1.On the basis of a Report by the Comptroller and Auditor General, we took evidence from HM Revenue & Customs (HMRC) on its performance in 2017–18.
2.HMRC is the UK’s tax authority, responsible for collecting tax from individuals and businesses, and providing support to families and individuals through Personal Tax Credits (Tax Credits) and Child Benefit. The cost of running HMRC was £3.5 billion in 2017–18. HMRC raised £605.8 billion of tax revenues in 2017–18 (some 85% of total revenues raised by government), an increase of £30.9 billion (5.4%) compared to 2016–17. HMRC estimates the value of its activities to collect and protect tax revenue in 2017–18 was £30.3 billion, which exceeded its target by 8.2% (£28.0 billion) and its performance in 2016–17 (£28.9 billion). HMRC’s total forecast of the costs of tax reliefs – which reduce tax for particular groups, individuals or things - for 2017–18 is £416.8 billion, an increase of £13.1 billion (3.2%) on 2016–17. This estimate reflects the costs of 105 reliefs. HMRC assesses a further 80 reliefs have either a nil or negligible cost. It has not reported estimates of cost for the remaining 239 reliefs.
3.HMRC’s preparations for the UK’s exit from the EU include developing systems to handle customs declarations and postponed accounting for import VAT. We have written separately to HMRC to emphasise our continued concerns about the progress HMRC is making.
4.HMRC administers Tax Credits which are designed to support families with children; tackle child poverty; and help to make sure that work pays more than welfare. HMRC paid out £38.0 billion in Tax Credits and Child Benefit in 2017–18, approximately one-fifth of the government’s total expenditure on benefits. Error and fraud in Tax Credits has been a significant challenge for HMRC since they were introduced in 2003. HMRC has a target for error and fraud in Tax Credits to be no more than 5% of Tax Credit payments. HMRC’s most recent statistics on error and fraud in Tax Credits cover the period up to 2016–17. The statistics show that, in 2016–17, losses due to error and fraud amounted to £1.3 billion; 4.9% of HMRC’s total expenditure on Tax Credits. The rate of error and fraud was lower than the 7% to 8% HMRC had forecast in 2017. HMRC also underpaid claimants by £200 million in 2016–17. In our January 2018 report on HMRC’s performance in 2016–17, we concluded that HMRC lacks an incentive to reduce Tax Credits error and fraud in the transition period to Universal Credit.
5.HMRC told us that error and fraud in Tax Credits was lower in 2016–17 than it had forecast because it had overestimated both the impact of a new test on self-employed peoples’ Tax Credits, and the impact of the termination of its contract with Concentrix. The contract was terminated in November 2016 following a decline in performance. HMRC told us that it had estimated that the ‘commercial with a view to profit’ test applied to self-employed peoples’ Tax Credits from April 2016 would increase the error rate by 1% in 2016–17. However, HMRC admitted that the sample it had used was too small to provide a robust estimate of the likely increase in error. HMRC said the actual increase in the error rate from the new test in 2016–17 was 0.2%. HMRC had also forecast that termination of the Concentrix contract would add 1% to the error rate as: Concentrix did not undertake the agreed level of interventions; HMRC staff were redeployed from business as usual activities onto compliance cases previously dealt with by Concentrix and subsequently bought in-house; and, following the termination of the contract, there was a reduction in the total number of staff tackling error and fraud in Tax Credits. HMRC told us the termination of the Concentrix contract did not have an impact on the level of error and fraud in 2016–17, but that it still believes the termination of the contract will add 1% to the error rate in subsequent years.
6.The level of error and fraud in Tax Credits in 2016–17 was similar to the level in 2015–16 (4.8%). We asked why the level of error and fraud was not reducing. HMRC asserted that there were three reasons. First, the design of the benefit was systemically flawed as the cycle of annual assessments does not suit those people whose employment and other circumstances change more frequently. Second, fewer resources are being put into administering Tax Credits since the termination of the Concentrix contract. Third, the appetite within government to change existing systems and processes “is very low” as Tax Credits are being replaced by Universal Credit in the period to 2023, and “are not a permanent feature of the landscape”. HMRC confirmed that it was not proceeding with some transformation projects that would have helped people claim Tax Credits and notify HMRC of changes of circumstances which affect their entitlement to Tax Credits. But HMRC said it had learned lessons from the Concentrix contract on how to prevent and respond to error and fraud. HMRC told us that, since the contract was terminated, it had taken a number of actions, including working with customer and third sector groups to look at how to encourage customers to inform HMRC in a timelier manner about changes in their circumstances.
7.We asked HMRC whether in its future statistics on Tax Credits error and fraud it would separate out the impact of the termination of the Concentrix contract. It agreed to do so. HMRC said the termination of the contract was the primary driver of why it expected the rate of error and fraud to rise to 5.5% of Tax Credits payments in 2017– 18 and 6% in 2018–19. We also asked whether HMRC could separate error and fraud in its statistics as the actions HMRC should take to address mistakes are very different to those needed to address claimants deliberately misrepresenting their circumstances. HMRC agreed to give more information in the narrative to future statistics. In the statistics it published in June 2018, HMRC reported separate figures for the value of error and the value of fraud, but it did not include separate analysis on factors causing error and the factors causing fraud.
8.HMRC is gaining new powers and access to information which should help it tackle risks to tax compliance and collect tax revenue, including country-by-country reporting by multinational enterprises. Country-by-country reporting came into effect in 2016 and aims to help HMRC better assess international tax avoidance risks. The regulations require that UK multinational corporations give HMRC financial information, including how much tax they have paid in each of the jurisdictions where they have a presence. We asked whether country-by-country reporting had helped HMRC understand how companies were allocating their profits. HMRC said the reports provide information up-front which helps its risk assessment. It considered the reporting “makes a marginal difference” as in general it would have been able to obtain the same information by opening an enquiry, but the information will help it to target its inquires better and make them faster.
9.The Government is due to introduce, in 2021, a new register of the owners of overseas companies which own property in the UK. The register is intended to make it easier for regulators, legitimate businesses and the general public to know who the true owners of UK property are. HMRC told us that the register was important for its management of tax avoidance and evasion. It explained that there were two implications of foreign company ownership of property that the register might help it address. First, the property might have been purchased with money on which a UK taxpayer had not paid the right tax. Second, the information on the register will help HMRC to tackle avoidance or evasion of stamp duty land tax or capital gains tax by people hiding behind entities.
10.We asked HMRC whether it expects to undertake enforcement activities for the register. HMRC said it was not aware that it would be directly responsible for enforcing compliance. We also asked HMRC whether it would have a role in validating information on the register. HMRC said it was not aware it would have such a role. It told us that the project was being led by the Department for Business, Energy & Industrial Strategy. HMRC said that its fraud investigation team was part of a cross-departmental working group that was considering enforcement of the register. HMRC also told us that it had not assessed the increase in tax revenue it expects from the introduction of the register.
11.Tax reliefs are designed to benefit the economy or support wider government aims by reducing tax, for example, for particular groups, individuals or things. HM Treasury is responsible for the design of tax reliefs, which HMRC administers. We previously examined HMRC’s administration of tax reliefs in our 2016 report on HMRC’s performance in 2015–16. We concluded that, despite repeated recommendations, HMRC “still does not make tax reliefs sufficiently visible to support parliamentary scrutiny and public debate about areas where the UK chooses not to collect tax.” We recommended that HMRC include an analysis of tax reliefs and their costs in its annual report and make clear why it collects data for a minority of tax reliefs. HMRC accepted our recommendation. It now includes a summary of the cost of tax reliefs in its annual report, and publishes material showing the cost of individual reliefs, or stating the reason why it has not costed a relief.
12.HMRC’s total forecast of the costs of tax reliefs for 2017–18 is £416.8 billion. HMRC’s annual report explains this estimate reflects the costs of 105 reliefs. HMRC assesses a further 80 reliefs have either a nil or negligible cost. It has not published estimates of the cost of the remaining 239 reliefs. HMRC told us that for 179 of these 239 reliefs, information on the use of the relief is not required in tax returns and HMRC judges the cost of collecting usage data would be disproportionate.
13.We asked how HMRC knew whether tax reliefs provided value for money if it did not know how much 239 reliefs cost. HMRC told us there were two types of reliefs: those that are structural and define the tax base (such as personal allowances); and those that are considered ‘tax expenditure’ as they are designed to achieve particular policy objectives (such as entrepreneurs’ relief). HMRC asserted that for structural reliefs “the question of whether they are value for money does not necessarily arise in the normal way.” It said it determines which of the ‘tax expenditure’ reliefs to evaluate by considering their nature, costs, risks and issues, and political priority. Where HMRC does not have cost data it can “make some estimates” or “do a one-off exercise” to collect some data. However, it said the volume of reliefs means that it cannot undertake such exercises every year for every relief. As some of the 239 reliefs are likely to impose large costs, we remain unconvinced that HMRC is giving sufficient attention to ensuring all tax reliefs are providing value for money.
14.We asked HMRC whether it had looked at the number of reliefs and made recommendations to help HM Treasury simplify the system so that it is easier for people completing their tax returns. It said the Government recognises, such as by having the Office of Tax Simplification, that the tax system needs to be simplified. However, the list of reliefs continues to grow as the number of new ones added by Governments and Parliament exceeds those removed.
15.Pay As You Earn (PAYE) is the system through which employers and pension providers (known as ‘intermediaries’) deduct Income Tax and National Insurance contributions from wages and pension payments. Employers’ responsibilities for PAYE include: deducting Income Tax and National Insurance from earnings; reporting information on earnings and deductions to HMRC; and paying HMRC the deductions they have made. HMRC and the Department for Work and Pensions use the information provided by intermediaries to administer Tax Credits and Universal Credit respectively.
16.HMRC told us that the standard of PAYE administration by employers is variable and some employers do not administer PAYE to the standard that it expects. It told us that this was particularly true for those employers with large workforces where the nature of employment and the hours worked are fluid which can lead to poorer administration of PAYE than where the workforce is more stable. HMRC explained that poor administration of PAYE by employers leads to errors in payslips which then leads to many taxpayers contacting HMRC about their tax. It said that better administration of PAYE by intermediaries would be beneficial for taxpayers. HMRC also explained that poor administration flows through into the quality of PAYE data it receives from employers, and also into Tax Credits. It acknowledged that poor quality PAYE data “raises questions” for Universal Credit.
17.HMRC told us that it had drawn up a list of employers whose administration of PAYE was not good enough. The list includes organisations in the retail and food sectors, and also some NHS trusts. HMRC said it had “gone chief executive to chief executive” with a couple of employers. But it told us that it did not have sanctions that it could apply to those whose administration of PAYE was poor, and neither was there a merit award or gold standard for PAYE administration. HMRC said it was preparing proposals to improve PAYE administration by intermediaries for Ministers to consider as part of the Spending Review.
1 Report by the Comptroller and Auditor General, Session 2017–19, HC 1222, 12 July 2018
2 page 8
3 paras, 2, 6, 1.16, 1.27
5 paras 2, 3.1, 3.5
6 , Updated 13 July 2017
7 paras 3.7, 3.10; HM Revenue & Customs, Personal Tax Credits Statistics: 21 June 2018
8 Committee of Public Accounts, Session 2017–19, HC 456, 12 Jan 2018
9 Q 10; para 3.11; Committee of Public Accounts, , Session 2017–19, HC 456, 12 Jan 2018, para 28
10 3.11; page 84
11 Qq 10, 34
12 Q 11, para 3.7
13 Q 39
14 Qq 10, 34–35
15 Qq 37–38
16 HM Revenue & Customs, , 21 June 2018
17 Qq 1, 3
18 Q 3; HMRC Revenue & Customs,
19 Qq 2–3
20 Department for Business, Energy & Industrial Strategy, , July 2018, page 2
21 Qq 104, 110
22 Qq 105–111
23 , Session 2017–19, HC 1222, page 47
24 Report by the Comptroller and Auditor General, , Session 2013–14, HC 1256, 7 April 2014, para 1.2
25 Committee of Public Accounts, , Session 2016–17, HC 712, 02 Dec 2016, page 5
26 HM Treasury,, 19 Oct 2017, page 188
27 , page 47
28 para 1.16; page 47
29 Q 96
30 Qq 96–98
31 Qq 100, 103
32 Q 12; , March 2015, page6
33 Qq 12,19, 22
34 Qq 14–16, 20–22
Published: 2 November 2018