Welcome, everyone, to this afternoon’s session of the Public Accounts Committee. The Chair, Sir Geoffrey Clifton-Brown, sends his apologies; I am chairing this session instead of him. We will be talking this afternoon about large business tax compliance. Large businesses account for tax collected by HMRC totalling £337 billion. Given the scale of that, HMRC has a particular approach to large businesses called intensive co-operative compliance, for about 2,000 large businesses. When examining this, the National Audit Office found a strong return from HMRC in terms of collecting the tax, but also identified certain important questions about how HMRC determines risk, uses data, resolves cases efficiently and applies lessons more widely. HMRC has a wide range of powers, including a special measures regime, though it has not used those powers on large businesses since it gained them in 2016. That is something, in particular, we want to explore in our session this afternoon, as well as whether the approach taken with large businesses, which seems to be successful, could have wider implementation across businesses as a whole. Our witnesses this afternoon are from HMRC. Perhaps, John-Paul Marks, as the permanent secretary and chief executive, you could begin by introducing yourself and your colleagues to the Committee.

I would be delighted, Chair. Good afternoon to your Committee. I am John-Paul Marks and I am the first permanent secretary and chief executive of HMRC. I will let my colleagues introduce themselves.

I am Penny Ciniewicz, the director general for customer compliance.

I am Nicole Newbury, the director of large business at HMRC.

My name is Jonathan Athow. I am director general for strategy and policy.

Thank you all very much for coming this afternoon. Sarah Olney has a couple of opening questions.

Thank you, Chair. Before we begin the main session, we want to ask a couple of questions about the loan charge. Mr Marks, we know that there are about 32,000 outstanding loan charge cases and that HMRC is spending about £31 million a year pursuing those outstanding cases. Can you tell us when you hope to resolve them?

Yes. Thank you for the question. We currently estimate that there are 37,000 customers with outstanding loan charge liabilities. As you know, the Government published a new settlement opportunity in the Budget, and we believe that most individuals could see reductions of at least 50% in what they have to pay. We think that a third of individuals may have nothing now to pay, and the vast majority of people will have seen a reduction. Our message to those 37,000 is: “Please work with your case manager through that settlement opportunity, so we can bring closure to the loan charge.” We have contacted all customers to notify them of their eligibility for the new settlement opportunity; we did that at the beginning of this year after the Government announced the settlement. All the individuals affected have a named case manager to work with them personally to resolve their case, and the regulations will conclude shortly. We are seeing inquiries increase as people engage with that settlement. Obviously, we are not in perfect control on the conclusion of that, because it will be driven by individuals’ behaviour. The Government scored the cost of the settlement scheme over a period of five years. Clearly, we would like to bring closure to those customers sooner.

Five years from now?

From the Finance Act, for five years over the spending review period, so it will have been from the Budget, when it was announced. Clearly, we would like to close cases much faster than that, and we think that this will bring closure to the 37,000 people affected and, as you say, be better value for money for the taxpayer, too. We will work as efficiently as we can to support customers to access the new settlement scheme and understand what it means for them, and then hopefully bring closure to the loan charge.

I declare an interest as a member of the all-party parliamentary group on the loan charge. My colleague said that you are spending £31 million a year on loan charge compliance. To date you have settled only £52 million of a total £1.7 billion tax liability, so how can you give us assurance that the approach you are taking is value for the taxpayer?

If I look overall at settlements with employers and individuals of disguised remuneration schemes more generally, that has brought in £4.4 billion in revenue to the Exchequer. That is over the period of 2016 to March ’25. The Government have set out the revised settlement scheme for those 37,000 people affected by the loan charge who have yet to settle. We want to support them in doing so in as efficient a way as we can. Where we can take steps to reduce the costs of administration, we will, but the Government—after what is now the second review—have determined that that is the right settlement scheme to bring closure. If you would like us to write with some more precise details on any further estimates we have on timeframe or the revenue that we expect to bring in as a result, we would be happy to do so to make sure that the Committee has all the latest detail.

Thank you.

We will move on to the issue of child trust funds.

We have received your letter to the Committee, particularly in response to the Committee’s recommendation—from some time ago, I think—that you should evaluate the child trust fund scheme to understand what it has achieved. More pressingly, we are concerned to know what is happening about the 46% of accounts that remain unclaimed. What is HMRC doing to make sure that the intended recipients of the money that sits in those funds receive it?

I agree with you: that is the pressing issue that requires further action. According to the latest data that I have available, just over 3 million child trust fund accounts have matured, and 2.2 million of those accounts have been claimed or transferred to an ISA so far, so approximately 758,000 matured accounts remain unclaimed. Those have an estimated total value of £1.5 billion, with an average account value of around £2,000. Our message to those 758,000 young people and their families—of course there are more accounts maturing all the time—is: “You can go online with your NINO and trace your fund for free.” The Economic Secretary to the Treasury convened a roundtable on 21 April 2026 to create a taskforce that brings together account providers, industry, Government and others to see what more we can do to trace account holders, strengthen communications for young people, pull customer insight and develop industry commitments to support reunification. Along with the research that you referenced that I set out in my letter to you, HMRC will be writing to all 21-year-olds with unclaimed accounts later this year, to alert them to the fact that the fund is still there, they have not claimed it, it is accruing interest in their account, it is legally theirs and we want to do everything we can to support them to match up with it. We will wrap comms and awareness around that, too. We are hopeful that the claim rate that you referenced will improve, particularly once we have done that mailshot. As you say, the research will gather qualitative evidence looking at why young people are not aware and not accessing the money, what the barriers are, and what more we can do about it. We will undertake that research and then happily write to the Committee with the conclusions and what more we can do on the back of it.

I have two quick follow-ups. First, if you know who the money belongs to, what is stopping you just releasing it to them?

HMRC does not hold the money. The money is in the individual’s name in the financial provider’s accounts. If it was child benefit-type benefit money, I would have control and be able to do direct payment-type solutions in a way that I cannot with the child trust fund under current legislation.

Understood. Secondly, the PAC recommendation was that HMRC should be looking at what has been achieved as a result of the child trust fund, so it is a little bit broader than what you have said your investigation will be looking at.

I hope that you can see in my letter to you that we accept your recommendation. We agree that the first, most pressing issue is to improve the matching and unification of young people with their accounts. Of course, the life-chance effect of asset-based welfare will take time, so the next stage would be to return, with Ministers and policy colleagues and the Treasury, to the question, “Would you like to do longitudinal behavioural analysis to understand the effect of the child trust funds?” But, particularly in terms of administration, we think that the most important first thing is to get the money to the young people.

There is no doubt about that, but shouldn’t you have started tracking some of those impacts already, given that some young people will already have claimed their matured fund?

I think that is what we would take into the next round of consideration on our research programme. As I say, 2.2 million young people have now accessed the child trust fund, so—you are quite right—we have a sample who we can start engaging with to see whether that supported them to attend university, buy their first home or whatever it might be. As you know, underlying the child trust fund are asset-based welfare, life chances and trying to provide equal opportunity, so I think there is a good case to do that behavioural research, but given the scarcity of our research budget, we focus first on that first priority.

I may have misheard you. Did you say that you were going to write to all those with unclaimed— John-Paul Marks indicated assent.

Is there a wider comms plan beyond writing to them?

Absolutely. If it would help, we could write to set out all the other comms that we will put around it. You are quite right: there is a bit of me that is—you know—but it is one mechanism that we have, so we will use it. But absolutely, we will do multi-channel, social media—TikTok, Facebook and all the rest of it—to try to raise awareness.

Thank you for clarifying that.

Have you got a target?

No, but I am happy to report progress. If I am honest, it is hard to know how successful it will be, but we know it is a pressing issue.

And you will report back to the Committee.

Happily.

Absolutely. We will be able to do things around measuring the impact of different campaigns and schemes in terms of engagement on the website and whether we get more hits, and, ultimately, we will be able to see the claim rate. The point of the taskforce, though, is also for people to recognise that HMRC can only do so much; we need the financial providers also to make sure that they are exhausting all their comms and marketing efforts to connect with young people. If the Committee is content, I would be happy to write to set out the timetable for the letters, the communications that sit alongside that, and how we will seek to measure success.

Thank you for that. We look forward to that further correspondence from you. Let us move on to large businesses. I suppose you are going to tell us that what you have been doing with large businesses has been successful and collected quite a lot of tax; indeed, I think you collect £95 for every £1 spent on staffing. Are you really saying that you are doing an excellent job and there aren’t any problems?

No. First of all, thank you to the NAO for the Report. We welcome it and accept the recommendations in full. We think it helps contribute to our determination to continue to improve our compliance services and HMRC services more generally. We are absolutely not complacent about the fact that there is more to do. As you say, Chair—I will do this briefly—some of the headlines are encouraging in terms of good value for money, with the co-operative compliance approach at the heart of the model. Penny and Nicole can say a bit more about how we seek to test and learn opportunities to expand that in the future, subject to the evidence. As you say, £337 billion came from the 2,000 largest businesses, which is 39% of all our receipts. The compliance yield delivered in 2024-25—£15.8 billion—is more than double what it was in 2021-22, when it was £7.1 billion. As a return on investment, £95 for every £1 spent is good. The large business tax gap is currently £5.8 billion, which is 0.7% of theoretical liabilities, down from the 1.7% it was 20 years ago. We are driving that tax gap down for large business and we need to keep the large business tax gap down. There is also a set of recommendations from the NAO around providing certainty to business when we request information, and improving our case management systems so that we can exploit data in a more segmented and joined-up way. We have made progress on satisfaction and on timeliness of inquiry, which is down from 35 months to 17 months, but we want to sustain that productivity improvement further.

We will probably come back to some of those issues in the rest of the session. Looking at the success, and the fact that every £1 spent brings in £95, why don’t you spend more?

We are increasing the size of our organisation to, indeed, spend more to increase additional tax revenue. Through the last set of fiscal events, the Government have announced both 5,500 additional compliance officers and 2,400 additional debt officers. Taken together, that will bring additional tax revenues of £10 billion in ’29-30. For the out-turn for this financial year, compliance yield will be around £50 billion, which we think will be a record high, but it needs to increase by £10 billion a year by the end of this spending review period. That is not all driven by capacity; some of it is transformation, third-party data, risking—the team can say more about that—but a significant investment into capacity is being made.

In terms of capacity, how much of that extra investment in staffing will go to the large business directorate?

In the expansion that JP just mentioned—the 5,500 full-time equivalent members that our teams will grow by—because 60% of the tax gap stems from small business non-compliance, we are, understandably, growing initially into our small business compliance. As we bring people in and train them in our foundation learning, that will free up more capacity for experienced colleagues to work on more complex areas. Specifically, in relation to large business, we really welcome the NAO recommendation about exploring the expansion of the co-operative compliance approach. The NAO’s Report recognises that it was a really successful approach in terms of return on investment, so in response to that recommendation, from September we will launch three specific test and learns, which will enable us to really test the cost and value of the co-operative compliance approach to a greater proportion of our business colleagues.

We will come on to the issue of spreading that approach to other aspects of your work. To come back to large businesses, how much extra resource are you putting into the large business directorate? That is the question that I asked.

We do not have that number to hand, so we will have to write to you to give you the precise number. As Nicole says, a large proportion of the 5,500 will be focused on the higher volume, higher value compliance activity, which is necessary to ensure that we get the small business tax gap falling. We are increasing capacity in the complex case teams and the large business teams, but it is in the hundreds of headcount, not the thousands.

Could you let us have the figures, then?

We will give you the precise figures. It is in the hundreds.

What we can say is that, last year, the first part of growing our compliance workforce was recruiting somewhere in the region of 40 to 50 external tax specialists to supplement the growth that we are training ourselves to focus on our priority tax risks. We are forecasting to do the same this year, and we are also growing that into other priority tax risk areas. It is relatively small compared with the 5,500, but it is absolutely focused on the areas that we think need it the most.

So the £95 figure is going to go up in the future, is it?

In terms of value for money and impact—

In terms of the amount that you collect for every £1 spent.

We will see. If we add a bit more headcount, obviously, that increases the admin costs, but once it is productive we should ultimately see some more output. As we say, we have also got the large business tax gap down. To be honest, I suspect the value for money equation will stay relatively stable.

But the amount collected will go up because you are putting more resources in.

We expect our yield to increase from £50 billion to £60 billion per year by the end of this SR period.

Will the increased investment in resource help to deal with the growing bit of HMRC’s work, which is the tax under consideration? We have seen that ratchet up for large business tax disputes, and it is continuing to increase. It is increasing above inflation; we think it is a little over £70 billion now. Can we expect that big pot of money currently being disputed to shrink with this increased resource?

I will let Nicole come in, because she leads the large business directorate. Ultimately, we want to ensure that Nicole has all the resources and capabilities she needs to continue to ensure that we fulfil our legal duty, which is to collect all the tax that is legally due. The directorate is currently resourced to be able to work, at any one time, with about half the large businesses in this country. It has substantive coverage. As you say, that then translates into tax coverage and, ultimately, what is collected.

The important thing to understand about the tax under consideration figure is that it is not a pot of money that is waiting for us to work through and secure. When we open an investigation, it is an estimate of the maximum amount of tax that potentially will be secured. It is a guide for us to prioritise our resource deployment and ensure we are focusing on the highest value. The increase from £42 billion to £70 billion, which is the figure from last October, recognises that we have invested a huge amount in recent years in increasing the amount of time we spend with our customers, making sure we are doing thorough risk assessments and using the data that we get from other jurisdictions and our customers effectively to spot the tax risk. It would not be accurate to describe that as a pot of money waiting to be converted into cash receipts. An awful lot of work needs to go in to understand to what extent that is recoverable or legally due.

We have already raised the possibility of looking at a wider use of the approach for large businesses in other directorates. Sarah Olney wants to explore that.

Penny, given the good results from the large business team, what are you doing to share that learning with other teams in HMRC?

My operational senior team works closely together. All our directorates have plans to improve productivity. We apply the lessons from the large business area in different ways in different contexts, but the basics—be clear about where your risk lies, apply your resources to highest risk, reinforce a focus on expectations within the business and make sure you are deploying skilled resources to the right problems at the right time—are fundamental to all the compliance work we do in HMRC. If you are working across a very large population with an individual single risk, it is a very different operating model than it would be for co-operative compliance, but we share very freely between us the lessons we learn from each other’s work on a regular basis.

What areas of good practice that the large business team have adopted are you expecting other areas of HMRC to adopt in due course?

I will come back to the pilots, but first I will touch on the question of test and learn. One of the things we are going to look at over the next 12 to 18 months is what aspects of co-operative compliance might be applied to the population of businesses below the current large business population. We will examine whether the full-blown CCM model used in large business full stop would work, whether a graduated model would work, or whether there is a sectoral approach that might work. I guess lessons have been learned—we already use a temporary CCM model for our mid-size business. At particular points or events in the life cycle of a business, we would apply a temporary CCM to those problems and work with the business to resolve them. That approach is successful; we just need to test and learn whether it pays dividends in terms of throwing up a more productive result. It is a very intensive relationship and therefore costs quite a lot in terms of resource.

We will move on to the use of the powers that you have.

Back in 2016, HMRC gained legislative powers for your special measures regime. For 10 years, you have had the power to sanction poor behaviour by the worst offenders. Why have you never used those powers?

As you rightly point out, we have not used the special measures powers in the last 10 years, but what we have done is a huge amount of work with our large business customers to tackle the more aggressive end of the tax planning market used by a small subsection of that group. The special measures legislation from 2016 was really tightly drafted to have a deterrent effect and to provide another mechanism to incentivise our large business customers to adopt responsible tax risk strategies. We are finding that the deterrent effect of the legislation, together with the operational activity we already undertake, proves to be highly successful in shifting the behaviour of this customer segment. As the NAO Report explained, we have a high risk corporates programme, for example, which has been running for a number of years now. That has delivered from 70 multinational groups £32 billion of additional tax revenue that would otherwise have gone unpaid, and resolved over 4,000 risks over that time. That programme is small, but really impactful in terms of working in a focused, intensive way with a small group of large business customers that have a large amount of tax at stake and typically adopt an aggressive tax stance. That is proving successful, but in all cases in that programme we consider the application of the special measures legislation. That programme, however, has succeeded with a 96% success rate of shifting the behaviour into a more compliant arena, which is a good outcome. We value the NAO’s recommendation to consider what barriers are getting in the way of applying the legislation, and we have already started having conversations with stakeholders and our policy colleagues to explore how the legislation could be strengthened. For example, in the current high risk corporates programme, we have three cases where we think special measures could apply, but the work that we are doing with them is already shifting their behaviour into a more compliant space. We think the barriers lie in the policy design. In the current policy design, the legislation sets out two hurdles. One is where an aggressive avoidance stance is taken; the legislative bar for that is either the general anti-abuse rule or the disclosure of tax avoidance scheme, which is quite a high bar. The second way it would apply is if there is enduring non-compliance. In the case that has been in our high risk corporates programme for the longest time, the customer has an aggressive tax strategy, but it does not meet the GAAR or DOTAS threshold, and the customer is actually really co-operative—probably one of the most co-operative customers that we work with. That is why we welcome the recommendation, and we will work with stakeholders to explore how we should amend the legislation to shift the behaviours of those really small number of customers that do not shift from their aggressive tax strategies.

Thank you. I will stop there, because my colleague has some follow-ups.

I suppose what I am finding a little difficult to understand is that you said it is a very small number, but obviously a number of them are engaging in such egregious behaviour. If a small number is there—probably the sums of tax at stake that could be owed are significant—why has this special regime never been used, especially considering HMRC officials came in front of this Committee in 2016 and told us that it would fill a gap in your armoury if you had this special regime in place? Why is that? Were the officials then incorrect, or are the officials now, sitting in front of us, incorrect? Which is it?

It is a really important part of our armoury. In the original design, it was intended to have a deterrent effect by being on the statute book, and I think that we can say that it has achieved that, because over time we have seen large businesses adopt a much more cautious risk approach to their tax compliance. To that extent, I think that the policy intent behind the regime has been achieved. That, coupled with the strong operational activity that we have deployed, has had, as I said, 96% success in shifting the highest-risk corporates into a more compliant space. The deterrent effect in the policy intent has been achieved, coupled with the strong operational performance, but that is not to say that we do not recognise that the policy needs to be revisited to make sure that it can work in the small number of cases where we still see issues.

If the deterrent has been effective, are no concerns or alarm bells raised with you by the corporation tax gap not actually decreasing in absolute monetary terms since this regime came into effect? If you look between 2021 and today, that gap has gone up by some £2 billion. Also—I mentioned this earlier—we have a very large amount in tax under consideration, where it is being disputed. Do you not think that actually deploying the special measures regime might help to reduce the corporation tax gap, and might start to deal with the very large tax-under-consideration issue that you have?

I think that we are seeing a shift in the tax gap in large businesses across the UK. If you look at the latest published tax gap estimates, the breakdown of behaviour shows that 50% of the large business tax gap stems from behaviour that we characterise as “legal interpretation”, which is a broad spectrum. In 2005-06, the tax gap for large businesses stood at 1.7% and a much bigger proportion was related to avoidance. Now, only 5% of that tax gap relates to avoidance and 2% to evasion, which means that that threshold set in the special measures legislation—with the general anti-abuse rule and the disclosure of tax avoidance scheme—is quite a high bar. That absolutely does not cover all the 5% of the avoidance tax gap in the large business estimate, and therefore I think that the nature of the tax gap has shifted over time. The regime was necessary at the time; it is still necessary now as a deterrent; and it is worth revisiting to see if we can strengthen it.

What I am finding difficult to get my head around is the fact that in recent years HMRC has actually been given a number of additional powers. We have already discussed the fact that the special measures regime introduced in 2016 has never been used. HMRC has only once taken action to prevent the facilitation of tax evasion, and that was not against a large business; it has never done so against a large business. HMRC has never enforced a penalty against an enabler of tax evasion, despite having those powers for some nine years. I am finding it difficult to get my head around—I imagine that plenty of my constituents are as well—whether these deterrents are actually robust, and are ever being implemented. Is this a case of very wealthy individuals or large businesses being let off the hook by HMRC?

Perhaps I can say something about the work we do with our fraud investigation service in respect of large businesses, and Penny may then want to come in more broadly. We work really closely with our fraud investigation service, and we have developed counter-fraud capability within large business. Over the last few years, we have sent over 200 referrals of suspected fraud, and we are contributing to a number of operations that our fraud investigation service is undertaking. In relation to the corporate criminal offence and the failure to prevent tax evasion, we have 43 ongoing investigations at the moment. We have an increasing range of powers, and we are seeking to use them in the most appropriate cases, recognising that we are working with a broad spectrum of behaviours.

On that point, can you understand the public frustration that it has never once been use against a large business? I think the one case where it was introduced involved a small accountancy firm in Stockport, but it has not been deployed against any large business, and you have had plenty of time to get on with it. Can you understand why that causes significant public frustration?

I completely take on board the challenge. We took that challenge away when we came here to give evidence on the Report on value for money and wealthy tax compliance. We are working on that, and we will come back to the Committee in the autumn with a response to the recommendation on the use of powers. I can assure the Committee that there is no let-up in the work that goes on, particularly in the case of the corporate criminal offence. We have looked at more than 125 specific instances, and there may have been specific reasons in each of those cases why they did not meet the bar for taking the case forward under those powers. That does not necessarily mean that we are not acting in other ways. For instance, the Committee will be aware of the case where we were supported by the CPS on a deferred prosecution agreement with Entain a year or so ago. We are not afraid to take on the biggest businesses whose behaviours are evasive or deliberately dishonest, but we have specific powers and we need to explore—we do actively explore this—whether those powers are useful in specific cases. As Nicole said, we have a number of those cases on the books at the moment.

I am aware. I suppose it would be helpful for the whole Committee if you can reassure us that you have a strong enough body of evidence that supports such limited use of these powers. How can you reassure us that the body of evidence is sufficiently robust that you can say, “We have never, or rarely, used these powers, but it’s okay because they are effective deterrents, and they are making sure that large businesses are tax compliant”? How can you reassure us of that?

I suppose we would look to the evidence in the NAO Report itself, which I think tells the story of the work that we have done in the context of things such as the high risk corporates programme that Nicole described. It sets out the shift in behaviours that we have achieved through that work. That does not mean that, as Nicole said, the deterrent effect of those powers on the statute book is not important, and we actively consider the use of them in that programme. Just because we do not apply them, that does not mean that they do not have an effect. I cannot evidentially prove to you with a piece of maths that it has an effect in a specific case, but the change in the shape of the tax gap is evidence, in itself, of the behaviours that we have seen. Special measures were specifically directed at the kind of high-profile, high-risk avoidance that is tackled under the general anti-avoidance regulations. That kind of avoidance is not prevalent or really seen in the contexts of large businesses now.

Where you do have that very small number, which officials here have already referred to, of particularly persistent aggressive tax planning, non-compliance or egregious behaviour—whatever we want to call it—are you confident, and can you reassure this Committee, that even in those cases, where big sums of potential public money are at stake, that deploying more enforcement powers would not make a significant difference to dealing with those issues and collecting the right amount of tax that is owed?

We will constantly look at the powers that we have. As Nicole said, in the context of those particular powers, we will look to see if there could be change within them that would help us further in the shifting shape of the tax gap and the behaviours. I do not want us to remain closed to that notion and think there is nothing else we can do. We constantly look to see if there is more that we can do that can strengthen the arm of our compliance investigators and that will enable us to do a better job of managing tax compliance.

The assurance that we can provide is that, in relation to the 2,000 or so groups that the large business team deal with, we have an intensive relationship with them. We undertake thorough business risk reviews regularly and from them, we identify those that are low, moderate, moderate-high or high risk. The latest statistics show that over half of them are low risk. We understand that we currently have 10 that are high risk, and that is where we actively consider the high risk corporates programme and special measures, and work with our fraud investigation service on joint operations, where that is appropriate. There was one recently where we worked together and where, in a large business context, there was pay-as-you-earn fraud operating. We worked with the fraud investigation service and we considered the corporate criminal offence powers that we have, but the best way to tackle the non-compliance that we found was the individuals who were operating within that corporate environment and the intermediaries that were selling it. That was the best way to address the fraud that we had identified, so I am hoping that, as a Committee, you can take some assurance from that in-depth knowledge and understanding, and that in the appropriate small number of cases we are using the breadth of our powers appropriately.

That is helpful. Having discussed the use—or lack of use, in some cases—of criminal enforcement powers by HMRC, I would like to briefly touch on civil investigations of tax fraud. Could you tell us how many civil investigations into fraud you have completed in large businesses?

I am afraid we can’t. We have a process where, as I said before, we have developed an expert counter-fraud capability within large business. That team is responsible for making sure that where we identify instances of suspected fraud, we work closely with our fraud investigation service: we set flags, make referrals and work together to identify the best way that we can use our counter-fraud capability across the whole of HMRC to bring that to bear on that small number of cases. As cases move across into our fraud investigation service, their case management system to monitor that is not capable of being cut by customer segment. When we were working with the National Audit Office, we were able to track a handful of cases that involved large businesses, but we cannot get systemic data on how many large business cases are subject to fraud investigation.

Just to be clear, does that information not exist? Do HMRC not know?

No, but as part of the transformation investment that we mentioned earlier—we are investing in improving our management information capability; that was another of the recommendations from the National Audit Office—

Do you think that not having that information creates some gaps within HMRC’s thinking? You cannot really have a full understanding of the scale of potential tax fraud within these 2,000 large businesses if you do not have an effective way of measuring how many civil investigations you have even opened or processed.

We have an understanding from the tax gap, which says that 2% of the tax gap in relation to large business stems from evasion behaviours. We have the over 200 cases of suspected fraud that we know we have referred to our fraud investigation service, and we know the cases that are in the public domain and that we can track through. It would undoubtedly be true to say that our management information understanding will be enhanced when we can cut our fraud case data by customer segment, and we will have that capability within the next 12 months. I think that will be an enhancement to what we can already do.

We publish annually the fraud investigation service’s output, as you know. We have the commitment to increase annual prosecutions by 20% in this spending review. We have about 1,000 in the pipeline, so we have significant activity, but we have more to do to improve the quality of our internal systems to support the MI.

Permanent secretary, my frustration—others on the Committee may share it—is that if we see very limited or non-existent use of the actual criminal proceedings that you can take forward, and then sitting alongside that, we see what I would describe as a limited understanding of the number of civil investigations that you are taking forward against large businesses, it makes it quite hard for us to have absolute confidence that HMRC is using all the powers at its disposal effectively and knowingly to try to collect the right amount of tax. How can you reassure us that, despite limited use of those criminal proceedings and not having a clear understanding of civil investigations, that is not the case?

It is a fair challenge. You have made this point to me before. In my first year, I looked for proof points. You have heard about a number from Nicole and Penny. The NAO Report is one of those. If I look at penalties issued by the large business directorate, 164 increasing fourfold to 636 is a proof point. If I think about governance and assurance that the processes and consideration are being applied properly and well, the NAO found well-established arrangements consistently applying governance, and no evidence of unfair settlements. That was true for 100% of the cases it looked at. On special measures, I have looked at cases with the team. I am very assured on the deterrent effect. I think the team are explaining that to you well, but we hear the frustration: “You’ve not used it.” We know that it was a big deal to be used, because of the brand impact that it would have on a corporate. It is a very high bar, and the legislation constrains its use, but we are not afraid to use it. The team RAG-assess the 2,000 cases and we, of course, look at the red cases all the time to make sure that we are doing everything appropriately. We know that we have to improve some of the data. The final thing is that we are very committed to transparency. We want, in our annual report and accounts, to be as clear as possible on where we have used powers, where we have not but what we think the deterrent effect has been, or what more we need to do to make sure that we have everything we need to drive the tax gap down. I think the large business directorate has made a good job of doing that in the years gone by, but we are not complacent at all about the importance of continuing to do it. I am very happy to report back to the Committee. I will write to you ahead of the next annual report and accounts and we will continue to try to put more data in there about the use of powers—civil or criminal investigation—and the impact. I recognise that you will keep asking us until we can give you reassurance and show you the proof points that that is robust. I see lots of good indicators of that, but we are very happy to continue to respond as well as we can.

We will move on to Rupert Lowe to look at the administrative burden, and probably one or two other issues as well.

Thank you, Chairman. In preparing for today’s meeting, I came across a very good quote that I would like to remind you all of from the very eminent US Chief Justice in the 19th century, John Marshall, who said that the power to tax is the power to destroy. I think that is a very powerful quote that you can perhaps take away today. On the face of it, this Report looks good, but I think we are looking at quite a soft target in big businesses. Most of the tax you are not collecting is not coming from big businesses; it is coming from small businesses, particularly businesses like Turkish barbers and vape shops, which you are doing absolutely nothing about because it is probably more difficult. The cheap deliverables are much easier to deliver when you are looking at big businesses, where you weaponise people’s financial advisers against them and they are effectively acting for you, not for the company. I noted that, in the Report, only 49% of large businesses surveyed by HMRC in 2024 agreed that the administrative burden relating to tax compliance was reasonable. That means that 51% thought it was unreasonable. Why do you think that is?

I will kick us off. You are absolutely right that in the last large business survey, 49% felt that the administrative burden of the UK tax system was reasonable. It was 29% that felt it was unreasonable. The balance did not comment—

But 51% were not entirely convinced it was—put as much spin on that as you like.

However, 29% did nail their colours to the mast and said that they thought it was unreasonable. That has to be taken with the other elements of the large business survey, where 91% of our customers said that they valued and had a good relationship with their customer compliance manager, intimating that they value the certainty and contact that we have with them. Overall, 81% of them said that they had a good experience in their dealings with HMRC. What I have heard from the customers that I work with is that the tax system overall—thinking about the implementation of pillar 2 and the other mechanisms and measures that we have put in place—means that there are burdens on the business community. However, I would say—as the NAO Report does—that there is feedback from our customers that we ask for a large volume of information in the course of an investigation or inquiry. Automating the collection of data will make sure that we can risk-assess much more accurately and avoid costly investigations further down the line. One example is the introduction of the international controlled transactions schedule measure announced last year. The feedback in the NAO Report says that when we have opened an investigation into transfer pricing—as one common risk we see in the multinational population—in order to understand the position and base the comparables under the OECD guidelines, we need a huge amount of information. Our customers tell us that costs a lot and they are not always clear why we are asking for it. As an example, the introduction of the international controlled transactions schedule means that we will follow suit with many other G7 countries that have already implemented it. It will not add to the administrative burdens of our customers but will enable us to be more effective and targeted in the work that we are doing—hopefully reducing the amount of large-scale information requests that we sometimes need to use with our customers.

They are expensive for people to gather. In my experience of HMRC—again, I asked some parliamentary questions on this—it takes you 26 and a half minutes to answer your telephones, or it did in 2024-25, whereas it took the DWP eight minutes to answer theirs. That says a lot about the people paying the tax compared to the people who are being handed out the tax that is being collected. When people are trying to sort an issue out, it does not help when you do not answer your telephones. Why can you not get that right?

The average speed of answer at the moment is about seven minutes. I agree with you that in 2024-25—

I have done it myself. Sometimes I give up, as do a lot of people. They waste a lot of time, and that costs money.

We regret that, of course.

Are you going to improve that? Is it a priority?

It is a priority and it is improving. Our average speed of answer in March, April and May has been in single digits—so below 10 minutes. That is half what it was a year ago. We are on track. We will write to the Committee ahead of our annual report once we have finalised the outturn data, but we look set to have hit our telephony standard for the first time in five years. I agree that it is very important that the front door is delivering exceptional customer service and that people can get through and increasingly resolve their inquiry first time, either online or over the phone.

I think that that is only fair for the businesses, because some of them are up against it anyway. We will go on to talk about other burdens that you put on businesses, which—in terms of my point from the John Marshall quote—are arguably holding back a lot of British enterprises, because they are spending their time collecting tax for you. But at the end of the day, I am pleased about that—hopefully the phone lines can be answered, like most people’s phone lines, immediately within three rings, but probably that is asking for too much. Figure 4 on page 19 of the Report was very illuminating. I thought those charts were fascinating because, in my experience of dealing with HMRC, you do not know your own rulebook. Your staff turnover is massive and very often your staff do not understand the rulebook either. That is my experience of it. When you drill down and see that you have 200 tax manuals totalling 80,000 pages, and that the UK tax code is 22,000 pages in two volumes, it is hardly surprising. For large businesses, when you look at the mauve bar, legal interpretation is a big issue. In my experience, you have often come at me saying that something is, by your own admission, a grey area, but you are going to raise the tax anyway. You use your larger covenant, effectively. You have the taxpayer behind you—businesses do not; they actually have to make profits and pay taxes. I think, actually, you bully people using this wretched rulebook, which has got more loose ends than a plateful of spaghetti. This is a very interesting chart and, even if you look at the light blue bit, that is error. Well, it is an error because your rulebook is so complicated, in my opinion. On the burden on businesses, why can’t you simplify your rulebook? It is too complicated. Do you agree that is a fair comment?

Let me say something about what I see in terms of the legal interpretation—the mauve bit—and then maybe Jonathan will want to come in on the rulebook more generally. Legal interpretation—the mauve bit of figure 4 on page 19—is a really big spectrum of behaviour. At one end, you have the bits of the tax code that have been there for a long time and have not kept pace with modern working practices. For example, where there is a permanent establishment and everyone is working from home in different places now, the tax code has not kept pace. At the other end of the spectrum, we see in a small number of cases what we determine as boundary pushing: taking positions that are legally arguable but against the intention of Parliament. For example, we see some of the VAT reliefs being claimed in inappropriate circumstances.

We are slightly getting into the weeds, because on working from home, you are not entirely clear what your own rules are, so you picked a bad example. I am interested in whether you think that part of the issue you have with these big businesses is that they cannot work out the rulebook. What is more interesting is that, when you look at all taxpayers, a lot of people do not understand the rulebook at all, which is why the mauve bit is so small there. They get confused by it and that is why you are able to open up all sorts of issues. When I say your staff are not good enough, I have had tax investigations in businesses where your staff have changed three times. That means that you have to re-engage with your staff, and they get loads more costs and loads more administrative burdens. I think you are doing a lot to damage British enterprise. It could be made a lot simpler, and people could live their lives. Most people want to pay their tax, and they want to be honest; they do not want to be dishonest. But your rulebook is so complicated that half the time they try and do the right thing, but you might take a different view to this wretched homunculus of a book, which is not clear. That may work to your advantage because you may then be able to tax people more, unfairly. We have to get to the bottom of this if we are to solve issues. My colleagues talk about collecting more tax and prosecuting more people; I am more interested in making tax clear so that people do not need to waste time interpreting your wretched rulebook.

I will say two things and then Jonathan may want to come in. We do respond to our customers’ feedback. When they tell us that there is an area that is unclear, we issue something called “Guidelines for Compliance”—I think we are on publication 17 of that. A big part of the mauve bit is about the complexity of multinational corporations’ tax affairs when they are operating across multiple jurisdictions. Transfer pricing is one example I mentioned. That is part of it, and we issue guidance to help our customers get their tax right. In terms of staff turnover, for our customer compliance managers in large businesses, 71% have been in post for more than three years and 40% for more than five years. We recognise the impact of turnover, and we try to minimise that wherever possible to give consistency.

In general, there are a number of drivers of why you have a more complex tax system. Sometimes, as I said, the economy is changing so the tax code needs to now adopt cryptoassets. Cryptoassets did not exist a few years ago when many of the basic tenets on taxation were laid down, so we do sometimes need to put out additional guidance to make sure that we correctly reflect economic innovation. The economy has changed drastically in the last 20 years. It is more international and more invisible.

Eighty thousand pages. Doesn’t that need slimming down?

Well, if it can be slimmed down, we want to slim it down, but we also want to—

But you don’t slim it down, do you, Jonathan? I do not think it gets slimmed down.

When we can, we will do so, but sometimes that reflects the will of Parliament in terms of legislation. If new legislation comes forward, new taxes come forward. We have two new taxes being developed at the moment: vaping duty, which was launched relatively recently, and the carbon border adjustment mechanism. They will require legislation and guidance to go alongside them. That guidance will also reflect what is going on in terms of the policy environment. Again, the policy environment is set by the Government of the day—

Do you know how long the Hong Kong tax code is?

No, I don’t.

Five hundred pages. We have 22,000 pages. If you take anything away, I think you could make life a lot easier for you and for our businesses if you made your tax manuals much simpler.

We will always look for simplicity, but shortness does not always imply simplicity.

I know I am going to use all my allotted time—a bit like in the Senate in America—so I would like to push on now because I have a couple of other things to ask you.

Just remember that we are dealing with large businesses.

I am talking about large businesses. If you look at this chart, either they are making errors because it is so complicated or their legal interpretation does not agree with what HMRC’s rulebook is saying. There is your issue. That is a burden on our businesses. It is a massive burden, Clive—massive. They have to engage tax lawyers who get paid a fortune, and that is a further drag on their businesses because they have to consider not only time but cost. While we are on the burden on businesses, you already burden people with the collection of pensions, collection of national insurance, collection of PAYE, collection of VAT, collection of—

We are straying into policy areas here.

We are not. We are talking about the burden on businesses, Clive. This is a further burden—they gift wrap it and send it to you.

This is not an HMRC decision; it is a decision of the current Government, so we have to be careful. We are not here to question whether they think the Government policy is right.

Now we are moving on to another cost on businesses, which, again, all my businesses are having to suffer: Making Tax Digital. That is another cost that, by your own admission, is going to cost businesses an awful lot of money, so it is a burden.

I think we are straying. I understand that large businesses are already doing this. It is a step forward for many small businesses, but we are not dealing with that area. We are dealing with large businesses, which I suspect are already digitising most of their work anyway.

I am not straying far from these charts on page 19. I think they are highly illuminating.

We can respond to a lot of that if it is helpful, but I appreciate that time is tight. I am very happy to set out, for example, what we are doing to tackle non-compliance on the high street. You raised that point. We can do that either now or in writing. We agree that complexity at the heart of the tax system is a problem. It adds a challenge for myself in trying to administer it, and for everybody—individuals or corporates—in understanding it. That is why we need to do everything we possibly can, given Government policy, to administer it as carefully and as well as we can. If I take the large business report, the satisfaction rate for large businesses on their relationship with their case manager is really high. Around 90% say that how we are working upstream to help them to get their tax right first time is very good. As you say, we are making sure that they are not lost in the complexity of the system and that they are supported to get it right; legal interpretation is one example. We are committed to a radical transformation of the tax system to make it digital-first, and to try to use technology to make it easier to comply. As you say, the vast majority of people want to get their taxes right first time.

Quarterly, not annually. That is what you are asking for.

Making Tax Digital is a transformation from a legacy approach where taxes would be reported in arrears with high error and without integration with accounting, invoicing or other operating systems. When it was done for VAT, in the evaluation there were good behavioural responses, which Jonathan can talk about if you are interested in hearing the evidence. We are confident that Making Tax Digital will lead to more real-time reporting, the integration of invoicing with transactions and accounting, and quarterly returns that are enabled by software. It will improve the digital capability of our economy, support productivity, reduce error and support yield. Ultimately, although I appreciate the quote at the beginning, my teams are collecting the revenues that fund the nation—our NHS, our schools and hospitals, our transport system—

Let’s not get into a political debate about how much of it is wasted.

So it is important that we recognise—

If you could put it in writing, that would be fantastic—I really would appreciate that.

That would be helpful.

I would be delighted to.

In the end, these burdens are damaging the wealth of Britain and damaging the amount of extra tax you could collect.

I think we have to move on now. We normally have a break at this time in our session so that everyone can collect themselves and come back refreshed. We will have a five-minute break and restart at 4.35 pm. Sitting suspended. On resuming—

We will move on to further questions, starting with Lloyd Hatton.

I would like to spend a little of time on how we ensure that HMRC works in a way that most people would believe to be fair and consistent. I know that HMRC has reassured us repeatedly in recent years, in the public domain, that the era of sweetheart deals is over, and I believe that the NAO Report reflects that, but could you reassure me that in the case of HMRC and GE, for example, the way that HMRC operated was fair and consistent, and did not deliver poor value for money to the taxpayer and when it came to making sure that the right amount of tax owed was paid?

I will start with the system and then the team might want to comment on the case, taking care in doing so. There is very robust governance across HMRC that has been developed and is transparently set out in terms of our litigation and settlement strategy. We publish our tax assurance commissioner’s report as part of our annual report and accounts, and we will write to the Committee again in the next couple of months with the latest on that as well. I am grateful to the NAO for looking at cases and being able to provide the assurance that, for that sample, good governance was robustly applied, consistent with that approach. Penny can say a bit more about the governance of the compliance operation and the different levels, and how that is run very carefully. As you say, consistency really matters. Fairness sits right at the heart of our charter, and I look for proof points on that all the time. The NAO Report is one of those. Similarly, at the other end of the spectrum, when cases go to litigation, HMRC has a success rate in litigation of over 90%, so the court is reinforcing back to us that we are getting it right in the vast majority of cases. Of course it is right that we test the law where we need to—where there is uncertainty—but obviously we need to maintain that high success rate too. We are looking for indicators and proof points all the time that the governance is consistently applied. Nicole made reference earlier to training, capability and retention, which are also fundamental. Penny has been doing a huge amount of work on our tax academy, ensuring that we are embedding the capabilities that we need to operate at the highest standard. We think this is fundamental to our charter. If there are particular points of interest or concern—you mentioned a particular case—we can try to provide some assurance on that.

Will someone come back on that case now, to provide some reassurance to the Committee?

Does anyone want to say anything on that? No. As you know, there is a duty of confidentiality, which must apply to all individuals and all businesses in the UK, and the team cannot comment on those cases. If there is a specific feature of that case that you want to write to me about, where you are looking for reassurance in terms of the policy or the compliance operation, we could respond, potentially, but we cannot speak about specific cases in a public domain like this.

Okay, so if I was to write, you would be happy to write a private response to the Committee, but you are not happy to speak publicly to the Committee today.

We would reply not about the case, but in more general terms about the concern that you might have. For example, you raised value for money. If I take large business, I would reply letting you know that, on average, large business value for money is £95 for every £1 invested, and that we have a very good confidence level on our hit rate, or our yield per FTE rate, for our large business investment.

I will try not to stray too much, Chair, but it would be useful to know this. A great deal is in the public domain—tax groups such as TaxWatch have put a lot of information into the public domain—in relation to the GE case that I mentioned. Is there nothing you can say in front of the Committee today, in relation to what is already in the public domain, to reassure us that HMRC did not in any way inadvertently walk into a sweetheart deal because of the way that it conducted those investigations?

The legal duty that relates to taxpayer confidentiality is at the heart of our charter, and it pertains to all individuals and all companies, including where information is in the public domain, so we cannot offer public commentary at any time.

Okay. Moving on, one of HMRC’s biggest ongoing tax disputes—again, this is all in the public domain—is with the UK oil giant Glencore. I believe there is a 15-year-long tax dispute between Glencore and HMRC, with around £1.5 billion at stake in terms of potentially owed tax. Again, that is in the public domain. I am keen to glean from officials, if possible, whether you feel that this case is ever going to be settled and whether that £1.5 billion that may be owed is ever going to be collected by HMRC. Do you share my concerns that, with a lot of this, we will not know the outcome until it is a fait accompli, and that we might actually find that none of the tax that was potentially owed is given over to HMRC, because the case essentially all gets done in an overseas court?

There is a lot in that question. We cannot talk about the case at all. I guess there are a few things that I could just repeat. First, when HMRC engages in litigation, we are successful more than 90% of the time. On your point about timeliness, aged cases—those over three years—have come down this year by 38%, so we do of course recognise the importance of efficient, timely inquiry and bringing that to a conclusion. On value for money, for those highly complex cases, whatever the outcome, the team will of course always engage in asking, “What is the learning that we derive from this? What does that mean in terms of the wider system? Is there a policy consideration, or whatever, that we need to reflect on?” But it is impossible for the team to comment on the specifics of the case.

I suppose what is slightly odd there is that a huge amount is already in the public domain in relation to the tax dispute between Glencore and HMRC, largely because of a number of associated court cases. Given that such a significant amount of money is being considered, and that the dispute has been going on for such a long time, can you give the Committee any reassurance that it will be concluded in the near future?

I don’t think my team can comment on this case in the public domain in that way. The litigation and settlement strategy is geared towards ensuring settlement in as efficient a way as possible that is accurate according to the law. We have a duty to ensure that the right amount of tax is collected under the law, and that is a duty that we must fulfil. The duty of confidentiality applies to all individuals and companies, whether or not the information is in the public domain, and there can be no exception to that. I do not know whether Jonathan wants to say anything in addition to that.

In terms of our oversight of the largest cases, we have something called the tax assurance commissioner, which is a role created in 2012. Cases involving more than £100 million of tax or £500 million of adjustment are automatically referred to the commissioner, so the largest cases are given an extra degree of scrutiny in our process. You referred to overseas courts. There is also something in the mutual agreement process. That is not about the application of UK law but about the application of taxation rights under double taxation treaties. That is a separate process from UK law that determines, under tax treaties that have been approved by Parliament, how those taxation rights should apply. That is a slightly different process from determining the overall quantum of tax.

Thank you for that extra detail. The point I was making is that when you have big sums of money at stake and long-running tax disputes that are very much in the public domain, as the Glencore one is, it is worrying when we see no end in sight to these investigations, which have been going on since I had a full head of hair and no GCSEs. Here I am now as a Member of Parliament, and that dispute is still ongoing. It is in the public domain—you can find all this stuff online very easily—so it would be useful to have a clear understanding of when we can expect to see these big, long-running disputes concluded. It does not look like you will be able to give a detailed answer today, but perhaps you will be able to provide a little more information confidentially in writing. Taking a step back from the two cases I have raised, how can you reassure the Committee, Parliament and the public at large that large businesses that likely artificially divert profit are pursued by HMRC in a fair and consistent way, so that they pay their fair share of tax?

I think the NAO Report reflects some good assurances in that regard. Nicole, do you want to say something in terms of the large business directorate?

Absolutely. Making sure that large businesses pay the right amount of tax in the UK that fairly represents the profits generated here is a priority for us. Whether that is through transfer pricing, thin cap risks or corporate interest restrictions, we have been at the forefront of work on global taxing rights through the OECD base erosion and profit shifting work. In terms of how we work with large business customers and make sure that we adopt a consistent and fair approach, we do a range of things. Of the £70 billion or so tax under consideration that you mentioned earlier, about £21 billion is in relation to international tax risks—making sure that we are securing the right level of UK taxation. That work is undertaken by international tax specialists, some of whom we have trained internally through trebling our funding for the advanced diploma in international taxation. As I mentioned before, the growth in external recruitment of external experienced hires has supplemented the work that we do on that. Those colleagues work together in regional and national governance that oversees the consistency of how we apply diverted profits taxation, thin cap legislation and transfer pricing OECD guidelines, so that we are consistent nationally and making evidence-based decisions in all those areas.

Do you think you have got a grip on it? Does the current way of working mean that we do not have large, often multinational companies able to artificially divert their profits, so that HMRC is not able to collect the amount of tax that it otherwise would?

Internationally, in the last 15 years or so, there have been huge strides on this. We work with partners around the globe to make sure that we have fair taxing rights. Pillar 2 is the most recent example of that. We are on track to implement the UK legislation, with the first returns due next month. So, yes, we have made huge progress. It still remains a significant risk that we tackle in the work of our teams, but we have growing expertise in that area and, as Jonathan talked us through, we have really sound and robust governance and oversight of the decisions we are making in those cases.

That leads us on nicely to pillar 2 rules.

Another good segue. Pillar 2 is about complying with the OECD agreement to levy a minimum tax. What is HMRC doing to ensure there is sufficient support to those companies that will come under the pillar 2 regime to make sure they are compliant?

We have undertaken a huge amount of work to support our customers who are within the remit of pillar 2. As you said, pillar 2 is an OECD activity that we have been at the forefront of. It applies to multinational corporations that have a turnover of over €750 million. We have a global minimum tax. In the UK, we have been working with international partners and the OECD to make sure we have clear legislation and guidance that apply internationally wherever possible. We are an active participator in the community of experts internationally, and last October we hosted the Amsterdam dialogue, where we had 50 jurisdictions and the UK, so we play our part on the international stage. Domestically, working with our UK large business customers, when we first started legislating for the introduction of pillar 2 in 2023 we set up a dedicated team, and we have increased the size of that team to make sure we are supporting our customers. We have done that through targeted mail—we said earlier that that is one mechanism—and we supplemented that through adviser workshops, working with third-party software providers. There has been a series of webinars, and the team have dealt with almost 2,000 technical and practical queries about how customers are getting ready. We are at an almost 90% registration rate for groups that we think are within scope of pillar 2, which is good. The IT platform that we have had to develop has been open for registration, payment and tax return submission since last year, and we are on track to have the GloBE Information Return IT platform ready ahead of the June deadline. We have invested a huge amount to make sure our customers are well supported with implementation, recognising the technical complexity of it. Jonathan, do you want to add anything more?

No, not really. It is a new responsibility for businesses, so we are asking them to do things differently, and it does require a different computation from how they would normally do their corporation tax. We have reached out, and we are also working with the OECD where we can to make sensible rationalisations to the rules. Obviously, since the rules were started, there have been iterations and ongoing discussions. Where we can, we are looking to simplify and take out requirements on businesses that are no longer needed. It is a living and breathing policy development.

Mr Athow, to what extent are businesses having to incur undue cost to comply?

It is a different basis of computation. It is an alternative minimum tax type of arrangement, so they have to compute tax in a different way from how they would for UK corporation tax.

Are these side-by-side computations, or is this new one replacing the other one?

Essentially, there is an OECD-wide approach that must be implemented by every jurisdiction within scope of pillar 2. We operate the corporation tax system, and they also have to calculate—

So it is two separate calculations.

It is two separate calculations, on a slightly different basis. As long as you are above the floor, you don’t need to pay any additional tax. It is a different approach, but it is internationally comparable. It is an extra requirement that we have asked businesses to do. As I said, where we can take out steps in that process because they are no longer needed, we will look to do so. It is a different requirement, and it is an extra requirement on those businesses to report in two different ways, and then make certain that they are consistent with that minimum tax.

Ms Newbury, how is the new agreement from January this year likely to impact compliance with the pillar 2 rules?

On the side-by-side agreement that was reached in January, we published updated impacts of that at the spring forecast. It has reduced the benefit—the additional tax that will be paid in the UK—by about £600 million a year. The forecast for what pillar 2 will bring into the UK has now reduced to £1.6 billion a year, so there will be a monetary impact.

This is the top-up payment where companies were paying less than 15%—

Yes, a minimum effective rate of 15%.

And then the new agreement and the way it will be calculated results in a reduction.

Because the US already has minimum taxation rates domestically, the side-by-side agreement removes US-headed groups from the calculation, but US-headed groups are still responsible for reporting and paying pillar 2 in the UK in relation to UK profits. So it does have an impact, and it only applies from 2026—from the implementation of the side-by-side agreement.

Does that change introduce more cost for businesses?

That is not feedback that I have heard from businesses. I engage with my customers and their representatives frequently on this, given how close we are to implementation, and I have not heard feedback that that has materially introduced more cost.

Are you forecasting any change to the level of compliance as a result of the change?

No. We have been doing a lot of planning with the specialist team that I mentioned, which we set up in 2023. We have been doing an awful lot of planning and forecasting around what our compliance activity might look like. We think that there will be less activity to do on some groups, but there will still be proportionate and reasonable compliance activity following implementation.

We will now move on to compliance interventions.

I want to ask about the time it takes for these cases to be resolved. The Report says that the large business directorate has been closing compliance interventions more quickly in recent years, and that we should congratulate you on that, but it is still taking 17 months, or at least it was when the Report was written. What further work are you doing to bring that number down?

I am really proud of the team and the work that they have done on that. There are a range of factors that contribute to how long an investigation lasts. Some of it will be about the complexity of the tax issue under consideration and how settled the law is; some of it will be about how collaboratively we work with the customer; some of it will be about the volume of information we need to seek to check a tax position. Of course, the time will be increased significantly if we have to litigate and go through the tribunal process. There are a lot of factors that impact it, but we have been, and continue to be, relentless in making sure that we are as quick as we can be with the work that we need to do.

Do you expect that figure of 17 months to reduce?

Yes. The forecast for the time taken, which we anticipate publishing in July as part of our annual report and accounts, shows a dip to 16 months, but there is more that we can do. We are continuing to focus on our older cases and making sure that we are focusing particularly on the legal and specialist support that we need, and on making sure that we are proportionate in the amount of information we are requesting so that we can make timely decisions. The other thing worth noting is the opportunity for investment that we have had as part of the spending review. We are thinking about how we can use technology, and AI in particular, to speed up the work that we are doing. I have mentioned a few times the work that we do on transfer pricing and some of the international work, where we often have to ask our multinationals for huge amounts of unstructured data and information. We are using tools that are commercially available, which we are tailoring and training with large language models, to make sure that we can extract the information and analyse—with human involvement and oversight—what that is telling us much quicker, so we are hoping that we can speed up those inquiries even further.

A few moments ago, you mentioned litigation. I know it is a much smaller cohort of cases, but where businesses decide to litigate, their cases are taking significantly longer. I thought I had misread this figure, but they were taking 97 months in 2024-25. That is a long time. What work are you doing to bring that number down?

We have already talked a bit about our litigation settlement strategy. As part of that, and as part of our code of governance, we are really clear that we recognise, from a time perspective and a cost perspective, that we want to be working openly, collaboratively and transparently with our customers and that, wherever possible, we want to resolve civil cases through agreement, with the appropriate oversight and governance that we have already discussed. Doing that to get to a position where we understand the respective positions and avoid litigation is absolutely what we aim to do. We are increasingly using alternative dispute resolution as a mechanism to try to avoid litigation, which is yielding benefits in terms of reducing the time taken and the number of cases going through litigation. Litigation remains a really important safeguard for our customers in terms of due legal process where they disagree with the view that HMRC has. I do not envisage that we will ever remove the need for litigation, but we are trying, wherever possible, to work within our compliance professional standards, use alternative dispute resolution and make sure that, through that collaboration, we can minimise litigation.

I am not clear on how that will help to bring that number down. Are you saying that it is unlikely to come down?

Do you mean the time taken to settle cases or the proportion of cases going to litigation?

There are a couple of things that might help. I am with you: the numbers are too high. As Nicole is setting out, one is about whether we can support a settlement process that gets the tax that is due paid earlier without litigation, so we can reduce demand into the court process. That could help. There is also the post-pandemic court system and recovery—

Some of it is out of your hands.

Yes, some of it is outside our control, but I think that could help. What else can we be doing in our own legal compliance operation to make sure that everything is prepared and ready? We want to get it right the first time, obviously, when cases are particularly long, such as when they have gone to appeal or whatever. Our general counsel is very focused on this. They work very closely with the compliance team. We know that, generally, large business timeliness has improved a lot. That is good. It is 16 months in the annual report, and we hope to keep going. We do not want to stop there. Obviously, we will get to a point where it will get harder to keep bringing it down. Generally, across the compliance operation on aged cases, that is down about 38% this year, so it is not just for large business; we want to do this for the whole compliance operation. For those cases that have to go into the litigation process because we have not been able to settle them any other way, hopefully we could reduce that demand a bit. Partly, the co-operative compliance model supports that, and then of course we will be working with our CPS colleagues to ensure that everything is right and can go through as efficiently as possible.

We will now move on to the extra investment that you are going to make in IT infrastructure.

I think it is too early to say. Jonathan might want to come in after me, because it really is a question of policy. What we are seeing is that the legislation has set quite a high bar, in terms of the level of avoidance behaviours that need to be demonstrated to apply special measures. You might want to consider setting a threshold that is not tied to the general anti-abuse rule or the disclosure of tax avoidance scheme behaviours. You might want to lower that threshold, but I think it is too early to say specifically. There is more work we need to do with our customers and stakeholders. Jonathan, is there anything you want to add?

I would just add that if you do change where you put the bar, the question is where you go on these decisions. I would draw a parallel with a recent change that we have made outside large business, in respect of agents. We have moved the bar for challenging agents from dishonest behaviour to behaviour that is leading to a loss of tax. That was quite controversial, because how you codify that in legislation is quite challenging. If you do go down that particular route, you want to find you have a clear bar. To go back to a point that we have talked about elsewhere, you also want certainty. We do not want to create uncertainty for business by putting down a line if people do not quite know which side they need to be on. We want to create certainty so that businesses know what they should do before they trigger the rules. Those are the sorts of issue that you would need to think about if you were going down that route.

I think we would want to review it, but again, I just want to be very clear that lowering the bar here is not an easy decision to make. If you were to want to change it because you decided that actually there were benefits—

I wonder whether it is worth drawing a parallel with uncertain tax treatment. About four years ago, we introduced uncertain tax treatment as a way of asking businesses to report interpretations that they had taken that disagreed with HMRC’s known position in guidance, to enable us to tackle the legal interpretation element of the tax gap more effectively. In the first three years after implementation, we had 30 notifications of uncertain tax treatment, with about £1 billion of tax advantage reported. Because the legal interpretation of the tax gap is significant, as we have talked about, we are currently consulting on whether to expand that. That is not to say, from the evaluation, that it was not working. It is working, as there have been 30 notifications and £1 billion of tax advantage, but there is potential for it to go further. There is a parallel here with special measures. Yes, we think that it has had a deterrent effect and, together with the operational activity, it has had an impact. However, it is right, in the light of the NAO recommendations, that we consider whether there are changes that would make it more effective.

If we lowered the bar, we would want to use it, and we would think that we could use it.

Again, we are only at the early stages of this work, which was referred to by the NAO. It is not something immediate; I think it would take a long time to consider it. We want to consult fully on this because, as we said, we also do not want to create uncertainty as a result of businesses not knowing where the bar is. It is something that we would approach very carefully, recognising that we do not want unintended consequences, and we do not want a chilling effect on businesses making sensible tax decisions. We would want to approach this very carefully, with proper consultation.

I think that any decisions on this are likely to be next year. Given where we are in the planning cycle, it will take some time to work through them. As I said, we want to consult informally and formally on all of this.

We may have started this year, but I am not sure whether we will have concluded it. I would expect a conclusion to be some time next year, given the complexity of the issue, and the fact that we need to be very careful to avoid unintended consequences. I am expecting us to work to a timetable next year.

I agree. As we have tried to set out, we think that it provides a good deterrent effect. It is a high bar, and quite reasonably there is a challenge to say, “Are you going to use this power?” If the bar were slightly different”—maybe lower, depending on how you define it—“would it be more feasible to use it effectively, not just as a deterrent?” I think that is what the team and the NAO quite reasonably call out. We need to respond to the NAO, and part of that will be to look at it again. If we were to change it, as Jonathan said, there would be a proper process of consultation, advice from Ministers and engagement with Parliament.

If I might interrupt at lightning speed, Chair, it has taken 10 years to reach the point at which you are saying that you think a review is a good idea, and it is going to take another year to conclude that review. Is this not a glacial way of working by HMRC when it comes to deploying your powers to stop egregious tax planning and aggressive tax avoidance? Is this not a really worryingly slow pace of change in the way HMRC works? It has taken 10 years of your never using it, and only now—a decade in—are you saying, “Right, we will spend another year reviewing whether we are going to change it, with the hope that we might then start using it.”

We have just concluded the NAO review, and we have set out the evidence on the effectiveness of this regime as a deterrent effect to reduce the tax gap and raise yield. We also recognise the points made about the burden on business and trying to be as proportionate as possible. The question of whether the bar is right has been raised. We think it is reasonable that it has been looked at. It is quite right that if we are to change it, we should not just do that overnight; there should be proper engagement with business, industry and others, and advice for Ministers and the scrutiny of Parliament. That can get going now, but it does not need to conclude in a few months; it should be done carefully, if we are going to change it. I think it is feasible that that could be done next year, but there is a balance to be struck between burden on business, proportionate powers and whether the deterrent effect is sufficient from the existing regime. Ministers may not want to change it at all, but we would provide advice based on the evidence.

We published a transformation road map shortly after the spending review, which set out the multi-year plan for achieving the benefits of that investment. I will give you the headline view, but I am happy to dig into it a bit more. If you take our first strategic objective, to close the tax gap, the additional tax revenues of £10 billion additional in ’29-30 are partly a function of additional capacity and partly a function of transformation. Penny can say a bit more about that, but it is everything from third-party data through to risking and case management, with more in between. Secondly, it is about improving customer service. We have talked about the telephony experience today. By the end of that period, we want to be digital-first, with 90% or more of our customer interactions through digital channels. It is currently just under 80%; it has gone up well this year. We are ahead of plan, but only marginally, and there is work to do to shift the channel. On our security maturity, our latest index is at 3.3; we want to get that to 4, so that at the end of this SR period, additional tax revenues and yields would be about £60 billion; we would be digital-first; and the organisation would be at that level that we would consider to be best practice for us as an organisation in security maturity. I am sure, however, that there will still always be more to do. Finally, on efficiency, we have committed to 15% efficiency against the SR baseline, much of which is of course also technology-enabled. To give you a sense of progress, we are hoping to provide an update and publish that for Parliament this year. I would describe the portfolio as, overall, amber. We are live in Making Tax Digital. We went live today with agent registration. We are just concluding our processes for onboarding our eCRM and our CCaaS systems. We have onboarded our hyperscaler to exit our legacy data centres so that we can be a cloud-first organisation. We have good indicators and are moving forward with the road map. We will continue to publish progress against that. We are clear on the outcomes that should be true when we get to the end of it.

I think the NAO Report talks about the progress we have made over the past few years. What this investment will enable us to do is make the next step change in the performance of our teams. We want to modernise our approach to data risking and case management, and we want to make sure that we are a data and intelligence-led operation that automates a huge amount of work.

I have a couple of examples that might be helpful. We have the senior accounting officer regime, which puts personal responsibility on a senior accounting officer for the accounting systems that enable them to get their tax right. At the moment, that is a pretty manual process. Senior accounting officers can submit their certificates or any qualifications via email or post. We are digitising that, and later this year we will go live with a digital service for senior accounting officers. That will enable us to reduce the manual handling for us and our customers, which will be more secure and easier to use and will improve customer experience. But from a risking perspective, it will enable us to look across the population, understand the risks in terms of underlying accounting systems, and enable our risk assessment to be more effective. That is a relatively minor, small-scale example, but it is a really important one in terms of the foundational elements for the large business customer. I mentioned the work we are doing on using AI in analysing large caches of email and other corporate documents when we are looking at international tax risks, for example. We are already doing that to a degree, but the new tools we are testing and the new operating model we are developing will be online later this year or early next year. That will help us to be more efficient and quicker at the work we are doing in that area. Across HMRC, we have also already deployed Microsoft Copilot and we are actively looking at use cases for Copilot agents, for example. There is more I could say on specifics, but it is starting to come to fruition.

Let me address that for the portfolio, and then I might invite Penny to say a bit. I think that all the portfolio for compliance transformation is integrated, but there is a very important opportunity in terms of our risking enablement and moving more of that capability upstream to get it right first time and data enabled. Penny can maybe give you more delivery confidence on that. In terms of the portfolio generally, my excom team reviews it every month: milestones, deliverables, benefit forecasts. Are we on track? If not, we look at RAG assessments. We will have programmes that might face a bump in road and there is something to do to get them back on track. Similarly, there may be others that are just on their way and under control. The overall portfolio is amber. We delivered the commitment for the ’24-25 benefit out-turn from the portfolio, so it is on track in terms of the outcomes, which is encouraging. We will put the latest into our annual report, and we will publish progress each year so you can see the out-turn of benefits as well. On level 1 milestones, I mentioned MTD: it was live on time and take-up is over 300,000, which is ahead of our central scenario. We are hitting the level 1 milestone against our top 20 programmes. The director of my transformation unit reviews those constantly, and where we find there is a challenge we obviously intervene quickly to try and keep those top programmes on track. Again, we are happy to keep the Committee updated, but Penny can maybe say a bit more about the compliance transformation.

I think that this transformation is highly ambitious, high risk and very complex, with a lot of dependencies. I currently judge it as amber; I think it is feasible. I need lots of support from anyone I can get my hands on to keep it on track—for example, to keep the scope as stable as possible and to try to avoid new scope, more shocks, disruptions or whatever. I made the point earlier around retention of talent and making sure we have the capabilities to execute it. We will quite regularly step through mini resets within the portfolio, when we need to do things where we can see a turning point and we need to keep the particular project on track. If I stand back and look at the whole of the transformation and those outcomes that I listed at the beginning, do I think it is feasible to get the 90% digital and the yield to 60, be secure at 4.0 maturity and deliver that intermediated digital-first enterprise? I do, but I cannot pretend that it is anything but ambitious. There is absolutely a whole enterprise focus on executing the transformation, so I am sure there will be bumps in the road as well.

I really appreciate that, because we want to create an empowered environment where our risk appetite is higher and we can execute with a higher velocity. We have a pioneering, advanced technology programme at the moment that is seeking to accelerate our adoption of advanced technology. Nicole made reference to our AI deployment, which is seeking to do the same thing. We have a set of accelerators in our organisation at the moment that are about optimising the operating environment to accelerate delivery. I mentioned that agent registration is live today and MTD went live last month on time, so we are pushing the team all the time to say that, despite what might have been okay in the past around delay, descoping or whatever, although we obviously need to keep things secure, we are determined to do everything we can to stay on track. The proof, I appreciate, will be in the pudding.

I would agree with JP; it is an incredibly ambitious programme. In compliance, we are looking to transform the way we work end to end, starting with the data, tools and capabilities we will have. We already use AI in a lot of our work, but we want to advance our tooling and capabilities to use data to identify risk sooner, to change the way that that can then be flowed down to caseworkers to support them with better tools and guidance, and to integrate the case management system so that their job is more straightforward and they can spend more time on the value add. We also have investment in our counter-criminal capabilities, so it is really end to end in all our work in compliance; I expect to see a lot of change. That said, it is quite early days.

You alluded to it there, so we will just touch briefly on the issues you may see now, and the possibilities to improve your activities, with regard to bulk data profiling and the utilisation of analytical tools, including AI. Where do you see that going, and what benefits do you think that can bring?

There are a few examples. I mentioned the onboarding of our new customer contact system and our new eCRM case management system. Those two projects are on track, and that onboarding process is under way. Both of those are market-leading software-as-a-service products, and they will have AI integrated for us to deploy at scale. In real terms, that means an agent picking up the phone to speak to you when you ring up about your tax case. In future, the system will summarise the call for the agent as it is happening and pull the next best action out from the guidance to support the customer service agent in ensuring that we resolve the inquiry once-and-done at the point of contact, so that there is first-call resolution. It will capture the record of the call as well, so that we can hopefully answer more calls per agent per day at a higher once-and-done rate, which would reduce the average speed of answer, given the capacity I have. If you fast-forward from that, in a self-serve, multi-channel environment you would want the digital services to be so intuitive that they meet need on the self-serve channel so that there is no need to call in the first place. We are making progress with that. I mentioned that digital customer interactions were around 75% or 76% last year, and that has increased to just under 80% now, so we are moving up in terms of enabling customers to get it right first time online and stay online. To get to 90% and achieve self-serve, CCaaS and eCRM are going to be two fundamental platforms that we will reuse across the enterprise so that we can then exploit the data. If you have told HMRC something once, we will then be able to use it multiple times, rather than asking for the same information again.

Is the aim to collect more tax or to provide better customer service?

Ultimately, the two things should be aligned. You are seeking to help customers, recognising that the vast majority are honest and want to get things right first time. You are seeking to make that as easy as possible, so that somebody on their phone can see, “Here’s my income, here’s my national insurance, here’s my PAYE and here’s whether I have any other savings income, dividends or whatever it might be. This is what I know about myself, and this is what HMRC knows about me.” We are able, through nudges and prompts, to keep you on track. For example, if the customer has told us something that does not look right, the system can say, “Are you sure you want to put that number in, because it doesn’t look within the mean?” So you can put the risking upstream into the self-serve channel, which helps to keep customers in that channel on track first time, rather than it ending up downstream in a compliance investigation that takes time, costs money and adds a burden to the business or the individual. We are trying to take the compliance model upstream and get it right first time, and that also supports yield and customer service.

Will there always be a person someone can speak to if they want to?

Absolutely. In fact, we are increasing the size of our additional support team. We are very committed to our voluntary community sector partnerships, and I am hopeful that we will do more of those as well. We are clear, in a digital-first operating model that is 90% customer digital interactions, that 10%, for whatever reason, will still need to get in touch with us through a different channel. As to the performance of that channel, single-digit average speed of answer would be optimal, in terms of higher satisfaction and higher trust. For those customers who do not want to speak to us on the phone—we recognise that that happens—we have an advocacy of voluntary sector partners we fund across the country to provide that very targeted additional support.

Moving on to productivity, our final questions are from Rupert Lowe.

Bad luck! I am afraid the Chairman gave me this question as well, so you have me to finish off with. One of the mild criticisms I picked up in the Report is that it is difficult to see how you can improve your productivity when you are not necessarily collecting the data in a way that allows you to do that. To pick out a couple of things, you have not recorded cases or you have inconsistently recorded them. In an era when data is everything, you have access to huge amounts of data. With the cases you are looking at, you could arguably log this data and use it much more effectively. Is there is a reason why that data is being collected only inconsistently or sporadically, however you want to look at it? I think that question is probably for Nicole, but it may be for you, Penny.

Perhaps I should pick it up, because it is a cross-cutting issue. We have a lot of legacy systems—systems for individual areas of taxation—that do not necessarily speak to one another. A lot of the investment that JP has described is about enabling us to join up our data in the first place and then to feed that into our risking capability. We have some incredibly able analysts, who do a very good job with the tools we have. We are uplifting the tools we have, but most importantly, yes, we need to knit that data together and get it into the right place in the right structure, so that it can be used in multiple ways—to help large business, but also to help right across the system. As JP was saying earlier, it can be used to help with things like nudging and prompting customers. It definitely should not all be about our investigations at the other end of the process; it is about helping customers to get it right first time.

Without wishing to labour the point, I wondered whether it was because of the complexity of your rulebook, which makes it difficult for you to work out what data you need to capture. I suspect that that might be part of it.

It is definitely a thing you hear in complex systems. I used to be permanent secretary of the Scottish Government, and you also hear it in the NHS. I used to run Jobcentre Plus, and you would hear that it is very hard to compare the productivity of one jobcentre to another. We do absolutely oversee a complex system, but we talk about productivity a lot. We talked about aged cases; they have reduced by 38% this year, which is great. There is more to do, but the timeliness of inquiry is coming down. We talk about things like yield per FTE. It is a simple input-output measure. It is obviously slightly crude, because it is hard to compare “large business” with “small business” with “wealthy”. But if you are running “wealthy”, it is still interesting to be able to look across your teams and to see yield per FTE, or hit rate, if you like. We are trying to bring together a cluster of indicators, but with a level of empowerment for different directorates to set their own stretch targets as they try to improve their performance. That might be because it is AI enabled. It might be about lifting capability or retention of talent. It could be about better systems. To be honest, it is all those things. Ultimately, at the end of the spending review, HMRC will be broadly the same size as it was at the beginning, but we expect our yield to be £10 billion more per year, and our efficiency 15% more. A lot of that will come through deployment of a whole cluster of change that improves our productivity.

Collecting the data correctly could help to reduce the burden we talked about earlier. You say that HMRC “is planning to agree engagement expectations” with large businesses “at the start of an intervention.” It can also “share an action plan” that sets out “how it expects to investigate the issue, including what data it needs and why.” Is that going to make data collection easier and make not collecting it less justifiable?

Yes. We have made good progress on that since the NAO was completing its fieldwork last autumn. We now have digital action plans in place for 97% of our cases, which has been good progress. In terms of productivity, that responds to the feedback that customers were not clear why we were asking for large volumes of data—that makes it really clear. It also makes clear the expected timeframes that we are all working to, for both our large business customers and our compliance colleagues, so that we can be really transparent and hopefully reduce those times even further.

Thank you for answering my questions. Unlike a lot of people, I genuinely think that if you make things easier for businesses, you will collect more tax. The more aggressive and opaque you are in everything you do, the more you will effectively close down Britain. I am a great believer in the Laffer curve, which you may or may not know about. I am a great believer in cutting businesses free. The easier you make it for them, the more tax you will collect. There is a danger at the moment that you are arguably flogging a dead horse. As you know, the private sector accounts for an increasingly small percentage of GDP. I suspect that our economy is turning down because a very large share of it is now driven by the state, which is raising wages and raising its spend every week. That is why the economy is stubbornly appearing to be strong, but if you look at your businesses in the private sector, they are not strong; they are suffering. The quote I gave you earlier needs to be borne in mind.

I do not think you have to answer that question, if it was a question.

I was thanking them, Clive. If everybody at HMRC sounded as credible as JP, we would not have a problem.

One final point. You had a long conversation with Lloyd Hatton about information and tax cases. I know it is very sensitive and that there is confidentiality, but we want to encourage people to come forward and raise concerns when they think fraud is going on. Whistleblowers are an important part of the system. If there is never any feedback to people in that situation—and I challenge the permanent secretary at the DWP on this as well—is there not a disincentive because people just feel that they report into a black hole and nothing ever happens? Can you do anything to address that concern?

I think we can, and I would be happy to come back to the Committee and give you a more detailed response in writing about what we already do and what more we want to do. We announced in the last Budget our high-value scheme, learning from the experience of the IRS in America and from global best practice. It is just getting going, and it is about working with people who give us information and ensuring that there is an incentive to work with us. Penny might have the details on the scheme if you want to hear about it, but we think it is an exciting opportunity. More generally, when somebody is in touch with us we have that duty of confidentiality, so when it leads to an investigation, sometimes it is difficult to engage and say, “This is what the outcome was.” I am conscious that I sign off many comms every week and they often say, “My constituent wrote and raised a concern of x. What has happened, please?” We are very happy to set out what we are already doing and what we could do to be more open in that regard.

That would be helpful. I thank all our witnesses for attending. An uncorrected transcript of the hearing will be published on the Committee’s website in the coming days. We will consider the evidence provided and produce a report in due course. Thank you very much again for attending this afternoon and for answering our questions.