Whitehall Monitor 2026: Part 2 - The state of the civil service

The size of the civil service

Civil service growth is starting to level off as fewer people join

The civil service has grown every year since 2016. In Q3 2025 it was 35% larger than at its low in Q2 2016. Over the last year of available data (between Q3 2024 and Q3 2025), the civil service grew by 5,355 (1%) to reach 520,440.

The past year’s growth has been much slower than the 3.8% recorded a year earlier. Numbers fell for only the second time since 2016 in Q4 2024, while growth was modest in Q1 and Q2 2025. The latest data, for Q3 2025, however, saw a sharp rise of 3,490 people, 65% of the past year’s total growth. This was driven mainly by increases of 1,185 and 695 at HMRC and the Cabinet Office respectively.

Civil service growth since 2016 has largely been a choice made in response to Brexit and the pandemic, and, while not inevitable, much was explicable. The rationale for continued expansion, and indeed why much of the increase associated with Brexit and the pandemic has not unwound, is less clear. That is concerning, particularly given the continued growth is in spite of sustained and explicit ministerial commitments to reduce the size of the civil service, including Jeremy Hunt’s imposition of an “immediate” headcount cap in 2023. ⁴⁹⁹ Markson T, ‘Hunt announces civil service headcount cap’, Civil Service World, 2 October 2023, www.civilserviceworld.com/professions/article/treasury-announces-civil-service-headcount-cap  

The current government has also said it wants a smaller civil service but is taking a different approach to its predecessors. Soon after the election in 2024, the government lifted Hunt’s headcount cap. The government has – so far – sensibly resisted targeting a specific number for the size of the civil service. It is instead focusing, as the Institute for Government has previously advocated, on “pounds not people” ⁵⁰⁰ Thomas A, Cutting the civil service: How best to slim down and save money, Institute for Government, 3 November 2022, www.instituteforgovernment.org.uk/publication/report/cutting-civil-service  by setting ambitious administrative budget savings targets in this year’s spending review.* In last year’s Whitehall Monitor we said that Labour’s emerging approach to cutting the civil service would be “welcome if borne out” ⁵⁰¹ Worlidge J, Grama T, Urban J and others, Whitehall Monitor 2025, Institute for Government, 16 January 2025, https://www.instituteforgovernment.org.uk/publication/whitehall-monitor-2025  – it is good, therefore, that the government is sticking to its plans here. 

The government is combining this approach with exit schemes to incentivise staff to leave the civil service – comprising voluntary exit, and voluntary and compulsory redundancy schemes ⁵⁰² Comptroller and Auditor General, Government exits and redundancies, National Audit Office, 25 September 2025, www.nao.org.uk/wp-content/uploads/2025/09/government-exits-and-redundancies-1.pdf  – and reportedly several recruitment freezes.**^{, 503} Markson T, ‘MoD perm sec announces recruitment freeze’, Civil Service World, 1 October 2025, www.civilserviceworld.com/professions/article/mod-perm-sec-announces-recruitment-freeze  These schemes vary considerably in their scope and type. The Cabinet Office and the Department for Business and Trade (DBT), for example, have announced plans to cut 1,200 and 1,500 jobs respectively. ⁵⁰⁴ House of Lords, Hansard, ‘Cabinet Office: Redundancy’, PQ HL7122, 1 May 2025, questions-statements. https://questions-statements.parliament.uk/written-questions/detail/2025-05-01/HL7122/ ^{, 505} Dunton, J, ‘DBT’s plans for 1,500 job cuts and office closures dubbed ‘irrational’’, Civil Service World, 11 November 2025, www.civilserviceworld.com/professions/article/dbt-job-cuts-1500-office-closures-bristol-leeds-glasgow-pcs-union  The Department for Transport (DfT), on the other hand, is running a voluntary exit scheme for 300 staff. ⁵⁰⁶ Markson, T, ‘Nearly 300 DfT civil servants to leave under voluntary exit scheme’, Civil Service World, 23 April 2025, www.civilserviceworld.com/professions/article/dft-voluntary-redundancy-scheme-300-civil-servants-leave  

So far, the slowdown in staff growth has been driven by a fall in entrants, rather than a rise in people leaving the civil service (see the graph above). The number of entrants to the civil service decreased by over 30% from March 2024 to March 2025, the largest year-on-year decrease since the fall between 2009/10 and 2020/11 – likely due to recruitment freezes. Given the time lag involved in voluntary exit schemes, their effect has not yet been reflected in official exit data. 

Recruitment freezes are, however, a blunt tool and can be hard to enforce over time. Departments will always need at least some leeway for some recruitment, either to replace critical roles or skills that have been lost, or to respond to new pressures, where these cannot be met by existing civil servants. The increase in staff numbers from Q2 to Q3 2025 is testament to this. While recruitment freezes and voluntary exit schemes should reduce civil service numbers, they come at a high cost of restricting new talent and the risk of losing high performers. A longer term solution for the civil service that keeps bringing in new talent and avoids having to re-run expensive exit schemes is required.

The government should be using this moment as a reset – removing low performers and cutting numbers especially quickly in some departments – and then continue its welcome focus on budgets over arbitrary headcount targets. Next year’s data should reveal the longer term impact of this approach, and reveal whether the government is succeeding in getting numbers down once recruitment freezes (which are rarely in place for multiple years) have ended.

Above all, however, what is missing is the civil service’s long-promised strategic workforce plan. That plan should be the way the government evaluates, then articulates, what it needs the size and shape of the future workforce to be, and give leaders the tools to manage it – including dismissing poor performers and bringing in external talent. The plan was first scheduled for summer 2025 and then due by the end of 2025; the government has now committed to publishing it in the first half of 2026.

The newest departments have grown the fastest

Departments differ vastly in their size. The largest (MoJ) dwarfs the smallest (DCMS), with over 40 times as many FTE staff. But while the relative size of departments remains broadly the same, there have been large changes in the absolute size of some.

Nine departments grew between Q3 2024 and Q3 2025, the most striking of which was the 17.7% increase in the size of the DSIT. This is in large part due to a transfer of some 930 staff from the Cabinet Office’s digital teams to it in Q1 2025, setting DSIT up as the ‘digital centre of government’.*** HMRC experienced the largest absolute increase, adding 2,960 staff. The Cabinet Office’s chief operating officer, Cat Little, explained to parliament that this growth was due to increased hiring to “tackle fraud and error”. ⁵⁰⁸ House of Commons, Public Administration and Constitutional Affairs Committee oral evidence, ‘The work of the Cabinet Office’, 16 December 2025, https://committees.parliament.uk/event/25586/formal-meeting-oral-evidence-session

The other departments established by Rishi Sunak in his 2023 machinery of government changes continue to see the fastest proportional growth, as might be expected for younger departments. DBT grew by over 8% and DESNZ by 5.2%. DCMS’s 10.9% growth reflects its small absolute size relative to other departments – it increased by 215 people.

Seven departments shrank over the past year, with the largest decreases in the Department for Environment, Food and Rural Affairs (Defra, -3%), Ministry of Defence (MoD, -2.8%) and the Department for Transport (DfT, -2.3%). For the MoD and DfT, this is in line with longer term trends. The Department for Education also saw a reversal of recent growth, shrinking by 1.6%. The decrease at DfE is caused by fewer entrants rather than more leavers, reflecting a broader pattern across the civil service: 195 staff joined DfE in 2024/25, compared to 770 in 2023/24 and 505 in 2022/23.

This is also the case at the Home Office, which has seen a dramatic reversal of its recent growth in the past year. It shrank by 1,115 (-1.7%) last year, having increased by 5,940 (14.6%) from 2021/22–22/23 and 8,405 (18%) from 2022/23–23/24. In 2022/23 and 2023/24, there were 7,790 and 9,100 entrants respectively. This figure was 67% lower in 2024/25, at 3,025. The 2024/25 figure was in line with pre-2022/23 levels, suggesting that the high number of entrants in 2022/23 and 2023/24 were outliers, likely driven by increased recruitment for caseworkers to process the UK’s record asylum backlog during that period. ⁵¹⁰ Cuibus M, Walsh P and Sumption M, ‘The UK’s asylum backlog’, The Migration Observatory, 28 April 2025, https://migrationobservatory.ox.ac.uk/resources/briefings/the-uks-asylum-backlog/

Departments differ sharply in the proportion of civil servants that work in the ‘core’

Departmental groups within the civil service are split into two categories: the core department and other organisations**** overseen by the core department.

The proportion of civil servants working in the core department is very different across the civil service. The primary reason for this is whether operational staff are classified as working in the core department or outside of it. In the Department for Work and Pensions (DWP), for example, 97% of staff are in the core department – though this includes those working in Jobcentres. In HM Revenue and Customs (HMRC) call handlers work in the core department too. However, the opposite is true of MoJ, where operational staff tend to work outside the core department in HM Prison and Probation Service; for example, as prison officers.

This means the classification of ‘core’ and wider department does not always shed helpful light on the type of work that staff are carrying out.*****

The grade composition of the civil service has changed markedly

As the size of the civil service has changed over the past 15 years, so too has its grade composition. The most junior roles (administrative assistants and officers, AA/AOs) made up most of the decrease in the number of officials between 2010 and 2016. Their numbers continued to fall once the civil service as a whole began to grow after 2016. There are now 119,000 officials at these grades, compared to more than 224,000 in 2010.

The post-2016 growth of the civil service was instead led by the continued expansion of more senior grades. Growth in the ranks of Grades 7 and 6, for example, began as early as 2013. This cohort saw the highest proportional growth of any grade group between 2017 and 2023 – with an astonishing expansion of 13.3% between 2020 and 2021 alone – though its rate of growth has been falling since.

There are now 132% more Grade 7s and 6s than in 2010. The senior civil service (SCS), meanwhile, has expanded by half (52%); numbers of higher and senior executive officers have increased by 57%; and those of executive officers have only just surpassed their 2010 level. Between 2024 and 2025 – the latest data available – these trends continued. AA/AO numbers fell by 4%, G7 and 6s expanded by 5%, and the SCS grew by 3%.

Some of these changes can be attributed to the changing nature of work in the civil service (for example, administrative roles being automated, ⁵¹² Worlidge J, Grama T, Urban J and others, Whitehall Monitor 2025, Institute for Government, 16 January 2025, https://www.instituteforgovernment.org.uk/publication/whitehall-monitor-2025  and growth in senior-leaning policy roles around Brexit and the pandemic), but there remains clear evidence of grade inflation – where staff are offered promotions as a makeweight for overall pay restraint in the past decade or so.******

As a result of these changes, the civil service is now significantly more skewed towards the mid-level and more senior ranks than that of 2010. AA/AOs now comprise 24% of officials – down from 46% in 2010 – while G7 and 6s make up 17%, up from 7%. This pattern broadly applies across departments. The proportion of the workforce made up of AA/AOs fell in every department between 2010 and 2025, while the proportion consisting of G7 and 6s increased in all. The scales, however, differ significantly – the proportion of the MoJ consisting of AA/AOs, for example, fell by 27 percentage points during this period, compared to just 4 in DHSC.

_{* See the chapter Departmental spending and efficiencies.}

_{** See the chapter Turnover, leaving routes and exit schemes.}

_{*** See the chapter Digital transformation.}

_{**** Unless otherwise stated the data and graphs in Whitehall Monitor refer to the whole departmental group, that is, both the core department and other organisations. Further detail on how we work with core departments and departmental group (including other organisational) data throughout Whitehall Monitor is available in the Methodology.}

_{***** See the chapter Professions and functions.}

_{****** ‘Grade inflation’ is the practice of civil servants being promoted more quickly than they otherwise would have been, and roles being advertised at higher levels than previously, in an attempt to recruit and retain the best in an environment of inflation-eroded grade-specific salaries. For more information see the Institute for Government’s Whitehall Monitor 2025.}

Professions and functions

The civil service workforce is categorised into professions and functions. Professions are about skills, grouping together civil servants with particular expertise, either specialist (such as veterinarian) or broad (such as policy). Most civil servants are part of a profession, and can belong to more than one. Functions are about activities, usually cross-government such as commercial or project delivery. Central units set functional standards, and members carry out their work to those standards across government.

Only 30% of civil servants report being part of a function, so the Institute for Government tends to use professions data: 95% of civil servants report a named profession. In recent years, the government has classified the professions into functional professions (those aligned with a function, such as commercial), specialist professions (such as intelligence analysis or veterinarian), operational delivery, and policy. ⁵¹⁴ Cabinet Office, ‘Government professions’, updated 4 September 2025, www.gov.uk/government/publications/government-professions/government-professions

Most civil servants work in operational delivery, concentrated in a handful of departments

More than half of civil servants are part of the operational delivery profession (54%); given its size, this year’s Whitehall Monitor looks at the profession in more detail.

Four in five of those operational delivery staff (82%) work in just five departments: MoJ, DWP, HMRC, MoD and the Home Office. These five departments are the largest overall, accounting for 68% of total civil service staff numbers, and operational delivery staff make up a large proportion of those large workforces. They account for more than 70% in the Home Office, DWP and MoJ, and, while a smaller proportion of their workforce, operational delivery is the single largest profession in the MoD and HMRC.

Change in the operational delivery profession has the biggest impact on civil service numbers

Professions data has historically been subject to data quality issues, including unexpectedly large increases in some professions year-on-year, probably due to staff being encouraged to record and report their profession.* Nevertheless, we are able to use the data to understand the broad roles and shape of the civil service, and can do so increasingly accurately as reporting has steadily improved.

Since 2016, and because of its sheer size, the growth in officials working in operational delivery has contributed to 52% of the total growth of the civil service, despite the profession having only grown 31% during that time period. In contrast, the government digital and data (GDD) profession and the policy profession have had the highest proportional growth, at 152% and 116% respectively since 2016, but because of their smaller relative size have contributed less to overall growth in civil service numbers. The proportional growth of those professions does not show signs of slowing: between 2024–25, the GDD profession grew 13.5% and the policy profession grew 5.7%.

The growth of the operational delivery profession since 2021 is larger than that of all other professions combined, and not one that is entirely explained by improved reporting.

The departments where the profession is focused are managing issues that are increasingly under pressure, politically salient, or both. For example, prisons are currently struggling to cope with demand, and the government has enacted emergency release measures to avert an impending capacity crisis. ⁵¹⁸ Rowland C, Public services performance tracker 2025: The criminal justice system, Institute for Government, 2025, www.instituteforgovernment.org.uk/publication/performance-tracker-2025/criminal-justice/overview  The number of operational staff in prisons has had to grow significantly over the last eight years, though this is from historic lows. ⁵¹⁹ Ibid.  Previous cuts in the number of operational staff mean that the number of staff per prisoner is still down 7% on 2009/10, and prisons are struggling to retain experienced staff. ⁵²⁰ Ibid.  

Similarly, the government has made ‘secure borders’ one of its three underpinning foundations, and with the government’s desire to clear the asylum backlog comes pressure to find the operational delivery staff to do so – from Border Force operatives to asylum caseworkers. This pattern is being replicated across the large operational delivery departments, and in the context of a government seeking a smaller workforce these pressures will feel particularly acute. 

Operational staff tend to be junior, and their work varies considerably

Most of the roles in the operational delivery profession are at more junior grades – 85% of the most junior AA/AO grade is made up of the profession. Only 13% of senior civil servants are in the operational delivery profession, compared to 26% in the policy profession, despite operational delivery making up more than half of the civil service and policy only making up 6.9%.

There is no single definition of ‘operational delivery’, but it is best described as “the point at which citizens experience the government”. ⁵²⁵ Thompson J, ‘Top things to know about the Operational Delivery Profession’, blog, Civil Service Quarterly, 5 July 2018, https://quarterly.blog.gov.uk/2018/07/05/top-things-to-know-about-the-operational-delivery-profession/  This covers a range of civil service roles, usually public-facing: administering benefits or the tax system, officials issuing passports, vehicle inspectors, immigration and asylum caseworkers, and prison officers. The distinction between operational delivery officials who happen to be civil servants, and those classed as outside the civil service as public or crown servants, such as teachers or police officers, is somewhat arbitrary and based on historical decisions and happenstance.

Operational delivery staff are often thought of as the ‘front-line’ roles in the civil service, though even this definition is a slippery one, and is increasingly framed as the opposite of ‘back office’ staff by the government, usually when talking about cuts. ⁵²⁶ Morton B and Howard J, ‘Reeves confirms 15% cut to Civil Service running costs’, BBC News, 23 March 2025, www.bbc.co.uk/news/articles/cy5nzy403l0o  In health care contexts, for example, ‘front line’ will apply to doctors and nurses (who are not civil servants). Within the civil service, in areas such as welfare, prisons and immigration, it tends to apply to operational delivery staff. When Pat McFadden, then the chancellor of the Duchy of Lancaster, spoke in March 2025 about his plans to prioritise front-line jobs, his examples were border guards and prison officers. ⁵²⁷ Whannel K and Watson I, ‘Civil Service reforms will be radical, minister vows’, BBC News, 9 March 2025, www.bbc.co.uk/news/articles/c9q4nr42z20o  Both are civil service operational delivery roles.

But the operational delivery profession also includes people who fall into ‘leadership and management’ and ‘business management and operational support’ roles. ⁵²⁸ Operational Delivery Profession, ‘Job Families and Role Clusters’, 24 July 2024, www.odp.civilservice.gov.uk/skills-framework-new/job-families-and-role-clusters  These staff provide support and management, without necessarily interacting with the public on a daily basis. It is unclear whether the government’s idea of the ‘front line’ includes these supporting roles, or how it aligns with the goals the government has set itself on immigration, security and – more broadly – delivery.

If the government’s aim is to protect – or even grow – operational delivery (or ‘front-line’) capacity while cutting the civil service, very large proportional reductions in headcount across departments would be needed to balance the impact of the profession’s growth on overall numbers.

The government needs to be specific about what it is looking to achieve by growing the ‘front-line’ capacity at the expense of other professions – and what it means for operational delivery roles that might not meet this definition. When ministers make the case for trade-offs between ‘front-line’ and ‘back office’ roles, they need to explain what the impact of such changes would have on government achieving its objectives – and what cuts to ‘back office’ roles mean as well. The forthcoming workforce plan will need to clarify this, and set out a rationale for the size and seniority of professions in government.

_{* In perhaps the clearest recent example, the number of staff in the operational delivery profession in the MoD reportedly went from 820 in 2022 to 10,820 in 2023. There is a drop of about 15,400 in the number of civil servants recorded with a profession of ‘unknown or other’ between 2022–23, coinciding with this increase.}

Turnover, leaving routes and exit schemes

Turnover has fallen sharply after years of high churn

High turnover in the civil service has long been a problem. It damages institutional memory, expertise and productivity, and brings with it higher costs of recruitment and training. ⁵³⁰ Sasse T, Moving on : The cost of high staff turnover in the civil service, Institute for Government, 15 January 2019, www.instituteforgovernment.org.uk/publication/staff-turnover-civil-service  

This turnover, or ‘churn’* – and in particular the rate of inter-departmental transfers, where civil servants move between departments – has been more volatile since 2020 than in the 10 years before. The pandemic probably accounts for the lower number of leavers in 2020/21 as civil servants stayed in their jobs, and also for the higher number of moves in 2021/22 as pent-up demand to move jobs could be acted upon. The second spike, in 2023/24, is at least in part attributable to the creation of the Department for Energy Security and Net Zero (DESNZ), the Department for Science, Innovation and Technology (DSIT) and the Department for Business and Trade (DBT) in 2023.

But the sharp fall in total turnover in 2024/25 – from 12.7% in 2023/24 to 8.9% – is driven not by fewer people leaving the civil service, but predominantly by civil servants moving between departments less frequently. There was a 3ppt drop in internal transfers between 2023/4 and 2024/25.
 

This could simply indicate a return to the pre-2016 level of internal turnover. However, the scale of the decrease in a single year suggests that another mechanism might be at play. Recruitment freezes** may have resulted in fewer officials moving between roles, either because fewer roles are available or because those roles are first advertised within the department, and only if they cannot be filled within the department are they advertised across the civil service or externally. Data on moves within a department is not publicly available, and it is possible that this type of churn has increased.

The government should not mistake lower turnover in a period where there are impending exit schemes (discussed below) and ongoing recruitment freezes for evidence that they have fixed the long-term problem. Instead, it needs to focus on tackling the incentives that encourage excessive churn in normal times: chief among these are that many civil servants feel they must move job either to ‘climb the ladder’, or simply to get a pay rise.

There are some examples of government addressing this. The digital and data profession already lets high-performing civil servants earn more each year while doing the same role (pay progression), and the FCDO has in place minimum service terms for some roles, which may be helping it record some of the lowest rates of turnover of all departments.

The Institute for Government has called for both to be implemented across the civil service, ⁵³³ Worlidge J, Urban J, Clyne R and Thomas A, 20 ways to improve the civil service, Institute for Government, July 2024, www.instituteforgovernment.org.uk/sites/default/files/2024-07/civil-service-reform.pdf  and the government should at least evaluate the success of these existing pilots with a view to expansion. It is also a welcome step that the Cabinet Office is exploring pay progression for the senior civil service. ⁵³⁴ Civil Service, ‘Government evidence to the Senior Salaries Review Body on the pay for the Senior Civil Service, GOV.UK, 30 October 2025, https://www.gov.uk/government/publications/government-evidence-to-the-senior-salaries-review-body-on-the-pay-of-the-senior-civil-service  

Levels and type of turnover vary between departments

Both the proportion of the workforce leaving – either for other departments or leaving the civil service entirely – and the split between those two routes, varies between departments.

In some places, high turnover is unsurprising. In the centre of government and in particular for the Treasury and Cabinet Office, higher levels of total turnover are expected (if not always welcomed) given they often bring in civil servants with experience of other departments on secondment. This model probably also accounts for at least some of the high proportion of Treasury leavers who stay in the civil service when the leave, finishing their stint in the centre of government before returning them to their ‘home’ departments. As noted, DCMS’s smaller absolute size relative to other departments means a small number of officials moving roles can equate to higher proportional recorded turnover, and while DfT’s 14.7% is high, it is in line with the department’s historic churn.

The other department to note is DHSC, where the drop to a 10.5% turnover rate brings it back into line with other departments, having seen a total turnover rate of over 50% in 2022/23 and 25% in 2023/24 as the influx of staff to the department during the pandemic sought other roles outside of DHSC. This is accompanied by a significant increase in the department’s morale scores.***

Most departments see a fairly even split between internal transfers and leavers, although it is notable that the MoD and DWP staff appear to be much more likely than other civil servants to leave the civil service entirely.

Resignations and retirements continue to account for the majority of leavers

The total number of officials leaving the civil service fell slightly in 2024/25, to 38,616. The vast majority of those leavers continue to be made up of resignations and retirements (79%). The proportion and number of retirements rose in 2024/25, and the proportion and number of resignations fell. Fewer resignations could be a sign that some officials have delayed resigning while they waited to hear more about, and see if they are successful in applying for, voluntary exit schemes and accompanying payouts. However there are other labour market factors, including a continued fall in the number of vacancies in the UK, ⁵³⁶ Office for National Statistics, ‘Vacancies and jobs in the UK: November 2025’, 11 November 2025, www.ons.gov.uk/employmentandlabourmarket/peopleinwork/employmentandemployeetypes/bulletins/jobsandvacanciesintheuk/latest  that may have contributed to this drop in resignations.

Leaving routes tend to vary by department and over time. A snapshot from 2024/25 shows how, even while resignations and retirements together account for the vast majority of leavers, those proportions vary significantly between departments. Retirements are highest in some of the big operational departments: DWP, HMRC and MoD in particular. It is at DWP, HMRC and the Home Office – with their large operational delivery workforces – that dismissals are also the highest.
 

The number of civil servants who want to stay in their organisation for at least three years is in steady decline

Across the civil service as a whole, the majority of employees in 2024**** planned to stay in their organisation for at least three years (50.4%). The long-term trends in this response to the annual People Survey are striking, with fewer people wanting to stay in their organisation for at least three years, and more wanting to stay for at least a year. Both of those main trends reversed slightly in 2024, but not by much.

New exit schemes are ambitious by recent standards

In December 2024 the Cabinet Office became the first department in this round to publicly announce a voluntary exit scheme – aiming to reduce headcount by 400. ⁵⁴² Dunton J, ‘Cabinet Office launches voluntary-redundancy drive’, Civil Service World, 12 December 2024, www.civilserviceworld.com/news/article/cabinet-office-launches-voluntaryredundancy-drive  By May 2025 it had tripled that target to 1,200. ⁵⁴³ House of Lords, Hansard, ‘Cabinet Office: Redundancy’, PQ HL7122, 1 May 2025, https://questions-statements.parliament.uk/written-questions/detail/2025-05-01/HL7122/  The National Audit Office (NAO) found that as of August 2025, 33 departments and arm’s length bodies had sought to introduce some form of exit scheme, targeting 8,586 leavers over the next two years. ⁵⁴⁴ Comptroller and Auditor General, Government exits and redundancies, National Audit Office, 25 September 2025, www.nao.org.uk/wp-content/uploads/2025/09/government-exits-and-redundancies-1.pdf.  Those figures have now likely increased.

The NAO sets out three types of exit scheme:

• Voluntary exit – staff apply, requesting to leave in return for a payment. There is no guarantee staff will be successful in their application.

• Voluntary redundancy – where roles are being reduced staff can be offered a payment in return for leaving. In the civil service this must be offered before compulsory redundancy, and payments are usually more generous.

• Compulsory redundancy – where voluntary redundancy schemes have not incentivised enough staff to leave, the civil service may run compulsory redundancy rounds with payments to staff based on length of service and salary.

In March 2025 the government also announced a ‘mutually agreed exits scheme’ for civil servants, which could be used to incentivise under-performing staff to leave. ⁵⁴⁵ Whannel K , Watson I, ‘Civil Service reviews will be radical, minister vows’, BBC News, 8 March 2025, www.bbc.co.uk/news/articles/c9q4nr42z20o  The government has not released further information, but the NAO confirmed a small Cabinet Office pilot; as of August 2025 there had been applications to the Cabinet Office from departments for just 30 employees to leave under a mutually agreed exit. ⁵⁴⁶ Comptroller and Auditor General, Government exits and redundancies, National Audit Office, 25 September 2025, www.nao.org.uk/wp-content/uploads/2025/09/government-exits-and-redundancies-1.pdf

We currently understand that voluntary exit schemes are being run in over a third of core departments: Cabinet Office, DBT, Defra, DESNZ, DfE, DHSC, FCDO and the Ministry for Housing, Communities and Local Government (MHCLG). More departments may be running exit schemes, but there is no requirement for government to report them publicly.

Looking at the civil service as a whole gives a sense of the scale of ambition for these schemes compared to those that have been run in the recent past. The average number of leavers through redundancy and voluntary exit schemes between 2018/19 and 2024/25 has been 1,678, and in five of those seven years fewer than 1,000 staff left through those schemes. As the NAO reported, as of August 2025 there were applications for 8,586 such exits (applications made since April 2024), with staff expected to leave before 31 March 2027.

This number is likely now higher; as the NAO published its report last year there were additional exit schemes planned, but these were not included as the data was incomplete at the time. On average over the course of this financial year and the next we should expect over two and a half times the number of average leavers through exit schemes – more than 10 times as many leavers as those who took up exit schemes in 2024/25.*****

Looking in more detail at just the Cabinet Office gives an idea of the scale of change that will be felt in some departments. The Cabinet Office expects 1,200 people to leave by 2026/27, through a mixture of voluntary exit schemes and some natural attrition (not replacing roles when they are vacated). The planned 1,200 exits is far higher than the Cabinet Office’s historic numbers of leavers, not only through previous bulk exit schemes where numbers are in the tens, but also from all leaving routes: 2023/24 saw 456 Cabinet Office leavers (of which 16 were from an exit scheme) and 2024/25 saw 454 (of which 19 were from an exit scheme). As of May 2025, 540 Cabinet Office staff had been approved to leave through a voluntary exit scheme. It is at best unclear the extent to which attrition can meaningfully contribute to the remaining 660 roles. In Q3 2025, for example, the Cabinet Office grew by 694 (FTE). Much of that was the September intake of the fast stream graduate programme, but even discounting that the department grew by 245 (FTE).
 

The current schemes are targeting an ambitious number of people in comparison to recent similar schemes. However, to meet its own administration budget savings targets, we estimate that the civil service would need to shrink by something between 29,000 and 40,000 staff by 2029/30.******

It appears that exit schemes, alongside the focus on administration budget savings targets, have, through tighter recruitment practices, started to make an impact on civil service numbers. Nevertheless, it appears that even accepting the 8,586 figure as an underestimate, voluntary exit schemes alone will not achieve the scale of headcount reductions that the government wants.

Exit schemes must improve long-term civil service capability if their costs are to be justified

It is not clear that the government has reckoned with the costs – either in monetary or capability terms – of these exit schemes. The NAO estimates that 8,586 exits will cost over half a billion pounds in redundancy payments alone. The Treasury has set aside £150 million of the spring statement’s £3.25bn transformation fund for exit schemes, ⁵⁵¹ HM Treasury, Spring Statement 2025, GOV.UK, 26 March 2025, https://assets.publishing.service.gov.uk/media/67e3ec2df356a2dc0e39b488/E03274109_HMT_Spring_Statement_Mar_25_Web_Accessible_.pdf  but this is match funding for departments who will still need to find their own funds.

As well as those up-front payouts, there is a cost to exit schemes in the likelihood of losing good civil servants (‘regrettable losses’ ⁵⁵² Review Body on Senior Salaries, Forty-Seventh Annual Report on Senior Salaries, Report No.98, May 2025, https://assets.publishing.service.gov.uk/media/682f20dec054883884bff424/SSRB_47th_Report_2025_Web_Accessible.pdf  in the euphemistic terminology). The government should adopt its own test and learn philosophy here, and publish aggregated data on the civil servants who leave under these schemes, including their most recent assessments of performance, to inform future schemes.  

Costs may be justified if the exit schemes act as an effective reset moment for the civil service, creating a cultural shift of expectations on staffing numbers. One of the NAO’s tests for an exit scheme is whether measures are taken to “prevent staff numbers creeping up again”. ⁵⁵³ Comptroller and Auditor General, Government exits and redundancies, National Audit Office, 25 September 2025, www.nao.org.uk/wp-content/uploads/2025/09/government-exits-and-redundancies-1.pdf.  The Institute for Government has previously argued for a far stronger culture of robust performance management; now is the right time for the civil service to introduce a systematic process to more easily remove staff who are under-performing.

That could be through the new mutually agreed exits process, which government will need to show can prove value for money when operated at scale and remove the poorest performers, or through regular compulsory redundancy rounds (in the civil service these would currently have to be preceded by voluntary redundancy rounds) where selection criteria is based on past performance, or through improving the current dismissals process for under-performance.

Regardless of how it is done, regular exit or redundancy rounds based on performance ⁵⁵⁴ Thomas, A, ‘Pat McFadden is right: the civil service needs to lose poor performers’, Institute for Government, 9 March 2025, www.instituteforgovernment.org.uk/comment/civil-service-mcfadden-performance  would keep numbers down while opening up the civil service to fresh ideas through new entrants and sharpening performance management. The government should use its current opportunity to implement changes that might have seemed radical in the past but are now urgent.

_{* In this chapter, 'turnover' and 'churn' are used interchangeably.}

_{** Recruitment freezes are not formally defined and vary between departments, but typically mean fewer roles are advertised and recruitment is done from inside the civil service in the first instance.}

_{*** See the chapter Morale and engagement.}

_{**** This data is taken from the Civil Service People Survey – it is carried out between September and October each year and published the following January. The latest figures therefore reflect civil servants’ plans for staying in their roles as of autumn 2024.}

_{***** It should be noted that this data is from two different sources. The exit figures up to 2024/25 are from Cabinet Office, ‘Civil Service Statistics’ and record civil servants who have left through those routes. The NAO figure of 8,586 is the number of exits departments had applied for since April 2024 and as of August 2025. It is possible some of those 8,586 staff left in 2024/25 and are captured in the 2024/25 exits figure. The 8,586 figure is an underestimate.} 

_{****** See the chapter Departmental spending and efficiencies.}

Pay

Civil service pay is recovering from deep real-terms cuts

Between 2010 and 2023, a combination of pay restraint and high inflation brought about large real-terms pay cuts for all civil service grades. Some of this ground has been made up in recent years.

For 2023/24, departments could make average pay awards to junior officials (also known as delegated grades, those below the SCS) of up to 4.5%, with the flexibility to pay up to 5% to the lowest paid, and were also able to make a one-off, across-the-board payment of £1,500. Senior civil servants all received an increase of 5.5%, with a further 1% available for those who were lower paid. ⁵⁵⁸ Cabinet Office, ‘Civil Service Pay Remit guidance, 2023 to 2024’, guidance, GOV.UK, 14 April 2023, www.gov.uk/government/publications/civil-service-pay-remit-guidance-2023-to-2024  This was followed by a 5% uplift for both delegated grades and the SCS in 2024/25. ⁵⁵⁹ Cabinet Office, ‘Civil Service Pay Remit guidance, 2024 to 2025’, guidance, GOV.UK, 29 July 2024, www.gov.uk/government/publications/civil-service-pay-remit-guidance-2024-to-2025

As a result of these decisions, most grades have seen real-terms increases for the past two years (see figure below).* These are, by design, most generous for the lowest paid. The most junior AA/AO ranks saw their pay increase by more than 6.4% in real terms between 2024 and 2025, and by 10.6% since 2023.

Over the entire period from 2010 to 2025, median pay in the civil service has roughly kept pace with median pay in the private sector, although pay for individual grades has fallen in real terms. The picture for the civil service as a whole is in part due to a shift in the grade structure, more detail on which is below. By 2025, median pay in the civil service was 0.4% higher in real terms than in 2010.

Pay data published later in 2026 will reflect the Labour government’s decision to award a 3.25% pay rise to both delegated grades and SCS for 2025/26. The pattern in future years is likely to be shaped by the Civil Service Reward Strategy, which is in development. ⁵⁶⁰ Civil Service and Review Body on Senior Salaries, ‘Government evidence to the Senior Salaries Review Body on the pay of the Senior Civil Service (October 2025) (HTML)’, transparency data, GOV.UK, 30 October 2025, https://www.gov.uk/government/publications/government-evidence-to-the-senior-salaries-review-body-on-the-pay-of-the-senior-civil-service

Civil servants are unhappy with their pay

Officials’ satisfaction with their pay and benefits has been persistently low, and has long been the lowest of the nine ‘theme scores’ in the People Survey. Satisfaction on this metric has – on average – hovered around the 30% mark since 2010, though it reached a high of 37.6% in 2021 before falling almost 10 percentage points to 28.7% in 2022. The measure stood at 31.7% in both 2023 and 2024.

Unsurprisingly, officials’ satisfaction with their pay and benefits has – since 2010 – been strongly correlated with actual median pay.** There are some anomalous years, however – satisfaction in 2021 (marked as 2021/2022 in the figure below as it corresponds to pay data from 2022; see chart notes) was relatively high given the level of real-terms pay. This may well have reflected civil servants’ satisfaction with the security of their pay and benefits compared to peers during the pandemic, rather than with their pay itself.

As we have noted in previous editions of Whitehall Monitor, the trend of relatively low pay has also had a significant impact on the structure of the civil service itself. The need to attract, retain and incentivise officials in such a difficult pay context has led to some officials being promoted sooner than they otherwise would have been, and the same roles being advertised at higher grades; the impact of this ‘grade inflation’ is shown above in the first figure in this chapter.

This has contributed to a shift in grade structure of the civil service, which since 2010 has become significantly more weighted towards mid-level and more senior roles.*** It is possible that the current trend of rising pay could mitigate future grade inflation, but the damage of skewed perceptions around what work is done at what grade is likely to be already done.

Pay varies substantially between departments

Civil servants are paid very differently depending on what department they work in. Even at the same grade, pay can vary significantly. The starkest discrepancy between core departments is seen at the most junior ranks (AA/AOs), for whom the highest median pay (DfT, £38,460)**** is almost 60% more than the lowest (MoJ, £24,200).

Such discrepancies are the result of a mix of factors: different proportions of individual grades in different departments; departments setting their own pay policy for delegated grades; external hires including specialists being offered higher salaries (rather than starting at the bottom of the relevant pay scale); and the legacy of automatic progression through a pay scale linked to time served, before this policy was stopped in the 2010s.

But regardless of the cause, these discrepancies have consequences – including incentivising officials to move between departments as they seek higher pay elsewhere (a persistent issue in the high turnover across the civil service).***** The Institute has recommended that an expert pay review body should be established for the delegated grades, ⁵⁶² Worlidge J, Clyne R, Urban J and Thomas A, 20 ways to improve the civil service, Institute for Government, 31 July 2024, www.instituteforgovernment.org.uk/publication/ways-improve-civil-service  in part to help to rectify this.

There are also marked variations between departments in officials’ satisfaction with their pay. In 2024, for example – the latest available data****** – satisfaction with pay and benefits among core departments varied from 25% in the Treasury to 39% in DfE.*******

In some departments, satisfaction on this measure has shifted dramatically in the latest data. Between 2023 and 2024 there were sizeable increases in satisfaction in DESNZ (12ppts) and Defra (8ppts), as well as increases of around 5ppts in DCMS, DSIT, the Treasury and MHCLG. However, there were (more modest) falls in satisfaction with pay and benefits in some other departments. DWP (4ppts), MoJ (3ppts) and the Home Office (2ppts) all saw falls.

Of the five core departments where satisfaction with pay and benefits fell in 2024, three (DWP, HMRC and the Home Office) are the largest in the civil service, with high numbers of officials in the operational delivery profession.******** By contrast, the largest increases in satisfaction on this measure are seen in smaller, policy-focused departments. This is mirrored by satisfaction with pay and benefits across the whole civil service falling by 1.2ppts in the operational delivery profession while rising by 3.4ppts in the policy profession.

The ‘broken’ system for SCS pay is becoming an urgent problem

Between 2010 and 2023 the SCS suffered the most dramatic real-terms pay cut of any grade, falling by more than 26%. And despite recent increases, it remains 24% lower in real terms than in 2010.

The Institute for Government has previously noted the difficulties this poses for the civil service. ⁵⁶⁸ Clyne R, Savur S, Pope T and others, Whitehall Monitor 2023, Institute for Government, 30 January 2023, www.instituteforgovernment.org.uk/sites/default/files/2023-01/whitehall-monitor-2023.pdf  With less opportunity to ‘benefit’ from grade inflation – particularly for directors general and permanent secretaries – the pay available at these grades has become increasingly less attractive and more uncompetitive with both the private and wider public sectors. This is making it harder to recruit and retain talented individuals.

The Senior Salaries Review Body (SSRB), the independent body that makes recommendations to the government on SCS pay alongside that of other groups of senior public sector workers, has often highlighted the problem of inadequate senior civil service pay. In its 2025 report (making recommendations for the 2025/26 financial year) it did so in particularly direct terms, describing the SCS pay system as “broken” and noting its “considerable frustration” with the government’s failure to act. ⁵⁶⁹ Senior Salaries Review Body, Office for the Pay Review Bodies, Cabinet Office and others, Senior Salaries Review Body Report: 2025, independent report, GOV.UK, 22 May 2025, www.gov.uk/government/publications/senior-salaries-review-body-report-2025  It took the “unusual step” of making a formal recommendation for a “fundamental review and ‘reset’ of SCS pay and reward frameworks”. Without this review, the SSRB considers the specific recommendations it can make on pay as little more than sticking plasters.

The government accepted almost all of the SSRB’s recommendations for 2025/26, including its call for a review of SCS pay and reward.*********^{, 570} House of Commons, Hansard, ‘Civil Service Workforce’, 22 May 2025, col 40WS, https://hansard.parliament.uk/commons/2025-05-22/debates/25052261000015/CivilServiceWorkforce  There was limited further detail provided on the review in the government’s evidence to the SSRB for the 2026/27 pay round, published in October – the Cabinet Office said that, while work on the fundamental review was progressing “at pace”, ⁵⁷¹ Civil Service and Review Body on Senior Salaries, ‘Government evidence to the Senior Salaries Review Body on the pay of the Senior Civil Service (October 2025) (HTML)’, transparency data, GOV.UK, 30 October 2025, https://www.gov.uk/government/publications/government-evidence-to-the-senior-salaries-review-body-on-the-pay-of-the-senior-civil-service  it would form part of its overall strategy for the SCS, which the SSRB has also called for. It did, however, set some “immediate priorities” for SCS pay reform:

• Setting clear reward principles for the SCS
• Addressing the current pay band spans and overlaps
• Developing a pathway to pay progression.

The final priority – enabling pay progression within the same role based on performance – is particularly welcome, as is the government’s intention to look at expanding such a system to the wider civil service. Allowing officials to progress through a pay band within the same role, linked to performance rather than time served, would help to tackle the problems of both churn and grade inflation. Also welcome are some of the government’s indications as to what the “clear reward principles” should be – including supporting a “geographically dispersed” and “increasingly technical and specialist” workforce.

Much of this – including the overarching strategy for the SCS – is positive. But work on such a strategy began under the previous government, and progress has been disappointingly slow (even if it was, in October last year, “being finalised”). ⁵⁷² Ibid.  On pay reform specifically, the government is intending to start making changes as part of the 2026/27 pay award, with full reform of the SCS reward structure achieved over three years. It is disappointing that in the five months since the government accepted the need to urgently reform SCS pay, little more than a handful of principles – however encouraging – have been produced.

Civil service pay reform is a notoriously complex area. It has been promised and abandoned before. Ministers and civil service leaders (particularly the cabinet secretary, Chris Wormald, and the civil service chief operating officer, Cat Little) must accept that this is too important to get lost – or buried in an ever-expanding set of workforce strategies – as other pressures inevitably arise.

_{* Data in the first chart in this chapter captures pay on 31 March at the end of the financial year. This means that the figures given for 2024, for example, reflect the impact of the 2023/24 pay award, and those for 2025 the impact of the 2024/25 pay award.
** R² = 0.6862, p = 0.0001. See Methodology (regression 1) for full details.
*** See the chapter The size of the civil service.
**** Despite AA/AO being the most junior grade, in DfT the median AA/AO salary (£38,460) is higher than the median EO salary (the next most junior grade, at £33,900). This is because employees of the Government Car Service, who are contracted to work for longer hours than the norm and therefore have a higher annual salary, are included in the AA/AO grade.
***** See the chapter Turnover, leaving routes and exit schemes
****** This data is taken from the Civil Service People Survey, which is conducted in autumn each year and the results published the following year. The latest available data, published in 2025, captures civil service morale in autumn 2024.
******* Almost all Civil Service People Survey data we refer to relates to core departments. DfE is the exception – DfE figures are for the whole departmental group.
******** While the MoJ is also a very large department – indeed the largest – and has a very high proportion of staff in the operational delivery profession, both of these facts are a feature of its ALBs rather than the core department, which is what the People Survey analysis focuses on. For further information on the operational delivery profession see the chapter Professions and functions.
********* This included the substantive recommendation on SCS pay – for a 3.25% increase to base pay from April 2025. It asked for one recommendation (changes to pay band maxima for SCS bands 1 to 3) to be deferred in order to form part of the fundamental review of pay and reward.}

Morale and engagement

The civil service tracks the morale and views of its workforce through the annual Civil Service People Survey. The voluntary survey, which covers a range of topics from wellbeing to how departments handle change, is carried out between September and October each year and published the following January – the latest data is therefore a snapshot from the autumn of 2024, a few months after Labour took office.

Civil servants’ responses of course reflect far more than their feelings about ministerial leadership, but we note that those first few months of the new government saw a concerted effort by ministers to reset what had become an increasingly antagonistic relationship between politicians and officials.

The prime minister recorded a video of himself telling the civil service personally how much they were valued – “from the get go I want you to know that you have my confidence, my support and, importantly, my respect” ⁵⁷⁵ Starmer K, ‘A message from Prime Minister Keir Starmer to the Civil Service’, speech online, GOV.UK, 8 July 2024, www.gov.uk/government/speeches/a-message-from-prime-minister-keir-starmer-to-the-civil-service  – and new secretaries of state reiterated that message in town halls across departments. However, it should be noted that these survey responses were collected before Starmer’s December 2024 comments about some civil servants being too comfortable in the “tepid bath of managed decline”. ⁵⁷⁶ Starmer K, ‘PM speech on Plan for Change: 5 December 2024’, speech at Pinewood Studios, GOV.UK, 5 December 2024, www.gov.uk/government/speeches/pm-speech-on-plan-for-change-5-december-2024

Civil service morale ticked up again in 2024

The headline measure of the People Survey is the employee engagement index – it is a composite measure that captures civil servants’ feelings about how things are done in their organisation, and their pride in where they work.

This saw almost a decade of steady improvement from 2010, followed by three consecutive years of decline from 2021 to 2023. The results from 2024 saw morale start to increase – but only just, moving from 60.7 to 61.2%.

That headline increase has been reflected reasonably consistently across departments (due to data availability, all departmental analysis of the People Survey looks at core departments only, in contrast to the departmental groups looked at in most of the report).

Most departments have seen a small (less than 2ppts) increase in their employee engagement scores. But there are a few outliers. DHSC and DESNZ stand out as particularly strong improvers, with large increases of 5 and 7ppts respectively. And while morale in the Cabinet Office rose by only just over 2ppts, this came on top of a 4ppt rise in 2023, which itself turned around four consecutive years of falling scores.

Only four departments saw decreases in morale in 2024. Three, the FCDO, HM Revenue and Customs (HMRC) and Ministry of Defence (MoD), all saw falls of less than 1ppt.

DfT is the only department that seems to have suffered a larger (3ppt) drop in morale. Of all departments, between 2023 and 2024 DfT saw the largest falls (and by a large margin) in scores for a variety of questions related to change management and leadership. These included 13ppt falls between 2023 and 2024 for the questions “when changes are made in my organisation they are usually for the better”, and “I have the opportunity to contribute my views before decisions are made that affect me”. The department saw falls of 9–10ppts for “I think it is safe to challenge the way things are done in my organisation”, “I believe that change is managed well in my organisation”, and “I believe that senior managers in my organisation will take action on the results from this survey”.

Most departments saw scores in these departments rise, and those that did see falls were not of the magnitude of DfT’s – the second biggest falls ranged from 2–5ppts across those same questions. These responses are reflected in the fall of 9ppts in 2024 on the theme score for ‘leadership and managing change’ (more on which below).

It is not clear exactly what is driving this, but two pieces of context are worth noting. First, staff would have been aware of the department’s voluntary exit scheme, which aims to release around 300 staff.*^{, 579} Markson, T, ‘Nearly 300 DfT civil servants to leave under voluntary exit scheme’, Civil Service World, 23 April 2025, www.civilserviceworld.com/professions/article/dft-voluntary-redundancy-scheme-300-civil-servants-leave  DfT is not the only department in this position, however, so this is unlikely to be the full story.

Second is home working. DfT staff reported a 30ppt drop in 2024 respondents agreeing with the statement “I have a choice in deciding where I do my work”, its biggest reduction for any single question, and a far bigger fall than in any other department. In 2023 DfT staff were expected to spend at least 40% of their week in the office; in 2024 this rose to 60% for all departments. ⁵⁸⁰ Government People Group, Civil Service People Plan 2024-2027, GOV.UK, 10 January 2024, www.gov.uk/government/publications/civil-service-people-plan-2024-2027  Whether it was the edict, or the way it was enforced, the return to office mandate has been particularly poorly received in DfT.

Civil servants are least happy with pay and with leadership

The survey asks questions related to nine themes relating to civil servants’ work. With the exception of ‘my work’, which has seen a very small dip, all other theme scores have either held level or – as is the long-term trend for most scores – continued to increase in the past year. Most of these theme scores have consistently scored higher than the overall engagement index, but civil servants are comparatively less satisfied with learning and development, leadership and managing change, and pay and benefits.

Departmental scores for each theme differ considerably, and the biggest variation is in ‘leadership and managing change’. DfT saw this theme score fall by 9ppts, while satisfaction on the measure increased by 6ppts for both DHSC and DESNZ (whose engagement scores had the biggest increases in 2024). The latter two departments also outperformed all others in their gains on ‘organisational objectives and purpose’ – a possible sign that the NHS and clean energy missions have provided some early clarity of purpose.

People’s day-to-day work and the quality of leadership drive positive engagement

Satisfaction with some elements of civil servants’ working lives drives overall engagement levels more than others. Analysed by core department, the ‘my work’ (comprised of questions about challenge, pride, and interest that people feel in their work) and ‘leadership and managing change’ theme scores have strong positive correlations with overall employee engagement.** Theme scores for ‘pay and benefits’ and ‘learning and development’ are particularly uncorrelated with overall engagement.

While satisfaction with it has improved over time, civil servants’ views on the ability of senior civil servants to lead their teams and to manage change in particular has consistently been poor, and well below the employee engagement score. This is dragging down morale, but if reversed also has the potential – as departments such as DESNZ have shown this year – to boost it considerably.

Senior leaders should learn from where leadership and change management skills are being developed and deployed, and developing those skills should be prioritised as part of future talent planning – including in the forthcoming SCS strategy.

_{* Exit schemes are discussed further in the chapter Turnover, leaving routes and exit schemes.}

_{** This is true both when looking at 2024 figures in isolation and when carrying out a two-way fixed effects regression on 2010–24 figures. See Methodology (regression 2) for full details.}

Diversity

Over the past 20 years, and following the long-term prioritisation of increasing demographic diversity, the civil service has become more diverse and representative of the wider UK workforce. In its figures for sex, minority ethnicity, disability and sexual orientation, the civil service as a whole is now much closer to or has overtaken the UK benchmark of the economically active population.*

The Institute has previously made three main arguments for why it is important to increase diversity in the civil service: to improve talent and performance, bring forward new ideas and perspectives, and increase its trust and reputation by reflecting the society it serves. ⁵⁸³ Worlidge J, Grama T, Urban J and others, Whitehall Monitor 2025, Institute for Government, 16 January 2025, https://www.instituteforgovernment.org.uk/publication/whitehall-monitor-2025 ^{, 584} Bishop M, A crossroads for diversity and inclusion in the civil service: Assessing the 2022 D&I strategy, Institute for Government, 6 December 2022, https://www.instituteforgovernment.org.uk/publication/diversity-inclusion-civil-service

There is more age diversity in the civil service

Though the median age of the civil service has been stable at 44 over the past four years, it has been becoming more diverse in age. The combined under-30 and over-60 age range has made up an increasingly larger share of the overall workforce, rising from 18% in 2012 to 28% in 2025. Meanwhile, the share of 40–59 year olds making up the civil service has slowly decreased from 58% in 2010 to 48% in 2025.

But differences in approach to work between generations – from expectations of career paths to flexible working – is something the civil service will need to remain aware of and adapt to so that it can get the most from its workforce.

The representation of female and LGB+ staff has met or surpassed the population benchmark

Female civil servants have consistently made up a higher proportion of the civil service as a whole than they do in the economically active population. Since 2002, they have made up more than 50% of the entire civil service workforce – but this has been concentrated at more junior grades. In recent years, the proportion of female civil servants in the senior civil service has also become more representative of the economically active population.**

The representation of LGB+*** staff in the civil service has also continued its upward trajectory. Across every grade, in all years for which the ONS provides data, staff identifying as LGB+ have been above the population benchmark (5.6% in 2023, the most recently available data).

LGB+ staff have in the past been represented in the senior civil service (SCS) at a higher rate than in the civil service as a whole. This is the first year that the proportion of SCS and all civil service LGB+ staff has been the same, and the first time that the proportion of HEO/SEO and G7/G6 staff identifying as LGB+ has reached similar proportions to those in the SCS.

Representation of minority ethnic staff in the senior civil service is stalling

Overall representation in the civil service of minority ethnic staff has tended to track that of the UK’s economically active population. The share of minority ethnic civil servants increased from 16.6% in 2024 to 18.0% in 2025 – nearly double what it was in 2010 (9.2%). But the proportion of minority ethnic staff in senior civil service roles has stalled at around 10.5% since 2021, well below the growing population benchmark, and failing to keep track with improving representation in the civil service as a whole.

Representation of staff with disabilities continues to grow

In 2025, the representation of staff with disabilities met the benchmark for the first time, although this is partly because the benchmark was lower than in 2024. The proportion of staff with a disability has continued increasing both across the civil service as a whole and in the senior civil service. Representation in the senior civil service, however, has not kept pace with the rest of the civil service: the gap has slowly grown from 2022, when it was at 5ppts, to 6ppts in 2025. But the overall trajectory – in contrast to the proportion of senior civil service staff from minority ethnic backgrounds in the past years – is a positive one.

Staff from higher socio-economic backgrounds continue to dominate

While progress on protected characteristics has made the civil service as a whole more demographically diverse, it is still lagging on socio-economic background. Despite commitments from successive governments, and a prominent focus in the 2022–25 Diversity and Inclusion Strategy, ⁵⁸⁶ Cabinet Office, Civil Service Diversity and Inclusion Strategy: 2022 to 2025, 24 February 2022, www.gov.uk/government/publications/civil-service-diversity-and-inclusion-strategy-2022-to-2025  there appeared to be little to no improvement in 2024, or in previous years, on this metric.****

We can infer***** from the survey that officials from higher socio-economic backgrounds continue to dominate the civil service: they make up the majority of each grade, and at every increase in seniority the proportion of officials from higher socio-economic backgrounds grows. Over two thirds of the SCS came from a high socio-economic background in 2024 (71%).

The government does not have as clear a picture of the socio-economic background of its staff as it does on other diversity characteristics. Data for socio-economic background has been collected in the People Survey since 2019 ⁵⁹⁴ Kelly B, ‘Social mobility and the People Survey’, blog, Civil Service, 18 October 2019, https://civilservice.blog.gov.uk/2019/10/18/social-mobility-and-the-people-survey/  but, as the Social Mobility Commission has pointed out, the optional survey “is not intended as a workforce reporting mechanism”. ⁵⁹⁵ Social Mobility Commission, Action plan: How to improve socio-economic progression within the Civil Service, GOV.UK, 20 May 2021, https://www.gov.uk/government/publications/navigating-the-labyrinth/action-plan-how-to-improve-socio-economic-progression-within-the-civil-service  It has recommended that the Cabinet Office publishes data on socio-economic background in the annual workforce statistics, as it does for the other diversity characteristics. ⁵⁹⁶ Ibid  But this has not been implemented.

The government has sent signals that it is keen to improve social mobility in the civil service. In August 2025 it announced that applications to the summer internship programme would be restricted to applicants from lower socio-economic backgrounds. ⁵⁹⁷ Cabinet Office, ‘Internship Scheme To Get More Working Class Students Into Civil Service’, press release, GOV. UK, 1 August 2025, https://www.gov.uk/government/news/internship-scheme-to-get-more-working-class-students-into-civil-service  (The internship was previously restricted to undergraduates from under-represented groups, until that restriction was temporarily removed in 2023.)

The Institute for Government has written about how restricting eligibility again – this time only to students from lower socio-economic backgrounds – is a worthwhile experiment that could help demystify government and the civil service. ⁵⁹⁸ Dunlop H, Keenan H and Thomas A, ‘The socio-economic background of civil servants needs to change’, blog post, Institute for Government, 1 August 2025, https://www.instituteforgovernment.org.uk/comment/socio-economic-background-civil-servants-needs-change  But while it may open up some junior roles to people from lower socio-economic backgrounds, an internship scheme alone will not be enough to make a long-term difference to social mobility in the civil service. It is necessarily limited by its scale – taking just 200 people per year – and by its narrow focus on the fast stream to grade 7 pathway. Across all grades, more needs to be done to make recruitment accessible.

For a start, the Institute for Government has argued that the civil service ‘success profiles’, which outline, among other things, the ‘behaviours’ (such as Delivering at Pace, or Seeing the Big Picture) and ‘strengths’ (such as Decisive, Networker, or Visionary) of preferred candidates, should be replaced. ⁵⁹⁹ Urban J and Thomas A, Opening up: How to strengthen the civil service through external recruitment, Institute for Government, 1 December 2022, https://www.instituteforgovernment.org.uk/publication/civil-service-external-recruitment  Tangible change at all grades also requires solving the problems that exist around progression once people are part of the workforce, which is currently very poor.

The previous government did not adopt the action plan laid out by the Social Mobility Commission in 2021, which included recommendations to “demystify the policy profession” and make access to high-profile roles that fast-track progression, like working in private office, more equal. ⁶⁰⁰ Social Mobility Commission, Action plan: How to improve socio-economic progression within the Civil Service, GOV.UK, 20 May 2021, https://www.gov.uk/government/publications/navigating-the-labyrinth/action-plan-how-to-improve-socio-economic-progression-within-the-civil-service  This government should use the internship as a springboard to a more thorough plan, informed by the commission’s work, to address some of these deep-seated problems with social mobility in the civil service.

_{* All benchmarks in this chapter are taken from the UK economically active working age population.
** Since Whitehall Monitor 2025, we have moved from using static UK population benchmarks for socio-demographic characteristics (using the latest available data) to using moving yearly benchmarks that reflect changes in the socio-demographic characteristics of the UK population. See Methodology for full details on how the benchmark is calculated.
*** The Cabinet Office invites civil servants to record their sexual orientation as ‘heterosexual/straight’, ‘gay or lesbian’, ‘bisexual’ or ‘other’. Our use of the term LGB+ refers to staff who report belonging to one of the last three groups. The term ‘LGBT+’ is not used because this data refers only to sexual orientation. The civil service records data on sex separately to sexual orientation.
**** Figures for socio-economic background are taken from the Civil Service People Survey, for which the most recent data is 2024.
***** See Methodology for details of how socio-economic background data is extracted from the People Survey.}

Location

The geographical distribution of officials has long been uneven, and weighted towards London. To some extent, this is to be expected – large numbers of civil servants will always need to be in the capital to be close to ministers, parliament and other departments.

The latest data (from March 2025) shows that there are more than 107,000 officials based in London* – 21% of the whole civil service, and far more than any other region. Relocating officials out of London has now been a consistent focus of civil service reform for several years, across Conservative and Labour governments.

Some relocation efforts have been successful

The most recent prominent drive to increase the number of officials working outside the capital came under the broad banner of the Conservatives’ ‘Places for Growth’ programme. ⁶⁰⁶ Cabinet Office, ‘Places for Growth 2030’, collection, GOV.UK, 12 December 2023, www.gov.uk/government/collections/places-for-growth  The Labour government inherited this and, in May last year, announced a new set of targets and ambitions around civil service relocation. ⁶⁰⁷ Cabinet Office, ‘Thousands of civil service roles to be moved out of London in latest reform to the state’, press release, GOV.UK, 14 May 2025, www.gov.uk/government/news/thousands-of-civil-service-roles-moved-out-of-london-in-latest-reform-to-the-state

• To reduce the number of London-based civil servants from 95,000 (FTE) to 83,000
  by 2030
• An ambition for the fast stream to have 50% of placements offered outside London
  by 2030
• Reaffirming the Conservatives’ target for 50% of the senior civil service to be based
  outside of London by 2030.

This continued political support for the relocation agenda is welcome. Institute for Government research has previously shown that the relocation efforts of recent years have been positive for the civil service. Where such efforts are accompanied by strong leadership and attention to building meaningful campuses they have attracted different types of recruits, enabled different regional voices to influence policy making, and provided a limited economic boost to areas hosting new offices. ⁶⁰⁸ Worlidge J, Grama T, Urban J and others, Whitehall Monitor 2025, Institute for Government, 16 January 2025, https://www.instituteforgovernment.org.uk/publication/whitehall-monitor-2025  These conclusions have been supported by the government’s own research ⁶⁰⁹ Cabinet Office, ‘Places for Growth Formative Evaluation Report’, corporate report, GOV.UK, 31 October 2024, www.gov.uk/government/publications/places-for-growth-formative-evaluation-report  into the Places for Growth programme.

Relocation efforts have been successful on a numerical basis. Data published in 2025 showed that, as of Q2 2024, a total of 23,249 (FTE) civil service roles had been relocated** outside of London. This means that the target initially set by the Conservative government in 2020 ⁶¹⁰ HM Treasury, ‘Budget 2020’, policy paper, GOV.UK, 11 March 2020, www.gov.uk/government/publications/budget-2020-documents/budget-2020  – to relocate 22,000 FTE roles by 2027 (the target date having been brought forward from 2030) – was met years ahead of schedule.

The focus on relocation is evident in the proportional change in the number of civil servants in each region over time. In both the most recent period (2024 to 2025) and since the Conservative government first announced a relocation target in 2020, growth in the number of officials in London has been outstripped by growth in most other regions and nations.***

The overall regional distribution of officials has barely changed

Such figures, however, look very different when viewed over a longer time frame. This is because the slimming down of the civil service between 2009 and 2016, and subsequent post-2016 expansion, affected different areas of the UK in markedly different ways. As the civil service shrank, London saw a more limited decline in numbers of officials than any other region or nation. Then as it grew, London also saw a greater proportional increase than elsewhere – partly because of the demand for more policy-focused roles needed to respond to the UK’s departure from the EU, and during the pandemic.

Indeed, even though the civil service is now significantly larger than it was in 2010, four regions of England, as well as Scotland, still have fewer officials than they did then. London, meanwhile, has 24% more.

The relatively limited impact of recent relocations is also evident in the changing proportions of the civil service in each region. Despite the success in relocating tens of thousands of roles, between 2020 and 2025 the overall proportion of the civil service located in London fell by less than a percentage point. Only in the North West did the proportion of officials located there change by more than 1 percentage point in this time frame (1.1%).

It is against this backdrop that the government’s intention to reduce the number of London-based civil servants from 95,000 to 83,000 by 2030 ⁶¹² Cabinet Office, ‘Thousands of civil service roles to be moved out of London in latest reform to the state’, press release, GOV.UK, 14 May 2025, www.gov.uk/government/news/thousands-of-civil-service-roles-moved-out-of-london-in-latest-reform-to-the-state  (using FTE rather than headcount figures) must be considered. This is the first time that there has been a specific target to reduce the number of officials in the capital.**** But while this makes it unambiguous, it is also more of a blunt instrument than the previous government’s more loosely defined 22,000 relocations target. And it will be difficult to reverse a long-running trend: on the government’s preferred FTE basis, the number of officials in London increased from 98,000 in 2021 to almost 103,000 in 2025.*****

The government is not on track for half of all senior civil servants to be outside of London by 2030

Different grades in the civil service are distributed very differently across the country. The proportion of each grade in London increases with seniority, from 9% of AA/AOs to 65% of the senior civil service (SCS).******^,******* And the two most junior grades – AA/AOs and EOs – are the only ones where the largest concentration of the grade is not in London. In both cases, the largest proportions of the grade are based in the North West.********

Addressing the imbalance in the distribution of senior civil servants was a specific aim of the Places for Growth programme, and is crucial if government efforts to establish more regional hubs and campuses with ‘end-to-end careers’ (discussed later this chapter) are to succeed. But progress has been limited. In 2025, two thirds of the senior civil service (on a headcount basis) were based in London (65%) – significantly lower than the 73% in 2020, but unchanged on 2024.

This stall in progress means this government, like the last, is not on track to meet its 50% target. Using the Institute’s measure of the proportion of the SCS based in London (which differs from the government’s, but brings the government closer to its target), and assuming – generously********* – that the proportion continues to fall at the average annual rate that it did between 2020 and 2025, 56% of the senior civil service will be based in London by 2030.**********

How relocations are managed is more important than headline numbers

The government has also outlined how future moves will be handled and which locations it will focus on. In May 2025, for example, ministers announced plans for three ‘major’ new regional government campuses and confirmed the locations of the first two. ⁶²⁰ Ibid  Manchester will host a government digital and AI innovation campus, and Aberdeen an energy campus – it is good to see the government acting on our 2024 recommendation for the latter.

The locations are perhaps not surprising – both cities already host government offices,*********** and indeed Manchester has been the biggest single beneficiary of relocations since 2020/21. The cities were also two of the 13 locations to which the government said it would relocate officials to create cross-government ‘regional hubs’, ⁶²¹ House of Commons, Hansard, ‘Places for Growth 2030’, 14 May 2025, col 10WS, https://hansard.parliament.uk/commons/2025-05-14/debates/25051471000007/PlacesForGrowth2030  while closing 11 central London offices over the spending review period.

The move to establish new campuses is welcome, and appears to draw on the success of the Darlington Economic Campus (DEC). Previous work by the Institute has identified several factors that were crucial to the DEC’s success, and to relocation generally. ⁶²² Worlidge J, Grama T, Urban J and others, Whitehall Monitor 2025, Institute for Government, 16 January 2025, https://www.instituteforgovernment.org.uk/publication/whitehall-monitor-2025; Urban J, Pope T and Thomas A, Lessons from the Darlington Economic Campus for civil service relocation, Institute for Government, 9 June 2023, www.instituteforgovernment.org.uk/publication/darlington-civil-service-relocation  The new campuses appear likely to benefit from several of these – including co-location of officials from different departments, a focus on a clearly defined theme, a retained focus on national policy making (given the themes of the campuses), and senior political and civil service support.

It is also welcome that last year’s relocation announcements came with a focus on developing career paths outside of London – another of the factors we identified as important. Pat McFadden, then chancellor of the Duchy of Lancaster, said that the 13 regional hubs would have “end-to-end careers across the civil service”, ⁶²³ House of Commons, Hansard, ‘Places for Growth 2030’, 14 May 2025, col 10WS, https://hansard.parliament.uk/commons/2025-05-14/debates/25051471000007/PlacesForGrowth2030  highlighting the ambition to have 50% of both fast stream placements and SCS roles outside of London by 2030, a new apprenticeship programme to start in 2026 focused on Birmingham, Manchester and London, and an interchange programme between the civil service and local authorities. ⁶²⁴ Cabinet Office, ‘Thousands of civil service roles to be moved out of London in latest reform to the state’, press release, GOV.UK, 14 May 2025, www.gov.uk/government/news/thousands-of-civil-service-roles-moved-out-of-london-in-latest-reform-to-the-state  The latter is one example of the government trying to link relocation to its ideas for how the civil service can work better – it has also explicitly linked the three new campuses to the delivery of its ‘missions’ and placed an emphasis on more interdisciplinary working and collaboration with local communities. ⁶²⁵ Ibid

It is not clear what practical impact these vaguer aspirations will have, and the government also faces an uphill struggle on more tangible targets. Its new aim to reduce the number of officials in London will be far from straightforward, as will both SCS and fast stream relocation.************ There are other risks too – including of pursuing relocation in isolation, rather than as a complementary part of wider reforms to the workforce, which will be particularly challenging at a time when ministers are aiming to reduce the size of the civil service.

Ministers and senior officials must also resist the urge to simply move officials out of London by any means to reach the targets. Future relocations must be carefully planned; for example, by following the Institute for Government’s key lessons for successful relocations: ⁶²⁶ Worlidge J, Grama T, Urban J and others, Whitehall Monitor 2025, Institute for Government, 16 January 2025, https://www.instituteforgovernment.org.uk/publication/whitehall-monitor-2025

• Ministerial and senior civil service buy-in are essential to success, and for relocated offices to be successful they should host a critical mass of senior roles.
• There are large potential benefits to co-locating departments in a single office outside Whitehall, with the fresh culture and physical proximity of officials helping to break down traditional departmental barriers.
• A ‘themed campus’, where roles cluster around a single policy area, helps provide staff with development opportunities in a single location, channels their career paths in a way that encourages the development of subject-specific knowledge, and facilitates cross-departmental work.
• The labour market in the relevant location must meet the civil service’s needs, which – as demonstrated by the DEC – can be the case in well-connected towns.

_{* This figure refers to the headcount of civil servants based in London. Unless otherwise stated, in this chapter all such figures are headcount figures. While the government uses FTE figures in relocation reporting, the IfG uses headcount because we think the success or otherwise of relocation efforts is best understood by assessing the movement of individuals, regardless of their working pattern. 
** Under the government’s methodology, a ‘relocated’ role does not necessarily mean that a specific individual or role has moved from London to another region. See Methodology for the government’s full definition.
*** The Places for Growth programme uses its own methodology and data to assess progress on relocation. For example, the Cabinet Office has told us that not all London-based roles are in scope of the new relocation target referenced above. This methodology and data are not publicly available; for our analysis we make use of publicly available sources.
**** As outlined above and in Methodology, while the previous government had a target to relocate 22,000 roles outside of London, this did not necessarily require individuals to leave London. Even if it did, it wouldn’t necessarily have meant that the overall number of roles in London would have fallen.
***** The source of the government’s 98,000 FTE figure is not clear.
****** There are two different definitions of the senior civil service; see Methodology. The publicly available statistics used in this chapter refer to one definition, while the government maintains separate figures, which are not publicly available. This should be borne in mind throughout this chapter.
******* The Institute has adopted a method of measuring the proportion of each grade based in different nations and regions that differs from that used by the government. Unless otherwise stated, all figures given in this chapter use the Institute’s measure. Most notably, the Institute uses headcount figures while the government uses FTE. For full details of the Institute’s measure and how it differs from the government’s, see Methodology.
******** This is largely a result of the geographic spread of the operational delivery profession, which has large numbers of more junior officials. Of all AA/AOs based in the North West, 94% work in the operational delivery profession. For EOs in the North West, the figure is 82%.
********* It is unlikely that the proportion will continue to fall at this rate. The rate at which it declined fell in 2024, and it is likely that the SCS roles easiest to relocate have already been moved.
********** Extrapolating from the government’s figures rather than the Institute’s, 58% of the SCS will still be based in London in 2030.
*********** The second HQs of DSIT and DCMS are based in Manchester. DESNZ’s second HQ is in Aberdeen.
************ While there are no published figures showing the regional distribution of all new fast stream placements, the
scheme has long been heavily weighted towards London.}

Public bodies and appointments

Arm's length bodies

Arm’s length bodies (ALBs) are a subset of the wider category of public bodies, which are (at least partially) state-funded organisations that carry out a public or government function. ALBs consist of executive agencies, non-ministerial departments and non-departmental public bodies (NDPBs). Understanding ALBs is important as 57.7% of the government’s resource daily expenditure limit (RDEL), that is, day-to-day spending on resources, is channelled through them.*

Keir Starmer announced the abolition of NHS England (NHSE) – a vast NDPB once dubbed “the world’s biggest quango” ⁶³² Timmins N, ‘The World’s Biggest Quango’: The first five years of NHS England, Institute for Government, 22 May 2018, www.instituteforgovernment.org.uk/publication/nhs-worlds-biggest-quango  – in a speech in Hull in March 2025. And he used his speech to call out the use of public bodies more generally, remarking:

“Over a number of years, politicians chose to hide behind a vast array of quangos, arm’s length bodies, regulators, you name it. A sort of cottage industry of checkers and blockers, using taxpayer money to stop the government delivering on taxpayer priorities.” ⁶³³ Starmer K, ‘PM remarks on the fundamental reforms of the British state: 13 March 2025’, speech at Reckitt, GOV. UK, 13 March 2025, www.gov.uk/government/speeches/pm-remarks-on-the-fundamental-reform-of-the-british-state-13-march-2025

It is not the first time that public bodies have found themselves thrust into the spotlight when the government has raised issues of state reform – David Cameron’s famous ‘bonfire of the quangos’ was a central plank of his austerity plans – and Starmer’s targeting of NHS England is another prominent example. However, the ALB landscape is complex and varied, and simply reducing their number is far from a silver bullet for any reform plans.

The number of arm’s length bodies has remained flat

The total number of ALBs ⁶³⁴ Dalton G and Gill M, ‘Public bodies’, Institute for Government, 25 January 2022, www.instituteforgovernment.org.uk/explainer/public-bodies  has remained broadly steady since 2017. This contrasts with dramatic decreases seen during the early 2010s (in that ‘bonfire of the quangos’). In fact the total number of ALBs grew by just two in the year to March 2025 to hit 306, reflecting the establishment of two new bodies: GB Energy – Nuclear and the Infected Blood Compensation Authority. No abolitions were completed in that period.**

This year’s small increase came in the context of Labour’s manifesto pledges to set up several new bodies, many of which are still works in progress. ⁶³⁵ Chivukula S and Gill M, ‘Tracker: The government’s proposed new public bodies’, Institute for Government, 13 March 2025, www.instituteforgovernment.org.uk/explainer/public-bodies-tracker  These range from the Fair Work Agency to the Independent Football Regulator. The current number proposed, if it is not increased substantially over the parliament, would be commensurate with other recent governments: some public bodies have been set up each year to deliver on their priorities, even as they have abolished others. ⁶³⁶ Gill M and Bishop M, How to set up a public body, Institute for Government, 20 November 2024, p. 10, www.instituteforgovernment.org.uk/publication/how-to-set-up-a-public-body

Total government funding to ALBs fell

The total government funding allotted to ALBs in 2023/24 (the most recent year for which official data is available) ⁶³⁸ Cabinet Office, Public Bodies 2024, GOV.UK, 29 May 2025, /www.gov.uk/government/publications/public-bodies-2024  was £394.8bn, a 0.61% (or £2.4bn) decrease in real terms compared to 2022/23. Most of this funding cut was felt across NDPBs (excluding NHSE), whose collective funding fell by 7.6% in 2023/24, the first decrease for NDPBs since 2018/19. HMRC’s funding also fell, continuing the trend seen since its inclusion in our dataset. NHS funding, which is channelled through NHSE and included in its funding figures, has, by contrast, increased by £7.8bn in real terms. This is in line with previous years: the NHS in England has had real terms increases in funding every year since its establishment in 2012.

Government funding is heavily skewed towards a small number of ALBs, 10 of which received 93% of all funding in 2023/24. NHSE alone received 47% of the total (£186.8bn), making it by far the largest ALB in these terms. NHSE was 2.4 times larger than the next biggest ALB by funding, the now abolished Education and Skills Funding Agency (the figures for which similarly include the funding it provided to schools and colleges).

At the other end of the spectrum, there are 109 ALBs that do not receive any government funding at all. These are mostly advisory bodies established to provide independent advice to government on specific topics and often funded by levies on their sectors, such as the Advisory Committee on Animal Feedingstuffs.

Staff numbers have continued to increase across all categories of ALBs. The total number of full-time equivalent (FTE) staff in ALBs increased by 6,302 (1.6%) to 397,256 in the year to 31 March 2024. Of these, 246,224 are civil servants*** – ALB staff therefore constitute a sizeable part of the overall civil service.****

Public body reform may become more piecemeal over time

Public body reviews have been conducted systematically since the beginning of the ‘bonfire’ in 2010. The latest iteration of these reviews, the ‘public body review programme’, concluded in March 2025. Data provided to us by the Cabinet Office indicated that by the time it ended, the public body review programme had launched reviews of 104 bodies, of which 75 were completed (72%). These reviews resulted in claimed efficiency savings of £172m.

Around the same time as the end of the public body review programme, the government launched two initiatives aimed at reforming the landscape of ALBs. First, an action plan to cut regulatory burdens, including by axing some regulators. Second, Pat McFadden, then chancellor of the Duchy of Lancaster, wrote to departments shortly after the announcement that NHSE would be abolished, asking them to swiftly identify other public bodies that could potentially be closed. ⁶⁵¹ Gill M, ‘Not another bonfire of the quangos…’, blog, Institute for Government, 10 April 2025, www.instituteforgovernment.org.uk/comment/pat-mcfadden-not-another-bonfire-quangos

The government currently has no plans to reinstate an ongoing public bodies review programme after this one-off exercise. However, some departments are continuing to carry out reviews, and the Treasury has also committed to commission periodic performance reviews of regulators, with a particular focus on how they enable economic growth. ⁶⁵² HM Treasury, Regulatory Action Plan – Progress Update and Next Steps, GOV.UK, 21 October 2025, p. 19, https://www.gov.uk/government/publications/a-new-approach-to-ensure-regulators-and-regulation-support-growth

These regulatory reviews should learn lessons from the public bodies review programme, in particular that:

• departmental oversight should be reviewed as well as the bodies themselves
• identifying sector-wide opportunities for reform requires reviewing multiple related bodies at the same time
• independent recommendations may not be fully owned by those reviewed, so tracking adoption and progress against them is key.

The government should also set out how reviews of non-regulatory public bodies will be co-ordinated on an ongoing basis, and how adequate departmental oversight will be maintained.

The responses to McFadden’s letter have not yet led to many abolitions, but instead to reasonable proposals for consolidation in some key areas (for example, in water, coinciding with the Cunliffe review into the water sector regulatory system in England and Wales). ⁶⁵³ Gill M and Haile D, ‘Water reform: From Ofwat to what?’, blog, Institute for Government, www.instituteforgovernment.org.uk/article/comment/water-reform-ofwat-what  In the coming months the government plans to consider further changes thematically – grouping regulators for review by area – as we have previously recommended.

But as the government considers further abolitions, it should recall that institutional change is costly and distracting and should only be pursued where there is a clear benefit – as set out in the Institute’s previous research on how to abolish a public body. ⁶⁵⁴ Dalton G, Amos N and Gill M, How to abolish a public body: Ten lessons from previous restructures, Institute for Government, 8 March 2023, www.instituteforgovernment.org.uk/publication/abolish-public-bodies  This lesson was clearly not learned when the government made the rushed and ill thought out decision to abolish NHSE – ostensibly to bring decision making closer to ministers but which, more than nine months on, is still a major source of uncertainty within government and frustration among staff. ⁶⁵⁵ Hoddinott S, Performance Tracker 2025: NHS, Institute for Government, 14 November 2025, www.instituteforgovernment.org.uk/publication/performance-tracker-2025/nhs/overview

Finally, it is important to note the reduced role the Cabinet Office is currently playing in public bodies policy. We have separately flagged the risk of the Cabinet Office having little involvement in regulatory reform ⁶⁵⁶ Gill M, Regulation and growth: Will the government’s strategy deliver? Institute for Government, 13 November 2025, www.instituteforgovernment.org.uk/publication/regulatory-action-plan  – and across the public bodies landscape more broadly there are risks to the consistency and rigour with which bodies are set up, abolished and reviewed if the centre lacks the capacity to develop strategy, co-ordinate decision making, disseminate best practice and oversee compliance.

Public appointments

Public appointments refer to ministerial appointments, usually for a chair or non-executive director on the board of a public body or for a member of an advisory committee, and are regulated by the commissioner for public appointments. Getting appointments right is important for the smooth running and effective operation of ALBs.

An updated Governance Code on Public Appointments was published by the Cabinet Office in October 2025. ⁶⁵⁷ Cabinet Office, Governance Code on Public Appointments, GOV.UK, 30 October 2025, www.gov.uk/government/publications/governance-code-for-public-appointments  It makes several welcome changes to the previous code, which the Institute for Government has been advocating for since the publication of our 2022 report, Reforming public appointments. ⁶⁵⁸ Gill M and Dalton G, Reforming public appointments, Institute for Government, 18 August 2025, https://www.instituteforgovernment.org.uk/publication/reforming-public-appointments  Among these changes are a reduced requirement for ministerial involvement in the appointments process, and a firmer and more realistic target for concluding senior appointments within four months. There has also been important progress on tracking public appointments across Whitehall.

These changes will need careful implementation, but should both reduce delays to appointments and improve transparency. Delays are rife and continue to get worse, so change cannot be implemented soon enough. Our other long-standing recommendations remain, such as that the scope of regulation should be extended to all ministerial appointments by default (including, for example, appointments to boards of executive agencies, which are not currently regulated).

Delays to public appointments continue to be damaging

Delays to appointments can leave unfilled roles on boards, which both undermines their functioning and undermines public trust. They can also deter good future candidates from applying.

Performance against the three-month target for completion of appointments, as defined prior to the recent governance code update, dropped again in 2023/24 to 13%, down from 16% in 2022/23 and 25% in 2021/22. ⁶⁵⁹ Commissioner for Public Appointments, Annual Report 2023/24, OCPA, 2024, https://publicappointmentscommissioner.independent.gov.uk/wp-content/uploads/2025/01/CPA-Annual-Report-2023-24.pdf  Even the 2021/22 figure represents a significant deterioration in performance since the previous commissioner for public appointments identified delays as a serious problem in his 2019 thematic review (at which point the figure he presented was 53%, albeit based on a sample of appointments). ⁶⁶⁰ Commissioner for Public Appointments, Thematic Review: Concluding competitions within three months of the closing date, OCPA, July 2019, https://publicappointmentscommissioner.independent.gov.uk/wp-content/uploads/2019/07/Final-Thematic-Review-The-Three-month-aspiration.pdf  A recent data report from the Cabinet Office (which covers new appointments only) has the 2024/25 figure even lower, at 12%. ⁶⁶¹ Cabinet Office, Public Appointments Data Report 2024/25, GOV.UK, 2 December 2025, p. 7, www.gov.uk/government/publications/public-appointments-data-report-202425

This data, collected via the Cabinet Office’s Public Appointments Digital Service (PADS) since 2023, is much more detailed and of higher quality than previous appointments data. Usefully, information is now available on where in the appointments process delays are happening. For 2024/25, this was mainly after interview: the average number of days between interview and offer acceptance was 133, compared to 50 between sift and interview, and 29 between the close of job advert and sift. ⁶⁶² Ibid, p.10.

In exceptional circumstances, direct appointments can be made by ministers to positions regulated by the commissioner for public appointments without a competition being run. This is often done to fill positions that are left vacant due to delays in the usual process. The number of exceptional appointments therefore offers some indication of the impact of delays to public appointments, although exceptional appointments also happen for other reasons (for example, if a board member steps down unexpectedly).

The number of exceptional appointments rose for the third consecutive year in 2024/25 to 56, the highest since the pandemic. As of September 2025, there had already been 28 such appointments since the start of the financial year, which implies that the rate has not decreased so far. The Cabinet Office is considering what future data it might collect on the reasons why exceptional appointments are being made.

Welcome reforms to public appointments should reduce delays

A long-standing source of delays to the appointments process has been the number of points at which either a decision by a minister is required, or civil servants and special advisers need to check if a minister would like to take a decision at all. Whereas the previous code said ministers must be “involved at every stage”, ⁶⁷³ Cabinet Office, Governance Code on Public Appointments (old), GOV.UK, 8 February 2024, p. 3, https://publicappointmentscommissioner.independent.gov.uk/wp-content/uploads/2025/11/governance_code_on_public_appointments.pdf  the updated code instead identifies a minimum of three points at which a minister must make a decision (defining the role, agreeing the advisory assessment panel, and making the final decision). ⁶⁷⁴ Cabinet Office, Governance Code on Public Appointments, GOV.UK, 30 October 2025, p. 5, www.gov.uk/government/publications/governance-code-for-public-appointments  Ministers can still choose to be involved at other points, but they must confirm which decisions they wish to make at the beginning of the appointment process.

The revised governance code reinforces the fact that using PADS is mandatory for departments. This will enhance the consistency with which departments administer their appointments – which the commissioner’s review of the role of DCMS in appointing the chair of the Independent Football Regulator has shown is necessary. ⁶⁷⁵ Gill M and Howes D, ‘Public appointments are being streamlined. But will it work?’ blog, Institute for Government, 17 December 2025, www.instituteforgovernment.org.uk/comment/public-appointments-streamlined-will-it-work

The updated governance code also requires that the prime minister’s interest list – a list of appointments “made by or of interest to the prime minister” – is published, which has already been done. ⁶⁷⁶ Cabinet Office, Public Appointments made by or of interest to the Prime Minister, GOV.UK, 30 October 2025, www.gov.uk/government/publications/public-appointments-made-by-or-of-interest-to-the-prime-minister/public-appointments-made-by-or-of-interest-to-the…  This is important both for transparency and for the progress of appointments. Anecdotally, No.10 has been a significant cause of delay to major appointments in the past, but this part of the process has not been fully acknowledged.

The code also revises the previous “aim” that appointments conclude within three months of the role advertisement closing. This aim was rarely being met. Instead, there is now a firmer “expectation” that senior appointments (such as chairs and chair-equivalent roles) conclude within four months, and other board member appointments within three, alongside the end of the process being redefined as the date an appointment is accepted (rather than when it is publicly announced). ⁶⁷⁷ Cabinet Office, Governance Code on Public Appointments, GOV.UK, 30 October 2025, p. 12, www.gov.uk/government/publications/governance-code-for-public-appointments

Publication of performance data from PADS, the first iteration of which was published in December 2025, will enable the impact of the changes on delays to be tracked and remaining blockages to be identified. The Institute for Government has been advocating variations on all of the above changes since 2022, ⁶⁷⁸ Gill M and Dalton G, Reforming public appointments, Institute for Government, 18 August 2025, www.instituteforgovernment.org.uk/publication/reforming-public-appointments  so this progress is welcome.

Progress on diversity in public appointments is mixed

Over recent years, both the Cabinet Office and the Commissioner for Public Appointments have been keen to promote diversity in public appointments. The diversity of public appointees and re-appointees in 2024/25 improved in several areas, but still lags behind in others. ⁶⁷⁹ Commissioner for Public Appointments, Annual Report 2024/25, OCPA, December 2025, http://publicappointmentscommissioner.independent.gov.uk/wp-content/uploads/2025/12/2025-12-17-CPA-2024-25-Annual-Report-FINAL.pdf

The proportion of female appointees (52%) and appointees identifying as LGB+ (6.6%) both increased to exceed their respective population benchmarks for this year (48.9% and 5.6%). While the percentage of appointees living outside London or the South East ticked up slightly to 62%, it remains below the benchmark of 69.7%. The proportion of appointees declaring a disability dipped to 4% against a benchmark of 17.8%, suggesting the accessibility of the appointment process and board activities remains poor.

Data on socio-economic background was released for the first time in the Cabinet Office’s December report. While the sample size is small (just 18% of new appointees provided this information), it is a welcome development to see this data being recorded. Of the new appointees who declared socio-economic background, 27% were from a working-class background, with 83% having attended a non-fee-paying school. ⁶⁸⁰ Cabinet Office, Public Appointments Data Report 2024/25, GOV.UK, 2 December 2025, pp. 31–33, www.gov.uk/government/publications/public-appointments-data-report-202425  It will take some years to build up a useful time series, but this will help to assess progress in reducing some of the barriers we have identified for lower socio-economic groups both in applying for, and discharging, public appointments. ⁶⁸¹ Gill M and Dalton G, Reforming public appointments, Institute for Government, 18 August 2025, pp. 35–42, https://www.instituteforgovernment.org.uk/publication/reforming-public-appointments

As well as demographic diversity, data is also collected on political activity – that is, whether an appointee has been employed by or otherwise significantly supported a political party prior to applying for a role. Declared political activity is higher than in recent years, at 7%, but the data is poor. A long-standing limitation to the data is that the governance code’s definition of political activity is quite narrow – excluding, for instance, people who are party members but not significant donors, or who have held only minor office with a party. ⁶⁸² Cabinet Office, Governance Code on Public Appointments, GOV.UK, 30 October 2025, p. 10, https://www.gov.uk/government/publications/governance-code-for-public-appointments  We have not published this chart since 2023 as a result.

Due to a change in data collection method following introduction of PADS, political activity was not broken down by party in 2023/24 or 2024/25 but this should still be possible in future years. ⁶⁸⁶ Commissioner for Public Appointments, Annual Report 2023/24: Diversity, OCPA, 2024, p. 23, publicappointmentscommissioner.independent.gov.uk/wp-content/uploads/2025/01/CPA-Annual-Report-2023-24.pdf

In addition, while the previous government committed to publishing details of all direct ministerial appointments before the election ⁶⁸⁷ Cabinet Office, Strengthening Ethics and Integrity in Central Government, GOV.UK, 20 July 2023, www.gov.uk/government/publications/strengthening-ethics-and-integrity-in-central-government  – a move the Institute welcomed ⁶⁸⁸ Dalton G and Gill M, Public appointments in 2023: What has changed – and what still needs to? Institute for Government, 7 November 2023, www.instituteforgovernment.org.uk/publication/public-appointments-2023  – this has not yet been done. Such appointments are more open to being made on purely political grounds, and are not included in the above figures, so this disclosure is an important aspect of making public appointments more transparent (and should include all unregulated appointments made by ministers).

_{*The government’s total RDEL spend in 2023/24 was £498bn, while RDEL spending through ALBs was £298.7bn (both figures expressed in 2024/25 prices). See HM Treasury, ‘Public Spending Statistics: July 2024’, GOV.UK, 30 August 2024, www.gov.uk/government/statistics/public-spending-statistics-release-july-2024/public-spending-statistics-july-2024}

_{**Our ALBs analysis is based on Cabinet Office data. This year’s data release was accompanied by a new online dashboard, which has greatly improved both granularity and ease of use of the data (Cabinet Office, Public bodies 2024, GOV.UK, 29 May 2025, www.gov.uk/government/publications/public-bodies-2024).}

_{***Staff of executive agencies, non-ministerial departments and crown NDPBs are civil servants. Staff of other NDPBs are public servants.}

_{****See the chapter The size of the civil service.}

Digital transformation

In 2025, the Labour government – like many before it – set out its ambitions for ‘digital transformation’ to change how the state works. ⁶⁹⁵ HM Treasury, Spending Review 2025, GOV.UK, updated 30 June 2025, www.gov.uk/government/publications/spending-review-2025-document  The government said the state should work “more like a start-up” ⁶⁹⁶ McFadden P, ‘Reform of the state has to deliver for people’, speech at University College London, Stratford, 9 December 2024, www.gov.uk/government/speeches/reform-of-the-state-has-to-deliver-for-the-people  and be “using the power of tech to modernise the system”. ⁶⁹⁷ Ibid.  Tom Loosemore, who co-founded the Government Digital Service (GDS), defined digital transformation as “applying the culture, processes, business models and technologies of the internet era to respond to people’s raised expectations”. ⁶⁹⁸ Loosemore T, ‘Our definition of digital’, blog post, Public Digital, 28 June 2017, https://public.digital/about-pd/our-definition-of-digital

Transformation is more than digitisation, which replicates analogue processes in a digital format. Instead, it involves teams and processes fundamentally changing in response to the opportunities presented by digital technologies. The aim to overhaul how the state works with new digital technologies has been around since GDS’s establishment in 2011. Since that date, strategy after strategy has promised culture change across government that ‘puts people first’.

GDS has achieved some real successes on digital transformation in its 15-year lifetime, not least GOV.UK Pay and Notify, which have standardised how the public sector handles small financial transactions and sends messages. But while effective, individual programmes such as these will not transform the state. Achieving that vision is a long-term project that requires rethinking how processes and teams across the civil service work. This is the major task facing the relatively young Department for Science, Innovation and Technology (DSIT).

DSIT has correctly diagnosed the problems facing digital transformation

Established by Rishi Sunak in February 2023, DSIT began as a small, policy-focused department. One of the early acts of the new Labour government was to move GDS from the Cabinet Office to DSIT, where it would sit at the heart of a new ‘digital centre of government’. The Central Digital and Data Office (CDDO), the Incubator for AI (i.AI), and officials from the Responsible Technology Adoption Unit and the Geospatial data team were all brought into DSIT and merged under the single banner of GDS at the time. ⁶⁹⁹ Davinson J, ‘Same name, new ambitions’, blog post, Government Digital Service, 27 January 2025, https://gds.blog.gov.uk/2025/01/27/same-name-new-ambitions/  The aim was for DSIT to drive through the government’s transformation ambitions.

In Q1 2024, the core department had 2,275 FTE staff;* 44% of these were from the policy profession, and only 6% from the Government digital and data (GDD) profession. When GDS transferred from the Cabinet Office, it added around 930 employees, the majority of whom were likely to be from the GDD profession – a significant addition to a small department like DSIT.**, ⁷⁰⁰ Department for Science, Information and Technology and Cabinet Office, ‘DSIT bolstered to better serve the British public through science and technology’, press release, GOV.UK, 8 July 2024, www.gov.uk/government/news/dsit-bolstered-to-better-serve-the-british-public-through-science-and-technology

In establishing itself as the new digital centre of government, DSIT released a review (State of digital government) and a strategy (A blueprint for modern digital government) in January 2025, together setting out how the government intended to transform the state. It secured backing from the centre in this, receiving a generous settlement in the spending review later that year.

The State of digital government review was a convincing consolidation of the challenges that digital transformation efforts have faced in government. ⁷⁰⁹ Department for Science, Innovation and Technology and Government Digital Service, State of digital government review, Department for Science Innovation and Technology, 21 January 2025, www.gov.uk/government/publications/state-of-digital-government-review/  It named five ‘root causes’:

• leadership – leaders are not rewarded for prioritising digital agendas
• structure – the way organisations across the public sector plan and develop their services is fragmented and sometimes duplicative, making it harder for the state to benefit from digital at scale
• measurement – there is no consistency in how digital performance is measured
• talent – recruitment is challenging because working in the public sector is a comparatively uncompetitive career path for digital professionals
• funding – models in government are unsuitable for digital and data projects.

The accompanying strategy, A blueprint for modern digital government, is an ambitious response to some of the issues from the review. ⁷¹⁰ Department for Science, Innovation and Technology and Government Digital Service, A blueprint for modern digital government, Department for Science, Innovation and Technology, 21 January 2025, www.gov.uk/government/publications/a-blueprint-for-modern-digital-government/  It outlines five ‘kickstarter’ projects which include: creating a GOV.UK App and GOV.UK Wallet, piloting GOV.UK Chat, and an “AI accelerator upskilling programme”. These are valuable projects to prepare government for the future.

The government has committed to growing the GDD profession

To support the digital transformation agenda, the government has also committed to increase the size of the GDD profession. The prime minister announced a new apprenticeship scheme, TechTrack, ⁷¹¹ Prime Minister’s Office, 10 Downing Street and Department for Science, Innovation and Technology, ‘Prime Minister: I will reshape the state to deliver security for working people’, press release, GOV.UK, 12 March 2025, https://www.gov.uk/government/news/prime-minister-i-will-reshape-the-state-to-deliver-security-for-working-people  which aims to have 2,000 apprentices in departments by 2030. He also announced that 1 in 10 civil servants would work in tech and digital roles within five years. ⁷¹² Ibid.  This is an ambitious vision for the GDD profession, especially in a wider context of staff reductions.

The number of GDD professionals across government has increased by 152% since 2016. In 2024, GDS included information on chief digital and data roles in its profession capability framework. ⁷¹³ Government Digital Service, ‘Government Digital and Data Profession Capability Framework’, GOV.UK, updated 28 November 2025, https://ddat-capability-framework.service.gov.uk/  This has provided more visibility on career paths to the senior civil service, which is good for recruiting and retaining staff at all levels. But the profession has not yet met the 6% target set by the 2022–25 roadmap from the previous government – at 5.3% of the overall civil service workforce. ⁷¹⁴ Central Digital and Data Office, Roadmap for digital and data 2022 to 2025, 9 June 2022, updated 29 November 2023, www.gov.uk/government/publications/roadmap-for-digital-and-data-2022-to-2025  As pointed out in DSIT’s review, this is below benchmarks of 6% for other governments and 8–12% in “regulated private sector industries”. ⁷¹⁵ Department for Science, Innovation and Technology and Government Digital Service, State of digital government review, Department for Science Innovation and Technology, 21 January 2025, www.gov.uk/government/publications/state-of-digital-government-review/

Increasing the GDD profession will further build out departmental digital teams, which are well-developed – many create their own systems and guidance – to support DSIT’s work. ⁷¹⁶ Clyne R, Savur S, Pope T and others, Whitehall Monitor 2023, Institute for Government, 30 January 2023, www.instituteforgovernment.org.uk/sites/default/files/2023-01/whitehall-monitor-2023.pdf
 

DSIT needs to establish its leadership as a digital centre

Despite DSIT’s accurate diagnoses and an increasing digital capability in government, it is not clear that the department currently has the capability to escape the fate of previous reform efforts – promising culture change and failing to deliver – and to translate this ambitious strategy into real-world change. Initiatives in the blueprint such as a service transformation team to “look at whole public sector service transformation” will require investment from departments and officials across government. ⁷²³ Department for Science, Innovation and Technology and Government Digital Service, A blueprint for modern digital government, Department for Science, Innovation and Technology, 21 January 2025, www.gov.uk/government/publications/a-blueprint-for-modern-digital-government  That in turn requires DSIT to take on a co-ordinating role which it has not yet had to manage at this scale.

GDS itself has faced challenges since 2016. The UK was in seventh place on the 2024 UN E-Government Survey, which compares nations’ abilities to deliver digital public services every two years, ⁷²⁴ The United Nations E-Government Development Database, ‘United Kingdom of Great Britain and Northern Ireland’, web page (no date), retrieved 25 November 2025, https://publicadministration.un.org/egovkb/en-us/Data/Country-Information/id/182-United-Kingdom-of-Great-Britain-and-Northern-Ireland  having been top in 2016. In 2018, the government announced it would stop funding the GOV.UK Verify, which would have allowed citizens to prove their identity online. It soon became a high-profile failure, running up excess costs and facing an investigation by the NAO. ⁷²⁵ Comptroller and Auditor General, Investigation into Verify, Session 2017-2019, HC 1926, National Audit Office, 5 March 2019, www.nao.org.uk/report/investigation-into-verify  Leadership was disrupted, with seven heads in ten years, none of whom stayed in post for much longer than three years. In 2021, CDDO was created to take on a co-ordinating role for digital strategy and standards – it has now been brought back into GDS.

One of the ways to help consolidate DSIT’s position at the digital centre would be to establish stable digital leadership. But, much like the post-2016 experience of GDS, this has not been achieved. The interim government chief digital officer (GCDO) has left DSIT, and will not be replaced, ⁷²⁶ House of Commons Science, Innovation and Technology Committee, Oral evidence: Digital centre of Government (HC 790), 19 November 2025, Q296 and Q297, https://committees.parliament.uk/oralevidence/16745/pdf/  with the department’s permanent secretary taking on the role’s responsibilities. Meanwhile, a permanent government chief data officer was only appointed in November 2025, with the government chief technology officer stepping down shortly after; recruitment for a successor only offered a12-month fixed term contract. ⁷²⁷ Dunton J, ‘DSIT launches hunt for next government chief technology officer’, PublicTechnology.net, 29 October 2025, https://www.publictechnology.net/2025/10/29/education-and-skills/dsit-launches-hunt-for-next-government-chief-technology-officer/  And as of December 2025, DSIT’s directors general for artificial intelligence and digital, technology and infrastructure are both interim appointments. ⁷²⁸ Department for Science, Innovation and Technology, ‘Our management’, GOV.UK, (no date), retrieved 25 November 2025, www.gov.uk/government/organisations/department-for-science-innovation-and-technology

The blueprint also proposed changes to the role and status of the most senior digital professionals to bolster digital leadership across government. Executive committees in all public sector organisations would need to have a ‘digital leader’ by 2026. Among the core departments, only seven currently do.*** All boards would also be required to have a digital non-executive. The blueprint calls for the position of the GCDO to be raised to second permanent secretary level, and all chief digital information officers (CDIOs) across government would report into them.

But without a GCDO, and huge demands on the permanent secretary’s time, DSIT will need to be clear about how it will build sufficient digital representation at the highest levels of government and ensure DSIT itself has the capacity to lead and foster cross-government co-operation.

Digital transformation depends on better engagement between GDS, departments and professions

In its early years, GDS transformed how government delivered information and services by launching GOV.UK, which replaced individual departmental websites to form one single access point for citizens. The success of this transformation – and the subsequent creation of hundreds of new digital services – was attributed to the collaboration between the digital centre and departments. ⁷³¹ Bracken M, ‘It’s not about us, it’s about collaboration’, blog post, Government Digital Service, 10 September 2015, https://gds.blog.gov.uk/2015/09/10/its-not-about-us-its-about-collaboration/

GDS will need to rediscover that spirit. Its success is heavily reliant on collaboration across departments and professions – and being supported by senior leaders within departments who understand the purpose of digital transformation. Flagship projects like GOV.UK OneLogin**** – which recently had its deadline pushed back by three years and whose deliverability depends on “cross-government alignment and sustained engagement” – will fail without cross-government leadership. ⁷³² National Infrastructure and Service Transformation Authority, NISTA Annual Report 2024-2025, 11 August 2025, http://www.gov.uk/government/publications/nista-annual-report-2024-2025

Of the current senior civil service, only a small portion is part of the GDD profession, which makes up 6% of the senior civil service, compared to 26% who are in the policy profession. Digital transformation cannot therefore rely solely on GDD leadership; senior leaders from other professions must develop the skills to work effectively with GDS.
 

There has been no shortage of attempts to upskill the senior civil service. But DSIT’s review reported that only 20% say they consider themselves ‘digitally upskilled’ relative to the digital and data essentials framework. ⁷⁴¹ Department for Science, Innovation and Technology and Government Digital Service, State of digital government review, Department for Science Innovation and Technology, 21 January 2025, www.gov.uk/government/publications/state-of-digital-government-review/  This includes understanding concepts like cloud computing and working in agile.*****, ⁷⁴² Central Digital and Data Office, ‘Digital, Data and Technology essentials for senior civil servants’, GOV.UK, 31 May 2022, https://www.gov.uk/government/publications/digital-data-and-technology-essentials-for-senior-civil-servants/digital-data-and-technology-essentials-fo…  The previous roadmap for digital and data had set a target of improving the digital and data skills of 90% of senior civil servants. The Digital Excellence Programme, run by Apolitical, was created specifically to meet the target. ⁷⁴³ Apolitical, ‘Digital Excellent Programme: Are you ready to lead in the new digital age?’, Apolitical, (no date) retrieved 25 November 2025, https://apolitical.co/learning-hub/digital-excellence-programme-uk-gov/  By September 2024, departments had purchased nearly 2,500 licences but, as the civil service itself acknowledges, “not all departments have taken the opportunity to enable leaders to develop by facilitating access to the programme”. ⁷⁴⁴ Winkle O, ‘Upskilling Senior Civil Servants with the Digital Excellence programme’, blog post, Cabinet Office, 2 September 2024, https://moderncivilservice.blog.gov.uk/2024/09/02/upskilling-senior-civil-servants-with-the-digital-excellence-programme/

DSIT’s review also remarked on a “culture gap” between policy and digital professionals that needs to be bridged. ⁷⁴⁵ Department for Science, Innovation and Technology and Government Digital Service, State of digital government review, Department for Science Innovation and Technology, 21 January 2025, www.gov.uk/government/publications/state-of-digital-government-review/  The review noted that the gap was so large that “the formation of true cross-functional teams, founded on mutual respect, was rare”. ⁷⁴⁶ Ibid.  There have been some examples of the professions working in teams together to design policy with user-centred design methods ⁷⁴⁷ Collier J, ‘What has User Centred Policy Design achieved in the MOJ?’, blog post, Justice Digital, 27 February 2019, https://mojdigital.blog.gov.uk/2019/02/27/what-has-user-centred-policy-design-achieved-in-the-moj/  – but these remain isolated and struggle with the problem of who is accountable.

There are a few things that the professions could do to resolve this problem. Introducing secondments between the teams would increase knowledge of how the other works. In addition, digital teams should undertake training on the policy making process. In the longer term, introducing more intra- or inter-departmental joint teams to tackle specific policy problems, with shared accountability, could provide different perspectives on how to address policy objectives.

Transformation is not a short-term goal. The government needs to give DSIT the time to build digital understanding among non-digital leadership and support the department to be in the lead on the government’s digital transformation agenda.

A concerning recent example of the current lack of coherence is digital ID. In October 2025, the government announced that the Cabinet Office would be given responsibility for “policy development, legislation and strategic oversight” of the newly announced digital ID programme – with DSIT remaining responsible only for “the technical design, build and delivery”. ⁷⁴⁸ Cabinet Office and Department for Science, Innovation and Technology, ‘Machinery of government: digital ID’, 23 October 2025, www.gov.uk/government/news/machinery-of-government-digital-id  A single Cabinet Office minister, Josh Simons, has been given responsibility for the whole scheme, which is in principle a good sign, but it is hard to see a strong rationale for taking any responsibility for an explicitly digital scheme out of the self-proclaimed digital centre of government.

_{*This figure is taken from the Cabinet Office, ‘Civil Service Statistics’ to maintain consistency between the reported profession sizes within the department.}

_{**Data taken from the ONS's Quarterly Public Sector Employment Release, because the 2025 'Civil Service Statistics' release does not reflect the GDS move to DSIT.}

_{***The blueprint defines a “digital leader” as a “Chief Digital Information Officer (CDIO), Chief Technology Officer (CTO), Chief Data Officer (CDO) or leader in service transformation, product or customer experience with deep digital expertise”.}

_{****GOV.UK OneLogin is the successor to the failed GOV.UK Verify. When it is implemented, citizens will be able to use the same email and password to access government services.}

_{*****‘Agile’ is a form of project management used to build and run digital services in government.}
 

Artificial intelligence and data

The government has big ambitions for savings from AI

In its broadest definition artificial intelligence (AI) refers to technology able to carry out tasks usually associated with humans. The government sees AI as one of the key ways to unlock productivity in the public sector and the civil service. At the launch of the AI Opportunities Action Plan in January 2025, the prime minister called it a “force that will turbocharge every single element of our Plan for Change”. ⁷⁸⁶ Prime Minister’s Office, 10 Downing Street, ‘PM speech on AI Opportunities Action Plan: 13 January 2025’, speech at University College London East, 13 January 2025, www.gov.uk/government/speeches/pm-speech-on-ai-opportunities-action-plan-13-january-2025  The State of digital government review estimated that £45bn of savings could be achieved, mainly from “process simplification” and “AI-driven automation of manual tasks”. ⁷⁸⁷ Department for Science, Innovation and Technology and Government Digital Service, State of digital government review, Department for Science, Innovation and Technology, 21 January 2025, www.gov.uk/government/publications/state-of-digital-government-review

However, the basis for the figure, and the grounds for the government’s level of ambition, is ambiguous at best, and many commentators are highly sceptical. ⁷⁸⁸ Wilkes G, ‘The hunt for the £45bn’, blog post, Freethinking Economist, 30 April 2025, https://freethinkecon.wordpress.com/2025/04/30/the-hunt-for-the-45bn/  It does, however, highlight the extent to which government is pinning its hopes on AI.

There is political and financial backing too. The chancellor announced £2bn at the spending review to support the AI action plan. ⁷⁸⁹ HM Treasury, ‘Spending Review 2025 speech’, GOV.UK, 11 June 2025, www.gov.uk/government/speeches/spending-review-2025-speech  The prime minister has appointed a new AI adviser to help “unlock the benefits”. ⁷⁹⁰ Department for Science, Innovation and Technology and Prime Minister’s Office, 10 Downing Street, ‘Appointment of Jade Leung as the Prime Minister’s AI adviser’, press release, GOV.UK, 15 August 2025, https://www.gov.uk/government/news/appointment-of-jade-leung-as-the-prime-ministers-ai-adviser  She reports directly to him and the secretary of state for science, innovation and technology. Starmer also sponsors an AI exemplars programme, an initiative for government organisations to experiment with AI technologies. ⁷⁹¹ King S, ‘The Prime Minister’s AI Exemplars programme: A portfolio approach to public sector innovation’, blog post, Government Digital Service, 18 August 2025, https://gds.blog.gov.uk/2025/08/18/the-prime-ministers-ai-exemplars-programme-a-portfolio-approach-to-public-sector-innovation/

Experimentation abounds, but there is no centralised way to understand and share lessons

The AI Opportunities Action Plan formalised the government’s desire to innovate further with AI. The plan sets out a public sector ‘Scan > Pilot > Scale’ approach. ⁷⁹² Department for Science, Innovation and Technology, AI Opportunities Action Plan, Department for Science, Innovation and Technology, 13 January 2025, www.gov.uk/government/publications/ai-opportunities-action-plan/ai-opportunities-action-plan  This involves developing prototypes or pilots in-house or through procurement before evaluating and publishing the results and scaling successful pilots. Currently, pilots are being driven by a mixture of individual departments and the ‘digital centre of government’ (DSIT, home to the Government Digital Service, GDS). ⁷⁹³ Rhodes C, AI in UK government departments, House of Commons Library, 4 April 2025, https://researchbriefings.files.parliament.uk/documents/CBP-10236/CBP-10236.pdf , ⁷⁹⁴ Incubator for Artificial Intelligence, ‘Projects’, web page (no date), retrieved 26 November 2025, https://ai.gov.uk/projects/  There has been one cross-government pilot of Microsoft Copilot organised by GDS. ⁷⁹⁵ Government Digital Service, Microsoft 365 Copilot Experiment: Cross-Government Findings Report, Government Digital Service, 2 June 2025, www.gov.uk/government/publications/microsoft-365-copilot-experiment-cross-government-findings-report

These many signs of AI use are an encouraging display of departments’ willingness to trial new technology, but make for a confusing landscape. GDS (itself home to the Incubator for Artificial Intelligence, or i.AI) has tried to steer how departments run these pilots. It has created the AI Playbook for UK Government, which contains 12 principles for using AI responsibly. ⁷⁹⁶ Government Digital Service, Artificial Intelligence Playbook for the UK Government, Government Digital Service, 10 February 2025, www.gov.uk/government/publications/ai-playbook-for-the-uk-government/  i.AI has also launched a ‘Knowledge Hub’, including a prompt library, guidance and a use case library, for teams to start from when looking to use AI, ⁷⁹⁷ Incubator for Artificial Intelligence, ‘Explore AI in government’, web page (no date), retrieved 26 November 2025, https://ai.gov.uk/knowledge-hub/  while more broadly DSIT supports teams to identify and mitigate risks associated with their AI use cases through workshops. ⁷⁹⁸ Department for Science, Innovation and Technology, The Model for Responsible Innovation, Department for Science, Innovation and Technology, 14 November 2024, www.gov.uk/government/publications/the-model-for-responsible-innovation

But GDS appears to be struggling to establish cross-government leadership and so is missing opportunities to share lessons. Government reporting on the use of algorithmic tools (of which AI is one) through the mandatory algorithmic transparency reporting standard (ATRS) has improved, but even GDS has not published information about all the AI tools it has trialled. ⁷⁹⁹ Cabinet Office, Department for Science, Innovation and Technology, and Government Digital Service, ‘Find out how algorithmic tools are used in public organisations’, web page, GOV.UK, retrieved 28 November 2025, www.gov.uk/algorithmic-transparency-records  DSIT said it was not possible to manage the Microsoft Copilot trial centrally since it took “substantial effort” to drive adoption. ⁸⁰⁰ Government Digital Service, Microsoft 365 Copilot Experiment: Cross-Government Findings Report, Government Digital Service, 2 June 2025, www.gov.uk/government/publications/microsoft-365-copilot-experiment-cross-government-findings-report

GDS does not need to know everything about what officials in individual departments are doing, and nor should it assert control without good reason – departments should be able to pilot AI independently. But DSIT (as the digital centre of government) should have some oversight and collect lessons learned to avoid high-profile failures, or at least to learn from them.

Indeed, the most well-documented and successful scaling of AI in government remains one that was done without GDS – the Government Communications Service’s (GCS) roll-out of GCS Assist, a tool that produces first drafts of communications products. As of May 2025, GCS reported that there was a 70% adoption rate of the tool across all eligible users. ⁸⁰¹ Cabinet Office, The People Factor: A human-centred approach to scaling AI tools, Cabinet Office, 4 June 2025, www.gov.uk/government/publications/a-human-centred-approach-to-scaling-and-de-risking-ai-tools/the-people-factor-a-human-centred-approach-to-scaling-…  Following the roll-out, the team behind Assist published two reports about what they had learnt. ⁸⁰² Ibid. , ⁸⁰³ Cabinet Office, The Mitigating ‘Hidden’ AI Risks Toolkit, Cabinet Office, 4 June 2025, www.gov.uk/government/publications/a-human-centred-approach-to-scaling-and-de-risking-ai-tools/the-mitigating-hidden-ai-risks-toolkit-html  These show an organisation that has begun to properly grapple with the challenges and opportunities of rolling out AI.

GCS’s roll-out of Assist and its transparency over its process and experiences (as well as the mandatory training that accompanied it for new users) is a template for the successful scaling of AI – and won it the AI award in the government and public sector category at the National AI Awards. ⁸⁰⁴ Government Communication Service, ‘Assist, the AI tool built for government communicators’, blog post, 14 October 2025, updated 10 November 2025, www.communications.gov.uk/guidance/assist-the-ai-tool-built-for-government-communicators/ , ⁸⁰⁵ Civil Service, ‘Government Communications scoop National AI Award’, blog post, 23 October 2025, https://civilservice.blog.gov.uk/2025/10/23/government-communications-scoop-national-ai-award/  The government should make efforts to enable other teams and departments to learn from it. ⁸⁰⁶ Ayar-Majeed S and Gill M, ‘Whitehall and AI: how can government move from promising pilots to real results?’, blog post, Institute for Government, 2 September 2024, www.instituteforgovernment.org.uk/comment/whitehall-ai-government-pilots  Better leadership from DSIT and GDS here will be key.

Productivity gains are more complex than time savings alone

The government hopes that implementing AI in the civil service will significantly improve its productivity. ⁸⁰⁷ Department for Science, Innovation and Technology, ‘Landmark government trial shows AI could save civil servants nearly 2 weeks a year’, press release, 2 June 2025, www.gov.uk/government/news/landmark-government-trial-shows-ai-could-save-civil-servants-nearly-2-weeks-a-year  AI is mentioned in almost three quarters of departmental efficiency plans* and some departments explicitly mention AI efficiencies. For example, the Treasury plans to find £6m in efficiencies from AI and automation by 2028/29. ⁸⁰⁸ HM Treasury, Departmental Efficiency Plans, GOV.UK, 11 June 2025, www.gov.uk/government/publications/departmental-efficiency-delivery-plans

In addition to the open approach on GCS Assist there have been a few welcome departmental reports on AI pilots: GDS published a cross-government findings report on the use of Microsoft Copilot (although it is unclear how well this is translating into lessons learned), ⁸⁰⁹ Government Digital Service, Microsoft 365 Copilot Experiment: Cross-Government Findings Report, Government Digital Service, 2 June 2025, www.gov.uk/government/publications/microsoft-365-copilot-experiment-cross-government-findings-report  and DBT published its own report on it. ⁸¹⁰ Department for Business and Trade, The Evaluation of the M365 Copilot Pilot in the Department for Business and Trade, Department for Business and Trade, 28 August 2025, https://assets.publishing.service.gov.uk/media/68adbe409e1cebdd2c96a19d/dbt-microsoft-365-copilot-evaluation.pdf  The Home Office has evaluated its asylum case summarisation and asylum policy search tools, ⁸¹¹ Home Office, Evaluation of AI trials in the asylum decision making process, Home Office, 29 April 2025, www.gov.uk/government/publications/evaluation-of-ai-trials-in-the-asylum-decision-making-process/evaluation-of-ai-trials-in-the-asylum-decision-makin…  and DSIT reported on an AI coding assistant trial. ⁸¹² Department for Science, Innovation and Technology and Government Digital Service, AI coding assistant trial: UK public sector findings report, Department for Science, Innovation and Technology, 12 September 2025, www.gov.uk/government/publications/ai-coding-assistant-trial/ai-coding-assistant-trial-uk-public-sector-findings-report#results-and-findings  Looking at these four evaluations reveals a mixed picture of potential productivity gains.

All trials reported time savings. Almost three quarters of users reported that Copilot made them more productive (72%) – for 10 of 15 professions in the report, ‘improves productivity’ scored the highest of a list of most common benefits. ⁸¹³ Government Digital Service, Microsoft 365 Copilot Experiment: Cross-Government Findings Report, Government Digital Service, 2 June 2025, www.gov.uk/government/publications/microsoft-365-copilot-experiment-cross-government-findings-report  A sizeable minority of users in DSIT’s AI coding assistant trial reported that they would be less productive if they could no longer use the tool (39%). ⁸¹⁴ Department for Science, Innovation and Technology and Government Digital Service, AI coding assistant trial: UK public sector findings report, Department for Science, Innovation and Technology, 12 September 2025, www.gov.uk/government/publications/ai-coding-assistant-trial/ai-coding-assistant-trial-uk-public-sector-findings-report#results-and-findings

These figures suggest an optimistic picture for government’s hopes for productivity. But departments need to be especially wary of self-reported time savings, which the Copilot and the coding assistant trials were based on – one (non-government) study with experienced developers showed they took longer with AI, but respondents believed they had performed the task quicker. ⁸¹⁵ Becker J, Rush N, Barnes B and Rein D, Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity, Model Evaluation and Threat Research, 10 July 2025, https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/

More importantly, even when AI is creating time savings, the DBT report notes that they “did not find robust evidence to suggest that time savings are leading to improved productivity”. ⁸¹⁶ Department for Business and Trade, The Evaluation of the M365 Copilot Pilot in the Department for Business and Trade, Department for Business and Trade, 28 August 2025, https://assets.publishing.service.gov.uk/media/68adbe409e1cebdd2c96a19d/dbt-microsoft-365-copilot-evaluation.pdf  The government must not treat AI as a productivity panacea, but recognise that it is a tool that is being deployed into an organisational system in which there are pre-existing problems. GCS has some helpful lessons to share here. Its report on hidden risks looks at AI roll-out from a systems perspective, exploring how to mitigate the unintended consequences that occur when people, teams or imperfect systems interact with AI. ⁸¹⁷ Cabinet Office, The Mitigating ‘Hidden’ AI Risks Toolkit, Cabinet Office, 4 June 2025, www.gov.uk/government/publications/a-human-centred-approach-to-scaling-and-de-risking-ai-tools/the-mitigating-hidden-ai-risks-toolkit-html

The government must not further lock in reliance on third-party suppliers

Poor quality data and legacy IT – where departments rely on outdated IT systems – are two fundamental barriers to the effective adoption of AI in the public sector. As of March 2025, around 80 legacy IT systems that the government were still actively using posed a critical level of cybersecurity risk. ⁸¹⁸ Trendall S, ‘Legacy IT reviews have found almost 80 government systems at ‘critical’ risk level’, PublicTechnology.net, 13 March 2025, www.publictechnology.net/2025/03/13/defence-and-security/legacy-it-reviews-have-found-almost-80-government-systems-at-critical-risk-level/  The Public Accounts Committee wrote that the cyber risk to government was now “extremely high”. ⁸¹⁹ House of Commons Committee of Public Accounts, Government cyber resilience (HC 643), 9 May 2025, https://publications.parliament.uk/pa/cm5901/cmselect/cmpubacc/643/report.html  A 2024 NAO report also found that 62% of organisations responding to its survey struggled to implement AI because of lack of access to good quality data. ⁸²⁰ Comptroller and Auditor General, Use of Artificial Intelligence in Government, Session 2023-24, HC 612, National Audit Office, 15 March 2024, www.nao.org.uk/reports/use-of-artificial-intelligence-in-government  DSIT’s A blueprint for digital government suggested that DSIT would create “tailored funding models for legacy remediation and risk reduction” and encourage better data sharing. ⁸²¹ Department for Science, Innovation and Technology and Government Digital Service, A blueprint for modern digital government, Department for Science, Innovation and Technology, 21 January 2025, www.gov.uk/government/publications/a-blueprint-for-modern-digital-government  But it will take time to fix these problems.

AI requires high-quality data to work effectively, but lots of government data is often ‘locked away’ in legacy systems. ⁸²² House of Commons Committee of Public Accounts, Use of AI in Government (HC 356), 26 March 2025, https://publications.parliament.uk/pa/cm5901/cmselect/cmpubacc/356/report.html  Much was procured at a time when government outsourced most of its IT development. As time passes, the systems have become more outdated, requiring rarer and more specialised skills to maintain them – and increasing their vulnerability. This has resulted in reliance on specialised suppliers who are the only ones who can manage the – increasingly costly – system. In its rush to roll out AI, the government may be missing the impetus to free up and improve its data, and instead risks rolling out AI on to poor data held together by third-party suppliers, exacerbating government reliance on them at a time when government should be minimising it.

The road to an AI-enabled British state is not straightforward. Taking advantage of AI will require tackling deep-seated problems like legacy data and anticipating emerging ones. New technologies provide real opportunities, but they do not solve pre-existing problems alone: this requires the familiar work of focusing on improving organisational systems and processes to make government function more effectively.

_{*See the chapter Departmental spending and efficiencies.}

Consultancy and temporary staff spending

Government departments often bring in external support to supplement their permanent workforce. Consultants may be hired to assist in digital transformation efforts, for example, while temporary staff* are frequently deployed to assist with backlogs or peaks in workload, such as in processing migrant casework or passport applications.

As we have argued before, ⁸³⁸ Worlidge J, Grama T, Urban J and others, Whitehall Monitor 2025, Institute for Government, 16 January 2025, www.instituteforgovernment.org.uk/sites/default/files/2025-09/whitehall-monitor-2025.pdf  the use of such services is sometimes justified – where consultants are used to bring an external perspective or expertise that the civil service cannot reasonably be expected to have in-house, or temporary workers are used to plug unexpected resourcing gaps. All too often, though, departments have displayed a reliance on such provision, and failed to build up necessary capacity internally, leading to wasteful spending. ⁸³⁹ Guerin B, Thomas A, Clyne R and Vira S, Finding the right skills for the civil service, Institute for Government, 16 April 2021, www.instituteforgovernment.org.uk/publication/report/finding-right-skills-civil-service

The government is targeting reduced consultancy spending

Consultancy spending has attracted particular attention in recent years, mostly when it has spiked during times of challenge for the civil service, like Brexit ⁸⁴⁰ Comptroller and Auditor General, Departments’ use of consultants to support preparations for EU Exit, Session 2017–19, HC 2105, National Audit Office, 7 June 2019, www.nao.org.uk/wp-content/uploads/2019/05/Departments-use-of-consultants-to-support-preparations-for-EU-Exit.pdf  and the pandemic – in 2021, for example, it was reported that the Covid Test and Trace team had paid £1m a day to Deloitte consultants. ⁸⁴¹ Consultancy.uk, ‘Test and Trace spends £1 million per day on Deloitte consultants’, Consultancy.uk, 2 July 2021, www.consultancy.uk/news/28374/test-and-trace-spends-1-million-per-day-on-deloitte-consultants

There have been efforts to reduce departments’ spending on consultancy. A renewed drive to do so was a prominent early announcement of the Labour government: as part of the “immediate savings” ⁸⁴² HM Treasury, ‘Fixing the foundations: public spending audit 2024 -25’, policy paper, GOV.UK, 29 July 2024, www.gov.uk/government/publications/fixing-the-foundations-public-spending-audit-2024-25  identified in the July 2024 Public Spending Audit, the Treasury said that stopping all non-essential government consultancy spend in 2024/25 would save £550m in-year. Halving total consultancy spending in future years would, the government said, save £680m in 2025/26. ⁸⁴³ Ibid.  The spending review then updated these estimated savings to more than £700m per year by 2028/29, which it said would reduce spending by 50% compared to the “average spending on consultants from 2017–18 to 2022–23”. ⁸⁴⁴ HM Treasury, ‘Spending Review 2025 document’, policy paper, GOV.UK, 11 June 2025, www.gov.uk/government/publications/spending-review-2025-document  The government has said ⁸⁴⁵ Cabinet Office, ‘New controls across government to curb consultancy spend and save over £1.2 billion by 2026’, press release, GOV.UK, 14 November 2024, www.gov.uk/government/news/new-controls-across-government-to-curb-consultancy-spend-and-save-over-12-billion-by-2026  it intends to achieve these savings through a new consultancy spending control** (in addition to the existing wider requirement for departments to gain central approval of any spending over certain limits) and a new Crown Commercial Service framework agreement, although this is optional for departments to use and the NAO has found that most spending on consultancy does not go through it.***, ⁸⁴⁶ Comptroller and Auditor General, Government’s use of external consultants, Session 2024-26, HC 1381, National Audit Office, 21 November 2025, www.nao.org.uk/insights/governments-use-of-external-consultants

Consultancy spending is very difficult to track. As we noted in Whitehall Monitor last year, the data has long been poor, beset by confused and differing definitions and a lack of consistent data collection methods. ⁸⁴⁷ Worlidge J, Grama T, Urban J and others, Whitehall Monitor 2025, Institute for Government, 16 January 2025, www.instituteforgovernment.org.uk/sites/default/files/2025-09/whitehall-monitor-2025.pdf  The NAO recently summarised these challenges, noting that departments’ annual reports and accounts (ARAs), various internal government systems, and private sector analysis of publicly released invoices all capture different (though overlapping) items of consultancy spending – some of which are bundled together with spending on other professional services – and so give very different overall figures. ⁸⁴⁸ Comptroller and Auditor General, Government’s use of external consultants, Session 2024-26, HC 1381, National Audit Office, 21 November 2025, www.nao.org.uk/insights/governments-use-of-external-consultants

Our approach is to use departments’ ARAs to collate data on consultancy spending (as well as on temporary staff spend). This is because they provide the only consistently published source of data on such spending across both departments and time. The government also recently confirmed to the Public Accounts Committee (PAC) that this is the measure it uses internally. ⁸⁴⁹ House of Commons, Public Accounts Committee oral evidence, ‘Government’s use of external consultants’, 15 December 2025, https://committees.parliament.uk/event/25033/formal-meeting-oral-evidence-session  But as the NAO has cautioned, ⁸⁵⁰ Comptroller and Auditor General, Government’s use of external consultants, Session 2024-26, HC 1381, National Audit Office, 21 November 2025, www.nao.org.uk/insights/governments-use-of-external-consultants  even ARAs do not necessarily provide a reliable estimate of such spending, because different departments use different definitions of consultancy, which may differ from formal cross-government definitions.**** In their recent evidence to PAC, Cat Little and others suggested that, rather than departments actively using different definitions, the challenge lies with some having difficulty classifying spending correctly against the agreed cross-government definition. The Cabinet Office also confirmed it was intensifying work to address this issue. ⁸⁵¹ House of Commons, Public Accounts Committee oral evidence, ‘Government’s use of external consultants’, 15 December 2025, https://committees.parliament.uk/event/25033/formal-meeting-oral-evidence-session

In any case, the government has said that it is using the average annual total government consultancy spending between 2017/18 and 2022/23 as its baseline, and that a 50% reduction will save £700m. ⁸⁵² HM Treasury, ‘Spending Review 2025 document’, policy paper, GOV.UK, 11 June 2025, www.gov.uk/government/publications/spending-review-2025-document  Our analysis of departments’ ARAs is consistent with this, showing that average annual consultancy spend between 2017/18 and 2022/23 was £1.4bn. It is notable that the government has chosen a baseline that includes the years of the pandemic, when consultancy spending was higher than average – which should make it easier to reach the target.

Overall consultancy spending is falling

Because the government has defined its target for consultancy spending with reference to the 2017/18–2022/23 average, we have analysed the government’s overall spending from 2017/18 onwards, reaching a total by summing the spending of all the major departmental groups.*****

These figures show that spending on consultancy increased by 42% during the pandemic, from £1.2bn****** in 2019/20 to £1.7bn – the peak – in 2021/22. Spending has since been falling year on year, reaching £1.1bn in 2024/25.

These trends in overall government spending on consultancy are mostly driven by a small number of departments. DHSC, for example (whose figures include the NHS), is a big spender and between 2019/20 and 2021/22 increased its spending by £385m (105%) as it responded to the pandemic.******* This accounted for 75% of the growth in overall government spending over that period. Its spending subsequently fell by £494m (66%) between 2021/22 and 2024/25 – accounting for 84% of the £590m fall for the government as a whole in that period. DHSC’s accounts attribute much of this to moving the Test and Trace programme to the UK Health Security Agency (UKHSA) in October 2021. UKHSA consultancy spend is reported as part of DHSC’s consultancy spend, so we assume this move was accompanied by transferring work from consultants to UKHSA staff.

The Home Office has also seen dramatic fluctuations in its consultancy spending in recent years. It more than doubled between 2020/21 and 2021/22, increased by 62% the following year and by 86% in 2023/24 – reaching £253m in that year. While successive ARAs during this period refer to pressures around migration and asylum, ⁸⁵⁸ Home Office, ‘Home Office annual report and accounts: 2021 to 2022’, corporate report, GOV.UK, 14 July 2022, www.gov.uk/government/publications/home-office-annual-report-and-accounts-2021-to-2022; Home Office, ‘Home Office annual report and accounts: 2022 to 2023’, corporate report, GOV.UK, 19 September 2023, www.gov.uk/government/publications/home-office-annual-report-and-accounts-2022-to-2023; Home Office, ‘Home Office annual report and accounts: 2023 to 2024’, corporate report, GOV.UK, 30 July 2024, www.gov.uk/government/publications/home-office-annual-report-and-accounts-2023-to-2024  they mostly do so in connection to temporary staff (see below), rather than consultancy spending.

It appears very likely that it was the same pressures driving the increase. The Home Office’s 2023/24 ARA does refer to an impact on consultancy spending from migration pressures, and specifically refers to the Rwanda policy and a package of changes to migration policy. ⁸⁵⁹ Home Office, ‘Home Office annual report and accounts: 2023 to 2024’, corporate report, GOV.UK, 30 July 2024, www.gov.uk/government/publications/home-office-annual-report-and-accounts-2023-to-2024  Consultancy spending subsequently halved in 2024/25 – and while that year’s ARA does not explain the fall, it also does not mention either of those specific points ⁸⁶⁰ Home Office, ‘Home Office annual report and accounts: 2024 to 2025’, corporate report, GOV.UK, 17 July 2025, www.gov.uk/government/publications/home-office-annual-report-and-accounts-2024-to-2025  – suggesting that the new government’s changed policy in this area could be why consultancy spending fell. Notably, the 2025 budget specifically referred to “clamping down” ⁸⁶¹ HM Treasury, ‘Budget 2025 document’, policy paper, GOV.UK, 26 November 2025, www.gov.uk/government/publications/budget-2025-document  on consultancy in the Home Office – indicating that this had already happened, with funding “repurposed for front-line police”. ⁸⁶² Ibid.

A few other departmental trends over this time period are notable. The Treasury’s spending fell continuously between 2017/18 and 2021/22, by a total of £193m (87%), before fluctuating in more recent years. The sustained fall until 2021/22 appears to be largely due to the winding up of the operations of UK Asset Resolution (the ‘bad bank’ established by the government to handle risky assets after the financial crisis). ⁸⁶⁹ HM Treasury, ‘HM Treasury annual report and accounts 2020 to 2021’, corporate report, GOV.UK, 20 July 2021, www.gov.uk/government/publications/hm-treasury-annual-report-and-accounts-2020-to-2021; HM Treasury, ‘HM Treasury annual report and accounts 2021 to 2022’, corporate report, GOV.UK, 19 July 2022, www.gov.uk/government/publications/hm-treasury-annual-report-and-accounts-2021-to-2022

The spending of the Department for Business, Energy and Industrial Strategy, meanwhile, rose by £139m (197%) between 2019/20 and 2021/22. While its ARAs make no mention of why this might be, it is reasonable to assume that this was connected to the pandemic. Defra’s spending more than trebled in 2022/23 alone, which is attributed in its ARA to changes in how some spending was classified ⁸⁷⁰ Department for Environment, Food & Rural Affairs, ‘Defra’s annual report and accounts 2022 to 2023’, corporate report, GOV.UK, 26 October 2023, www.gov.uk/government/publications/defras-annual-report-and-accounts-2022-to-2023  – giving an insight into the difficulty that government often has with properly classifying this type of spend.

Departments’ 2024/25 figures also give the first indication of the progress being made towards the government’s target. It is notable that, except for the three relatively new departments (DBT, DSIT and DESNZ), consultancy spending fell in every department in 2024/25. Overall, spending on consultancy fell by £179m (14%) between 2023/24 and 2024/25. This was primarily due to the reduction in Home Office spending discussed above, but also due to reductions in Defra, HMT, DfT and DHSC.

It is not clear if these reductions, albeit minor compared to the Home Office (whose £136m dwarves the £15–20m for each of those four other departments), are a direct result of the government’s determination to lower costs or would have happened in any case. Indeed Defra’s lower spend is largely due to anticipated changes connected to the advancement of an Environment Agency programme.********, ⁸⁷¹ Department for Environment, Food & Rural Affairs, ‘Defra’s annual report and accounts 2024 to 2025’, corporate report, GOV.UK, 12 November 2025, www.gov.uk/government/publications/defras-annual-report-and-accounts-2024-to-2025  And in HMT’s case, while there is no specific reason given for the lower spend in 2024/25, ⁸⁷² HM Treasury, ‘HM Treasury annual report and accounts 2024 to 2025’, corporate report, GOV.UK, 21 July 2025, www.gov.uk/government/publications/hm-treasury-annual-report-and-accounts-2024-to-2025  its 2022/23 report references high spending, which was expected to fall in future – including on scaling up the UK Infrastructure Bank, and time-limited economic and IT consultancy. ⁸⁷³ HM Treasury, ‘HM Treasury annual report and accounts 2022 to 2023’, corporate report, GOV.UK, 20 July 2023, www.gov.uk/government/publications/hm-treasury-annual-report-and-accounts-2022-to-2023

In addition to these cases of spending falling by chance timing, the government will be helped towards its target by the fact that overall spending was already falling in 2023/24 – the year before it came to office. In that year, there were significant reductions in spending in DCMS, DfT and the MoD. While the latter two departments again offer little information about why this might be, the fall of £102m (86%) in DCMS is mostly attributable to the department no longer incurring costs connected to the 2022 Commonwealth Games. ⁸⁷⁴ Department for Culture, Media and Sport, ‘DCMS annual report and accounts 2023 to 2024’, corporate report, GOV.UK, 21 November 2024, www.gov.uk/government/publications/dcms-annual-report-and-accounts-2023-to-2024

This shows how much consultancy spending in any given department can fluctuate year to year for idiosyncratic reasons.********* This makes it difficult to assess whether the overall reduction seen in 2024/25 has anything to do with the government’s target, and whether the target itself is achievable.

Total government spending on consultancy is clearly falling. But the government has already reaped the benefits of post-pandemic falls in spending in DHSC, and possibly from altered migration policy in the Home Office. And there is no shortage of complex projects in government for which departments may choose to seek support from consultants in the coming years. On the basis of the information currently available, then, reaching the government’s target looks far from a given.

Temporary staff spending has also fallen

Unlike consultancy spending, the government has not stated an intention to reduce spending on temporary staff. Spending in this area rose continuously between 2017/18 and 2021/22, peaking in that year at £8.8bn. Most of this is DHSC – and within that, NHS – spending, with the former accounting for 70–80% of this spending each year since 2017/18. Post-pandemic trends saw this fall in 2022/23 and 2023/24, and then – more dramatically – to £5.9bn in 2024/25.
 

As was the case for consultancy spending, then, the increases in overall temporary staff spending up to 2021/22 were driven by DHSC and the NHS’s understandable activity during the pandemic. Overall government spending on temporary staff increased by £3bn between 2017/18 and 2021/22, and that of DHSC by £1.9bn. This is also true for the subsequent fall in such spending since 2021/22 – lower spending in DHSC accounts for 82% of the reduction in overall government spending between 2021/22 and 2024/25. And notably, between 2023/24 and 2024/25 alone, DHSC’s spending in this area fell by £1.3bn (24%).**********

DWP is another department where a pandemic impact is evident – with a spike in spending in 2020/21 and falls in subsequent years.

The MoD also stands out in having dramatically increased its spending on temporary staff – by £437m (more than 300%) – between 2018/19 and 2021/22, followed by a decrease of £284m (49%) between 2021/22 and 2024/25. Its ARAs attribute this to various causes – including recruitment challenges in Defence Equipment and Support and demands within strategic command, including digital and data work. ⁸⁸³ Ministry of Defence, ‘Ministry of Defence annual report and accounts 2019 to 2020’, corporate report, GOV.UK, 22 October 2020, www.gov.uk/government/publications/ministry-of-defence-annual-report-and-accounts-2019-to-2020; Ministry of Defence, ‘Ministry of Defence annual report and accounts 2020 to 2021’, corporate report, GOV.UK, 20 January 2022, www.gov.uk/government/publications/ministry-of-defence-annual-report-and-accounts-2020-to-2021; Ministry of Defence, ‘Ministry of Defence annual report and accounts 2021 to 2022’, corporate report, GOV.UK, 14 July 2022, www.gov.uk/government/publications/ministry-of-defence-annual-report-and-accounts-2021-to-2022  Its 2022/23 ARA refers to programmes requiring this support transitioning to a more permanent workforce, ⁸⁸⁴ Ministry of Defence, ‘Ministry of Defence annual report and accounts 2022 to 2023’, corporate report, GOV.UK, 20 July 2023, www.gov.uk/government/publications/ministry-of-defence-annual-report-and-accounts-2022-to-2023  and later ARAs simply to “savings measures implemented across the Department”. ⁸⁸⁵ Ministry of Defence, ‘Ministry of Defence annual report and accounts 2023 to 2024’, corporate report, GOV.UK, 30 July 2024, www.gov.uk/government/publications/ministry-of-defence-annual-report-and-accounts-2023-to-2024; Ministry of Defence, ‘Ministry of Defence annual report and accounts 2024 to 2025’, corporate report, GOV.UK, 4 November 2025, www.gov.uk/government/publications/ministry-of-defence-annual-report-and-accounts-2024-to-2025

The Home Office is also notable for its temporary staff spend, which largely echoes its consultancy spending – increasing from 2020/21 onwards until a peak in 2023/24, and a fall in 2024/25. Its ARAs again suggest that this reflects pressures around migration and asylum policy, as well as backlogs in passport applications. ⁸⁸⁶ Home Office, ‘Home Office annual report and accounts: 2021 to 2022’, corporate report, GOV.UK, 14 July 2022, www.gov.uk/government/publications/home-office-annual-report-and-accounts-2021-to-2022; Home Office, ‘Home Office annual report and accounts: 2022 to 2023’, corporate report, GOV.UK, 19 September 2023, www.gov.uk/government/publications/home-office-annual-report-and-accounts-2022-to-2023; Home Office, ‘Home Office annual report and accounts: 2023 to 2024’, corporate report, GOV.UK, 30 July 2024, www.gov.uk/government/publications/home-office-annual-report-and-accounts-2023-to-2024  The cause of the decrease in spending in 2024/25 is not clear, although as with consultancy spending, a comparison of ARAs suggest that policy changes including discontinuing the Rwanda policy could be a factor. ⁸⁸⁷ Home Office, ‘Home Office annual report and accounts: 2023 to 2024’, op. cit; Home Office, ‘Home Office annual report and accounts: 2024 to 2025’, corporate report, GOV.UK, 17 July 2025, www.gov.uk/government/publications/home-office-annual-report-and-accounts-2024-to-2025

As with consultancy spending, then, these examples demonstrate just how significantly spending on temporary staff can fluctuate year to year. So while spending does appear to be on a downward trend, it is too early to tell whether this will continue.

This year, the civil service’s forthcoming (delayed from last year) Strategic Workforce Plan presents an opportunity for the government in this area, as the future of the civil service workforce and the future trajectory of spending on consultancy and temporary staff are closely linked. The repeated use of temporary staff in some areas suggests that more capacity is needed, while for their part some departments have reported that pay restraint and restrictions on hiring have made it more difficult to attract candidates with the skills they need – forcing them to instead reach for more expensive consultants. ⁸⁸⁸ Comptroller and Auditor General, Government’s use of external consultants, op. cit; Cabinet Office, ‘Cabinet Office Annual Report and Accounts: 2023 to 2024’, corporate report, GOV.UK, 12 December 2024, www.gov.uk/government/publications/cabinet-office-annual-report-and-accounts-2023-to-2024

If the government is to reach its target for consultancy spending, while also reducing the size of the civil service, it will need to anticipate the skills that the civil service will need in the future – including those for which departments currently reach for consultants – and ensure that civil servants develop them now. It will also need to get far better at understanding where these skills already exist in the civil service. The government has said that these issues will be covered in the workforce plan, with Cat Little specifically mentioning skills gaps in digital and technology, and project delivery and management in evidence to the PAC. ⁸⁸⁹ House of Commons, Public Accounts Committee oral evidence, ‘Government’s use of external consultants’, 15 December 2025, https://committees.parliament.uk/event/25033/formal-meeting-oral-evidence-session  But some necessary steps are already clear – such as loosening eligibility requirements for pivotal role allowance to better retain specialists,*********** and making it easier to recruit external talent into the civil service, as we called for in our 2024 report 20 ways to improve the civil service. ⁸⁹⁰ Worlidge J, Urban J, Clyne R and Thomas A, 20 ways to improve the civil service, Institute for Government, 31 July 2024, https://www.instituteforgovernment.org.uk/publication/ways-improve-civil-service

_{*Other terms used to describe this type of support include ‘contingent labour’ and ‘agency workers’. We use the term ‘temporary staff’ throughout this chapter.}

_{**The previous consultancy and professional services spending control was managed by the Cabinet Office. It required ministerial approval for contracts of more than £120,000 or lasting for more than three months and Cabinet Office approval for contracts of more than £600,000 or nine months in duration. This was discontinued in 2023, in an effort to speed up decision making and shift accountability away from the centre and on to departments. Under the new approach, announced in late 2024, the Cabinet Office has required departments to establish their own controls – with permanent secretary approval for contracts lasting more than three months or worth more than £100,000, and ministerial approval for contracts lasting more than nine months or worth more than £600,000. The NAO has reported that departments differ in how they have approached this.}

_{***The framework is designed to provide public sector organisations with a centralised list of suppliers offering cost-effective management consultancy services. The value of the framework was reduced from a planned £5.7bn over four years to £1.7bn over two years.}

_{****The NAO report gives the Cabinet Office’s definition of consultancy as “providing advice to identify options or recommendations, or advice to assist with implementing solutions”. In oral evidence to PAC, officials referred to a more detailed definition provided in The Consultancy Playbook, which sets out guidance for how departments should engage consultants. The playbook, which is to be updated and reissued early this year, in turn refers to a more detailed definition of consultancy set out in the now-defunct consultancy spending control. A further, also more detailed, definition is available in the Treasury’s financial reporting manual. It is this that departments should abide by in preparing ARAs.}

_{*****See Methodology for details of which departments are included.}

_{******From this point, all figures in this chapter for consultancy and temporary staff spending are given in 2025/26 prices.}

_{*******Aside from the pandemic response, the department’s 2021/22 ARA also attributed this increase to the then government’s New Hospital programme.}

_{********Defra’s 2024/25 annual report and accounts states that spending on the EA’s Flood and Coastal Erosion Risk Management programme moved from resource-based spend (including consultancy) to capital-related construction work.}

_{*********Some departments’ 2024/25 ARAs further demonstrated this issue by reclassifying some of their spending. DWP, for example, reclassified much of its reported 2023/24 consultancy spending as ‘professional services’; HMRC referred to spending in both 2023/24 and 2024/25 that “should be classified as contracted out services [rather than consultancy]”; and DBT revised its 2023/24 figures following “an internal review of accounts codes” that led to “a re-allocation of costs between consultancy costs and other professional and legal services”. Such revisions by a number of departments in the same year are unusual, and suggest that a renewed focus on consultancy spending may have led departments to analyse it in more detail.}

_{**********Budget 2025 referenced efforts to reduce such spending in DHSC – referring to efforts to reduce the spending of NHS trusts on agency staff by “at least 30% in this financial year”, following an “almost £1 billion reduction in costly NHS agency spend in 2024–25”. It further referenced plans to “eliminate the use of all agency staff by the end of the Parliament”.}

_{***********The 2025 budget removed some of the controls on these allowances, which was a welcome start.}

Major projects

The Government Major Projects Portfolio (GMPP) brings together the most complex, high-cost and long-term projects across government departments. ⁸⁹⁶ HM Treasury, NISTA Annual Report 2024/25, 11 August 2025, www.gov.uk/government/publications/nista-annual-report-2024-2025/nista-annual-report-2024-25  The 213 projects in the portfolio have a total lifetime cost of almost £1 trillion (in today’s prices) and range from the construction of Dreadnought military submarines to national programmes of employment support. The aim of the GMPP is to enable central oversight and support for these large and strategically important programmes.

Until recently, the GMPP was overseen by the Infrastructure Projects Authority (IPA), a joint Cabinet Office and Treasury unit. The IPA worked alongside the National Infrastructure Commission (NIC) – an independent executive agency of the Treasury founded in 2015 with an independent chair to provide advice on infrastructure strategy.

In April 2025, the IPA and NIC were merged to form the National Infrastructure and Service Transformation Authority (NISTA) – as trailed in Labour’s election manifesto. Again a joint Treasury and Cabinet Office unit, NISTA reports to the chief secretary to the Treasury and will, in the department’s own words, bring “strategy and delivery under one roof”. ⁸⁹⁷ HM Treasury, ‘Government ushers in new era for UK infrastructure delivery’, GOV.UK, 1 April 2025, www.gov.uk/government/news/government-ushers-in-new-era-for-uk-infrastructure-delivery

NISTA has taken on responsibility for the GMPP, which, alongside the 10-year infrastructure strategy and MHCLG’s planning reforms ⁸⁹⁸ 203 HM Treasury, UK Infrastructure: a 10 year strategy, CP1344, The Stationery Office, HM Treasury, https://assets.publishing.service.gov.uk/media/6853c5db99b009dcdcb73649/UK_Infrastructure_A_10_Year_Strategy_Web_Accessible.pdf ^{, 899} Metcalfe S and Mitchell M, ‘Five changes set out in the Planning and Infrastructure Bill’, Institute for Government, 24 March 2025, www.instituteforgovernment.org.uk/comment/five-key-changes-planning-and-infrastructure-bill  the government hopes will help address some of the strategic and delivery problems that lead to frequent cost and time overruns on big infrastructure projects.

The whole-life cost of the Government Major Projects Portfolio is at a record high of £1 trillion

NISTA’s first annual report sets out the status of the GMPP up to 31 March 2025, immediately before NISTA was formally established. ⁹⁰⁰ HM Treasury, NISTA Annual Report 2024-2025, GOV.UK, 11 August 2025, www.gov.uk/government/publications/nista-annual-report-2024-2025/nista-annual-report-2024-25  Although this is the first GMPP report since Labour came to power in July 2024, only 2% of projects have a start date after the 2024 general election. Given the average project duration of 11.3 years, the data outlined below is largely a baseline which this government will build on, and have to manage, over the rest of the parliament, rather than the result of its own policy decisions.

The whole-life cost of all 213 projects in the GMPP has reached a record level of £998bn. This is more than double the real-terms cost in 2013, but the number of projects has only increased by 12% (from 191) in that time – meaning the increase in overall portfolio cost has largely come from increased cost per project.

There are 56 projects in the GMPP for which costs are not published; for example, due to national security or commercial grounds. In 2025, these made up 43% of the portfolio’s whole-life cost, up from 7% in 2021. The overall increase in the portfolio’s whole-life cost since 2021 has come almost entirely from exempt projects, rather than those for which cost information has been provided.

The GMPP also contains 17 projects that do not provide a delivery confidence rating – the majority led by the MoD or DESNZ. These include some very large projects that are exempt from publicly disclosing detailed information, again largely on national security or commercial grounds, such as the Dreadnought submarines and Sizewell C nuclear power plant, which have estimated whole-life costs of £31bn and £38bn respectively, according to recent reports. ⁹⁰³ Clun R and Issimdar M, ‘Sizewell C nuclear power plant costs rise to £38bn’, BBC News, 22 July 2025, retrieved 26 November 2025, www.bbc.co.uk/news/articles/cev03wer0p2o ^{, 904} Mills C, Replacing the UK’s Nuclear Deterrent: Progress of the Dreadnought Class, House of Commons Library, 5 August 2025, https://commonslibrary.parliament.uk/research-briefings/cbp-8010/

The number of projects in the portfolio subject to these exemptions has gone up from 20 in 2021 to 56 in 2025, which may reflect the inclusion of projects from an earlier stage when costs may not be known or cannot be disclosed due to the risk of this biasing procurement. However, the average whole-life cost per exempt project has also more than tripled in that period. Providing support for these projects, particularly given their ballooning costs, will likely demand an increasing proportion of NISTA’s resources. The exemption of some projects with particular sensitivities from public reporting is understandable, but the growing whole-life cost of exempt projects merits thorough scrutiny of whether they are delivering good value for money.

In the absence of published information to support external scrutiny, NISTA should ensure both internal processes are robust, but also sufficient access to information for the Public Accounts and Intelligence and Security Committees of parliament to effectively scrutinise this spend.

Overall confidence in delivery has weakened

The proportion of the GMPP that has a delivery confidence rating of ‘unfeasible’ (by whole-life cost) has reached its highest level since 2020 – though at the same time, there was also a small increase in the proportion of the portfolio with delivery confidence rated as ‘highly likely’.*

There is substantial variation of these costs and ratings by department. However, where departments have notably higher whole-life cost of projects rated as either ‘unfeasible’ or ‘highly likely’ this is often driven by a single project. For example, the £25bn Farming and Countryside Programme in Defra is rated as ‘unfeasible’, and DWP’s ongoing Universal Credit programme with whole-life cost of £17bn is rated as ‘highly likely’. It is these large projects in particular that can have an outsized impact on a department’s budgets if costs overrun – and where central oversight and delivery support can have the most impact.

Evaluation of major projects has improved but has a long way to go

In 2019, only 8% of GMPP spending was covered by a suitable impact evaluation plan. ⁹¹² Comptroller and Auditor General, Evaluating government spending, Session 2021-22, HC 860, National Audit Office, 2 December 2021, www.nao.org.uk/wp-content/uploads/2021/12/Evaluating-government-spending.pdf  A review by the Evaluation Task Force (ETF) last year showed a substantial improvement, with a third of all projects, representing 38% of GMPP whole-life cost, now having good quality evaluation plans in place. ⁹¹³ Evaluation Task Force, Government Major Projects Evaluation Review, GOV.UK, 23 April 2025, www.gov.uk/government/publications/government-major-projects-evaluation-review/government-major-projects-evaluation-review-html#section-1-coverage-an…  However, another 29% provided evaluation plans of insufficient quality, while more than a third of projects provided no evaluation plan at all.

The 2025 review set out an action plan to improve evaluation governance and capability, and ensure evaluation is embedded into project delivery across the board – which the ETF, NISTA and the Treasury will be jointly responsible for delivering. ⁹¹⁴ Evaluation Task Force, Government Major Projects Evaluation Review, GOV.UK, 23 April 2025, www.gov.uk/government/publications/government-major-projects-evaluation-review/government-major-projects-evaluation-review-html#section-1-coverage-an…

NISTA has an opportunity to get a grip of the ballooning major projects portfolio

There are some welcome signs that this government is taking a more considered and coherent approach to infrastructure strategy and delivery. Officials have said that the 10-year Infrastructure Plan has encouraged longer term thinking and the government has said that business cases for major projects will be published within four months of approval – although this is a commitment previous governments have failed to deliver on. ⁹¹⁵ Pope T, Tetlow G and Pattison J, Capital spending in public services, Institute for Government, 2024, www.instituteforgovernment.org.uk/publication/capital-spending-public-services  It is good that the first of these has been published; ⁹¹⁶ HM Treasury, ‘Business case publications: collection’, 15 December 2025, retrieved 16 December 2025, www.gov.uk/government/collections/business-case-publications-collection  more must follow over the coming months.

The Office for Value for Money’s report on ‘megaprojects’ was an honest public assessment of recurring delivery problems and cost overruns, and the government has accepted its recommendations on streamlining decision making and between-year funding flexibility. ⁹¹⁷ Office for Value for Money, Value for money study: governance and budgeting arrangements for mega projects, GOV.UK, 19 June 2025, https://assets.publishing.service.gov.uk/media/6853c3fc235ba1380b6aa6e5/OVfM_Mega_projects.pdf  Given that almost two thirds of the portfolio’s whole-life cost** is from projects that are due to expire beyond this parliament, sustaining this long-term focus in the face of short-term priorities will be crucial.

Similarly, despite many projects having no published whole-life cost or delivery confidence rating, NISTA’s first report on the GMPP provides more detail in other areas than past data releases, which support more nuanced and earlier support, and greater accountability for the projects. This includes ranges for costs and expected end dates for some projects, the inclusion of a small number of projects at a very early stage (before the strategic outline case has been finalised), and centrally providing the name of the civil servant in charge of each project (the senior responsible officer) for almost all projects.

At the autumn 2025 budget, the government announced that the GMPP will be slimmed down to 80–100 projects so that “central expertise and ministerial bandwidth is focused on the most strategically important, complex and risky projects”. ⁹¹⁸ Office for Value for Money, Reforming the spending control and accountability framework, 26 November 2025, GOV.UK, https://assets.publishing.service.gov.uk/media/6925f3a122424e25e6bc31b3/Controls_report.pdf  This makes sense. The challenge will be determining which projects are included, and ensuring that ones that would most benefit from NISTA’s input are not left out.

There are some big potential wins to setting up NISTA as a single unit inside central government with responsibility for both major projects strategy and delivery. Strong ministerial oversight from the chief secretary could bring political direction to difficult project decisions. Bringing relevant strategic and delivery expertise together should enable NISTA to provide more coherent advice, and being an internal unit, as opposed to the independent agency model of the NIC, should enable closer working with those overseeing budgets in the Treasury.

The risk is that the government loses the NIC’s ability to provide and publish honest, independent advice and bring external expertise to an area where government has a poor track record. Whether this can be mitigated will depend on ministers and NISTA’s civil service leadership being open to expertise and challenge, even when that is politically difficult.

An early sign that a breadth of expertise will be valued is the appointment of Becky Wood as NISTA’s chief executive. She brings with her extensive experience managing large infrastructure projects in both the public and private sectors, such as Crossrail, across the UK and internationally. Further signs of success would include publication of more business cases, substantially more projects having good evaluation plans in place and addressing delivery problems with projects currently rated as ‘unfeasible’ such that their delivery confidence rating improves.

^{* The delivery confidence rating indicates the likely success of a project achieving its aims and objectives on time and on budget. Green = highly likely; amber = feasible, however significant issues exist; red = unachievable. A more detailed description of each delivery confidence rating can be found in Appendix A of HM Treasury, NISTA Annual Report 2024-2025, GOV.UK, 11 August 2025, www.gov.uk/government/publications/nista-annual-report-2024-2025/nista-annual-report-2024-25}

^{** Of projects for which a whole life cost is provided.}

Departmental spending and efficiencies

The government had to move quickly in 2024 to provide clarity on spending plans

The government faced an unenviable fiscal inheritance. No spending plans had been set for 2025/26, and the figures pencilled in for the years ahead, by the outgoing chancellor Jeremy Hunt, were implausibly tight. ⁹²⁹ Hoddinott S, Rowland C, Davies N and others, Fixing public services: priorities for the new Labour government, Institute for Government, 2024, https://www.instituteforgovernment.org.uk/publication/fixing-public-services-labour-government  Rachel Reeves quickly set one-year budgets for 2025/26 at the 2024 autumn budget. That autumn budget also saw substantial uplifts in both day-to-day and capital spending – enabled by raising taxes and changes to the fiscal rules.

The increase in capital spending was a welcome recognition that a lack of investment has held back long-term economic growth and public services performance. ⁹³⁰ Davies N and Haile D, Public services performance tracker 2025: Cross-service analysis, Institute for Government, 19 November 2025, www.instituteforgovernment.org.uk/publication/performance-tracker-2025/cross-service-analysis  Real-terms day-to-day spending increased by 4.3% in 2025/26 – the second largest increase of any spending round since 1997. ⁹³¹ Hoddinott S, Davies N, Pope T, Dellar A and Nye P, Austerity postponed? The impact of Labour’s first budget on public services, Institute for Government, 12 November 2024, www.instituteforgovernment.org.uk/sites/default/files/2024-11/impact-labour-first-budget-public-services.pdf  That budget also set the spending envelope up to the end of the parliament – within which the 2025 spending review would set departmental settlements – but with much less generous average annual real-terms increases of just 1.3% between 2026/27 and 2028/29.

There were some welcome process changes in the latest spending review

In 2024, the government also made some important improvements to how it makes spending decisions, including a regular cycle of spending reviews and providing longer term certainty on capital spending. ⁹³² HM Treasury, A strong fiscal framework, CP 1344, The Stationery Office, 2024, https://assets.publishing.service.gov.uk/media/67211bf34da1c0d41942a8bf/A_strong_fiscal_framework.pdf These changes gave the government a good foundation from which they could have gone for bigger reforms in 2025 – but these were not, entirely, achieved last year.

On launching the 2025 spending review, back in December 2024, the chancellor set out her ambition to take a new approach to setting departments’ budgets, ⁹³³ HM Treasury, ‘Every pound spent will deliver Plan for Change’, press release, GOV.UK, 12 December 2024, www.gov.uk/government/news/chancellor-every-pound-spent-will-deliver-plan-for-change  and move away from the siloed and insufficiently strategic approach of the past. ⁹³⁴ Bartrum O, Paxton B and Clyne R, How to run the next multi-year spending review, Institute for Government, 2024, www.instituteforgovernment.org.uk/publication/how-run-next-multi-year-spending-review  This would include using the five missions as a framework for prioritisation, with greater cross-government collaboration around them, a more in-depth ‘zero-based review’ of existing spend to drive better value for money, and enabling innovation with new Test, Learn and Grow pilots (see Box 2.1 below).

In the event, however, it is not clear a substantially different approach was taken at the 2025 spending review. While some in government say they took a genuinely new approach, to many others it felt similar to past processes. This aligns with our own assessment – outlined below – that despite some improvements, this was far from a ‘total rewiring’ of how government approaches spending decisions.

There was no step-change in collaboration during the spending review process

The 2025 spending review failed to deliver a substantially more collaborative approach. Confusion about the role that mission boards and mission clusters* were supposed to play in the spending review hindered progress. Some found the various mission structures improved understanding of departmental interdependencies on spending, but others found they added little value. The former chief secretary to the Treasury, Darren Jones, said in December 2025 that mission boards, as well as departments, were asked to submit bids. ⁹³⁵ House of Commons, Public Administration and Constitutional Affairs Committee oral evidence, ‘The work of the Cabinet Office’, 16 December 2025, https://committees.parliament.uk/event/25586/formal-meeting-oral-evidence-session  Allocations were ultimately made along departmental lines – even if some specific funding was labelled as mission-related. To many people involved, the process did not feel substantially less siloed than in the past.

The inability of these changes to systematically improve collaboration was perhaps the result of insufficient incentives. Previous Institute for Government work has recommended allocation of joint budgets, shared across departments, to provide this incentive. ⁹³⁶ Bartrum O, Paxton B and Clyne R, How to run the next multi-year spending review, Institute for Government, 2024, www.instituteforgovernment.org.uk/publication/how-run-next-multi-year-spending-review  Doing this should have been possible within existing structures. However, the spending review 2025 announcement included very few clearly joint bids or subsequent shared budgets. ⁹³⁷ HM Treasury, Spending review 2025, GOV.UK, 30 June 2025, www.gov.uk/government/publications/spending-review-2025-document/spending-review-2025-html

In the absence of joint budgets, a comprehensive planning and performance framework, with shared outcomes that departments were jointly accountable for delivering, could also have incentivised collaboration in the spending review. No such framework was published alongside spending settlements, although the government has now said it will publish ‘strategic plans’ that cover “strategic outcomes against spending” in spring 2026. ⁹³⁸ House of Commons, Public Administration and Constitutional Affairs Committee, oral evidence, ‘The work of the Cabinet Office’, 16 December 2025, https://committees.parliament.uk/event/25586/formal-meeting-oral-evidence-session

One valuable improvement that could still improve collaboration – in future – was establishing a programme of thematic value-for-money reviews between spending reviews. This follows a recommendation by the Office for Value for Money, a temporary Treasury unit established to drive forward improvements to the spending framework. If done well, these exercises could enable more in-depth policy development around cross-cutting topics going forward.

Box 2.1 Test, learn and grow

The Test, Learn, Grow (TLG) programme aims to drive public service reform and was backed by £100m in the 2025 spending review. It is designed to “bring policy makers closer to the front-line” to deliver “people-focused” public services. ⁹⁴⁷ House of Commons, Hansard, ‘Public Service Reform: Test, Learn and Grow Programme’, 16 July 2025, col.16, https://hansard.parliament.uk/Commons/2025-07-16/debates/25071645000010/PublicServiceReformTestLearnAndGrowProgramme  It comes in two parts: ‘test and learn’ and then ‘grow’. The overarching idea is to pilot new ideas, use the results to improve or pivot plans, then scale approaches that work.

Wave 1: Partial lessons on iterative, locally driven reform were valuable

The programme has been characterised by two waves. Wave 1 functioned as a rapid proof-of-concept of the programme prior to the 2025 spending review. Delivered in 12 weeks with minimal funding and pro-bono secondments, it was a test of ‘test and learn’ and used to inform the design of the programme. Pilots ran in four locations across two policy areas: temporary accommodation and family hubs, the latter tied to government’s opportunity mission – the pilots for which were particularly promising.

In Manchester, ‘innovation squads’ ⁹⁴⁸ Cabinet Office, ‘Communities across the country to benefit ‘from innovation squads’ to re-build services’, GOV. UK, 16 July 2025, www.gov.uk/government/news/communities-across-the-country-to-benefit-from-innovation-squads-to-re-build-public-services  collaborated with local government officials and communities to map barriers to accessing family hubs, identify potential solutions and test outreach approaches. The government has yet to publish formal evaluation of the proof of concept but officials reported that this initial pilot drove an increase in new families engaging with hub services – a critical vehicle for meeting government’s target to increase the number of young children achieving ‘good learning development’ standards. ⁹⁴⁹ Davison N and Metcalfe S, Labour’s Best Start in Life Strategy has got off to a good start – now comes the hard part, Institute for Government, 11 July 2025, www.instituteforgovernment.org.uk/comment/labour-best-start-life-strategy  Officials also experimented with ‘test and learn’ methods in other areas, noting the value of collaborating with service users and ‘failing fast’ to save money.

In Sheffield, local government officials also piloted new outreach methods and reported a 25% increase in uptake of family hub services. ⁹⁵⁰ Lord E, ‘Test, Learn, Grow: Bridging the centre and the edge’, Capacity, 17 July 2025, https://thisiscapacity.co.uk/test-learn-grow-16-july-2025/  They also described a shift from a hierarchical relationship with central government representatives to a more equal, collaborative partnership.

Wave 2: Test, Learn and Grow will need to scale up to fulfil its promise

Wave 2 saw the programme move from proof of concept to full delivery. Announced in July and formally launched in October 2025, it has expanded the programme across 10 locations and missions: health (establishing neighbourhood health services), growth (getting more people into work) and safer streets (tackling violence against women and girls). ⁹⁵¹ House of Commons, Hansard, ‘Public Service Reform: Test, Learn and Grow Programme’, 16 July 2025, col.16, https://hansard.parliament.uk/Commons/2025-07-16/debates/25071645000010/PublicServiceReformTestLearnAndGrowProgramme

If the government is to use these teams as a core part of delivering the preventative, people-focused services government wants, it will need to tackle two major challenges.

• Visibility: To achieve its ambition to “change the centre”, ⁹⁵² Nick Kimber, LinkedIn post, February 2025, www.linkedin.com/feed/update/urn:li:activity:7311312816701952000/  TLG will need a clear strategy for raising its profile and embedding lessons across departments as the programme scales. The programme’s existing steering
  group may be the right mechanism for building senior buy-in across Whitehall, raising awareness among ‘rank and file’ officials will require greater transparency. A first step is publishing learnings from Wave 1.
• Scaling: Success of the initiative relies on local political appetite as well as capacity boost from the ‘innovation squads’, ⁹⁵³ Cabinet Office, ‘Communities across the country to benefit ‘from innovation squads’ to re-build services’, GOV. UK, 16 July 2025, www.gov.uk/government/news/communities-across-the-country-to-benefit-from-innovation-squads-to-re-build-public-services  Cabinet Office officials deployed to work with local government. The programme will need to set out a strategy and explain how it will scale to areas with weaker political appetite and engagement, ensuring the model works beyond early adopters and without the added capacity of the innovation squads. Steps are being taken to address this challenge through a Test, Learn and Grow Network, which will “provide a space for peer learning, sharing best practice and sustaining momentum” for participating councils and others outside the programme. ⁹⁵⁴ Cabinet Office, ‘Communities across the country to benefit ‘from innovation squads’ to re-build services’, GOV. UK, 16 July 2025, www.gov.uk/government/news/communities-across-the-country-to-benefit-from-innovation-squads-to-re-build-public-services

Test, Learn and Grow is itself a test of the government’s ambitions: outcome-focused, iterative and collaborative. The real test will come in seeing whether it can deliver on its promise to really change how the state works.

The extent to which missions drove spending review allocations is unclear

The chancellor gave a speech to kick-off the multi-year spending review process in the week after the Plan for Change set out more detail on the government’s missions. ⁹⁵⁷ HM Treasury, ‘Every pound spent will deliver Plan for Change’, press release, GOV.UK, 12 December 2024, www.gov.uk/government/news/chancellor-every-pound-spent-will-deliver-plan-for-change  As part of the spending review process, departments were asked to review spending line-by-line to ensure it was aligned with the Plan for Change. In practice, this meant business cases and spending bids setting out how plans would support delivery of the missions.

To some extent, this appears to have worked. The missions did shape many departments’ submissions to the Treasury, and negotiations thereafter. And the settlements received by individual departments and budgets most directly responsible for the missions – DHSC, DESNZ, DfE and the police budget** – received real-terms increases in both resource and capital spending.***

Another six departments also saw both categories of spending increased. This included the MoJ, which is key to the wider safer streets mission, and DBT, DWP and MHCLG, which play important roles in the growth mission in leading on the industrial strategy, tackling economic inactivity and the housebuilding milestone respectively. DSIT received the largest relative increase in day-to-day spending, albeit from a low level of annual baseline spending (~£300m), reflecting its position as the new ‘digital centre of government’. Defence spending, one of the three foundations in Plan for Change and highlighted as one of Reeves’ top three priorities in her spending review speech, was also increased. In contrast, the departments that received less generous settlements are less directly relevant to the government’s five missions.

However, these decisions were also not obviously that different to those taken by past chancellors. DHSC was prioritised, as it has been at almost every recent spending reviews, and education was kept in the middling position it has tended to occupy since the early years of the last Labour government. In contrast, MoJ and local government have tended to lose out to other departments in the past, and are somewhat prioritised relative to others this time round. Although not mission-lead departments, both do have key roles in delivering the wider missions. The Home Office as a whole has not been prioritised, but this belies the diverging trajectories of planned increases for police spending and other Home Office spend; police spending had seen year-on year increases since 2018/19, ⁹⁵⁸ Rowland C, Performance Tracker 2025: Police, Institute for Government, 23 October 2025, www.instituteforgovernment.org.uk/publication/performance-tracker-2025/criminal-justice/police  and cuts to migration spending. With DESNZ only existing since 2023, straightforward comparisons of past and future spending are not possible.

Overall, this does not reflect a process that was ‘rewired’ around the five missions. Even if the departmental allocations could suggest a degree of prioritisation in line with missions, the reality is that their breadth, and the addition of three ‘foundations’ and DSIT as an ‘enabler’, means the 2025 spending review does not appear especially targeted at the missions themselves.

Ambitious departmental efficiency targets were accompanied by more detailed plans

Finding efficiencies was a key priority for spending review 2025. The chancellor set departments a headline target of “at least 5% savings and efficiencies” over the spending review period, ⁹⁶² HM Treasury, Spending review 2025, GOV.UK, 30 June 2025, www.gov.uk/government/publications/spending-review-2025-document/spending-review-2025-html  the same target used by Rishi Sunak as chancellor in 2021 ⁹⁶³ HM Treasury, Autumn budget and spending review, 2021, GOV.UK, October 2021, www.gov.uk/government/publications/autumn-budget-and-spending-review-2021-documents  (but which were not reported against). Taking 2025/26 day-to-day spending as the baseline, this means that departments will have to deliver annual savings and efficiencies of at least £21.9bn by 2028/29 (in 2025/26 prices).****

Alongside this headline target, Reeves asked departments to set out “stretching and realistic” plans for at least 1% per year ‘technical efficiencies’ – that is, delivering more output for the same input, or the same or more output for less input. Departmental efficiency plans were published setting out in much more detail than the past how the government hopes to achieve these technical efficiencies. ⁹⁶⁴ HM Treasury, Departmental Efficiency Plans, GOV.UK, 11 June 2025, https://www.gov.uk/government/publications/consolidated-budgeting-guidance-2025-to-2026  These set out plans (exceeding the 1% target) for technical efficiencies of 4% per year by 2028/29 overall, equivalent to £13.8bn in nominal terms.***** Given the emphasis on finding savings within tight budgets, it’s notable that this £13.8bn of efficiencies includes non-cash releasing efficiencies (e.g. a new IT system, which saves officials time that they then spend on other related tasks) as well as those that release cash, which could directly lead to lower spending (e.g. reduced energy costs from installing more efficient lighting).

All but four departments set out plans to meet or exceed this 3% technical efficiencies target (1% a year until 2028/29). The Home Office fell just short (2.9%), while the MoD set out plans to reach 2.3%. The MHCLG and DfE plans were for technical efficiencies of less than 1.5%, but these excluded efficiencies from front-line delivery spend, which make up a substantial proportion of those departments’ budgets. Almost two thirds of all planned technical efficiencies across government come from the DHSC’s ‘NHS productivity plan’, which it hopes alone will deliver almost £9bn of efficiencies a year by 2028/29.

While each department categorised their planned technical efficiencies differently they can be grouped into some broad categories, as outlined in the figure below.

The category of ‘workforce, corporate and digital reforms’ makes up the majority of planned technical efficiencies in over half of departments.****** AI is mentioned in almost three quarters of departments’ plans, and consolidation of corporate functions into shared services forms part of planned efficiencies for most departments.

The departmental efficiency plans also give an estimate of the investment needed to deliver their planned efficiencies. While the details of this investment are not specified, the explicit recognition that delivery of efficiencies often depends on investment – such as purchasing new IT infrastructure or renovating buildings – is welcome. Indeed the need for up-front investment to generate accumulating benefits from efficiencies is reflected in the planned level of investment remaining relatively flat over the spending review period, while efficiencies net of investment steadily increases.

Within the overall target is of at least 5% savings and efficiencies, equal to £21.9bn in 2025/26 prices, less than £13.0bn is due to come from the technical efficiencies discussed here. This means several billion pounds of the savings and efficiencies needed to reach this target are largely unspecified. These are likely to come from allocative efficiencies (reallocating resources to more efficient activities) and non-efficiency savings (stopping activities all together). Implementing these savings and efficiencies can be politically difficult, and will only become harder the closer we get to the next general election.

Ambitious administration budget savings targets have been set by the centre

Alongside whole department savings and efficiency targets, the chancellor asked departments to make substantial cuts to their administration budgets, with planned savings of 11% by 2028/29 and 16% by 2029/30.

These budgets cover non-front-line, or what is sometimes referred to as ‘back-office’, spending – for example, salaries for officials providing policy advice, technical support and office services, and the procurement of goods and services they use such as IT equipment and consultancy services. ⁹⁷⁰ HM Treasury, Consolidated budgeting guidance: 2025–26, GOV.UK 27 February 2025, www.gov.uk/government/publications/consolidated-budgeting-guidance-2025-to-2026/consolidated-budgeting-guidance-2025-26#administration-budgets

That most departments are set to hit both targets for administration budgets exactly, though a few are planning to exceed them, suggests a lack of detailed work on how exactly this will be achieved. However, there is also some variation in the trajectory of savings between departments, which suggests that there has been at least some planning at a departmental level to set the potential path of savings. This includes three departments with plans for administration budgets to rise in 2026/27 before falling thereafter.

The spending review did not produce departmental plans for administration budget savings. In many cases, technical efficiencies set out in the departmental efficiency plans will also lower administration budgets. But they are unlikely to deliver the full scale of savings necessary – departments will still need to reallocate resources to areas where they can achieve the same but with less, or stop doing some activities altogether.

Smaller administration budgets will require headcount cuts

Staff costs made up 82% of administration budgets in 2024/25; ⁹⁷¹ HM Treasury, Public Expenditure Survey Analyses 2025, GOV.UK, 17 July 2025, https://assets.publishing.service.gov.uk/media/6874fa6f92691289bdb7d393/Public_Expenditure_Statistical_Analyses_2025.pdf  meeting these savings targets will unavoidably require headcount reductions. Indeed, this is a desired result of imposing these savings targets – and is a far preferable approach to the perverse incentives caused by blanket headcount targets. ⁹⁷² Worlidge J, Grama T, Urban J and others, Whitehall Monitor 2025, Institute for Government, 16 January 2025, www.instituteforgovernment.org.uk/sites/default/files/2025-09/whitehall-monitor-2025.pdf  The Institute for Fiscal Studies has estimated that hitting the target of 16% administration budget savings by 2029/30 would require as much as an 18% fall in the headcount of civil servants covered by such budgets.******* ⁹⁷³ Harvey-Rich O and Warner M, ‘How could planned cuts to administration budgets affect public sector productivity’, blog, Institute for Fiscal Studies, 22 September 2025, retrieved 26 November 2025, https://ifs.org.uk/sites/default/files/2025-09/IFS_Report_The_Outlook_for_Public_Sector_Productivity_0.pdf

There is no published data on the total headcount of civil servants whose salaries are covered by administration budgets; however, an 18% headcount cut is, by our own analysis, approximately equivalent to a reduction of between 29,000 and 40,000.******** The government hopes that exit schemes will help deliver some of these headcount reductions. However, the departments do not seem to have consistently factored in the up-front costs for these schemes into their spending plans. And moreover, current exit schemes are planning to provide less than a third of these exits (as of August 2025, 8,583 exits had been applied for by departments, although this figure has likely grown since then). ⁹⁷⁴ Comptroller and Auditor General, Government exits and redundancies, National Audit Office, 25 September 2025, www.nao.org.uk/wp-content/uploads/2025/09/government-exits-and-redundancies-1.pdf  So while early indications are that the government is moving in the right direction to bring down civil service numbers, further workforce reforms will be required. This makes the long-awaited Strategic Workforce Plan all the more crucial in the context of such significant cuts.

Ministers have sometimes said that they are reducing administration budgets to move spending to ‘front-line services’. Administration spend, however, makes up only a tiny fraction of all day-to-day spending (2.7% in 2025/26). The planned administration budget savings of £1.6bn a year by 2028/29 (in 2025/26 prices) are just a small proportion of the planned programme budget increase of £21.4bn over this period.

The extent to which lower administration spend could help to increase programme spend (one approximation for the ‘front line’) differs between departments. In large delivery departments, administration budgets make up only a fraction of overall day-to-day spending – less than 3% in DHSC, DfE and the Home Office. In these departments – usually considered front-line departments – the 10% cuts to administration budgets will make little difference (less than half a percentage point) to the balance of programme and administration spend within their budgets.

However, administration spending makes up a larger proportion of budgets in policy-focused departments such as the Treasury and DSIT. A 10% cut to these administration budgets within a set spending envelope implies a substantial restructuring of these (non-front-line) departments’ budgets, with programme spending making up a much larger proportion of the total than it did before. For example, programme spend in DSIT will go from being just under half of day-to-day spending (45%) in 2025/26, to just under two thirds (64%) in 2028/29 – a 19 percentage point increase.

That does not mean administration budget savings should not be made. The civil service has grown in recent years, and savings can be made – but these should be made based on where resources are best allocated, not driven by a blanket desire shift them to the front line.

There must be accountability for departments delivering these efficiencies

There is a long history of governments planning to do more with less, but struggling to put that aim into practice. The publication of departmental efficiency plans and planned trajectories of administration budget cuts could help to drive better internal planning, and will help enable external accountability for delivering them going forward. The Treasury should fulfil the commitment it has made to publish departmental efficiency plans again at spending review 2027. ⁹⁷⁸ HM Treasury, Departmental Efficiency Plans, GOV.UK, 11 June 2025, www.gov.uk/government/publications/departmental-efficiency-delivery-plans

In July, the Treasury confirmed that departments will be required to include efficiency reporting in the annual reports and accounts from 2026/27. ⁹⁷⁹ Bowler J, ‘Cabinet Office functional savings’, 28 July 2025, https://committees.parliament.uk/publications/49304/documents/262536/default/  To avoid this process being ‘gamed’, the Treasury has worked with the NAO to develop assurance arrangements. This includes departments providing evidence to ensure efficiency definitions are being met in good faith. Reporting should include an assessment of where efficiencies have been achieved as planned, where they have not, and include any additional efficiencies achieved that were not in the spending review 2025 plans. Where particular investments have led to efficiencies, or unexpectedly failed to do so, this should also be included.

At the 2025 autumn budget, the chancellor announced £2.8bn further efficiencies and savings in 2028/29 ⁹⁸⁰ HM Treasury, ‘Budget 2025’, policy paper, GOV.UK, updated 28 November 2025, www.gov.uk/government/publications/budget-2025-document/budget-2025-html#public-sector-receipts-and-spending  – but without specifying how they will be achieved.********* Following the more considered approach to efficiencies taken at the spending review 2025, this cart-before-the-horse approach is a disappointing step backwards. The fact these savings and efficiencies are due just before the next general election further brings into question their political credibility. Departments have been asked to appoint non-executive directors to help them find these efficiencies, but this is no substitute for analysis that can inform the level of efficiencies planned and the approach taken to delivering them.

^{* The mission boards are interministerial groups chaired by the lead secretary of state for each mission with a “remit to oversee and drive progress on the relevant mission”. In addition, departments were asked to meet in “mission clusters” during the spending review process to “agree priorities and links” around the missions as they put together their spending plans.}

^{** The Treasury is the lead department for the growth mission. However, it is difficult to assess the prioritisation of growth in this way since many of the policy levers sit across other departments or are non-spending measures.}

^{*** Real-terms police core spending power is due to increase by 1.7% on average each year over the spending review period, even as overall Home Office day-to-day spending is due to fall by 1.7%, driven by a planned reduction in spending on migration and asylum. Figures for planned capital spending on policing is the figure for the Home Office as a whole, as separate plans for policing were not published.}

^{**** The 2025 spending review did not provide a specific figure for the baseline so we have taken the 2025/26 baseline as RDEL excluding depreciation, Barnett consequentials and the reserve. 5% of this baseline is taken as the savings and efficiencies target.}

^{***** This is £13.0bn in 2025/26 prices. Nominal terms are used for ease of comparability with HMT’s reported figures. The 4% figure relates to the 2025/26 RDEL baseline, as set out in the departmental efficiency plans. This baseline includes Spring Statement 2025 plans (excluding the Reform and Innovation Fund, the Transformation Fund and ODA), plus machinery of government changes.}

^{****** DHSC did not break its £9bn of efficiencies down by category in a way that is comparable with other departments, and is excluded from the figure titled "Technical efficiencies in 2028/29 planned at spending review 2025, by department and category (2025/26 prices)".}

^{******* This assumes that staff costs stay constant as a share of the overall administration budget, and that average civil service pay follows the OBR forecast of 0.5% per year real-terms increase in public sector pay.}

^{******** This estimate takes the total administration budget paybill (£11.2bn) and divides it by estimates of the paybill costs for one median civil servant (including salary, civil service pension contributions and National Insurance contributions). The range depends on whether you take the median salary for all civil servants (£35,680), including those working in arm’s length bodies, or the median salary of civil servants working in the median core department for pay (£49,055). Core department civil servants are more likely to be covered by administration budgets than the former group, but due to a lack of data our estimate is less accurate than that for all civil servants, which is supplied by the Cabinet Office. See Methodology for further details.}

^{********* These include £1.4bn from the Ministry of Defence (MoD) and DHSC, which can be reinvested back into departmental budgets, and £1.4bn of cash-releasing savings and efficiencies from all other departments.}