Guidance

Changes in the fraud and error in the benefit system: financial year 2022 to 2023 estimates

Published 13 April 2023

Introduction

This statistical notice is about changes to the Fraud and error in the benefit system publication and to provide advance notice of 3 revisions.

The changes and revisions in this statistical notice will take effect for the Financial Year Ending (FYE) 2023 publication, due for publication on 11 May 2023.

The revisions are:

• the proportion of State Pension claims with an overpayment or underpayment and the proportion of claims with a de minimis overpayment or underpayment

• the monetary value and rate of State Pension overpaid

• the monetary value and rate of Housing Benefit overpaid and underpaid

The changes are:

• inclusion of estimates of fraud and error on Cost of Living Payments

• publishing a new measure on the total proportion of claims with an error and net reporting on the proportion of claims that are overpaid or underpaid. This includes publication of a new table and reordering the existing tables

• applying the 10p de minimis, introduced for State Pension uprating errors in FYE 2022, across all benefits in the incorrectness tables relating to the proportion of claims with an overpayment or underpayment

• updating the proxy benefits used to estimate benefits which have never been reviewed

• capping Housing Costs fraud and error on Universal Credit at the amount of housing element the claimant is receiving

• refining the classification of claims which fail to provide evidence using additional data

• updating the estimate of the Universal Credit advances omission in the background information document

These changes are made as part of the continuous review of the methodologies used to produce the statistics.

Revised figures for FYE 2022 will also be published within the Fraud and error in the benefit system: FYE 2023 estimates. This will allow a like-for-like comparison with the newly published figures for FYE 2023.

Due to the amount of resource required to apply the methodology changes to all historic data, we are unable to provide an updated time series prior to FYE 2022.

Further detail on all these changes will be presented in the background information document that will accompany the Fraud and error in the benefit system: FYE 2023 estimates.

If you have any queries on these changes, please send them to: caxtonhouse.femaenquiries@dwp.gov.uk

Background

The Department for Work and Pensions (DWP) uses a complex methodology to measure the levels of fraud and error in the benefit system. Read more information on the methodology used in the background information document.

Revisions

Revision of the proportion of State Pension claims with an overpayment or underpayment and the proportion of claims with a de minimis overpayment or underpayment

Last year was the first full review of State Pension since FYE 2006.

We have continued to review our categorisation methodology for State Pension against the benefit’s regulations and legislation. We have identified that we had incorrectly recorded errors (the majority of which were 1p or 2p) on some claims that should have been recorded as correct.

We will revise the FYE 2022 State Pension figures affected in the upcoming publication. It should be noted that the affected figures, on proportion of claims with an overpayment error and proportion with an underpayment error, will also be impacted by the upcoming methodology change in applying the 10p de minimis. Further information on this change is provided within the ‘changes’ section of this document.

Revision of the monetary value and rate of State Pension overpaid

As the reviews for State Pension only cover claims from Great Britain, we also include an additional amount of Claimant Error to estimate the impact of non-notification of death on State Pension claimants living overseas. Last year we conducted a full review of State Pension claims for the first time since FYE 2006. As part of that we updated the methodology used to generate this additional amount. This estimate was based on the life certificate and death exchange data. Last year this estimate accounted for around 40% of the total State Pension expenditure overpaid.

This year, whilst looking to update and refine the new methodology developed last year, we have found inconsistencies and gaps in the underlying data which has led us to question its quality. If this had been identified last year, we would not have made the methodology change that we did. Therefore, whilst we work to develop an updated estimate, we will return to using the rate we had been using before FYE 2022 for this additional amount. This rate relates to a measurement exercise conducted in 2004. We will revise the figure published in FYE 2022.

Revision of the monetary value and rate of Housing Benefit overpaid and underpaid

Last year we revised our estimates for FYE 2021 and FYE 2020 due to a change made to the Housing Benefit expenditure data. The Housing Benefit expenditure data used in the calculations previously included Universal Credit as a non-passporting benefit. We adjusted the expenditure data to categorise Universal Credit as a passporting benefit because people in receipt of Universal Credit who are in supported, sheltered or temporary housing are treated similarly to those claimants in receipt of other income-related benefits.

Whilst we used the best information available to us at the time to estimate the expenditure splits, over the past year work has been conducted to formalise this methodology which has resulted in minor changes to the estimates. We will be using this data for FYE 2023 and in future publications. To allow consistent comparisons with the previous year we are revising our figures for FYE 2022.

Effect

The 2 revisions to State Pension led to slight changes in overpayment and underpayment monetary amounts and the associated rates of expenditure overpaid and underpaid, and small changes in the proportion of claims with an overpayment or underpayment. However, they have a larger impact on the proportion of State Pension claims with a de minimis overpayment or underpayment.

Making the revision to Housing Benefit led to small changes in overpayment and underpayment monetary amounts and the associated rates of expenditure overpaid and underpaid.

The combined impact of these revisions may have an impact on the Global (all benefits combined) estimates.

Changes

Inclusion of estimates of fraud and error on Cost of Living Payments

Cost of Living Payments were made by the department for the first time in FYE 2023. Guidance can be found in the Cost of Living Payments 2022 document. Since our statistics give an estimate of fraud and error on all benefit expenditure, we are including an estimate of the amount of fraud and error associated with these payments in the upcoming publication.

We have not conducted benefit reviews on Cost of Living Payments, so the estimate will be included within the unreviewed benefits section of the supplementary tables. Given the public interest in these payments, we will include a section within our main publication document that will provide information specifically relating to Cost of Living Payments (separated out from the other unreviewed benefits).

If a claimant is not eligible to receive a qualifying benefit, then they would also not be eligible to receive a Cost of Living Payment. Therefore, to derive an estimate for the rate of fraud and error on these payments, we will use the rate of claims that lose entitlement on the qualifying benefits (the majority of which have been measured in the current or recent years).

Publishing a new measure on the total proportion of claims with an error and net reporting on the proportion of claims that are overpaid or underpaid

We currently report the proportion of claims with an overpayment error and the proportion with an underpayment error, which we call incorrectness. However, a benefit claim can have multiple errors found on it, which can be a combination of overpayments and underpayments. Our methodology for monetary amounts nets overpayments and underpayments against each other, but this is not currently carried out when we calculate incorrectness.

To improve consistency of the measures within our publication, and to make our statistics more easily understandable for users, we will be including additional figures on net incorrectness. The net incorrectness figures are the proportion of claims that are ultimately overpaid or underpaid, and a total incorrectness figure which provides an estimate of the proportion of claims that are incorrect.

These additional figures will be included in a new table within our supplementary tables. To ensure that the supplementary tables remain in a sensible order we will be reordering them, grouping them into tables related to monetary amounts, tables related to incorrectness and finally the table on net loss.

Applying a 10p de minimis across all benefits for the proportion of claims with an overpayment or underpayment

Last year we applied a 10p de minimis to State Pension uprating errors. This meant that we removed all uprating errors of 10p or less from the tables relating to the proportion of claims with an overpayment error and the proportion of claims with an underpayment error and reported them separately within our main publication.

In the FYE 2023 publication we will be applying the 10p de minimis to all errors, on all benefits. The proportion excluded by the de minimis for each of the headline breakdowns will be included in the following supplementary tables: the proportion of claims with an overpayment error; the proportion of claims with an underpayment error; the new table on proportion of claims overpaid/underpaid discussed above.

We will be revising the FYE 2022 figures to align with this methodology change to allow for direct comparisons.

Update to the proxy benefits used to estimate benefits which have never been reviewed

For benefits which we have never reviewed we use rates from similar benefits as a proxy.

Over the last year we have conducted a review of the benefits which we use as proxies. Previously we used the rates for whole benefits as proxies. For example, we used the Employment and Support Allowance rate as a proxy rate for Maternity Allowance. However, we are now taking a more refined approach and only using the elements of the proxy benefit that relate to the particular unreviewed benefit. For example, for Maternity Allowance we are continuing to use Employment and Support Allowance as it has similar eligibility requirements about work and National Insurance contributions, as well as covering the target working age population. However, we will only take the error rates found on Employment and Support Allowance relating to Abroad, Conditions of Entitlement, Earnings and Contributions as they are the only error reasons that could occur on Maternity Allowance.

This will impact on the unreviewed benefit lines in the supplementary tables as well as potentially having a small impact on the “All benefit” lines.

Other changes

We are making 3 other smaller changes as part of the regular review of our methodology for consistency and accuracy purposes. The changes are as follows:

• incorporating the use of more data when categorising frauds that are classified as ‘failure to provide evidence’. This will allow us to categorise more claims with known reasons for Fraud

• capping Housing Costs fraud and error on Universal Credit at the amount of housing element the claimant is receiving. Previously we have been overstating the amount of Housing Costs Fraud on those claims that do not provide evidence to us as part of their review. If this then leads to suspension and termination of the claim, we assume that the claimant was not entitled to the benefit. If the only evidence not provided is a tenancy agreement, we have been allocating all of the Fraud to Housing Costs. We are changing our methodology so that the Housing Cost Fraud will be capped at the housing element of the claim and the remaining Fraud will be allocated to the ‘failure to provide evidence/fully engage in the process’ category

• updating the estimate of the Universal Credit advances omission in the background information document. We have previously quantified the omission using a National Audit Office publication from 2020 based on FYE 2019 data. However, last year this was felt to no longer be representative of the current situation. A new methodology has been developed to estimate fraud and error on advances and this was used to place an estimate in DWP’s Annual Reporting and Accounts last year. We plan to use the new methodology to quantify this omission in future

The first 2 changes will not have any impact on the overall figures but will change lower-level reporting on the ‘reason’ for fraud and error. To allow for a direct comparison, we will be revising the FYE 2022 figures for all benefits, at this lower level.

The final change on the Universal Credit advances omission does not directly affect our reported figures and is used to provide added context.