19
Examination of witnesses
Dr Andy Summers, Bill Dodwell, Kevin Hart, Samantha O’Sullivan and Joshua Toovey
Q60 The Chair: Welcome to the second session today of the Finance Bill Sub-Committee. I am Lord Leigh of Hurley. Before we ask the witnesses questions, could you introduce yourselves and explain who you are and where you come from?
Joshua Toovey: Good evening. I am senior research and policy and adviser for IPSE, the Association of Independent Professionals and the Self-Employed. We are a membership body that exclusively represents the self-employed.
Bill Dodwell: Good evening. I am representing HMRC’s Administrative Burdens Advisory Board, which is an independent board chaired by Dame Teresa Graham. I am also the former tax director of the Office of Tax Simplification.
Dr Andy Summers: Hello. I am an associate professor at the London School of Economics Law School. I work using HMRC data accessed via the HMRC Datalab, which is a facility for academic researchers to use de-identified tax data for research purposes. I should say that I am speaking in an independent capacity and not for HMRC; the work is independent of HMRC and is not for the LSE either.
Samantha O’Sullivan: Hello. I am the policy lead for the Chartered Institute of Payroll Professionals, the CIPP. We are the UK’s professional body for payroll, pension and reward professionals. I head up the policy and research team, and we act as the voice of our members and the wider payroll professionals to address any concerns they may have.
Kevin Hart: Good evening. I am chief executive and chairperson of BASDA, the trade body for business software developers in this country. We specialise in finance, payroll, HR and pensions in particular, but I represent many others. HMRC is a major stakeholder between businesses, consumers and our members, and I have been working with it probably for longer than I care to recall. I have been very involved in PAYE and Making Tax Digital system and programme boards in a number of areas, and I work closely on a number of these matters, but I am here representing our members and more widely. I have worked very closely in payroll bureau, so hopefully embellishing my colleagues’ comments for the wider perspective as well.
The Chair: Thank you. If feels a bit like a “University Challenge” line-up. We have no points on offer, sadly, but we do have Lord Altrincham with the first question.
Q61 Lord Altrincham: The question is on the topic of employee hours that you are aware of. There have been mixed messages from HMRC concerning the purpose of collecting information about employees’ hours. What is your view of how valuable this data will be and the uses to which it could be put? Do you think HMRC should be more transparent about the purpose of collecting this information, and what would your view be if HMRC was to go on to disclose that information to other government departments?
The Chair: It is probably best to start with Mrs O’Sullivan.
Samantha O’Sullivan: The CIPP absolutely welcomes the wider policy perspective. Improving the data that HMRC collects can only be a good thing. An area of concern is that the approach needs to be considered extremely carefully, ensuring that the correct data is collected and submitted to HMRC and that a massive additional administrative burden is not put on to payroll teams, which are already extremely busy.
The data is invaluable. It is the basis of multiple calculations with regard to payroll administration, holiday pay, and determining whether employers are paying and adhering to national minimum wage regulations. If it is shared with other government departments or agencies, that could result in a more holistic and compliant approach, which again could only be a good thing.
Dr Andy Summers: I mainly want to emphasise the potential utility of systematically collecting employee hours data for policy development purposes, not only in HMRC but across government departments.
Just to set the scene for you, what we know currently about hours worked in the economy comes from one of two main sources: the Labour Force Survey, which is a household survey of about 0.2% of the population and has a very high non-response rate at the moment, so it is providing quite an unreliable picture of what people’s hours are doing in the economy at the moment; and the Annual Survey of Hours and Earnings, which is a 1% sample of HMRC’s PAYE records that contacts the employer to request information but still has a significant non-response.
I have with me a recent quote from Paul Johnson, director of the Institute for Fiscal Studies, who says that our labour market stats “have gone bonkers” recently, essentially since Covid; that there is a completely different charted path between these data sources, and we basically do not know what is going on in the economy in terms of hours worked. Another IFS economist said that we are basically flying blind on policy at the moment. I want to stress the importance of the usefulness of these data for essentially statistical analysis.
Going to your question about sharing data with other government departments, my own view is that there should be a lot more sharing of microdata—in other words, individual-level data—between government departments, obviously appropriately pseudonymised so that there is no risk to the privacy of the individuals. Even leaving that aside, if statistics were produced in HMRC on, for example, changing regional pictures in hours or how people’s hours have responded to changes in the tax system, that sort of evaluation would be extremely useful for other government departments, even if those departments were not receiving the raw data themselves. We could get a great policy evaluation resource from this.
Bill Dodwell: As a general point, we would certainly support HMRC gathering more data. The principle in the UK, of course, is that roughly 34 million people pay income tax, but only about 12 million people send in a tax return, which is completely different from the situation in the United States, for example. That means that, if our system is to be effective overall, we need our tax authority to collect lots of data from third parties—obviously employers, in this case, but other sources too—that will help them to get individuals’ tax affairs right and to provide a better service to people more generally. So, yes, we very much support the idea.
The key principle, as I think others have said, is to make sure that data is collected in a manner that is convenient for the supplier of that data. If you are asking for data in connection with a return, that is a much more convenient time to do it than if you come along later and put in a detailed data request. So it is better to do it in that sort of way.
Joshua Toovey: Collecting data from the self-employed on the hours worked is notoriously difficult. Many contractors do not keep a record of the hours they work, and working out who would be responsible for submitting this data to HMRC could be problematic. Would it be the client who is then responsible for submitting the hours worked by the contractors? They may not know themselves. So we would be quite keen if the self-employed were excluded from this proposal.
Kevin Hart: It is probably a little contentious, to be honest, because we struggled to understand the problem statement. It was not really clear what issue this is looking to resolve. It certainly did not seem to have a fiscal benefit, while the costs, which we will perhaps come to in a moment, are not insignificant. That raises the question: if it is for statistical purposes, is this the best way of collecting it?
To your question about other departments and agencies using the data, one has to be careful about unintended consequences. If you are an employer, you are collecting this data and passing it on. Hypothetically, if that were to result in an impact on someone’s universal credit, they would then challenge that by saying, “How did you get that information?” and would get the reply, “It was from your employer”, so there could be quite some friction between employees and employers in that regard.
What is the problem that it is looking to fix, and what is the benefit? Payroll teams are already incredibly busy—I agree with a number of my colleagues here—particularly the bookkeepers, accountants and bureaux that are collecting from multiple employers. Historically, it has been hard to get the right data at the right time. Quite often, data that is collected is late or delayed. It is not unusual for people to have left the business, and the poor payroll team are some of the last people to find out. It is notoriously difficult to collect data in the first place just to get the payroll done, and another piece of data to collect could be a significant burden on them, particularly, as Joshua was saying, for those like bureaux that are operating on behalf of third parties, whether microbusinesses or small businesses. We therefore question the validity of this measure, because it could have unintended consequences, and it will be costly to implement and deliver.
Q62 Baroness Valentine: I want to follow the point about collecting for statistical purposes. I have two questions, primarily for Dr Summers. First, if this relates to broad governance statistics, is HMRC the right department to collect the information? I imagine it has something to do with productivity, but whatever.
Secondly, regarding the accuracy of the data, contracted hours are one thing, but how many hours people actually work are another. We have heard about the self-employed, but in many different industries people are not working their precise contracted hours, so I would have thought that there would be a problem with the accuracy of the data as well. Do you have any observations on either of those points?
Dr Andy Summers: Yes, I can speak to both of those. Taking the second point first, I noted in quite a lot of the written responses that there was concern about collecting information on actual hours. I had anticipated that this would be a collection of the contractual hours. For most statistical purposes, that would suffice; it would certainly strike the right balance between usefulness for policy purposes and the burden of collection. I can see the force in the concerns about how you estimate someone’s actual hours and employers needing to ask questions of their employees that they are not currently asking, whereas with contractual hours—perhaps Kevin can speak to this—I would have thought that information was more readily available already to employers and so would be more straightforward for them to provide.
In fact, under ASHE, the data source that I mentioned earlier, contractual hours are already collected. It is just that it is for a very small sample, which means that we could do any of the regional analysis that we might want to do for various productivity or levelling-up agendas. We cannot break it down in a fine-grained way into top earners, women with young children and so on, which is the benefit that we get from making this a requirement for all employers.
To your first question, yes, HMRC is definitely the right government department to collect that information, for the simple reason that it already interacts with employers through the monthly RTI report. The most natural way to execute this would be to integrate it into that system, with contractual hours reported via RTI. No other government departments have that regularity of interaction with employers, so it should definitely be HMRC.
Bill Dodwell: I agree. I will simply add that the Government’s response to the consultation was that they would just be collecting contractual hours, plus actual hours where people worked on an hourly basis.
Kevin Hart: That was not clear. In retail, for example, there is a huge variety of hours worked, and it varies almost day by day and week by week. Again, this comes back to the question of need. There are some sectors where that information is just not available, to the best of my understanding. In agriculture, for example, quite often the employer may instruct set payments, but there are not necessarily hours associated with them. One size definitely does not fit all here, and one needs to be mindful of that. It would be useful to have clarity. If it is just contracted hours, that is quite a different picture.
Samantha O’Sullivan: By being transparent you will get earlier buy-in from employers when it comes to sourcing this additional data. It is taken as a given that this data is recorded somewhere, but in most cases it will not be. If it is contractual hours, what happens when someone is on statutory leave for sickness or maternity? If someone is doing a phased return to work, a number of hours are being worked, but that combines with taking time off sick. The CIPP would push for extremely clear guidance and for setting out the intention behind the requirement for the data, in order to get buy-in in the first instance.
The Chair: The question has been asked, “Why are we collecting this?” Do you think the Government are seeking the information to see who is breaching minimum wage legislation or longevity-of-hours work? We certainly know of people in the City who regularly work 70 hours or more. Is that what is behind this?
Bill Dodwell: But they are not close to minimum wage, are they?
The Chair: No. Mr Hart was saying that he does not know why. Maybe Dr Summers knows why. Can you unravel that a bit?
Dr Andy Summers: I think there is some force in the objection about the level of clarity about the objectives here. HMRC is in a difficult position. Historically, it can only collect data that is relevant to the administration of the current tax situation, but, in my view, a lot of the best justifications for collecting this employee-hours data goes to policy evaluation that would benefit lots of other government departments—and, in the long run, all of us—through a better understanding of how the economy works. Frankly, though, it is difficult for HMRC to make that case. Maybe it would like to do so, but it has not felt able to lay cards on the table that that is one of the best reasons for doing this.
There are operational reasons. You can look to enforcement of national minimum wage legislation, but, to your point about City workers, if you are earning over £100,000, there is no way you can be in breach of minimum wage legislation, however many hours you work, so that will not justify the collection of data across the board. HMRC ought to be a bit more on the front foot about the usefulness of collecting this data for policy purposes, because a lot of what I am hearing is sympathy for the objective but, so far, not for the way it has been couched.
Lord Altrincham: Is the challenge not that the desire of data for analysis is infinite, and that people like you would love to get a lot of data for a lot of research? That is well understood. However, there are costs in assembling that data which fall on the private sector, and there could be economic consequences from collecting it. The very action of collecting the data might lead an employer to try to pull back the number of hours that people are working, for example. If they were to have longer hours but they are contracted hours and they do not want to be in breach of any kind of regulations or HMRC problems, they will tend to want to pull back.
We as a country need people to work more. We need work to happen, not to not happen. So if we set up systems that tend to reduce the amount of work that goes on, we all bear a large economic cost, even larger than the friction costs of collecting this kind of data. So this is not a cost-free exercise and, if it is somewhat touristic, that would be a shame, because there has to be a benefit, as there will be costs in doing this.
Dr Andy Summers: I certainly agree that there is a trade-off here, although I suggest that the information problem goes in the other direction: the direct costs of collecting these data are relatively certain. Individual payroll operators and so on can quantify those costs and there is an estimate—I know it is disputed—so we can reasonably reliably put a figure on the cost side.
I admit that the benefit side is a lot more difficult or diffuse, but we are about to have the Autumn Statement on Wednesday, and the Chancellor will stand up and make some tax policy announcements. I obviously do not know what they are, but they will hopefully be informed by evidence and, if its quality is poor, the policy is likely to be poor. In the tax context, these are often policies that are costing or raising billions of pounds. In my view, in that context, the cost of collection, although it is concrete, is worth the benefit from making a better policy.
Lord Altrincham: Can you specifically address the topic that the very action of collecting changes behaviour?
Bill Dodwell: If they are collecting only contractual hours—that is what the Government’s response clearly said—they will collect actual hours for people who are paid by the hour, but contractual hours for those who may indeed work 70 or 100 hours a week, although their contract says 37 and a half, or whatever it is. So they collect just the contractual hours, which would not influence behaviour.
Kevin Hart: I will bring in the perspective that we may have a particular vision of City traders, but there are many salaried and contracted staff whose hours are just not measured, so it is almost a meaningless exercise. Some of the less contracted hours are perhaps where one would think that, from a policy point of view, the data would be more beneficial and therefore more meaningful.
For a minimum wage-type piece, for example, we are not convinced that this is the best method to achieve this. If I were to paraphrase, I would say that it feels like using a sledgehammer to crack a nut in this context. If it is contractual, government may be missing a trick in relation to the data and the value of that data. As I say, much of the data will not be valuable: you will get an HR system’s lifting of what was in an employee’s offer letter, letter of engagement or letter of employment, which will have the number of contracted hours, but the actual number of hours they work will not be tracked and therefore not reported in many cases. So we need to be mindful and have our eyes wide open to this.
The Chair: I can see that that would be meaningless. Given your answer, I am not sure I phrased my question right, so I will rephrase it: is this an attempt to get the data to target individual firms and employees, or is it purely to determine policy?
Dr Andy Summers: I cannot speak directly for HMRC on that. I am saying that there is a strong case to be made for collecting these data for what I would call policy evaluation purposes. “Statistical purposes” sounds rather academic and abstruse, given the concrete costs involved. I would say that this is about the capacity to design better policy for the benefit of all taxpayers and citizens. It is fair to say that HMRC has not made the case in those terms, and I am here to make it for it. I hope it takes it up, but you will have to ask it what its plans are for individual compliance measures. I am not sure.
Bill Dodwell: I doubt whether the data will actually help detailed minimum wage compliance. I think it will probably flag up that the information was already there where HMRC needed to conduct an investigation.
Q63 Lord Roborough: I declare an interest as a small business owner in the agricultural sector, going back to Mr Hart’s point.
I had a general question about how onerous this will be for employers, but I think we have addressed that. However, I have a specific question for Mr Hart. What systems changes are likely to be necessary to enable employers to supply data on hours? What further information will be needed in order to develop these, and what is the likely timescale from the point at which the detail of the proposal is known?
Allied to that, what sort of coverage of businesses is there by software that may be able to report this data to HMRC? How many people are doing things on an envelope with a pencil and may need to buy a system to do this?
Kevin Hart: With our circa 1.9 million to 2 million PAYE employers, if one were to take the UK as a triangle split into three tiers, there would be a distorted pyramid. The micro and very small businesses would be at the bottom, the small to medium businesses would be in the middle, and the larger businesses and corporates would be at the top. The latter would have full-blown systems that do all their finance, payroll and HR, and they would be able to integrate that, but not at insignificant cost; just to buy the licence for the software alone for those systems is in the £1 million-plus range, so you can imagine that the change cost is not insignificant.
If there is a legislative change, most older software provides that as part of its package, because it complies with payroll and associated law. However, the costs come in when you want to install that software, test it in parallel with your live service, and make sure that everything is performing. Although that is one small change, they must make sure that it does not break anything else. A number of these larger businesses will have strong nuances to their operation, whether they be multinational, UK-only or multi-industry types. So it is a significant piece of work for them just to implement a new piece of software, and they tend to have set release cycles in the year because of that. You can imagine that the consultants are charging anything between £1,200 and £1,500 a day just to go in and work on some of these systems. I could not give you a broad-brush figure for how many days’ worth of work there would be, but it would not be insignificant.
Those in the middle tier quite often have different systems. They may have an HR system, a payroll system and an accounting system. They may talk to each other in a simple fashion, but the HR data is stored in the HR system, so they would have to onward develop the linkage between the HR and the payroll systems. That has a cost, which may be bespoke for quite a lot of individual providers.
Down at the bottom, there is a complete mix. You also have those with pieces of paper and simple spreadsheets. HMRC has a simple old payroll product called the basic payrolls toolkit. The last time I looked, something like 400,000 employers were using it. It is a simple product, and there is no calculation at all for how many hours have been worked; you literally plug in the numbers for who you want to pay. Coming to your very good question, that information is on paper or, at best, on a spreadsheet.
Alongside that, there are a lot of simple low-cost software solutions, but they are typically payroll, which quite often does not have the number of hours worked in a number of cases; that will be stored somewhere else. The software acts more on an instruction—for example, “Pay employee 1, who did this number of standard hours and this number of overtime hours”. So these products will need to be upgraded. Again, because it could be a legislative update if it is implemented in this fashion, the software is likely not to have additional cost, but they still have to implement it, test it, train on it and make sure that nothing is broken.
At that lower end, you have a lot of bookkeepers, accountants, and payroll bureau processing payroll on behalf of all these businesses. We did some very broad-brush numbers, so please do not hold me to this, but we would say that nearly two-thirds of the 1.9 million to 2 million PAYE offers go via bookkeepers and accountants, so they basically contract out the responsibility for doing the payroll. Our broad-brush numbers suggested that probably close to 1 million actual end-users of software are having to do this piece of work. HMRC’s impact analysis quoted £35 million to implement this, which basically means that they average £35 per business to implement it. Hopefully, given what I have just shared with you, it is not unreasonable to draw the conclusion that £35 is nowhere near enough for this.
Then there is a significant ongoing cost of trying to get this data on every pay period, and there are still a lot of weekly, bi-weekly, four-weekly as well as monthly staff. So there are significant operational costs. As I mentioned earlier, a lot of that data flows in much later, so one has to be mindful of the fact that this might be out of date or wrong. So one has to be sure that one has robust data with which to form any policy decisions.
I hope I have answered all those questions, but please challenge me back.
Lord Roborough: Thank you. That was very thorough. Quickly, on the timescale, how much notice do you think the industry needs to adopt this?
Kevin Hart: To your great question earlier, it is the software industry and all the accountants and bookkeepers who all have to be familiar with all this and implement it. There is certainly insufficient time to do this for April 2024. Most of the software updates have already been compiled and are just waiting for the Autumn Statement in case there is any change in thresholds, percentages et cetera. They start to implement these solutions in January, so there is no time to achieve this for the coming April.
It is a difficult question to answer, but it could be six to nine months. It depends on the complexity of the customers, but they probably need to be providing information and technical data on it at the beginning of next year to stand a good chance of hitting April 2025.
Lord Altrincham: Kevin, you took us to the 1 million providers not costing £35 a head. What is your estimate of the cost?
Kevin Hart: Forgive me, but that is a really tough question to answer. I explained the three areas of the pyramid. It probably would not be unreasonable to say maybe £50 per bureau per customer at the low end. At the top end, you could be talking thousands or tens of thousands of pounds to implement and test it.
Lord Altrincham: You made the assumption of £1,000 across 1 million. Is that a crazy estimate?
Kevin Hart: That would probably be excessive, but several hundred pounds might not be unreasonable. That is not research-based; that is just instinct and experience.
Lord Altrincham: Thank you.
Q64 Lord Roborough: Thank you, Mr Hart. Mrs O’Sullivan, I have a specific question for you as well. What about employers who outsource their payroll? What do the new requirements mean for both the employer and the business providing the payroll service?
Samantha O’Sullivan: Mr Hart explained that really well. Just to bolster his argument in response to the previous question, if these reporting requirements were made mandatory, and if, therefore, the number of hours were not reported on the submissions made to HMRC, meaning that your payroll could not go through on time, would bookkeepers, and employers, make a wise estimate and just guess the number of hours to ensure that the payroll could be processed on time if they did not have the hours readily available to them? That would undo the whole purpose of what we are trying to achieve.
The CIPP is a critical friend of HMRC, and we have run some think tanks for it recently specifically with outsourced payroll professionals. Some of our key findings there were that service providers simply do not receive the total number of hours worked. They may say, “Pay Samantha £1,000 this month”, but there are no hours there, so again you are essentially plucking a figure out of thin air and putting it on the real-time information submission just to get your payroll processed. Especially if they wanted to see this for individuals paid weekly or hourly, they would really need to look at that data. That is where it would link into national minimum wage compliance if we were going down that route. Our members were also concerned that their clients would not be very forthcoming with the information unless they knew specifically what this data was going to be used for, which links back to our first point.
On the impact on service providers, they may need to change their data-gathering processes. It is not just about what the system needs; it is the starting process. You sign up for a client—“These are your KPIs. This is the information we need off you”. That could take time to develop and implement, alongside the financial implications. So we would stress again and again that sufficient time needs to be given to implement this and to ensure that it is implemented successfully.
Regarding businesses, once again—I know I am repeating myself, but it is a really valid point—people need to know why this information is being requested. They may need more investment. They might not want to use a legacy piece of payroll software and might want to go out to tender and look at new software providers. That is time and cost which the business then obviously needs to pick up.
Support and guidance really need to be brought forward here to explain the new requirements and to get the employers’ buy-in. However, to summarise, again, the data is not held right now. There are lots of individuals out there who get a fixed amount of pay, per pay period, regardless of the hours worked. Again, as Mr Hart stressed earlier, for those on a zero-hours contract or who work flexibly, you are more likely to have the hours they have worked to hand versus those of someone who is on a full-time contract working 35 hours a week but who more than likely works about 50 hours but those are hours are not recorded anywhere.
The Chair: Thank you.
Q65 Lord Palmer of Childs Hill: It does seem from those comments that a great number of businesses will take rough estimates because they have to put something on paper or on the computer. If there are no hours but there is a fixed sum, they know that they have to put some hours in because the software demands it. So would you agree that there will be a lot of imagined estimates in here, not for any illegal purposes but because people are just trying to fill in the information?
Based on that, and on what Mr Hart was saying about the costs of providing the information on employees’ hours, HMRC, despite what various of you have said, estimated the cost to be £35 million. It that, in your view, completely unrealistic? That seems to have been said. If that is the one-off cost, is there an estimate of the ongoing cost? It might be a lovely academic exercise, but will business and industry gain anything from this?
Dr Andy Summers: Can I speak to that?
Lord Palmer of Childs Hill: Yes. You are the academic.
Dr Andy Summers: Concretely, for example, probably the headline measure in the last Budget was the changes to nursery funding. The whole policy of that very expensive measure was to increase return to work by mothers and fathers. How are we going to evaluate that policy? At the moment, we do not know whether a mother who is in work is working full time on a very low salary or part time on the same wage that she was on before. It is not possible to make a sensible evaluation of that kind of billion-pound policy if, as the director of the IFS said, you are flying in the dark with your statistics.
I take the point about the challenges with these data—if they are collected—not being perfect. That is the nature of almost all data analysis. However, what is currently being worked with is a survey where the Office for National Statistics cold calls someone and says, “Would you like to answer some questions?” Most people just say, “No thanks, I haven’t got time”. It then takes those results and multiplies them by 300 to get the supposedly representative population-level estimate, because it is a 0.3% sample of the population[1].
So the information we use at the moment is not suitable for making policy, and Chancellors are guessing on billion-pound measures. I accept all of what has been said about the costs—I am obviously not competent to gainsay those—but I emphasise again the very real benefits of collecting proper data or, to put it another way, the total inadequacy of what we collect at the moment.
Kevin Hart: Building on that, and coming back to the question, it feels like we are trying to get a square peg into what we think is a round hole but, because it is completely dark, we are not sure what the shape is. I have mentioned this to many, even the executive at HMRC, on many occasions, and I continue to say, “Please share your problem statement with us: namely, where’s your pain point, and what are your policy needs? Let’s work together”. I ask whether we in the software industry can help—it may be not us but others, like those in the payroll professional industry, but we count as software—so that we can work together and explore the solutions that we may be able to work collectively on to achieve the objective, as opposed to us effectively not understanding the high-level goal and questioning how achievable the objective is. This feels like the starting point.
I completely understand the process we are running here, because this is obviously a new amendment to the Finance Bill, but it does feel a little like we are walking in reverse here. We completely believe in supporting policy development, but the question is: is there a better way of achieving it?
Q66 Baroness Valentine: What guidance do you feel is necessary to help owner-managers and the self-employed comply with the requirements that apply to them?
Bill Dodwell: Is that the disclosure of start and close dates?
The Chair: It is also dividends.
Bill Dodwell: Separately, it is where you earn company dividends. The start date will be easy to define for some self-employed individuals but less clear for others. Something might start off as a hobby and then run up against the trading allowance—less than £1,000—and still not be disclosed. Only after that would it end up as taxable income and finally appear on a tax return. Helping people to understand what the relevant date to disclose is will be useful, and guidance will be needed to support that, but I see no reason why HMRC should not be able to do that. I would have thought that the closure date is much easier: people are much clearer about when they have stopped doing something, so that would not be too hard.
On the dividends point, that strikes me as relatively straightforward. For close companies—controlled by five or fewer people, together with their associates—it is a reasonably straightforwardly defined class of people, so getting that disclosure does not strike me as hard at all. Provided that it is done on the tax return, which I am sure it will be—in fact, both of these will be—that should be relatively straightforward to do.
Joshua Toovey: I agree with Mr Dodwell. We do not believe it will be particularly onerous for the self-employed to comply with these new reporting requirements. However, HMRC needs to be clear about what it intends to do with this data. We are not entirely convinced that it will use it for future pandemic preparation, as the consultation outlined, and the self-employed will naturally be quite nervous that HMRC might have different intentions for this data going forward, given falling levels of trust in HMRC and wider government tax policy in recent years. For us, transparency from HMRC here is the most important thing.
The Chair: We wondered why HMRC wants this in this way, given that dividends have to be declared on personal tax returns anyway, and the Economic Crime and Corporate Transparency Act requires iXBRL-format company accounts that show dividend payments, so why is there this extra information?
Bill Dodwell: It is because they are not directly linked into the tax system in that sense, which is the problem with trying to take Companies House data; it is just not good enough, even if it is supplied under iXBRL.
Dr Andy Summers: I can speak in a bit more detail on that problem. I agree that, wherever possible, HMRC should use data that is already collected for other purposes, and link it so that it does not have to request more information on a separate occasion. The problem is that, across our systems of government, we do not have consistent identifiers for individuals, so it is currently not straightforward to link up Companies House records.
Unfortunately, in my view, there was a major missed opportunity in the economic crime Act. On the one-off shareholder register, which would provide a percentage shareholding for all shareholders, including close company directors, you would think—going broader—that that would be a fantastic opportunity for HMRC to link that to the dividends that are already being reported, and we would not need any more. Unfortunately, the only thing that is reported on that shareholder register is name and address; not even date of birth is reported. So, yes, you can link that to tax records at scale if a tax inspector is trying to do it case by case, but for compliance purposes, where you are trying to analyse risk across the whole population, that is not straightforward. I agree that more could be done to make data better linked, which would ultimately reduce the reporting burden.
I have one concern about how HMRC proposes to collect the close company dividend information. From the consultation, it was suggested that this would be via a change to the SA102 form, which is the form for employees and directors, but any close company shareholder who is receiving dividends but is not a director—or is one, but is not remunerated as one—is currently not filling out SA102. That means that the information HMRC would get from this measure is incomplete, which is a problem in itself, but also a problem because it will not then be possible for HMRC to reconcile these reports at the company level. I would have thought that that is the main compliance purpose of doing this: you want to see whether there are people who are reporting the close company dividends that they should. But if you cannot cross-check what is reported against the amount of dividends shown in the company accounts as having being distributed, you do not have the major benefit of the policy.
So, in my view, there ought to be reporting by close company directors of the company numbers for the companies that they are close company directors of. All the rest of the information could then be obtained from that link. I would like to hear others’ comments, but I would have thought that, if you are a close company director, it is not too onerous to find out your company number. You can look that up if you do not have it on your documentation. That strikes me as a better option than getting people to add things on to SA102.
The Chair: So will it be on not a corporate tax return, but on the SA102.
Dr Andy Summers: The proposal I read is that it will be on the individual tax return—not in the section for reporting dividends, which is SA100 and SA101, but SA102, which is for directors. That is fine if you are a close company shareholder and director, but it does not work if you are not a director, or if you are an unremunerated director.
The Chair: Again, is the purpose of this to determine policy, or is it to discover directors who are not declaring their dividends on their tax return?
Dr Andy Summers: Again, I think HMRC would have to answer that. It could be more on the front foot about the compliance issue here. We know from the tax gap statistics that self-assessment among owner-managers is a large share of the tax gap. Hardly any of the tax gap comes from PAYE; it comes mostly on the individual side from this group. Obviously, most people are completely honest about their tax affairs, but this is an area where compliance is a major issue.
At the moment, the problem is that you cannot distinguish between dividends where the individual controls the distribution themselves and passive investments. You do not know whether the dividend is coming from shares held in an FTSE 100 company or from a company where you are the sole owner and director, and you decide how much is distributed when. That seems like an obvious area where you could improve compliance significantly if you had more information.
Q67 The Chair: Lord Palmer and I probably remember corporation tax, which would have solved this, but let us not go there. On the general point, the detail of the reporting requirement will not be known until later on. What consultation do you consider necessary, and how realistic is all this in respect of a start date for implementation in 2025?
Bill Dodwell: I would have thought it was perfectly achievable by 2025, because that gives us 15 months’ notice. Leaving aside the employee hours that we have just covered, for the other measures it is simple to understand what you are trying to do. But you will need HMRC to release the forms in plenty of time, because that will need to be incorporated into software. Kevin will explain better than me the update cycle for all that.
Kevin Hart: I agree with those points. We have a few things to consider, particularly on the dividend side. There are more complex areas. A number of my members provide software that is very functional across these areas but, as was mentioned earlier, there are different classes with different dividend entitlements. Again, going back to why we need to do this, we need to be clear about that and work it through.
Mr Dodwell was quite right from a time point of view. Even if the details were available early next year, my concern is that it may still take some time to work through the scenarios to make sure that we are able to capture all that information. One needs to be clear on the detail of how would need to operate. There will be a cost to doing that, exactly as I mentioned before on deploying this. There will need to be awareness and understanding, particularly by accountants and bookkeepers, of how they need to operate this and what it means if there are any other consequences. It is not just an up-front development piece; there is some training and familiarisation involved. In all these areas, as Mrs O’Sullivan mentioned, there needs to be a transparent government briefing on these points.
HMRC already knows the start date, so I am puzzled as to why it would be asking us to provide it. That should be a piece of information that it pre-populates in the interactions with our software. It could perhaps say that someone in the most recent financial year would have two if they started in that period, but HMRC already has that information, so it is puzzling why it is asking for it.
On the question of Covid, small business relief and so on, HMRC is also championing Making Tax Digital for income tax and self-assessment. The benefit statement is that, when it is introduced, it should give an indicative quarterly cycle of businesses’ income and expenditure, and therefore profit; that is an indicative figure, not a categorical one. That is deemed to be hugely beneficial, rather than waiting for potentially up to 18 months for a self-assessment filing.
To Mr Dodwell’s point about April 2025, one needs to be mindful that in these areas of self-assessment across dividends and dates collation, it is all hands to the pump in delivering Making Tax Digital for income tax and self-assessment. Some significant changes are likely to be announced in the Autumn Statement, which will mean a wholesale change to a lot of software. We are already on a tight timeline to make that available for public beta in April 2025 and to go live in its entirety in April 2026. To be candid, we would not welcome even more obligations to incorporate here, unless—dare I say it?—there was a strong benefit statement associated with it.
Samantha O’Sullivan: I would like to come in on employee hours. Obviously, the CIPP is here to represent employed individuals. If this were to be implemented, we would like to see some pilots run. I believe that the challenge is more in an outsourced environment. Between us, we could probably come up with a list of recommended companies that you could use for legislative pilots. Lastly, I agree, and I suggest no earlier than April 2025, in line with the new tax year.
Q68 Lord Altrincham: This is all the same topic. Are there less burdensome ways in which the relevant data could be obtained? Could HMRC get any of this data from existing sources, for example?
Dr Andy Summers: The answer is yes, via the short step of making it possible to link individuals’ Companies House records and HMRC records. Many people would be astounded at the idea that HMRC cannot straightforwardly do that already, but if you work with HMRC data, you find that, sad to say, a lot of things that you think ought to be straightforward data links turn out to be anything but.
That seems to be a big missed opportunity to streamline the way these data are used. The Companies House record contains information about shareholder percentages from the one-off shareholder register, and you have information in the accounts about distribution. If we could link those to tax records, we would not need the taxpayer to self-report the amount of dividends because HMRC would already know it. In Scandinavia, things already work like that across government departments. There may be a short-term imperative to get this information, but I would do that by creating links to existing data rather than getting taxpayers to do another self-report of something that is already held elsewhere.
Bill Dodwell: That would require that each individual had a unique identifier. Simply relying on name and address will never be good enough to link something in one department to something in a different one. We in the UK do not have a sufficient system of unique identifiers for tax for individuals. A high proportion have a national insurance number, but not everyone in the taxpaying population does. Another set of people have a unique tax reference, a 10-digit number, but millions of people do not have one of those either. If we are going to find alternative and better ways of doing this, as Andy has outlined, we need to make some progress on the question of unique identifiers. There was a separate government consultation on third-party data that included part of that, but it has not yet been taken forward to legislation.
Lord Altrincham: From a statistical point of view, if there were a transitional period where it was voluntary, you might still be able to collect a lot of statistical data, although it would not be complete. Unique tax identifiers and those kinds of datasets are massive, and obviously that would be a huge exercise, but as we look for cheaper ways to achieve a good statistical outcome, maybe some organisations already gather this data and have it on their systems, more or less, and would be happy to share it voluntarily for the time being.
Dr Andy Summers: If a provision had been added to the economic crime Bill regarding the one-off shareholder register that said, “In addition to name and address you need a taxpayer reference”, and every shareholder of a UK company needed to be allocated a UTR if they did not have one already, that would be that. It is not difficult, and we missed an opportunity in the last big piece of legislation on this.
On voluntary reporting, it is worth taking baby steps towards the outcome that you want. The problem is that, any time something is voluntary, for it to be at all useful for statistical purposes you need to understand the bias that results from the people who choose not to provide the information; otherwise, what you get from that is not a statistically representative sample.
To go back to the point about agricultural workers, there are steps that you can take statistically to account for that kind of bias if that information is not reported—that is tractable—but if the question is, “Do you want to report or do you not?”, it is hard to understand the reasons among the population who do not want to, and to make that representative. So I would be cautious about using any information that is on a voluntary basis for statistical purposes. For compliance, it is even more problematic.
Bill Dodwell: I suspect that the dividends point has a big compliance emphasis. Also, the Government have reasonably made the point that if they had only had more data, they could have awarded owner-managers of companies support money under the Covid schemes, but clearly they could not do that.
Q69 Lord Roborough: I have a question specifically for Mr Hart again. On Wednesday, the Government could announce that they were going ahead with the merger of R&D relief from April 2024. You have talked quite a bit about product cycles and how much time they require in software. Is that enough time for the software developers to make the necessary changes to tax software? How much time would they need in reality?
Kevin Hart: I can be very brief with my first answer, and it may not surprise you. The answer is no. As I mentioned earlier, a lot of the software is going through various cycles and maturing in those development cycles. It takes time to get correct specifications from HMRC along with the provision of all the necessary coding requirements and the test systems. The software developers then have to factor that into their road map when they are developing different products and feature sets. It is very much our view that that is an artificial date and it is just too tight—particularly for those who, as I mentioned earlier, are making Making Tax Digital products. It does not feel meaningful and achievable. April 2025 certainly does, but again on the basis that sufficient information is declared very early next year to allow them to analyse, assess and develop accordingly, particularly since my understanding is that this new R&D system is likely to be a hybrid of two, so it is not as if we are just going to tweak a few areas; it could mean quite significant rewrites in that regard.
Bill Dodwell: It is worth adding that R&D is applicable only to companies, while Making Tax Digital is applicable only to individuals at this stage. So those are slightly different products.
Kevin Hart: That was more from the point of view of the workload on the companies that are producing these solutions.
Bill Dodwell: Developers will certainly be busy generally, yes.
Kevin Hart: You make a very good point, but if you set it aside and consider those who are not doing this, it would still create an artificial priority that we believe would be damaging to the mainstay of those services for HMRC.
Q70 Lord Palmer of Childs Hill: We can change the subject a bit, which may be a relief. I would like your views on the draft Finance Bill proposals to tackle the promoters of tax avoidance. Do you think they will be an effective deterrent?
Bill Dodwell: First, it is important that we have as many deterrents as possible against promoters of quite aggressive tax avoidance. As you know, there is a stop notice system—in place since 2021, I think—whereby HMRC may issue a notice to a promoter that they may not continue marketing a particular scheme, and they need to tell the previous customers that it is subject to a stop notice from HMRC. That is targeted at promoters of mass-marketed schemes; it is not necessarily someone who is doing any form of bespoke, tailored tax planning that is specific to an individual, company or business generally.
Stop notices are really important as part of that, because trying to shut down the market for aggressive tax avoidance is a good thing for all of us. It is really good for the taxpayers involved, the individuals, because it means that they do not get caught up with all the unpleasantness and complexity of a very detailed HMRC inquiry, which is terrible for everyone. I am sure you have had evidence about some of the tax avoidance problems in the past. That is just a challenge for the system. It is a challenge for HMRC to deal with, but it is really unpleasant for the individuals involved, so trying to shut this stuff down is really important.
Fourteen stop notices have been issued so far. One is no longer in effect, so 13 are still in effect, and there is a public page listing all of them. They are all about so-called loan charge schemes whereby an individual provides labour services, and the intermediary, the promoter of the scheme, pays them the national minimum wage or the national living wage and then makes some form of loan or other advance, trying to claim that it is tax free. It is not tax free; it is completely taxable. End of story, really, but the trouble is that not all individuals know about this straightforwardly. They probably eventually realise that their income should indeed be subject to tax, but the key with a stop notice is to stop the whole process from kicking off in the first place.
It is really about taxpayer protection, although there is also a bit of Exchequer protection and HMRC protection too. If by adding a criminal sanction we make it even harder for anyone to ignore a stop notice and carry on marketing their aggressive avoidance scheme, that can only be a good thing.
Joshua Toovey: We welcome these proposals. We are aware that many contractors have fallen foul of these tax avoidance schemes in recent years. As Mr Dodwell said, many have been caught up in the loan charge, which has had a devastating impact on their lives and their families’ lives. It is quite right that the Government are now looking at going to the source of this issue rather than going after the contractors themselves, who quite often are unaware that they are inadvertently falling foul of these tax avoidance rules.
The Chair: Mr Dodwell, are you saying that you would like to see the legislation amended to incorporate the bespoke schemes? Are there any other aspects that need to be addressed?
Bill Dodwell: I do not think it needs to be amended to incorporate a bespoke scheme. The purpose of the legislation is to stop mass marketing. You cannot stop people doing tailored tax planning like that. You can do so through the compliance process, maybe the general anti-abuse rule and maybe action before the courts, but the key is to stop someone selling the same thing over and over again to thousands of individuals and, as Joshua and I agree, ruining people’s lives.
The Chair: So you are happy with the proposed legislation.
Bill Dodwell: Yes.
Lord Palmer of Childs Hill: It is the artificiality of the schemes that is the real problem, is it not?
Bill Dodwell: Yes. Exactly.
Lord Palmer of Childs Hill: It is not real business. It is making an artificial situation in order to claim some tax advantage. Do you agree?
Bill Dodwell: I agree. That is who this measure is aimed at. If you look at the list of those with stop notices, you will be as appalled as everyone else that this stuff is out there at all. I am a member of HMRC’s advisory GAAR panel, and some of this stuff gets referred to us. It is all old by that stage, but when you see these things, in most cases you think, ”How on earth can anyone even have thought this had a chance of being successful?” Guess what? It is not. But the fact is that people have paid a lot of money to do it and there is a lot of aggravation, and that is not something that we want to see continue.
The Chair: If there are no other comments on that point from anyone, I thank you all for contributing so helpfully to our deliberations. We are very grateful to you.
