Whitehall Monitor 2026: Methodology

We set out below our methodology behind various analyses that have contributed to this year’s Whitehall Monitor. For questions on the below, or regarding any of our findings, please get in touch via email at enquiries@instituteforgovernment.org.uk.

Publication of data

The charts presented in this report draw principally on publicly available government data, which we have further processed and analysed. The data for the majority of the charts is available through clicking the ‘Download the data’ link, which can be seen in the chart footer (HTML version only). For the remaining charts that do not have this functionality built in, the final data inputs are published on our website alongside this report. If you have any questions about the analysis of this data, please get in touch via email at the address given above.

Throughout the report

Most of the analysis in this report is based on the Cabinet Office’s annual Civil Service Statistics publication. This dataset rounds numbers to the nearest five and reports numbers less than five as ‘[c]’. We replace these suppressed figures with ‘3’. Unless otherwise stated, all figures for numbers of civil servants are FTE rather than headcount figures. FTE figures count part-time staff according to the proportion of full-time hours they work. For example, somebody working three days a week would be counted as 0.6 FTE. In most cases, this metric is more meaningful than headcount, which does not distinguish between part- and full-time staff.

Defining civil servants

We define ‘civil servants’ as politically impartial, appointed officials of the UK Home Civil Service, which supports the work of the UK’s central government departments. This includes agencies that employ civil servants such as executive agencies, non-ministerial departments and some non-departmental public bodies.

Our definition includes staff of the three Whitehall-based territorial offices that manage the UK’s relationships with Scotland, Wales and Northern Ireland, though for simplicity we do not include them in much of our analysis that compares different departments. It also includes the civil servants that the devolved governments of Wales and Scotland employ, but not the staff of the Northern Ireland Civil Service, which is administratively distinct.

In this way, civil servants are defined more narrowly than public sector workers – police officers, teachers, NHS staff, members of the armed forces and local government officers are not counted as civil servants. Nor do we include the UK’s diplomatic service in our analysis since it too is administratively separate from the UK Home Civil Service.

Defining the senior civil service

There are two definitions of the senior civil service. The annual Civil Service Statistics publication (the source of much of the data in this report) contains information on ‘SCS level’ employees. This measure includes “a number of health professionals, military personnel, and senior diplomats that are not part of the Senior Civil Service”. ¹³ Cabinet Office, ‘Quality and Methodology Information for Civil Service Statistics 2025’, GOV.UK, 29 October 2025, https://www.gov.uk/government/statistics/civil-service-statistics-2025/quality-and-methodology-information-for-civil-service-statistics-2025

By contrast, the Cabinet Office collects data on ‘the senior civil service’ in the SCS database. Because this is not publicly available, by necessity we use the data in the annual Civil Service Statistics publication and for simplicity refer to this as ‘the senior civil service’ throughout this report.

These two measures give different figures for the size of the senior civil service. The 2025 release of Civil Service Statistics recorded 7,775 individuals at ‘senior civil service level’ (7,525 FTE). While the SCS database is not publicly available, the government’s evidence to the Senior Salaries Review Body on SCS pay (published in October 2025) put the size of the senior civil service at 6,905 (6,715 FTE). ¹⁴ Civil Service, Senior Salaries Review Body, ‘Government evidence to the Senior Salaries Review Body on the pay of the Senior Civil Service (October 2025)’, 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

Defining departments

Where possible, we categorise bodies into ‘departmental groups’ according to where ministerial responsibility lies, even when these are reported under a separate departmental heading in government data. For example, we group Ofsted with the Department for Education departmental group, and not as a separate department, as it is reported in the original Office for National Statistics Public Sector Employment Data.

In such cases where source data reports organisations as independent from core departments, we have identified the departmental group to which those organisations belong by using the ‘sponsor department’ identified in the most recent public bodies report (for 2023/24) published by the Cabinet Office, ¹⁵ Cabinet Office, Public Bodies 2024, GOV.UK, 29 May 2025, www.gov.uk/government/publications/public-bodies-2024  or by government statements (such as on machinery of government changes) to parliament.

Unless otherwise indicated in the notes for a chart, the figures used throughout the report are for departmental groups, and therefore include the civil servants who work in the core departments as well as the agencies they oversee.

A table listing the departments and their associated agencies and non-ministerial departments is presented below. We have not included organisations that no longer exist – for example, because they have been merged with other bodies or renamed. However, historic organisations are counted in our figures for change over time, and details of those used in our analysis are available upon request. Where the names of departments differ between the ONS and Civil Service Statistics data releases we use the ONS classification – for example, the Export Credits Guarantee Department is listed as UK Export Finance by the Civil Service Statistics release.

These departments and organisations constitute the entire civil service, and they are all included, unless stated otherwise, in analysis that looks at the civil service as a whole. For analysis conducted at a department level, we tend to focus on the 17 major Whitehall departments – we do not include the AGO, OSSS, OSSW, NIO, or the Scottish or Welsh governments. We include HMRC, which is a non-ministerial department, because of its unique size (around 63,000 FTE staff), when we do not include other non-ministerial departments.

List of departments and associated organisations

Interpreting regression results

In several places, we use univariate regression analysis to identify relationships between variables. These regressions do not let us say that one variable causes another, but they do let us see if a high level in one variable tends to occur with a high or low level in another.

We report the R² for univariate ordinary least squares (OLS) regressions (those with only one independent variable). This is the proportion of the variation in the dependent variable that can be explained by variation in the independent variable. We interpret these as follows:

• Strong correlation: R² > 0.35
• Moderate correlation: 0.1 < R² <= 0.35
• Weak correlation: 0 < R² <= 0.1
• No correlation: R² = 0

We use asterisks to represent the level of significance of an independent variable, as determined by the p-value from our regressions. Our thresholds are as follows:

• *: Weakly significant, 0.01 <= p < 0.05
• **: Moderately significant, 0.001 <= p < 0.01
• ***: Highly significant, p < 0.001

Inflation/real-terms figures

For government spending information that spans multiple years, we use the GDP deflator to present figures consistently. The GDP deflator measures economy-wide inflation and so is an appropriate metric for tracking changes in government spending. We use the GDP deflator published alongside the autumn budget 2025.

For pay figures, we use the Consumer Price Index (CPI) to present numbers in consistent prices as this is the relevant measure to understand how much pay packets are worth to households. We adjust pay figures using the ONS’s CPI measure for October of the relevant year and used the CPI data last updated on 19 November 2025. October was chosen as the most relevant month to use in statistical analysis we have carried out of pay – discussed below – and has then been used for consistency in other analysis of pay.

Specific chapters

Size of the civil service

To analyse the overall size of the civil service, we use Table 9 of the ONS’s quarterly Public Sector Employment data release. This dataset contains staff numbers (headcount and FTE) in all public organisations employing civil servants. As in the rest of the report, unless otherwise stated we use FTE figures.

The ONS public sector data rounds figures to the nearest five, and report numbers less than five as ‘..’. We replace these suppressed figures with ‘3’.

Our figures exclude temporary Census field staff. Staff from Central Government Security (formerly Security and Intelligence Services) have been excluded from civil service statistics since Q2 2016. We adjust for this by manually excluding Central Government Security staff from our datasets before this date.

Professions and functions

Our analysis of the civil service professions uses Table 8A in the Cabinet Office’s annual Civil Service Statistics. The professions in the civil service have changed over time. For example, the ‘government digital and data’ profession used to be the ‘digital, data and technology’ profession. Where this has happened, we refer to the profession according to its latest name.

In this year’s Whitehall Monitor, we have grouped civil service professions according to the civil service groupings: ‘operational delivery’, ‘policy’, ‘functional professions’ and ‘specialist professions’. The official statistics record ‘Inspector of Education and Training’ as a profession, but we cannot categorise it because the government does not publish this information; as a result, we include it in the category ‘unknown or other’. It should be noted that the civil service relies on civil servants self-reporting their profession for this data.

1. Operational delivery
2. Policy
3. Functional professions
   • Actuary
   • Commercial
   • Communications
   • Counter fraud
   • Debt
   • Digital and data
   • Economics
   • Finance
   • Geography
   • Grants
   • Human resources
   • Internal audit
   • Legal
   • Project delivery
   • Property
   • Operational research
   • Risk management
   • Security
   • Social research
   • Statistics
4. Specialist professions
   • Corporate finance
   • Intelligence analysis
   • Knowledge and information management
   • Clinical
   • Occupational psychology
   • Planning
   • Planning inspection
   • Science and engineering
   • Tax
   • Veterinary

Professions data has historically been inconsistently reported. We have avoided correcting this, except for in one instance. The Department for Work and Pensions (DWP) did not report the professions of over 95% of its staff in 2018, 2020 and 2021. Given the size of the department, this would significantly skew the trend data, particularly for the operational delivery profession. For this reason, we interpolated DWP’s professions numbers for these years based on its reported numbers in 2017, 2019 and 2022.

Throughout our analysis, we combine the figures stated in the source data under ‘not reported’ and ‘other’ in the category ‘unknown or other’. In 2021, 56,789, or 13%, of civil servants were recorded in this category. By 2025 this had decreased to 27,345, or 5%. Some of the change appears to be driven by civil servants newly being classified as belonging to the operational delivery profession, but some of it also appears to be civil servants being classified as belonging to functional and other professions.

Turnover and leaving routes

Data on civil service staff turnover is derived from the annual Cabinet Office Civil Service Statistics dataset (Tables 20, 33 and 34). We use headcount rather than FTE for all staff turnover calculations. Some figures relate to core departments only and not the wider departmental group; where this is the case it is made clear in the chart notes.

External staff turnover is calculated as the number of civil servants who left the civil service entirely over the course of a given year, as a percentage of the average civil service headcount during that period. Average civil service headcount is calculated as the mean of civil service headcount at the beginning and end of the interval (for instance, headcount in March 2021 and March 2022 for the period 2021–22). We use average headcount to account for the fact that the number of civil servants changes over the course of the year. Internal staff turnover is calculated as the number of civil servants who transferred to another department over the course of a given year, as a percentage of the average civil service headcount during that period. Data on moves within a single department are not publicly available.

Civil Service Statistics lists 17 leaving causes. For ease of analysis we have grouped these into the following categories:

• Resignation
• Retirement
• Dismissal
• End of appointment
• Redundancies and voluntary exits
• Other

Consolidation of categories is as follows:

• Resignation
• Retirement
• Dismissal
• End of appointment: (which the dataset records as ‘End of casual, Period,
  Conditional or Provisional Appointment’)
• ‘Redundancies and voluntary exits’, which combines the following leaving causes reported in the data: ‘Voluntary Exit Scheme: With Payment’; ‘Voluntary Exit Scheme: With an Unreduced Pension’; ‘Voluntary Exit Scheme: Terms Not Recorded’; ‘Voluntary Redundancy Scheme: With Payment’; ‘Voluntary Redundancy Scheme: With An Unreduced Pension’; ‘Voluntary Redundancy Scheme: Terms Not Recorded’; and ‘Compulsory Redundancy Scheme’
• ‘Other’, which combines the following leaving causes reported in the data: ‘Death in Service’; ‘Transfer of Function to Private Sector’; ‘Secondment to Organisation External to Civil Service’; ‘Transfer to Non-Civil Service Public Sector’; and ‘Other Leaving Cause’.

In previous editions of Whitehall Monitor ‘compulsory redundancy’ has been recorded under ‘Other’. This year it is reported under ‘Redundancies and voluntary exits’ (previously named ‘Voluntary exit/redundancy’). This is in order to allow a more accurate comparison of previous exit schemes, including compulsory redundancy rounds. Civil service compulsory redundancy rounds must be preceded by voluntary redundancies, and departments apply for exit schemes that often encompass leavers under both voluntary and compulsory redundancies.

Pay

We refer to the results of statistical analysis of median salaries and pay and benefits theme scores from the Civil Service People Survey in this chapter. The pay and benefits theme consists of two questions on pay levels and one on the overall benefits package (pay, benefits and pension). See ‘Morale and engagement’, below, for further details of the People Survey.

Pay data is as at 31 March, while the People Survey is carried out in the autumn. We join People Survey data to pay data from the following calendar year (i.e. roughly six months later). For example, 2024 People Survey data is joined to 2025 pay data – these are the data releases in which effects of pay awards for 2024/25 are most likely to show.

We put pay data for all years into 2025 prices using October CPI figures. We use October figures to align with the timing of the People Survey, as people will be making judgments on the satisfaction with their pay and benefits without knowing what will happen to inflation in the rest of the financial year. This makes a material difference to the results – primarily because of the effect very high inflation at the end of 2021/22 following the full invasion of Ukraine by Russia in February 2022 has on 2022 pay figures if these are deflated using a CPI figure from the financial year end.

Morale data is not available by organisation and grade, therefore People Survey figures used relate to civil servants of all grades. We use mean scores, as described in the ‘Morale and engagement’ section below.

Regression 1: Pay and benefits theme scores compared to real-terms median salary since 2010

Using whole civil service figures from the People Survey and Civil Service Statistics and looking at the period 2010–24 (morale data)/2011–25 (pay data), we carried out a simple linear regression.

Results are shown in the figure showing civil servant satisfaction with pay and benefits compared to real-terms median salary, and are as follows:

See ‘Interpreting regression results’, above, for details of how we interpret R² values and p-values for ordinary least squares regressions.

Morale and engagement

The Civil Service People Survey is a survey carried out annually in autumn, which civil servants in more than 100 organisations are invited to take part in.

As well as aggregate responses to individual questions, a headline ‘employee engagement index’ is published, drawing on five questions:

• I am proud when I tell others I am part of [my organisation]
• I would recommend [my organisation] as a great place to work
• I feel a strong personal attachment to [my organisation]
• [My organisation] inspires me to do the best in my job
• [My organisation] motivates me to help it achieve its objectives.

It is the employee engagement index that we use when analysing headline civil service engagement.

People Survey results are also published for nine ‘themes’: my work; organisational objectives and purpose; my manager; my team; learning and development; inclusion and fair treatment; resources and workload; pay and benefits; and leadership and managing change.

In our People Survey analysis we focus on mean scores, to better reflect the performance of the civil service overall. This does mean that larger organisations count more than smaller organisations in the figures we use, unlike if we analysed median scores.

Regression 2: Civil servant engagement compared to People Survey theme scores for core departments for 2024

Using figures from the People Survey for the 17 core departments* that are the main focus of Whitehall Monitor (see ‘Throughout this report’, above) and looking at 2024, we carried out simple linear regression on employee engagement index scores and individual theme scores. There are statistically significant linear relationships between employee engagement index scores and certain theme scores only.

Partial results are shown in the figure showing civil service engagement compared to satisfaction with selected aspects of their working experience. The full results are as follows:

See ‘Interpreting regression results’, above, for details of how we interpret R² values and p-values for ordinary least squares regressions.

Where relationships are statistically significant for this single year, 2024, there are also statistically significant within-organisation relationships in a two-way fixed effects regression for the same organisations for the period 2010–24 (p [F-statistic (robust)] = 0.0000 in all cases). This included the fixed effects of time and department, but no other controls.

Diversity

The population benchmarks for female, disabled and minority ethnic staff are calculated using ONS’s Labour Force Survey (more details below). Falling response levels in recent years have led to concerns about the quality of ONS’s estimates. Short-term changes in benchmarks should be treated with caution.

Unless otherwise specified, the representation of a demographic characteristic in the civil service is calculated as a share of the number of staff for which that characteristic is known – i.e. excluding those for whom it is ‘not declared’ or ‘not reported’.

Population benchmarks

Female, disabled and minority ethnic staff

For the proportion of female staff in the civil service, the population benchmark is calculated from the ONS’s Labour Force Survey (Table A02). Since the Civil Service Statistics dataset captures the state of the civil service annually on 31 March, we make the benchmark as closely comparable as possible by dividing the number of economically active women aged 16–64 in the February–April quarter of every year by the whole economically active population in the same quarter.

The population benchmark for disabled individuals is calculated from Table A08 in ONS’s Labour Force Survey, using non-seasonally-adjusted figures. To get a figure that covers the 31 March date of Civil Service Statistics, the number of economically active people who meet the ‘harmonised standard definition of disability’ in both the January–March quarter and April–June quarter of every year is divided by the number of economically active people whose disability status is known (‘harmonised standard definition of disability’ and ‘people who do not meet the harmonised standard definition of disability’ combined) in the same quarter. The resulting two figures are averaged to obtain a benchmark.

The population benchmark for minority ethnic individuals is also calculated from the ONS’s Labour Force Survey (Table A09), using non-seasonally-adjusted figures. The number of economically active people belonging to ‘all other ethnic groups combined’ (i.e. not white) aged 16–64 in the January–March and April–June quarters of every year is divided by the number of economically active people whose ethnicity is known (‘all other ethnic groups combined’ and ‘white’ combined) in that same quarter. The two resulting figures are averaged to obtain the benchmark figure.

Sexual orientation

The population benchmark is calculated from the ONS’s Sexual Orientation, UK dataset. The number of people identifying ‘Gay or lesbian’, ‘Bisexual’ or ‘Other’ aged between 16 and 64 is divided by the number of people aged 16 and 64 whose sexual orientation is known (‘Gay or lesbian’, ‘Bisexual’, ‘Other’ and ‘Heterosexual or straight’ combined). The latest figures available are for 2023.

Socio-economic background

In the absence of robust workforce statistics, our analysis of civil servants’ socio-economic background uses the annual Civil Service People Survey. The Cabinet Office publishes the results of the People Survey by socio-economic background. This data breaks down responses by officials in the national statistics socio-economic classification categories ‘never worked’, ‘routine’, ‘intermediate’ and ‘high’, based on questions about the main income earner in their household when they were 14 years old. Under each of these categories, the data states the number of survey responses received from officials in each grade.

We have combined the number of survey responses received under the ‘never worked’ and ‘routine’ categories into the ‘low’ socio-economic background category, in line with the terms that the Social Mobility Commission employs, and used the resulting data to estimate the socio-economic background of officials in each grade.

Consultancy and temporary staff spending

For this chapter, we have analysed spending on consultancy and temporary staff (or ‘contingent labour’ or similar wording) reported in departments’ annual reports and accounts. Figures for such spending are often restated in subsequent annual reports, and we have used the latest figures available in all cases.

As mentioned in the text, different departments apply different definitions to these categories of spending, even if this is not clear from their annual reports. Some caution should therefore be exercised in directly comparing figures between departments.

Departments included in the analysis and treatment of machinery of government changes

We have analysed the consultancy and temporary staff spending of the current 17 major Whitehall departments (see the ‘Defining departments’ section of Methodology) and their predecessors. The sum of these departments’ spending is shown in the figures showing overall government spending.

Where departments have closed or merged, new ones opened or names changed, in most cases we have presented the data reported by each department in existence in each year. The two exceptions are MHCLG/DLUHC and the FCDO. The former underwent a simple name change, so for simplicity we have referred to the department as MHCLG throughout the period. The FCDO was formed from a simple merger, so data for 2018/19 is the sum of that reported by the FCO and DfID. The 2019/20 figures are taken from the FCDO’s first (2020/21) annual report.

Location

Unless otherwise stated, all figures in this chapter are based on headcount of civil servants, rather than FTE. While the government uses FTE figures in relocation reporting, the Institute uses headcount because we think relocation efforts are best understood by assessing the movement of individuals regardless of their working pattern.

Except for the figure showing civil servants by region (the first in the chapter), all analysis in this chapter excludes officials working for the Scottish and Welsh governments. This is because the UK government has no control over where they work. For the table showing civil servants by grade and location, in excluding officials working for the Scottish and Welsh governments we assume that the excluded officials are all based in those respective nations. Data limitations prevent us from doing otherwise.

Civil servants based overseas, as well as those whose location (and grade, where relevant) were not reported, have also been excluded from the analysis.

Proportion of the SCS based in London

The final figure in the chapter shows both the government and IfG measures for the proportion of the SCS based in London over time. The IfG measure is calculated using the above method – and to do so we use figures from the Cabinet Office’s annual Civil Service Statistics publication. The government’s measure uses FTE, is updated quarterly (though with a significant lag), and is a division of the total SCS based in London by the total number of SCS.

This figure also includes an extrapolation of both measures up to Q1 2030. To do this for each measure, we have taken the total change in the proportion of SCS based in London between 2020 and 2025 (for the IfG measure) and Q2 2020 and Q4 2024 (for the government measure) and divided it by the relevant number of years or quarters. We have then applied this average annual / quarterly change to the measure until we reach 2030.

Definition of relocation

The Cabinet Office included a definition of ‘relocation’ in the Places for Growth (PfG) Formative Evaluation Report, published in 2024. It set out that any of the following can count as a relocation:

• Civil servants that physically relocate along with their roles from London into a non-London region across the UK
• Internal civil service recruitment, in which a civil servant moves department or location into a non-London role that was previously based in London
• New joiners to the civil service that acquire a non-London role that was previously based in London
• Newly created roles that could have been recruited into London, but the successful candidate (internal or external) is recruited into a non-London region in the UK.

The report acknowledged that the above definition (particularly the fourth option) was applied and interpreted inconsistently by a small number of departments. It also said that the PfG programme had chosen to accommodate different reporting methodologies from different departments.

Arm’s length bodies

In this report we use the term arm’s length bodies (ALBs), a subset of public bodies. The classification of ALB includes non-departmental public bodies, executive agencies and non-ministerial departments. The wider term ‘public bodies’ includes public corporations, unclassified bodies and parliamentary bodies. Data on these is not included in the Cabinet Office data, which informs the charts and analysis presented here.

We use the Cabinet Office’s annual ‘Public Bodies’ reports from 2013–20 and 2023–24 to analyse the number of ALBs, their funding and their staffing. Data on the number of ALBs in 2021/22 and 2024/25 was directly provided to the Institute by the Cabinet Office. This dataset, which only includes ALBs, differs from the GOV.UK list of departments, agencies and public bodies, which has a broader scope and is not always an exact match. ¹⁶ GOV.UK, ‘Departments, agencies and public bodies’, GOV.UK, no date, www.gov.uk/government/organisations

For details on how population benchmarks for characteristics of public appointees are calculated, see the ‘Diversity’ subsection above. To calculate the population benchmark of individuals of working age (aged 16–64) living outside London and the South East, we used the mid-2024 edition of the ONS dataset ‘Estimates of the Population for the UK, England, Wales, Scotland, and Northern Ireland. The sum of the working-age populations of London and the South East isdivided by the working age-population of the UK to give the proportion of those living in London and the South East. We subtract this from 100% to get the proportion of working-age population living outside London and the South East.

Departmental spending and efficiencies

The IFS estimate of the 18% headcount reduction required to achieve the 16% administration budget savings by 2029/30 set out at spending review 2025 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.

To estimate the number of civil servants whose salaries are covered by administration budgets, we use the PESA 2025 figure for total administration paybill (£11.2bn) and divide this by estimated paybill costs for the median civil servant. Our estimate of paybill costs includes salary, 15% employer National Insurance contributions (above the secondary threshold of £5,004) and 28.97% employer pension contributions.

Using median salary (£35,680) for all civil servants, including those working in arm’s length bodies, gives an estimated 218,402 civil servants with salaries covered by administration budgets. However, civil servants in core departments are more likely to be covered by administration budgets than those in arm’s length bodies – and also have higher salaries on average. Since data for median salary for core department civil servants is only available by department, not overall, we use an approximation of the median salary of civil servants in the median department for pay (£49,055). This gives an estimated 158,854 civil servants with salaries covered by administration budgets. The true number of civil servants covered by administration budgets is likely to be between these two figures. Therefore an 18% headcount cut would be approximately equivalent to a reduction of between 29,000 and 40,000.

_{* Figures used relate to core departments, except for figures for the Department for Education, which relate to the DfE group, as department-only figures are not published.}