About the statistics

The statistics allow people to see how many sanction decisions have been made in relation to Universal Credit (live and full service). It also explains the methodology for the benefit sanction duration statistics and sanction rate that is used from the February 2024 release for:

• Universal Credit (UC)

• Jobseeker’s Allowance (JSA)

• Employment and Support Allowance (ESA)

The sanction decisions statistics are published in the Department for Work and Pensions’ (DWP) collection of benefit sanctions statistics (BSS):

• from November 2017 for UC live service
• from May 2019 for UC full service (UCFS)

These statistics are released every quarter in February, May, August, and November.

In response to feedback from the users of sanction statistics, and to improve public understanding of them, the sanctions statistics publication strategy was published by the DWP on 19 April 2016, where changes were proposed to the publication of sanction statistics.

One of these changes was to include Income Support and Universal Credit sanctions statistics in the publication alongside JSA and ESA. This was achieved as of May 2017.

These statistics covered UC live service only until May 2019.

UC Live service statistics were published within the quarterly benefit summary (QSS) until August 2017.

The underlying data for live service statistics is available through Stat-Xplore.

Benefit sanctions statistics to April 2020 (Official Statistics) and beyond cover the time period after March 2020 and include data for sanction decisions during the coronavirus (COVID-19) pandemic. Due to legislative changes to disapply work-search and work availability requirements between 30 March 2020 and 30 June 2020 (in response to the pandemic), comparisons during the coronavirus pandemic should not be made to other times.

At the beginning of the pandemic, conditionality was paused for UC claimants and face to face appointments were suspended, to enable the Department for Work and Pensions (DWP) to focus on processing the unprecedented volume of new claims. While DWP reintroduced conditionality in July 2020, face to face appointments were gradually reintroduced from April 2021 for all claimants. As the return to face-to-face appointments was a gradual process and claimants were only subject to sanction when they failed to meet a mandatory requirement, there was a delay between the start of the return to face to face appointments in April 2021 and the increase in the UC sanction rate in June 2021. Conditionality was reintroduced from 1 July 2020 on a phased approach and as capacity allowed.

The series for sanction durations and median sanction length, and Universal Credit sanction rate data for earlier than April 2019 have been reinstated with effect from the February 2024 release. These data were suspended from the November 2020 release while an investigation and review of the code was carried out.

Following this review, methodological improvements were made to the complex code used to derive the suspended sanction durations, median sanction length and UC sanction rate measures. Extensive Quality Assurance activities were undertaken to ensure improvements are robust before they were reinstated in the February 2024 release of the Benefit Sanction statistics.

Users referencing releases earlier than February 2024 should be aware that the methodology for benefit sanction duration and the sanction rate was different. Users can access previous versions of this document explaining previous methodologies from the National Archives.

Context of the statistics

Sanctions statistics comprise 4 benefits:

• Jobseeker’s Allowance (JSA)
• Employment and Support Allowance (ESA)
• Income Support (IS)
• Universal Credit (UC)

Prior to the introduction of Universal Credit, sanctions could be applied to the following benefits- JSA, ESA and IS. This document focuses purely on Universal Credit. For information on background and methodology for JSA, ESA and IS see methodology for the JSA, ESA and IS statistics.

Universal Credit

To claim Universal Credit, a Work Coach will set out with the claimant what is required of them in their Claimant Commitment. If they fail to meet each of their responsibilities that they agreed in their Claimant Commitment without good reason, they may be subject to a sanction. A sanction will reduce their UC standard allowance until they comply with their Claimant Commitment again and, depending on the number of similar failures by the claimant and their age, for an additional fixed period.

The length of the sanction depends on the level category that the reason for the sanction falls into (see Table 3) and the number of previous failures within the year.

Sanction levels in UC

There are 4 sanction levels in UC.

1. Lowest level

Failure to attend or take part in a Work-Focused Interview. The sanction lasts until the claimant attends or takes part in one, or moves to either the Working – no requirements or No Work Requirements conditionality regimes.

2. Low level

The sanction lasts until the claimant does what they previously failed to do and were sanctioned for (for example, failing to attend a training course) or because either the requirement is no longer appropriate or an alternative compliance condition has been met, plus 7, 14 or 28 days for the first, second or third low level sanction in any 12-month period.

3. Medium level

The sanction lasts 28 days for the first sanction in any 12 month period, and 91 days (approximately 3 months) for a second medium level sanction. Medium level sanctions apply, for example, where the claimant has to meet the work availability requirement, but has failed to be available to attend an interview or start work.

4. High level

The sanction lasts for 91 days (approximately 3 months) for the first sanction in any 12-month period and 182 days (approximately 6 months) for a second high level sanction. High level sanctions apply, for example, where a claimant refuses the offer of a job.

The length of the sanction will be less than that stated above if the claimant is under the age of 18 at the point of the sanctionable failure.

Decision Making Process

For Universal Credit, the Decision Making Process follows 4 stages.

1. Referral

Work Coach in a Jobcentre, or Employment Scheme provider, identifies that the claimant has failed, in some way, to meet the requirements placed upon them. The benefit will remain in payment until the Decision Maker makes an adverse decision. The Work Coach will gather information to support the referral from the claimant and, where applicable, the reasons for their failure, to enable a quality decision to be made. The case is then referred to the Decision Maker. If the referral is from an employment scheme provider, the Decision Maker will gather this information from the claimant.

2. Original decision

The Decision Maker will weigh up all of the information and evidence presented to them, including the claimant’s reasons for failing to meet the requirements placed upon them. They will then make a decision. If the decision is to apply a sanction (adverse decision), the claimant is notified of the outcome in writing.

3. Mandatory Reconsideration (MR)

If a claimant disagrees with a decision made about their benefit, they can ask the department to look at it again. A Mandatory Reconsideration must be undertaken before the claimant can appeal. There are strict time limits for asking for a Mandatory Reconsideration. The claimant must ask for a Mandatory Reconsideration within one month of the date on their decision letter, however, they can apply outside of this period if they show good reason for the delay, or have this period extended by 14 days if they receive a written statement of reasons.

If the DWP changes its original decision (non-adverse MR decision), any arrears of benefit due are paid less any hardship payments. Once the MR has been undertaken, the claimant will receive two copies of a Mandatory Reconsideration Notice (MRN). The MRN tells the claimant the outcome of the reconsideration and what to do if they wish to appeal.

4. Appeal

If the department does not change the decision at Mandatory Reconsideration, the claimant can appeal directly to an independent tribunal.

A major difference between sanctions policy for UC compared to JSA is that under UC, where a claimant fails to attend a Work Coach meeting, they can be sanctioned. This implies that UC claimants may be in receipt of other parts of UC, such as housing and child care, which would not be subject to a sanction and so they would remain on benefit. The amount deducted is calculated daily as a percentage of the standard allowance and is dependent on the claimant’s current personal circumstances and conditionality regime. Whereas under JSA claimants would receive a sanction, or have their claim terminated if they did not make contact within 5 working days of their failure.

Table 1: Percentage deducted

Claimant	Deduction (%)
Single claimant in either the Searching for Work, Working – with requirements, Working No Requirements[footnote 1] or Preparing for work Conditionality Regime	100%
Claimant in a couple in either the Searching for Work, Working – with requirements, Working No Requirements or Preparing for work Conditionality Regime	50%
Single claimant in the Planning for Work regime or in No Work requirements regime on the grounds of childcare responsibilities, adoption or pregnancy	40%
Claimant in a couple in the Planning for Work regime or in No Work requirements regime on the grounds of childcare responsibilities, adoption or pregnancy	20%
Claimant in No Work requirements regime with limited capability for work related activities	0%

UC sanctions

When a sanction is imposed on a UC claim and the sanctioned claimant or couple have had their standard allowance reduced at the daily rate equal to 100% – or 50% for a couple – for a high, medium or low level sanction, hardship payments may be available for claimants.

These payments are equivalent to about 60% of the UC standard allowance and have to be paid back under UC.

For UC claimants, a hardship payment is only available following the first reduced payment and where the claimant complies with all requirements placed on them. They are available to any claimant who can demonstrate that they would suffer financial hardship if the benefit was not paid. Claimants are expected to apply for a hardship payment for every assessment period where a sanction has been applied.

In May 2019, UC full-service statistics were added to the BSS publication which covered a wider range of claimant types. Full service was gradually introduced to jobcentres from 2016 and was available in every jobcentre across Great Britain and Northern Ireland by December 2018. When full service became available in a Jobcentre, existing Universal Credit claimants on live service were transferred to full service within 3 months.

As of December 2018, all Jobcentre Plus Offices have rolled out to full service, meaning that no new claimants will join live service. As of March 2019, less than 1% of the Universal Credit caseload was on live service. As of April 2019, the Live service was fully closed down.

Sanctions figures for live service are shown from August 2015 onwards, this being the earliest date from which we can assure the quality of the data and allow for additional functionality on the UC system to be captured.

Figures for full service adverse decisions are shown from May 2016 onwards, as this was the date from which roll out began to expand to areas of the country other than the initial trial areas (Sutton, Southwark, Croydon, Hounslow and Musselburgh).

At present, we do not have data on non-adverse, reserved or cancelled decisions for full service.

The following updates have been made to the UC full service sanction decisions statistics since they were initially released in May 2020:

1. From November 2019, Northern Ireland sanction decisions were removed from the statistics, as were sanctions given in error. All adverse sanction decisions remain in our full service data source, even when they are deleted from the front end system. This means that prior to November 2019, sanctions added in error and then subsequently removed were included in the count of adverse decisions. We have now changed our methodology, allowing us to remove these decisions, and improving the accuracy of our statistics.

2. In February 2020, we added demographic breakdowns relating to age, gender, and regional group for adverse decisions on full service. In addition, a measure to determine the number of claimants receiving multiple adverse decisions in the past year was added.

3. In February 2020, it was identified that previously published ‘all sanction decisions’ statistics for UC full service were missing some data as a result of how information is recorded. While the data on the number of adverse sanctions given is correct and the sanction rate is correct, some other sanction related decisions have been undercounted (i.e. where a sanction was not given). We removed the ‘all sanction decisions’ measures from our upcoming publications and from the previously published documents from May 2019, August 2019, and November 2019.

4. While implementing a methodological improvement in May 2020, it was identified that previously published ‘adverse sanction decisions’ statistics for UC full service had been undercounted since November 2019. We previously calculated UC full service adverse decisions by selecting cases where the decision date fell within the period a service centre was active, but found that some valid adverse decisions were missed because they occurred outside of this date range. The new methodology counts all adverse sanction decisions.

In August 2017, a monthly rate of UC claimants who receive a deduction as a result of a sanction was published alongside information on the duration of these deductions in the quarterly statistical summary, then in the benefit sanctions statistics publication from November 2017. Further information on these measures can be found in the background and methodology document.

Due to increasing demand from users, DWP have made improvements to published geography information in their Official statistics releases, including data released via Stat-Xplore. The following changes were implemented in the benefit sanctions statistics publication from February 2018, which involved:

• updating geographical information with the most recent geographies, currently the 2024 COAs.
• assigned geographies using an improved methodology, which involved using more reliable data to source address information

New claims to Universal Credit live service ceased in January 2018, and since then all live service cases have been migrated to Universal Credit full service.

At the end of March 2019, the systems used to administer live service cases were shut down. Due to this, data for any original Universal Credit live service sanction decisions has been frozen from this point. Users should be mindful that this is also evident in the combined Universal Credit live and full-service sanctions durations and rate data, and caution should be taken when comparing data before and after 1 April 2019.

Work is currently underway to develop more data for Universal Credit full service, and the aim is to release this as and when it is available.

Data sources

UC Live Service

UC live service was gradually introduced to Jobcentres from 2013. It was available in every Jobcentre (excluding Northern Ireland) by May 2016. All claimants were moved to UC full service by April 2019. New claims to UC live service ceased in January 2018, and since then the remaining UC live service cases were gradually migrated to UC full service. By 1 April 2019, the systems that were used to administer live service cases were shut down.

The main bulk of the data is taken from the Work Services Platform (WSP), the system on which Work Coaches record doubts and one of the systems Decision Makers use to record decisions made. This is then complemented with data from the Decision Makers and Appeals Case Recorder (DMACR) system, and the Payment Manager Systems (PMX). PMX is the system used to calculate a claimant’s entitlement.

The data is collated and put on Stat-Xplore for analysis. The variables include all sanction decisions made, and decisions on individuals. These can then be broken down by:

• month decision made
• age
• gender
• geography
• decision type
• outcome
• sanction level
• referral reason

Stat-Xplore is available to the public to enable the build of tables of specific interest.

UC Full Service

UC full service is the digital system that administers UC for the full range of claimant groups. It was gradually introduced to Jobcentres from 2016 and was available in every Jobcentre (including Northern Ireland) by December 2018. It is worth noting that whilst UC full service is available in Northern Ireland, these statistics exclude these cases.

Universal Credit sanction data is sourced from central systems used to administer Universal Credit. These systems are Universal Credit (UC) live service, including Payment Management Systems (PMX), and monthly data feeds of Universal Credit (UC) full service administrative data.

Revisions to the statistics

DWP has a policy for planned revisions describing how we will handle revisions and give confidence that all revisions will be handled in a transparent manner. For these statistics, to reflect any updates to the figures, the full historic statistical series is refreshed each time the figures are released.

Benefit Sanction durations and sanction rate calculation

Data source

The data on both UC live and full service systems shows the date of which the claimant did not comply with their requirements, the date when a decision maker determined the sanction is to be applied and, for each assessment period the sanction was active, the number of days the sanction was applied in the assessment period (only UC full service splits this into open and fixed days served).

Due to the availability of some data on UC live service it has not been possible to apply the methodology described in this live service document to all UC sanctions before January 2017. Therefore, statistics are only presented from January 2017.

The data on UC live service and UC full service are brought together into one data file. As there are no consistent identifiers to link the sanctions that were still active when a claimant transferred from UC live service to UC full service it has been necessary to utilise data matching techniques using consistent information between the two systems to estimate sanctions that are the same on the two systems. It is possible that not all sanctions that transferred from UC live service to UC full service have been identified in this process and that a small number have been incorrectly linked. These limitations have no material impact on the statistics.

The Quality Assurance of Administrative Data (QAAD) details the quality assurance measures and processes on the data used in the production of the Benefit Sanctions statistics.

Unit of measurement

Durations measure the length of each completed UC sanction from when the sanction started to be applied to when it stopped being applied and the length of the sanction is counted in the month in which the sanction ended.

Where a sanction is a combination of ‘open-ended’ and ‘fixed’ days, the duration is the sum of those days. The duration is not necessarily the length of time between a claimant non-complying and re-complying as a fixed period may be added on to the period where they were not compliant with their requirements. Where a claimant has multiple benefit sanctions running one after the other, these are each treated as separate sanctions.

The sanction rate measures the number of claimants undergoing a sanction on the second Thursday of the reference month (the count date) divided by the number of UC claimants in conditionality regimes where sanctions can be applied.

All UC sanctions in progress on the count date are measured, including those where there is no reduction in the claimant’s UC payment. A claimant may not have a reduction in their UC payment if their standard allowance has been reduced to nil by other means, for example because of the earnings taper. Therefore, the number of people being sanctioned should not be read as the number of people that have had their UC payment reduced because of a sanction.

UC sanction start date

A start date for the sanction has been derived based on when the claimant did not meet their requirements and when the decision maker confirmed the sanction to be applied. The following rules are made to determine the start date:

• if the non-compliance date and decision date to impose the sanction are in the same assessment period, the start date will be the non-compliance date

• if the decision date to impose the sanction is in a later assessment period than the non-compliance date, the start date will be the start of the assessment period in which the decision is made

• if the claimant is already serving another sanction on the start date (as determined in the above bullets), then the start date will be the day after all existing and deferred sanctions have completed

In a small number of sanctions, the methodology will derive a start date that is before the non-compliance date. This occurs when a claimant will be serving a sanction entirely within one assessment period and there are insufficient days remaining in the assessment period from the non-compliance date to serve the fixed days. This is a decision taken in the methodology of the statistics and does not occur in practice.

During an assessment period, a claimant will have an amount deducted from their Universal Credit payment corresponding to the number of days they do not fulfil the requirements of their claimant commitment. Additionally, an amount for the fixed days will be deducted if there is sufficient standard allowance remaining after the reduction for the open-ended part of the sanction.

Sanction end date and length of sanction

The end date for the benefit sanction is the number of open and fixed days sanctions were imposed as shown in the data added to the sanction start date, inclusive of the sanction start date.

The length of the sanction is the number of days between the start and end of the sanction.

Overturned sanctions

The purpose of these statistics is to provide a picture of the number of people who were serving a sanction at the time of the original payment for the assessment period. When a sanction has been overturned at a later date and the payment for the assessment period adjusted the overturned sanction will still be included in these statistics as it reduced the UC payment in the original payment for the assessment period.

As an overturned sanction may bring forward the start dates of subsequent sanctions to be imposed, the methodology applied in these statistics may show the claimant serving two sanctions at the same time. This is a compromise in the methodology to derive these statistics and not a feature that occurs in practice. Where this occurs we will only count the claimant being sanctioned once when calculating the sanction rate. However, each sanction will be counted at the end date of the sanction for durations.

Median

The median length of a sanction is calculated by sorting all sanctions by length from shortest to longest, and then identifying the middle value. The decision was made to use the median value – rather than the mean – to represent sanction lengths. This is because there are outliers in the datasets which cause the mean values to be skewed, resulting in them not being accurate representations of the data. The median values for sanction lengths are therefore more representative of the central values in the data.

Data issues of UC benefit sanction duration and rate statistics

The following issues have been identified in our data for the UC sanction rate and are under investigation.

There are a small number of cases missing from the Universal Credit caseload dataset. The sanction rate uses the UC caseload data to calculate both the number of UC claimants undergoing a sanction (numerator), and the number of claimants in conditionality regimes that can be sanctioned (denominator). While there is a small impact to the UC caseload, there is no material impact to the sanction rate. However, users should apply some degree of caution when considering underlying figures for numbers of sanctions or caseload counts.

There is a processing issue with a portion of the conditionality regime data affecting a small number of claimants in the statistical datasets. This issue causes some of the affected claimants to be recorded as being in a different conditionality regime in the statistics to what they were actually in on the reference date. The number of claimants in conditionality regimes that can be sanctioned is used in the denominator for the sanction rate. It is not thought that this has a material impact on the sanction rate.

Change to the methodology

The above commentary documents the methodology in use since February 2024. This improved methodology was introduced to better align the statistics with policy and operational practice and make better use of the available data.

A brief summary of the differences between the February 2024 methodology and earlier methodologies is provided in the following table.

Measure	February 2024 methodology	Earlier methodology	Reason for change
Duration being measured	Individual sanction	Period of reduction in payment from one or more sanctions	More accurate reflection of policy and better use of available data
Sanction start date	Non-compliance date, or – Start of assessment period (AP) of later AP in which decision is made, or – Immediately after completion of exiting sanction being imposed	Payment date (7 days after end of AP) for 1st AP in which reduction is made, or – Immediately after completion of existing sanction being imposed	More accurate reflection of policy and better use of available data
Assessment Period with sanction applied but no reduction in UC payment	Sanction continues to be counted in rate. Period of no reduction in payment is included in duration	Sanction spell would end and not restart until there is another reduction in payment from a sanction. Not counted in sanction rate	More accurate reflection of policy

The effect of these changes is to:

• count most sanctions earlier by 1 month as sanctions are now counted in the assessment period in which they occur and not from the payment date after the end of the assessment period
• count more people being sanctioned by including those with no payment reduction who were excluded from earlier methodologies thus the sanction rate will be higher than previously. There is no change to the number of claimants being sanctioned
• change duration lengths as sanctions are not being joined together to form a continuous spell or split up to separate spells where there is no reduction in payment

Sanction rate by ethnicity - Methodology

Measuring disparities between ethnic groups

The sanction rate measure is the most appropriate way to make direct statistical comparisons of outcomes by ethnic group relative to the relevant caseload. This is because the sanction rate measures direct outcomes and thus enables us to make direct comparisons. The sanction rate also takes into account the duration of sanctions, which is not considered when looking at adverse sanction decisions.

The statistical testing carried out to identify any differences that are significant between ethnic groups follows the methodology used as the standard approach for assessing disparities of outcomes by ethnicity, developed and recommended for use across Government by the Cabinet Office (Race Equality Unit) published in 2020. It is used in other government departments for considering potential disparities of outcomes by ethnicity and is endorsed and supported by the Office for Equality and Opportunity.

Understanding the relative likelihood index

The Relative Likelihood has been used in the Benefit Sanction Statistics which is a statistical approach used for the comparison of the relative difference in sanction rates between two ethnic groups.

A relative likelihood indicates the extent to which two groups differ in their likelihood of experiencing an outcome.

The baseline ethnic group must be independent from the comparator ethnic groups. The white group has been taken as the baseline in comparisons, which is why the overall sanction rate would not be suitable to use for comparisons

However, differences in relative likelihoods should only be interpreted as evidence of disparity where statistical significance is observed concurrently with a test of effect size.

Statistical Significance testing

The confidence intervals around the comparator ethnic group relative likelihoods provide individual tests of statistical significance. For details of the calculation used for the confidence intervals, see the guidance published by the Race Equality Unit.

Benefit Sanction Statistics uses a 95% confidence interval to test for statistical significance. This means that we can be 95% certain that the relative likelihood for the ethnic group is between the lower bound and upper bound of the confidence interval range. Confidence intervals that do not cross parity (a relative likelihood of 1.0) indicate the ethnic group has a statistically significant disparity compared to the baseline ethnic group.

Testing the size of disparity and the ‘four-fifths’ rule

Some difference in outcomes between ethnic groups is to be expected due to natural variation. As such it is essential to not only understand if such differences are statistically significant, and not the result of chance alone, but also whether the scale of impact of those differences is meaningful. With larger sample sizes, significance testing has the power to detect small effects. Considering the scale of the effect size allows us to assess whether such differences would have a notable material meaningful impact. This then guides interpretation as to whether such differences require monitoring, investigation or action.

To test the effect size (i.e. the size of any differences in relative likelihoods), a four-fifths rule has been applied to the relative likelihood estimates. Any relative likelihood estimates that fall outside a range of 0.80 to 1.25 indicate the impact of the disparity is large enough to be considered adverse from a policy perspective (practically significant).

The four-fifths rule is used in the US by the Equal Employment Opportunity Commission, to review decisions such as those about hiring, promotion or other employment decisions. The Commission has taken the position that evidence of adverse impact exists when the selection rate for any race, sex, or ethnic group is less than four-fifths (80 percent) of the rate for the group with the highest rate.

The four-fifths rule has been used by the Ministry of Justice to identify disproportionality in the criminal justice system, to focus on those outcomes which are further from equality. published by the Race Equality Unit

Example

Worked example (not real data):

V, X, Y, Z representing different ethnic group

In this example, the V ethnic group are 1.30 times as likely (30% more likely) to be sanctioned than the baseline ethnic group. In comparison, the Z ethnic group are 0.60 times as likely (40% less likely) to be sanctioned than the baseline ethnic group. 

The confidence interval range for Y ethnic groups does not cross 1 so is statistically significant. The relative likelihood is in the shaded (grey) area so while the disparity from the baseline ethnic group is statistically significant it is not notable. 

The Z ethnic group has a relative likelihood outside of the shaded area (lower than 0.8) and the range of the confidence interval does not cross 1 so the disparity of the relative likelihood for the Z ethnic group compared to the baseline ethnic group is statistically significant and notable. Further information: Using relative likelihoods to compare ethnic disparities (Race Quality Unit)

Sanction decision statistics

The Universal Credit sanctions (Official Statistics) show numbers of decisions that have been referred to sanction. Breakdowns are available by

• age
• gender
• decision outcome
• decision type
• geography
• ethnicity

Universal Credit Sanction decisions by reason charts

Within the Benefit Sanction statistics publication there is a chart which breaks down the sanction decisions by reason. For the purposes of this chart, the reasons are grouped together, with the groupings shown below in Table 3 along with the sanction level to which each reason belongs. The table shows the different reason groups and their corresponding Sanction levels.

Universal Credit Adverse Sanction decisions by ethnicity

Data Source

Universal Credit claimants are asked to answer some optional equality questions This includes asking about their ethnicity, amongst some other protected characteristics (religion, marital status, sexual orientation and disability status).

A prompt to complete the equality questions appears in the claimants’ UC account when they complete their UC claim. If no response is given after 40 days the questions are cancelled. It is not compulsory for claimants to complete this survey, and their claim may still be progressed without this information. Agents cannot see this data when administering the claim. Furthermore, for each question the claimant may opt for the ‘prefer not to say’ response rather than provide the requested information. If they do not provide a response for any questions, their equalities record is not stored.

The level of ethnicity declarations for claimants included within the sanction rate measure, and other measures, have not yet reached the minimum level of 70% in any month, so ethnicity breakdowns of the sanction rate will not be available until this is reached. We will continue to monitor the level of ethnicity declarations in relation to the sanction rate, and other measures with the goal of publishing statistics on sanction rate and ethnicity once the 70% threshold is reached. The regular sanction rate data will still be available.

Summary statistics for Adverse sanction decisions by ethnicity and UC claimants in conditionality groups where sanctions can be applied are made available through ODS tables.

Ethnicity is made available as a breakdown into the 6 higher level ethnic groups including Unknown, which are consistent with the recommended Government Statistical Service (GSS) harmonisation ethnicity groups:

• Asian/Asian British
• Black/African/Caribbean/Black British
• Mixed/Multiple ethnic groups
• White
• Other Ethnic Group
• Unknown

The ‘unknown’ category is comprised of:

• claimants that did not answer the equality questions
• claimants that responded ‘prefer not to say’
• claimants whose ethnicity data was lost

These groups are split further into the ethnic sub-groups: for example, the ethnic group ‘Asian/Asian British’ is split into five subgroups: Bangladeshi, Chinese, Indian, Pakistani, Any other Asian background and Unspecified subgroup. Claimants can declare a higher-level ethnic group but choose not to state a detailed ethnic sub-group.

Given that the equality survey is optional, ethnicity is only recorded for a portion of the adverse sanction decisions. This means that there is an element of non-response which is quantified in the response rate. The response rate for ethnicity is calculated as the number of adverse sanction decisions giving an ethnicity response divided by the total number of adverse sanction decisions, including those who did not respond or chose ‘prefer not to say’ as their response.

Based on UC equality data analysis and wider stakeholder engagement, our standard minimum threshold for calculating representative percentages is a response rate of 70%. The response rate of the equality survey is monitored, and ethnicity and religion statistics will only be published if this is met.

The level of non-completion both indicates the level of uncertainty around the figures and means the likelihood of non-response bias is more prominent. Non-response bias is a particular type of data collection bias that typically happens when people who are unwilling or unable to respond to a survey have different characteristics to those who do respond. This can mean that different groups may be over or under-represented in the resulting survey estimates.

The effect of non-response bias could be quite substantial, and this should be considered when using Benefit Sanction ethnicity statistics. It should also be considered that ethnicity data is self-reported and cannot be verified. Ethnic identity can change over time and therefore claimants may now identify with a different ethnic group than the one they stated in their equality response.

If claims have more than one ethnicity declaration in the same day then we are unable to determine which is the latest. In these cases, if their information is consistent across all responses, we are able to include this information, but if there are conflicting submissions (for example, different ethnic groups selected each time), we code them as unknown ethnic group.

If an individual has declared different ethnicities for multiple adverse sanction decisions, then we have used their most recent ethnicity to populate all of their historic adverse sanction decisions.

Work is on-going to improve the response rate. The approach to capturing ethnicity information is being addressed to ensure that claimants are supported in providing the information. Claimants can be reassured that ethnicity information is used solely for statistical purposes in an aggregate fashion, non-attributable to individuals.

Limitations of the statistics

UC Live Service

Every quarter, when the latest figures are added to the previous sanctions figures, the previous sanctions figures will be updated to take account of any changes, such as decisions being challenged or changed, that have occurred. Only the latest time point of each decision in the process of sanctions: decisions, Mandatory Reconsiderations and appeals, is kept.

When collecting and presenting the statistics in this way it can cause confusion over the number of original decisions to sanction and therefore the proportion that have been changed. It also means that, as some cases move on to Mandatory Reconsideration and appeal, there will always be retrospective adjustments to earlier publications.

When it came to deciding how to count decisions, there were 2 options. The first option was to record only the latest decision on a case, the second to record every single decision. The first option gives information about the sanctions count, whilst the second option gives information on the decision making process. It was decided that the first option, the sanctions count, was the most important information in terms of public interest and informing sanction policy, with the only drawback that we cannot know when the original decision was taken.

Data on sanction decisions where the Decision Type is an appeal should be treated with caution. The data is currently under investigation due to a potential issue with the under-recording of Appeal decisions where the Decision Outcome is adverse.

Data published prior to November 2017 contained an issue whereby a small number of Mandatory Reconsideration decisions were incorrectly recorded with either a Cancelled or Reserved Decision outcome. This issue was corrected in the November 2017 release and affected 0.3% of Mandatory Reconsideration decisions and 0.2% of all decisions.

Users should note that there is a new methodology for allocating a claimant to a Census Output Area (COA) as of our February 2018 publication.

Residency based geographies are derived from address information as recorded on the Customer Information System (CIS). CIS is a more reliable source of addresses as it links to all of the DWP benefit systems and contains the most up to date address for each individual.

These addresses are then put through a data cleansing procedure which makes sure postcodes are formatted correctly and the address fields are populated correctly. 2011 COAs are then assigned to claimants using the ONSPD (ONS Postcode Directory), starting with a direct postcode to COA lookup and then working through a logical allocation routine. These COAs are then used to merge on higher level geographies from the National Statistics Postcode Look-Up (NSPL).

This file is then matched to the sanction decisions data, ensuring that the date of decision is between the address start and end of dates. If no known address can be matched for a period of time, a previous known address for that claimant is used, as this is more accurate than random allocation. There may still be a very small number of records that are still recorded as unknown.

All higher level geographies (Lower Layer Super Output Area/Data Zone, Middle Layer Super Output Area/Intermediate Zone, Local Authority, Region, Country) are derived from COA.

UC Full Service

It should be noted that UC full service statistics are produced using a different data source and methodology to that used to produce the live service statistics, and therefore the two should not be compared.

When the latest figures are added to the previously published UC full service sanctions figures, the sanction figures for all previous months will be updated to take account of any changes.

Unlike the live service data, we are currently unable to report on numbers of non-adverse, cancelled or reserved decisions. This is due to the fact that only information on adverse decisions is captured on the UC full service build. Figures on adverse decisions are recorded differently on the system than those on all decisions. In February 2020, it was determined that the previous figures on all decisions for UC full service did not include non-complex decisions, due to the way information is recorded. For this reason, this measure was withdrawn.

Within the adverse decisions data, we are aware that in any given month, up to 10% of the data may be made up of errors (4%), non-adverse Mandatory Reconsideration outcomes (5%), and non-adverse Appeal outcomes (1%).

Investigation of the data is ongoing and we aim to continue to improve our methodology in order to allow us to identify these cases and report them separately.

Whereas in the live service data, we only report the latest decision on each sanction, it is our aim to include all decision outcomes at each stage within our statistics for full service once development is complete. This means that if a claimant reaches appeal stage, the outcome of their original decision and Mandatory Reconsideration will continue to be included within the data, rather than being overwritten at each stage. This will allow us to more accurately follow the journey of the claimant through the sanctions process.

As previously indicated, at present, we do not have reliable data relating to Mandatory Reconsiderations or Appeals on Universal Credit full service. We are continuing to investigate the available data and hope to be able to produce statistics on Mandatory Reconsiderations and Appeals in the near future. For this reason, any adverse decisions reported in the full service statistics relate to only original, adverse decisions, other than the 10% described previously.

The data used to produce the full service statistics is taken from the Universal Credit full service build, which is the front end system used by both agents (Work Coaches and Decision Makers) and claimants. At present, the build is undergoing continuous development in order to make it as user-friendly as possible for both agents and claimants to use. For this reason, the variables we receive in each monthly data snapshot are not always consistent – variables may be added or removed at any time, which may affect the statistics we are able to produce.

Note that sanction decisions are often made and processed after the date that the claimant failure occurred. This means that it is possible for a sanction decision to be recorded against a month differing from the one in which the claimant originally failed to meet their conditionality requirements.

Table 3: Reason Group and Sanction Level

Reason Group	Sanction Level
Work Focused Interviews – Fail to comply with an interview requirement	Low
Work Focused Interviews – Fail to comply with an interview requirement (Self Employed)	Low
Work Focused Interviews – Fail to comply with a Work-Focused Interview requirement	Lowest
Availability for Work – Fail to accept a job	High
Availability for Work – Fail to apply for a job	High
Availability for Work – Fail to be available to take up work	Medium
Availability for Work – Fail to comply with a work preparation requirement	Low
Availability for Work – Fail to undertake all reasonable work search action	Medium
Availability for Work – Fail to undertake particular, specified Work Search action	Low
Availability for Work – Other Availability for Work reason	Low
Employment Programmes – Fail to participate in an Employment Programme	Low
Employment Programmes – Fail to participate in training	Low
Employment Programmes – Fail to undertake Mandatory Work Activity	High
Employment Programmes – Fail to undertake work experience or work placement	Low
Reason for Leaving Previous Employment – Leaving employment Voluntarily	High
Reason for Leaving Previous Employment – Loss of employment through Misconduct	High
Other (Not Included in Summary Charts) – Fail to comply with requirement to provide evidence or confirm Compliance	Low and Lowest (see note)
Other (Not Included in Summary Charts) – Fail to comply with requirement to report specified change in circumstances relevant to Work requirements	Low and Lowest (see note)
Other (Not Included in Summary Charts) – Lose pay through Misconduct	High
Other (Not Included in Summary Charts) – Lose pay Voluntarily	High

Note: The level is Lowest for those in the Planning for work conditionality regime, otherwise the level is Low.

Comparisons between the statistics

The nature of the process used to update the data (retrospection) means that it will be difficult to compare data between years as the data set is constantly changing as and when mandatory reconsiderations and appeals are resolved. This means that the previously released data will be updated every quarter as new information about each sanction is added to it.

The data shows that decisions can change significantly after one quarter. For live service, decisions can still change a year after the initial sanction has been applied, though the number of these changes is minimal.

Users should be aware that there are differences between sanctions policy in UC and the other benefits that will affect comparisons.

Whilst other countries also apply sanctions to their versions of Universal Credit, their systems differ too much to allow any comparisons to be made.

Rounding

Rounding has been applied to the sanctions statistics figures in the BSS. The headline figures and those accompanying the graphs have been rounded as follows.

Table 2: Rounding

From	To	Round to nearest
0	1,000	10
1,001	10,000	100
10,001	100,000	1,000
100,001	1,000,000	10,000
1,000,001	10,000,000	100,000
10,000,001	100,000,000	1,000,000

An exception has been made to the rounding in some sanction statistics tables. Some tables contain a range of numeric values and in these instances, the lower rounding band has been applied across the whole table.

We allow some time for additional information to be incorporated into DWP’s data systems. This time period is referred to as retrospection. Information may be submitted, corrected or resubmitted sometime after the event occurred. This means that data presented in each release is subject to some retrospection and figures may be revised in subsequent releases, especially the most recent months.

For example, if a decision to apply a sanction is made, the data may then be published before the next steps are commenced. If all the steps are gone through in order, it may be some time before the final stage is reached, sometimes a year after the initial decision to sanction has been made.

The methodology for how residence-based geographies are derived, as outlined in the Limitations of the statistics section, changed in the February 2018 release of the statistics. This means that figures, by residence-based geography (for example, local authority), released from this point will not be directly comparable to those published previously. The impact for UC will be minimal. The exact level of revision will vary dependent on the time period and type of geography selected. Users are therefore advised to recreate any figures that include a residence-based geography breakdown using the new data.

Official Statistics

These statistics are designated Official Statistics.

We have reviewed the Official Statistics in Development label on the Benefit Sanctions statistics, following the reinstated duration measures and rate methodology improvements. We have determined that the UC data and sanctions methodology is now stable and fit for purpose, and as such we have removed the “in development” label. As of 14 May 2024, these statistics are now published under the “Official Statistics” label.

We will continue to progress developments to meet user needs through the DWP Statistical Work Programme.

Quality statement

Our statistical practice is regulated by the Office for Statistics Regulation (OSR).

OSR sets the standards of trustworthiness, quality and value in the Code of Practice for Statistics that all producers of official statistics should adhere to.

In June 2023 the BSS publication underwent an OSR compliance check. This details our work to comply with the code of practice and our continued development to maintain our compliance.

You are welcome to contact us directly at epass.team@dwp.gov.uk with any comments about how we meet these standards.

Alternatively, you can contact OSR by emailing regulation@statistics.gov.uk or via the Office for Statistics Regulation website

The BSS are published quarterly in February, May, August and November. Statistical outputs are produced in a timely manner and dates of release are pre-announced in accordance with the Code of Practice. Should pre-announced dates of release need to be changed, or any issues occur impacting the coverage of the bulletin or delays to any aspect of the bulletin, we will provide an explanation on the statistics at DWP page. Information on pre-release access is published on the statistics at DWP page.

BSS are published through a range of dissemination methods including Stat-Xplore which is an online tool that lets you create and download customised statistical tables, and view the results in interactive charts. Methodologies, processes, and practices are documented, and documentation is kept up to date and made available publicly where appropriate. This documentation is regularly reviewed to follow any emerging best practice or to learn from experience. Currently the BSS is conducting a quality assurance report of Administrative Datasets used in the production of the BSS statistics, once complete this will be published.

Information on suspensions of data and measures can be found in the latest release of the BSS publication. Recent and ongoing development of the BSS statistics can be found on the DWP statistical work programme. Limitations of our statistics are also publicly published in this document. Please see limitations of these statistics section.

Definitions within the statistics

There are certain terms used in the Benefit Sanctions statistics that need explanation.

Sanction decisions

A claimant can be subject to sanction when they do not meet a condition of their benefit. The Decision Maker looks at the available information about the claimant and their referral and decides on an outcome. The decision made can be:

• Adverse
• Non-Adverse
• Reserved
• Cancelled

Definitions of decision types

Adverse

A decision to apply a reduction in payment if a claimant has not met the requirements of their claimant commitment Sanction duration.

In UC, the number of days where a reduction can be applied to a standard allowance. In legacy benefits, the length of a sanction as determined by the period of a reduction in benefit payment.

Non-adverse

Decision to not apply a sanction or lift one already applied.

Reserved

Decision to apply a sanction cannot be imposed because the claimant is currently not entitled to/claiming the UC standard allowance.

Cancelled

The referral is cancelled without a sanction being applied.

Standard Allowance (UC)

In UC, claimants are entitled to a Standard Allowance. This is the basic UC benefit amount dependent on age and single/couple status before any extra elements are added on such as support for children or disability and housing costs.

Assessment period

The monthly period over which the amount of claimants entitlement to UC is assessed and paid.

Non-compliance date

The date in which the claimant does not meet the requirements of their claimant commitment for their UC claim.

Decision date

The date in which a decision maker approves the sanction to be applied.

Feedback

To give feedback on the Universal Credit Sanctions Official Statistics email: epass.team@dwp.gov.uk

Or write to:

Useful links

Read the most recent and previous quarterly statistical summaries.

Read the JSA and ESA, UC and IS sanction statistics, including the links to the most recent statistics in data tables.

Use Stat-Xplore to tabulate any of the information available on sanctions.

Read the publication strategy for sanctions statistics.

Read all the statistics available from DWP.

Read a schedule for upcoming DWP statistical releases.