Fraud and Error Contents

---

1Government’s understanding of fraud and error risks and the resources needed to tackle them

1.On the basis of a report by the Comptroller and Auditor General, we took evidence from Cabinet Office, the Department for Business, Energy and Industrial Strategy (BEIS), the Department for Work and Pensions (DWP), HM Revenue & Customs (HMRC) and HM Treasury about fraud and error.1

2.The Cabinet Office told us that fraud accounts for 40% of all crime committed across the UK.2 It estimated that in 2018–19 fraud and error cost the taxpayer £29.3 billion to £51.8 billion, around £26.8 billion of which was attributable to the tax and benefits system. The rest is based on Cabinet Office’s assumption that fraud and error is likely to be in the range of 0.5% and 5% for the £503 billion of government expenditure where fraud and error is not measured.3 This range is based on academic research and fraud measurement activities in the United States government, the European Union and the private sector, and 24 measurement exercises undertaken since 2015 by Cabinet Office across £3.6 billion of expenditure.4

3.Each Department is responsible for managing its own risks of fraud and error which has led to varying approaches and capabilities across government.5 In 2018, the Cabinet Office established a Government Counter Fraud Function which works to increase the understanding of fraud risks and threats to government by instilling professional standards and bringing together the 16,000 counter fraud professionals across the public sector.6 HM Treasury is responsible for setting out the counter fraud requirements for government departments and approving spend on initiatives subject to fraud and error and on counter fraud and error activities.7

Fraud in COVID-19 support schemes

4.The government introduced several large-scale schemes in response to the COVID-19 pandemic which provided vital support to many vulnerable businesses and individuals. However, the NAO concluded that the risk of fraud and error has risen significantly as a result of the government’s response to the COVID-19 pandemic.8 This was the result of a variety of factors, including that government had spent more on areas that were prone to fraud and error, such as welfare, business support and grants, and that it had often prioritised the need for speed when setting up new initiatives over reducing the risk of fraud and error. The increased risk was also due to providing support to people and businesses that government did not have a prior relationship with, relaxing or modifying controls in place to prevent or detect fraud and error, prioritising its COVID-19 response over business-as-usual compliance activity; and a general increased risk appetite for fraud and error, as shown by ministerial directions accepting the risks identified by the civil service. Early indications are that fraud and error has risen by billions as a result of the COVID-19 response, but the actual amount will only become clear as departments measure the level of fraud and error across specific initiatives.9

5.BEIS, DWP and HMRC are responsible for some of the support schemes Cabinet Office identified as having the highest risk of fraud or error.10 We have previously reported on several of these schemes, including their consideration of fraud and error risks. For DWP, we concluded that COVID-19 will lead to further increases in fraud and error, but it has an opportunity to learn from the impacts of its control easements.11 In our report on the Bounce Back Loan Scheme, we found that Government lacks the data needed to assess the levels of fraud within the Scheme.12 In October 2020 we reported that although HMRC was yet to see the full effects of COVID-19 on taxpayer compliance, it was already estimating up to £3.5 billion of fraud and error in furlough payments.13 In December 2020 we reported our concerns that HMRC still did not know the actual level of fraud and error in the Coronavirus Job Support Scheme (CJRS) and the Self-Employment Income Support Scheme (SEISS) schemes and would not have a complete estimate until the end of 2021 at the earliest.14

6.We asked HMRC, DWP, and BEIS how the government’s response to the pandemic had changed the risk of fraud and error within their areas and how they had responded. BEIS told us that launching multiple large-scale support programmes had markedly changed the fraud risks it needed to manage. It explained that prior to the pandemic it had considered itself a “relatively low risk environment for fraud” and that it had good controls in place to deal with the very traditional risks that it faced, for example around procurement.15 However it told us that it had found itself in a new environment as a result of the pandemic, which required bigger interventions, and was now “doing a lot more work on fraud”.16 BEIS recognised that loan programmes, such as the Bounce Back Loan Scheme, had left it reliant on banks that lacked incentives to prevent, detect and correct fraud and error given it is not their money on the line.17 As previously reported, BEIS estimates it could lose £16 billion to £27 billion through fraud or credit risks on loans issued under the Bounce Back Loan Scheme.18

7.BEIS is also responsible for several government support schemes delivered by local authorities, such as the Small Business Grant Fund and Retail, Hospitality and Leisure Support Grant Fund. In our recent report on local government finances we noted that government support schemes during the pandemic were not always designed with sufficient knowledge of local government finance or input from the sector, and that the focus on speed of delivery did not always take account of the need for assurance against fraud.19 In written evidence the Fraud Advisory Panel also told us that for the COVID-19 schemes local authorities were asked to make payments ‘as quickly as possible’ leading to a high likelihood of fraud and error.20 When we asked about counter fraud capabilities in local authorities, BEIS told us there is a variable level of practice but as local authorities regularly deal with fraud as part of their day-to-day business so they are “probably even better equipped” than BEIS to handle these risks.21

8.Fraud and error within Universal Credit rose by £3.8 billion to an all-time high of £5.5 billion between April 2020 and March 2021. DWP told us it that during lockdown it found opportunities for innovation by adapting traditional controls, such as face-to-face meetings with claimants, for remote working. As an example, DWP described how from June 2020 it introduced enhanced biographical questions from June 2020 to verify identity over the phone, leading to fewer cases referred to its fraud checking service.22 We have previously reported that HMRC has carried out fewer compliance investigations since lockdown began in March 2020, as it had to prioritise the implementation of COVID-19 support schemes and be responsive to the needs of taxpayers struggling with the impacts of the pandemic. The number of completed civil compliance checks fell from 62,000 in the first quarter of 2019–20 to 40,000 in the first quarter of 2020–21. HMRC estimates up to £3.5 billion of furlough payments made by 16 August 2020 may have been fraudulent or paid in error.23

Understanding of fraud and error across government

9.We have regularly reported on HMRC’s and DWP’s efforts to tackle fraud and error in the tax and benefit systems. DWP’s accounts have been qualified every year since 1988–89 due to material levels of overpayments and underpayments in benefits expenditure.24 In our report on their 2019–20 accounts we concluded that while DWP is good at understanding the types of fraud and error in the benefits system it is unable to demonstrate the cost-effectiveness of its counter fraud and error controls. We also reiterated our intention to hold DWP to account for improving its performance, aided by the fraud and error targets it has agreed to set.25 We have previously criticised HMRC for basic errors in financial forecasting including its uncertainty around what its estimate of fraud and error from tax credits should be and delays in producing a more rigorous estimate of the level of fraud and error associated with the Research & Development relief.26 We have previously concluded HMRC’s COVID-19 support schemes have led to a major reprioritisation of its resources and it has needed to reduce compliance activity while under lockdown. This has adversely affected HMRC’s core compliance activities and led to a backlog of investigations.27

10.We asked Cabinet Office about fraud and error risks outside the tax and benefits system. Cabinet Office told us that the diversity of risks is “very, very large”, because government spends £850 billion and collects around £800 billion in income. It explained that this covers all the different grants, procurement, and services that the public sector offers, which have a range of different fraud risks.28 Cabinet Office explained that the COVID-19 response has changed the fraud risk profile for many departments who are now exposed to higher levels of risk and loss than they had experienced previously.29 It acknowledged that it still needs to do more work to understand the broader fraud and error picture outside the remit of DWP and HMRC.30

11.The Counter Fraud Function undertook a Global Fraud Risk Assessment across 206 COVID-19 response schemes, with an estimated total value of £387 billion. It risk-assessed 16 of these schemes as having a high or very high fraud risk, representing 57% (£219 billion) of the £387 billion.31 As part of this Global Fraud Risk Assessment, the Counter Fraud Function identified four COVID-19 response schemes as having the highest risk of fraud and error:

a)Universal Credit, as DWP accepts that a doubling of the caseload and relaxing controls will lead to a further increase in fraud and error levels. There is uncertainty around exactly how much fraud and error will rise but NAO believes that the increase is likely to be substantial.

b)Coronavirus Job Retention Scheme, as HMRC’s planning assumption estimated fraud and error losses of 5% to 10% of scheme expenditure.

c)The Bounce Back Loan Scheme, as BEIS and the British Business Bank have estimated between 35% and 60% of the loans may not be repaid due to fraud or credit issues.

d)Coronavirus Response Fund - Funding for the NHS: Procurement of medical equipment (including additional ventilators), as there is a high risk of fraud in procurement of personal protective equipment.

12.It also assessed a large number of other COVID-19 schemes as potentially at high risk of fraud but believed scheme owners needed to do more work to fully quantify those risks.32

Counter fraud resources across government

13.We asked HM Treasury and Cabinet Office about the capability and capacity of counter fraud resources across the public sector. Cabinet Office explained that traditionally Departments have decided individually what level of capability they need, but as fraud is a quickly evolving and complex crime this approach is not optimal.33 Cabinet Office told us that there are currently 16,000 members of the Counter Fraud Function, and 91% of these individuals work for HMRC and DWP on tax and welfare as these are government’s highest areas of known loss. The Government Counter Fraud Profession has 6,823 members from 42 organisations across central and local government and policing.34 Cabinet Office told us that 77% of these professionals work in DWP or HMRC.35

14.We asked whether counter fraud expertise is adequately deployed across the rest of government. Cabinet Office provided information on where counter fraud expertise is deployed across the rest of government but it is clear that some of the major departments have relatively few counter fraud resources.36 Cabinet Office told us that because the Counter Fraud Profession does not have a structure to assess people individually, it is “not yet able to give a view” on whether the capabilities are all in the right place. However, it is “confident” that the Counter Fraud Function is developing the structures to show where the different types of counter fraud capability are.37

15.HM Treasury told us it has ongoing dialogue with departments and supports the funding of counter fraud activity if it is cost effective. We asked HM Treasury why it took 12 months to approve the £100 million funding for a Taxpayer Protection Taskforce within HMRC. HM Treasury told us it wanted to see the assessment of the potential benefits that should accrue, and to be assured that the money is going to be directed in the right place, though was keen to assure us that throughout the 12 months there was “active discussion” around the funding required. HMRC told us it did not wait until March 2021 to deploy counter fraud and error resources on CJRS and SEISS, but the additional investment enabled it to open a further 20,000 one-to-one investigations on top of the 10,000 that it had already opened, as well as more criminal investigations.38

16.HM Treasury told us that other government functions have a critical role in helping with fraud risk. It told us the Finance and Commercial Functions must operate as an effective second line of defence in advising and supporting fraud risk. It suggested that the combination of these functions operating under the expert of guidance of the Counter Fraud Function “ultimately has the bigger impact” in tackling fraud and error.39 HM Treasury also highlighted that the Counter Fraud Function is “fairly immature” compared with some of the older functions so it is important that government keeps up the professionalism and assessment work currently under way.40