Guidance

Benefit sanctions durations and sanction rate calculation: explanation of methodology

Updated 14 November 2023

This guidance was withdrawn on 13 February 2024

This background information and methodology document is no longer in use. Read the methodology for the benefit sanction duration statistics and sanction rate that is used for statistical releases from February 2024 onwards.

As of the August 2017 publication of the Quarterly Statistical Summary (QSS), statistics on the length of time a reduction in benefit due to a sanction became available. This has enabled, for the first time, durations of a sanction to be produced for:

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

Income Support (IS) durations will become available in the future, when further data is available.

Further work is ongoing to continue to refine and improve the methodology used to calculate sanction durations.

Context of the statistics

These statistics are comprised of 3 benefits:

• JSA
• ESA
• UC

It is intended that IS will be incorporated in a future release.

The statistics for JSA and ESA were primarily developed by using payment data to identify payment drops that result from sanctions. The start date of the sanction was identified as the date of the first reduced payment from a sanction, and the end date was identified as the date where there was a pause or stop in deductions, or the claimant flowed off benefit. Any durations or rates analysis presented for JSA and ESA contains information regarding sanctions that have already ended by the time of analysis. This is because a claimant’s payment might drop for a number of reasons, and not only as a result of a sanction (for example, the claimant might have a change in circumstances which results in their payment being reduced). Due to this, the methodology used for calculating sanction durations and rates only includes payment drops which are followed by a subsequent payment increase (or followed by the claimant flowing off benefit). This helps to prevent the methodology from falsely counting a sanction, where a claimant’s payment has been reduced for any reason other than them being sanctioned. As a result, sanctions that are still ongoing at the time of analysis are excluded from calculations.

The underlying data for durations can be accessed through Benefit sanctions statistics supporting tables.

UC data already includes relevant information on payment drops. The data used to calculate sanction durations includes both full and live service sanctions.

For UC there can be cases where a claimant stops receiving any standard allowances due to high earnings, meaning that there is no remaining standard allowance in which a deduction can be taken, or where they have moved to a conditionality regime in which they can no longer receive a deduction. For these cases the first point in which these deductions stop will be classed as an end point. This means that one decision could be counted twice in these statistics when a claimant then goes on to receive further deductions for one sanction following a pause in deductions. Unlike the other UC sanctions statistics, these statistics cover both live and full service, thus capturing all those on UC.

JSA and ESA sanctions decisions data does not include information on benefit payments, and therefore work was required to match sanction decisions to payments drops by linking data. This was achieved by matching the decisions to sanction to the National Benefit Database (NBD) to look for dates that the claimants flow on and off benefit.

The benefit payment drop data from General Matching Service (GMS) scans was merged on to the sanctions decisions data. This enabled us to determine the amount of benefit paid every week. Drops and increases in benefit are identified from the payment data and linked to a sanction decision. The claim start and end dates were also combined with the data.

When linking a sanction decision to a payment on GMS, we looked four months either side of the sanction decision to identify the corresponding payment change/drop. This enabled us to determine the start and end date for a sanction, and hence the sanction duration.

For all benefits we have produced the median length of a sanction as a key measure; 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.

The sanction durations that are published from August 2017 are the length of time that claimants have a reduction in their benefit due to a sanction. To note, when a claimant leaves benefit following a sanction start, but before the sanction is served, the claim end date is taken to be the sanction end date also. In the cases where a sanction end date is the same date as the sanction start (the claimant serves zero days of sanction), the cases are not counted as a sanction served and are excluded from the durations data.

Another advantage of identifying sanction start and end dates is that it allows a Point-in-Time rate of claimants undergoing a sanction to be produced.

Sanction Rates

In May 2016, a Monthly Rate of Claimants Sanctioned was first published in the Quarterly Statistical Summary for JSA and ESA sanctions. The rate was calculated to give an estimate of the number of sanctions relative to the number of JSA claimants and the number of ESA claimants. These statistics were classed as official statistics in development as the methodology was new and subject to change, if an improvement could be found for calculating the rate.

Previously the sanction rate was calculated for JSA as the number of decisions to apply a sanction (adverse)^{[footnote 1]}. in a full calendar month divided by the monthly JSA claimant count (from NOMIS), which relates to the number of people claiming JSA on the second Thursday of the month.

Similarly, the sanction rate for ESA was calculated as the number of decisions to apply a sanction (adverse)^{[footnote 2]}. in a full calendar month divided by the ESA claimants in the WRAG group on the last day of a calendar month. This claimant count figure is based on the quarterly claimant count figures, and the intervening months are derived from internal monthly datasets. The ESA monthly counts are now published as part of the ESA Supporting Tables (14).

The start and end date sanction data allows for an improved Point-in-Time rate to be produced. This rate counts the number of people with an ongoing sanction at each of the claimant count dates and presents it as a proportion of the number of people on benefit. This rate uses the same claimant count data as the previous rate for JSA and ESA. For ESA only claimants in the WRAG group are counted. Note that sanctions that are still ongoing at the time of analysis are excluded from calculations – see the ‘Context of the statistics’ section above for further information.

Up until November 2018, for UC, claimants in all conditionality regimes were taken from the published People on Universal Credit statistics^{[footnote 3]}. It was chosen to include all conditionality regimes due to the possibility of claimants with a deduction changing conditionality regimes between the point in which a deduction was calculated and the count date. The People on Universal Credit figure is based on the number of people claiming UC on the second Thursday of the month. However, as Universal Credit full service continues to roll out, a larger proportion of claimants are moving into the “Working – no requirements” and “No work requirements” conditionality regimes. A more accurate methodology was developed for the November 2018 publication in which all claimants with a drop in benefit due to a sanction are divided by all claimants in regimes where a sanction can be applied. This reduces the number of claimants in the bottom number (denominator) of the equation as it removes claimants who are in the working- no requirements and no work requirements regimes, and gives a more accurate rate. Unlike the methodology used for JSA and ESA (WRAG), the methodology for UC full service includes open sanctions in the rate.

Where the previous methodology used Adverse decisions in a month to calculate the rate, the new methodology uses the date of the claimant count for each benefit to calculate the number of claimants with a reduction to their benefit as a result of a sanction on the same day in the month that the claimant count was taken, a Point in Time rate.

In November 2021 a change was made to the UC sanction rate definition to bring this in line with operational implementation of sanction policies. Claimants in the “working with requirements” (light touch) conditionality regime have their work-search and work-preparation requirements switched off. As such, the definition used to calculate the rate has been changed to properly reflect the implementation of the policy, removing claimants in this regime from the count of UC claimants subject to conditionality (which is used in the rate calculation). There is no material impact on the long-term trend as a result of the definition change.

The underlying data for the durations statistics and calculation of the rate are published as part of the Benefit sanctions statistics supplementary tables.

Purpose of the statistics

These statistics have been developed for three reasons:

• to improve the understanding of the actual length of each sanction
• to improve the understanding of how many claimants are undergoing a sanction at a particular point in time
• to allow the development of work looking at the next destination of claimants who have been sanctioned (have they gone into work, moved benefit, remained on the benefit)

Limitations of the statistics

There are some limitations with the methodologies used in calculating sanction start and end dates and durations. These are outlined below:

JSA and ESA

1. The payment data had to be merged on to the sanction decisions data. This left some cases where an adverse decision could not be matched to a corresponding payment drop (around a fifth of JSA and ESA adverse decisions). Some of the reasons for the lack of payment data include: adverse decisions that are then changed (for example, a decision review) before the benefit is sanctioned, claimants receiving hardship payments, meaning that no drop in payment is observed, claimants are immediately sanctioned, meaning that they receive a reduced payment from the start of their claim and no drop in payment is observed, or claimants leaving benefit before their benefit is sanctioned.

2. When claimants receive multiple sanctions over a short period of time, the sanctions run concurrently (at the same time). This means that a sanction starting a few days after a previous sanction (assuming they are the same length) will end the same number of days after the original sanction. Therefore, any multiple sanctions that have no obvious start and end dates are treated as one sanction. This has two impacts on the data. One, that occasionally multiple sanctions will only show up as one sanction, and two, that these sanctions will have their length extended beyond what would be expected.

3. For JSA, non-complex Failure to Attend decisions that are made by a Jobcentre Decision Maker are not included in the durations data as they are recorded on the LMS database. This accounts for a small percentage of all decisions made since the regime change in October 2012. They may be included in the future when further work has been done on matching to the LMS database.

4. A further limitation of the current system is that some week-long JSA and ESA sanctions are missing from the data. This is due to an issue with tracking payment drops of a week or less as they do not always appear on the benefit payment database.

5. Where there are small drops in payments (less than 5%) which may/may not relate to a sanction (this happens less than 0.1% of the time), in a very small amount of cases we may be starting the sanction earlier than it is supposed to. This will lead to a slight overestimation of sanction duration for these claims.

6. Where a payment drop is followed by a second payment drop, the methodology correctly identifies the start dates relating to each drop and records 2 sanctions. One of the sanctions will be prematurely ended when the other ends – as we identify end dates when a payment increase occurs. This means we may be slightly underestimating one of the sanction durations in this situation.

7. We end a sanction when the claimant flows off benefit. Sometimes they flow off before the payments stop, which means we don’t capture the earlier end point, rather the end date relating to their next claim. This results in an overestimation of sanction duration for these cases. To get around this, claims that have JSA sanction durations over 3 years (the maximum possible) are removed from the final dataset on quality grounds. These occur in less than 1% of the total number of cases. However, this does mean that some sanction durations may still be overestimated due to a lack of matching NBD end date.

8. Recorded sanction length is impacted by benefit off-flows and the specific nature of the benefit received. We have found that people are more likely to flow off and on JSA, therefore immediate JSA off-flows are excluded to ensure the sanction length used is representative of the full sanction duration. Off- and on-flows are rare on ESA so the full sanction length is more likely to be captured in the first instance.

9. The sanction duration methodology does not capture any outstanding portion of sanctions unspent on return to benefit.

10. There have been some retrospective changes to our reported JSA and ESA rates as a result of operational processes, which have occurred due to the migration of legacy claimants to UC. Claimants are included in the JSA or ESA rate only when their sanction has ended, or their claim has closed. This is because the claimant must have had a drop in payment (due to a sanction) and a corresponding rise in payment (when the sanction ends) in order to calculate the sanction length. An increasing number of JSA and ESA claims are now being closed when claimants are being moved onto UC Full Service, which has caused an increase in the number of claimants included in the rate, predominantly for long-term, open-ended sanctions that have been ongoing for some time. This has primarily occurred with ESA sanctions, where some sanctions are open-ended until the claimant either complies or flows off benefit. As the maximum length of a JSA sanction is three years, there has not been any retrospective change as a result of this issue prior to 2014.

Universal Credit

1. When claimants receive multiple sanctions over a short period of time, the sanctions run consecutively (one after the other). This means that as soon as one sanction finished, any other sanction will start immediately. Therefore, any multiple sanctions that have no obvious start and end dates are treated as one sanction. This has two impacts on the data. One, that occasionally multiple sanctions will only show up as one sanction, and two, that these sanctions will have their length extended beyond what would be expected.

2. If a claimant serves a sanction which later goes on to be overturned, data used to calculate the sanction duration is removed from the system on the front end. To capture overturned sanctions in our statistics, we merge the data produced during our previous release to include any sanction spells for sanctions that have since been overturned. Due to the fact that the Benefit Sanctions Statistics are produced quarterly, our methodology is not able to capture portions of sanction spells that are served after the latest release and overturned before the following one. This means we might be slightly underestimating sanction durations.

3. Universal Credit claimant count statistics released in October 2017 incorporated improvements to the processing of data for the people on Universal Credit. As a result of these improvements, the People on Universal Credit series was revised back to December 2014. In turn, this change in the claimant count affected the rate of sanctions calculation back to August 2015, though the effects have been slight.

For all benefits once the methodology is finalised, the rate will not change retrospectively, which means that any adverse decision that is subsequently overturned will still exist as an adverse decision at the point at which the sanction was served. This is done to help the final sanction durations and rate statistics represent the true impact that sanctions have had on claimants and their benefit awards, regardless of whether their sanction was later overturned or not.

4. We have become aware that there are 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, overall there is no material impact to the sanction rate, however some degree of caution should be applied when considering underlying figures for numbers of sanctions or caseload counts.

As these are official statistics in development, we hope to address some of the above issues as we become more familiar with the data and underlying patterns.

Comparisons between the statistics

Comparisons should not be made between the durations statistics and the statistics using the decisions methodology as the underlying data is not compatible.

The decision data uses all decisions made within a month which means they look at original decisions, decision reviews, Mandatory Reconsiderations, and Appeals, as well as adverse, non-adverse, cancelled, and reserved decisions.

The durations data does not distinguish between stages or outcomes of decisions, but instead looks at reductions in benefits due to a sanction. This means that a claimant’s benefit may be reduced/removed due to an original adverse decision which is then overturned upon reviewing the decision. Whilst in the decision data, this person would appear as a changed decision at decision review stage, in the durations data they will just appear as having lost their benefit for a time. The aim of this is to give a truer reflection of the effect that a sanction decision has on a claimant. Another difference between the decisions data and the durations data is that a claimant will only appear once in the decision data as only the latest decision for each claimant is kept. However, in the durations data, the claimant will appear in as many months as it takes for their sanction to end. This may be one month or less, but it may be 6 months if it is a six-month sanction. This makes it impossible to relate the decision and durations data.

Source of the statistics

The data comes from the following systems.

• JSA and ESA sanctions come from the Decision Making and Appeals System (DMAS).
• JSA and ESA benefit payments information comes from General Matching Service (GMS)
• JSA and ESA flows on and off benefit comes from National Benefit Database (NBD)
• UC live service sanction and benefit payment information comes from the Payment Manager Systems (PMX)
• UC full service sanction and benefit payment information comes from the Universal Credit full service

Definitions within the statistics

Sanction

A reduction or loss in payment until a claimant re-engages, or for a fixed period

Disallowance

The claim to benefit is ended, and claimants have to re-apply for the benefit (JSA only)

Intermediate Sanction

In JSA, a claimant will receive a disallowance to benefit for intermediate level sanctions. The reasons for an intermediate sanction are: Not Actively Seeking Employment, and Not Being Available for Work

Duration

The length of a sanction as determined by the effect it has on the claimant’s benefit

Standard Allowance

In UC, claimants are entitled to a Standard Allowance. This is the basic UC benefit amount before any extra benefit is added on such as child benefit, housing costs, and work allowance.

WRAG – Work-Related Activity Group for ESA

Claimants in this group are expected to prepare for work, as opposed to those in the Support Group who have no requirements placed upon them.

Revisions to the statistics

DWP have 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 sanction statistics, to reflect any updates to the figures, the full historic statistical series is refreshed each time the figures are released.

Status of the statistics

Official statistics in development

In line with recent Office for Statistics Regulation (OSR) guidance, all statistics previously referred to as ‘experimental’ will now be referred to as ‘official statistics in development’.

The duration statistics and monthly rate calculation are classed as official statistics in development. They are classed as such because they are recently developed statistics and may be subject to revision. Read more information on official statistics in development.

Quality Statement

These statistics have been developed using guidelines set out by the UK Statistics Authority, and are new Official Statistics undergoing evaluation. They have, therefore, been designated as official statistics in development. Users are invited to comment on the development and relevance of these statistics at this stage.

Feedback

We welcome feedback

To give feedback on the duration statistics or the monthly rate of sanctions:

Email: epass.team@dwp.gov.uk

Write to:

Tracy Hills
Development Team Client Statistics
Digital, Data and Analytics
Room BP5201
Benton Park View
Longbenton
Newcastle-upon-Tyne
NE98 1YX

Useful links

See the Benefit Sanctions Statistics publication for sanction statistics from November 2017 onwards

See the Quarterly Statistical Summary where benefit sanctions were previously published

See the JSA, ESA, IS and UC sanction statistics, including the links to the most recent statistics in Excel tables, which include the durations underlying statistics, the link to this document, the JSA, ESA and IS Background Information and Methodology document, the UC Background Information and Methodology document, and the benefit destinations methodology document.

Read information on official statistics in development.

1. available from Stat-Xplore ↩

2. available from Stat-Xplore ↩

3. available from Stat-Xplore ↩