Corporate report

DWP single departmental plan: 2018 to 2022 headline indicators technical detail

Updated 17 February 2020

This was published under the 2016 to 2019 May Conservative government

This corporate report was withdrawn on 15 July 2021

This has been replaced by the Department for Work and Pensions Outcome Delivery Plan.

The headline indicators are used to demonstrate DWP’s progress against its strategic objectives set out its single departmental plan. This technical annex shows how each of the headline indicators is calculated.

1. Overall UK employment rate

1.1 Indicator description

This indicator shows the proportion of the 16 to 64 year old population of the whole of the UK who are in employment.

1.2 Formula

This indicator is measured by dividing the number of 16 to 64 year olds who are in employment by the total number of 16 to 64 year olds.

1.3 Good performance

Generally a statistically significant increase in the indicator will demonstrate an improvement in the labour market allowing for changes in the population. External factors such as economic conditions will need to be taken into account when interpreting changes to this measure.

1.4 Comparability

Data is seasonally adjusted and available monthly on a rolling quarter basis, and therefore quarter on quarter comparisons can be made. It is based on Labour Force Survey (LFS) data which uses internationally agreed concepts and definitions, so is internationally comparable.

It is not possible to compare overlapping quarters (consecutive data points) as this is effectively the same as comparing single month figures which is not considered to be robust.

1.5 Collection frequency and details

Monthly with a 6 week time lag. Data is available on a comparable basis from 1971.

1.6 Data source

The employment rate, calculated from the LFS, is published on a monthly basis by the Office for National Statistics (ONS) in the UK labour market statistical bulletin. A full time series of the data can be found in the background tables and using the tools available on the ONS website.

1.7 Robustness and data limitations

The data is a National Statistic, produced to high professional standards. The quality of National Statistics products is assessed on a regular basis by the independent UK Statistics Authority (UKSA).

Along with other users, the Department for Work and Pensions (DWP) is represented on groups that monitor the quality and relevance of the underlying data (LFS Steering Group) and the related National Statistics outputs (Labour Market Statistics Theme Group).

The confidence intervals around the indicator are published each month. The latest confidence interval for a single quarterly estimate of the indicator is +/- 0.4 percentage points. The confidence interval for a quarterly change is +/- 0.4 percentage points and for a year on year change is +/- 0.5 percentage points.

The survey data is based on a random sample of approximately 40,000 responding UK households and 100,000 individuals per quarter. It is used for a wide range of National Statistics and is considered to provide robust estimates of labour market outcomes.

1.8 How indicator can be broken down

Data is published by:

• gender
• age (16 to 17, 18 to 24, 25 to 34, 35 to 49, 50 to 64 and 65+)
• nationality
• country of birth
• education status for those aged 16 to 24
• geography at a national, regional and local level, however local level data is often based on small sample sizes and therefore considered less robust

1.9 Further guidance

Published in the monthly ONS UK labour market statistical bulletin.

2. Young people (18 to 24 year olds) not in full-time education in employment

2.1 Indicator description

This indicator shows of the 18 to 24 year olds not in full-time education, what proportion are in employment.

2.2 Formula

This indicator is measured by dividing the number of 18 to 24 year olds not in full-time education who are in employment by the total number of 18 to 24 year olds who are not in full-time education.

2.3 Good performance

Generally a statistically significant increase in the indicator will demonstrate an improvement in the labour market position of young people. However, external factors such as economic conditions will also need to be taken into account.

2.4 Comparability

Data is seasonally adjusted and available monthly on a rolling quarter basis and therefore quarter on quarter comparisons can be made. Although the data are seasonally adjusted, it is a more reliable indicator of performance to compare changes over the course of a year rather than quarter. It is based on LFS data which uses internationally agreed concepts and definitions, so is internationally comparable.

It is not possible to compare overlapping quarters (consecutive data points) as this is effectively the same as comparing single month figures which is not considered to be robust.

2.5 Collection frequency and details

Monthly with a 6 week time lag. Data is available on a comparable basis from 1992.

2.6 Data source

This statistic is calculated using the LFS, and is published monthly by the ONS in the UK labour market statistical bulletin. A full time series of the data can be found in the background tables (Table A06) on the ONS website.

2.7 Robustness and data limitations

Analyses are National Statistics produced to high professional standards. The quality of National Statistics products is assessed on a regular basis by the independent UKSA.

Along with other users, DWP is represented on groups that monitor the quality and relevance of the underlying data (LFS Steering Group) and the related National Statistics outputs (Labour Market Statistics Theme Group).

There is no published data for the confidence intervals around the indicator. However, by looking at the variation in the raw data and by applying a design factor provided by the ONS, it is estimated that the confidence interval for a single quarterly estimate of the indicator is +/- 1.4%.

The confidence interval for a year-on-year change is wider, because it is based on 2 independent estimates and so subject to 2 sources of uncertainty. The confidence interval for a year on year change is approximately +/- 2%.

The survey data is based on a random sample of approximately 40,000 responding UK households and 100,000 individuals per quarter. It is used for a wide range of National Statistics and is considered to provide robust estimates of labour market outcomes.

2.8 How indicator can be broken down

Data is published by gender.

2.9 Further guidance

Published in the monthly ONS UK labour market statistical bulletin. It is available in Excel format in the background tables (Table A06).

3. Children living in workless households

3.1 Indicator description

The number of children living in workless households as a proportion of all children. A workless household is a household that includes at least one person aged 16 to 64, where no-one aged 16 or over is in employment.

‘Children’ refers to all children under 16.

3.2 Formula

This indicator is measured by dividing the number of children living in workless households by the number of children living in all households.

Worked example using October to December 2016 data

1,309,000 / 12,456,000 = 10.5%.

3.3 Good performance

An improvement would be indicated by a statistically significant fall in the indicator.

As the LFS is a sample survey, it is subject to a margin of uncertainty, as different samples give different results. Therefore small changes over time in estimated indicators may not be statistically significant (they may have arisen from the different samples by chance).

The magnitude of the fall required for a statistically significant change depends on the sampling variability around both the current and previous data points. For example, for there to have been a statistically significant fall between October to December 2015 and October to December 2016, the proportion of children living in workless households had to fall by at least 1.0 percentage point.

Significant changes in the indicator may be observed more easily over a longer time period. For example, 2 consecutive year-on-year changes, neither of which are statistically significant, may combine to show a significant change over the 2 year period. Looking at a series of estimates over time will aid interpretation of trends.

External factors impacting on the prevalence of parental worklessness, for example general economic conditions, may affect this indicator but be outside of the department’s control.

3.4 Comparability

Data are not seasonally adjusted, so it is recommended to use only year on year comparisons.

Eurostat, the EU statistics agency, publishes data on the proportion of children living in jobless households in EU countries, including the UK. The data is not directly comparable with the ONS published figures as it uses a different age definition for children (defined as aged 0 to 17) and adults (aged 18 to 59) to enable comparisons between different countries.

There is a time series of data from 1996 for April to June data, from 2004 for October to December data and a continuous quarterly time series from April to June 2014.

3.5 Collection frequency and details

Quarterly, around 2 months after the end of the reference quarter.

3.6 Data source

Published in the ONS release on working and workless households using the Household Labour Force Survey (HLFS). See Table K for details on children.

3.7 Robustness and data limitations

Data are National Statistics produced to high professional standards. The quality of National Statistics products is assessed on a regular basis by the independent UKSA. DWP is involved in the quality assurance process for the HLFS by assessing results for plausibility.

The survey data is based on a sample of around 40,000 households. However, only about a third of these households contain children, so small sample sizes can limit analysts’ ability to identify statistically significant trends.

As the HLFS is a sample survey, it is subject to a margin of uncertainty as different samples give different results. For example the confidence interval around the estimate for October to December 2016 is +0.7 percentage points.

Along with other users, DWP is represented on groups that monitor the quality and relevance of the underlying data (LFS Steering Group) and the related National Statistics outputs (Labour Market Statistics Theme Group) and is able to feed in any concerns about the data collection process.

3.8 How indicator can be broken down

The headline data can be broken down by:

• whether all members of the household are unemployed, all are inactive, or a combination of the 2 – the data can further be compared with households where all members are working, and with households containing both working and workless members
• household type – whether children are living in a couple household, lone parent household, or other type of household
• region or country of the UK
• housing tenure

Additional breakdowns are also available for age, gender and ethnicity, but are not routinely published.

3.9 Further guidance

See the indicator in Excel and CSV format.

4. Number of employed disabled people

4.1 Indicator description

This indicator shows the numer of 16 to 64 year old disabled people who are in employment.

Disability is defined according to the Government Statistical Service (GSS) Harmonised Principle, in line with the Equality Act 2010.

The core definition covers people who report having at least one physical or mental health condition or illness which:

• has lasted, or is expected to last, 12 months or more
• reduces their ability to carry out day-to-day activities

4.2 Formula

This indicator is taken directly from an independent published source, without further calculations.

4.3 Good performance

Generally a statistically significant increase in the indicator will demonstrate an improvement in the labour market performance of disabled people. It is an outcome measure, reflecting the real-world changes that the government wants to see, rather than a direct measure of government performance in isolation. As such, it is affected by, and ensures government policy is responsive to, external factors such as economic conditions and changes in reported disability prevalence.

4.4 Comparability

Data is not seasonally adjusted and available on a quarterly basis so only year-on-year comparisons can be made, between the same quarter each year. It is based on Labour Force Survey (LFS) data which uses internationally agreed concepts and definitions of employment.

A series break in 2013 due to definitional changes means year on year comparisons can only be made as far back as April to June 2013. The first comparable January to March data point was in January to March 2014.

4.5 Collection frequency and details

Quarterly, 6 weeks after the reference quarter. Data is available on a consistent basis starting from the April to June quarter in 2013.

4.6 Data source

Estimates based on the Labour Force Survey (LFS) are published by the ONS in their Labour Market Overview bulletin (Table A08).

4.7 Robustness and data limitations

Analyses are National Statistics produced to high professional standards. The quality of National Statistics products is assessed on a regular basis by the independent UK Statistics Authority.

Along with other users, DWP is represented on groups that monitor the quality and relevance of the underlying data (LFS Steering Group) and the related National Statistics outputs (Labour Market Statistics Theme Group).

The latest estimates for the confidence interval for a single quarterly estimate of the indicator is approximately +/- 100,000 people. The confidence interval for a year on year change is approximately +/- 140,000 people. The confidence intervals around the indicator are not published by ONS.

The survey data is based on a random sample of approximately 40,000 responding UK households and 100,000 individuals per quarter. It is used for a wide range of National Statistics and is considered to provide robust estimates of labour market outcomes.

In April 2013, changes were made to the wording of the disability questions to bring the LFS into line with the GSS Harmonised Standard for questions on disability, which in turn are based on the core definition of disability in the Equality Act 2010.

The information is self-reported and subjective.

4.8 How indicator can be broken down

ONS routinely publishes breakdowns by gender only.

4.9 Further guidance

Labour market status of disabled people(Table A08) is published quarterly by the ONS in their Labour Market Overview bulletin in Excel format.

ONS have completed their investigations into Labour Force Survey estimates of disability in 2017, concluding in their report that there was no discontinuity and lifting the health warnings previously in place.

5. The percentage of disabled people with a low income

5.1 Indicator description

This indicator measures the percentage of individuals in families where someone is disabled with household incomes below 60% of median income in any particular year before deducting housing costs.

Incomes are equivalised to adjust for family size and composition, so different household types can be compared in a reliable way.

Disabled people are identified as those who report any physical or mental health condition(s) or illness(es) that last or are expected to last 12 months or more, and which limit their ability to carry out day-to-day activities a little or a lot. This is in line with the Equality Act 2010.

5.2 Formula

This indicator is measured by:

1. producing an equivalised income before housing costs (BHC) for all individuals
2. calculating the median
3. looking at how many individuals in a family containing someone with a disability fall below a threshold of 60% of that median

Worked example

The equivalised median real income BHC in 2016/17 was £494 a week for all individuals. Thus the 60% of median income threshold was £296 a week.

Anyone with a household income of £296 or less per week would therefore be described as ‘low income’ using this measure.

5.3 Good performance

Generally a statistically significant decrease in the indicator will demonstrate that an improvement has been achieved, but external factors such as wider economic conditions also need to be taken into account.

5.4 Comparability

Measuring low income using 60% of equivalised median income is widely used internationally as a way of measuring relative low income.

Relative low income is the headline low income measure, but there are other measures of living standards that might be used including absolute low income and persistent low income.

Data available from 1995/96. The means of identifying people with a disability has changed over time however, with different criteria applied for 1994/95 to 2003/04, 2004/05 to 2011/12, and 2012/13 to date. As such, changes over time in the number of individuals with disabilities could be affected by the changes in the disability questions. Further, different individuals may also have different interpretations of particular health conditions or question wording, meaning that changes to the disability questions may have had a different effect on different groups. Therefore, comparisons between figures from 2012/13 onwards with earlier years should be made with caution.

5.5 Collection frequency and details

Annually, published around one year after the end of the survey period.

5.6 Data source

Published in the households below average income statistics (HBAI), which use the Family Resources Survey (FRS).

5.7 Robustness and data limitations

Data are National Statistics produced to high professional standards. The quality of National Statistics products is assessed on a regular basis by the independent UKSA.

In 2016/17, interviews were completed in over 19,000 households in the UK. The measure does not include care home residents as the sample used for the survey consists of private households only. Relative to administrative records, the FRS is known to under-report benefit receipt. However, the FRS is considered to be the best source for:

• looking at benefit and tax credit receipt by demographic characteristics not captured on administrative sources
• looking at total benefit receipt on a benefit unit or household basis

The FRS also collects information on other (non-benefit) income sources.

The data source, the FRS, is sponsored by DWP, and has been designed to meet departmental needs for understanding of the household income distribution and changes over time.

Confidence intervals can be calculated for the headline statistics. This means that we can be 95% sure that between 17.0% and 21.1% of individuals in families containing someone who is disabled are in relative low income BHC. For 2016/17, our central estimate is 19%. The Institute for Fiscal Studies also independently quality assures the underlying data.

5.8 How indicator can be broken down

There are a wide range of additional breakdowns available. For example by age band (child, working age or pensioner).

5.9 Further guidance

See the latest HBAI release, methodology document and supporting tables.

6. Eligible employees in a pension scheme sponsored by their employer

6.1 Indicator description

This indicator measures the number of employee jobs held by eligible individuals. Eligible individuals are defined as:

• aged at least 22 and under State Pension age
• earning above the earnings threshold for automatic enrolment
• participating in a pension scheme sponsored by their employer

An individual may have more than one job.

6.2 Formula

This indicator is measured through a count of the number of employee jobs (including those affected by absence) with an employer sponsored pension (defined benefit, defined contribution, group personal pension, stakeholder pension and unknown pension type) which are held by people aged between 22 and State Pension age who have annual gross earnings greater than the earnings threshold for automatic enrolment.

Previous years’ data have been adjusted to account for the annual revisions to the earnings thresholds used to determine automatic enrolment eligibility.

The £10,000 threshold (in 2015/16 earning terms) has been applied in 2016.

The £10,000 threshold (in 2014/15 earning terms) has been applied in 2015.

The £9,440 threshold (in 2013/14 earning terms) has been applied in 2014.

The £8,105 threshold (in 2012/13 earning terms) has been applied in 2013 and deflated by average weekly earnings (AWE) from 2006 to 2012.

6.3 Good performance

An increase of at least 100,000 employee jobs, based on unrounded data, would demonstrate that an improvement has been achieved.

6.4 Comparability

Due to eligibility criteria it is not comparable to other pension participation measures or international comparisons.

6.5 Collection frequency and details

Annually. It is available in a consistent format from 2006. The data collected relates to a specific reference date in April and is published around 3 months after the period.

6.6 Data source

Analysis is performed by DWP analysts using the Annual Survey of Hours and Earnings (ASHE) and is published in the workplace pension participation and saving trends statistics. The ASHE is published by the ONS and is an important source of information on workplace pensions in the UK.

6.7 Robustness and data limitations

There are no known quality issues. Analyses are Official Statistics produced to high professional standards, and there is a good fit between the data and the indicator being measured.

The department has oversight of the controls operated by the ONS as it is represented on the survey’s cross government user group.

The ASHE is based on a 1% sample of employee jobs taken from HM Revenue & Customs (HMRC) PAYE records. It does not cover:

• the self-employed
• employees not paid during the reference period

An indication of the quality of estimates is included as part of tables provided by the ONS. This is measured by its coefficient of variation, which is the ratio of the standard error of an estimate to the estimate.

6.8 How indicator can be broken down

Breakdowns are available by:

• sector
• industry
• employer size
• occupation
• earnings band
• gender
• age band
• region
• working pattern

6.9 Further guidance

See:

• further methodology and information
• additional methodology and quality information for ASHE

7. Total pension saving of eligible savers

7.1 Indicator description

This indicator measures the total amount saved by eligible individuals in a pension scheme through:

• employee contributions
• employer contributions
• tax relief

Eligible individuals are defined as:

• aged at least 22 and under State Pension age
• earning above the earnings threshold for automatic enrolment
• participating in a pension scheme sponsored by their employer

An individual may have more than one job.

7.2 Formula

This indicator is measured by the total amount saved into a pension scheme by eligible employees within that tax year through a combination of:

• employee contributions
• employer contributions
• tax relief

Previous years’ data have been adjusted to account for the annual revisions to the earnings thresholds used to determine automatic enrolment eligibility.

The £10,000 threshold (in 2015/16 earning terms) has been applied in 2016.

The £10,000 threshold (in 2014/15 earning terms) has been applied in 2015.

The £9,440 threshold (in 2013/14 earning terms) has been applied in 2014.

The £8,105 threshold (in 2012/13 earning terms) has been applied in 2013 and deflated by AWE from 2006 to 2012.

7.3 Good performance

An increase of at least £1 billion saved annually, from 2012, would demonstrate that an improvement has been achieved.

7.4 Comparability

Due to eligibility criteria it is not comparable to other pension participation measures or international comparisons.

7.5 Collection frequency and details

Annually. It is available in a consistent format from 2006. Data collected relates to a specific reference date in April and is published around 3 months after the period.

7.6 Data source

Analysis is performed by DWP analysts using the Annual Survey of Hours and Earnings (ASHE) and is published in workplace pension participation and saving trends statistics. The ASHE is published by the ONS and is an important source of information on workplace pensions in the UK.

7.7 Robustness and data limitations

There are no known quality issues. Analyses are Official Statistics produced to high professional standards, and there is a good fit between the data and the indicator being measured.

The department has oversight of the controls operated by the ONS as it is represented on the survey’s cross government user group.

The ASHE is based on a 1% sample of employee jobs taken from HMRC PAYE records. It does not cover:

• the self-employed
• employees not paid during the reference period

An indication of the quality of estimates is included as part of tables provided by the ONS. This is measured by its coefficient of variation, which is the ratio of the standard error of an estimate to the estimate.

7.8 How indicator can be broken down

No further breakdowns are available.

7.9 Further guidance

See further methodology and information and additional methodology and quality information for ASHE.

8. Percentage of pensioners with a low income

8.1 Indicator description

This indicator measures the percentage of pensioners with incomes below 60% of median income in any particular year after deducting housing costs.

Incomes are equivalised to adjust for family size and composition, so different household types can be compared in a reliable way. The after housing costs (AHC) low income measure is preferred for pensioners, as around three quarters of pensioners own their own home. Particularly when pensioners’ incomes are compared to those of younger individuals the AHC measure allows for more meaningful comparisons of income, both between groups and over time.

8.2 Formula

This indicator is measured by:

1. producing an AHC equivalised income for all individuals
2. calculating the median
3. looking at how many pensioners fall below a threshold of 60% of that median

Worked example

The equivalised median real income AHC in 2015/16 was £425 a week for all individuals. Thus the 60% of median income threshold was £255 a week.

A pensioner with a household income of £255 or less per week would therefore be described as ‘low income’ using this measure.

8.3 Good performance

Generally a statistically significant decrease in the indicator will demonstrate that an improvement has been achieved, but external factors such as wider economic conditions also need to be taken into account.

8.4 Comparability

Measuring low income using 60% of equivalised median income is widely used internationally as a way of measuring relative low income. The preferred measure for pensioners differs from the most commonly used international measures, as it takes incomes AHC. This however is appropriate in the UK context as most pensioners own their own home. A before housing cost measure is also available. Relative low income is the headline low income measure, but there are other measures of living standards that might be used including absolute low income and material deprivation.

Data available on a comparable basis from 1994/95.

8.5 Collection frequency and details

Annually, published around one year after the end of the survey period.

8.6 Data source

Published in the households below average income statistics (HBAI), which use the FRS.

8.7 Robustness and data limitations

Data are National Statistics produced to high professional standards. The quality of National Statistics products is assessed on a regular basis by the independent UKSA.

In 2016/17, interviews were completed in over 19,000 households in the UK. The measure does not include care home residents as the sample used for the survey consists of private households only.

Relative to administrative records, the FRS is known to under-report benefit receipt. However, the FRS is considered to be the best source for:

• looking at benefit and tax credit receipt by demographic characteristics not captured on administrative sources
• looking at total benefit receipt on a benefit unit or household basis

The FRS also collects information on other (non-benefit) income sources.

The data source, the FRS, is sponsored by DWP, and has been designed to meet departmental needs for understanding of the household income distribution and changes over time.

Confidence intervals can be calculated for the headline statistics. This means that we can be 95% sure that between 13.8% and 17.8% of pensioners are in relative low income AHC. For 2016/17, our central estimate is 16%.

The Institute for Fiscal Studies also independently quality assures the underlying data.

8.8 How indicator can be broken down

There are a wide range of additional breakdowns available. For example by:

• age band
• families with someone disabled
• ethnic group

A range of other geographical and family characteristics are also available subject to sample size constraints.

8.9 Further guidance

See the latest HBAI release, methodology document and supporting tables.

9. Children in couple-parent families reporting relationship distress

9.1 Indicator description

This indicator shows inter-parental relationship quality. A couple-parent family is classified as experiencing relationship distress if either parent states, in response to questions about their relationship with their partner, that most or all of the time they:

• consider divorce
• regret living together
• quarrel
• get on each other’s nerves

9.2 Formula

The Understanding Society survey collects information about the quality of couple relationships though the following 10 questions:

1. how often do you have a stimulating exchange of ideas? (ideas)
2. how often do you calmly discuss something? (discuss)
3. how often do you work together on a project? (work together)
4. how often do you and your partner “get on each other’s nerves”? (nerves)
5. how often do you consider divorce/separation? (divorce)
6. do you ever regret that you married or lived together? (regret)
7. how often do you and your partner quarrel? (quarrel)
8. do you kiss your partner? (kiss)
9. do you and your partner engage in outside interests together? (interests)
10. overall, how happy are you with your relationship? (happiness)

Each of the 10 questions have been analysed by exploring the association between negative responses to the questions and a range of indicators which are directly or indirectly associated with outcomes of children.

The top 4 questions, regret, divorce, quarrel and nerves, were chosen to inform the indicator.

The final relationship indicator has been constructed such that, if either adult answers negatively to any of the 4 questions, the relationship is considered to be ‘distressed’.

9.3 Good performance

A statistically significant reduction would represent an improvement. Children exposed to frequent, intense and poorly resolved conflict between parents, whether they are together or not, are at elevated risk of negative outcomes in the short and long-term.

9.4 Comparability

This measure was developed by DWP analysts using Understanding Society survey data. The underlying sample for this section was children who were present in all waves of the survey. Children could only join or leave the sample in subsequent waves if they were born into the sample, or they stopped being a dependent child. For basic descriptive statistics, all children were included in comparisons. At present only 3 data points are available for comparison, 2011-12, 2013-14 and 2015-16.

9.5 Collection frequency and details

Every 2 years.

9.6 Data source

Published in the Improving Lives evidence pack using Understanding Society survey data. Future publications of this data may come under a different heading.

9.7 Robustness and data limitations

Understanding Society is a nationwide household survey, following 41,000 households across the UK from 2009 to 2010 onwards.

Surveys gather information from a sample rather than from the whole population. The sample is designed carefully to be as representative of the general population as possible, given practical limitations such as time and cost constraints. However, results from sample surveys are always estimates, not precise figures. This means that they are subject to a margin of error (sampling error) which can affect how changes in the numbers should be interpreted, especially in the short-term. Year-on-year movements should be treated with caution.

Regression analysis of relationship distress confirmed that the elements of relationship distress captured by the new relationship distress measure were associated with separation and child conduct problems. This was the case even after we controlled for other important forms of disadvantage, family characteristics and demographics which existing research suggests have a role in explaining these outcomes.

9.8 How indicator can be broken down

There are several additional breakdowns of this data including:

• parental worklessness
• age of child
• age of youngest parent
• whether a parent is ill or disabled
• prevalence of other parental characteristics

9.9 Further guidance

See the Improving Lives Evidence Base section 2 and methodology.

10. Percentage of separated families with a maintenance arrangement

10.1 Indicator description

This indicator shows the percentage of separated families in Great Britain with a child maintenance arrangement.

10.2 Formula

This indicator is derived by dividing an estimate of the number of separated families in Great Britain with a child maintenance arrangement by an estimate of the total number of separated families in Great Britain.

10.3 Good performance

An increase in the indicator may indicate that a higher proportion of separated families have a child maintenance arrangement. However, the statistical significance of any changes cannot be assessed as the indicator is an estimate derived from various data sources.

When interpreting this indicator, it should be noted that there is no requirement for separated families to have a child maintenance arrangement, families may not want an arrangement, and in some cases an arrangement may not be possible. This indicator also does not measure the effectiveness of child maintenance arrangements.

If parents do want an arrangement, the policy intention of the reformed child maintenance scheme is to encourage co-operation and support separated parents to agree a non-statutory arrangement themselves without the involvement of the Child Maintenance Service (CMS) where possible. Breakdowns of the indicator showing the prevalence of non-statutory and statutory arrangements are published separately (see how indicator can be broken down).

10.4 Comparability

Annual comparisons are possible, and trends over time will be monitored. However, the statistical significance of any changes cannot be assessed as the indicator is an estimate derived from various data sources.

10.5 Collection frequency and details

Annually, published around 2 years after the end of the Family Resources Survey survey period (see data source). Data are available on a comparable basis from 2014/15.

10.6 Data source

Published in the separated families population statistics publication. This indicator is derived primarily from FRS data. Where appropriate, FRS responses are checked for accuracy against DWP’s CMS and Child Support Agency (CSA) administrative databases and an adjustment for the under-reporting of statutory and non-statutory arrangements is applied.

10.7 Robustness and data limitations

This indicator is based on experimental statistics which are subject to various limitations. These are explained in the relevant background information and methodology note. An important limitation is that it is not possible to assess whether any year-to-year changes in the indicator are statistically significant because the statistics are estimates derived from various data sources. However, any apparent trends over time will be monitored in line with policy intent.

10.8 How indicator can be broken down

Data is published broken down by type of child maintenance arrangement. The types of arrangement include:

• statutory arrangements (arranged with the CMS or its predecessor the CSA)
• non-statutory arrangements (all other types of arrangement including family-based arrangements)

10.9 Further guidance

Published in the annual separated families population statistics publication.

11. Net loss due to fraud and error as a percentage of overall benefit expenditure

11.1 Indicator description

This indicator measures the estimates of the levels of overpayment as a percentage of benefit expenditure, due to fraud and error across the benefit system in Great Britain.

The net overpayments (the amount of benefits overpaid minus the amount of overpayments recovered) value is calculated as the estimated monetary value of fraud and error (MVFE) overpayments minus the actual monetary value of overpayments recovered in the same financial year.

The net overpayments rate is calculated by dividing the net overpayments value by the total benefit expenditure for that year.

Recoveries refer to money recovered in the same financial year as the overpayment estimates, regardless of the period the debt is from.

This includes debt recovered by both the department and local authorities, for Housing Benefit only.

11.2 Formula

The indicator is calculated from sample data, adjustments are made and data is grossed up to provide an estimate for the whole benefit population.

Worked example: fraud and error 2018 to 2019 estimates

The estimate of total overpayments due to fraud and error across all benefits is £4.1 billion. This is 2.2% of the total benefit expenditure, which was estimated to be £183.5 billion in 2018 to 2019.

The estimate of total underpayments due to fraud and error across all benefits is £2 billion. This is 1.1% of the total benefit expenditure in 2018 to 2019.

In the same year, the department recovered £1.1 billion of overpayments, meaning the net loss to the department was £3 billion, equating to a rate of net loss of 1.6% of benefit expenditure.

These estimates are subject to statistical sampling uncertainties as detailed in Robustness and data limitations.

11.3 Good performance

A statistically significant decrease in the percentage of overpayments and underpayments would demonstrate improved performance. The estimates are based on a random sample of the total benefit caseload and are therefore subject to statistical uncertainties. The uncertainty is quantified by 95% confidence intervals around the central estimate, which are published alongside the main fraud and error estimates.

For the 2018 to 2019 net overpayment figure, the central estimate is 1.6% with the 95% range of confidence between 1.4% and 2%. Changes within this range are not considered significant and caution should be taken when interpreting changes within this range, as any changes are just as likely to be due to sampling variation rather than real change.

As claimants move from tax credits to Universal Credit, the demographic of the population the department serves will include more working families with more changeable circumstances, and this is driving an upward trend in fraud and error.

11.4 Comparability

HM Revenue and Customs produce statistics on the levels of error and fraud in child and working tax credits.

11.5 Collection frequency and details

Annual figures published once a year, with estimates usually available in May.

11.6 Data source

Published in the fraud and error in the benefit system publication, using a combination of survey data and internal data from the debt management system.

We obtain data on Housing Benefit overpayment recoveries from the Housing Benefit recoveries and fraud National Statistics.

11.7 Robustness and data limitations

The indicator is taken from a National Statistics publication and is a recognised standard. It is recognised by the UKSA.

As the fraud and error estimates are published National Statistics, the net overpayment estimates go through a rigorous quality assurance process. The finalised estimates are scrutinised by a quality assurance group bringing together:

• the measurement team
• the error checking teams
• other analysts
• operational, policy and strategy staff

Recoveries of overpayments for most benefits are included in the net overpayment calculation. Even if the department no longer administers a certain benefit, if the overpayment is recovered in subsequent years, it is included in the calculation. Several sources of recoveries are excluded from the calculation:

• administrative and civil penalties (these are penalties that are paid rather than a recovery of the overpayment value)
• Social Fund loans and short-term advances (these do not relate to overpayments)
• Direct Payment After Death (these are not included in the MVFE calculation)

Some recoveries have no associated overpayments for the same period, as these benefits are no longer administered by the department. This is because the debt relates to expenditure from previous years. When this occurs, we subtract the recoveries from the total MVFE.

11.8 How indicator can be broken down

Breakdowns for benefit overpayments and underpayments are published and are available for error type (fraud, customer error and official error) and category of error. Time series data are available, and some estimates are split by gender and age band. The fraud and error estimates can also be used for:

• obtaining an estimate for the amount over and underpaid in total and by benefit, and broken down into fraud, claimant error and official error, across the benefits administered by DWP and local authorities
• obtaining estimates for the amount over and underpaid by benefit, broken down into the types of fraud, claimant error and official error, across Universal Credit, Housing Benefit, Personal Independence Payment, Employment and Support Allowance, Pension Credit and Jobseeker’s Allowance

11.9 Further guidance

See fraud and error in the benefit system statistics for these estimates and accompanying background, methodology, technical and quality information.

The resulting estimates are published in reports and Excel spreadsheets on the above page, and the net loss figures can be found on tab 13.

12. Customer and claimant opinion of departmental services

12.1 Indicator description

This indicator measures data from the DWP Claimant Service and Experience Survey to generate a pan-departmental score of overall customer satisfaction with the department’s services.

The indicator measures the proportion of the following claimants who had meaningful contact with the department in the 3 months prior to the fieldwork, who are either fairly or very satisfied with the service they received for:

• Jobseeker’s Allowance
• Employment and Support Allowance
• Income Support
• Disability Living Allowance
• Attendance Allowance
• Carer’s Allowance
• State Pension
• Pension Credit
• Personal Independence Payment
• Universal Credit live service

12.2 Formula

The indicator is derived from a combination of the satisfaction scores from claimants who had meaningful contact or interaction with DWP services. Satisfaction is assessed through a single item relating to general perception of satisfaction.

So, thinking about all the services provided by Jobcentre Plus/the Pension Service/the Disability and Carers Service, overall how satisfied or dissatisfied are you with the service. Are you…?

The proportion of respondents who report to be ‘fairly satisfied’ or ‘very satisfied’ with the DWP service they received are counted as being satisfied.

The score is weighted to reflect the extent of contact from different benefit groups. For example if JSA claimants have a greater amount of contact than other benefit claimants, then the overall satisfaction score will be more heavily influenced by the score for this group.

12.3 Good performance

An increase in the proportion of satisfied respondents would indicate an improvement in DWP service. Due to the large sample size and the rounding to whole percentage points, any improvement or decline is likely to be significant.

12.4 Comparability

DWP satisfaction has been measured since 2003 in different forms, but a variety of changes to the welfare system, sample and methodology mean direct comparison to previous years is not possible. The survey has been broadly consistent since 2014-15.

In 2009 to 2010 Pension, Disability and Carers Service (PDCS) moved to a combined survey to include both pensions and disability carers. However comparisons longitudinally are limited due to changes in sample design, questionnaire and business structures. From 2010 to 2011, data was combined from the PDCS and Jobcentre Plus surveys to generate a pan-departmental score which was available from autumn 2011.

The satisfaction measure could potentially be compared to other organisations if overall satisfaction of those organisations is measured with the same question item and response scale.

Direct comparison to other countries cannot be made due to differences in welfare policy and delivery.

12.5 Collection frequency and details

Published scores are available annually, on a financial year basis, approximately 6 to 9 months after the close of the year.

Overall satisfaction score for the financial year is available approximately one month following the conclusion of the final quarter of fieldwork (concludes in May). Therefore an early estimation of overall satisfaction is usually available and published in the DWP Annual Report and Accounts.

12.6 Data source

DWP Claimant Service and Experience Survey.

12.7 Robustness and data limitations

The survey took its population as all claimants who had contact with the service in the 3 months prior to each data collection rather than all claimants in receipt of benefits. The rationale for this was that claimants who had no recent contact would not be able to provide useful information about the current state of the service, if they were able to offer any opinions at all.

Around 40,000 people are identified as being valid to be surveyed with around 15,000 achieved interviews forming the final survey.

The survey results have been weighted to be representative of contacting DWP claimants rather than the overall benefit recipient population. This means that claimants on benefits for which there is more regular contact with the department are present in larger numbers than those on benefits for which there is less contact but greater numbers, for example State Pension. The procedure for arriving at the DWP weight mainly involves calculating the population of contacting claimants for each benefit and then applying a weight to respondents within each benefit according to the number of contacting claimants.

More details on survey information, including the sampling approach, weighting and breakdown of number of people interviewed can be found in Further guidance.

12.8 How indicator can be broken down

The overall satisfaction score can be broken down by benefit group and region.

The following breakdowns can be obtained, base sample size permitting, but are not currently routinely published:

• type of transaction
• age
• gender
• ethnicity
• disability
• sexual orientation
• marital status
• religion
• whether claimants have children or not
• English as a second language

12.9 Further guidance

See further details of the survey, sample and methodology.

13. New claims processed within planned timescales

13.1 Indicator description

The annual aggregate measure is derived from monthly data relating to processing times for a range of new benefit or service claims cleared within planned timescales. Specifically, the percentage of new claims processed within 2018 to 2019 timescales for:

• Jobseeker’s Allowance (JSA) – 10 days
• Employment and Support Allowance (ESA) – 10 days
• Income Support – 5 days
• State Pension – 10 days
• Pension Credit – 35 days
• Disability Living Allowance (children) – 20 days
• Personal Independence Payment (PIP) – 75 days
• Child Maintenance Service 2012 scheme – 6 weeks

The aggregate measure is constructed using a weighting of the performance for each benefit using the total number of claims processed for each benefit over any particular year. It is defined to be an overview of benefit processing performance.

The department processes millions of new benefit and service applications each year. Each different benefit or service requires different levels of input and tasks to process and each of the benefits listed may differ in structure. All of the timescales listed are based on regular planning exercises and performance is monitored on a monthly basis.

13.2 Formula

This indicator is measured by dividing the number of new claims cleared within planned timescales by the total number of new claims processed.

13.3 Good performance

Good performance is indicated when the percentage:

• increases over time
• is greater than or equal to the planning assumptions

13.4 Comparability

This aggregate measure was derived to give an overview of how the department is performing in terms of processing. It was brought in in 2015 to 2016 as part of the single departmental plan and originally only included the main out of work benefits. From 2016 to 2017 this has been expanded to include more benefits and services. Therefore the data for 2016 to 2017 onwards is not directly comparable with the years before. In the current 2018 to 2019 measure, a number of products (ESA, JSA, Income Support, State Pension and PIP) are reported on a cohort basis – here performance is measured against the claims due to have been processed rather than those which have actually been processed. This has only a minor impact on the annual figures and does not render the 2018 to 2019 measure incomparable to previous years. Clearance time measures are used in operations to manage performance and customer satisfaction. A number of other benefits are monitored each month, but use an average clearance time and are therefore not comparable.

Each benefit processing time varies based on the type of benefit and number of tasks and inputs needed to process the claim. For example JSA is a reactive benefit and the processing time is based upon how many working days it takes to make a decision from the claim date. Comparatively State Pension is proactive and can be made at any time up to 4 months prior to pension age. Therefore it is a measure of how many claims are processed by the ‘claiming from’ date and also those received within 10 days of the ‘claiming from’ date, how many of them have been processed within 10 days. It is not a measure of how many claims were cleared within 10 days of receipt of the claim.

While Universal Credit full service has been rolled out, it is not comparable with other benefit lines from the aggregate timeliness measure because it utilises a % full payment made on time (first assessment period) measure as opposed to an X in Y measure. It is therefore not consistent with the measures of other benefits and so is non-comparable. Universal Credit full service payment timeliness will be published separately.

13.5 Collection frequency and details

Published annually within the Annual Report and Accounts for that financial year.

13.6 Data source

Various internal data systems.

13.7 Robustness and data limitations

Data collection is automated via internal reporting systems and checks are in place to ensure accuracy of data extraction from system. Each of the measures undergoes a planning stage to set the processing time to ensure that it is set at an appropriate level. In 2015 to 2016 for example the number of days to process an ESA claim was reduced from 16 to 10 days following analysis. A number of other benefits are excluded as they are not measured in a set format that allows a measure of simple number processed or total volume. Where possible we will seek to include these benefits in future, so that we can give the broadest possible overview of performance across the whole department.

Each year we will evaluate whether additional benefits can be added to this aggregate measure.

13.8 How indicator can be broken down

Some processing times are included within publications for that benefit, see further guidance for details.

13.9 Further guidance

The department publishes a large amount of information on the benefits and services it provides and many of these publications include detail on the number of claims processed and processing times. See:

• DWP statistical summaries
• Personal Independence Payment statistics
• Statistics on the 2012 statutory child maintenance scheme
• Housing Benefit: statistics on speed of processing

14. Proportion of Universal Credit new claims that were paid (full payment in time)

14.1 Indicator description

From November 2018, the statistics on households on Universal Credit were expanded to include information on payment timeliness for households on Universal Credit full service.

Universal Credit entitlement is calculated over monthly assessment periods. Claims which are paid on time will receive payment by the payment due date, which is 7 days after an assessment period ends. Payments can be made late for a number of reasons, including:

• verification processes not being completed on time (either by DWP or the claimant)
• claims being amended at a late stage

It is assumed that all claims that are due a payment will be paid: either on time or late. Some claims will not receive a payment, and these claims are excluded from the calculation, as they are neither paid on time nor late. There are a numerous reasons why a claim might not be in payment, for instance because the claimant is not eligible for support, or because their earnings during the monthly assessment period are sufficiently high.

The statistics show, for households that have an assessment period that spans the count date (the second Thursday of the month) and that have been paid, the proportion which received either full or partial payment on time. The statistics for ‘all claims’ give a full picture of payment timeliness for UC full service. The figures for ‘new claims’ reflect the timeliness of the first payment to claimants (for example, relating to the first assessment period of the claim). There are a number of one-off verification processes that must be completed by the claimant and by DWP at the start of the claim – to confirm the current circumstances of the claimant (or both claimants in a joint claim) and their entitlement to UC. Payment timeliness is therefore liable to be lower for new claims.

Claims are recorded as receiving ‘some payment on time’ if at least some of the UC award has been paid by the date it was due. In some cases the claimant may receive an additional payment after the payment due date.

Claims are recorded as receiving ‘full payment on time’ where the full UC award has been paid by the date it was due (for example, the claimant does not receive any additional payments, for the same period of the claim, after the payment due date).

One circumstance where a payment could be made in multiple instalments is where verification processes have not been completed for a particular element of a claim. For example, if verification has not been completed for the housing element, the basic payment may be paid first, followed by a payment in respect of housing once the claimant’s entitlement to this element is confirmed. If the basic payment was made by the due date but the housing element was paid later then we would determine that ‘some payment’ was made on time, but that the ‘full payment’ was not.

More information, including the explanation of key concepts, is available from the UC background and methodology document.

14.2 Formula

The proportion of claims which were paid on time is calculated as:

Proportion of claims paid on time = Claims paid on time / (Claims paid on time + Claims paid late)

14.3 Good performance

An increase in the payment timeliness measures will generally indicate an improvement in operational performance. However other factors need to be taken into account. If there are delays in making the first payment, this can be due to outstanding verification issues, such as providing bank statements or proof of rent. It can also be due to a claimant not signing their claimant commitment.

14.4 Comparability

The payment timeliness measure in the Universal Credit Official Statistics is not directly comparable to the statistical ad-hoc that was published in September 2017. This is because, although the same method of calculation has been used for both, the ad-hoc was based on the week that payment was due, rather than the count date.

Universal Credit full service is not comparable with other benefit lines from the aggregate timeliness measure because it utilises a percentage of full/part payment made on time measure as opposed to an X in Y measure. It is therefore not consistent with the measures of other benefits and so is non-comparable. For example, Jobseekers Allowance (JSA) is a reactive benefit and the processing time is based upon how many working days it takes to make a decision from the claim date. Comparatively, State Pension is proactive and can be made at any time up to 4 months prior to pension age. Therefore, it is a measure of how many claims are processed by the ‘claiming from’ date and also those received within 10 days of the ‘claiming from’ date, how many of them have been processed within 10 days. It is not a measure of how many claims were cleared within 10 days of receipt of the claim.

14.5 Collection frequency and details

Published quarterly as part of the official statistics on households on Universal Credit.

14.6 Data source

Internal Universal Credit full service data.

14.7 Robustness and data limitations

Universal Credit household statistics are subject to revision in future publications. These statistics are based on the status of each claim as the latest count date. Information which is provided or verified after this date can result in further late payments being made to these households. Therefore payment timeliness figures are likely to be revised downwards as more payment data becomes available.

14.8 How indicator can be broken down

The department is planning to add payment timeliness information to Stat-Xplore later this year which will allow users to create various tabulations based on geography, entitlements, family type and more.

14.9 Further guidance

The department publishes a large amount of information on Universal Credit including official statistics on people on Universal Credit and various ad-hoc publications. See:

• Universal Credit Statistics
• Universal Credit statistics: background information and methodology

Further information

Read more about DWP research and how it supports our policies and services.

Read about statistics at DWP.

The list of latest statistics from DWP allows you to refine the results using keywords and other filters.

Read the latest list of publications and research from DWP

Our Stat-Xplore tool allows you to download and customise DWP statistics.

For information on Equality and Diversity at DWP, please see our dedicated page.

You can find Health and Safety statistics on the Health and Safety Executive website.

Read previous DWP business plan transparency indicator measures and Single Departmental Plan indicators.

The previous indicator measures are also on the National Archives website.