Thirty-Third Report - HMRC performance in 2021–22

Resourcing HMRC’s compliance work to maintain rather than reduce the tax gap means the government is missing out on billions in lost revenue. HMRC estimates that the tax gap—the difference between the amount of tax that should, in theory be paid to HMRC, and what was actually paid—was £32 billion in 2020–21, or 5.1% of all tax liabilities, the same proportion as in 2019–20. This masks changes in the tax gap for each category of tax, with the tax gap for VAT decreasing while the tax gap for Corporation Tax, excise duties and income tax Self-Assessment increasing in 2020–21. HMRC bases its compliance performance and resourcing on maintaining the tax gap and stopping it from growing. However, there remains scope for reducing it; for every £1 that HMRC spends on compliance activities, it recovers £18 in additional tax revenue. The pandemic has created more uncertainty in the data that HMRC uses to estimate the tax gap, but HMRC does not currently report the range of uncertainty in its headline estimate. HMRC told us that this would be difficult to do, but possible. Recommendations: • HMRC should set out what level of investment in its compliance teams would be needed to reduce the size of the tax gap, and confirm what, if any, intention it has to pursue this. • HMRC should also calculate and report an uncertainty range for its headline tax gap estimate to provide more transparency to users of the estimate.

HMRC’s plan to only recover a quarter of losses due to fraud and error on its COVID-support schemes does not go far enough. HMRC estimates that total error and fraud across the lifetime of the COVID-support schemes is £4.5 billion, representing 4.6% of the £96.9 billion total support provided. This is lower than HMRC’s previous estimate, though the actual level of fraud and error remains very uncertain. HMRC has drawn together a wider set of data to improve its estimate, but 6 HMRC performance in 2021–22 this is limited by the shortage of data that HMRC collected on taxpayers’ working patterns at the time the schemes were running. HMRC has been given £100 million to fund a temporary taskforce to investigate fraud and error on the schemes and has opened about 40,000 investigations so far. However, of the £4.5 billion in fraud and error losses, HMRC forecasts by the time the taskforce closes it will have recovered only around £1.1 billion, with the rate of return for the funding expected to be less than if invested in tax compliance. HMRC is yet to demonstrate it has done all it reasonably can to recover the losses and avoid the dent to public finances. HMRC risks rewarding those taxpayers that were dishonest if it does not pursue more of the losses than currently planned. Recommendation: In determining what further recovery action to take on fraud and error on the COVID-19 support schemes, HMRC should: • keep under review the return on investment of spending more resources on recovery; and • set out how it will ensure it maintains a level playing field for individuals and businesses that did not abuse the schemes, rather than being seen to reward those that were dishonest.

We are concerned that HMRC may be lagging behind other established tax authorities in preventing fraudulent VAT registrations. HMRC is constantly changing its processes to prevent criminals from exploiting the tax system. VAT is particularly susceptible to fraud and criminality as it can involve HMRC repaying large amounts to taxpayers. We raised concerns about a case where a criminal used a legitimate company’s details to apply for a VAT registration number and make fraudulent VAT repayment claims. We understand that the criminal was successful in the UK, but that safeguards adopted by the German tax authorities appeared more effective at identifying similar fraudulent activities at an earlier stage. Recommendation: HMRC should engage with its international counterparts to understand what lessons it can learn in preventing fraudulent VAT registrations and minimising the impact on honest taxpayers.

Taxpayers and their agents are still not receiving an acceptable level of customer service. In the last five years, HMRC has reduced its customer service staff numbers from 25,500 to 19,500. During the pandemic, HMRC’s performance in replying to post or handling calls fell significantly, partly because it did not have sufficient customer service staff to manage the pressures that the pandemic brought. We were surprised to learn that at times in the past, HMRC has simply closed its telephone line when it could not cope with demand. It is not acceptable not to answer calls from people who are trying to pay the government money. HMRC’s plan for improving customer service is to continue digitalising the tax system, moving people away from phone and post onto online systems. Taxpayers report being more satisfied with HMRC’s digital services than its phone and post services. However, we are not convinced that its plans will sustainably reduce demand for traditional channels or deal with the unacceptable level of service that taxpayers and agents are currently suffering. The move to online services will not happen quickly and will not be appropriate for all circumstances or customers. HMRC performance in 2021–22 7 Recommendation: HMRC should write to the Committee setting out its plan to improve customer service to adequate levels as quickly as possible, and within three months, including: • the metrics HMRC will use to monitor its customer service performance, including metrics it needs to demonstrate it can answer calls and deal with post in a timely manner; • the level of customer service taxpayers and their agents can expect to receive over the next three years against each of these performance metrics; • how it will support customers who are unable to engage digitally or have a preference for post or telephone contact; and • its contingency arrangements if its plans to reduce demand for traditional channels are unsuccessful or take longer to implement.

HMRC has further to go until it can differentiate between taxpayers who are genuinely struggling, and those who can afford to meet their liabilities but are choosing not to. Total tax debt in August 2022 was £46 billion, less than at the height of the pandemic in March 2021, but significantly higher than before the pandemic. HMRC’s data suggest that the tax debt has started to increase again as the economy slows down and taxpayers feel the effects of the cost-of-living crisis. HMRC has previously taken a standardised approach to debtors, but is now trying to vary its approach depending on whether a debtor is in genuine financial distress. This segmentation requires good data on the behaviour of its debtors, as well as sufficient capacity to take a tailored approach. HMRC has increased its debt management service by 700 people and is experimenting with data from credit reference agencies to gain insights into its debtors. However, HMRC’s ability to understand debtors’ circumstances will be limited until it has completed its single customer account project, which will join up taxpayers’ records that are currently held in its different digital systems. Recommendation: HMRC should set out how it will strike the right balance between providing support to taxpayers who need it, whilst ensuring that those able to meet their liabilities are doing so. HMRC should also set out when its single customer account will be ready and consider how it can bring the implementation of it forward.

Research and development tax reliefs are costly, prone to abuse and provide questionable benefit to the UK economy. The government has a target of 2.4% of GDP to be invested in research and development. In 2021–22, HMRC spent £9.5 billion on research and development reliefs, with expenditure having continued to grow year-on-year. HMRC told us that analysis it has undertaken demonstrates that these reliefs increase research and development investment, albeit marginally for small and medium-sized enterprises, but the benefit to the UK economy of that investment is not clear. Based on HMRC’s latest estimates, abuse of the reliefs cost the taxpayer £469 million in 2021–22, though work to establish the true scale of the problem remains ongoing. Abuse of the small and medium-sized scheme is particularly worrying, with 7.3% of claims for relief estimated to be fraudulent or erroneous. HMRC has undertaken work to reconcile the £21.6 billion gap 8 HMRC performance in 2021–22 between HM Treasury and Office for National Statistics estimates of research and development expenditure, but discrepancies remain. HMRC is introducing new measures from April 2023 to strengthen its compliance approach and bear down on abuse of the schemes. Recommendation: HMRC should develop its analysis of the additional research and development expenditure its relief schemes result in, to consider what impact that expenditure has on the UK economy. HMRC should report to the Committee on its findings within 12 months. HMRC performance in 2021–22 9 1 Managing error and fraud, compliance and tax avoidance

On the basis of a report by the Comptroller and Auditor General, we took evidence from HM Revenue & Customs (HMRC) on its performance in 2021–22.1 HMRC published its annual report and accounts for 2021–22 on 18 July 2022, shortly before Parliament’s summer recess. It had published its annual report and accounts for 2020–21 in November

HMRC’s 2020–21 estimates of the tax gap are more uncertain than usual due to the impact of COVID-19 on the data it uses to inform its estimates. For example, HMRC said it has had to make some assumptions about the underlying level of write-offs and remissions that might happen. HMRC does not publish an overall range around the tax gap at the moment, and said it was quite difficult to do given the different methodologies used to calculate the tax gap for each individual component. In its reporting, HMRC comments on the confidence it has in each of these components, and sets out the sensitivity associated with any assumptions used. We asked whether HMRC was confident that the tax gap had actually stayed the same in 2020–21. HMRC said that it was confident that the 2020–21 estimate is the best possible estimate at the moment given the data available.12 6 C&AG’s report, para 1.25 7 Qq 24–29 8 Qq 13–14 9 C&AG’s report, para 5 10 Q 15 11 Q 46 12 Qq 15, 19–23; C&AG’s report, para 5 HMRC performance in 2021–22 11

With the resources it was given in the 2021 Spending Review, HMRC’s aim is to maintain the tax gap and prevent it from growing. For every £1 that it spends on compliance activity, it said it recovers £18 in additional tax revenues. HMRC sometimes gets additional resources from HM Treasury for particular purposes in return for additional tax revenues. HMRC said that typically this investment recovers between £10 and £20 in tax revenue for every £1 it spends.13 Fraud and error in the COVID-19 support schemes

HMRC estimates that total fraud and error across the lifetime of the COVID-support schemes was £4.5 billion, representing 4.6% of the total support provided. This is lower than the estimate HMRC included in its 2020–21 accounts. HMRC has drawn on new data to improve its estimate, from a random enquiry programme, more compliance activity and Self Assessment returns for 2020–21. The estimates are, however, still very uncertain. HMRC said that the most uncertain element was around claims for people working while their employer claimed support for them. Given the speed at which HMRC had to work to get support to individuals and businesses, it said some level of fraud and error was inevitable. HMRC accepted that it needed more timely data on taxpayers’ working patterns at the time the schemes were running, rather than activities after the event that relied on people recalling their working patterns from some time ago. It told us it would look into drawing together its lessons from the schemes to support other parts of government that are tasked with disbursing funds at pace, such as the Household Support Fund.14

HMRC has been given £100 million to fund a temporary taskforce to investigate fraud and error on the schemes and has opened about 40,000 investigations so far. HMRC forecasts that by the time the taskforce winds down, it will have recovered around £1.1 billion, around a quarter of the fraud and error losses. HMRC expects some of the more complex cases to still be unresolved by that point. HMRC expects to also recover further losses through its usual tax compliance activity, though it did not articulate how much more it will realistically be able to recover.15

We raised the point that there was a moral duty and fairness issue to pursue fraud with, for example, some companies getting an unfair competitive advantage through their abuse of the schemes. HMRC said that it looks into every single allegation of fraud, drawing together other supporting data such as business performance. HMRC said that its controls had been effective at preventing organised crime from abusing the support schemes. It has found that a higher proportion of the losses than it first thought is due to error, rather than fraud, by a large number of fairly small concerns. HMRC said this affects the cost-effectiveness of recovering the losses, because it has to investigate a larger number of people. It said the rate of return it can expect from this activity is less than if it spent the same investment on tax compliance.16 VAT registration

We asked HMRC about a case where a criminal used a legitimate company’s details, apparently successfully, to apply for a VAT registration number and make fraudulent VAT 13 Qq 14, 33, 43–45 14 Qq 80–83, 92–95 15 Q 84 16 Qq 89–91, 113 12 HMRC performance in 2021–22 repayment claims. This fraud was only picked up when the legitimate company received communication from a debt collection agency requesting payment. We raised concerns that safeguards adopted in other countries appear more effective at identifying this type of fraud at an earlier stage, with the German tax authorities apparently refusing the same criminal a VAT registration and reporting the incident to the police.17

HMRC told us that VAT is a particular target for organised crime as it can involve HMRC repaying large amounts to taxpayers. HMRC described it as an ‘arms race’, with HMRC constantly changing its processes and risk assessment to prevent criminals from abusing the tax registration process, and criminals constantly finding new ways around them. HMRC felt it was effective at identifying these attempts at fraud early and adjusting to them. It said it needs to balance making the system secure with making it as quick and easy as possible for legitimate companies to register. HMRC said it had put in extra steps to verify the identification of individuals and companies, which has meant it has been able to reject a large number of VAT registrations. However, the extra steps have also meant that there had been delays in legitimate applications for VAT registration in the last year.18 17 Q 9 18 Qq 9–10 HMRC performance in 2021–22 13 2 Supporting taxpayers Customer service

HMRC’s post and call handling performance fell significantly during the pandemic. For example, in 2021–22 HMRC responded to 39.5% of post within 15 days, compared to 70.3% in 2019–20. The average speed of answering calls to HMRC helplines was 12:22 minutes in 2021–22, compared to 6:39 minutes in 2019–20.19 The Chartered Institute of Taxation submitted evidence to us that highlighted concerns about the difficulties both advisers and taxpayers face getting timely responses and action from HMRC.20 HMRC said that its ability to manage shocks such as the impact of COVID-19 was impacted by reductions in staff: in the last five years its number of customer service staff has reduced from 25,500 to 19,500 as HMRC has moved more customers to digital systems.21 We raised concerns that this reduction in staff was premature and were effectively cuts rather than efficiency savings. HMRC said that it had achieved efficiencies, but that the additional pressure of responding to the pandemic meant that performance had suffered, for instance through a backlog of 3.3 million items of post building up. HMRC sees further digital improvements as key to moving customers away from phone and post and making further efficiencies.22 Beyond what it has already planned, it said it would be difficult to deliver any new efficiencies before 2024–25.23

HMRC said that it limits the number of callers that can join the call queue, which can have an impact on its call waiting time measure. Anyone not able to join the call queue is played a “busy message” and is forced to end the call. HMRC told us that in previous years it would play this message about 11% or 12% of the time. It has now reduced this to 5.5%, as its customers have told HMRC that they would prefer to wait in a queue for longer than keep calling back. HMRC said that these changes mean that it is not fair to compare average call waiting times year on year. Instead it is focusing on the proportion of customers who need to speak to an agent that get through to an agent the first time they call. HMRC has set itself a service standard of 85%, although is currently achieving 77%.24

HMRC’s expects to improve customer service by continuing digitalise the tax system, for instance by developing its digital assistant and single customer account. In this way HMRC hopes to reduce the demand for phone and post by servicing more people through online systems. HMRC’s customers report being more satisfied with HMRC’s digital services than its phone and post services. We raised concerns that taxpayers were being forced to interact with HMRC online rather than by phone or post. HMRC said it recognises some customers will always need to speak to someone, either because they are unable to deal with HMRC digitally or because of the nature of their enquiry. But HMRC believes there is a lot of “low-value demand” from taxpayers that can be removed through better customer self-service or better HMRC processes.25 19 C&AG’s report, para 8 20 Written evidence ARA0002 – Chartered Institute of Taxation, para 1.10 21 Q 114 22 Qq 2, 120–121 23 Qq 4, 79 24 Qq 114–118 25 Qq 60, 120–122; C&AG’s report, para 8 14 HMRC performance in 2021–22 Segmenting taxpayers in debt

The total tax debt owed to HMRC in August 2022 was £46 billion. This is less than at the height of the pandemic in March 2021 but significantly higher than before the pandemic. HMRC told us that while the size of the debt fluctuates throughout the year, the trend continues to be upwards. It said that the cost of living crisis meant that more people are unable to pay their debts or feel they have to prioritise other spending. It raised concern that some taxpayers had started using HMRC as an overdraft facility.26 In October 2022, for taxpayers on agreed Time to Pay arrangements HMRC was charging interest for late payment at 2.5 percentage points above the Bank of England base rate, far lower than what taxpayers would get charged on a credit card.27

HMRC said it is dealing with a larger number of smaller debts than it is has dealt with before. It told us that it had previously took a one-size-fits-all approach to managing debts, but is now trying to vary its approach depending on whether a debtor is in genuine financial distress or is prioritising other spending over paying HMRC. HMRC tries to put those in genuine distress into affordable arrangements to repay the debt, while taking a firmer approach with those that need it. It has increased its debt management service by 700 people to support this, and as of March 2022 had 851,000 taxpayers on Time to Pay arrangements. HMRC now allows taxpayers to create their own Time to Pay arrangement online if they meet certain criteria. The average length of a Time to Pay arrangement has increased from six months to 14 months.28

HMRC said it is trying to improve the data it uses to understand the behaviour of its debtors. It is experimenting with data from credit reference agencies.29 It is also developing a comprehensive single customer account, using funding it received in the 2021 Spending Review.30 HMRC’s digital services are currently separate and not integrated. The single customer account aims to bring together the data that HMRC has on a customer from across the separate services into a single account.31 Research and development tax reliefs

The government has a target of 2.4% of GDP to be spent on research and development.32 To encourage this investment, HMRC administers tax reliefs to companies via two schemes – one for large companies and one for small and medium-sized enterprises. HMRC said that the scale of the schemes is growing all the time. HMRC said that it has published research that demonstrates that the two schemes increase research and development investment, albeit marginally for small and medium-sized enterprises. While this research demonstrates what additional research and development activity was stimulated, it does not quantify the benefit to the UK economy.33 26 Qq 47–49; C&AG’s report, para 7 27 Qq 58, 66 28 Qq 30, 47–48, 51, 55 29 Q 51 30 Q 60 31 Q 64 32 Q 96 33 Qq 103–106 HMRC performance in 2021–22 15

In 2021–22, HMRC granted research and development tax reliefs worth £9.5 billion.34 HMRC said the relief is an attractive target for abuse, whether by companies that have not carried out any research and development or by advisors encouraging companies to push the boundaries of what expenditure is eligible to be claimed for under the schemes.35 HMRC estimates that the level of fraud and error on research and development relief was £469 million, or 4.9%. This is an increase from 3.6% in 2020–21, although HMRC believes this is due to improvements in its methodology rather than non-compliance worsening.36 HM Treasury has found UK companies claimed tax relief on £47.5 billion of research and development expenditure in 2019, but the Office for National Statistics estimated that businesses only carried out £25.9 billion of privately financed research and development in the UK.37 HMRC has recently published a joint paper with the Office for National Statistics that explains some of this disparity, but HMRC said that it is still working to resolve the remaining gap.38 Some of the gap can be explained by tax relief funding research and development costs incurred outside of the UK. We asked HMRC whether this was a legitimate use of the funding. HMRC said that the objective of the relief schemes was to fund research and development that benefits the UK economy, and that it has made some reforms to the small and medium enterprise scheme that better targets research and development undertaken in the UK.39

Fraud and error is particularly high on the scheme for small and medium-sized enterprises, estimated at 7.3% in 2021–22. HMRC said it gets about 76,000 claims on this scheme every year, providing a lot of activity for HMRC to police.40 It said it is introducing new measures from April 2023 to strengthen its compliance activity on the schemes, including requiring more information from businesses when submitting their claim, such as the name of their advisor and a named person within the business accountable for the information in the claim.41 34 C&AG’s report, para 3.26 35 Q 96 36 Q 99; C&AG’s report, para 12 37 C&AG’s report, para 3.35 38 Qq 97–98; Office for National Statistics, Comparison of ONS business enterprise research and development statistics with HMRC research and development tax credit statistics, 29 September 2022 39 Q 98 40 Q 102 41 Q 96 16 HMRC performance in 2021–22