1. Main points

  • Total public service productivity grew for the seventh consecutive year in 2017, but growth of 0.3% was the joint slowest annual rate since 2010

  • Healthcare remains the service area with the strongest contribution to overall public service productivity in 2017, owing to its large share of total expenditure and positive growth of 0.9% compared with 2016

  • Total inputs grew by 0.7% and total output grew by 1.0% in 2017; this was principally driven by changes in inputs and output growth from healthcare

  • Three out of the four services areas with a quality adjustment saw an improvement in quality in 2017 (healthcare, education and adult social care), whereas public order and safety quality declined

  • Because of methodological improvements, productivity is estimated to have grown at a slightly slower rate over the time series compared with previous estimates

Back to table of contents

2. Total public service productivity

Productivity is the measure of how many units of output are produced from one unit of inputs, and is calculated by dividing total output by total inputs. Public service productivity grew for the seventh consecutive year in 2017, increasing by 0.3% from 2016. The index for productivity and the year-on-year growth rates are shown in Figure 1.

From 1997 up to and including 2009, average annual growth in total public service productivity was negative 0.1%. This period included years of both positive and negative growth in productivity. From 2010 to 2017, the average annual growth rate was 0.7%, reflecting more stable and positive growth. However, growth of 0.3% in 2017 was the weakest in the post-downturn period with the exception of 2015.

Nine different service areas are included in these estimates of productivity growth. Figure 2 shows the areas, listed in order of their share of total government expenditure in 2017. Healthcare is the largest, followed by education and then other government services (which comprises the variety of smaller services delivered such as recreation and housing). The different colours within each bar show the proportions of that service area that are directly or indirectly measured, and whether each of these are adjusted for quality – more detail can be found in Section 15. The preferred method is to use direct measures and quality adjustments.

Overall, 50.6% of total public services are adjusted for quality (the sum of the light blue and the dark blue bars in Figure 2). A further 11.9% of output is non-quality adjusted and directly measured (the yellow bars in Figure 2), and the remaining 37.5% is non-quality adjusted and indirectly measured (the mid-blue bars in Figure 2).

To further explore the increase in productivity from 2010 onwards, Figure 3 shows growth rates of inputs and output each year. It demonstrates that since 2010 there was a clear shift, with both inputs and outputs growth falling, but output now growing more rapidly than inputs, which was not the case prior to that point. This reflects a period of constrained government spending from 2010, in the wake of the economic downturn in 2008.

The reduction in government spending since the economic downturn in 2009 is commonly known as “austerity”. Government departments have been encouraged to make efficiency savings, and have also cut services in some cases. During the period of “austerity”, our statistics suggest an improvement in public service productivity, as output growth has exceeded input growth in every year.

Various research by think tanks and research bodies support the idea that productivity gains were made in the delivery of public services in the “austerity” period. In particular, the NHS and local government services have been highlighted as areas where innovative solutions have been found to funding constraints. For instance, the Police Foundation recognises the need for innovation in policing, and Localis reports the innovative efforts of local council. The Health Foundation provide evidence of innovation early in the “austerity” period.

However, many public services are labour intensive, and opportunities for innovation may be more limited than in the private sector1. While also labour intensive2, healthcare may be the service area for which innovation and productivity gains are most achievable through the use of specialist machines and drugs. There are also clear market competitors for healthcare, unlike many other service areas, which may drive productivity gains.

Most local government services are currently indirectly measured in our statistics, which means productivity changes cannot be observed. Even in directly measured service areas, there are substantial challenges in output measurement for public services. Increases in output (especially in the quality adjustment) may only be seen in years following increases in inputs, giving a mixed productivity picture. This makes drawing definitive conclusions on the relationship between “austerity” and innovation difficult. It should also be noted that there is necessarily a time lag between public spending and improved outcomes, and there will likely be a limit to public innovation in the interest of delivering core services and reducing risk.

Alternative productivity measures

The headline public service productivity figure uses the quality adjusted output series. It also includes the indirectly measured service areas. If the non-quality adjusted output series are used instead, productivity is estimated to have fallen by an average of 0.2% per year, rather than rising by an average of 0.2% per year if including the quality adjustments. This is shown in Figure 4.

Indirectly measured service areas, which are included in the headline estimate of productivity, have a “dampening” effect – pulling the aggregate towards zero growth. If service areas measured in this way are excluded, this also affects the trend, slightly increasing the average annual growth since 1997. This is also shown in Figure 4. Note that this series is based on the quality adjusted output series.

The non-quality adjusted output series generally follows the quality adjusted series but its growth rates are always lower, whether the growth is positive or negative (although in some years it is close, such as 2016). This divergence is particularly noticeable from 2002. The difference between the non-quality adjusted output series and the quality adjusted output series is because of the positive growth of the quality adjustment during this period.

Notes for: Total public service productivity

  1. The labour share of income in the whole economy is typically taken to be around two-thirds, with the remaining third because of capital (ignoring intermediate inputs). For public services, if ignoring intermediate inputs, labour accounts for over 80% and capital less than 20%, as currently measured. This may be partially because of the accounting conventions with regard to non-market sector capital, which prevent a return on capital being recorded – this makes capital measurement inconsistent between the market and non-market sectors. If the capital contribution to inputs in public services were in reality twice as large as currently measured, this would make the labour share in public services around 75% – still considerably higher than in the rest of the economy. If including intermediate inputs, public services still appear more labour intensive than the rest of the economy. While intermediate consumption makes up nearly half of the value of gross output in the economy as a whole, it accounts for only 42% of total inputs for public services in 2017. On this basis, labour inputs account for around half of total public service inputs, but make up only around a third of input costs in the economy as a whole.

  2. The only public service areas with labour shares comparable to the market sector are defence (which relies heavily on military equipment which is classified as capital in the national accounts), and other government services (which includes the value of public infrastructure, as well as a large amount of office-related capital equipment). Healthcare is labour intensive, but also uses large amounts of intermediate goods and services such as drugs and medical supplies.

Back to table of contents

3. Contributions to total public service productivity

Four of the service areas (healthcare, education, adult social care, and public order and safety) are adjusted for quality, the preferred method, and consequently the productivity estimates for these service areas are more reflective of their respective outcomes than the other five areas. The productivity indices for the four quality adjusted areas are shown in Figure 5.

Since 2004, healthcare is the only service area that has consistently outperformed total public service productivity, and is the primary reason for the increases in the aggregate since 2010 (due also to its large weight). Education productivity tends to be comparatively flat, and tracks the aggregate reasonably consistently. Adult social care (ASC) and public order and safety (POS) have consistently drawn total public service productivity downwards.

Another way to illustrate the concept of service area trends and how they affect the total is using contributions to growth (Figure 6). This takes the productivity growth for each service area and weights it by its expenditure share each year. For example, in Figure 5 POS appears to significantly drag total productivity downwards. However, POS has a small expenditure share (3.3% in 2017) and as such contributes less to the total productivity estimate than Figure 5 may initially suggest. Three service areas (police, defence and other government services) have been excluded from Figure 6 as they are measured using the inputs-equals-output convention and consequently they have no impact on estimates of productivity growth.

Despite POS’ smaller expenditure share, it still does noticeably affect the total productivity estimate. Over the time series, its average contribution is negative 0.1 percentage points. Healthcare’s average contribution is 0.3 percentage points. Education, despite its large expenditure share, has a flatter productivity trend and as such its average contribution is only marginally positive, although there are clear larger movements in some years.

Back to table of contents

4. Total public service inputs

Unlike other measures of productivity produced by the Office for National Statistics (ONS), public service productivity includes goods and services, as well as labour and capital, as inputs. This is necessitated by the fact that public service output measures are gross output (total output) measures, rather than value added measures as used in labour productivity and multi-factor productivity.

Figure 7 shows the growth for total public service inputs, broken down by the type of input. Total inputs grew 0.7% in 2017, after falling by 0.1% in 2016. This is in line with the average annual growth rate of inputs since 2010, and therefore continues the period of historically weak growth in inputs as a result of constrained government spending in the post-downturn period.


Labour inputs are the volume of labour used to provide public services, covering all workers that contribute to producing public services, such as teachers, doctors, and administrative staff. We previously published analysis of the types of workers who work in the public sector.

Most public services are labour intensive, and labour is the largest input to public services by weight, at around 48% in recent years. Labour inputs have grown relatively slowly over time, at an average rate of 1.3% per year. Growth since 2010 has been more modest, but growth in 2017 of 1.8% is the fastest since 2009.

Intermediate inputs

Intermediate inputs cover all goods and services purchased in the course of delivering public services, except fixed capital assets (which last for more than a year). This includes utilities, energy, professional services and medical supplies, among others. The volume of this input is calculated by removing price growth from expenditure growth using appropriate price indices (known as deflators).

Unlike labour, these inputs have grown rapidly over time, and accounted for 41% of total inputs in 2017, up from 34% in 1997. This partially reflects changing methods of delivering public services: more public services have been delivered by the private sector over time, and the costs of this are recorded as intermediate inputs. In the case of healthcare, this trend has reversed somewhat in recent years, with NHS “bank” staff (recorded as labour inputs) replacing agency staff (recorded as intermediate inputs) – see the separate healthcare publication for details. Total intermediate inputs fell by 0.7% in 2017, after falling by 1.4% in 2016, but was offset by the stronger than usual growth in labour inputs.


Capital inputs reflect the use of fixed capital assets which last more than a year, such as buildings, equipment, computers and software. Most public services are not very capital intensive, and capital has a small weight over of around 10%. However, this masks considerable variation by service area: in defence, it accounts for over 20%, but in adult social care and children’s social care only around 1%.

Improvements in the measurement of capital consumption in the National Accounts (described in Section 13) have changed this significantly in this release. Capital is now estimated to have grown in every year between 1997 and 2017, at an average rate of 2.9%. This has slowed to 2.0% in years since 2010, as a result of weaker capital investment. The growth of 1.4% in 2017 continues this weaker trend.

A driver of productivity growth is the amount of capital inputs available per unit of labour input, as workers that have more capital at their disposal can be expected to be more productive. Figure 8 shows the ratio of capital to labour inputs in total public services, healthcare, and the market sector (taken from the ONS multi-factor productivity1 release) as indices relative to 1997.

Capital per worker fell (known as capital shallowing) in public services between 1997 and 2005, and by 2010 was barely above the level in 1997. However, since then it has increased substantially (known as capital deepening), mainly as a result of a decline in labour inputs rather than a faster growth in capital. For the market sector the trend is the reverse; capital deepening (more capital per worker) until 2009, shortly after the economic downturn in 2008, and capital shallowing (less capital per worker) since. This may help to explain the relatively stronger productivity performance seen in public services since 2010 as compared to the weak productivity growth (known as the “productivity puzzle”) in the rest of the economy.

The trend in capital to labour input for healthcare is similar to the rest of the market sector in recent years. In spite of this, healthcare productivity has been the strongest performing of public service areas, as shown in Figure 5. The Health Foundation also identified this and similar trends for healthcare – the UK lags international comparators on capital investment, capital stocks and capital per worker in healthcare.

Inputs growth can also be decomposed by service area, as shown in Figure 9. The growth in 2017 has been mainly driven by healthcare, in line with recent years.

Notes for: Total public service inputs

  1. The measures in public services and market sector differ slightly, but are a reasonable approximation.
Back to table of contents

5. Total public service output and quality change

Figure 10 shows the contributions from the nine service areas (with police, defence and other government services combined into the indirectly measured category) to total output growth. From 2016 to 2017, total public service output grew by 1.0%.

Healthcare is a noticeable positive contributor to output growth throughout the time series, even in later years where total output grows more slowly. Public order and safety (POS) has contributed negatively in every year since 2010, and the indirectly measured areas have also contributed negatively in five of the years since 2010. Education contributes positively through most of the series, but on a small scale considering its relatively large expenditure share.

The growth of quality adjusted (QA) output can be split into contributions from non-quality adjusted (NQA) output and the quality adjustment. This is illustrated in Figure 11.

Most of the growth in quality adjusted output is because of growth in non-quality adjusted output, particularly after the economic downturn. However, since quality improvements are cumulative, small increases over a long period help to explain the large divergence between the quality adjusted and non-quality adjusted productivity series in Figure 4.

Figure 12 shows the contribution to the growth of the quality adjustment by service area.

From 1998 to 2010, the education adjustment contributes positively to the total adjustment, before continuing a positive contribution on a smaller scale in more recent years. More details on this are in Section 7. The POS adjustment has contributed negatively for the past eight years whereas ASC has contributed positively every year since it was implemented in 2011.

Back to table of contents

6. Healthcare

Healthcare is the largest service area included in these statistics. Output is estimated for four sectors, which are then weighted together to form the overall output index. These four sectors are Hospital and Community Health Services, Family Health Services, GP prescribing, and non-NHS provision. This last sector illustrates the importance of the definition of these estimates as “public services”, not “public sector” – non-NHS provision includes services that may not be delivered by the public sector, but they are paid for by Government. Of all healthcare output, around 79% is adjusted for quality. More information is available in Public service productivity estimates: healthcare QMI.

Healthcare contributes positively to total public service productivity growth consistently over the time series, although it does make notable negative contributions in 2002 and 2009. In healthcare, inputs and output have both grown rapidly over the past twenty years, reflecting increased demands on the healthcare system.

Productivity growth has fluctuated over the period, but has been more consistently positive since the economic downturn. Output has consistently grown at a faster pace than inputs, however the growth rates of both have slowed since 2010. In 2017, output grew by 2.4% and inputs by 1.5%.

More details are available in the accompanying healthcare productivity article. Some important improvements have been made to the methodology for this year’s set of estimates, which are explained in Methodological developments to public service productivity: healthcare, 2020 update.

Back to table of contents

7. Education

The education service area is the second largest by expenditure share in 2017. Output is measured by the number of students in the educational system, at various stages, and is adjusted for quality using attainment measures. It includes some nursery places (if funded by the government), schools, further education, Initial Teacher Training and some health courses at university (again, only those funded by the government). Schools includes students at primary, secondary and special schools, including academies in England, the number of which have greatly increased over the past decade as local authority-maintained schools transition to become academies.

Inputs and output indices tend to increase over time, as shown in Figure 14. However, output growth fluctuates less year to year than inputs growth.

In 2017, productivity was unchanged compared with 2016, because inputs and output both grew by 1.1%.

The methodology for education productivity has been updated, with more detail available in Improved methods for total public service productivity: total, UK, 2017. These improvements concern the quality adjustment. Attainment at GCSE level or equivalent is used as an indicator of the quality of the education system. New data sources to improve consistency and coverage over all four countries of the UK and a new “cohort split” method have been developed and used for the first time in these annual estimates.

The effect of these changes is demonstrated in Figure 15, which shows the non-quality adjusted and the quality adjusted series as published in last year’s article and as published in this article.

The “cohort split” method has applied attainment data to the five years of secondary school, rather than solely the last year. This better reflects the accumulation of knowledge and skills for learning as pupils progress through the education system. As a result, the quality adjusted series is smoother than previously, and improvements in exam results show as gains in quality earlier in the series, attributable to improved teaching in earlier years. New data sources have also improved the series.

The effect of these changes can also be seen in Figure 16, which shows the growth rate of the education quality adjustment. The steeper peak in 2009 and trough in 2013 and 2014 are significantly reduced.

Education output is measured as the number of students in the education system, adjusted for attendance rates. For schools, this is closely related to the birth rate in the UK. An increase in births tends to increase the number of children starting school some four or five years later, and as these children move through the school system, output increases. Accordingly, the non-quality adjusted output trend tracks the birth rate with a time lag. There are differences between the two, though, because of factors such as private education, children migrating in and out of the UK, and home schooling. The birth rate is a factor that schools cannot easily control, whereas the quality of education is likely more responsive to school actions. However, it could be that at times of rising pupil numbers, especially when this translates to rising class sizes (a decreased pupil to teacher ratio), schools are less able to improve the quality of education.

Back to table of contents

8. Adult social care

Adult social care (ASC) output is partially directly measured, and both indirect and direct measures of output have a quality adjustment applied. Activities that contribute to ASC output include residential and nursing care and home care. The quality adjustment covers the time series from 2011 onwards and considers the improvements to the quality of life for the recipients of the service, accounting for external factors that affect this. More information is available in Public service productivity: adult social care, sources and methods, 2019 update.

ASC productivity has been decreasing over the time series, although in some years it does show positive growth. In 2017, inputs grew by 0.6% while output was unchanged compared with 2016. Accordingly, productivity fell by 0.6%. Over the entire time series, the average annual growth rate of productivity is negative 0.5%.

Figure 12 shows that quality in ASC has grown every year since 2011, the year in which the quality adjustment begins. This contributes positively to the growth of quality adjusted output. However, as shown in Figure 17, inputs have increased at a faster pace than output, causing productivity to fall (albeit at a slower rate than pre-2011).

This release incorporates methodological improvements first used in the adult social care article in February 2019. Figures for England up to financial year ending 2019 will be published in a separate ASC productivity article, to be published on 5 February 2020. The previous version, for England up to financial year ending 2018, is also available.

Back to table of contents

9. Public order and safety

Public order and safety (POS) includes courts, probation services, the prison service and the fire and rescue services. The police service is treated as a separate service area within this publication.

Output is measured on an activity basis using data on fire service incidents, fire prevention work, the volume of work in the courts system and the prison population. Output is also, where appropriate, quality adjusted using data on recidivism, prison safety, custody escapes and courts timeliness1. More details on the quality adjustment are in Quality adjustment of public service public order and safety output: current method.

POS saw productivity decrease by 6.4% in 2017. Inputs increased in 2017 for the first time since 2010 while output continued to fall. The increase in inputs was primarily driven by increases in intermediate consumption for courts and prisons. The fall in output was mostly driven by falling quantity output for courts, and some worsening quality in courts and prisons because of higher re-offending rates. Analysis by the Institute for Government on prisons and criminal courts support these findings of falling inputs in recent years and reduced quantity of output, with service quality also falling according to their measures – see their 2019 Performance Tracker for criminal courts and prisons.

The increase in inputs in 2017 may reflect government policies to improve prison safety, as a result of rises in prison violence in the preceding few years. Public service productivity estimates may be viewed as a long-run indicator of change, and increases in inputs in 2017 may lead to increases in quality adjusted output in subsequent years, which are not yet visible in our statistics.

Overall quality continued to decline within POS, however the fall in quality was less pronounced than the general trend since 2009. While re-offending rates have worsened, prison safety and timeliness in Crown and Magistrates Courts have improved. Quality adjusted productivity and non-quality adjusted productivity have also returned to broadly comparable levels, with quality improvements within POS before 2009 offset by declining quality since 2009.

Notes for: Public order and safety

  1. Note, the fire service and county courts are not quality adjusted.
Back to table of contents

10. Children’s social care

Children’s social care includes, in its output measures, looked after children, children in need, Sure Start schemes, adoption and other activities.

Currently around two-thirds of children’s social care is indirectly measured – that is, the “output-equals-inputs” convention is used. The remaining third of output is directly measured, although no quality adjustment is applied1. This direct measurement is of looked after children only. Therefore, inputs and output growth follow the same general trend over the time series. The directly measured proportion of output shows more subdued growth than inputs, so total output grew at a slower pace. The similarity of the inputs and output growth rates leads to a fairly flat trend in long-term productivity growth, especially from 2005 onwards.

The fall in productivity from 2016 to 2017 was driven by inputs increasing at a faster rate than output, with growth rates of 1.0% and 0.5% respectively. The service area is a negative contributor to total public service productivity growth, but its small share of total expenditure (2.8% in 2017) means that it has little effect compared with healthcare and education.

Notes for: Children’s social care

  1. We intend to develop our measures of children’s social care in the coming year to provide improved insights for this service area in future releases.
Back to table of contents

11. Social security administration

Social security administration (SSA) covers activities related to the administration of social security, such as unemployment benefits. This estimate of productivity covers the work of various departments, including the Department of Work and Pensions (DWP), HM Revenue and Customs (HMRC) and others.

As demonstrated in Figure 21, productivity over time for SSA fluctuates, more so because of changes in inputs than changes in output. From 2016 to 2017, productivity increased by 9.6%. Output rose by a comparatively small 0.3%, so the productivity growth was primarily driven by a fall in inputs of 8.5%.

From 1997 to 2017, the average annual growth rate for SSA productivity is negative 2.5%. This is the lowest of the services areas with the exception of public order and safety, which is negative 3.4%.

SSA can be significantly affected by events in the wider economy. For example, output grew in 2008 and 2009, coinciding with the economic downturn. In 2007, the UK’s unemployment rate was 5.3%, whereas in 2009 it had grown to 7.6% (although it would not reach its peak of 8.1% in this century until 2011). The greater number of unemployed people likely necessitated more social security claims.

Back to table of contents

12. Indirectly measured service areas

The three indirectly measured service areas are those where output is assumed equal to inputs, and therefore productivity growth will always be equal to zero. Inputs (and thus output) for all three areas – police, defence, and other government services – are shown in Figure 22.

Police includes such activities as incident response, crime investigation and neighbourhood policing. Defence is the administration, supervision and operation of military affairs and land, sea, air and space defence forces. Other government services covers spending in many different areas, including general government administration, environmental protection, housing, and recreation.

From 2017 to 2016, inputs and output fell in all three services areas: by 0.6% in police, 0.4% in defence and 0.3% in other government services. Since the economic downturn, negative growth has been particularly noticeable in other government services.

Defence inputs have been revised upwards from the 2016 estimates of public service productivity primarily as a result of the changes to capital. Police inputs have been revised this year as we now estimate the volume of local government labour directly from data on full-time equivalent employees (FTEs) and relative salaries for different groups (as explained in Improved methods for total public service productivity: total, UK, 2017). The effect of this has been to increase police inputs growth in the mid-2000s, shifting the overall police inputs series upwards.

Back to table of contents

13. Methodology changes in this release

The updates to the methodology used for the first time to estimate total public service productivity in this article were pre-announced, as set out in the Code of Practice for Statistics, in Improved methods for total public service productivity: total, UK, 2017 in November 2019.

The methodological improvements to Education were discussed in Section 7 and the impacts of the changes illustrated in Figures 15 and 16. Other methods improvements were considerably smaller.

Some significant data source changes have been introduced this year. To reflect the latest improvements in the UK National Accounts, particularly on the measurement of capital consumption, we have used government expenditure data consistent with Blue Book 2019. Previously, we had used a Public Sector Finances dataset known as European System of Accounts Table 11 (General government annual expenditure), which was consistent with Blue Book 2017 for the 2016 publication. As a result, our data in this release take on improvements in two Blue Books, and in some places are significantly revised compared to last year’s estimates.

One of the largest data changes was as a result of the improvements to measures of capital consumption in Blue Book 2019. A review was undertaken of asset lives for capital consumption, which in the majority of cases found them to be shorter. This necessitates the faster replacement of these assets, which in turn increases estimates of capital consumption. The Office for National Statistics (ONS) has also improved its method for estimating government purchases of goods and services, for which the Government does not pay Value Added Tax. However, to maintain consistency with private sector purchases in the UK National Accounts, the cost of paying the tax is estimated. The ONS has widened the scope of this method.

These updates have increased the consistency and accuracy of inputs data. These and other methodological changes to Blue Book 2019 are detailed in an impact article.

The effect of these changes was in part explained in Section 5 in relation to capital inputs. For an overarching view, Figure 23 shows the impact of the revisions on the total public service productivity index. The two indices (the headline estimates from this article and last year’s article) track fairly closely until 2009. From there, the new estimate is lower, although stronger year on year growth from 2012 to 2015 means that the gap between the two indices narrows.

The contributions to the revisions by service area are shown in Figure 24. Education makes some significant contributions to revisions, negatively in 2009 to 2011 and then positively in 2012 to 2014. This mirrors the new quality adjustment in relation to the previous adjustment for education. Healthcare also makes some notable downwards contributions, including 2016 which was due partly to the introduction of NHS bank staff into inputs, a staff group that grew quickly in 2016. There have also been numerous changes to year-on-year growth rates for healthcare because of the introduction of a “days adjustment” to account for the effect of leap years and changes in the annual number of working days on output. For more details see Methodological developments to public service productivity: healthcare, 2020 update.

In public order and safety, we now use separate deflators for the prison service and the fire service that better reflect the composition of staff and associated wage inflation within these areas. Police inputs measurement has also been improved, with local government labour inputs now measured directly, using FTE data, instead of indirectly.

Back to table of contents

14. Glossary

Public services

These are services delivered by or paid for by government (central or local). If paid for by the government, they may be delivered by a private body – for example, the provision of nursery places by the private sector, where these places were funded by the government.

Direct output measurement

Using a cost-weighted activity index to estimate the quantity of a service provided, such as the number of students in state schools, adjusted for attendance to produce an estimate of total hours of schooling delivered each year. Differs from indirect output measurement, where output is assumed equal to inputs.

Quality adjustment

A statistical estimate of the change in the quality of a public service, using an appropriate metric, such as safety in prisons as part of the public order and safety adjustment.


The Classification of the Functions of Government (COFOG), which is the structure used to classify government activities. It is defined by the United Nations Statistics Division.

Service area

The way we refer to the breakdown of public services into nine areas, closely following COFOG.

Intermediate inputs

Also referred to as “goods and services”, or “intermediate consumption” (the UK National Accounts term). Intermediate inputs include goods and services used up in the provision of a public service, such as utilities, energy, professional services and medical supplies, among others.


A price index used to remove inflation effects from estimates of expenditure to provide a volume estimate. Tailored wherever possible to maximise relevance and accuracy for the expenditure being measured.

Back to table of contents

15. Data sources and methodology

Productivity is the measure of how many units of output are produced from one unit of inputs, and is calculated by dividing total output by total inputs.

Inputs can be broken down into three components. They are labour (which can either be measured directly through means such as number of staff or indirectly by measuring service area expenditure on staff), intermediate consumption (expenditure on goods and services) and consumption of fixed capital. These inputs, as appropriate, are adjusted for inflation using a suitable deflator.

We measure output either directly or indirectly. If output is measured directly, this means that we have some measurable indicator for the amount of service provided. For example, direct output measurement for schools uses data on the number of students, adjusted for attendance. If output is measured indirectly, as it is for other government services (a service area that includes various smaller activities such as spending on housing and recreation), police and defence, then output is assumed to equal inputs. Therefore, there is no productivity change in these areas as the ratio of inputs to output is always one.

To account for changes in the quality of the service provided, we also use adjustments for quality in some service areas. More information on quality adjustment is available in A guide to quality adjustment in public service productivity measures. It is preferable to use direct measures of output and incorporate an adjustment for quality. Work to develop these methods for different service areas is ongoing and follows the principles of the Atkinson Review (PDF, 1.08MB).

The range of possible output measures are summarised in Table 1.

Growth rates of output and inputs for individual service areas are aggregated by their relative share of total government expenditure (expenditure weight) to produce estimates of total public service output, inputs and productivity. Service areas are defined by Classification of the Functions of Government (COFOG) rather than administrative departments or devolved administrations. As a result, estimates presented cannot be taken as direct estimates of departmental productivity. For example, estimates of social security administration productivity are not the same as an estimate of Department for Work and Pensions productivity. Lastly, it should be noted that these estimates do not measure, for example, the value for money in public services, or the true effectiveness of the services (quality adjustment includes some measurement of this but coverage is limited).

As well as estimates of public service productivity, the Office for National Statistics (ONS) also produces estimates of labour productivity and multi-factor productivity. Estimates of public service productivity cannot be compared to these other measures, because of differences in how we measure output and the scope of inputs. As explained in our blog post, public services are provided free at the point of use and often have no market price. Different approaches are therefore needed to measure public service output than those used to measure output in the private sector. We also measure inputs differently to labour productivity where the only input is labour, whereas we also include capital and intermediate inputs.

Estimates of public service productivity are published each year, and on a calendar year basis for consistency with the UK National Accounts. There is a two-year time lag associated with the estimates, because of the timeliness of our data, which come from administrative sources. A timelier estimate is available in the Public service productivity: quarterly, UK statistics, which have a two-quarter time lag. However, the quarterly estimates are experimental, whereas the annual estimates are badged as National Statistics. This means that they meet certain quality criteria, listed in the Code of Practice from the UK Statistics Authority.

Estimates in this bulletin cover the UK and, where possible, are based on data for England, Scotland, Wales and Northern Ireland. Where data are not available for all four countries, the assumption is made that the available data are representative of the UK.

More detail on methodology for these statistics is provided in Public service productivity: total, UK QMI (where QMI is Quality and Methodology Information) and Sources and Methods for Public Service Productivity Estimates: Total Public Services (PDF, 111KB).

The data used in this article is available to download. Data used in the construction for each chart is listed as a download under each one, alongside an option to download the chart as an image.

Back to table of contents

16. Authors and acknowledgements

Josh Martin, Leah Harris, Sophie Barrand and Jon Gardner, Office for National Statistics

We are grateful to colleagues in various government departments for making their data available for the compilation of these statistics, and providing helpful comments.

Back to table of contents