Guidance

The employment of disabled people: background information and methodology

Updated 26 October 2023

Context of the statistics

In 2017, the government set a goal to see a million more disabled people in work by 2027. This goal was reached in 2022. The command paper, Improving Lives: the future of work, health and disability, set out the government’s intention to monitor progress on the goal by tracking the number of working aged disabled people in employment, as recorded by the Labour Force Survey. It was recognised that a range of statistical information is needed to give a rounded picture of employment trends among disabled people, particularly as disability status and employment status are affected by each other and may change concurrently.

The statistics also provide context for the government’s commitment to reduce the gap between the employment rates of disabled and non-disabled people (known as the disability employment gap). The government reaffirmed this commitment in the Health and Disability White Paper published alongside the Spring Budget 2023 and stated their intention to set a new disability employment ambition.

Purpose of the statistics

This employment of disabled people publication describes trends over time in employment of disabled people of working age. It looks at the characteristics of people in employment, and flows into and out of work, comparing the experiences of disabled people with non-disabled people.

These statistics allow external stakeholders to examine employment circumstances and trends among working age disabled people in more detail than has previously been publicly available.

The statistics include:

• overall trends in employment of disabled people compared with non-disabled people, as measured by:

  • numbers in work
  • employment rates
  • the disability employment gap (the difference between the employment rate of disabled people and that of non-disabled people)

• estimates of the factors underlying change in in the number of disabled people in work under different hypothetical scenarios – the components of change

• employment rates for disabled people and non-disabled people according to several individual and work-related characteristics – employment characteristics

• characteristics of people moving into or out of work by disability status

These statistics use and build upon statistics published by the Office National Statistics on the employment of disabled people, as part of its ‘Labour Market Overview: UK’ release (Table A08), which present overall trends. It uses complementary data sources which enable patterns to be validated, and a broad range of statistics to be presented.

Source of the statistics

The statistics in the report are based on two key data sources:

• Labour Force Survey
• Annual Population Survey

The Labour Force Survey

The Labour Force Survey (LFS) is a large-scale household survey produced by the Office for National Statistics (ONS) designed to enable monitoring of labour market issues. The sample consists of approximately 40,000 responding UK households and 100,000 individuals per quarter. Respondents are interviewed for 5 successive waves at three-monthly intervals with 20% of the sample being replaced every quarter. As used in ONS’ Labour Market Overview release, the LFS is the source recommended for employment-related statistics, such as estimates of the number of people in employment, and is the key source for trend data on different measures of disability employment.

The achieved sample size for the UK Labour Force Survey (LFS) during January to March 2023 (JM23) was 57,876 individuals in 26,712 households. Please note that there were no NHS households in this period. Compared with the previous quarter October to December 2022 (OD22) this represents a decrease of 7.9% in achieved person interviews and decrease of 7.5% in household interviews.

The Office for National Statistics (ONS) are transforming the Labour Force Survey to increase the sample size (making it more representative of the population as a whole) and improve the methods for collecting data, including the flexibility to quickly change the questions so that they reflect the key needs of the day. The survey will move to an on-line first approach, supported by telephone collection and ‘knock to nudge’. This will provide more robust insights on the detailed characteristics of those within and outside work. The latest update on progress can be found at: Labour market transformation – update on progress and plans – Office for National Statistics (ons.gov.uk)

The Annual Population Survey

The Annual Population Survey (APS) is a continuous household survey, covering the UK. The APS is produced by ONS. It is not a stand-alone survey; it uses data combined from two waves of the main Labour Force Survey (LFS), alongside a local sample boost. The APS is a recommended source for employment statistics for smaller groups of the population.

The two-year longitudinal Annual Population Survey (APS) has been used to examine movements into and out of work. Each individual in the data is interviewed at two time points, one year apart. This allows us to see the number of people changing their employment status over time.

Read more information on both the APS and LFS.

Analysing components of change

The report provides estimates of changes in the number of disabled people under different hypothetical scenarios. The number of disabled people in employment can be considered as the product of the number of disabled people in the working age population and the proportion of those who are in work (their employment rate). In turn, each of these could be broken down further into another two components:

1. The disabled population is the product of:

• the total working age population, and

• the proportion who are disabled (the disability prevalence rate)

2. The employment rate of disabled people is influenced by both:

• general labour market trends – as measured by either the overall employment rate or the non-disabled employment rate

• how labour market trends differ for disabled people – as measured by the employment rate gap between disabled people and non-disabled people

To understand how each of these components might affect the overall number of disabled people in employment, 14 hypothetical scenarios were created, designed to shed more light on this issue, alongside examining actual changes in all four components. Each scenario is based on holding certain components constant at the level or rate observed in April to June 2013 (the baseline for this analysis), while allowing other components to increase in exactly the same way they did in the observed data. This means that the numbers presented are not estimates of real-world outcomes.

Interaction effects are defined as the additional impact on the number of disabled people in work, caused by two or more components of change, over and above the impact of changing each component one at a time.

The 15 different scenarios respectively consider the following.

Change in one component

1. Disability prevalence only

2. Employment rate gap only

3. Non-disabled employment rate only

4. Working age population numbers only

Interactions between 2 components of change

5. Both disability prevalence and employment rate gap

6. Both disability prevalence and non-disabled employment rate

7. Both disability prevalence and working age population numbers

8. Both employment rate gap and non-disabled employment

9. Both employment rate gap and working age population numbers

10. Both non-disabled employment rate and working age population numbers

Interactions between 3 components of change

11. Disability prevalence, non-disabled employment rate and employment rate gap

12. Disability prevalence, non-disabled employment rate and working age population

13. Disability prevalence, employment rate gap, and working age population

14. Non-disabled employment rate, employment rate gap, and working age population

Actual changes in all 4 components

15. Disability prevalence, non-disabled employment rate and employment rate gap, working age population numbers.

Scenerio 1

For example, scenario 1 models what happens if the disability prevalence rate were to increase to the April to June 2023 level of 23.2%, while other components remained at their April to June 2013 levels. Excluding people whose disability status in unknown, in this scenario we have the following:

40.2 million working age population (at April to June 2013 level)

multiplied by 23.2% disability prevalence (at April to June 2023 level)

equals 9.3 million working age disabled people (a);

and

76.8% non-disabled employment rate (at April to June 2013 level)

minus 33.1 percentage points employment rate gap (at April to June 2013 level)

equals 43.6% disabled employment rate (b).

Multiplying (a) and (b) results in a hypothetical 4.1 million disabled people in work in April to June 2023, around 1.2 million more than in April to June 2013, as a result of the observed increases in the disability prevalence rate alone.

Considering interactions, for example in scenario 5, the impact of the interaction between the disability prevalence rate and the employment rate gap is calculated from the combined impact of those two components minus the impact of prevalence alone (scenario 1) and impact of the gap alone (scenario 2).

Interaction effects between the non-disabled employment rate and the employment rate gap are zero, because these 2 components operate independently of one another in this analysis.

The scenarios give a broad idea of the scale of impact from each component. They should not be interpreted as what would have happened in practice, not least because not all interactions between components of change will be captured by this sort of analysis. Nor do the results measure a direct causal relationship between the components of change and the number of disabled people in employment. For example, an increase in disability prevalence could impact on the disability employment gap, even without any changes in employment, if prevalence increased faster among those already in employment than among those out of work.

This analysis is based on annual rather than quarterly intervals in order to avoid sampling and seasonal variations associated with short-term changes. It uses a Q2 (April to June) baseline so that future releases can include the latest data. Shorter-term comparisons would be more susceptible to sampling error and short-term movements.

Examining movements into and out of employment

The method used to examine movements into and out of employment by disability status, follows that used in previously published statistics, in Annex D of Improving Lives: the future of work, health and disability. It uses the two-wave longitudinal APS, where each individual is interviewed twice, one-year apart. Individuals aged 16 to 64 in both years were selected and their disability and employment status in both of the years were recorded.

For all further breakdowns in the flows of employment, such as age and gender, this analysis registers the disability and employment status in the first interview and the employment status in the second interview, one year later. In other words, it does not measure changes in disability status, but purely changes in the employment status of people who started off disabled or started off not disabled. However, as seen in Table FLW001, analysis has been done on the overall flows of employment to include the disability status in the second interview, to help assess the impact of changes in disability status over time.

The analysis includes yearly data over a multi-year period, with averages shown to improve the robustness of the analysis, allowing trends to be examined over time.

This analysis gives an indication of the relative rates at which disabled people and non-disabled people move in or out of work from one year to the next and there are several caveats to consider, noted below in the section on limitations.

Limitations of the statistics

Breaks and fluctuations in the LFS time series on disability employment

While LFS trend data on disability employment have been available since 1998, the disability questions on the LFS have undergone several changes since 2010, making comparisons over time difficult. Two sets of changes in particular that have resulted in discontinuities in time series:

• in January 2010, a rewording of the introduction to the section of the survey covering disabilities

• in April 2013, changes to the wording of the disability questions in order to bring the LFS more into line with the definitions and questions used in other household surveys in the UK (see Table 1, below)

Consequently, estimates of or based on disability employment rates of levels from 2010 onwards are not directly comparable with those for previous years.

In addition, from 2010 onwards it has been possible to produce estimates for women aged 16 to 64, after the definition of working age was changed due to state pension age equalisation between men and women. Prior to 2010, estimates for women were only available for those aged 16 to 59, when women’s state pension age was 60.

In the latest release, we have focused our analysis from April 2013, due to the change in definitions and this will allow direct comparisons to be made over time.

Quarterly LFS time series statistics on disability employment rates, levels, and the disability employment rate gap are subject to short-term fluctuations, due to seasonal variations and random sampling variation. This includes some unexplained quarterly changes, such as an apparent discontinuity in the number of disabled people recorded in the LFS between Q2 and Q3 2017, which was investigated by the producers of the data, ONS. ONS concluded that while the timing of variation could not be explained, comparisons based on longer-term were valid, and recommended that a more reliable picture of change can be gained from considering longer-term trends.

Read more about Analysis of the large increase in the Labour Force Survey disability data: April to June 2017 to July to September 2017.

Estimates in this release are based on versions of the Annual Population and Labour Force Surveys using the latest population weights. These have been reweighted to new populations derived using growth rates from HM Revenue and Customs (HMRC) Real Time Information (RTI), to allow for different trends during the coronavirus (COVID-19) pandemic. The reweighting gives improved estimates of both rates and levels.

All figures are not seasonally adjusted

As noted above, there can be seasonal fluctuations in employment and disability, so any trend data should be interpreted with caution. Same-quarter comparisons between different years are recommended to enable meaningful interpretation of change over time.

Limitations of longitudinal analyses on movements into and out of work

As analyses of movements into and out work are based on longitudinal survey data, from the APS, the precision and accuracy of these estimates can be affected by response errors, sampling errors and attrition bias.

The analysis does not capture any movements before or after this annual period, or any short-term moves that may have been reversed between the two snapshot interviews. The data gives an indication of the size of the flows into and out of work, from one year to the next.

Due to sample rotation and non-response, the sample sizes of the two-wave longitudinal datasets are smaller than the regular Annual Population Survey. This means that the results from the longitudinal Annual Population Survey differ somewhat from the regular Annual Population Survey, and the results should be treated with some caution.

In addition, in order to capture the labour market experience of all disabled people, the method used to examine movements into and out of work by disability status does not account for changes in disability status (apart from in Table FLW001). The analysis registers the disability and employment status in the first interview and the employment status in the second interview, one year later. In other words, it holds constant the disability status of the first period. Some people might change their disability status between years one and two. This analysis does not enable us to explore the reasons for any transition into or out of work, or the sequencing between any changes in employment or disability status. However, some analysis that includes disability status in the second interview, is done to show the general scale of changes in disability status over time.

General limitations of survey-based statistics

More generally, estimates in this publication are subject to potential limitations inherent in all surveys, including:

• Survey design: For example, the LFS uses a rotational sampling design, whereby a household, once initially selected for interview, is retained in the sample for a total of 5 consecutive quarters. The interviews are scheduled to take place exactly 13 weeks apart, so that the fifth interview takes place one year on from the first

• Sampling error: The fact that only a sample of the population has been selected and a different sample would probably produce a different estimate. This will vary to a greater or lesser extent depending on the level of disaggregation at which results are presented

• Non-response error: Systematic bias due to non-response by households selected for interview. In an attempt to correct for differential non-response, estimates are weighted

• Survey coverage: The error which arises because some units are either excluded or duplicated on the sampling frame used to identify members of the population of interest

• Measurement error: Made up of 4 types:

  • interviewer error arising from both conscious and unconscious differences in the way interviewers administer a survey, and also from the reactions of respondents to different types of interviewers

  • respondent error arising from the inability or unwillingness of a respondent to produce a correct answer

  • instrument error which reflects the effect of question wording, response categories and form design on responses, and

  • mode error which describes the effect of different methods of administering a questionnaire on the recorded responses

• Processing error: This consists of systems error and data handling error. Systems errors are errors in the specification or implementation of systems needed to carry out surveys and process results; system errors on the LFS can creep in when derived variables are specified and/or amended. Data handling errors are errors in the processing of survey data

• Sample size: Although the APS and LFS have a relatively large sample sizes for a household survey, small sample sizes for particular breakdowns may mean that specific analysis is not robust enough to report

For further information on the Labour Force Survey, see the following performance and quality monitoring report, including data on sample sizes and response rates: Labour Force Survey performance and quality monitoring report – Office for National Statistics (ons.gov.uk)

Comparisons between the statistics

As previously noted, these statistics use and build upon statistics published by the Office National Statistics on the employment of disabled people, as part of its Labour Market Overview: UK release (Table A08), which present overall trends.

Information on the Characteristics of disabled people in employment: April to June 2017 has previously been published by DWP.

Analysis of flows into and out of work, were presented in Annex D of Improving Lives: the future of work, health and disability, showing a similar pattern to the current release. In addition, the 2017 Stevenson/Farmer review of mental health and employers contained similar analysis that showed around 300,000 moves out work, from one quarter to the next, among people with a long-term mental health condition. The difference in the approach used in that source is described in Annex D of Improving Lives.

The estimates of employment flows in this release also differ from the ‘employment retention’ rates published in Health in the workplace: patterns of sickness absence, employer support and employment retention in 2019. The latter focused purely on people whose disability status did not change from one year to the next, whereas this release offers a more complete picture of the disabled population as a whole.

Comparing Family Resources Survey to Labour Force Survey estimates

The FRS uses the same GSS harmonised questions and definition of disability that the LFS has been using since April 2013, and the same definition of employment. However, the estimates published in the Family Resources Survey report cover people below the contemporary State Pension age at the time of interview, which for women has been increasing over time. In contrast, LFS estimates since 2010 have covered all people aged 16 to 64, for consistency over time. This release analyses FRS data for 16 to 64 year-olds, for consistency with LFS estimates. The FRS analysis does still exclude dependent children aged 16 to 19 (defined as those who live with their parents, are not married and are in full-time education), but the impact on comparisons with LFS estimates should be small.

Leading estimates of disability employment, published by ONS, are based on the Labour Force Survey (LFS). Estimates are also available from the Family Resources Survey (FRS), using the same definition of disability (from 2013 to 2014 onwards), but with a different survey methodology and a smaller sample than the LFS. FRS estimates are based on financial years (April to March), so the average of the corresponding four quarters in the LFS was used as comparison.

Figure A: Percentage of disabled people in employment by data source, people aged 16 to 64, UK, 2013 to 2020.

The estimated employment rate for disabled people tended to be slightly higher in the FRS than in the LFS. However, recent figures show the difference between the sources to be small. For 2019 to 2020, both sources gave an estimate of 53.4% for the disability employment rate.

Both sources suggest that around 1 in 5 (19%) of working age people in the UK reported a disability in 2019 to 2020, having increased by around 3 percentage points since 2013 to 2014.

Figure B: Disability employment gap by data source, people aged 16 to 64, UK, 2013 to 2020

For the disability employment rate gap, both sources show a similar trend with the gap having fallen from 2013 to 2020, but it was slightly larger in the FRS than the LFS (29.8 percentage points and 28.6 percentage points, respectively).

It is clear that these 2 data sources present a consistent picture of employment trends for disabled people in recent years, despite some small differences in estimates for specific years, caused by differences in methodology. Therefore, the FRS estimates increase our confidence in the trends observed in the LFS elsewhere in this release.

Definitions and terminology within the statistics

Disability

The Government Statistical Service (GSS) harmonised standard definition of disability is used for estimates from Apr to Jun 2013 onwards. In summary the core definition covers people who report: current physical or mental health conditions or illnesses lasting or expected to last 12 months or more; and the conditions or illnesses reduce their ability to carry out day-to-day activities.

Measuring disability can be particularly difficult. Many conditions change over time and some people have more than one condition at the same time. There are also different ways to consider disability. For example, we could consider disability according to legal definitions, medical conditions, or societal barriers.

We pledged to review and research the topic area with the aim to ensure that disability standards meet user needs, and to explore any potential improvements. In the report the ONS identified weakness of the current standard, including subjective definitions of the difference between “a little” and “a lot”, definition of “day to day activities” and incomplete impairment and conditions options. The ONS reviewed their research and engagement work to create plans for the short, medium, and longer-term to mitigate these issues.

DDA-disabled

All people with a long term health problem or disability that limits their day-to-day activities. This applies to any estimates prior to Apr to Jun 2013.

Disability employment gap

The difference between employment rates of non-disabled and disabled groups.

In employment or in work

People of working age who either: did paid work in the reference week (as an employee or self-employed); had a job that they were temporarily away from; were placed with employers on government-supported training and employment programmes; or doing unpaid family work.

Long-term health condition (LTHC)

An individual is defined as having a long-term health condition if they reported having a physical or mental health condition or illness that lasts, or is expected to last, 12 months or more.

Health conditions

Types of health condition are self-reported by LFS respondents by selecting any number of conditions from the list below, before indicating which one they consider to be their main health condition:

1. problems or disabilities (including arthritis or rheumatism) connected with your arms or hands

2. …legs or feet

3. …back or neck

4. difficulty in seeing (while wearing spectacles or contact lenses)

5. difficulty in hearing

6. a speech impediment

7. severe disfigurements, skin conditions, allergies

8. chest or breathing problems, asthma, bronchitis

9. heart, blood pressure or blood circulation problems

10. stomach, liver, kidney or digestive problems

11. diabetes

12. depression, bad nerves or anxiety

13. epilepsy

14. severe or specific learning difficulties

15. mental illness or suffer from phobias, panics or other nervous disorders

16. progressive illness not included elsewhere (for example, cancer not included elsewhere, multiple sclerosis, symptomatic HIV, Parkinson’s disease, Muscular Dystrophy)

17. Autism

18. other health problems or disabilities

Mental health (MH) condition

This is a group of conditions including depression, bad nerves and anxiety (common mental health problems) and manic depression, schizophrenia and other serious mental health problems.

Musculoskeletal (MSK) condition

This is a group of conditions including back pain, neck and upper limb problems and other musculoskeletal problems.

Working age

For estimates from 2010 onwards, this refers to people aged between 16 and 64 years old. Prior to 2010, it was defined as men aged 16 to 64 and women aged 16 to 59.

Longitudinal surveys

Longitudinal surveys have repeated observations of the same variable over a period of time interviewing the same cohort of people

Out of work

This refers to people who do not meet the definition of ‘in employment above’. In other words, they are either ILO unemployed or economically inactive.

Disability Prevalence

Disability Prevalence is the proportion of the working age population who have a disability. People who do not report whether or not they have a disability are excluded from the calculation.

Industry

Industry: is based on respondent’s main job and is classified according to the Standard Industrial Classification (SIC) 2007.

Occupation

Occupation: is based on the respondent’s main job and is classified according to the Standard Occupational Classification (SOC) 2010.

Full-time or part-time employment

Full-time or part-time employment: is not defined according to a fixed number of hours worked, but is self-reported by survey respondents.

Quarters

Labour Force Survey data is based on three-month periods known as quarters. For each year, the first quarter is based on January to March (Q1), followed by April to June (Q2), July to September (Q3) and October to December (Q4).

Reporting of the statistics

Statistical significance

All results in the commentary are statistically significant at the 95% level, unless stated. Differences in employment of groups with given characteristics are not necessarily caused by those characteristics.

We can use statistical significance to decide whether we think a difference between two survey-based estimates reflects a true change in the population rather than being attributable to random variation in our sample selection.

Statistical significance helps us to establish what observed changes or relationships we should pay attention to, and which apparent changes may have occurred only because of randomness in the sampling.

A result is said to be statistically significant if it is likely not caused by chance or the variable nature of the samples. A defined threshold can help us test for change. If the test of statistical significance calculated from the estimates at different points in time is larger than the threshold, the change is said to be “statistically significant”.

A 5% standard is often used when testing for statistical significance. The observed change is statistically significant at the 5% level if there is less than a 1 in 20 chance of the observed change being calculated by chance if there is no underlying change.

Within the commentary of our statistical bulletins, we will avoid using the term “significant” to describe trends in our statistics and will always use “statistically significant” to avoid any confusion for our users.

Rounding

Figures contained within the statistical bulletin are subject to additional rounding unless otherwise stated. The level of rounding applied, which is dependent on the magnitude of the figure being quoted, is shown in the table below.

Table showing statistical bulletin rounding policy

Range	Rounded to the nearest
0 to 1,000	10
1,001 to 10,000	100
10,001 to 100,000	1,000
100,001 to 1,000,000	10,000
1,000,001 to 10,000,000	100,000
10,000,001 to 100,000,000	1,000,000

Users should note that percentages shown within the statistical bulletin are calculated using numbers prior to rounding and rounded to one decimal place. Percentages therefore may not add to 100%.

Revisions to the statistics

These Official Statistics use are based on data from the:

• Labour Force Survey (LFS) and Annual Population Survey (APS) produced by the Office of National Statistics (ONS)

Any revisions made to the survey by data producers will be included in the next release.

Status of the statistics

Official statistics

These statistics are official statistics and have been produced in compliance with the Code of Practice for Statistics.

Quality statement

The analysis in this statistical publication is mainly based on data taken from the LFS and APS which are closely monitored for their methodology and quality. Read more about the Labour Force Survey performance and quality monitoring.

Data from the Family Resources Survey have been used to corroborate trends in employment rates for disabled people seen in the LFS.

These statistics have been developed using guidelines set out by the UK Statistics Authority in the Code of Practice for Official Statistics. This details the necessary principles and practices to produce statistics that are trustworthy, high quality and of public value. More details are given in the ‘statement of compliance’ at the end of the main release.

Feedback

We welcome feedback

Please let us know what you think of the presentation and content of our statistical release by emailing: team.workandhealthanalysis@dwp.gov.uk.

Useful links

A08: Labour market status of disabled people

Labour market overview, UK

Family Resources Survey

Characteristics of disabled people in employment: April to June 2017

Outcomes for disabled people in the UK: 2020

Census - Office for National Statistics – (ons.gov.uk)