Could covid support schemes have been better targeted, MPs ask

The MPs also questioned HMRC on why the scheme for the self-employed seemed more generous than the one for the employed, and put CIOT concerns about the claims process to the officials.

The PAC session was much anticipated with the National Audit Office (NAO) having published recently a report looking at the value for money of the Coronavirus Job Retention Scheme (CJRS) – also known as furlough - and the Self-Employment Income Support Scheme (SEISS). The NAO report and this session with PAC did not consider other COVID-19 interventions designed to support businesses, such as Eat Out to Help Out, or business loans and benefits.

The PAC last looked at these schemes at the end of 2020, when it was discussing how HMRC was going to clamp down on fraud and error. HMRC have since published an interim report on how they are dealing with fraud and error and they are working on a final report for next year (2023).

Anne Marie Morris, Conservative, said that 18.4 million people who were eligible for the schemes did not benefit from them. Jim Harra replied that some businesses did not ask for help and some single-employee businesses felt the scheme was not for them. Harra said that it was too big a risk initially to help people who had recently started a business but they did receive help in later phases.

Harra justified the lack of help for directors of companies who took their incomes through dividends rather than through PAYE by saying there was no way of distinguishing dividend income that was in lieu of earnings from dividend income that was just a return on investment. But HMRC are looking at asking for data on dividends to be broken up between dividends from one-man companies and other types of dividends in the future.

Beth Russell, Second Permanent Secretary at the Treasury, told Morris that there is not an evidence base that shows that the inactivity in the jobs market, especially in relation to over-50s, is particularly related to these covid schemes – but HMRC are still reviewing this area. Russell argued that there was not a big uptick in redundancies from those on the schemes.

Harra told Morris that the increase in creditor insolvencies now is partly a consequence of the legal embargo on creditor insolvencies of companies during the pandemic.

Was the length of the schemes actually too long? asked Morris. Russell replied that the employer contribution to the furlough scheme was an example of balancing supporting jobs with not ending up in a position where the schemes either incentivise people to not work when they could have done or to not move around the economy to areas where there was demand.

Nick Smith, Labour, was told by Harra that there is still a very high level of indebtedness in the tax system compared with normal times, through a combination of the pandemic and events that have happened since then in the economy.

Russell told Smith that HMRC are still trying to understand why the payments were so much higher for each self-employed person than for each person who was furloughed. The MP questioned whether it was good value for money for the SEISS grants to increase claimants’ average profits by 14 per cent but Russell said that without it, there is a lot of evidence that people would have had a drop, in some cases a very large drop, in incomes.

Looking back at the pandemic, what would you do differently? asked Committee Chair Dame Meg Hillier, Labour. Jim Harra told her there is evidence that a financial impact test in furlough scheme might not have been well targeted in terms of preventing relief from going to people who did not need it, because those employers were saying that they did need it. On the other hand, there is also evidence from the employers we surveyed who did not use the scheme that where ‘they did not need it, they did not use it’, he said.

Beth Russell added that the decision not to introduce a financial impact test had been taken by ministers. They felt it would have added complexity into a scheme that was relatively straightforward for businesses to use and that the data that HMRC had to administer a financial impact test was always going to be out of date. She said she would look at Hillier’s suggestion that a final evaluation covers how many people on furlough were able to get a second job, especially their age profile.

Asked by Hillier how to avoid a repeat of the self-employed scheme excluding 0.2 million freelancers, Harra said there is some evidence from overseas, although it is quite difficult to get evidence, that self-certification did result in a higher level of error and fraud but the Making Tax Digital programme is intended to get more regular information about people’s turnover and expenses which could help in the future. HMRC will also look to collect data about what sectors people work in or where they are in business or employed geographically. Hillier suggested it would have been easier if HMRC had had earlier a clear test that was based on income at the time, for the self-employed grants. Russell cautioned that in any year, businesses doing ‘more or less well has to do with lots of different factors; it’s not just related to the lockdowns’.

Harra told Flick Drummond, Conservative, that there is little evidence that companies and the self-employed are now restructuring and taking income as opposed to dividends. He said he would be surprised if anyone is taking less dividend income and more salary income in anticipation of such a scheme being reintroduced. There are a number of different factors in the tax system and elsewhere that influence how people structure businesses. The new rules on how HMRC administers IR35 have undoubtedly resulted in a significant number of people no longer working through personal service companies and therefore no longer taking their income in the form of dividends, he added.

The UK’s approach to these schemes was that HMRC wanted to minimise reliance on self-certification and maximise reliance on data they held. Other countries took a different approach. One thing that would be useful for the future, said Harra, if those countries are transparent and do their evaluations, would be to understand what the outcomes were there compared to here. HMRC did take a slightly different approach from some of those other countries, based mainly on a concern about error and fraud, he said.

Anne Marie Morris called for changes in future in how furlough worked, remarking that ‘you have those jobs that, frankly, were never going to survive, and it was throwing good money after bad—they just were not viable. Then there were others that were absolutely viable and did not need the support at all’. Harra replied that HMRC put the onus on the employers to decide which jobs were viable and which were not. But it was not simply a subjective decision, he insisted.

Harra said HMRC’s best estimate of fraud and error in relation to these schemes is £4.5 billion but he will have better data to do a further revised estimate because HMRC have a round of inquiry programmes and a new set of self-assessment returns coming in in January 2023.

Janet Alexander, Compliance Operations Director, HMRC, claimed HMRC have done a ‘fairly intensive’ range of compliance interventions. She said: “First, we designed it into our system - as we set up our system, we put checks and balances in. If you didn’t have a live pay-as you-earn scheme, you did not get in to start off with, and we stopped 100,000 claims at the front door. We then stopped another 67,000 claims by looking at pre-payment checks—so the claims that showed indications of criminal activity. That amounts to about £425 million. Then, we have done what we call post-payment compliance activity. Where we have got the claims in and we have paid them out, we risk-assess, using data, all those claims and say, ‘What’s the likelihood that this claim is incorrect?’ We then prioritise those and take action, starting with the most risky and going right through to the less risky.”

Harra attempted to reassure Nick Smith, Labour, that HMRC have opened about 46,000 inquiries into these schemes so far, which is more than the 30,000 that they said they would. Most of HMRC’s work on fraud on covid schemes is civilian, he said. Harra later said HMRC have 31 criminal investigations under way and have made 56 arrests in respect of these covid schemes.

Jonathan Djanogly, Conservative, asked why the NAO report concluded that HMRC could not prevent claims around employers claiming furlough for employees who were working and thus intended to recover sums after payment. Harra said it is quite difficult to prove at the time, but there are certainly ways of doing it after a while. The HMRC chief executive said they have a behavioural insights team and he is confident that every employer who went through that process knew well what the rules were regarding claims for people who were not working.

Djanogly replied: “Can I just stop you there? The evidence was from the Chartered Institute of Taxation. It said: “The claims process was very ‘affirmative’ in nature, encouraging claimants to complete the process, and make the claim, with no obvious mechanism to exit the claim if they were ineligible. There were insufficient ‘red flags’ in the process.”

Harra responded: “I will look at that. I was not aware that they made that criticism. I do not agree. Certainly, once you were in the process, there was no inevitability that you had to go through and complete the claim; you could leave it at any point. But we were clear with employers that, before they went into that system, they had to check their eligibility and compile their data, then they went in and inputted that data. Of course, it was one of our clear objectives to make this as easy and quick as possible for people to use. I am confident that any employer using that scheme was well aware of the rules. I think there is evidence from the very first phase of CJRS that leads us to believe that the level of error and fraud was higher than in the later phases. Some employers who had partially furloughed their staff bristled against the rule that said that they had to be fully furloughed. We have an expectation that, in that phase, employers claimed for people who did some work. Later, that was legitimised, and you could adjust your claim for their hours not worked. That was probably more responsive to the situation that employers found themselves in.”

Janet Alexander told Djanogly that HMRC had acted on reports and intelligence at the beginning of furlough, but those investigations were not particularly successful, because you need an additional source of data to counter employers when they deny the allegations.

Alexander said she has seen in the CJRS scheme that some large businesses did make errors, particularly at the start, but once HMRC identified them, they were keen to correct them very quickly. Harra added that as HMRC have got better data about error and fraud, not only has that caused HMRC to adjust the overall estimate, they have identified that there are more small errors in a larger number of cases than they originally predicted. That was the reason why the number of enquiries they will carry out has been taken up to 45,000 from the original 30,000.

Asked by Meg Hillier when Making Tax Digital for income tax is coming, Harra said: “It is on schedule; we are working hard on it. But you also, no doubt, will hear from external sources that they feel that that is quite pressured – both internally on HMRC being ready, but also the external readiness of software houses and small businesses. We keep our ability to deliver it under review.”

Harra told Hillier about the benefits of HMRC wanting to get more sectoral information about businesses, such as helping to risk assess cases and identify where HMRC needs to intervene, as well as sharing non-identifying aggregate data with other government departments.

Harra told Hillier that in the period between when the NAO reported and now, there has been an increase in the number of penalties and the value of penalties that HMRC have taken. By the end of September, HMRC had charged penalties in over 1,000 cases. In the three schemes, including Eat Out to Help Out, it is about £12.3 million, which is an uplift. But Hillier replied: “We are talking in the millions, when the fraud and error estimates are in the billions.” Beth Russell countered: “Obviously we would like to see more coming in, but we obviously rely on the advice of HMRC as to the best way of doing that.”

Harra said he is not disappointed because there is more error and less fraud than HMRC originally expected and there is an evidential difficulty of establishing deliberate behaviour in schemes where people could make a whole range of errors. He explained that HMRC’s top-down estimate of error and fraud is based on what they think is the likeliest behaviour to have occurred; it is not based on a bottom-up, “What can I actually prove?”.

Turning to data sharing, Hillier said that often members of the public expect that they tell government something and it will be shared. It is very helpful sometimes to share information with the DWP when constituents have challenges. She suggested HMRC ask ‘taxpayers to tick a box saying that they are happy for their data to be shared with other government departments’. But Harra said HMRC do not do that, partly because they want to share entire databases (as with real-time information and universal credit) not just parts where taxpayers have consented to data sharing.

Janet Alexander said HMRC are winding up the taxpayer protection taskforce between April and September next year. For them to be effective quickly, they were moved from existing tax compliance teams, but then we had funding to backfill them, she explained. While it takes time for that backfilled resource to be effective, this funding means that over the medium term - up to about 2025 - there should be no impact on managing the tax gap and on tax compliance. In terms of performance, the taskforce is obviously going to collect less money than HMRC originally forecast it would. That is because there is less error and fraud than originally forecast and it is spread over a large number of smaller cases.

Alexander told the committee that, in its first year, the taskforce had recovered £226 million. In the first six months of the second year, it recovered £202 million. It is on track to achieve somewhere between £525 million and £625 million during its lifetime. By the time the taskforce is unwound, HMRC will have recovered about £1.1 billion (excluding money paid back voluntarily), of the £4.5 billion errant fraud that HMRC estimate was there, said Harra. If you look at payback in strictly financial terms, the payback per staff unit in the taskforce is lower than it is on tax work, he went on to say.

Closing the session, Harra was asked about lessons learned. He praised the ‘really effective’ close working between policymakers and operational staff. But he also said HMRC need better data without putting undue burdens on people and he would like to compare outcomes in the UK with other countries. Janet Alexander said, in the first year, HMRC did one-to-many campaigns and wrote to people asking them to check their own claims: “we need to be a lot more specific about what they were asking people to check for. We have already taken that learning on how we carry out those claims.”

The full session is here.

By Hamant Verma, CIOT Senior External Relations Officer