Foreword

The Office for Budget Responsibility (OBR) was created in 2010 to provide independent and authoritative analysis of the UK’s public finances. In December 2013, the Government asked the OBR to take on additional responsibilities in relation to its newly announced cap on a subset of welfare spending. This request was in two parts: to assess the Government’s performance against the welfare cap and to “prepare and publish information on the trends in and drivers of welfare spending within the cap”, so as to facilitate open and constructive debate. Parliament formally included these requirements in the October 2015 edition of the Charter for Budget Responsibility. The January 2022 update to the Charter reduced the frequency with which our Welfare trends report (WTR) must be published from once a year to once every two years.

We have explored several issues in our successive WTRs, ranging from a broad historical sweep of trends in UK welfare spending and international comparisons of welfare spending in our first two reports; to topics including universal credit, disability benefits, the Summer Budget 2015 welfare spending cuts, and changes in welfare spending during and after recessions, across our subsequent six reports. With health-related economic inactivity and health-related benefit caseloads having grown substantially in recent years and the latter projected to rise further over our latest medium-term forecast, this ninth WTR focuses on incapacity benefits spending over the past half-century, and the drivers of changing incapacity benefits caseloads over the past 15 years in particular.

This WTR was initially scheduled for publication on 2 July. Due to the general election that was announced on 22 May and took place on 4 July, we postponed its release to 10 October. This delay did not alter the content of our report.

The analysis in this report represents the collective view of the OBR’s Budget Responsibility Committee. We take full responsibility for the judgements that underpin it and for the conclusions we have reached. We have, of course, been supported in this by the full-time staff of the OBR, to whom we are enormously grateful, as we are to officials in the Department for Work and Pensions that have provided their help and expertise. We are also grateful to external stakeholders who gave their time and shared their expertise. In particular, we would like to thank Ben Baumberg-Geiger at King’s College London, Tom Pollard and Sam Tims at the New Economics Foundation, Kayley Hignell at Citizens Advice, Tony Wilson at the Institute for Employment Studies, Deven Ghelani at Policy in Practice, Christina Beatty at Sheffield Hallam University, Iain Porter at the Joseph Rowntree Foundation, and Declan Gaffney.

As with all our reports, the WTR remains a work-in-progress. We would be pleased to receive feedback on any aspect of the content or presentation of our analysis. This can be sent to [email protected].

In line with our Memorandum of Understanding, we provided a final copy of this report to HM Treasury two working days in advance of publication.

The Budget Responsibility Committee

Richard Hughes, Professor David Miles CBE and Tom Josephs

Executive summary

1 Incapacity benefits – income-replacement benefits for people of working age who have difficulty performing day-to-day activities due to ill health – have moved into sharper focus in recent years.^[1] Significant increases in both caseloads and spending on these benefits have prompted a series of new policies partly designed to arrest and reverse this. The proportion of the working-age population in receipt of an incapacity benefit reached a post-financial crisis high of 7.0 per cent in 2023-24 and is forecast to reach an all-time high of 7.9 per cent in 2028-29. This reverses the steady decline in caseload prevalence from the early 2000s to mid-2010s. There have also been two major reforms to the structure of incapacity benefits over this period, with the introduction of employment and support allowance (ESA) in 2008 and the gradual rollout of universal credit (UC) replacing income-related ESA (and five other working-age, means-tested benefits) from 2016. In light of these trends and developments this year’s Welfare trends report:

• places these recent trends in the wider history of the four different incapacity benefits regimes of the past half century;
• zooms in on the role of the operation of the incapacity benefits system since 2010, looking at claimants’ journeys onto and off the different benefits over this period; and
• draws a set of conclusions for how we construct our incapacity benefits forecast in future.

A history of incapacity benefits

2 Incapacity benefits caseloads (Chart 1) have cycled through periods of growth and decline since the late 1970s. Over this period, there have been four distinct policy regimes:

• Invalidity benefit (1980s to mid-1990s): invalidity benefit (IVB) provided support to those with a longer-term inability to work and was complemented by shorter-term sickness benefits. After a period of relative stability in the early 1980s, the incapacity benefits caseload rose rapidly from 3.6 per cent of the working-age population in 1984-85 to 6.9 per cent in 1994-95. This primarily reflected two factors. First, labour market dislocation in former industrial areas saw many of those exiting jobs move into economic inactivity and onto incapacity benefits, with very little change in unemployment. Second, and relatedly, falling relative generosity and rising work-search requirements within unemployment benefits over this period appear to have channelled more claimants towards incapacity benefits, which had no work-search requirements and were relatively more generous.
• Incapacity benefit (mid-1990s to late 2000s): incapacity benefit (IB) was introduced in 1995 to replace IVB and sickness benefit. The main changes were reduced generosity relative to IVB, and tightened eligibility criteria via the introduction of a new medical test. Alongside a relatively strong economy through this period, these changes contributed to reductions in onflows relative to IVB, particularly among men. The result was the incapacity benefits caseload as a share of the working-age population largely flattening out by the late 1990s, reaching a peak of 7.6 per cent in 2002-03, and then falling modestly to 7.0 per cent by 2007-08.
• Employment and support allowance (late 2000s to late 2010s): ESA replaced IB for new claims from 2008, with existing IB claimants gradually reassessed for ESA through the early 2010s. The principal differences relative to IB were the introduction of a new, more work-focused medical assessment – the work capability assessment (WCA) – and the fact that this assessment could place claimants deemed not fit for work into two groups (with the less severe incapacity group receiving lower awards and facing some conditionality). The incapacity benefits caseload fell and then stabilised over this period: it stood at 6.7 per cent of the working-age population in 2008-09, and 4.9 per cent in 2017-18. This was driven in large part by the reassessment of the IB caseload increasing off-flows, somewhat offset by the rising female state pension age expanding eligibility at ages where people are more likely to claim.
• Universal credit (late 2010s to late 2020s): UC began replacing income-related ESA from 2016 and became the route for all new income-related incapacity benefits claims from late 2018.^[2] While there were several operational changes to the process for initiating claims and progressing to assessments, the switch from ESA to UC did not entail the same kind of shift in the assessment regime, generosity or relative conditionality of incapacity benefits that previous regime changes had (though awards were cut for the less severe incapacity group in both ESA and UC in 2017). The incapacity benefits caseload as a share of the working-age population increased by 2.1 percentage points to 7.0 per cent between 2017-18 and 2023-24 and is forecast to continue rising. As discussed in more detail below, this rise is likely to reflect a range of factors including how UC operates relative to ESA, changes to operational processes, and wider health and labour market trends.

Chart 1: Caseload prevalence of incapacity benefits

3 The history of incapacity benefits over the past half century shows that a range of factors combine to drive changes in caseloads and spending, falling into three broad categories: changes in population health; changes in the labour market and wider economy; and changes in the benefits system. The third of these relates in part to changes to the structure of incapacity benefits themselves, such as the assessment regime and the degree of conditionality. But it also relates to these benefits’ interactions with wider welfare policy, including the relative generosity and work-search requirements associated with unemployment benefits, and the overall level of spending on other parts of the working-age welfare system (which has sometimes moved inversely with incapacity spending). The rest of the report looks in greater depth at the role that changes in the benefits system have played in driving incapacity benefits caseloads and spending over the past 15 years, focusing on claimants’ journeys onto and off these benefits.

Incapacity benefits flows since 2010

4 Between 2008-09 and 2023-24 the incapacity benefits caseload (combining IB, ESA and UC) increased by 340,000 (13.2 per cent), the net result of it falling, levelling off, and then rising in roughly five-year intervals over this period. A large part of the 330,000 fall between 2008-09 and 2019-20 reflects falling prevalence among the under-60s. This was partly offset by increases in the state pension age in this period, particularly for women, which extended eligibility into ages at which people are more likely to claim incapacity benefits. Since 2019-20, incapacity benefits receipt rates have risen near-uniformly by around 2 percentage points across all age groups, pushing up the caseload by 670,000.

5 To understand the factors underlying incapacity benefits caseload trends since 2010, Chart 2 decomposes annual changes in the caseload into the contributions of annual onflows (purple bars) and off-flows
(yellow bars, shown as negatives), with the annual caseload change (orange diamonds) being the sum of the two. It shows, for example, that relatively high volumes of annual off-flows drove the falling caseload in the early 2010s, while caseload increases since 2018-19 are explained by the growth in onflows outstripping the growth in off-flows.

Chart 2: Annual changes in incapacity benefits caseload, onflows, and off-flows

6 Changes in onflows can be further decomposed into three stages of the claims process: the volume of initial claims submitted; the dropout rate (the share of claims that are withdrawn prior to reaching a WCA decision); and the approval rate (the share of WCA decisions that deem a claimant to have more or less severe incapacity). Chart 3 decomposes changes in onflows between three different points in time into these factors. It shows that, of the overall rise in annual onflows between 2010-11 and 2022-23, only a minority (20 per cent) is due to a higher number of initial claims, with 30 per cent of the rise due to lower dropout rates and just under 50 per cent due to the increase in approval rates.

Chart 3: Changes in onflows due to initial claims, dropout and approval rates

7 Overall caseloads depend not only on onflows but also on off-flows. To understand how initial claims, dropouts, approval rates and off-flows have contributed to caseload changes, we focus on three distinct periods: the early 2010s period of declining caseloads, the later 2010s period of stable caseloads, and the post-2018 period of increasing caseloads.

8 The falling caseload (by an average of 145,000 a year) in the early 2010s was largely the result of comparatively high off-flow rates – which spiked during 2012-13 – alongside lower but gradually rising onflows:

• Onflows rose by around 110,000 a year over this period.^[3] This increase was nearly all driven by a rising approval rate, which contributed 109,000 more onflows a year. This likely reflected responses to a series of Department for Work and Pensions (DWP) independent reviews of the relatively new WCA that led to some easements of assessment criteria and to extra support being provided to claimants.
• Annual off-flows averaged 300,000 a year over this period, nearly twice the annual level of onflows, and peaked at 390,000 a year in 2012-13. This was in large part driven by the policy to reassess the stock of IB claimants under the new WCA. 1.3 million IB reassessments took place between 2011-12 and 2014-15 and resulted in 250,000 claimants being deemed fit for work. Adding in those for ESA, reassessments explain just under half of total annual off-flows from 2011-12 to 2014-15.

9 The more stable caseload (which rose by an average of around 50,000 a year) in the late 2010s was the result of relatively consistent onflows and off-flows, with the former at a slightly higher level on average. This represents a relatively rare period of stability in the incapacity benefits regime over the past half century. Almost all of the caseload was on the same benefit (ESA), assessments and reassessments did not experience any significant operational changes, and there were no major external shocks.

10 The rising caseload (by an average of 160,000 a year) since 2018-19 reflected the growth in onflows outstripping the growth in off-flows amid the move to UC for all new claims, and the disruption of the pandemic and its aftermath:

• On the broader measure shown in Chart 3 onflows rose by around 270,000 a year from 2018-19 to 2022-23. This was driven by increases at all three stages of the process:
• Initial claims rose by 72,000 a year, which explains a quarter of the overall rise in onflows. This is likely to reflect the increase in the state pension age from 65 to 66 widening eligibility, alongside wider heath trends, labour market disruption due to Covid, and cost-of-living pressures together increasing eligibility and take-up. The greater generosity of incapacity benefits relative to unemployment benefits has not changed materially in this period, but the large gap may be a more important consideration for potential claimants when cost-of-living pressures are more acute.
• A falling dropout rate resulted in 129,000 more onflows a year. This could reflect worsening health conditions of applicants, changes to the length of time an assessment takes, rules changes under UC relative to ESA that allow claimants to do some work during the assessment phase without having to terminate claims, and changes to the processes for verifying and handling fit notes following Covid.
• A rising approval rate resulted in 70,000 more onflows a year. The increase in the approval rate was nearly entirely in the more severe incapacity group, with most onflows now receiving higher awards and having no conditionality. This could reflect claimants having more serious health conditions than in the past, along with the 2017 policy that reduced generosity for the less severe incapacity group creating higher incentives to be placed in the more severe group.
• Annual off-flows averaged 215,000 a year over this period, and rose by around 125,000. But as a share of the stock of claimants each year the rise was much less stark, and will partly reflect some renormalisation after the pandemic. There were very few reassessments in this period.

Conclusions

11 The analysis in this report, and particularly the detailed examination of flows onto and off incapacity benefits in recent years, provides a range of insights for our future forecasts. Recent trends reinforce the importance for our forecast of assessing all three of the key historical drivers of incapacity benefits caseloads and spending – working-age population health, the state of the labour market, and the structure of the benefits system. But several of the findings in the analysis particularly highlight the importance of the policy and operational features of the benefits system.

12 One of the most significant is that only a minority of the recent rise in incapacity benefits onflows reflects a higher number of people initiating claims. Changes in what happens after that as claimants move through the system – the number who drop out of the process and the final approval rate – have had larger effects. We cannot be certain about the drivers of these effects, but it is likely that they are at least in part due to the different rules under UC relative to ESA that allow claimants to do some work during the assessment phase without having to terminate claims, and the changes to the processes for verifying and handling fit notes during Covid that were subsequently adopted as the business-as-usual approach. They may also reflect changes in the characteristics of claimants such as the type and severity of their health conditions. In terms of off-flows, it is only reassessments that appear to have substantially changed caseload trends, in particular in the early 2010s. There is inconclusive evidence in relation to the effects on off-flows of employment programmes for incapacity benefits claimants.

13 These features of the incapacity benefits system have interacted with changes in the non-incapacity parts of the system, and wider health and economic trends, to drive recent incapacity caseloads and spending. And our broader historical sweep shows that, over the past half century, the incapacity benefits system has rarely reached any kind of steady state, making the basis from which to forecast future trends highly contextual and uncertain.

14 These and other insights provide areas for further research or information that will help inform future forecasts. The report identifies areas for improved data, in areas such as UC incapacity claimants’ primary health condition; claimants’ employment and benefit circumstances at the point of onflow; claimants’ off-flow journeys into work or elsewhere; and claimants’ interactions with health services.

15 The report also identifies risks to our forecasts that we will scrutinise in more depth in future. We plan to conduct greater scrutiny of our incapacity onflow and off-flow assumptions, working with DWP analysts. We will also ask DWP to monitor population health data in addition to incapacity caseload data to judge if our onflow assumptions remain central. And we will more closely scrutinise DWP’s operational provider performance, particularly in relation to reassessments, and delivery of employment support programmes when making decisions about the path of incapacity benefits caseloads.

Chapter 1: Introduction and context

1.1 Our biennial Welfare trends reports (WTRs) examine trends in working-age and pensioner benefits and tax credits. This year’s report focuses on working-age ‘incapacity benefits’ – defined as the various income-replacement benefits that have been available over the past half century for people who find it difficult to work due to ill health.^[4]

1.2 This part of the welfare system has undergone major reform twice within the past 20 years and recently has experienced a significant increase in claimants. Employment and support allowance (ESA) replaced incapacity benefit (IB) in 2008. From 2016, universal credit (UC) began to replace new claims for income-related ESA (and six other working-age, means-tested benefits and tax credits).^[5] The post-pandemic period has seen a rapid increase in claims for the health-related elements of UC alongside rising economic inactivity due to sickness. As a result, we have repeatedly made large upward revisions to our forecasts for incapacity benefits since March 2022.^[6] Our most recent forecast from March 2024 projects the proportion of the working-age population in receipt of an incapacity benefit reaching an all-time high of 7.9 per cent by 2028-29.

1.3 This report explores the potential drivers of this increase, with a focus on how changes in the operation of the incapacity benefits system may have contributed to trends in incapacity benefits claims, particularly in the period since 2010. In this chapter we introduce the metrics and methodological approach that we use to analyse the evolution of incapacity benefits over time. To provide context for the subsequent analysis we then briefly introduce trends in incapacity benefits caseloads and spending over the past 45 years.

How we measure and analyse incapacity benefits

1.4 Our WTRs focus on those elements of benefit and tax credit spending that are financed by central government as part of what the Treasury calls ‘annually managed expenditure’ (AME). Most are administered by three central government bodies:

• the Department for Work and Pensions (DWP) for most benefits in Great Britain;
• HM Revenue and Customs (HMRC) for the personal tax credits, child benefit and tax-free childcare systems across the UK; and
• the Department for Communities for most benefits in Northern Ireland.^[7]

1.5 The welfare system contains several benefits that aim to support people affected by sickness or disability. The main distinction we draw within these is between two categories of benefit, which can be claimed at the same time.^[8]

• ‘Incapacity benefits’ – the focus of this report – which provide income replacement to those who find it difficult to work due to ill health. In 2023-24 the government spent £24.9 billion on incapacity benefits, almost all of which went to ESA claimants and UC claimants either in the limited capability for work (LCW) or limited capability for work and work-related activity (LCWRA) categories.^[9] We also briefly assess those incapacity benefits previously available over the past half century: income support for sick and disabled claimants; incapacity benefit; invalidity benefit (IVB); sickness benefit; and severe disablement allowance (SDA).
• ‘Disability benefits’, which cover the extra costs of daily living and mobility associated with disability and long-term health conditions, regardless of whether the individual is working or not. In 2023-24 working-age disability benefits – personal independence payment (PIP) and disability living allowance (DLA) – cost £19.0 billion.^[10]

1.6 This report focuses on incapacity benefits because we covered disability benefits in our January 2019 WTR.^[11] That report showed that the prevalence of working-age disability benefits has grown steadily since the introduction of DLA in 1992, related to the growing prevalence of self-reported disability, broadening eligibility criteria, and higher take-up. By contrast, the incapacity benefits caseload has been somewhat cyclical. The prevalence of incapacity benefits grew rapidly in the late 1980s and early 1990s before falling in the mid-2000s, and starting to grow again after 2017-18, with the growth rate accelerating after the pandemic. These varying trends in incapacity benefits reflect changes in health, the labour market, and the structure of the welfare system – a more complex set of drivers than extra-cost disability benefits. We briefly summarise these trends and their drivers below.

Trends in incapacity benefits caseloads and spending

1.7 Chart 1.1 presents spending on incapacity benefits as share of GDP over the past half century (as we discuss below, this is clearly a partial picture of the working-age welfare system, with trends in non-incapacity spending being quite different, particularly in recent decades). It highlights a series of overlapping dynamics between economic, health, and benefit system changes over four distinct periods:

• 1980s to mid-1990s: after a period of relative stability in the late 1970s and early 1980s, spending rose rapidly from an average of 0.7 per cent of GDP in the 1980s to a peak of 1.3 per cent of GDP in 1995-96, shortly after the introduction of IB. The rapid increase reflects sharply rising IVB and income support caseloads, due to the rapid rise in male economic inactivity and the declining relative attractiveness of unemployment benefit in the late 1980s.
• Mid-1990s to late 2000s: There was a substantial fall in incapacity benefits spending as a share of GDP between the introduction of IB and the introduction of ESA. This reflects the stabilisation in caseload prevalence due to the tighter gateway onto incapacity benefits, and the increase in the employment rate. In contrast to the late 1980s and early 1990s, there was limited growth in claimants’ awards, due to reductions in the generosity of IB compared to IVB and low inflation resulting in low uprating.
• Late 2000s to late 2010s: Spending on incapacity benefits remained steady at 0.7 per cent of GDP between 2009-10 and 2019-20. Increases in average awards from uprating and a shift in the composition of the caseload towards the more severe incapacity group offset a small reduction in the caseload, which fell by an average of 0.4 per cent each year over this period.
• Late 2010s onwards: Spending has grown rapidly in recent years, increasing by 0.2 per cent of GDP between 2019-20 and 2023-24. This reflects rising caseloads and an ongoing rise in the share of incapacity benefits claimants joining the (more expensive) more severe incapacity group (discussed further in Chapter 3). The fall in spending as a share of GDP in 2022-23 and rapid rise in 2023-24 reflects lagged benefit uprating.^[12] Our forecast assumes that incapacity benefits spending will continue growing as a share of GDP until 2025-26, before stabilising.

Chart 1.1: Incapacity benefits spending as a share of GDP

1.8 Incapacity benefits caseloads have been a key driver of the trends in incapacity benefits spending as a share of GDP discussed above. Chart 1.2 shows that the proportion of the working-age population in receipt of an incapacity benefit increased sharply from the mid-1980s through to the mid-1990s, fell from the mid-2000s through to 2013-14, and has risen consistently since 2017-18. It reached a post-financial crisis high of 7.0 per cent of the working-age population in 2023-24. While instructive overall, these trends over long time periods have to be interpreted with some caution, given the different ways in which different benefits classify people over time. Most importantly, as a household benefit UC includes some partners of claimants in the incapacity benefits caseload, thereby increasing the number of individuals ‘counted’ as on the benefit relative to the predecessor ESA system.^[13]

1.9 Our March 2024 forecast assumes that incapacity benefits prevalence will continue to increase over the next five years, albeit at a slower rate from 2026-27, and reach an all-time high of 7.9 per cent in 2028-29. This is based on a judgement that part of the recent strong growth in incapacity benefits caseloads has been driven by cost-of-living pressures driving up applications and take-up, meaning growth rates will slow as real household disposable incomes recover.^[14] Despite this forecast rise in caseload prevalence, our forecast for incapacity benefits spending as a share of GDP stabilises after 2024-25 (Chart 1.1). This is because nominal GDP is forecast to rise faster than CPI inflation (the policy assumption for benefit uprating).^[15]

Chart 1.2: Caseload prevalence of incapacity benefits

1.10 Trends in incapacity benefits spending and caseloads have been driven by three broad factors: changes in the health of the working-age population; changes in the labour market and wider economy; and changes in the benefit system. It is hard to isolate the exact contributions of these factors to overall trends because changes in each have often happened simultaneously, as previous OBR work (and a wealth of external analysis) has explored. Taking each of the factors in turn:

• Changes in population health can change the proportion of the working-age population eligible to claim incapacity benefits. For example, our 2016 WTR found that part of the decrease in incapacity benefits caseloads in the 2000s reflected a decrease in the amount of time people spent in ill health. More recently, the rising share of incapacity benefits claims awarded for mental health in the 2010s mirrors the rise in the share of the population reporting mental health problems in survey data. And our 2023 Fiscal risks and sustainability report (FRS) suggested a link between the post-pandemic rise in incapacity benefits caseloads and the direct and (particularly) indirect impact of the pandemic on health outcomes.
• Changes in the labour market and wider economy can change the number of people eligible to claim incapacity benefits, given that these benefits are very largely awarded to those out of work. Our 2014 WTR showed how the rapid rise in the male IVB caseload in the 1980s was partly caused by industrial restructuring, particularly rising male unemployment in formerly industrial areas (discussed in more detail in Chapter 2). That report also noted that falling incapacity benefits caseloads in the 2000s reflected the strength of the wider economy at that time. Our 2023 FRS showed that rising health-related inactivity (strongly associated with rising incapacity benefits caseloads) in recent years has been particularly concentrated among those previously working in high-contact service sectors most affected during the pandemic. And as discussed above, our recent forecasts have been based on the judgement that cost-of-living pressures in recent years have also contributed to rising incapacity benefits caseloads, by widening means-tested eligibility and incentivising take-up. Finally, a key policy-driven labour market change – the rising state pension age during the 2010s – has contributed to rising incapacity benefits caseloads and caseload prevalence by shifting eligibility into ages at which people are more likely to claim (Box 3.1 in Chapter 3 explores this in detail).
• Changes in the benefit system change the incentives to take up, continue claiming, and move off different benefits, from the perspective of both claimants and advisors. Our 2014 WTR concluded that rising incapacity benefits spending in the 1990s reflected policies to tighten conditionality for unemployment benefits and to make them less generous relative to incapacity benefits, increasing the incentives to claim the latter. Similarly, our 2023 FRS concluded that rising incapacity benefits caseloads in recent years have been associated with the expansion of conditionality and sanctioning in the non-incapacity parts of the working-age welfare system. And rising spending since 2017 may be associated with a reduction in the generosity of payments to the less severe conditionality group, pushing claimants (or assessors) towards the more generous more severe incapacity group (discussed further in Chapter 3). Incapacity benefits spending also needs to be seen in the context of total non-incapacity spending. Chart 1.3 shows that the two have sometimes moved in different directions to one another over the course of the 21^st century. This could imply some substitution between different parts of the system in the eyes of both policymakers and claimants, as discussed in Chapter 2.

Chart 1.3: Working-age spending, split by incapacity benefits and other spending

1.11 We explored the health and economic drivers of incapacity benefits caseloads in detail our 2023 and 2024 Fiscal risks and sustainability reports. While this report discusses the role of all three of the above factors, it focuses in particular on aspects of the third – exploring in detail how changes in the operation of the incapacity benefits system have shaped recent trends in caseloads. Within this we focus on the period since 2010 following the introduction of ESA, which was then followed by the introduction of UC. This is a period for which we have more complete data and so can explore in detail the role of onflows, off-flows, and the journey of claimants through the benefit system, in driving caseloads.

Structure and approach of this report

1.12 The remaining chapters of this report look at the history of incapacity benefits in Great Britain since the 1970s and explore trends in claims, assessments, and outcomes since the introduction of ESA in 2008 in greater detail. The report is structured as follows:

• Chapter 2 sets out a brief history of incapacity benefits over the past half century;
• Chapter 3 discusses rising caseloads since 2010, decomposing changes into the role of onflows and off-flows, and providing a detailed exploration of claimants’ ‘journey’ through the incapacity benefits system; and
• Chapter 4 draws some conclusions and discusses implications for our forecasts.

Chapter 2: A history of incapacity benefits

2.1 This chapter describes how the incapacity benefits system has evolved in Great Britain over the past half century, focusing on caseloads. We look in detail at what has driven trends in invalidity benefit (IVB) and incapacity benefit (IB) caseloads between the late 1970s and 2008. We also discuss the main trends in employment and support allowance (ESA) and universal credit (UC) caseloads after 2008, but Chapter 3 provides a more detailed decomposition of these.

2.2 There have been four main incapacity benefits since the introduction of invalidity benefit in 1971: invalidity benefit, incapacity benefit, employment and support allowance and universal credit. Table 2.1 describes the basic features of each benefit, and the following sections describe each in greater detail. In summary, the key factors in the transition from each iteration of the system to the next have been to tighten the assessment gateway to receive the benefit, reduce generosity for certain groups, move away from awards based on age and duration towards relative fitness for work, and more recently to introduce elements of conditionality.

2.3 In addition to these four main benefits, in some periods claimants have been able to receive other benefits related to incapacity. In brief these operated as follows:

• Sickness benefit was a contributory, short-term benefit for those with an inability to work for fewer than 28 weeks. From 1983 onwards sickness benefit was replaced by the employer-provided statutory sick pay (SSP). When first introduced, employers were able to reclaim the cost of the payments from the Government. This reclaim provision was subsequently made less generous and eventually abolished in 2014, shifting the full liability of SSP to employers.
• Severe disablement allowance (SDA) was a non-contributory, non-means-tested benefit for those disabled young or deemed to be 80 per cent or more disabled who were unable to work – and could not be claimed at the same as sickness benefit. SDA was paid at a relatively low rate, meaning about 70 per cent of claimants relied on income support as a top-up. SDA was abolished from 2001 and new claimants who would previously have qualified for SDA became entitled to IB instead. Existing SDA claimants were unaffected.
• Income support was a non-contributory, means-tested benefit that included disability premiums. This meant that claimants who had not paid enough National Insurance Contributions to receive incapacity benefits could receive means-tested income support on the grounds of illness or disability. Disability premiums were paid in addition to the basic rate of income support at different rates determined by the claimant’s circumstances (disability premium and severe disability premium). Income support was often received in conjunction with SDA. Income support claimed on grounds of illness or disability was replaced by ESA in 2008.

Table 2.1: The main features of incapacity benefits over the past half century

Invalidity benefit (1971-1995)

2.4 Invalidity benefit, made up of invalidity pension and invalidity allowance, was introduced in 1971 alongside the existing short-term sickness benefit. It was designed to provide more generous support for those with an inability to work for longer than 28 weeks. It was contributory and so provided support to those with sufficient National Insurance credits, while those without sufficient contributions received SDA, often alongside income support to top up incomes. The medical assessment focused on the ability to conduct “suitable work” and was carried out by GPs.^[16]

2.5 Chart 1.2 in the previous chapter showed that overall incapacity benefits caseload prevalence among the working-age population rose sharply from 1978-79 to 1994-95, from 3.6 per cent to 6.9 per cent (detailed caseload data is not available before 1978-79). It also shows that prevalence for IVB alone more than tripled over this period, from 1.6 per cent to 5.3 per cent. Chart 2.1 shows IVB (and SDA) caseload prevalence by age for men and women over this same period. It shows:

• IVB caseload prevalence doubled for men between 1978-79 and 1994-95. It nearly quadrupled, from a much lower rate, for women (top two panels of Chart 2.1), partly due to increasing numbers of women becoming eligible for contributory benefits, given both the big shifts in female working patterns over this period and the removal of the option to pay reduced national insurance contributions for women married on or after 6 April 1977. The female proportion of the overall IVB caseload therefore increased slightly over this period: from 22 per cent of claimants in 1978-79 to 34 per cent in 1994-95.
• Overall trends in caseload prevalence for men were reflected in trends for different age groups over the age of 40 (bottom-left panel). This meant that large differences in the likelihood of claiming between men of different ages became more pronounced in absolute terms. By the mid-1990s, a quarter of men aged 60-64 were in receipt of IVB, and 16 per cent of men aged 55-59, compared to only 2 per cent of men aged 20-29. Similar trends were seen for women in each age group over 40 (bottom-right panel), though prevalence rates for women in the mid-1990s remained significantly below those for men of the same age.

Chart 2.1: Invalidity benefit caseload prevalence by age and gender

2.6 There is broad agreement across the literature that the increase in male IVB prevalence during the 1980s and early 1990s primarily reflected structural changes in the labour market, rather than changes in underlying health.^[17] A number of studies have shown that in labour markets with reduced demand, men with underlying health conditions were more likely to exit jobs, usually onto incapacity benefits, whereas in stronger labour markets those with underlying conditions would continue to work.^[18] For example, a study looking specifically at labour market adjustments in the British coalfields, where 160,000 jobs were lost and only a quarter of these replaced, found that the largest adjustment was older workers moving onto incapacity benefits, with very little change in unemployment.^[19] Similar patterns were seen in areas where labour supply outstripped demand, such as in response to increases in the number of women participating in the labour market.^[20] For these reasons, incapacity benefits caseloads in these periods were often described as a form of ‘hidden unemployment’.^[21]

2.7 Compounding these labour market-related drivers, from the mid-1980s onwards, several policy factors seem to have channelled more individuals into incapacity benefits rather than unemployment benefit:

• Incapacity benefits became relatively more generous than unemployment benefit, for example due to the earnings-related additional pension effectively increasing the relative value of the former through the 1980s.^[22]
• The requirements to search for work for those on unemployment benefits increased, starting with the 1986 Restart reforms which introduced compulsory work-focused interviews with claimants, while there were no equivalent requirements for claimants of incapacity benefits. The stringency of job-search requirements then increased, including the introduction of employment centres and the 1989 removal of the requirement to be available for “suitable” employment, with individuals instead expected to be available for any job they could “reasonably be expected to do”.
• The administration of the incapacity benefits system remained passive, with few medical examinations relative to the number of claimants.

Incapacity benefit (1995-2008)

2.8 Incapacity benefit was introduced in April 1995 to replace IVB and sickness benefit for both new and existing claimants, with the differences between the two summarised in Table 2.1 above. Like IVB, IB was a contributory benefit and so required sufficient national insurance contributions, although the contributory principle was weakened with the abolition of non-contributory SDA for new claimants in 2001 and its merger into IB. The transition to IB reduced generosity via fewer and less generous age-related additions and the introduction of three different rates linked to the length of incapacity, as set out in Table 2.1 (although some claimants were eligible for income support as a means-tested top-up). The caseload for those claiming the disability premium of income support increased by two-thirds between 1995-96 and 2007-08, while spending on disability premiums increased by 57 per cent over the same period.

2.9 The transition to IB also tightened eligibility via the introduction of a medical eligibility test, known as the ‘all work test,’ which assessed claimants’ capability for work against a defined set of criteria. The all work test was initially conducted by the claimant’s GP, then from 1998 onwards began to be outsourced to the commercial sector. In 2000 it was replaced by the personal capability assessment (PCA) which included an additional ‘capability report’ intended to outline what a claimant could do despite their medical condition.

2.10 Chart 1.2 in the previous chapter showed that overall incapacity benefits caseload prevalence among the working-age population fell modestly between 1995-96 and 2007-08, from 7.9 per cent to 7.0 per cent. Chart 2.2 shows this prevalence for IB (and SDA) from 1999-00 onwards, disaggregated by age and gender (we cannot consistently disaggregate by age and gender prior to 1999-00 because the necessary data is not available for some parts of the caseload). They key trends are as follows:

• The prevalence of IB receipt for working-age men fell slightly, from a peak of 9.7 per cent in 2001-02 to 8.6 per cent in 2007-08 (top-left panel of Chart 2.2), while for working-age women it remained broadly flat over the period at an average of 6.8 per cent (top-right panel). Women made up 40 per cent of the IB (and SDA) caseload in 1999-00 and 43 per cent in 2007-08.
• The likelihood of claiming within each age group remained stable for most groups during this period (bottom two panels), aside from men aged 60-64, for whom prevalence fell by almost a third from 27 per cent in 1999-00 to 18 per cent in 2007-08. This largely reflects a cohort effect whereby many of the men who moved onto IVB during the 1980s and early 1990s reached retirement age and moved onto the state pension. There was also a more gradual reduction in the likelihood of claiming IB for men aged 55-59, from 18 to 14 per cent over the same period.

Chart 2.2: Incapacity benefit caseload prevalence by age and gender

2.11 The gradual decline in caseload prevalence during this period is generally thought to have been driven by the combination of:

• A substantial reduction in onflows, reflecting the reduction in generosity and tightening in eligibility of IB relative to IVB, alongside the relatively strong economy and employment growth over this period. This decline in onflows was more pronounced for men than women.^[23]
• A smaller reduction in off-flows. There is evidence that some interventions, such as the Pathways to Work employment programme,^[24] might have temporarily increased off-flows through claimants moving into work.^[25]

Employment and support allowance (2008-present)

2.12 Employment and support allowance was introduced for new claims in October 2008 to replace IB. ESA could be claimed on both a means-tested and contributory basis. ESA was only introduced for new claims leaving a substantial stock of IB claimants who gradually either left or moved to ESA (IB claimants began to be reassessed for ESA after October 2010).

2.13 The principal difference between IB and ESA was the introduction of the work capability assessment (WCA), which replaced the PCA within incapacity benefit (see Table 2.1). In a 2006 White Paper on welfare reform, the then Government stated that this revised WCA was intended to be “assessing people’s capability for work rather than just their entitlement to benefits”.^[26] The WCA could either deem a claimant fit for work (FFW), or place them in either the work-related activity group (WRAG) or support group (SG), depending on their ability to work. Unlike IB claimants who generally did not face any conditionality, WRAG claimants were expected to engage with a personalised programme of back-to-work support.^[27]

2.14 Chart 1.2 in the previous chapter showed that overall incapacity benefits caseload prevalence among the working-age population fell slightly between 2008-09 and 2015-16 – from 6.7 per cent to 4.9 per cent. Chart 2.3 shows the same caseloads by age for men and women. The key trends are:

• The prevalence of incapacity benefits receipt fell for men aged 50-64 and women aged 50-59, with particularly steep falls among older men. Prevalence did not decline significantly among men and women aged under 50. The prevalence of incapacity benefits was higher among men and women in their 20s in 2015-16 than in 2008-09.
• The substantial rise in the proportion of women aged 60-64 receiving incapacity benefits from 2010-11 reflects the gradual rise in the state pension age from March 2010.

Chart 2.3: Incapacity benefits caseload prevalence by age and gender (2008-16)

2.15 The modest fall in the prevalence of incapacity benefits receipt over this period was driven by a notable increase in off-flows, partly from reassessing the IB caseload under the stricter ESA WCA, as discussed in further detail in Chapter 3. As mentioned above, this was partly offset by the rising female state pension age expanding eligibility into ages at which people are more likely to claim incapacity benefits (see Box 3.1 in the following chapter for a fuller discussion).

Universal credit (2016-present)

2.16 Universal credit began to replace income-related ESA from May 2016, and became the benefit for all new claims from December 2018.^[28], ^[29] As of 2023-24, there are still 0.7 million legacy income-related ESA claimants, compared to 1.6 million claimants on health-related UC.^[30] We forecast this will gradually fall as some claimants choose to move to UC and the Department for Work and Pensions migrates those who do not choose to move.

2.17 Under UC, claimants who report a work-limiting illness or disability continue to undertake a WCA to assess their eligibility for either the limited capability for work (LCW), or limited capability for work or work-related activity (LCWRA) groups. These groups are analogous to those in ESA, so throughout the rest of this report, we refer to the ESA WRAG and UC LCW groups as the ‘less severe incapacity group’ and the ESA SG and UC LCWRA groups as the ‘more severe incapacity group’.

2.18 Unlike the shifts from IVB to IB, and from IB to ESA, the shift from income-related ESA to UC did not entail large changes in the assessment regime or relative conditionality. That said, awards were cut for new claims to the less severe incapacity group in both ESA and UC in 2017 to the same level as that for those found capable to work. This means it is now only the level of work-search requirements relative to unemployment-related claimants within UC that differs for less severe incapacity claimants, which is discussed further in Chapter 3. There are several smaller differences between UC and ESA, especially in terms of the process for initiating a health-related claim and progressing to an assessment, as well as payments being made to households rather than individuals, which are discussed in Chapter 3.

2.19 During the pandemic, the Government implemented several changes to the process of claiming the UC health element and ESA. A significant change to the assessment gateway resulted from the suspension of face-to-face WCAs from 17 March 2020. During this time, paper-based assessments were used, with telephone and video assessments introduced in May 2020. Face-to-face assessments did not resume until May 2021. Reassessments were temporarily suspended, as well as work search conditionality for the less severe incapacity group.

2.20 Chart 1.2 in the previous chapter showed that overall incapacity benefits caseload prevalence among the working-age population rose between 2019-20 and 2023-24 – from 5.4 per cent to 7.0 per cent. Chart 2.4 shows the caseload by age for men and women from 2019-20 to 2023-24,^[31] capturing the impact of the pandemic and the gradual introduction of UC. All age-and-gender groups show similar rises to the overall increase in caseload prevalence over this period.

Chart 2.4: Incapacity benefits caseload prevalence by age and gender (2019-23)

2.21 The rise in the prevalence of incapacity benefits receipt over this period was primarily driven by a rise in onflows outstripping a more gradual rise in off-flows. This is discussed further in Chapter 3, which considers in detail trends since 2010 in incapacity benefits caseloads, decomposing changes in onflows, off-flows, and the claimant journey.

Chapter 3: Incapacity benefits flows since 2010

3.1 This chapter describes in detail how the combined universal credit (UC), employment and support allowance (ESA) and incapacity benefit (IB) caseload has changed over the past 15 years. It decomposes changes into:

• onflows, defined as happening after a claimant has been through an assessment and been deemed eligible to receive incapacity benefits; and
• off-flows, defined as happening when a claimant becomes ineligible for incapacity benefits, either by: reaching a certain age;^[32] entering employment and either working above a set number of hours or earning above a certain amount;^[33] or losing eligibility after being reassessed as not eligible to receive incapacity benefits.^[34]

3.2 The analysis looks in detail at the three steps in a ‘claimant journey’ that result in an onflow: the initial claim; whether a claimant proceeds to an assessment or drops out of the process; and assessment outcomes. This provides new evidence on the drivers of incapacity benefits spending in this period. It also helps to inform our UC forecasting, which operates based on judgements about levels of onflows and off-flows over time.

The role of onflows and off-flows in driving caseloads

3.3 For the analysis in this chapter we focus on caseload levels, rather than the caseload prevalence rates (i.e. levels as a share of the working-age population) that we explored in the previous chapters. Chart 3.1 shows that both caseload levels and prevalence rates fell until 2013-14 and grew thereafter, with the pace of growth increasing after 2017-18. The slightly faster rise in levels terms partly reflects increases in the state pension age up to 2020, which expanded the eligible population (see Box 3.1). An important definitional point is that, since 2017, new claims in the less severe incapacity group no longer receive higher awards relative to unemployment-related claimants and those found fit for work at a work capability assessment. In line with the Department for Work and Pensions’ (DWP’s) approach to classifying health-related claims, and reflecting the fact that they still experience fewer work-search requirements, we still count this less severe incapacity group as part of the incapacity benefits caseload after this change (and we discuss its potential effects below).

Chart 3.1: Combined incapacity benefits caseload and caseload prevalence

Box 3.1: The effects of ageing and a rising state pension age on incapacity benefits caseloads

As shown in Chart A, the total incapacity benefits caseload increased by 330,000 people (12.3 per cent) between 2008-09 and 2023-24. Over this period, the state pension age (SPA) has been rising – from 60 to 65 for women between April 2010 and November 2018, and then from 65 to 66 for both men and women between December 2018 and October 2020.^a This box explores the role of the rising SPA in the rising incapacity benefits caseload.

Chart A shows the distribution of incapacity benefits claimants by age and sex in 2008-09 and 2023-24. It shows that of the total 330,000 caseload increase between these two points in time:

• 280,000 is made up of female claimants aged 60 and above. In turn 250,000 of this is women aged 60-64, so can be related to the rising female state pension age between 2010 and 2018. And 30,000 of this is women aged 65 and over, so partly reflects the increase in the state pension age from 65 to 66 between 2018 and 2020.
• 50,000 is made up of men aged 65 and over – which partly reflects the increase in the state pension age from 65 to 66.
• There is almost no net change among women under the age of 60 and men under the age of 65 (a rise of 210,000 women offset by a fall of 220,000 men).

Chart A: Incapacity benefits caseloads by age and sex

The above breakdown could lead one to conclude that the entirety of the increase in the incapacity benefits caseload since 2008-09 can be attributed to the rising state pension age. However, as discussed elsewhere in this chapter, this period covers a phase in which the incapacity benefits caseload and caseload prevalence was falling – up to around 2014-15; a period in which it was relatively stable – up to the late 2010s;^b and a period of rising caseloads since then. Chart B shows caseload prevalence by single year of age, up to the age of 64, at three of these points in time. It shows that:

• Caseload prevalence fell for those aged below 60 between 2008-09 and 2019-20, but rose for those between the ages of 60 and 64, reflecting the rising female SPA. Caseload prevalence fell particularly sharply between the ages of 50 and 60, although remained high. But this is an age group in which the population was growing: 50-60 year olds made up 22.5 per cent of 16-60-year-olds in 2008-09, and 25.5 per cent in 2019-20. As a result, despite caseload prevalence falling by 29.2 per cent for 50-60-year-olds over this period (from 11.6 per cent to 8.2 per cent), the caseload only fell by 16.6 per cent (from 940,000 to 790,000). So the rising SPA, and shifts in the population towards older working age beyond that, partly offset falling prevalence rates for the under-60s to limit falls in the caseload in the decade to 2019-20. Both of these factors also limited falls in the overall prevalence rate, given older people have much higher claim rates than average.
• Caseload prevalence then rose relatively uniformly at all ages (by around 2 percentage points) between 2019-20 and 2023-24. This relatively uniform rise across the age range reflects the drivers of the changing caseload discussed elsewhere in this chapter. At the same time, caseload prevalence at age 65 and above (not shown in Chart B) more than doubled between 2019-20 and 2023-24, related to the rising SPA.

Chart B: Incapacity benefits prevalence by age

In summary, the falls in the incapacity benefits caseload through most of the 2010s were significantly limited by two age-related factors. First, increases in the state pension age expanded the population eligible to receive incapacity benefits into age groups in which people are much more likely to claim. Second, a shift in the composition of the eligible population towards older working age offset the particularly sharp falls in caseload prevalence at older ages. By contrast, the rising caseload since just before the pandemic appears to have been felt relatively uniformly across the age range, and so cannot be explained by age-related factors.

a) As UC is paid to households, individual claimant age is no longer the sole determinant of eligibility for the UC health element. ‘Mixed-age couples’ on UC (households where one claimant is older than the state pension age and the other is of working age) are treated as ‘working age’ for entitlement to means-tested benefits.

b) We compare to 2019-20 here rather than 2018-19 as above because UC health data by age is not available prior to April 2019.

3.4 These trends in caseloads have been driven by substantial variation in onflows and off-flows over the post-2010 period. Chart 3.2 shows how the net effect of onflows and off-flows in each year results in the overall change in the caseload compared to the previous year, and Chart 3.3 presents onflows and off-flows as shares of the working-age population and the incapacity benefits caseload, respectively. These charts illustrate three distinct phases:

• 2009-10 to 2013-14: The falling caseload over this period was largely the result of comparatively high off-flow rates – which spiked during 2012-13 in particular – alongside low but gradually rising onflows. The high rate of off-flows was partly driven by the policy to reassess the stock of IB claimants under the new ESA work capability assessment, which began in October 2010. Of the 1.5 million IB reassessments between March 2011 and March 2024, 1.2 million happened between 2011-12 and 2013-14 and resulted in 240,000 claimants being deemed fit for work. These explain thirty per cent of total incapacity benefits off-flows over those years.
• 2014-15 to 2018-19: The caseload rose very gradually over this period due to relatively low off-flows and slightly higher, but relatively stable, onflow rates. Low off-flows and high onflows in the middle of the decade reflected the DWP’s decision to temporarily pause most ESA reassessments in 2014-15 and 2015-16 so as to bring waiting times for new claims down (discussed in further detail below). By mid-2016, this new claims backlog had largely been dealt with, which resulted in lower onflows in 2016-17 and 2017-18.
• 2019-20 to 2023-24: The caseload rose more rapidly over this period – which covers most of UC’s roll-out to date, the pandemic, and its aftermath – with the growth in onflows outstripping the growth in off-flows. Chart 3.3 shows the onflow rate increased by around 60 per cent from 2019-20 to 2023-24, while the off-flow rate only rose by around 40 per cent. The drivers of onflows and off-flows in this period are discussed in more detail below.

Chart 3.2: Annual changes in incapacity benefits caseload, onflows, and off-flows

Chart 3.3: Incapacity benefits onflow and off-flow rates

Onflows

3.5 This section considers the rises in incapacity benefits onflows in greater detail. It focuses on three key aspects: initial claims; claimants who drop out while waiting for an assessment; and assessment outcomes. The subsequent section explores changes in off-flows in greater detail, including changes in reassessments and work entry rates.

3.6 The previous section described how ESA and UC onflows have tripled from 2010-11 to 2022-23. Looking in more detail, Chart 3.4 shows that quarterly onflows increased by around 90,000 from 40,000 in May 2010 to 130,000 in November 2023.^[35] The growth has been much more pronounced in the more severe incapacity group,^[36] which made up 77 per cent of onflows in November 2023, relative to 26 per cent in May 2010.

Chart 3.4: Incapacity benefits onflows

3.7 Chart 3.5 decomposes changes in onflows (necessarily using a broader measure of onflows than our preferred one in Chart 3.2, above^[37]) into three stages of the claims process that are described in more detail in Box 3.2 below: initial claims, dropout rates,^[38] and approval rates. It shows that of the overall rise between 2010-11 and 2022-23, just over 20 per cent was due to a higher number of people initiating claims, just over 30 per cent was due to lower dropout rates, and just under 50 per cent was due to the increase in approval rates. Looking in more detail at the three time periods discussed in the previous section (paragraph 3.4, above):^[39]

• 2010-11 to 2014-15: Annual onflows were around 110,000 higher in 2014-15 than they were in 2010-11. This was mainly driven by a rise in approval rates. A rise in initial claims contributed 28,000 to the increase, but a rising dropout rate offset this by 30,000.
• 2014-15 to 2018-19: Annual onflows were around 29,000 lower in 2018-19 than in 2014-15. This was mainly explained by a fall in initial claims (accounting for 24,000 of the fall); with a small fall in the dropout rate (increasing onflows by 5,000) and a slightly larger fall in the approval rate (decreasing onflows by 11,000) explaining the rest.
• 2018-19 to 2022-23: Annual onflows were 270,000 higher in 2022-23 than they were in 2018-19. All three factors contributed to this increase: a falling dropout rate played the largest role accounting for roughly half of the increase (130,000), with higher initial claims and a higher approval rate each explaining roughly a quarter of the increase (around 70,000 each).

Chart 3.5: Changes in onflows due to initial claims, dropout and approval rates

3.8 The following section considers changes in these steps in greater detail, exploring the role of the operation of the benefits system (particularly in the transition from ESA to UC), the characteristics of claimants, and broader contextual factors in driving these trends. As background to this analysis, Box 3.2 describes the UC claimant journey in more detail.

Box 3.2: The universal credit claimant journey

There are several stages before a claimant becomes an ‘onflow’ to an incapacity benefit, as detailed in Chart C below. To join either the more or less severe incapacity groups in universal credit, a claimant typically:^a

• provides evidence of a restricted ability to work (typically a fit note provided by a member of staff in a claimant’s GP surgery) – an initial claim;
• if a fit note lasts longer than 29 days, DWP immediately refers a claimant for a work capability assessment. If a fit note lasts for less time that this, then DWP refers the claimant for a work capability assessment (WCA) after 29 days if a claimant does not drop out of the process;
• the claimant undertakes a WCA, after which the assessment provider issues a recommendation and DWP provides a final decision on eligibility.^b We define a decision of more or less severe incapacity as an approval;
• if a claimant disagrees with DWP’s decision, they can ask for a mandatory reconsideration (which can also be appealed); and
• if a claimant then joins the less severe incapacity group (LCW), some may eventually be reassessed by an assessment provider to see if they remain unable to work for health reasons. Following policy changes at the 2023 Autumn Statement, if a claimant joins the more severe incapacity group (LCWRA), most will not be reassessed.

Chart C: Claimant journey for the universal credit health element

The claimant journey up to the point of onflow in UC differs in some important respects to that in ESA – summarised in Table A. We explore the role these changes may have had in trends in initial claims, dropout rates and assessment outcomes in this chapter.

Table A: Differences between employment and support allowance and universal credit health element claimant journeys

a) This is a stylised representation of a claimant journey and there are exemptions at various stages. For example, not all claimants will be required to undertake a work capability assessment or reassessment.

b) Before issuing a final recommendation, DWP calls the claimant in a ‘decision assurance call’ where they verify details provided at assessment.

Initial claims

3.9 The number of people initiating an incapacity benefits claim within either ESA or UC health has fluctuated over time,^[40] settling at a higher level in recent years than at the beginning of the 2010s (Chart 3.6). The number of initial claims rose a little from 2010-11 (averaging 160,000 per quarter) through to 2014-15 (195,000 per quarter), dipped back close to where they started by 2017-18 (165,000 per quarter), and then rose to a higher level in 2018-19 (185,000 per quarter) and higher still in the years after the onset of the pandemic (reaching 215,000 per quarter in 2022-23). But as Chart 3.5 above shows, this rise in initial claims only explains a minority (20 per cent) of the overall rise in onflows between 2010-11 and 2022-23.

Chart 3.6: Initial claims to incapacity benefits

3.10 The factors driving this increase in initial claims are likely to include:

• Increases in the state pension age between 2010 and 2020 which expanded eligibility for incapacity benefits, as explained in Box 3.1.
• Wider health, labour market and economic trends. Our 2024 Fiscal risks and sustainability report (FRS) discussed the deterioration in a range of health conditions, and particularly mental health conditions, over the course of the 2010s, with the pandemic adding pressures onto health outcomes on top of these. The disruption of the pandemic and associated lockdowns also caused a spike in new claims across the different parts of UC. This may have been due to income concerns combined with greater public awareness of UC due to an increase in media coverage of it at the time. And while our 2023 FRS found no evidence of increasing generosity in incapacity benefits relative to unemployment-related benefits over this period, the large differences in awards may have become a more important consideration for potential claimants when cost-of-living pressures increased in recent years. Chart 3.7 shows these large differences, also showing that while relative generosity rose for those who didn’t receive the severe disability premium (SDP) prior to the switch to UC, it fell more substantially for claimants in the other two incapacity groups, who together represent the majority of incapacity benefits claimants. All of these broader factors may have contributed to rising initial claims in recent years.
• The structure and operation of UC. Unlike the legacy benefits system, existing UC claimants do not have to make a claim for a different benefit to begin receiving the UC health element if a health condition begins or worsens. It is possible that already being ‘in the system’ makes it easier to claim health-related UC, or increases its salience relative to the separation between benefits in the legacy system. Differences in the definition of the unemployment-related group (‘intensive work search’) in UC relative to the preceding unemployment benefit (jobseeker’s allowance) mean it is not possible to test this theory.^[41] That said, if there were more initial claims for incapacity benefits due to the structure of UC, we might expect higher dropouts and lower approval rates as UC rolled out. But actually we have seen lower dropouts and higher approval rates – as discussed further below.

Chart 3.7: The relative generosity of means-tested benefits over time

Dropouts

3.11 As explained in Box 3.2, ‘dropouts’ from ESA and health-related UC measure are defined as claimants who exit part-way through the process for claiming ESA or UC health.^[42] Up until the full rollout of UC, dropout volumes had a U-shaped increase, rising from 50,000 per quarter in 2010-11 to 90,000 per quarter in 2013-14 before falling halfway back down to 70,000 per quarter in 2016-17. With the introduction of UC, dropouts have fallen sharply, decreasing from 70,000 per quarter in 2018-19 to 40,000 per quarter in 2022-23 (a 43 per cent fall). Dropout rates, measured as a share of new claims, have followed a similar trend to volumes.

Chart 3.8: Incapacity benefits dropout volumes and share of new claims

3.12 The reasons for the changing dropout rate are not completely clear, but possible factors include:

• Duration of assessments: a longer time period between making a claim and receiving a final decision may increase the likelihood of a claimant dropping out. Chart 3.9 shows that following the introduction of ESA, the average duration of assessments increased to a peak of 192 days in August 2014. Durations then gradually fell back to an average of 82 days in the year preceding the pandemic. They then spiked upwards to a peak of 204 days in July 2021 for ESA assessments (UC peak was 329 days in August 2021), reflecting the pause of nearly all work capability assessments (WCAs) in March 2020 and their gradual reintroduction,^[43] and the reduced healthcare provider supply to do assessments as many were diverted to supporting the NHS. While durations in both UC and ESA have subsequently been falling, Chart 3.9 shows that the average duration for the UC assessment process has been longer than ESA by an average of 78 days, largely reflecting systemic differences outlined in Table A in Box 3.2. Across the period, there is some correlation between assessment durations and the dropout rate (both in the early 2010s as durations and dropouts rose, and from 2022 onwards where dropouts and durations have both fallen).
• Universal credit system changes: Chart 3.8 shows that, excluding the pandemic period, the dropout rate has decreased as UC has been rolled out. Compared to ESA, it is in some respects less burdensome for a claimant to make a claim for UC health: those already in the UC system do not have to make a claim for a new benefit, and can remain in work without any additional limits on how much they can work and earn while waiting for a work capability assessment (known as ‘permitted work’ rules, see Table A in Box 3.2).^[44]
• Changes in DWP’s operational approach: DWP changed their approach to verifying fit notes during the pandemic. Prior to the pandemic, claimants were directed to take their fit note to a jobcentre for verification. During the pandemic, DWP work coaches accepted a fit note without seeing it in person.^[45] DWP has since made this their business-as-usual approach. DWP also began issuing fit note reminders to claimants in May 2021 to ensure claimants received reminders that their fit note evidence was due to expire (fit note expiry would often result in them dropping out of the process). This reminder was further expanded in November 2023 to include two reminders (one a week before expiry and one two weeks after). Both these changes may have lowered the dropout rate.
• Changes in health: DWP qualitative research into ESA found that the main reason claimants dropped out was that their health improved.^[46] More persistent or severe health conditions among claimants could be a contributing factor to the falling dropout rate.

Chart 3.9: Median duration of assessment process for an initial claim

Assessment outcomes and approval rates

3.13 The ‘approval rate’ captures the share of claimants reaching an ESA or UC work capability assessment deemed eligible to receive incapacity benefits. The approval rate has increased significantly since the introduction of ESA, rising from 38 per cent in May 2010 to 57 per cent in May 2013 to 79 per cent in November 2019.^[47] Following the temporary spike to 96 per cent between August 2020 and February 2021 due to Covid easements, it has remained steady, at around 80 per cent, since early 2021.

Chart 3.10: Work capability assessment decisions and ‘approval’ rate

3.14 The following sections explore in more detail the factors that appear to have driven approval rate changes: changes in guidance; changes in assessment provider capacity; and reductions in awards for the less severe incapacity group. We also explore two factors that might be expected to have contributed to changing approval rates but do not appear to have done so: and changes in claimants’ age and conditions; and the introduction of mandatory reconsiderations.

Changes in guidance

3.15 The DWP has changed the operation of the WCA in response to department-led reviews, independent reviews, and broader public reception. The most notable changes occurred in the early 2010s and had the net effect of expanding the opportunities open to claimants to provide additional evidence, as well as partly expanding eligibility for incapacity benefits via easements to various assessment criteria.^[48] This is likely to have contributed to the increase in the approval rate in this period.

3.16 Separate to DWP guidance, as the WCA bedded in and became better understood, external advice on it from organisations like Citizens Advice become more widely accessible to prospective claimants, which may have improved claimants’ chances of being approved.

Changes in assessment provider capacity

3.17 Assessment provider capacity is also likely to have affected approval rates. The DWP initially contracted with Atos as the sole supplier of ESA WCAs after 2008. Atos struggled to cope with the volume of WCAs after the introduction of IB reassessments from 2011 onwards, resulting in a growing WCA backlog and longer waits for assessments for new claimants. In March 2014, Atos announced they would withdraw from the contract in February 2015. In the intervening period, there was a temporary spike in the approval rate. In October 2014, Maximus was appointed as the new provider for WCAs, commencing in March 2015. Their introduction as the WCA provider coincided with a slight fall in the approval rate.

Reductions in awards for the less severe incapacity group

3.18 A major policy change in April 2017 was the removal for new onflows of the higher award and benefit cap exemption for the less severe incapacity group.^[49] By making the financial distinction between more and less severe incapacity groups sharper,^[50] it may have led claimants, outsourced assessors, and DWP decision-makers to be more likely to conclude that there is a health risk to placing someone in the less severe incapacity group.^[51]

3.19 Potential evidence to support this is provided by the dashed line on Chart 3.10. In November 2023, the share of initial WCA decisions that resulted in someone being placed in the more severe incapacity group was 53 per cent. This is similar to the share of initial WCA decisions which placed a claimant in either incapacity group before the policy change (56 per cent in February 2017). This is consistent with the higher award approval rates eventually rising after the policy change, although it is not possible to fully isolate policy effects from wider changes in claimants’ conditions, as discussed above.

Age and condition composition

3.20 It is possible that changes in the characteristics of those reaching assessments have driven the approval rate up. We don’t have the data to test this comprehensively, but we do have information on the age of those reaching assessment across the whole period, and the primary health condition for the earlier years (ESA) only. Overall, we do not find evidence that in the period prior to around 2017, changes in composition by either age or condition contributed to the rising approval rate:

• Age: for those aged 25 and over, approval rates have been very similar at different ages over the entire period (varying from the overall approval rate by an average of just 3 percentage points each year in either direction). The age composition of those reaching assessment has largely been consistent over time too. The only major change was a rising share in the 55-and-over group and a falling share in the 45-54 group, which broadly mirror shifts in the population age structure over this period, plus the effects of the rising state pension age widening eligibility in the older age group. As a result, simple shift-share analysis shows that age composition does not at explain any of the rising approval rate between 2009-10 and 2022-23.
• Condition: similar analysis that very simply splits conditions into mental health and other conditions shows that composition by condition appears to play a negligible role up to 2015-16. Mental health and non-mental health approval rates are similar on average over this period, meaning the shift in the composition of assessments towards mental health as the primary condition – from 38 per cent of assessments in 2009-10 to 48 per cent of assessments in 2015-16 – does not have a compositional effect on approval rates. However, the rapid rises in the mental health approval rate from below to above average over that period (by 2015-16, the mental health approval rate was 71 per cent, compared to a non-mental health approval rate of 53 per cent), alongside a clear gradual upward trajectory in the share of WCAs in the mental health group, means that higher mental health claims could have played a role in rising approval rates up to just before the pandemic. Unfortunately, the absence of primary condition data for UC health means we cannot test this.

3.21 While the composition of health conditions was not driving the increasing approval rate under ESA, the severity of those health conditions may have been. As referenced in paragraph 3.10, our 2024 Fiscal risks and sustainability report discussed deterioration in a range of health conditions over the course of the 2010s, likely intensified by the pandemic. This wider increasing severity of health conditions is consistent with the rising approval rate seen across both ESA and UC.

The introduction of mandatory reconsiderations

3.22 The introduction of mandatory reconsiderations in 2013 was another major change in this period. Until 2013, claimants who disagreed with their benefit decision could appeal directly to an independent tribunal. Mandatory reconsiderations (MRs) added a new stage where DWP would first internally reconsider the benefit decision. In most cases, a claimant can only take their case to appeal if they are dissatisfied with the outcome of the MR.

3.23 Chart 3.11 shows that in 2012-13, the last full year before the introduction of MRs, there were 78,000 ESA appeals for ‘fit for work’ outcomes (FFW appeals), of which 27,000 resulted in an incapacity benefits onflow (11 per cent of total onflows in that year). In 2014-15, the first full year after the introduction of MRs, there were 80,000 ESA FFW MRs and 21,000 ESA FFW appeals, which together resulted in 23,000 incapacity benefits onflows (8 per cent of total onflows).

3.24 The introduction of MRs therefore does not appear to have had a significant effect on the incapacity caseload. While the introduction of MRs was associated with an increase in the volume of MRs and appeals combined, the number of cases overturned (resulting in an incapacity onflow) due to MRs and appeals remained in line with pre-MR trends. The level of cases overturned due to MRs and appeals in UC in recent years, while rising slightly over time, is substantially below that for ESA during the early 2010s.

Chart 3.11: Mandatory reconsiderations and appeals

Box 3.3: Trends in working-age disability benefit onflows

Personal independence payment (PIP) is the working-age disability benefit which helps individuals deal with the extra costs of having a disability. Unlike incapacity benefits, PIP is not means-tested and is available to those in-work. Only around one-sixth of PIP recipients are in work, a number that has been fairly stable over recent years.

There is significant overlap between disability and incapacity benefits. Around two-in-three claimants who received ESA or UC also received PIP or DLA, and around two-in-three claimants who received PIP or DLA also received ESA or UC in November 2020. However, the two types of benefits have different purposes and trends over time, as discussed in Chapter 1. This report mainly focuses on incapacity benefits, but there are interesting differences between trends in onflows for incapacity and disability benefits, with somewhat more granular data available for disability benefits. This box therefore summarises trends in PIP onflows, as a point of comparison for the incapacity benefits onflows trends discussed in this chapter.

Decomposing onflows to personal independence payment

Working-age onflows to PIP since 2018-19 have grown at a similar rate to incapacity benefits onflows, with both roughly doubling over that period. Working-age PIP onflows increased by 98 per cent from 2018-19 to 2022-23, rising from 175,000 to 348,000. Incapacity benefits onflows increased by 111 per cent over the same period (from 252,000 in 2018-19 to 533,000 in 2022-23).^a The drivers of higher onflows are very different across the two benefits, however. Chart D below decomposes these drivers:

• Initial claims: new claims for PIP increased by 310,000 (67 per cent) between 2018-19 and 2022-23, a much larger increase than new claims for incapacity benefits (which rose by 140,000, or 21 per cent). Higher new claims accounted for nearly nine-tenths of the increase in PIP onflows, compared to one-quarter of the increase in incapacity onflows.
• Dropout rate: the dropout rate for PIP claims fell from 26 per cent in 2018-19 to 14 per cent in 2022-23, accounting for one-fifth of the increase in onflows. In contrast more than half of the increase in incapacity onflows over this period is explained by a lower dropout rate.
• Approval rate: where the incapacity approval rate increased from 66 per cent in 2018-19 to 81 per cent in 2022-23, the PIP approval rate decreased slightly from 55 per cent to 51 per cent over the same period. This fall in the approval rate decreased PIP onflows by 17,000 between 2018-19 and 2022-23.

Chart D: Changes in incapacity and disability benefits onflows due to initial claims, dropout and approval rates

Understanding dropouts from the personal independence payment claim process

Dropouts from PIP can be split into the following groups: claimants withdrawing their claim; claimants failing to return the PIP questionnaire (FTRQ); claimants dropping out between submitting a questionnaire and DWP referring them for assessment; and claimants failing to attend their assessment (FTA).

As shown in Chart E, most claimants who dropout fail to return the PIP questionnaire (FTRQ). These ‘FTRQ disallowals’ explain most of the fall in the dropout rate after 2019-20, falling from 100,000 in 2019-20 (20 per cent of initial claims) to 60,000 in 2020-21 (12 per cent of initial claims). This decreasing trend in FTRQ disallowals has brought the dropout rate down to 13 per cent in 2023-24, half its pre-Covid level.

DWP changed the process for submitting the PIP questionnaire (‘PIP2′) in March 2020 to include a reminder and an automatic extension of the deadline for submitting the PIP2 form by two weeks if a claimant runs out of time. And since March 2020, claimants have been able to submit PIP2 forms over email rather than by post. These changes, which are likely to have reduced FTRQ disallowals, mirror changes in the UC fit note verification process discussed in this chapter. This evidence therefore supports the idea that a fall in the number of ESA and UC claimants dropping out due to expired fit notes similarly explains the falling incapacity benefits dropout rate.

Chart E: PIP dropouts by reason

Understanding the primary conditions of working-age personal independence payment onflows

PIP onflows can be broken down by their primary health condition. Chart F shows PIP onflows for mental health conditions and all other conditions. Up until 2019-20, the share of onflows that were for mental health conditions was roughly flat; following the onset of Covid, this share has risen by a third, up from 28 per cent in 2019-20 to 38 per cent of onflows in 2023-24.

Chart F: PIP onflows by primary health condition

Unlike PIP, we do not know if the rise in UC onflows relates to claimants with a primary mental health condition, because UC assessment data does not record the primary condition of claimants (it instead records all conditions, and only since January 2022). A consistent time series of data on UC incapacity claimants’ conditions – ideally including primary condition for consistency with ESA – would allow us to better understand changes over time, and whether the rise in UC incapacity onflows reflects the trends seen in PIP and wider health surveys.

a) Incapacity benefits onflows are defined as the sum of work-related activity group and support group decisions for ESA and limited capability for work and limited capability for work or work-related activity decisions for UC.

Off-flows

3.25 This section considers changes in off-flows in greater detail, focusing on reassessments and employment support programmes. Chart 3.3 above showed that off-flow rates, while volatile, fell from their 2012-13 peak through to 2018-19 and have been rising very slightly since, with Chart 3.12 presenting these trends on a quarterly basis.^[52] In terms of the level of off-flows, Chart 3.12 shows a spike in 2012-13 as DWP reassessed many IB claimants under the tighter ESA WCA. Off-flows then mostly fluctuated between 40,000 and 60,000 per quarter before the pandemic reflecting: changing volumes of reassessments; wider labour market conditions; and the rising female state pension age, which lowered off-flows by allowing claimants to stay on the caseload for longer. The slump during the pandemic likely reflects the limited labour market opportunities available to claimants during that period.

Chart 3.12: Incapacity benefits off-flows

Reassessments

3.26 Reassessments have driven a third of all incapacity off-flows since 2009-10. The level of reassessment off-flows has varied significantly across this period (Chart 3.13):

• Early years and IB migration (2009-10 to 2013-14): the introduction of IB reassessments in October 2010, alongside high volumes of ESA reassessments and a high reassessment disallowal rate,^[53] led to very high levels of reassessment off-flows, averaging 150,000 a year between 2011-12 and 2013-14.
• ESA steady state and introduction of UC (2014-15 to 2019-20): themajority of IB reassessments were completed by the end of 2013, and in March 2014 the Government announced that the DWP would focus on delivering assessments for those making new claims for ESA and claimants whose conditions had changed, substantially reducing the number of monthly reassessments. Alongside a lower average disallowal rate, off-flows due to reassessments fell to a pre-pandemic low of 35,000 in 2015-16. Reassessment off-flows temporarily picked up in 2016-17 and 2017-18 as the new assessment provider Maximus dealt with the reassessment backlog, but fell back once the backlog was addressed, reaching 28,000 in 2019-20.
• The pandemic and its aftermath (2020-21 to 2022-23): most reassessments for ESA and UC were paused in response to the Covid pandemic, and the disallowal rate fell below 1 per cent. Reassessment volumes continued to fall over the rest of the period, and despite an increase in the disallowal rate as some Covid easements subsided, there were only 3,000 reassessment off-flows in 2022-23. The previous Government’s response to the WCA consultation at Autumn Statement 2023 confirmed its policy intention not to reassess the vast majority of more severe incapacity cases, marking a significant shift in their reassessment policy compared to the previous decade.^[54]

Chart 3.13: Incapacity benefits off-flows due to reassessments

Employment support programmes

3.27 In addition to reassessments, DWP has also substantially changed the employment support offered to incapacity claimants during this period. DWP has offered some form of employment support to incapacity benefits claimants since the early 2000s – spanning Pathways to Work, Work Choice, Work Programme, and the Work and Health Programme – but the volume of support, and whether claimants are obliged to engage with it, has changed over time.

3.28 The evidence on DWP’s largest employment support programmes since the introduction of ESA shows some positive but limited results:

• Quasi-experimental evidence on the Work Programme – the DWP’s largest employment programme aimed at claimants in receipt of out of work benefits between 2011 and 2017 – found that the average effect of the programme on claimants who participated was to increase the days they spent in employment (by 46 days) and reduce days they spent in receipt of out-of-work benefits (by 70 days).^[55] However, the evaluation only covered jobseeker’s allowance (JSA) claimants – we do not yet know whether it had additional impact for people on incapacity benefits.
• Quasi-experimental evidence on Work Choice – a specialist disability employment programme available to JSA claimants with health conditions and ESA claimants not on the Work Programme between October 2010 and November 2017 found positive employment effects from claimants participating.^[56]
• The Work and Health Programme – which replaced the Work Programme and Work Choice and began to be rolled out from November 2017 – was evaluated in a randomised controlled trial. A quantitative evaluation in 2023 compared participants to a control group and found that voluntary participants were statistically significantly more likely than a control group of claimants receiving business-as-usual Jobcentre Plus support to report having done some work since starting the programme.^[57] DWP has not yet published a full quantitative impact analysis tracking employment, earnings, and benefit impacts.
• The previous government announced Universal Support – an employment support programme aimed at a similar group of claimants to the Work and Health Programme – in 2023. While it is too early to fully evaluate the programme, DWP and the Department for Health and Social Care conducted a randomised controlled trial of a health-led employment service,^[58] which informed the design of Universal Support. This evaluation found a statistically significant positive impact on the employment of participants in the West Midlands Combined Authority, but no impact on participants in the Sheffield City Region.^[59] The effect of the programme on benefit receipt was not measured.

3.29 In addition to these large programmes, some local authorities and combined authorities have developed their own employment support programmes for incapacity benefits claimants, but there is generally no robust evidence on employment entry and benefit off-flow from these programmes. Where there is evidence, it is often mixed,^[60] and these programmes are generally small-scale.

3.30 Employment support programmes open to incapacity benefits claimants are typically open to a wide group of benefit claimants,^[61] making it difficult within the evaluation evidence available to isolate impacts on incapacity benefits claimants in particular, and not all have undergone a quasi-experimental impact evaluation. That said, the limited evidence available does not suggest that employment support programmes have made a significant contribution to incapacity benefits off-flows.

3.31 This conclusion is supported by the work entry rates of UC claimants, which have remained stable since UC was rolled out for most new working-age benefit claims from December 2018. Approximately 2.0 per cent of the less severe incapacity group moved into work each month in the year before the pandemic, compared to 2.0 per cent in 2022-23. The equivalent figures for the more severe incapacity group also changed little, from 1.2 per cent in 2019-20 to 1.0 per cent in 2022-23. Work entry rates fell to 1.4 and 0.9 per cent for more and less severe cases respectively during 2020-21 at the height of the labour market disruptions caused by the pandemic.

Chart 3.14: Into-work entry rates by work capability assessment outcome

Chapter 4: Conclusions

4.1 Our forecasts are always subject to a range of risks and uncertainties, which we try to highlight in our Economic and fiscal outlooks (EFOs). In analysing the evolution of incapacity benefits caseloads and spending, this report illustrates several of the risks and uncertainties surrounding our most recent welfare spending forecast, both to the upside and the downside. In this chapter we conclude by briefly setting out some of the key risks and uncertainties for welfare spending that emerge from the preceding chapters.

Conclusions from this analysis

4.2 The analysis in this report, and particularly the detailed examination of flows onto and off incapacity benefits in recent years, provides a range of insights for our future forecasts. Recent trends reinforce the importance to our forecast of assessing all three of the key historical drivers of incapacity benefits caseloads and spending – working-age population health, the state of the labour market, and the structure of the benefits system. Several of the findings in this analysis particularly highlight the importance of the policy and operational features of the benefits system.

4.3 One of the most significant features is that only a minority (20 per cent) of the rise in incapacity benefits onflows between 2010-11 and 2022-23 reflects a higher number of people initiating claims. Lower dropout rates (explaining 30 per cent of the rise in onflows) and rising approval rates (explaining 50 per cent) have been more important. The reasons for the rise in initial claims, fall in dropouts, and higher approval rate are not clear-cut, but our analysis has identified several factors:

• Rising initial claims likely reflect a rising state pension age widening eligibility; evidence of deteriorating health; labour market disruption due to Covid; and cost-of-living pressures – combined with lower generosity and higher conditionality elsewhere in the benefits system – increasing eligibility and take-up.
• A falling dropout rate could reflect falling assessment durations in certain periods; rules changes under universal credit (UC) relative to employment and support allowance (ESA) that allow claimants to do some work during the assessment phase without having to terminate claims; and changes to the processes for verifying and handling fit notes following Covid.
• A rising approval rate reflects changes to Department for Work and Pensions (DWP) guidance that eased some assessment criteria, alongside improved external advice for claimants and potentially worsening health conditions. The 2017 policy to reduce generosity for the less severe incapacity group is likely to have contributed to the rising share of approvals for more severe incapacity.

4.4 In terms of off-flows, only reassessments appear to have substantially changed caseload trends, particularly in the early 2010s. There is inconclusive evidence in relation to the effects on off-flows of employment programmes for incapacity benefits claimants.

4.5 And our broader historical analysis shows that, over the past half century, the incapacity benefits system has rarely reached any kind of steady state, making the basis from which to forecast future trends highly contextual and uncertain.

4.6 These and other insights provide areas for further research or information, and issues that we will place greater emphasis on in our forecast scrutiny, working with DWP analysts.

Areas for further research and information

4.7 Analysing changes in incapacity benefits is made harder by limited data and research available on UC incapacity claimants. The most critical gaps are:

• UC incapacity claimants’ primary health condition: primary health condition data is available for all significant working-age incapacity and disability benefits apart from UC.^[62] The inability to look at UC work capability assessment (WCA) decisions or caseload data by primary health condition – the predominant condition a claimant first presents with – severely limits our ability to analyse trends in conditions over time. This data would allow us to better assess whether changes in the incapacity benefits caseload track other health-related survey evidence, and the extent to which changes in approval rates can be explained by changes in claimants’ health conditions.
• UC incapacity onflow origins: WCA decisions resulting in a claimant being placed in the UC health caseload are not split by prior benefit or employment status. Further research into claimant backgrounds and journeys would improve our understanding of the dynamics of the recent rise in onflows.
• The effects of relative conditionality and generosity on benefit flows: as discussed across this report, the wider benefits system – in particular the conditionality and generosity associated with incapacity benefits relative to other parts of the system – has affected incapacity benefits flows over time. But it is hard to quantify the extent to which these interactions between different parts of the system contribute to caseload trends, and to specify mechanisms. Further research into how relative conditionality and generosity affect flows, possibly exploiting variations in sanction rates or the rollout of UC, in addition to qualitative research with claimants and advisors, would improve our understanding in this area.
• UC incapacity claimants’ employment and off-flow journeys: more evidence on the employment and subsequent benefit status of UC incapacity off-flows – claimants’ earnings, the sectors they work in, and how long they stay in employment – would improve our understanding of the dynamics of off-flows.
• UC incapacity claimants’ interactions with health services: the recent rise in incapacity caseloads coincides with Covid-induced disruption to healthcare services.^[63] Further research into the UC incapacity claimants’ interactions with health services would improve our understanding of the role of the health system in driving onflows and off-flows, and the health conditions of claimants.
• Incapacity benefits claimants’ interactions with employment programmes: there is limited conclusive evidence on the effectiveness of employment support programmes for incapacity benefits claimants in particular. Further research into the effectiveness of employment support programmes would improve our understanding of the extent to which these programmes can support incapacity off-flows.

Risks to, and lessons for, our forecasts

4.8 Our analysis has identified several risks to our forecasts which we will scrutinise in more depth in future, notably:

• Incapacity onflow assumptions: we currently forecast rising UC incapacity onflows based primarily on a microsimulation model that is trained on 2022-23 data,^[64] which could prove too pessimistic or optimistic if population health and labour market opportunities improve or worsen, respectively. We will therefore ask DWP to monitor population health data in addition to incapacity caseload data and labour market trends, to judge whether our onflow assumptions remain central.
• Incapacity off-flow assumptions: we currently forecast a slight reduction in the rate of UC incapacity off-flows, which could prove too pessimistic if the volume of reassessments or employment support increases substantially. We will therefore more closely scrutinise DWP’s assessment provider performance, and delivery of employment support programmes, in the baseline forecast.
• Changes in operational processes: changes to operational guidance fall outside of the OBR’s policy costings process as they do not have direct fiscal impacts, but they appear to have contributed to changes in dropout and approval rates. Further operational changes could either increase or decrease onflows, for example the introduction of new provider contracts for WCAs in September 2024, given the changes to the approval rate that coincided with the move from Atos to Maximus in 2015. We will therefore more closely scrutinise DWP’s and providers’ operational guidance and communications, and changes in outsourcing contracts.

Index of charts and tables

• Executive summary
  • Chart 1: Caseload prevalence of incapacity benefits
  • Chart 2: Annual changes in incapacity benefits caseload, onflows, and off-flows
  • Chart 3: Changes in onflows due to initial claims, dropout and approval rates
• Chapter 1 Introduction and context
  • Chart 1.1: Incapacity benefits spending as a share of GDP
  • Chart 1.2: Caseload prevalence of incapacity benefits
  • Chart 1.3: Working-age spending, split by incapacity benefits and other spending
• Chapter 2 A history of incapacity benefits
  • Table 2.1: The main features of incapacity benefits over the past half century
  • Chart 2.1: Invalidity benefit caseload prevalence by age and gender
  • Chart 2.2: Incapacity benefit caseload prevalence by age and gender
  • Chart 2.3: Incapacity benefits caseload prevalence by age and gender (2008-16)
  • Chart 2.4: Incapacity benefits caseload prevalence by age and gender (2019-23)
• Chapter 3 Incapacity benefits flows since 2010
  • Chart 3.1: Combined incapacity benefits caseload and caseload prevalence
  • Chart A: Incapacity benefits caseloads by age and sex
  • Chart B: Incapacity benefits prevalence by age
  • Chart 3.2: Annual changes in incapacity benefits caseload, onflows, and off-flows
  • Chart 3.3: Incapacity benefits onflow and off-flow rates
  • Chart 3.4: Incapacity benefits onflows
  • Chart 3.5: Changes in onflows due to initial claims, dropout and approval rates
  • Chart C: Claimant journey for the universal credit health element
  • Table A: Differences between employment and support allowance and universal credit health element claimant journeys
  • Chart 3.6: Initial claims to incapacity benefits
  • Chart 3.7: The relative generosity of means-tested benefits over time
  • Chart 3.8: Incapacity benefits dropout volumes and share of new claims
  • Chart 3.9: Median duration of assessment process for an initial claim
  • Chart 3.10: Work capability assessment decisions and ‘approval’ rate
  • Chart 3.11: Mandatory reconsiderations and appeals
  • Chart D: Changes in incapacity and disability benefits onflows due to initial claims, dropout and approval rates
  • Chart E: PIP dropouts by reason
  • Chart F: PIP onflows by primary health condition
  • Chart 3.12: Incapacity benefits off-flows
  • Chart 3.13: Incapacity benefits off-flows due to reassessments
  • Chart 3.14: Into-work entry rates by work capability assessment outcome