Making Welfare Work

MAKING WELFARE

WORK

Summary

The welfare system is not working. Around one-in-ten working-age people in England and Wales are on PIP, the disability benefit.1 One-in-seven Universal Credit (UC) claimants are not British or Irish citizens.2 The number of children receiving disability benefits has doubled in just 10 years.3 The overall welfare bill, excluding pensions, cost the average family more than £5,600 last year. And Labour is making it worse. By 2031, the bill will reach almost £7,000 per family, per year.4 A system intended as a safety net has become a way of life, including for foreigners who haven’t paid in a penny. Today, the welfare system is just not fair on British workers. Reform UK will make sure that it is. Our changes will maintain, reform and enhance PIP for working-age claimants. We will: 1. Require all work-capable long-term claimants to contribute by working for their communities; 2. Make British citizenship a condition for accessing almost every form of benefit; 3. Reform disability benefits to make the system fairer, keep people connected to employers and gradually remove unconditional cash support for mild cases, while protecting: a. Those over the State Pension Age, b. The severely disabled, who will see their cash awards protected, c. Existing non-trivial, non-mental-health-related claimants, who will be exempt from reassessment under the new system for three years, and d. All children currently receiving child DLA 4. Substantially tighten up fraud and error enforcement. We will also make small technical changes to indexation to further reduce costs. As previously announced, we will also reintroduce the Two-Child Benefit Cap. By 2030, Reform’s welfare reforms, if implemented today, would get 241,000 British nationals back into work. The reforms will also save the taxpayer £50 billion a year. That’s £1,700 per family. These funds will support tax relief and public services for British families. This paper lays out, in detail, how we will achieve these savings. It has been developed over many months in consultation with experts, many of whom have previously served in DWP.

Foreword from Nigel Farage and Robert Jenrick

Britain used to be proud of our welfare state. It provided security to people who fell on hard times, dignity to those who could not support themselves, and help to those who could return to work. Just as importantly, it was fair to the people who paid for it. The current system is nothing to be proud of. Millions of working-age people are now dependent on benefits. Disability and incapacity claims have almost doubled in less than a decade. More than a million foreign citizens receive Universal Credit. And billions of pounds are lost every year to fraud and error. The result is a system that is failing almost everyone. It fails workers, who are asked to pay ever-higher taxes to support a growing welfare bill. It fails the severely disabled, because the system is overwhelmed by assessing and managing mild ailments. And it fails many claimants themselves. Work is about more than a paycheque. It provides independence and dignity. A welfare system that allows people who could work to remain dependent is not compassionate. It is a moral and fiscal catastrophe. Our constituents who need and rely on PIP should, and will, see their awards protected, but the system as a whole is not sustainable. Successive Conservative and Labour governments have known about these problems. Since Covid, neither has been willing to confront them. We will. Under Reform, the welfare system will be based on common-sense principles. If you can work, you should. If you cannot work because of a severe disability, society should support you. If taxpayers provide for you while you look for work, you should give something back to the community. Welfare funded by British taxpayers should be reserved for British citizens. And if you cheat the system, you should expect to be caught. This paper sets out how we will put these principles into practice. By 2030, we estimate that this programme will put 241,000 more people into work and save the average family £1,700 a year, if implemented this year. We achieve this while protecting those who need the State’s support, including pensioners and disabled children. Those savings will allow us to reduce taxes on working people and invest in public services. But the case for welfare reform is bigger than the numbers. It is a moral crusade to restore fairness to a welfare state we once prided ourselves on. We hope you will join us. Nigel Farage Leader

Reform UK

Robert Jenrick Shadow Chancellor

Reform UK

CONTENTS

Summary

Foreword from Nigel Farage and Robert Jenrick

Britain’s Welfare Problem Welfare to Work

British Welfare for British Citizens

Fairer Disability Benefits

Combatting Fraud and Error Technical and Already-Announced Changes

Overall Costing

Conclusion

Appendices

Endnotes

Britain’s Welfare Problem After the Global Financial Crisis, welfare reform and falling unemployment gradually reduced welfare spending. Since Covid, however, Labour and the Conservatives have lost control of the numbers. Today, with an unemployment rate of 4.9%, welfare spending per household is around where it was in 2009/10, when the unemployment rate was 7.9%.5 Out-of-Work Working-Age Benefit Claimants, Great Britain, Feb 2013 to Aug 20256 As of August 2025, 9.3 million working-age people in England and Wales were claiming some form of welfare. This is the highest number on record, surpassing the Covid-era peak. Of this, 6.6 million are on ‘out-of-work’ benefits.7 Roughly 300,000 of these claimants are ‘work-ready’ and have been for more than a year. 8 Currently, they must look for work, but they are not required to contribute actively to their community. This makes staying on benefits more attractive than it should be, and it is unfair to the taxpayers who work to support them.

Worse, many of those claiming British benefits are not British citizens. As of February 2026, 1.3 million people who are not British or Irish citizens (or otherwise have ‘right of abode’) were claiming Universal Credit.9 Between April 2022 and February 2026, we estimate that £55 billion in British taxpayers’ money was paid to foreign citizens through Universal Credit or equivalent.10 That is £1,900 per family. But perhaps the single biggest contributor to spending growth is the explosion in disability and sickness benefits. In 2018/19, 2.2 million working-age people were receiving incapacity benefits. In 2026/27, this is already 3.9 million. By 2030/31, it is expected to reach 4.1 million.11 British people have not become 77% sicker in the past 8 years: something is fundamentally broken in our disability system. Finally, fraud and error remain far too common in the welfare system. According to the Office for Budget Responsibility, 8.5% – or £1 in every £12 – of Universal Credit spending is fraudulent or mistaken.12 These are not separate problems. They are symptoms of a welfare system that has lost sight of its purpose. The welfare state should protect British people who genuinely cannot support themselves, while helping everyone else back towards independence, including those with temporary or mild health conditions. Instead, too many people remain on benefits for years with no pathway back to contributing. The following sections set out how Reform will put that principle into practice.

Welfare to Work At the last count, more than 330,000 work-capable adults were claiming Universal Credit on Seeking Work conditions and had been doing so for more than 12 months, as British firms cried out for workers.13 That is not sustainable. Reform will make it cheaper and easier for these people to find work. First, we will reverse Rachel Reeves’s tax on British workers. Second, we will reverse Angela Rayner’s so-called Employment Rights Act, which has stopped firms from hiring new workers. Third, we will reinvigorate British business by getting energy prices under control, raising the VAT threshold, and cutting the myriad red tape that gets in businesses’ way. But if people are still not working after these reforms, they ought to contribute to their local communities instead. That would be a fair return to the taxpayers that fund their lives. If taxpayers must work, so too should the recipients of their support, if they are able to. Requiring claimants to work will make long-term welfare dependence less attractive as a lifestyle option. This should encourage claimants to seek out work more actively. Meaningful work would also give claimants a chance to return to a work routine and show employers they can be productive employees. Our Welfare to Work plan would require work-capable welfare claimants who have claimed benefits for more than a year to work 20 hours a week in placements assigned by local councils. Tasks could include cleaning or beautification, clerical work, or minor construction and repair work. Placements could be with the council itself or with approved locally registered charities. This plan sets out the scheme’s fundamental parameters. Given the urgency of restoring fairness to our welfare system, we seek to begin legislating to achieve this within the first 100 days of a Reform Government. Accordingly, we have established a Policy Delivery Group (PDG) to develop detailed guidance to enable local authorities to implement this policy rapidly. Robert Jenrick MP will chair the PDG, with membership including Lee Anderson MP and current Council leaders and portfolio holders from Reform UK’s councils. Those members are: Cllr Stephen Atkinson, Cllr Diane Morton, Cllr Mark Arnull, and Ben Bradley. Policy Detail If training or certification were required for the task – such as a DBS check or a CSCS card – the local authority would provide them. This training or certification could help the employee find unsubsidised work in the future. Failing to participate, absenteeism, poor performance, or on-the-job misconduct would result in immediate sanctions. This could include the full non-time-limited revocation of benefits. This would be a much harsher sanction than those in previous workfare programmes both in Britain and abroad. The placements would be part-time to allow claimants to continue their job search concurrently. Work-search obligations would remain in place. The placements would be a condition of receiving benefits; they would not be employment for employment law purposes. This policy is modelled on Ireland’s highly successful Community Employment and Tús programmes.14 According to the OECD, these policies make it more likely that beneficiaries find work, increase their earnings, and reduce claim rates for disability benefits.15 By the fourth year after participating, a participant is 8 percentage points more likely to be employed than a non-participant.

Requiring claimants to work for their benefits is not a new idea in British politics. In fact, in some ways, it is the historic norm. John Major had ‘Project Work’, Tony Blair had ‘New Deal’, and Iain Duncan Smith introduced ‘mandatory work activity’ under the Coalition Government. Each of these schemes had its drawbacks, but they were directionally sensible. The schemes were shamefully abandoned, however, because the Conservative Government succumbed to legal challenges from activist left-wing lawyers. Those legal challenges took two forms: Technical challenges under domestic statutory frameworks and challenges under human rights or other international treaty frameworks. In R (Reilly and Wilson) v Secretary of State for Work and Pensions, the Supreme Court held that the 2011 regulations that introduced Iain Duncan Smith’s Mandatory Work Activity gave ministers too much discretion and failed to describe the schemes with sufficient specificity. They were therefore struck down. The Government then tried to retrospectively validate sanctions imposed under the defective regulations. In Reilly (No 2), this was struck down. That effectively marked the end of the Coalition’s workfare experiment. Despite the ballooning welfare claimant counts under successive Work and Pensions Secretaries, including, in particular, Mel Stride, the policy was never countenanced again. Reform will never be deterred from delivering common-sense proposals by the threat of legal challenge. To address this, we will introduce specific primary legislation enabling our policy within the first 100 days of a Reform Government. We will also leave the ECHR and repeal the Human Rights Act. That means that – unlike the Tories or Labour – we will be able to deliver a welfare system that is fair for all. Once it matures, this programme is expected to save taxpayers more than £1 billion a year. Work obligations would apply to Universal Credit claimants with the Seeking Work conditionality who have been continuously claiming that benefit for more than 12 months. Short breaks (less than 3 months) in benefit claims would not reset this count. The obligation would begin once the council assigned the claimant to a placement. Councils would assign tasks to claimants based on local needs. As in Ireland, tasks might include:

operations,

A full list of 87 specific role types that we have identified as being potentially suitable for inclusion in the scheme is in an annex to this plan. The Policy Delivery Group will meet and consult on a final list of suitable roles to be provided as guidance to local authorities at the time of the scheme’s implementation. The local authority would receive a grant of £3,200 per claimant per year to cover the cost of administering the programme, supervision, and training the claimant for their task. If a local authority could not find adequate work for claimants in its area, it could assign them to a neighbouring area.

In addition to the local council managing and supervising these tasks, registered charities could apply to manage scheme participants. This would allow good causes, whose importance has not hitherto been recognised in funding decisions, to benefit while expanding the list of possible functions. We estimate that the economic value of the work done by each claimant would be £9,000 per year, of which 50% might be expenditure that the council would otherwise have to incur. The remainder would deliver additional services to the community. Claimants would receive up to £30 per successfully completed week, on top of their existing benefit, to cover additional expenses such as transport and food. Any claimant who accepted an offer of non-trivial employment could leave their placement immediately. We would gradually roll out the programme to allow local authorities to find high-value uses for claimants, but by the fourth year, two-thirds of eligible claimants would be participating. Should a Reform Government be elected, this scheme will have the democratic mandate of the British people. Reform takes implementing their will seriously. Accordingly, councils that refused to implement this policy, failed to find sufficient placements, or failed to monitor scheme participants’ performance effectively would be penalised with lower or no central Government funding. We will also legislate to make its implementation a legal duty for councils. Costing Full details of methodology available in the appendix. The OBR forecast for unemployment, on which our long-term claimant forecast is based, concludes in 2031.

British Welfare for British Citizens Mass migration has overwhelmed the British welfare state. Many foreigners, with no particular ties to this country nor an individual record of contribution, are currently subsidised by British taxpayers. The Conservatives allowed welfare payments to foreign nationals to balloon. Over 1.3 million foreign nationals are now on Universal Credit, up from around 900,000 in 2022. Department for Work and Pensions data shows the share of Universal Credit paid to foreign nationals was 15.5% in January 2026 This has happened as our economy has become increasingly dependent on overseas labour while neglecting domestic skills, productivity and worklessness. Between January 2020 and December 2025, the number of non-EU under-25s on UK payrolls rose by 289,000, compared with an increase of just 11,000 among young UK nationals. Meanwhile, the number of young people outside employment, education or training rose by almost 200,000. Migration, labour-market and welfare reform must therefore be considered together: reducing reliance on imported labour requires stronger incentives to recruit, train and pay British workers, alongside a welfare settlement that prioritises British citizens. Reform will deliver all of these. By 2029, we forecast that around 1.5 million non-citizen adults will receive Universal Credit and around 215,000 will receive PIP.19 This is completely unsustainable. This is not just an economic catastrophe but a fundamental unfairness. British taxpayers pay – to the tune of thousands of pounds a year – for a welfare state on the understanding that it is there to support our own on a reciprocal basis. Foreign nationals have not contributed on an equivalent basis and accordingly, ought not to qualify for the benefits. While some temporary migrants are subject to ‘no recourse to public funds’ (NRPF) rules, access to welfare expands substantially once people obtain indefinite leave to remain (ILR) or qualifying status under the EU Settlement Scheme.

Under Reform, only British citizens will be entitled to almost all benefits. 20 In particular, we will restrict access to Universal Credit, Housing Benefit, most other non-contributory benefits, and taxpayer-funded free childcare to British citizen adults. These rules will apply to both new and existing claimants. Foreigners who made the United Kingdom their permanent home could access benefits only if they sought and were granted citizenship. In the coming months, Reform will set out further proposals to reform and strengthen that system. Only a limited number of exceptions will apply. Non-citizen residents will retain access to contributory State Pensions, Armed Forces compensation, and War Widows Pension. All recognise substantial contributions to the United Kingdom. No child will lose their eligibility for free school meals. Foreign nationals will retain access to the NHS, subject to an increased Immigration Health Surcharge – which will ensure the taxpayer is not left out of pocket. If only one spouse in a couple is British, only the citizen adult would remain eligible for welfare. However, their non-citizen spouse’s income and assets would continue to count against eligibility thresholds. A British-citizen child of non-citizen parents would retain their eligibility for Child Benefit. By 2029/30, this reform will save British taxpayers £21 billion a year, of which £13 billion comes from Universal Credit. These savings will compound over time, as the expected growth in foreign-national claims does not occur. Values in nominal £mn International Law Under the Conservative Party’s Brexit deal, the British government is required to treat EU citizens resident in this country before 2021 on largely equal terms with British citizens. We would renegotiate the treaty and disapply this provision for welfare eligibility. Entitlement in domestic law would be amended according to the table below. We have allowed a £500 million indexed contingency for the possible costs of British citizens in the European Union returning to the United Kingdom, if their benefit eligibility in those countries is withdrawn. There are only 1.3 million British citizens in the European Union, excluding Ireland.21 Our contingency allows for the costs of approximately 10% of them returning home and immediately drawing benefits; therefore, we do not anticipate it being exhausted.

The following benefits will be limited by citizenship in particular:

Situation British Citizens resident in the United Kingdom Entirely foreign household, under state pension age Foreign adults with a British citizen child British Citizen with Foreign Spouse Foreign pensioner with sufficient NI contributions

Benefits

*Excludes Armed Forces Benefits All currently available (subject to other changes in paper) Universal Credit

Housing Benefit

Jobseeker’s Allowance and other benefits* Tax-free Childcare Child Benefit Free School Meals for children Universal Credit

Housing Benefit

Jobseeker’s Allowance and other benefits* Tax-free Childcare Child Benefit Free School Meals for children All currently available (subject to other changes in paper) State Pension Pension Credit and other benefits Eligible under

Reform Plan

Yes No No No No No No No No No No Yes Yes British citizen retains eligibility Yes No New Eligibility Criteria

Costing In nominal £mn

Fairer Disability Benefits In 2025, 6.9 million people received disability benefits in Britain. That’s 2.5 million more than 20 years ago.22 Every month, more than 20,000 new people are granted PIP. And once they’re on PIP, they’re likely to stay there: Two-thirds of 2017’s new PIP recipients were still claiming it five years later.23 As a result, health and disability benefit spending is forecast to reach £110 billion a year – £3,600 per family – by 2031.24 Mental health claims have largely driven this growth. In 2002, 25% of disability benefit claimants had mental or behaviour conditions as their main complaint. By 2024, this had risen to 44%, or 1.6 million people.25 In Labour’s first year, almost 250 people were awarded benefits for anxiety or mood disorders every single day.26 The state must provide for our fellow citizens who cannot provide for themselves. Mental health conditions can be just as debilitating as physical ones. But we cannot accept an ever- growing share of British people being dependent on welfare and often out of work. It robs them of dignity, and it isn’t fair for working people. Our reformed disability benefit system will keep more people off the benefit rolls and reduce the burden of each claimant on taxpayers, while protecting the severely disabled. First, we will reduce the number of people ever entering the disability benefit system. A new system of ‘Return to Work Cover’ will cover the first two years after an employee signs off sick (after Statutory Sick Pay expires). During this period, employers or their insurers will be responsible for their employees’ benefits. This will ensure that the connection between employees and employers is not severed. The evidence from the Netherlands and Finland is that such a system is much more effective at helping people return to work. 27 Employers can make adaptations to welcome people back, while insurers have an incentive to invest in rehabilitation. Employer NICs will be cut (by 0.2pp, in the first instance) to ensure overall employer costs do not increase, and the requirement will not apply to employers with five employees or fewer. Second, we will implement a single rigorous Disability Needs Assessment to determine eligibility for both health-related income support and disability payments. This would replace the separate Universal Credit Health and PIP assessments. This will reduce the administrative burden for claimants and DWP alike. It will also ensure consistency across the system. Third, we will retain but fundamentally reform PIP for working-age claimants. Nothing will change for those with severe and enduring disabilities - those in real need. They will still receive Personal Independence Payments (PIP/HSA). Those in need of milder support will be able to receive PIP support via Disability Support Accounts that we will create. Such support will be administered by local authorities and pay only for verifiable additional costs arising from a disability. This will ensure that the welfare system does not encourage people to consider themselves disabled. Existing working-age PIP claimants for mental health or trivial conditions would be reassessed and brought into the new system over the next three years. Other claimants would only be reassessed after three years, in line with normal reassessment timelines. Once all reassessments have concluded, we estimate that 2.16 million claimants would retain their present cash entitlement. This would be concentrated among mental health and musculoskeletal claimants, where recent growth has been most dramatic. However, many people would retain access to a Disability Support Account as needed. We will also reform the children’s Disability Living Allowance for anxiety, depression, and ADHD to align it with our changes to the adult system. Between 2013 and 2023, the number of children receiving child DLA has doubled. Today, in an average class of 32 children, two

of children receiving child DLA has doubled. Today, in an average class of 32 children, two receive disability benefits. As with adult disability benefits, this has been driven by mental health claims. Our changes will bring this spending under control, but it will apply only to future claims for a subset of mental health conditions. We will make no changes to the rules affecting physical disabilities and children currently receiving support. Fourth, we will reform the health premium on Universal Credit to make it coherent with the reformed Personal Independence Payment system. Those with severe cases and in real need will retain existing support. The specific disability payment will cover the additional costs of being disabled; no additional income support will be needed on top of that. The Disability Needs Assessment will continue to determine whether health-related claimants must seek work. Finally, we will spend an extra £1.85 billion a year on interventions that are proven to help people stay well and off benefits. This includes cognitive behavioural therapy for mental health conditions and physiotherapy for musculoskeletal conditions. It also includes employment support funding. The rapid growth in disability benefits, overseen by both Labour and the Tories, is a moral and fiscal crisis. Only Reform can and will take the bold steps necessary to make the system fair again. These proposals will achieve that. How Did We Get Here England and Wales currently have two primary sickness or disability benefits for working- age adults. The first, the Personal Independence Payment, is meant to cover the additional costs of disability. It can be layered on top of other benefits or received while in work. The second, the Universal Credit Health Element, is designed for people out of work because of their condition. Neither is working as intended. PIP was introduced under the last Conservative government and was intended to make disability assessment more rigorous and consistent. Its predecessor, DLA, relied heavily on self-assessment, around 70% of recipients had indefinite awards, and there was no systematic process for reviewing the existing caseload. 29 Iain Duncan Smith argued that PIP would instead be based on a “more objective assessment of individual need”, while continuing to pay people both in and out of work. Meanwhile, Universal Credit’s health element replaced the Blair/Brown-era Employment and Support Allowance, which itself replaced the old Incapacity Benefit. It has two separate components that distinguish it from standard Universal Credit for those who are unemployed or on low incomes. First is the additional payment that recipients receive over standard Universal Credit. Second is the exemption from work-search obligations. UC was meant to ensure that work always paid and to reduce overall dependence. Since Covid in particular, both components of the system have failed. The UK has seen an explosion in disability and incapacity spending, which is unmatched in comparable countries, as even the Tony Blair Institute has acknowledged.30 The Government’s own review has found that PIP is “no longer fit for purpose”.31 The vast majority of the increase in PIP caseloads, in particular, has been driven by mental health conditions. The eligibility criteria for receiving PIP and enhanced-rate PIP have gradually expanded over time, by both ministers and judges. This has made claiming for mental health reasons easier. For instance, the judiciary has narrowed the distinction between physical and mental conditions by finding that the physical manifestation of psychological conditions can allow people to access mobility-related PIP. Similarly, they have substantially loosened what it means for a condition to appear ‘intermittently’.32 Together with the labour market shock of Covid, this created an environment ripe for claims.

Disability benefits prevalence by primary reported condition, England and Wales33 In addition to more claims, people are staying on disability and incapacity benefits longer. In 2025, less than 1% of Universal Credit health claimants left the benefit rolls to a job every month.34 In 2012, this was twice as high. The result is that more people are getting stuck on disability and incapacity benefits for longer.

The Reformed System in Summary Our reformed disability support system will separate short- and long-term benefits. This will ensure that a temporary injury or illness does not turn into a lifetime of dependence. Return to Work Cover will provide the first two years of coverage for those newly entering the system from work. If a disability persists for more than two years, the redesigned disability benefits system will step in. This will provide unconditional cash support only to severe cases. However, it will cover verifiable additional costs for others. The diagram below shows a claimant’s journey through the reformed system. Return to Work Cover The current British system pushes people too quickly into state-funded incapacity benefits. Statutory sick pay lasts only 28 weeks. After that, the link between employee and employer is often severed. This makes it much harder for a temporarily disabled person to return to work. The Dutch system requires employers to cover their employees’ costs when they become ill for up to 2 years. This encourages employers to make the necessary adjustments to bring someone back to work quickly. It also means that employers – not an uninterested bureaucrat – can provide rehabilitation and monitor recovery. When this system was introduced, it reduced applications for disability benefits by 40% and increased labour force participation by 1.2 percentage points.35 Reform will introduce a similar system, Return to Work Cover, designed to be cost-neutral for businesses. This is subject to consultation with employers and ongoing detailed design work.

Firms that are required to provide Return to Work Cover will receive a cut in employer- side National Insurance Contributions to cover the average expense to provide the statutory minimum cover. We estimate this will be a 0.2 percentage-point cut initially. We will ask the OBR to monitor the programme’s cost to employers. If it is higher than estimated, we will reduce NICs further to ensure that it imposes no net cost on employers. Most employers will be required to purchase Return to Work Cover to cover the first two years of an employee’s disability benefits. Insurers may experience-rate this cover to encourage employers to bring employees back quickly. However, anti-discrimination rules will ensure insurers cannot charge different rates to employers based on the individual health or welfare histories of their workers. Larger employers will be able to choose to provide Return to Work Cover themselves. For those in employment when they become ill or disabled, Return to Work Cover would provide cover for two years after Statutory Sick Pay expires. If the employee still has not recovered by then, they could seek a Disability Needs Assessment for state support. For the self-employed or those employed by small firms, DWP will provide Return to Work Cover directly. Return to Work Cover would provide an amount equal to the current Universal Credit – Limited Capability for Work-Related Activities lower rate, except in certified severe cases, when it would be equal to the HSA rate. Employers could choose to provide more support, if they wished. Employers – or their insurers – would be responsible for determining, in the first instance, whether an employee was fit to return to work. Disability Needs Assessment Reform will require all prospective claimants of incapacity-related benefits to undergo a single Disability Needs Assessment. This will replace both assessment against PIP’s ‘daily living and mobility descriptors’ and the Work Capability Assessment for Universal Credit in the current system. This one assessment would determine a series of separate questions: whether the claimant qualifies for protected health income through the PIP/HSA; whether they have verified additional disability-related costs that should be met through a Disability Support Account; and what work expectations should apply within Universal Credit, if they are a UC claimant. Our proposal goes beyond the Government’s current plans to abolish the Work Capability Assessment. Under the Government’s plans, everyone signed off as having no work obligations will receive additional PIP support, even if their disability imposes no verifiable additional costs. We will keep these issues separate, while combining the assessment. The assessment would begin with structured medical evidence, collected by a GP. This would cover diagnosis, severity, duration, prognosis, treatment history, expected recovery or deterioration, treatment burden, functional impact, mobility, supervision, safeguarding risk and evidenced additional costs. GPs would certify medical facts; the Disability Needs Assessment would determine entitlements to support for.

The first gateway would determine whether the claimant qualifies for the PIP/HSA. As laid out below, this will be reserved for serious cases. The second decision would determine whether the claimant should receive a Disability Support Account. This would be separate from the cash decision. Claimants who do not qualify for the PIP/HSA could still receive practical support where additional costs are independently verified. The third decision would determine the claimant’s Universal Credit conditionality band. Receiving a DSA would not automatically remove work expectations. Conditionality would depend on what activity is reasonable given the claimant’s health, capacity, treatment and recovery prospects. Some claimants would have standard conditionality, some would have health-adjusted requirements, and PIP/HSA recipients would have no ordinary work-search or work-preparation requirement. The new assessment would consider condition, functional impact, prognosis, treatment history, risk and medical evidence together. Certain conditions will automatically qualify for the highest level of support. This will include profound untreatable deafness, blindness, severe physical or intellectual disabilities, and terminal disease. Most conditions will be directly assessed against the criteria. Others will face enhanced scrutiny. This will include ADHD, anxiety, and depression. Claimants with severe or atypical presentations could still qualify, but would need stronger independent evidence of severity, durability, prognosis, risk and verified additional costs. Personal Independence Payment and Health Security Allowance Under the reformed system, cash support above standard Universal Credit rates would be reserved for severe cases and delivered through a new Personal Independence Payment and Health Security Allowance (PIP/HSA). This would include those suffering from terminal illness, severe permanent physical disability, profound learning disability, severe autism with high support needs, severe enduring mental illness, major neurological disease and serious brain injury. It would also include those who are highly dependent on support for daily living, have irreversible or highly unlikely-to-improve conditions, or otherwise suffer from serious, unstable or deteriorating conditions where prognosis, treatment burden or safeguarding risk justified continued cash support. Claimants receiving the allowance would not face ordinary work requirements, although they could work where they can and want to. As noted above, existing PIP and UC Health claimants for mental health or trivial conditions will be reassessed over a three year period under the reformed system. If they met the new severe, enduring or high-risk gateway, they would retain the cash value of their existing UC and PIP awards, subject to uprating. Other claimants will only be brought into the reformed system in three years, at the earliest, by which time they will already be facing reassessment. New claimants with severe conditions would receive a single flat rate payment. It would sit alongside the standard elements of Universal Credit and provide a protected income- replacement level for people whose condition is so severe, enduring or high-risk that they cannot reasonably be expected to support themselves through work. The initial rate would be £429.80 a month at 2026/27 rates, equivalent to the higher LCWRA element currently available to people with severe lifelong conditions. A single claimant aged 25 or over would therefore receive £854.70 a month through standard UC and this payment, before housing, child, carer or other relevant UC elements. New claimants would also be eligible for a Disability Support Account to cover the verifiable additional costs of their disability or equivalent cash payment at current PIP rates. Existing claimants that retained the cash value of their current awards would not receive a DSA in addition to that protected cash.

Disability Support Account Disability Support Accounts will cover the verified additional costs a claimant suffers due to their disability. They will be available to anyone who received a Disability Needs Assessment that determined they faced such costs, including PIP/HSA recipients. DSA support could include equipment, aids, adaptations, mobility support, transport, communication support, personal assistance, employment-related adjustments or other approved services. We propose using a test similar to that in the VAT Act, which has already been heavily litigated, to ensure that only truly additional disability costs are included. It would not include unmodified motor vehicles, as currently provided by the Motability scheme, which would remain only for PIP/HSA claimants. DSAs would be delivered through local authorities or mayors, funded by a ringfenced Local Disability Support Grant. The central government would set the national rules, including the severity gateway, eligible uses, evidence requirements, maximum allocations, review rights, and audit requirements. Local authorities would administer DSAs, commission services, and work with local providers. Using local authorities to deliver this funding will ensure that it is auditable and accountable to communities. It will also open up scope for local innovation and partnerships. The Policy Delivery Group of local government leaders established by Reform UK will develop additional guidance in the coming months so this policy can be rolled out at a pace commensurate with the urgency of this mission. Child Disability Living Allowance We will change Child DLA for anxiety, ADHD, and depression to reflect the changes we have made to working-age disability benefits. However, these changes will only apply to new applicants for the benefit. Nothing will change for existing claimants. It will also not affect claimants of child DLA for any other condition. We will reinvest around one-fifth of the child DLA savings into treatment options directly targeted at these mild conditions. Costing

Combatting Fraud and Error According to the OBR, £1 in every £31 (3.2%) of taxpayers’ money spent on welfare should not have been, even under today’s relatively permissive rules. This money is spent because of either deliberate fraud or error. The fraud and error rate for Universal Credit is even higher, at 8.5% – or £1 in every £12 spent.36 That’s £260 stolen or taken in error from the pockets of every family in Britain, every year.37 Some of this is fraud on an industrial scale. In 2024, five people were convicted of making almost £54 million in fraudulent Universal Credit claims. 38 But most is more prosaic. As the OBR reports, the main cause of overpayment is underreporting of income.39 There are existing government programmes that work to reduce this waste. But they are simply too small. In 2024/25, the DWP eliminated £2 billion in fraud and error with a £700 million investment in dedicated fraud and error activities.40 This 3-to-1 return on investment is an excellent result for taxpayers, so it should be scaled further. Reform would substantially expand the anti-fraud and error effort. First, we would reset the Targeted Case Review (TCR) back to the previous Government’s timetable and widen it from Universal Credit to all benefits. Second, we would increase the frontline counter-fraud headcount to 9,000 above its current baseline, adding 6,000 officers on top of the 3,000 uplift already announced. Artificial intelligence could multiply the impact on this force. This is notwithstanding our commitment to significantly reduce the overall size of the civil service. Costing

Technical and Already-Announced Changes Reinstating the Two-Child Benefit Cap As Robert Jenrick announced in February, Reform will fully reinstate the two-child benefit cap. Taxpayers cannot be expected to provide limitless support for ever-larger benefit- dependent households. It is simply unfair. Reintroducing the cap will save an average of £2.9 billion a year over the next eight years, with the funds pre-committed to funding our Save Our Pubs policy. Consequential Effects on the Benefit Cap Under current rules, a single adult outside Greater London cannot receive more than £1,229.42 in benefits each month. This ‘benefit cap’ is meant to ensure that work always pays. However, the benefit cap does not apply if a claimant or their immediate relative receives PIP or child DLA. As our policy reforms PIP and reforms child DLA, fewer people will be eligible for benefit cap exemptions. This will further reduce welfare spending. These effects are included in the overall costing of the disability benefit reform. Reforming Indexation The Consumer Price Index (CPI) used to uprate working-age benefits each year systematically overstates the true rate of cost-of-living inflation. This is because the Office for National Statistics (ONS) is behind both international best practice and the academic frontier in its calculation method. In particular, it does not adequately account for how people respond to higher prices for certain goods by switching to cheaper alternatives. Reform would establish an independent commission to consider updates to the CPI methodology. It would be similar to the British Johnson review, which recommended that CPIH be used as the main measure of inflation in 2015, and the American Boskin Commission, which found the American CPI was substantially overestimated. 41 This commission would review the academic literature and recommend ways for the ONS to improve the CPI. The exact savings will depend on the commission’s recommendations. Nonetheless, we estimate that correcting for three separate forms of substitution bias could reduce the annual measured inflation rate by 0.6 percentage points a year on average.42 Over time, this would add up to very substantial mismeasurement. We will use this new CPI measure to uprate working-age welfare benefits. The savings delivered by this change become significant quickly: Tapering Disability Benefits by UC Rules The new Health Security Allowance we propose above will combine both Universal Credit – Health (UC) and the Personal Independence Payment for those who are severely or permanently disabled. UC is currently means-tested, but PIP is not. This means taxpayers are providing substantial support to households that can support themselves. To make this fairer, the total Health Security Allowance will be means-tested under standard UC rules. This means that the amount people will receive will depend on their own earnings, their partner’s earnings, and their assets. Disability-related earnings – i.e., the Carer’s Allowance and contributory ESA – will not count towards these tests. This will not affect existing grants until reassessment.

This change saves an average of £1.8bn a year over the next 8 years, with the savings coming largely from high-income households. Devolution Consequences Welfare is generally not devolved to Wales or to mayoral authorities in England; therefore, the only areas where effects may differ are Scotland and Northern Ireland. In Northern Ireland, social security is devolved, but the Northern Ireland Act requires Westminster and Stormont to seek to achieve “single systems of social security, child support and pensions for the United Kingdom”. 43 Therefore, we assume all of our changes will be extended to Northern Ireland. In Scotland, only some parts of the welfare system are devolved. This means the effects of our reform will depend on which benefit is affected, as follows:

MEASURE EFFECT IN SCOTLAND

Welfare to Work Universal Credit is not devolved to Scotland; therefore Welfare to Work will apply without any changes. British Welfare for British Citizens The citizenship requirement would automatically apply to all reserved benefits (e.g., Universal Credit). For devolved benefits, this would be a matter for the Scottish Government. Fairer Disability

Benefits

The health element of Universal Credit is not devolved; therefore its reform will be extended to Scotland. PIP is devolved, so its reform will not necessarily be extended. The Scottish Government will choose whether to adopt these reforms. (The same is true for our limited reform to Child DLA, which is devolved) If the Scottish government does not reform PIP, Health Security Allowances for Scottish claimants will be lower to avoid duplicating devolved PIP support; however, they would still cover the reformed UC health elements. Scottish claimants would also not be eligible for Disability Support Accounts. Reforming Indexation Indexation reform will automatically extend to all reserved working-age benefits. The Scottish Government may choose to apply it to devolved benefits. Block Grant Adjustment Whether or not the Scottish Government follows our proposed reforms for devolved benefits, the Scottish block grant adjustment will be mechanically reduced to maintain equity between England and Scotland. We have applied this estimate using the latest population-based share used in the Block Grant Transparency report.

Overall Costing Note: These policies are primarily the newly announced package of measures in this plan, but also include previously announced Reform policies, such as the two-child cap. Forecast welfare spending comes from the OBR’s Economic and Fiscal Outlook and therefore only runs until 2030/31. Our costing assumes the plan is implemented in this fiscal year.

Employment Effects Many of our proposed measures will increase employment. This is largely because they make remaining on benefits less attractive or more difficult. Some measures also make employment easier. In total, we estimate that 241,000 more British nationals will be in employment by 2029/30 than would otherwise be the case, if this plan were implemented in this fiscal year.44 The fiscal benefits of this are accounted for in the costing above. Measure Employment Effect by 2029/2030 Welfare to work +24,000 Reformed Disability Benefits +198,000 Return to Work Cover +11,000 Tapering of Disability Benefits -14,000 Indexation Reforms +22,000 Total +241,000

Conclusion

The basic bargain of the welfare state should be simple: Those who can work should work. Those who need support because of severe disability should receive it. British taxpayers should support only British citizens. And those who defraud the system should expect to be caught. The current system fails to deliver that basic bargain. Reform would re-establish it. Long-term, work-capable claimants will be required to contribute to their communities while they look for employment. Our reforms to PIP and Universal Credit will ensure that disability benefits will once again distinguish between severe, enduring conditions and problems for which treatment, rehabilitation and practical support are more appropriate. Access to most non-contributory benefits will be reserved for British citizens. Stronger enforcement will ensure taxpayers’ money goes only to those genuinely entitled to receive it. These reforms would save taxpayers billions. By 2030, net savings would be £50 billion or £1,700 per family. That money can be returned to the people who earned it through lower taxes, while helping Britain meet essential commitments such as defence and national investment. But they are not just about saving money. They are about changing what the welfare system does. Instead of paying people to stay out of the workforce, it should help them return to it. Instead of treating an ever-growing range of health problems as grounds for permanent dependence, it should support recovery wherever possible. Instead of being an automatic entitlement for anyone who acquires the right immigration status, welfare should be a reciprocal obligation of British citizenship. Parties of both colours have overseen the welfare state’s explosion. Only Reform has the courage – and the plan – to actually fix the problem.

Appendices Throughout this paper, figures are nominal, and positive figures denote an improvement in public sector net borrowing. If units are not provided, they are in £mn. Comparison with the Conservative Proposals The Conservatives have announced a proposal to cut welfare spending. The entire public plan is a two-page section of their Alternative King’s Speech.45 They suggest it will save £23 billion but have provided no methodology. This is how it compares to the Reform plan. Issue Conservative Position Reform Position Total Savings £23 billion per year (methodology unclear) £53 billion per year by 2030/31 Workfare No position; the previous government abandoned workfare. Introduce part-time work obligations for all long-term work-ready claimants; underpinned by primary legislation. Foreign citizens accessing British benefits EU settled status claimants (58% of foreigners on benefits) retain their benefits. All non-CTA citizens lose access to British benefits. Short-term illness ‘Reform the fit note system’. New short-term Return to Work Cover to keep employees connected to their employers; proven to work in the Netherlands and Finland. PIP and Motability Access reduced for low-level mental health conditions. PIP reformed; maintaining secure cash funding for severe cases; support only for verified additional needs for others. PIP assessments Brought back in-person. Existing mental health or trivial condition claimants to be re- evaluated under reformed rules by GPs over 3 years. Other claimants to be re-evaluated after 3 years. Work Capability Assessments No position. Merged into Disability Needs Assessments, which will sign fewer people off the obligation to find work. The overall benefit cap ‘[L]imit the extra benefits households receive if they are on an exempting benefit or if there is an adult in the household who is able to work but not working’. Solve the underlying problem by restricting eligibility for disability benefits. Combatting fraud and error No position. A clear road-map to reducing wasted taxpayer money.

Costing Approach for Work Obligations We estimate the Exchequer impact of the Welfare to Work programme by combining a per-place unit costing with a cohort-based projection of how participants’ benefit and employment statuses change over time. The model is built in three layers: exogenous assumptions (macroeconomic, unit-cost, and behavioural); per-participant costing at 2026- 27 prices; and annual aggregation to 2033-34. The unit costing establishes the cost of one placement to the state for one year. Gross cost per place is £4,782, comprising £1,560 of participant costs (a weekly expense allowance equivalent to £30 a week) and £3,222 of delivery costs (supervision, materials, training and central administration). These are based on the Irish policy. Against this we set the value of council work displaced by participant labour. Each place delivers 854 productive hours; valued at a fully loaded council cost of £22.00 per productive hour, this implies £18,781 of gross labour value, of which we assume 50% substitutes for work councils would otherwise have funded and that participants work at 50% of the productivity of the staff they displace. The resulting offset is £4,695, giving a net cost of £87 per place. Two further unit values are estimated on the same basis: each person moved into unsubsidised employment yields £11,487, being £6,388 in new tax revenue and £5,099 in foregone Universal Credit; each person diverted onto UC Health costs £2,607. This may overestimate the benefits of employment, as not everyone will move into a full-time job at the National Living Wage; nonetheless, we likely underestimate the savings from welfare, as we assume each participant is only on the basic UC for single adults. Each person who leaves UC without entering employment simply saves £5,099. Unit values are uprated annually from their 2026- 27 base — cost and benefit-rate lines by CPIH, earnings-linked lines by the average weekly earnings index, both taken from the March 2026 EFO — so that all figures are nominal. Population flows are projected from the March 2026 EFO forecast for ILO unemployment. We hold the 2023-24 ratio of UC “searching for work” claimants to unemployment constant and scale it forward, then deduct those who find work or migrate to UC Health and those who remain in the programme to obtain the eligible population for new referrals. Referrals are set at 50,000 in 2026-27 and at 200,000 per year thereafter, with referrals entering the model as eight annual cohorts. Each cohort is tracked separately by member status. The five statuses available are ‘on UC Seeking Work’, ‘in placement’, ‘on UC Health’, ‘off (working-age) benefits and not in employment’, and ‘in unsubsidised employment’. Each year, we estimate where each cohort would have been without intervention and where it would be with intervention. In particular, we follow evidence from Ireland and assume the policy will gradually increase the probability of employment by 8pp by year 5, except that in year 0 it will reduce the probability of employment by 1pp (the so-called lock-in effect), as workers are less able to look for work. We assume that these are net new jobs in the economy. We assume the number of people switching to UC Health does not increase because of this policy, as other elements of this proposal will make entering UC Health harder. We assume the exit rate – both to employment and to no support – also increases by 10% due to a ‘disincentive’ effect – i.e., people look for work harder in order to avoid the placements. Our estimate for this comes from the Dorsett and Smeaton (2008) review of the Intensive Activity Period – an older UK workfare programme. The change in the number of people in each status is multiplied by the unit cost of that status (relative to staying on UC Seeking Work) to determine the policy’s total fiscal impact on a given cohort in a given year. The policy’s total fiscal impact is the sum of the cohort impacts. The profile is front-loaded in cost and back-loaded in benefit, as people have to go through costly placement before the positive employment effects can be realised.

Costing Approach for Foreign National Savings We estimate the Exchequer impact of restricting non-contributory welfare to British citizens by applying measured non-citizen incidence rates to the forecast spending for each affected benefit, projecting those incidence rates forward to reflect changes in the eligible migrant population, and then adding a small behavioural response. Universal Credit is the largest component. In 2029-30, forecast UC expenditure is £92.159bn. We estimate that 17.71% of UC expenditure will be in households containing a non-citizen claimant, giving £16.322bn of spending potentially in scope. This share starts from the observed 14.77% of UC claimants in February 2026 and is projected forward using explicit flows into and out of the five principal immigration-status groups, including settlement and citizenship. Given that naturalisation of net recipients will be restricted, fewer grants of citizenship add roughly 0.27 percentage points to the share each year. We then apply a policy factor of 0.8349 because mixed-citizenship households lose only the difference between the couple and single standard allowances rather than their whole UC award, and deduct £572m to remove the overlap with the separate health reform. This produces a UC saving of £13.055bn in 2029-30. Most remaining benefits are costed the same way. As with UC, the foreign shares are projected to 2029-30 using the same settlement and citizenship stock-flow model. Foreign shares are expected to increase by 17% for working-age benefits, 8% for pensioner benefits, and 18% for child-related benefits. In 2029-30, this method gives savings of £883m from remaining PIP claims (now paid through the unified disability benefit system, but treated separately for this), £814m from Housing Benefit, £323m from Carer’s Allowance, £689m from DLA, £360m from Attendance Allowance, £151m from contributory ESA, £24m from bereavement and Industrial Injuries benefits, and £776m from Pension Credit. Child Benefit and free childcare are treated separately because eligibility is attached to the household and child structure. Child Benefit expenditure is £13.449bn in 2029-30; applying a 10.02% wholly non-citizen-adult-household share and the child-linked drift gives a saving of £1.591bn. Free childcare expenditure in England is projected at £10.807bn; applying an 11.05% share of delivered hours gives £1.194bn. We then add the behavioural effect of withdrawing benefits. Around 1.05 million working-age adults are estimated to be affected, of whom roughly 630,000 are initially out of work. The model assumes that 15–35% of this group moves into employment, calibrated on evidence from large benefit withdrawals, while those already working increase taxable earnings as the effective marginal withdrawal rate falls. Together these effects are estimated to raise direct tax receipts by around £1.0bn a year at maturity. Against this, withdrawing funded childcare reduces labour supply among affected second earners by around £100m a year. The net dynamic addition is therefore about £0.9bn in 2029-30. Costing Approach for Fairer Disability Benefits Reformed Disability Benefits System We estimate the Exchequer impact of the reform by rebuilding the working-age PIP caseload from administrative data, combining it with the UC Health caseload, and projecting the amount of cash entitlement withdrawn as claimants are reassessed into the reformed system. We then adjust the resulting benefit saving for interactions with Universal Credit, appeals, movement onto other benefits and the cost of the reformed support system. Our estimated saving is £20.15bn by 2033-34. For PIP, we reconstruct the register across 186 condition × daily-living × mobility cells, pricing each at its observed mean award. The protected group comprises enhanced daily-living awards outside the treatment-responsive conditions and 60% of enhanced awards within those conditions, giving a protected share of 47% of spending. The remainder is allocated

between higher DSA, lower DSA and no-cash groups. The PIP spending envelope is projected using the March 2026 OBR forecast, with £31.2bn of working-age PIP expenditure in 2028- 29. The UC Health side uses the March 2026 caseload and again assumes that 47% of expenditure remains protected. PIP and UC Health overlap is estimated from the Family Resources Survey so that claimants receiving both are counted only once. Savings are realised as existing claimants pass through the new Disability Needs Assessment. There are 2.89 million unique affected people, who will be reassessed over time, beginning with mental health and trivial claims in the first 3 years. We then allow 9.8% of withdrawn cash to be reinstated through mandatory reconsideration and appeals, reducing savings by £2.24bn, and assume 8.5% reappears as spending on other benefits, reducing savings by a further £1.94bn. This leaves a final AME saving of £19.65bn in 2028-29. Against this we deduct the DEL cost of delivering the reformed system. Recurring provision costs £4.44bn a year at full rollout, comprising DSAs, treatment, homecare, employment support and administration. A further £3.11bn of reassessment and IT costs is spread evenly across the three-year implementation period, adding £1.04bn in each of the first three years. Total DEL is therefore £5.48bn in 2028-29. From 2029-30, the one-off implementation costs disappear, but a £0.52bn annual standing reassessment cost remains, resulting in a steady- state DEL of £4.96bn per year. Finally, we add the savings from reforming Child DLA in line with our proposals for adults. In particular, we would restrict child DLA awards for anxiety, depression and ADHD to the most severe 10% of cases. We would make no changes to other conditions. This would apply only to new applicants, not existing claimants. The starting point is forecast DLA expenditure of £9.354bn in 2029-30. Children are projected to account for 83.03% of this, or £7.767bn. We estimate that 43.9% of child DLA expenditure will relate to the three affected categories — Hyperkinetic Syndrome, Behavioural Disorder and Psychoneurosis — based on their observed growth within the child caseload. This gives around £3.41bn of DLA expenditure in scope. However, initially, most of this is for current claimants and so protected. The unprotected share grows over time to almost 100% by the eighth year, as children age out of DLA. However, restricting eligibility to the most severe 10% of cases does not remove 90% of this expenditure. Severe cases receive substantially larger awards, particularly because they are more likely to qualify for enhanced care and mobility components. On average, retained cases receive roughly 2.5 times the award of an affected claimant. We therefore assume that retaining the most severe 10% of cases preserves around 22% of expenditure and removes 78%. The reform also reduces expenditure on entitlements linked to a child’s receipt of DLA. Roughly 60% of such children are in Universal Credit households, where the loss of the disabled-child element, at a blended value of around £2,400 per year, produces a further saving of approximately £650m. Carer’s Allowance adds around £300m, based on the proportion of affected middle- and higher-rate care awards for which a carer is assumed to claim the benefit. Finally, the model includes a modest fiscal gain of around £100m from the parental labour- supply response as some parents increase employment or earnings following the withdrawal of disability- and caring-related support. Adding these linked and behavioural effects to the £2.659bn direct DLA saving gives an all-in Exchequer saving of £3.709bn in 2029-30. Return to Work Cover We estimate the fiscal effect of the employer-liability system in two stages: first, the size of the liability imposed on employers and/or insurers; and second, the reduction in benefit spending resulting from the incentive to retain sick workers. The starting point is the annual flow of new health-benefit cases, not the existing stock. We use 271,401 new non-ESA-transition LCWRA cases per year, excluding 671,000 ESA

into incapacity. We assume that 34% of new cases are relevant employment exits and that the employer or insurer remains liable for a qualifying exit for 2 years. Given an average present value of this liability of £17,448, the total cost to employers and/or insurers is £ 1.6 bn in 2029/30. The £1.6bn is not counted as an Exchequer saving. Employers receive it back through a flat reduction of roughly 0.2 percentage points in employer National Insurance, making it cost-neutral overall for businesses. The charge instead changes the marginal incentive: an employer receives the NI reduction regardless, but the firm or its insurer bears a direct cost if one of its workers leaves employment and moves onto long-term health support. We then estimate how much this incentive reduces entries onto health benefits. Our central estimate is a 2.7% reduction in permanent labour-market exit, based on the Dutch experience, but scaled down for the much lower liability imposed under our system. Each prevented exit is valued at around £16,164 a year to the Exchequer. For severe cases, this includes the standard Universal Credit award, housing support, and the HSA that would otherwise have been paid; for non-severe cases, it includes standard UC and housing support, plus the associated indirect-tax effect. The savings build gradually as successive annual cohorts of prevented exits accumulate, reaching around £70m in 2028-29 and approximately £170m per year once the system matures. The costing is therefore cohort-based: we estimate the number of people each year who would otherwise leave employment for long-term health benefits, apply the assumed reduction from the employer incentive, and value the benefit expenditure avoided over the period during which those people would otherwise have remained on support. Costing Approach for Fraud and Error We estimate the effect on public sector net borrowing by first calculating the gross savings from the two fraud-and-error measures, then deducting savings that are already included in current Government plans, interactions with other reforms, and the additional departmental cost of delivering the policy. This produces a peak improvement in PSNB of around £2.8bn in 2029-30, settling at around £925m a year in steady state. The first component is the accelerated and expanded Targeted Case Review. The existing UC-only review is scaled up to cover fraud across the wider benefits system, using the share of total benefit fraud currently accounted for by UC and overlapping benefits. The second component expands frontline counter-fraud staff from the 3,000 already announced to 9,000 in total. The staff saving is scaled proportionately from the Government’s existing scorecard estimate. These two savings cannot simply be added together because both policies attack some of the same stock of fraud. The model therefore reduces the combined saving to allow for this overlap. This gives the gross fiscal benefit of the reform package before considering what would happen anyway under existing policy. We then deduct the savings already due to be delivered by the Government’s existing UC Targeted Case Review and its previously announced 3,000 additional counter-fraud staff. We also remove the share of savings attributable to foreign-national claimants that is already captured by the separate removals policy, avoiding double-counting between the two reforms. Finally, we deduct the departmental spending required to employ the additional 6,000 officers. The benefit savings in the underlying scorecards are gross AME savings and do not include the cost of the staff who generate them, so we must add this expenditure separately when calculating the effect on PSNB. After these deductions, the package improves borrowing

by £10.4bn cumulatively over 2026-27 to 2033-34, with the unusually large savings around 2029-30 reflecting accelerated reviews rather than a permanently higher underlying yield. Costing Approach for Indexation We estimate the saving by first estimating how far the existing CPI overstates increases in the cost of living. Our central assumption is that a revised measure would grow at 0.6 percentage points less each year. That figure combines three separate sources of measurement bias: 0.16 percentage points from substitution between broad categories of spending, 0.30–0.40 points from the treatment of new varieties within categories, and 0.05–0.10 points from substitution between purchases and household production. Together these imply a range of 0.51–0.66 percentage points a year; we use 0.6 percentage points as the rounded central estimate. The first 0.16 percentage points is measured directly from UK data. The existing CPI largely uses expenditure weights based on what households bought before prices changed. This means that when, for example, food becomes relatively expensive, and households respond by spending more on other things, the index is slow to reflect that substitution. We reconstruct CPI using a Törnqvist index, which weights spending by households both before and after price changes and is therefore designed to capture this response. We do this across 38 ONS spending groups, covering around 96% of the CPI basket, from December 1996 to March 2026. Over that period, the conventional CPI rose by 4.76 percentage points more than the reconstructed measure, equivalent to an average bias of 0.16 percentage points per year. The estimate falls only slightly, to 0.154 points, if the unusual pandemic years are removed, while using only the post-2005 period produces a larger estimate of around 0.20 points. We therefore retain 0.16 points. The second component is 0.30–0.40 percentage points from increased product variety. Conventional price indices measure how the price of an existing product changes, but they are less good at capturing the benefit consumers receive when new products and varieties become available. We estimate this using the standard Feenstra approach, which relates the value of additional variety to how rapidly varieties enter a market and how readily consumers substitute between them. We calibrate the assumptions separately for the twelve major CPI divisions, based on empirical studies of the relevant markets. For example, the resulting annual corrections are 0.40 points for food, 0.81 for clothing and footwear, 0.84 for communications and 1.01 for recreation and culture, but only 0.03 for housing and 0.01 for education. Weighting each category by its share of the UK CPI basket produces an aggregate estimate of 0.40 percentage points a year. We then apply a substantial haircut to that variety estimate. ONS basket updates already capture some new products, so it would be inappropriate to assume that the entire estimated variety gain is currently being missed. Discounting the 0.40-point estimate by 25% gives 0.30 points. The fiscal model therefore uses 0.30–0.40 percentage points as the plausible range rather than selecting the higher estimate outright. The final 0.05–0.10 percentage points captures substitution between goods and services bought in the market and things households can instead produce themselves. Examples include cooking rather than eating out, repairing rather than replacing something, or providing care at home. The calibration starts with evidence from Aguiar, Hurst, and Karabarbounis that unemployed households shifted about 50 minutes per day into home production during the financial crisis. Compared with roughly 150 minutes of normal daily home production, this implies that home-production time responds meaningfully to changes in the opportunity cost of people’s time. The paper translates this into an elasticity of around 0.33. The model does not simply extrapolate the unusually large recession response into normal years. Instead, it allows roughly 0.03–0.05 percentage points for continuing improvements that make home production easier — such as online services, instructional content, telehealth and telework — and around 0.02 points for the smaller group of people who are persistently

unemployed or weakly attached to the labour market and therefore have more scope to substitute time for market purchases. This gives the 0.05–0.10 percentage-point range used in the headline calculation. Adding the three components gives a plausible annual correction of 0.51–0.66 percentage points. The lower end combines the directly measured 0.16-point substitution effect with the conservative 0.30-point variety estimate and the 0.05-point home-production estimate; the upper end uses 0.16, 0.40, and 0.10, respectively. The midpoint is about 0.59 percentage points per year, which we round to 0.6 percentage points per year for the central fiscal estimate. This is also close to the 0.6-point residual CPI bias estimate produced by Lebow and Rudd following earlier US statistical reforms. The correction is applied to the OBR’s expenditure forecast for working-age welfare, less the cuts our other policy proposals would result in. However, only around 89% of that spending effectively follows CPI: disability and incapacity benefits, the main UC allowances, child- related support and most other benefits do, while much of housing support does not. The 0.6-point correction therefore translates into an effective first-year reduction of about 0.53% across the whole working-age envelope. The saving then compounds. In the first year, the affected benefits are around 0.6% lower than under existing CPI uprating. In the second year, they are around 1.2% lower because the second year’s smaller uprating is applied to a benefit level already reduced in the first year, and so on thereafter. Annex – Welfare for Work Possible Roles This annex lists the council functions initially identified as suitable for placement of work- fit long-term unemployed claimants, together with the likely training and vetting each role requires. Durations are indicative and assume group delivery. Costs, where given, are per participant at 2026 commercial rates. Many of the regulations identified in this annex may be removed or reformed under a Reform government. This should be taken as a merely indicative list, based on today’s regulatory environment. Similarly, the list of roles will be actively considered by the Policy Delivery Group and adjusted in accordance with their advice about what councils across the country need help with. Requirements applying to every placement The following apply to all roles in this annex and are not repeated in the tables below:

of non-participation (half day).

procedures, before first attendance at each new location.

appropriate to the role.

whether the role involves contact with children or adults at risk.

before the placement starts.

Note on vetting terminology Disclosure and Barring Service checks are not available at the discretion of the employer. The level obtainable is fixed by the duties of the role under Part V of the Police Act 1997, the Police Act 1997 (Criminal Records) Regulations 2002 and the Safeguarding Vulnerable Groups Act 2006. Authorities must not seek a higher level of check than the role attracts, and the scheme specification should require each host authority to record the eligibility ground relied on for every enhanced check. In this annex:

unspent convictions subject to filtering, plus relevant police intelligence.

regulated activity with children or adults; required for the roles at A.8 and marked instances at A.7. No role in this annex involves regulated activity in the sense of personal, intimate or health care. Where an enhanced check with a barred list check is specified, it is because the role involves unsupervised contact with children in a specified place, not because it involves personal care. Universal exclusions No placement under the scheme may involve any of the following, whatever the host authority proposes:

foreseeable confrontation with a member of the public.

powered plant.

A.1 Street scene and environmental operations Role Training required Vetting and other requirements Fly-tipping clearance (labouring) Manual handling; COSHH awareness; UKATA Category A asbestos awareness (half day) Works under a trained operative. No handling of suspected asbestos, clinical or chemical waste. Tetanus advice issued. Graffiti and fly-posting removal COSHH awareness (half day); face-fit test where solvent respirators are used Ground level only; no ladders or work at height. Solvent stock controlled and signed out by supervisor. Leaf clearance and gully- grate clearing Manual handling; Chapter 8 signing, lighting and guarding awareness (1 day) where kerbside Traffic management layout set by qualified staff— no entry to chambers or confined spaces. Litter bin emptying Manual handling; infection control and sharps awareness (half day) Hepatitis B vaccination offered; puncture-resistant gloves; sharps disposal protocol. Weed control on hard surfaces PA1 plus PA6 handheld applicator certificate (2–3 days, approx. £400–£600);

COSHH

Minimum age 16 for PA certification. Application records held and countersigned by supervisor. Wheeled bin delivery, repair and cleaning Manual handling; team lifting (half day) No operation of council vehicles. Kerbside delivery only; do not enter curtilage where a dog is present. Household waste recycling centre (HWRC) assistant Site induction; manual handling; traffic marshalling / banksman (1 day) Customer-facing: conflict avoidance briefing. No cash handling, no plant or compactor operation. Reuse shop sorting and refurbishment Portable appliance testing (1 day); manual handling PAT records countersigned by a qualified tester. No pricing, sales or cash handling. Bicycle refurbishment Internal cycle mechanic course (2–3 days) Roadworthiness signed off by a qualified mechanic before any item is released or sold. Winter pavement clearance and salt bin restocking Manual handling; winter service briefing (half day) Seasonal and call-out driven; incompatible with fixed mandated attendance — operate as an opt-in supplement. Flood response: sandbagging and clearance Manual handling; flood-water hygiene briefing (half day) Excluded from water-edge working. Leptospirosis (Weil’s disease) card issued. Beach, riverbank and towpath cleaning Manual handling; water safety awareness (half day) Throw-line trained supervisor present; no lone working; tidal and flow briefing before each shift. Invasive/hazardous plant species control (balsam, ragwort) Species identification (half day); COSHH where herbicide-adjacent Report Japanese knotweed; do not treat. Biosecurity wash-down between sites.

A.2 Parks, open spaces and burial grounds Role Training required Vetting and other requirements Mowing and strimming City & Guilds / NPTC pedestrian mower and brushcutter units (1–2 days each) Face guard and hearing protection; exclusion-zone marshalling by a second person. Ride-on machinery excluded. Hedge, shrub bed and mulching work Manual handling; hand tool safety (half day) Powered hedge trimmer requires a separate NPTC unit. Nesting-season restrictions under s.1 Wildlife and Countryside Act 1981. Watering of newly planted trees Task briefing (half day) Bowser driven by council staff. Establishment survival rate recorded as the placement’s output measure. Tree and hedgerow planting; guard fitting and removal Planting technique briefing (half day) Seasonal (November–March). Staff will complete an underground services check before any auger or post- hole work. Rights of way clearance and furniture repair Brushcutter unit; basic joinery briefing (half day) No lone working. Landowner notification handled by the rights-of-way officer Playground visual inspection and cleaning RPII Routine Inspection course (1 day) Records countersigned by an Operational-level inspector. Defects reported, never repaired, by the placement. Sports pitch preparation and line marking Line-marking equipment briefing; COSHH for paints (half day) Goalpost erection requires a minimum of two persons and a documented lifting plan. Allotment site maintenance Manual handling; site induction Boundaries, haulage roads and common areas only. No entry to individual let plots. Cemetery grounds maintenance and memorial cleaning Brushcutter unit; bereavement sensitivity briefing (half day) Excluded from grave excavation, interment, exhumation and memorial topple-testing. Work suspended during services. War memorial and war grave upkeep Conservation cleaning briefing (half day) No abrasive or chemical treatment of stone without conservator sign-off. Meadow and scrub management; fencing assistance Hand tool safety; brushcutter unit Post-driver operation excluded. Protected species and badger sett briefing before site entry.

A.3 Highways and street furniture Role Training required Vetting and other requirements Sign cleaning, straightening and sight- line clearance Chapter 8 signing, lighting and guarding awareness (1 day); manual handling Works only within a traffic management layout designed and set by qualified staff. No live carriageway working. Street furniture washing and repainting COSHH; Chapter 8 awareness Ground level only; ladders prohibited. Lead paint awareness on pre-1970s ironwork. Footway sweeping ahead of surface treatment Chapter 8 awareness; manual handling Hi-vis to specification; staff to set footway closure signage. Gully and drainage channel clearing Chapter 8 awareness; manual handling Surface grate work only. No confined space entry or chamber access under any circumstances. Pothole and patching gang labouring CITB Health, Safety and Environment test plus CSCS Labourer card (1 day); hot materials briefing Supervising operative must hold NRSWA Unit 002— no plant, roller or hot box operation by the placement. Manual traffic, parking and footfall counts Survey protocol briefing (half day) Suitable for claimants with manual work limitations; lone-working assessment; body-worn alarm when isolated. Asset data collection and geolocation Mobile data capture application training (1 day); asset classification briefing. UK GDPR awareness; device issue and return protocol; no photography of identifiable individuals or interiors of private property. Bus shelter cleaning and timetable case updating Chapter 8 awareness; broken glazing handling briefing Step ladders excluded. Damage reported for specialist glazing repair. Footway and carriageway condition walking surveys Defect classification briefing (1 day) Photograph-and-log only. Placements must not conduct the statutory safety inspections relied on for the s.58 Highways Act 1980 special defence.

A.4 Council housing and estate services Role Training required Vetting and other requirements Void property clearance and deep clean Manual handling; COSHH; sharps and drug paraphernalia awareness (half day) Staff check the asbestos register and clear the property before entry. Hoarded and severely soiled properties excluded. Void garden clearance Manual handling; brushcutter unit Staff clear the property before entry; sharps kit on site. Void painting and decorating Internal decorating skills course (3–5 days) Ground floor, ladder-free work only. Lead paint awareness in pre-1970s stock. Communal area cleaning COSHH; infection control (half day) Fob and key control with daily sign-out. Escorted access to bin stores and plant rooms. Estate grounds maintenance Pedestrian mower and brushcutter units As A.2. Estate-based supervision; residents notified in advance. Removal of abandoned items from communal areas Manual handling; notice procedure briefing (half day) Placement removes only items that have been served with notice by staff under the Torts (Interference with Goods) Act 1977. Communal fire safety sweeps and logging Fire safety awareness (half day); briefing on the Fire Safety (England) Regulations Clearance and logging only — no risk assessment. Defined escalation route for obstructed or locked means of escape. Handyperson service for elderly and disabled tenants Basic tool and DIY course (2–3 days); safeguarding Level 1 Enhanced DBS. Paired working for first three months. No gas, electrical or work at height. No key holding. Tenant garden maintenance scheme Pedestrian mower and brushcutter units; safeguarding Level 1 Enhanced DBS. No entry to the dwelling under any circumstances, including for use of facilities. Estate walkabout and defect logging Mobile application training; defect classification briefing Accompanied at all times by a housing officer. No tenancy or arrears discussion.

A.5 Corporate property and facilities Role Training required Vetting and other requirements Cleaning of council premises COSHH; infection control (half day) Basic DBS; enhanced where the premises are a school, children’s centre or care setting. Lone working assessment for out-of-hours shifts. Internal painting and decorating Decorating skills course (3–5 days) Ladder-free work only; occupied buildings out of hours. Portering and office moves Manual handling; team lifting (half day) Powered pallet truck operation excluded. Sealed archive boxes not to be opened. Depot and car park maintenance; ticket machine cleaning Manual handling; site induction Machines opened only by staff. No access to cash vaults or coin mechanisms. Stock, inventory and asset tagging Inventory system training (half day) Supervised access to secure stores; two-person rule for controlled stock.

A.6 Libraries, culture and leisure Role Training required Vetting and other requirements Library shelving, returns and stock tidying Library management system training (1 day) Basic DBS; enhanced where regularly working in children’s library areas. No access to borrower records. Home library service delivery Safeguarding Level 1; welfare concern reporting protocol (half day) Enhanced DBS. Paired working; no driving of the service vehicle; no entry beyond the doorstep unless invited and paired. Digital inclusion support Digital champion training (2 days); safeguarding Level 1; UK GDPR awareness Enhanced DBS. Absolute prohibition on handling residents’ passwords, banking details, or completing benefit claims on their behalf. Reading group and homework club support Safeguarding Level 1; literacy support briefing (half day) Enhanced DBS with children’s barred list check. Never unsupervised with a child; supernumerary to any staffing ratio. Museum and heritage site invigilation Collections awareness (half day); emergency evacuation training Basic DBS. No handling of objects. Lone gallery cover excluded. Country park visitor centre and car park assistance Customer service briefing; emergency first aid appointed person (1 day) No cash or card handling. Two-way radio issued; check- in protocol for remote sites. Leisure centre support (reception, cleaning, kit setup) COSHH; pool plant awareness (non-operative); safeguarding Level 1 Enhanced DBS where children’s programmes run. No lifeguarding, no water supervision, no access to the plant room. Market operations and hall cleaning Manual handling; stall erection training (half day) No collection of pitch fees. Agree and fix the early-start rota in advance.

A.7 Adult social care adjacent (no regulated activity; no personal care) Role Training required Vetting and other requirements Meals-at-home delivery and welfare check Food hygiene Level 2 (1 day); safeguarding Level 1; welfare check protocol (half day) Enhanced DBS. No personal care, no key holding, no medication. Scripted escalation for no-answer calls. Day centre assistance Food hygiene Level 2; safeguarding Level 1; dementia awareness (half day) Enhanced DBS. No personal care, no medication administration, no moving and handling of people. Community transport passenger assistant MiDAS Passenger Assistant (1 day); emergency first aid at work (1 day); safeguarding Level 1 Enhanced DBS. No driving. Wheelchair-securing training and separate sign-off before the first solo shift. Telephone befriending and welfare calling Befriending and boundaries training (2 days); safeguarding Level 1; suicide awareness briefing Enhanced DBS. Calls made only from council premises on council lines; no exchange of personal contact details; supervised debrief after each session. Shopping and prescription collection Safeguarding Level 1; financial abuse awareness (half day) Enhanced DBS. No handling of resident cash or bank cards — council prepaid card or direct-billing account only. Pharmacy collection requires written resident authority. Activity and companionship support in sheltered and extra-care schemes Safeguarding Level 1; dementia awareness (half day) Enhanced DBS. Scheme manager on site throughout. No personal care. Winter warmth checks and warm hub staffing Energy advice briefing (1 day); safeguarding Level 1 Enhanced DBS. Paired visits only. Referral to a qualified adviser — no tariff, grant or eligibility advice given directly.

A.8 Children, families and schools Role Training required Vetting and other requirements School crossing patrol School crossing patrol training (2 days); emergency first aid (1 day) Enhanced DBS with children's barred list check. Warranted under s.28 Road Traffic Regulation Act 1984. Very high reliability requirement — unsuitable where attendance is volatile. Breakfast and after-school club assistance Food hygiene Level 2; safeguarding Level 1; paediatric first aid awareness Enhanced DBS with barred list check. Placements are supernumerary and must not be counted towards statutory staffing ratios. Holiday activity and food (HAF) programme assistance Food hygiene Level 2; safeguarding Level 1 Enhanced DBS with barred list check. Seasonal; supernumerary to ratios. Play scheme and youth club support Safeguarding Level 1; behaviour management briefing (half day) Enhanced DBS with barred list check. Never alone with a child; supernumerary to ratios. Reading and numeracy volunteer support Safeguarding Level 1; school induction Enhanced DBS with barred list check. Within sight of a teacher at all times. Family hub reception and room setup Safeguarding Level 1 Enhanced DBS with barred list check. No access to case records or referral systems. School grounds maintenance Pedestrian mower and brushcutter units Schedule to school holidays wherever possible; enhanced DBS required if any term- time presence.

A.9 Regulatory and community safety support Role Training required Vetting and other requirements Empty homes and derelict site surveys Survey protocol (half day); conflict avoidance (half day) External observation from the public highway only; no entry onto private land. Lone working assessment and body-worn alarm. Selective licensing and HMO street surveys Survey protocol; briefing on Parts 2 and 3 Housing Act Data capture only. No enforcement contact; no discussion with landlords or occupiers. Abandoned vehicle logging Procedure briefing under the Refuse Disposal (Amenity) Act 1978 (half day) Logging and photography only. No contact with the vehicle; authorised officers serve notices. Town centre vacancy and shopfront surveys Survey protocol; data capture application (half day) Public highway only; no entry to premises. Community warden patrol support Conflict avoidance and de-escalation (1 day); safeguarding Level 1; radio protocol Basic DBS. Paired patrol; observe and report only; mandatory withdrawal from confrontation — no SIA- licensable activity Alley-gate and Secured by Design walkabout support Survey protocol (half day) Accompanied by a community safety officer throughout. Home and fire safety leafleting Task briefing (half day) Letterbox drop only. No doorstep engagement, no entry to premises, no safety advice given. Environmental health administrative support Records system training (1 day); food hygiene Level 2 awareness Basic DBS. No attendance at inspections, no sampling, no contact with food business operators. Pest control group assistance COSHH; manual handling Carrying, access and record-keeping only. No handling, placing or recovery of rodenticide bait, which requires separate certification.

A.9 Administrative and back office Role Training required Vetting and other requirements Records digitisation and indexing Scanning and metadata training (1–2 days); UK GDPR awareness (half day) Basic DBS and signed confidentiality undertaking. Supervised room; no personal devices; special category material segregated and excluded. Electoral registration canvass Canvasser training to Electoral Commission guidance (1 day); conflict avoidance Basic DBS; identity badge. Signed impartiality undertaking; exclude any claimant standing for, employed by, or campaigning for a party or candidate at the relevant poll. Contact centre first-line support Telephony and CRM training (3–5 days); customer service; UK GDPR awareness Basic DBS. Scripted signposting only; no payment card handling (PCI DSS); all calls recorded and quality- assured. Post room, scanning, indexing and franking Mail handling and UK GDPR awareness (half day); suspicious package awareness Basic DBS. No opening of items marked personal, legal or privileged, or addressed to elected members. Blue badge and concessionary travel processing support Case management system training (1–2 days); UK GDPR awareness Basic DBS, enhanced where medical evidence is handled. No eligibility determinations and no assessment of medical evidence. Resident satisfaction and consultation fieldwork Survey protocol; Market Research Society Code of Conduct briefing (1 day) Basic DBS. Signed impartiality undertaking; verbatim script; no leading or supplementary questions. Freedom of information logging and triage FOIA 2000 and EIR 2004 awareness (1 day); UK GDPR awareness Basic DBS. Logging and acknowledgement only. No disclosure decisions and no application of exemptions. Data cleansing (single person discount, business rates, tenancy) System training (1–2 days); UK GDPR awareness; counter- fraud awareness Basic DBS. Read-and- flag only; no account amendments; no contact with residents; full audit trail retained.

Endnotes

1. DWP, Personal Independence Payment: Official Statistics to January 2026; ONS, Estimates of the population for England and Wales. 2. DWP, Universal Credit statistics, 29 April 2013 to 12 June 2025. 3. Resolution Foundation, Growing pressures: Exploring trends in children’s disability benefits. 4. OBR, March 2026 Economic and Fiscal Outlook, p. 70. Household count is 29mn in 2025 (ONS, Families and households in the UK: 2025) and 30.2mn in 2031, adjusted for OBR forecast population growth. 5. DWP, Benefit expenditure and caseload tables 2026; ONS, Unemployment. 6. DWP, Benefit Combinations: Official Statistics to August 2025. 7. DWP, Benefit Combinations: Official Statistics to August 2025. 8. DWP, Long-term out of work and “Searching for Work” claimants on Universal Credit, January 2023 to January 2024. 9. DWP, Universal Credit quarterly statistics, 29 April 2013 to 12 February 2026. 10. DWP, UC Immigration Status/Nationality Statistics to February 2026; DWP, Benefit expenditure and caseload tables 2026. We adjusted for very slightly lower average spending (2.1%) for non-British claimants, based on a March 2025 FOI. 11. DWP, Benefit expenditure and caseload tables 2026. 12. OBR, Welfare Trends Report – June 2026, p. 3. 13. DWP, Long-term out of work and “Searching for Work” claimants on Universal Credit, January 2023 to January 2024. 14. Citizens Information Board of Ireland, Employment support schemes. 15. OECD, Irish Department of Social Protection, & European Commission, Impact Evaluation of Ireland’s Active Labour Market Policies, pp. 171-216. 16. DWP, Universal Credit statistics, 29 April 2013 to 8 January 2026 17. MHCLG, Social housing lettings in England, tenants: April 2024 to March 2025 18. CSJ, 27 young non-EU migrants hired for every young Brit since 2020, analysis reveals. 19. DWP, Universal Credit immigration status and nationality statistics: 1.22 million non-citizen adults on Universal Credit at February 2026, 14.8% of the caseload, projected forward on the Home Office settlement forecast of 1.6 million grants between 2026 and 2030, with naturalisation restricted under this paper’s proposals. DWP publishes no nationality breakdown for PIP; our estimate applies Labour Force Survey and Family Resources Survey shares, calibrated to DWP administrative data, to the working-age caseload forecast in the Benefit expenditure and caseload tables 2026. 20. Reform’s policy to abolish new grants of ILR has already been announced. This policy extends ineligibility for benefits to existing ILR holders and migrants who have EU settled status, as well as a number of other immigration statuses. 21. House of Commons, Migration statistics. 22. House of Commons, UK disability statistics: Prevalence and life experiences, p. 30. 23. IFS, New disability benefit awards continue to fall but remain well above pre-pandemic levels. 24. OBR, March 2026 Economic and Fiscal Outlook, p. 70. Household count as above. 25. IFS, The role of changing health in rising health-related benefit claims. DWP, StatXplore. 26. CSJ, Disability benefits for anxiety hit new record high, analysis reveals. 27. Nynke de Groot and Pierre Koning, Assessing the effects of disability insurance experience rating. The case of The Netherlands, Labour Economics (2016); Pierre Koning, Experience Rating and the Inflow into Disability Insurance, De Economist (2009); Tomi Kyyrä and Juha Tuomala, The effects of employers’ disability and unemployment insurance costs on benefit inflows, Labour Economics (2023). 28. Resolution Foundation, Growing pressures: Exploring trends in children’s disability benefits. 29. DWP, Annex B to the Timms Review. 30. Tony Blair Institute, An Emergency Handbrake for UK Welfare. 31. DWP, The Timms Review of Personal Independence Payment: Interim Report. 32. Adam Smith Institute, Reverse Gear: Why Motability is Broken and How to Fix It. 33. DWP, The Timms Review of Personal Independence Payment: Interim Report. 34. Resolution Foundation, Government must focus on entrances, entitlements and exits if benefit reforms are to stem spending and boost living standards. 35. DWP, International Comparisons of Disability Benefits and Disability Employment. 36. OBR, Welfare Trends Report – June 2026, p. 3. 37. OBR, March 2026 Economic and Fiscal Outlook, p. 70. 38. DWP, Fraudsters behind £53.9 million benefits scam brought to justice in country’s largest benefit fraud case 39. OBR, Welfare Trends Report – June 2026, p. 25. 40. NAO, Tackling benefit overpayments due to fraud and error, p. 23–24. 41. Paul Johnson, UK Consumer Price Statistics: A Review; Robert J. Gordon, The Boskin Commission Report: A Retrospective One Decade Later. 42. Our estimate corrects for Törnqvist upper-level substitution bias, Feenstra within-category bias, and market-to-home substitution. 43. House of Commons Library, The Barnett formula and fiscal devolution, p. 31. 44. These estimates are based on standard elasticities from the literature, which show how people have empirically reacted to changes in benefit rates, as well as the empirical evidence of specific programmes, such as Ireland’s Tus scheme and the Netherlands’ disability benefit reforms. 45. Conservative Party, Alternative King’s Speech

MAKING WELFARE

WORK