PAC: DWP excuses for “unprecedented and unacceptable” levels of benefit fraud and error don’t stand up

PAC reports today that “levels of fraud and error in benefit expenditure are unacceptably high”, with the Department for Work & Pensions overpaying an “eye-watering” total £8.6 billion across all benefits in 2021-22, with £6.5 billion of that figure due to fraud.

Fraud and error were rising year-on-year before COVID-19 and the NAO has qualified DWP’s accounts every year since 1988–89. During the pandemic “fraud and error rose to historic levels across the benefit system” reaching an estimated 7.6% (£8.5 billion excluding the State Pension) overpayment rate across all benefits, compared with 4.7% (£4.4 billion) in 2019-20.    

DWP maintains that current fraud levels are still due to COVID-19 but is unable to say when levels of fraud and error will fall. It has repeatedly claimed that there is increasing propensity to fraud in society in general since the pandemic but is unable to point to convincing evidence why this should lead to increasing losses to the taxpayer. 

The Department assumes that all claims from Universal Credit claimants who choose “not to engage” with DWP’s fraud and error measurement exercises are fraudulent but admits that it has no statistically significant information to support this view. The Committee warns on DWP’s ability to strike the right balance between being robust in tackling fraud and ensuring claimants are treated fairly, sayingDWP’s narrative about fraud in society “could encourage a complacent attitude toward unprecedented and unacceptable levels of benefit fraud” and mean that people “come to see committing benefit fraud as normal”. 

At the other end of the problem, benefits underpayments can lead to severe hardship. The Department estimates that 237,000 pensioners have been underpaid a total of £1.46 billion in State Pension, with underpayments going back as far as 1985. Work to rectify this is behind schedule and “efforts to correct the systemic underpayment of State Pension are too slow to meaningfully put things right”, and “will be too little, too late for many affected pensioners”.

The Committee remains unconvinced that DWP’s systems overall are “adequate to detect further underpayments before they build up into major issues in future”. 

Dame Meg Hillier MP, Chair of the Public Accounts Committee, said: “DWP is blaming everything from the pandemic to ills in wider society for unprecedented and wholly unacceptable levels of fraud in the benefits system. But the truth is losses to the taxpayer to fraud and mistakes have been at record levels and rising for years. DWP didn’t have a plan to get a grip on the billions it was losing every year before the pandemic, and it doesn’t have one now.”

PAC report conclusions and recommendations: 

8. Benefit fraud and error remains at a record high and is yet to show any sign of falling back to pre-pandemic levels. The estimated overpayment rate across all benefits, excluding the State Pension, was 7.6% (£8.5 billion) in 2021-22, compared with 4.7% (£4.4 billion) in 2019-20 before the pandemic. The Department acknowledges that fraud and error levels remain far too high, but it does not fully understand why this is the case, and it cannot say when they will meaningfully reduce. It asserts that the reasons for fraud and error remaining so high include the timing of its measurement exercise during a period when fraud and error controls were relaxed due to COVID-19, and an increase in the propensity to commit fraud across society since the pandemic. But it has not been able to explain in detail to what extent these and other factors are driving fraud and error, and therefore what residual level of fraud and error it expects once these controls have been reintroduced. The Department has still not published a fraud and error target, despite previously saying in response to a recommendation by this Committee that it would do so after the Spending Review in Autumn 2021. We are unconvinced by the Department’s assertion that it cannot set a meaningful target given the uncertainty in baseline levels of fraud and error in the benefit caseload, and in society in general. 

Recommendation: The Department should, by the publication of its next Annual Report & Accounts: 

• Set out its forecasts of the future levels of fraud and error in benefits, including its assessment of the factors driving these trends. 
• Set a target for fraud and error, working with the NAO to develop commentary alongside it to explain the context in the absence of a stable baseline.  

9. The Department risks allowing high levels of fraud and claimants disengaging with its compliance processes to become normal. The Department has repeatedly claimed that there is an increasing propensity to fraud in society in general since the pandemic. It believes that this is in part driving the record levels of fraud and error in the benefit system. However, it is unable to point to convincing evidence that increasing fraud in society must lead to increasing losses to the taxpayer. The Department does not fully understand the reasons why fraud and error in benefit payments remain at record levels. We are therefore concerned that the Department’s narrative about fraud in society could encourage a complacent attitude toward unprecedented and unacceptable levels of benefit fraud and that people come to see committing benefit fraud as normal. We are also concerned about the sharp increase in the number of claimants choosing not to engage with the Department’s sampling exercise to measure fraud and error. These claimants accounted for 14.4% (£852 million) of all Universal Credit overpayments in 2021-22, an increase of £328 million compared with 2020-21. The Department assumes that all such claims are fraudulent but admits that it has no statistically significant information to support this view. It is therefore unable to say whether claimants who fail to engage have done so because they are fraudsters or have been unwilling or unable to engage for other reasons. Given this uncertainty, the Department should do more to reassure itself that it has struck the right tone in its communications with claimants to encourage trust and engagement with its services.  

Recommendation:  Before the end of January 2023, we expect the Department to write to us with a clear plan of how it intends to increase the number of claimants responding to its fraud and error sampling exercises. In doing so the Department should consider reviewing the tone and content of its communications with claimants to both encourage compliance and catch fraudsters.  

10. The success of the Department’s strategy to bring down fraud and error is dependent on highly uncertain assumptions.The Department has set out its strategy to tackle fraud and error in Fighting Fraud in the Welfare System. This includes a £613 million investment in counter-fraud activities. The plan is dependent on the Department’s recruitment and training of enough people to implement its fraud counter-fraud activities in a difficult hiring environment, while simultaneously making headcount reductions in line with wider cuts to civil service staffing. The legislation needed to grant new powers the Department needs to deliver its counter-fraud strategy was not included in the Queen’s speech. Without new powers to access third-party data it will have limited ability to crack down on fraud driven by claimants misreporting the value of their savings or capital. The Department will be dependent on close cross-government working, including with the new Public Sector Fraud Authority, and is reliant on HM Revenue & Customs’ (HMRC’s) progress in collecting more timely information about self-employment earnings to address this driver of Universal Credit overpayments. 

Recommendation:As part of its Treasury Minute response, the Department should set out in detail how its counter-fraud plan would be impacted if it is not able to get the staff, powers or HMRC data that it needs, and what its contingency plans would be in these circumstances. 

11. The Department has not set out in sufficient detail how it will assess whether it is achieving what it wants from its investment in fraud prevention measures. We have previously found that the Department lacks the ability to demonstrate that its counter-fraud activities are having the intended impact and are cost-effective, and recommended that it work with the NAO to develop a consistent framework for reporting savings it generates for the taxpayer. The Department expects its £613 million investment in counter-fraud measures to generate £4 billion of savings over five years, but it has not set out in detail how the specific activities funded will produce this impact. In its 2021-22 Annual Report the Department reported a newly developed estimate that suggests its counter-fraud activities generated £2 billion of savings for 2021-22. However, this estimate requires further development before it can provide an appropriate reporting framework. The Department has committed to developing better, stronger metrics to demonstrate the impact and cost-effectiveness of its counter-fraud measures. 

Recommendation:We again recommend that the Department work with the NAO to ensure that by the time of its 2022-23 Annual Report and Accounts it has in place an agreed framework to report on the impact and cost-effectiveness of its counter-fraud activities. 

12. The Department’s lack of transparency over its use of data analytics risks eroding public trust in the benefit system. The Department’s strategy to bring down fraud and error will depend increasingly on the use of data analytics and machine learning to identify potentially fraudulent claims. It has trialled a model to detect fraud in Universal Credit advances. This uses historical fraudulent claim data to predict which claims are likely to be fraudulent in future and flag these to caseworkers for their review. The Department is aware of the potential for data analytics methods to generate outcomes that could have an adverse impact on certain claimants. For instance, some cases flagged as potentially fraudulent will turn out to be legitimate claims. If the model were to disproportionately identify a group with a protected characteristic as more likely to commit fraud, it could inadvertently obstruct fair access to benefits. The Department has taken steps to evaluate the potential impact of data analytics and machine learning on groups with protected characteristics, but the results are inconclusive and it has not made them public. The Department expects it will need to regularly update its assessment of the potential impact on vulnerable claimants as it develops its data analytics over time. 

Recommendations: The Department should report annually to Parliament on its assessment of the impact of data analytics on protected groups and vulnerable claimants. 

The Department should also consider what role the Social Security Advisory Committee can play in supporting public trust over the use of data analytics in the welfare system. 

13. The Department’s efforts to correct the systemic underpayment of State Pension are too slow to meaningfully put things right. The Department now estimates that 237,000 pensioners have been underpaid a total of £1.46 billion in their State Pension. Despite these underpayments going back as far as 1985, the Department’s overall exercise to correct this issue is delayed from the end of 2023 to the end of 2024. The Department cannot be certain that its plan to deliver the exercise on schedule is achievable, as it is dependent on assumptions around recruitment, retraining, and automation. We are not convinced that the Department has done enough to ensure its communications to potentially affected pensioners are sufficiently clear. We are concerned that this may leave many pensioners lacking reassurance that they will receive meaningful and timely redress. The Department does not yet know the full extent of the underpayment relating to Home Responsibilities Protection, and it is dependent on HMRC to evaluate the impact of these underpayments on pensioners. The Department cannot be certain that it has identified all the underpayments implied by the results of its annual measurement exercise. Overall, we remain unconvinced that the Department’s control systems are adequate to detect further underpayments before they build up into major issues in future. 

Recommendations:As part of its Treasury Minute response, the Department should set out a credible plan to deliver the exercise to correct State Pension underpayments on schedule and explain how it will update its communications to reassure pensioners that they will be meaningfully compensated. In addition to this, the Department should: 

• as part of its Treasury Minute response, work with HMRC to fully evaluate the extent of the HRP underpayment as soon as possible and provide a timetable of when it expects each phase of this process will be completed 
• by the publication of its next Annual Report & Accounts, set out a plan and timetable for introducing a measure to report the total value of arrears payments that arise due to underpayments, and how it will review individual arrears payments to assess whether they are indicative of a systemic underpayment issue.