Managing tax compliance following the pandemic

HMRC does not expect to prosecute as many people for tax evasion as it did before the pandemic. HMRC prosecuted far fewer cases during the pandemic than before 2020. It has said publicly that no one will escape prosecution and that it has up to 20 years to follow up on cases of fraud. However, it also said it would not prosecute as many people as before the pandemic, despite a reduction of around 1,000 cases during the two pandemic years. HMRC says its strategy is to prosecute the most serious forms of evasion and criminal activity, and that it focuses on high- value and high-profile cases rather than large volumes of smaller cases. However, we are concerned that, without sufficient numbers of prosecutions, HMRC cannot demonstrate a credible deterrent effect. We have previously recommended that HMRC should measure the deterrent effect of its work, but it concluded that it could not do so. Recommendation 2: HMRC should develop a better understanding of the deterrent effect of its compliance work, for example by monitoring the future revenue benefit of prosecutions compared to those it decides not to prosecute. It should utilise the expertise of academics, if necessary, for example using the HMRC Datalab.

Compliance yield fell during the pandemic, and HMRC does not know what level it should be targeting with its current resources. In the five years before the pandemic, HMRC collected on average around 5.2% of tax revenues through its compliance work. This fell significantly during the pandemic, initially to 5.0% in 2020–21 and then 4.2% in 2021–22. This drop equates to a £9 billion reduction in compliance yield over the two years, compared with previous performance. HMRC maintains that no tax will go uncollected as a result of the pandemic. However, it would not be drawn on the future level of compliance yield it can generate beyond the current year’s target of £36 billion, citing uncertainty in the economy and 6 Managing tax compliance following the pandemic inflationary pressures. We are concerned that any compliance yield projections or targets it expresses in cash terms will not be sufficiently stretching during a period of high inflation. HMRC will need compliance yield to exceed 5.2% of tax revenues over the next few years to demonstrate it has caught up with the impact of the pandemic. Recommendation 3: HMRC should set a clear target of the compliance yield required to make up the shortfall during the pandemic, and specify a rolling target for compliance yield as a percentage of tax revenues.

HMRC is not doing enough to help those who want to pay their taxes correctly. Taxpayers who want to pay their tax sometimes need help to get it right, and both the pandemic and the economic situation since have put pressure on people and businesses. Tax debt levels have risen again—from £39.4 billion in April 2022 to £45.7 billion in November 2022—rather than decrease as HMRC initially expected. HMRC has made changes to its debt management practices, including tailoring its approach to each taxpayer’s circumstances and ability to pay, and informing this with new data from credit reference agencies. It has improved its productivity at recovering overdue debts, which we welcome. However, this has still not been enough to stop debt levels rising due to the number of people finding it difficult to pay. HMRC’s customer services are also struggling with high call volumes, particularly at certain times of year, meaning those who need help to get their tax affairs right cannot always get support. Taxpayers who want to put their affairs right can find it hard to do so due to inflexible repayment practices and confusing correspondence from HMRC. Recommendation 4: HMRC should ensure it is providing sufficient support to taxpayers, big and small, who want to pay their tax. It should look at whether the additional staff it has secured for debt recovery work are sufficient, given it is struggling to keep up with demand even with better productivity.

We are concerned that HMRC may be overstating the impact of its compliance work, and that it is overcharging some taxpayers. HMRC tests 400 of its completed compliance cases each year. In 2021–22, 80 of these cases had errors in the compliance yield recorded, which were skewed towards overstating yield. HMRC’s testing also found seven cases where it had overcharged the taxpayers by a total of £32 million. HMRC says that it has controls in place to prevent over-recording of yield but acknowledges that it does not know to what extent the errors identified are representative of compliance yield as a whole being overstated. HMRC also cannot estimate how many taxpayers it is overcharging following compliance enquiries, or by how much. It says that there has been a slight improvement in the quality of its casework in the past year but acknowledges that the number of overcharges identified in its sample testing is too high. HMRC has agreed to update its testing approach so that it can better estimate the extent of errors it makes. Recommendation 5: a) HMRC should develop statistically robust estimates of the level of error in its compliance yield measure, and how far taxpayers are overcharged. Managing tax compliance following the pandemic 7 b) HMRC should demonstrate it has taken all proportionate steps to identify and correct overcharges. It should make clear what compensation is available if taxpayers are overcharged.

There are signs that the tax gap may grow, and that HMRC does not have the operational resilience needed to deal with this. HMRC is funded to stop the tax gap from growing. The tax gap is an important measure of how much revenue may be missed—due to evasion, avoidance or non-payment—that could otherwise fund vital public services. Due to the way it is estimated, the tax gap does not yet reflect the full impact of the pandemic and will not do so for some time. HMRC’s latest estimate is that the tax gap was stable in 2020–21, but HMRC says that this has a much larger range of uncertainty than normal and may need to be revised as it gets more data. There is a significant risk that the tax gap will grow, in light of compliance yield dropping and levels of debt and non-payment rising. HMRC’s own planning estimates indicate that it is unlikely to generate enough compliance yield to stop the tax gap from growing in the next few years. However, it does not have a contingency plan if this happens. Since it takes four years to get compliance staff fully trained and up to speed, there could be a long lag if HMRC waits to determine whether it needs more resources. HMRC has up to 20 years to pursue cases of suspected non- compliance, but this only applies to fraud cases. For errors the period is less, and in some cases HMRC cannot go back more than four years. Recommendation 6: a) HMRC needs to build in more resilience to the tax system, with the tax gap at risk of growing and high returns available from compliance work there is a strong value for money case for increasing resources. b) At a minimum, HMRC should specify a contingency plan for bringing in additional compliance capacity to ensure increased levels of non- compliance can be tackled quickly, and before the window closes for investigating cases it did not pursue during the pandemic. 8 Managing tax compliance following the pandemic 1 Impact of the pandemic on tax compliance

On the basis of a report by the Comptroller and Auditor General, we took evidence from HM Revenue & Customs (HMRC) regarding the challenges to tackle non-compliance during and following on from the pandemic.1

During the two pandemic years of 2020–21 and 2021–22, ‘compliance yield’ (the additional revenues protected as a result of HMRC’s interventions) per staff member fell from £1.3 million a year to £1.1 million (in 2021 prices).10 HMRC told us this was due to several factors, including the fact that staff redeployed away from tax compliance and onto the COVID schemes were typically more experienced.11 HMRC’s operations were also affected by external factors such as court closures and requirements on homeworking, social distancing or personal protective equipment.12

HMRC did not initially recruit at scale but has more recently recruited 4,800 new compliance staff, leading to 2,500 more FTE than in 2021–22. HMRC expects staff productivity to increase over the next few years, but told us that new staff take up to four years to be fully productive. The return per staff member is therefore likely to stay lower than before the pandemic due to lack of experience among newer staff.13 Prosecutions

Over the two pandemic years, HMRC completed around 1,000 fewer prosecutions for tax-related offences than before the pandemic. In 2020–21 and 2021–22, it concluded just 163 and 236 prosecutions respectively, compared with around 700 a year in the two years before the pandemic.14 HMRC told us that court closures and reduced court capacity during the pandemic affected the extent to which it could progress prosecutions through the courts, which therefore constrained the numbers it could complete.15

However, HMRC also told us that it was already reducing numbers of prosecutions before the pandemic, and that is not planning on restoring the number of prosecutions to pre-pandemic levels, even with the backlog. Before the pandemic, the number of prosecutions had reduced from around 900 in 2017–18 to around 700 in each of 2018–19 and 2019–20. HMRC told us this was because it considered that its criminal prosecution powers were best used to tackle the most serious and complex cases, rather than large volumes of smaller cases.16

We asked whether such significant reductions in prosecutions risks weakening the deterrent effect of HMRC’s work, which could lead to greater levels of non-compliance. HMRC told us that while it wants to see serious tax crimes addressed in the most effective way, it is not seriously concerned that the reduction might weaken the deterrent effect. It argued that the publicity around high-profile convictions, as well as its use of other enforcement tools such as civil powers, is sufficient to maintain a credible deterrent.17 However, it does not have a way to monitor this. This Committee has previously recommended that HMRC should measure the deterrent effect of its work, but it concluded 9 Q 46; C&AG’s Report, para 9 10 C&AG’s Report, para 19 11 Q 46 12 Q 80; C&AG’s Report, para 1.10 13 Qq 44, 80; C&AG’s Report, para 19 14 Q 47; C&AG’s Report, para 16 15 Qq 48–52 16 Qq 53–54; C&AG’s Report, para 16 17 Qq 47, 55 10 Managing tax compliance following the pandemic that it could not do so.18 There may be scope for HMRC to make more use of the expertise of academics, for example using its Datalab which provides trusted researchers with secure access to anonymised data.19 Compliance yield

HMRC defines compliance yield as the additional revenue collected and protected that would have otherwise been lost to the Exchequer if not for HMRC’s interventions. It is the most direct measure of the impact of HMRC’s compliance work, and it covers both the broad effect of its measures to prevent non-compliance and the direct impact of its enquiries and investigations into non-compliance that has already happened.20

In the five years before the pandemic, HMRC collected or protected an average of 5.2% of tax revenues through its compliance work. In 2020–21, total tax revenues fell as a result of the pandemic, but compliance yield fell slightly further, dropping to 5.0% of total revenues. In 2021–22, total revenues rose by 20% to record levels following the reopening of the economy and other factors such as higher inflation. However, compliance yield did not recover at the same pace, dropping to 4.2% of total revenues. This drop equates to a £9 billion reduction in compliance yield over the two years, compared with previous performance.21 HMRC told us that this partly reflected the economic impacts of the pandemic, and that there was less yield to be had. However, it also acknowledged that the reduction was affected by reductions in its own work, and a reduced ability of taxpayers to engage with its enquiries.22

We asked HMRC whether, to catch up on compliance yield lost during the pandemic, it should expect to generate higher levels of yield than before the pandemic over the next few years. HMRC acknowledged that this should happen over time and suggested that no tax should go uncollected as a result of the pandemic because it can go back up to 20 years with its compliance enquiries. However, it told us that other risks of non-compliance may take priority. There are also timing issues between when taxes are owed and when HMRC records compliance yield, meaning the full impact of the pandemic will not be known for several years.23 HMRC would therefore not be drawn on what levels of compliance yield it should target in the coming years, either as a percentage of total revenues or in cash terms, beyond aiming for £36 billion in the current year. It told us this was due to uncertainty in the economy and inflationary pressures.24 18 Treasury Minutes, Government responses on the Thirty Fourth to the Thirty Sixth; the Thirty Eighth; and the Fortieth to the Forty Second reports from the Committee of Public Accounts: Session 2015–16, Cm 9323, July 2016 19 https://www.gov.uk/government/publications/hmrc-datalab 20 C&AG’s Report, para 2.2–2.3 21 Qq 35, 85; C&AG’s Report, para 11–12 22 Qq 35, 45–46, 85 23 Qq 85–86, 88; C&AG’s Report, para 2.8–2.9 24 Qq 89–95 Managing tax compliance following the pandemic 11 2 Understanding how to catch up following the pandemic Supporting taxpayers who want to pay their taxes correctly

HMRC told us that the pandemic created difficulties for taxpayers, and that the economic situation since has continued to put pressure on them.25 HMRC’s strategy in recent years has increasingly focused on preventing non-compliance occurring in the first place, including by helping taxpayers get their tax affairs right.26

Following our previous recommendations on HMRC’s debt management practices during and after the pandemic, HMRC has made changes to its approach aimed at supporting people and businesses to pay what they owe and recovering debts more quickly. This includes tailoring its approach to each taxpayer’s circumstances and ability to pay, and informing this with new data from credit reference agencies. HMRC has also added 500 staff to its debt management service. It told us that these efforts have improved its productivity at recovering overdue debts and that it has made strong progress to settle the tax debts amassed during the pandemic, which we welcome.27

However, HMRC acknowledged that, despite these improvements, its initial expectation that debt levels would reduce has not been borne out in practice. This is because of ongoing economic pressures affecting taxpayers, particularly small businesses. Instead, debt levels have risen again – from £39.4 billion in April 2022 to £45.7 billion in November 2022.28

HMRC can also support taxpayers through its customer service functions, by answering questions or otherwise providing helpful information in a timely way. However, its customer services have also been struggling with high call volumes, particularly at certain times of year such as the January deadline for filing self-assessment returns.29 We have previously reported on a deterioration in service levels in the past few years.30 If people cannot contact HMRC when they need to, then some of those who need help to get their tax affairs right will not always be able to get support. We have also heard examples of people who want to put their affairs right but find it hard to do so because of HMRC’s approach. Examples have included inflexible repayment practices and confusing correspondence.31 25 Qq 35–36, 40, 84 26 C&AG’s Report, para 5 27 Qq 37–41, 44; Letter from Jim Harra, 24 January 2023 28 Q 41; Letter from Jim Harra, 24 January 2023 29 Q 103 30 Committee of Public Accounts, HMRC performance in 2021–22, Thirty-Third Report of Session 2022–23, HC 686, 11 January 2023 31 Qq 79, 102–103 12 Managing tax compliance following the pandemic Understanding the impact of HMRC’s compliance work

HMRC’s compliance yield estimate is an important measure of the impact of its compliance work. A good understanding of this performance is essential for identifying whether HMRC has the overall resources it needs, as well as whether it is deploying these resources to where they are most effective.32

HMRC’s quality assurance arrangements involve testing 400 of its completed compliance cases each year. In 2021–22, 80 of these cases had errors in the compliance yield that HMRC had recorded. These errors included both overstated and understated figures, and occurred for a range of reasons including data input mistakes, incorrect calculations, or incorrect decisions on what should or should not be included. However, they were skewed towards overstating compliance yield. Out of £736 million of yield that was tested, the net effect of the errors was to overstate compliance yield by £59.7 million (8%).33

HMRC’s 2021–22 testing also found seven cases where taxpayers had been overcharged following the completion of a compliance enquiry, by a total of £32 million. Almost of all of this was a single large overcharge, which HMRC identified before any payment was taken.34 However, most of the smaller overcharges were for individuals or small businesses, who might therefore have struggled to pay had HMRC not identified the error and corrected the position.35 HMRC told us that it has controls in place to prevent over-recording of yield, and that its testing has found a slight improvement in the quality of its casework in the past year. However, it acknowledged that the seven overcharges it identified was too many.36

HMRC does not know to what extent the errors identified are representative of compliance yield as a whole being overstated. HMRC also cannot estimate how many taxpayers it is overcharging following compliance enquiries, or by how much. HMRC explained that it designed its sampling process to examine quality in all areas of its casework, rather than to provide assurance on the overall estimate. Its approach means it cannot accurately extrapolate the errors or overcharges it finds to produce an overall estimate of their impact. HMRC told us it expects compliance yield to be materially correct but it cannot provide assurance that this is the case. Errors should be rare, and finding a lot of them in a sample may indicate a bigger problem that merits further testing. HMRC told us that it is reviewing how it can update its testing approach so that it can better estimate the extent of errors it makes, based on suggestions from the National Audit Office.37 Stopping the tax gap from growing

HMRC is funded to stop the tax gap from growing. The tax gap is an important measure of how much revenue may be missed—due to evasion, avoidance or non-payment —that could otherwise fund vital public services.38 32 C&AG’s Report, para 2.2, 2.14 33 Qq 57–58, 64–65; C&AG’s Report, para 2.12 34 Qq 56, 59, 67–70, Letter from Jim Harra, 9 February 2023 35 Letter from Jim Harra, 9 February 2023; C&AG’s Report, para 2.12 36 Qq 59–60 37 Qq 61–66; C&AG’s Report, para 2.13 38 Qq 1, 44, 90, 106; C&AG’s Report, para 2.18 Managing tax compliance following the pandemic 13

Due to the way it is estimated, the tax gap does not yet reflect the full impact of the pandemic and will not do so for some time. HMRC’s latest estimate is that the tax gap remained stable in 2020–21, at 5.1% of taxes theoretically owed. However, HMRC’s tax gap estimates typically need to be revised over several years as it gets more data. HMRC told us that its 2020–21 estimate has a much larger range of uncertainty than normal, because of the impact the pandemic had on some of the key data sources it uses.39

Some measures that affect the tax gap have weakened since the start of the pandemic, indicating that the tax gap may grow in the coming years. Non-payment of taxes owed is one such component of the tax gap that HMRC expects will grow. Levels of debt and non-payment rose considerably during the pandemic, and debt levels are not reducing as quickly as HMRC expected them to.40 Compliance yield, which reduces the tax gap, also saw a significant drop during the pandemic.41 HMRC’s Customer Compliance Group (CCG) produces planning estimates of the level of compliance yield it thinks it will need to achieve to stop the tax gap from growing, and the range of yield it thinks it could achieve. Its own planning estimates indicate that it is unlikely to generate enough compliance yield to stop the tax gap from growing in the next few years. The gap between the required level and HMRC’s mid-point planning estimate equates to around £17 billion in total over the three-year period. HMRC would need to achieve its upper estimate of compliance yield to stop the tax gap from growing, but in practice expects to achieve somewhere between its upper and mid-point estimates.42

We asked whether HMRC needed more resources to catch up on the impact of the pandemic and to keep the tax gap from growing, and whether it was looking to recruit more. HMRC acknowledged that it will be more challenging to maintain the tax gap than it has been in the recent past but did not indicate what contingency plans it had if its resources were insufficient.43 HMRC told us that it will be gradually reintegrating staff from COVID-19 work to tax compliance between April and September 2023.44 It also told us that it has recruited 4,800 staff to its compliance group in the past two years. However, it takes up to four years for new staff to be fully trained and to gain enough experience to be fully effective, due to the complex nature of tax legislation and investigations which require skills ranging from technical tax knowledge to project management.45 HMRC also argued that it has up to 20 years to pursue cases of suspected non-compliance during the pandemic. However, this only applies to cases of suspected fraud. For errors the period is less, and in some cases HMRC cannot go back more than four years.46 39 Qq 82–83, 93–95 40 Qq 35–36, 41, 81, 84 41 Qq 35, 85; C&AG’s Report, para 12 42 C&AG’s Report, para 18 43 Qq 44, 85–87 44 Q 43 45 Qq 44, 80; MFP0001 46 Qq 25, 85–87 14 Managing tax compliance following the pandemic