cutting their support while leaving the underlying conditions unaddressed is the entire point.

'The UK cannot afford to increase welfare spending'

The UK cannot afford not to address child poverty, which generates long-run costs in health, education, criminal justice, and lost productivity that dwarf the short-run cost of adequate support. An honest fiscal argument would apply the same standard of scrutiny to a £150.7 billion universal pensioner transfer as it applies to a £13.3 billion child benefit budget. The difference is not the scale. It is whose needs the political system treats as priorities.

'Not all pensioners are wealthy — many live in genuine poverty'

Correct, and this is the most important internal distinction in the triple lock debate. Approximately 1.9 million pensioners live in poverty. They are disproportionately women with broken employment histories, ethnic minority pensioners, and renters who received none of the housing wealth windfall. The reform argument is not that pensioners are undeserving. It is that universal uprating — regardless of wealth — is the wrong instrument for supporting those in genuine need. A smoothed earnings link combined with enhanced Pension Credit for those below a defined income threshold would provide stronger protection for the poorest pensioners at lower total cost than the current arrangement.

11. Proposals for Change

The following represent the evidence-based proposals of this pillar, put forward for public discussion and challenge. They are not a programme for government.

State Pension and Triple Lock

• P1 Replace the triple lock with a smoothed earnings link modelled on Australia's system and recommended by the IFS Pensions Review. The state pension rises with earnings. In years when inflation exceeds earnings, the pension rises with inflation and the shortfall is made good in subsequent years. This protects pensioner living standards during downturns while eliminating the asymmetric ratchet costing £12 billion per year above earnings indexation.

• Enhanced Pension Credit for the poorest third of pensioners throughout any transition — specifically protecting the 1.9 million pensioners currently in poverty who are most vulnerable to any change in the uprating mechanism.

• Phase implementation over ten years with full parliamentary transparency and independent OBR monitoring.

Child Poverty

• P2 Establish a statutory child poverty target — defined as the proportion of children below 60% of median income after housing costs — with a binding trajectory to 10% by 2035, independently monitored.

• Restore Local Housing Allowance to cover the 30th percentile of actual local market rents and index it annually to rental inflation.

• Invest £3 billion per year in universal early years provision — funded from the fiscal saving generated by triple lock reform.

• Embed child poverty impact assessments as statutory requirements for all major fiscal decisions.

Disability and Long-Term Sickness

• P3 Introduce a statutory employer return-to-work obligation — employers fund and manage structured return-to-work for the first twelve weeks of sickness absence, modelled on the Norwegian system.

• P4 Establish an NHS-DWP waiting list bridge: people on waiting lists for conditions preventing employment receive active labour market support during their wait, with NHS treatment prioritised by employment impact.

• Commission an independent mental health workforce expansion programme — the PIP caseload for anxiety and depression represents a public health crisis requiring clinical response, not a benefits eligibility crisis requiring tighter assessment.

Universal Credit

• P5 Raise the UC standard allowance to 15% of median earnings within the first parliament, with a pathway to 18% over ten years.

• P6 Remove the benefit cap for households with dependent children.

• Reduce the UC taper rate from 55p to 45p per pound earned to strengthen work incentives.

Social Care

• Implement the Dilnot framework in full: a national cap on lifetime care costs of £86,000 (uprated to 2026 prices), a means test threshold raised to £100,000 and indexed to asset price inflation, and a minimum national fee rate. Fund through a hypothecated social care levy generating approximately £8 billion per year.

The Generational Reset is a non-partisan, public-interest project. It is not affiliated with any political party, does not accept corporate funding, and publishes all its work under open licence for public discussion and adaptation.

For public discussion. Not affiliated with any political party. | generationalreset.org

The Generational Reset | S1_03: Social Welfare | For public discussion. Not affiliated with any political party. | generationalreset.org