Agenda item 2 is consideration of the report entitled “Administration of Scottish income tax 2021/22”. I am pleased to welcome our witnesses from His Majesty’s Revenue and Customs and the Scottish Government. Alyson Stafford is director general of the Scottish exchequer, and Lorraine King is deputy director of the Scottish Government tax and revenues directorate. We are also joined by Jonathan Athow, who is director general for customer strategy and tax design at HMRC, and Phil Batchelor, who is deputy director for income tax policy at HMRC.
We have quite a number of questions that we want to put to the witnesses. However, before we get into those questions, I invite Alyson Stafford to make a short opening statement.
Thank you very much, convener, and good morning.
Scotland has had fiscal powers to vary tax rates and bands for non-savings, non-dividend income tax since 2017-18, and it relies on His Majesty’s Revenue and Customs for the administration of that partially devolved tax.
I draw to the committee’s attention three particular actions that have been taken forward since last year’s session with the committee. First, on compliance, the Scottish Government continues to proactively consider Scotland-specific compliance risks with HMRC, and it has commissioned a risk-based analysis of the scope of paper migration among the Scottish tax base.
Secondly, on outturn methodology, in response to the National Audit Office recommendations last year, the 2020-21 outturn calculation methodology was updated to reflect the impact of Covid-19 in HMRC’s calculation of the outturn receipts.
Thirdly, on data, we have updated the service level agreement between the Scottish Government and HMRC to streamline data requests and provision. That is important as we focus on building empirical evidence to support policy development for ministers and to inform any further Scotland-specific compliance needs in future.
To stay briefly with the theme of three, there are three key sources for assurance for the Scottish Government on the proper collection of the tax. The first is the operation of the service level agreement and the performance framework, which is vigilantly monitored.
The second is that HMRC has a designated additional accounting officer with responsibility for Scottish income tax and all that goes with that responsibility. That officer is Jonathan Athow, who is with us today to support the committee’s scrutiny of HMRC’s administration.
The third is the valuable third-party assurance that comes from the National Audit Office providing an audit opinion—which has confirmed, among other assurances, that the Scottish income tax outturn for 2020-21 has been fairly stated by HMRC—and the Auditor General for Scotland being satisfied that the findings and conclusions in that report are reasonable.
All of those are vital sources of assurance. We will continue to work with HMRC to ensure that Scottish public finances are underpinned by tax administration arrangements that function effectively and as intended, and to ensure value for money to the taxpayer.
Colleagues who are here today look forward to answering the committee’s questions.
Thank you very much.
As you have suggested, those are very much the kind of areas that we want to probe, question and better understand. However, before we get to questions, I invite Jonathan Athow to give us an opening statement.
I will try to be brief, as Alyson Stafford has covered many of the issues that I was going to touch on.
I will echo two or three points. First, we have been administering Scottish income tax for a number of years, and we have learned from that. You see some of that learning being translated into performance and different ways of working. We welcome a positive working relationship with the Scottish Government. We are also grateful to the auditors for their work and reassured by their opinions.
However, we are certainly not complacent. There is no room for being complacent about risks. HM Revenue and Customs sees a number of risks emerging in the tax system all the time. We need to spot them, identify whether they need to be tackled and, if they do, put in place mitigations. The greater divergence between the rates in Scotland and the rest of the United Kingdom beginning this financial year might be a spur for slightly different behaviour, and we will need to be on top of that. Obviously, we will not have any data on that for a while yet, but we are mindful of the point and we want to be on top of it.
Thank you very much.
The committee is aware that, in places, the reports prepared by the National Audit Office and the Auditor General for Scotland used language such as “continuing limitations” and “risk”, and identified some areas of concern. We will get to those shortly but, before we get into some of that detail, I will take you back to one of the fundamental issues raised in the audit.
The issue came out in the evidence session that we had on 9 February with the National Audit Office and the Auditor General. They drew our attention to the conclusion that the growth in Scottish income tax receipts in the financial year 2021-22 was expected to be 11.3 per cent, whereas the UK equivalent income tax receipts were expected to grow by 13.2 per cent. I turn to Alyson Stafford first. Can you give us an explanation of the Government’s thinking on why Scottish income tax growth has been lower than the growth in the UK as a whole in recent years?
You would expect me and the Government to be very interested in tax performance. The relationship between the economy and tax is complex and reflects a range of factors, some of which are to do with macroeconomic performance, some of which are regional, and some of which are sectoral. Demographics and the composition of the tax base, as well as Government policy, also have an impact.
Although Scotland’s labour market has matched, if not exceeded, UK performance in recent years, with high employment and low unemployment, Scottish income tax performance has certainly experienced the impact of the change in the oil industry. That has affected Scottish receipts in the north-east of Scotland at the same time as there has been strong growth in the rest of the UK, particularly in the receipts in London and the south-east, and that is a factor. The issue is those particular sectoral elements.
The change in the oil industry and what has happened more recently, such as the inflation changes, have been factors. We set that out in the medium-term financial strategy that was published last May to be transparent about the fact that those regional factors have been the driver of the earnings gap. That is a particular factor because of how the fiscal framework operates. It is about relative performance. To a degree, the fiscal framework gives a little bit of a cushion around that. The interesting fact is that a positive reconciliation impact has been built into the budget for the current financial year—2023-24—based on the tax performance and the operation of the block grant adjustment.
I admit that I do not have last May’s medium-term financial strategy in front of me or at my fingertips. When you describe the downturn in the oil and gas industry affecting income tax take receipts in the north-east, we can broadly understand that, but to what extent do you have data that absolutely backs that up?
We will have data about the overall position in 2020-21. In my opening statement, I mentioned the changes to the service level agreement and the extent to which the data is now becoming a little more readily available. Some data flows that will give us a bit more granularity have been coming through in the past few months. We do not have that granularity of data going back to the year in question and that the NAO 2020-21 report is based on, but we are now getting data that is based on the pay as you earn section of the tax collection regime. As you will know from the reports from the National Audit Office, the tax is categorised and administered in a combination of ways, and we are now starting to get some data on the PAYE element of that. Obviously, that will inform us much more about the Scotland-specific elements within that.
For the avoidance of confusion, are you going to get, or are you now getting, regional level data about Scottish income tax receipts?
We are getting more granular data that are not necessarily specific to regions but are more associated with different bands of taxpayers. That will help us, because we know that there is a regional distribution to those bands.
Okay. Thank you.
The other thing that you mentioned, which is of interest to not only the Public Audit Committee but the Parliament as a whole, is the fiscal framework and how that works. Of course, it works in a very particular way. If Scottish income tax receipts are at a certain level compared with UK income tax receipts, there are consequential effects on the operation of the fiscal framework, which can be advantageous but can also be disadvantageous. Can you update us on where the renegotiation of the fiscal framework is?
Yes—by all means. Part of the agreement when the fiscal framework was first set up was that it would go through a review, and one of the key elements of the planned review was a jointly commissioned independent report on the block grant adjustment arrangements. As you have said, the arrangements are complex. It is not just about tax take; it is also about how our block grant is adjusted. The block grant adjustment arrangements also called for stakeholder input. That is prior to the broader review of the fiscal framework.
The report was written by David Bell, David Eiser and David Phillips—those are three names that members might well know—and the specification for that independent report was agreed jointly with His Majesty’s Treasury and the Scottish Government. That has been one key element to inform the process. The discussions will resume with the new Deputy First Minister and the Chief Secretary to the Treasury in order to take that forward. The renegotiation has therefore moved back into an active space.
Okay. I am sure that that academic analysis will be of use but, in the end, that is a negotiation between two Governments, and it was due to be concluded in 2020, I think. That has been on the go for quite some time. Why has there been such a delay in that negotiation? When do you now expect the negotiations to be concluded?
You are absolutely right that the first fiscal framework involved a negotiation. The review will very much look at which areas need most attention. The Scottish Government, informed no doubt by lots of discussion that has taken place in different committees in the Parliament, will have a view on the particular things that are a priority, as will HM Treasury.
All that I can say is that the discussions are now resuming. A specific deadline has not been set or agreed, but we can keep Parliament informed as the situation develops.
Again, for the avoidance of doubt, have those discussions recommenced? You said that they are going to. Is that a future or a present description?
Basically, there has been a conversation in which it has been said that, yes, we want to put those discussions back on to the agenda. Therefore, they will be starting.
So you have had talks about talks.
You could put it like that.
10:15
Another area that the committee has been particularly keen to explore is the implications of a much higher inflation economy, one of the consequences of which is that people’s real-terms living standards are, in most cases, being cut. On the other side of that equation, pay settlements have been higher and people’s earnings—not in real terms but in monetary terms—have been going up, which presumably has an implication for the income tax that is collected.
I turn to Jonathan Athow first. Could you help us to understand a bit more about the impact of inflation on the collection of Scottish income tax?
As you said, inflation has two effects, which I will draw out. First, for any administrative organisation such as ours with a fixed budget, high inflation means that we are under more pressure. Our cost base is going up, but our budget has been fixed, so one challenge for HMRC as an organisation is on offering customer services.
The second effect is with regard to receipts. Most thresholds in the income tax system, either in the rest of the UK or in Scotland, have been either fixed or reduced. Higher inflation feeds through to higher nominal wage growth—that is, wage growth before inflation is taken into account—which brings more people into the tax system and more people into higher brackets of income tax. That process, which brings in higher receipts over time, is called fiscal drag.
The effect of inflation in relation to receipts is, therefore, actually positive—higher inflation feeds through into positive elements of receipts. However, some challenges come with that: the taxpayers who are involved are brought into higher tax brackets, so we get more people into more complex parts of the tax system. One example is that more people are starting to see a tapering away of their personal allowance—that begins at £100,000—which adds a degree of complexity to their tax affairs.
Although we are seeing more receipts, we are sometimes seeing more costs for taxpayers because of the increased complexity, as well as more costs for us in administering parts of the tax system that require taxpayers to do more to ensure that they are compliant.
As the Public Audit Committee of the Scottish Parliament, we are interested in whether any differential fiscal drag exists in Scotland, in comparison to other parts of the UK. Do you have any data that would help to illuminate that point?
There are two components of fiscal drag. The first is the macroeconomic picture, elements of which Alyson Stafford has already spoken about. If you have higher or lower earnings growth in Scotland in comparison to the rest of the UK, you will see higher or lower levels of fiscal drag.
In the oil and gas industry, for example, many of the jobs are very highly paid, so it is not just about the average but about the distribution—having more or fewer people at the top of the income distribution because of the changes in the oil and gas industry will feed through.
The second element is what is happening with the personal allowance and income tax bands. Although there are more tax bands in Scotland, I do not see huge differences in overall policies. The way that tax bands have changed in Scotland has broadly been consistent with what has happened in the rest of the UK. The additional rate band starting at £125,000 instead of £150,000 is the same in Scotland as in the rest of the UK.
Although the band structure is different, I do not think that it adds hugely to fiscal drag. The key determinants are probably the fact that most bands and rates are frozen in cash terms and what is happening to the underlying economy, where, again, you will see a difference between Scotland and the rest of the UK for the reasons that Alyson Stafford set out.
Just to be clear, do you have data to back up that analysis and help you reach the conclusion that there does not appear to be much of a differential impact in Scotland compared with the rest of the UK?
We will certainly see that in the outturn. Part of the challenge with the tax system, which Alyson alluded to, is that, although we get data for PAYE on a regular basis—the data is now reported monthly, various cuts of it are available and some of it is now provided to the Scottish Government—we do not get people’s self-assessment income tax returns until nine or 10 months after the year in question. Therefore, inflation is high now, but we will not get the file with the self-assessment tax returns for 2022-23 until January 2024. Some of the data always has a lag, and self-assessment is very important for higher income groups.
We will have some data for PAYE, but the data for self-assessment will inevitably lag. However, once we have that data, we can do that sort of analysis.
I have one final question in this area. Do you expect there to be a higher level of income tax per head in Scotland as a result of the process of fiscal drag, wages going up, inflation rising and so on?
Yes. Depending on whether the denominator is everyone working in Scotland or everyone with an income in Scotland, you would expect to see income tax per capita grow over time because of those elements of fiscal drag. The only caveat to that is what happens in employment. At the moment, employment is relatively robust but, if there are changes in employment, the situation could change. Other things being equal, fiscal drag will increase the tax take per capita or per worker.
We are a bit pressed for time, so we will move on to another area that has been of special interest to the committee. Craig Hoy will ask about that.
Good morning. In recent years, the main concern in Scotland has been the different tax thresholds, but we are now getting different rates potentially impacting behavioural patterns. I want to quickly put some scenarios to you to make sure that my thoughts are correct and perhaps to assist anybody who is watching. Am I correct in thinking that, if somebody lives in Berwick-upon-Tweed but works in Edinburgh—commuting into Scotland to do their job—they would not require an S tax code?
As long as they are properly resident in Berwick, yes, that would be the case.
If someone was living in Berwick and working remotely for a Scottish company, it would be entirely applicable for them to have a UK tax code.
Yes—it is where they are resident rather than where they are working that determines that. There are some exceptions but, broadly speaking, for most people it is residence. Obviously there are special rules, for example, for MPs representing Scottish constituencies in Westminster. Excluding those special cases, the code is mainly based on residence.
If I was the chief executive of a Scottish bank but lived in London, commuted up from London and stayed in a hotel five days a week, even though I was spending more time in Scotland than in England, I would still be resident in England. Is that correct?
I would need to look at the very detailed rules about what determines residence. We have specific rules for specific circumstances, so there might well be cases where people, because of the amount of time that they spend in Scotland, could be deemed to be resident.
It is not necessarily linked to someone’s primary residence in that respect. Is that correct?
In essence, we look at residence in the round, and there are a number of tests that we would look at in that situation. There are detailed rules, and people have different circumstances. For example, there are people who spend a lot of time offshore. We need to have special rules to cater for a wide range of purposes, but you could have somebody who, as you say, lives in the rest of the UK but spends a large proportion of their time in Scotland who could be considered resident outside Scotland.
I put those scenarios to you because behavioural and working patterns have changed dramatically since Covid. People no longer have to live near their work, and remote working is far more common than it was. To what extent are you or either Government looking at the impact of remote working in relation to the operation of the tax system?
Remote working affects income tax in a number of ways. It is much wider than where someone lives. There are rules around benefits in kind and other such things that are very important. Therefore, this new way of working will affect tax in a number of ways.
We are building a longitudinal data set, because we are keen to understand how behaviour changes over time. Particularly given changing rates, which you referred to, we want a longitudinal study so that we understand who is registered as a Scottish taxpayer and how that changes over time. That will allow us to monitor what happens with the number of people who are Scottish taxpayers and what happens to them. In that way, we will be able to monitor those patterns. We will also see patterns more generally through postcode matching, which we already do—we will be looking at that.
Earlier, I reflected on the fact that some aspects will depend on how employers understand the operation of the Scottish income tax system. The better understanding they have of that, the better able they will be to keep records up to date. We will be able to monitor the situation in a number of ways, but I hope that the longitudinal data in particular will allow us to understand what happens over time.
What are the emerging trends, and what is the emerging evidence telling you about behavioural patterns in Scotland? Is it too early to capture anything, or can we see that, for example, people are not taking promotions or doing the extra shift because the rate of tax is discouraging them?
We are not seeing anything particularly marked in terms of behaviour in Scotland that we would not expect to see elsewhere in the UK. The Scottish Fiscal Commission does the forecasts and costs—to use the jargon—how much additional revenue the tax rate will have brought in. It makes judgments about how people respond to higher tax rates, whether that is by changing their tax planning behaviour or their economic activity by not taking a promotion and so on. Therefore, that will already be factored into that analysis. We look at that sort of thing more generally across the whole of the UK tax system when there are changes, and we make allowances for behavioural change.
As I said, behavioural change is not always just economic change; it can be changes in people’s tax planning affairs. However, again, that will all be monitored. You will be able to evaluate that by looking at changing patterns of receipts. Some work has already been done, I think by the Fraser of Allander Institute, to look at some of those behavioural responses.
Is it fair to say that the greater the divergence, the greater the risk is that people’s behavioural patterns will start to change?
Yes. As I said in my opening remarks, the financial year has just begun and the differential in the additional rate has gone from 1 per cent to 2 per cent. Obviously, for very highly remunerated individuals, that could amount to a significant amount of additional tax. We need to be mindful of whether we are seeing behavioural change. In particular, we are interested in non-compliance—people trying to present themselves as being resident outside Scotland when they are actually resident in Scotland, which would be a key concern for us—and whether the changes drive other forms of non-compliance. As I said, as an organisation, we experience lots of unfortunate creativity in how people try to circumvent their tax obligations, and we need to be on top of that.
The discussion around the impact of the changes and whether it is sensible to have very divergent tax rates is, in effect, a matter for ministers. Alyson Stafford, do you want to comment on what work is being undertaken? We might find out about the problem only once it is too late—once too many upper rate taxpayers have moved or too many individuals are incorporated and, therefore, all the tax receipts are going to the UK Government rather than the Scottish Government. What work is being done to ensure that we are alert to the fact that we do not come to the issue once it is too late and has already had a material impact on tax receipts in Scotland?
10:30
One key bit of work still draws on evidence from when Scotland introduced the five tax bands, which was a major policy shift to having a more progressive tax regime here. As Jonathan Athow said, that work involved academics from the Fraser of Allander Institute and others. There was a collaboration across HMRC and the Scottish Government. That policy evaluation was published and is available for all to see.
Interestingly, more recently, the independent expert panel that was commissioned in November last year to reflect on what had been set out in the emergency budget and the UK Government’s tax changes at the time, stated:
“empirical studies suggest that tax-induced cross-border migration is small in magnitude.”
That is certainly what came through from the research on the 2018-19 income tax policy change, which found no evidence of significant behavioural change from that perspective.
As Jonathan says, people who are at the higher end of the earnings spectrum can afford to pay for advice and can legitimately change their tax affairs. Those behavioural factors are taken into account in our estimates when policies are being shaped. We all—Government and Parliament—also get information from the Scottish Fiscal Commission about the behavioural impacts.
The most dramatic change in Scotland’s taxation landscape was in 2018-19, and that has given us a basis to work from. In the collaboration with HMRC, it was helpful that the tax environment in the rest of the UK was fairly static, giving us better insights at that point.
As Jonathan Athow explained, the difference between the actual impacts and the SFC’s assessment of those impacts will only really be tested when we get the outputs.
Is there a risk that there is a lag, and that public policy is continuing to move in a more progressive direction, towards higher tax, when the impact of that will not be felt until two or three years after any fiscal change takes place?
Jonathan Athow may wish to comment on the analysis that is done south of the border, but that is very much the sort of thing that the Scottish Fiscal Commission looks at. There are regular data feeds for the surveys of personal income. One particular data feed that can go to the Scottish Fiscal Commission is called PUT—I am trying to remember what those initials stand for; someone from HMRC will tell me. There are sources of data that the Scottish Fiscal Commission can draw on. There is a combination of looking at the history and looking at more recent survey data feeds, and we will also do that after the evaluation.
The overall Scottish tax base relies heavily on a very small number of upper-rate taxpayers. It would not take too many of those high-end payers to significantly distort the overall tax take. To what extent should that be on ministers’ radar? I am thinking both of behavioural change and of the impact of inward and outward migration—of fewer people coming here or of more higher-rate taxpayers leaving—if that differential becomes significant. Are ministers alert to the potential double risks that might have an impact on the upper-rate tax take in Scotland?
What we know about the geographical shift comes from recent commentary from experts last November and from the data that we have. As we know, people base their decisions about where to live and work on a range of factors, not only on tax, and that concern is not borne out by the data that we have or by independent assessments.
As you say, the national strategy on economic transformation emphasises recognising the role of entrepreneurs. During the past few weeks, we have seen some real successes with inward investment into Scotland. The inward investment statistics that are produced separately by EY show that that is still a vibrant space. Other things are happening in Scotland that take a more holistic approach and do not just look at tax. All those things are obviously important to ministers as they shape how we can, in effect, raise money in a progressive manner to support valuable investment in public services across Scotland.
Director general, you are in danger of straying into an area of policy that is of great interest to me, which is the extent to which we will have to rely on foreign direct investment versus the extent to which we are boosting our indigenous business base. However, that is not for this morning’s conversation.
I will look forward to that on another day.
Absolutely. Any time.
Bill Kidd has some questions about transparency to put to you.
Thank you, convener. How about injecting a wee bit of excitement into the morning and having a couple of questions about Scottish income tax administration? That is always something to keep you on your toes.
The committee is already familiar with the service level agreement between the Scottish Government and HMRC. It is publicly available, reviewed annually and changed as necessary. Are the terms of reference and the minutes of meetings of the board publicly available? That was wondered and the question was asked, and the answer that was given to us was that there were no details of the income tax board or minutes in the public domain. Why is there no publicly available information on the Scottish income tax board? Would it be possible to publish details about agenda items and the minutes from Scottish income tax board meetings?
Jonathan Athow and I have just discussed that. I will respond to your question and Jonathan can also come in if that would be helpful.
To update the committee, HMRC and the Scottish Government are content to publish the minutes from future board meetings to help with the transparency of the governance arrangements for Scottish income tax. We will start publishing them from the current tax year, 2023-24. You will appreciate that the minutes will go to the subsequent meeting for approval and agreement. I am sure that the committee adopts that as good practice. Those minutes will be made available this year and, as part of that, it will be clear who are the members of the board.
That is perfectly reasonable. That will happen this tax year—is that right?
Yes.
In the interim, would it be possible for the committee to receive by correspondence a summary of issues covered by the board? The Finance and Public Administration Committee has such an arrangement and it receives a summary letter from the Cabinet Secretary for Finance when the quadrilateral meetings of UK finance ministers occur. Could that sort of correspondence be introduced to cover the period between now and when the minutes are published?
You are looking for some insight in the interim. There has been one meeting so far this year. The board meets quarterly, so the next meeting will be in July and you will get the minutes from the first meeting after the July board meeting.
As soon as that.
If that will meet your needs, we can work to that sort of timetable.
That is admirable transparency. Thank you. Does Jonathan Athow have anything to add?
No, I am happy to support what Alyson Stafford said. We want to be as transparent as we can be. We sometimes talk about sensitive operational compliance issues, so some discretion might need to be exercised around that, but in general we wish to be as open and transparent as we can be.
I seem to have been successful this morning—I mean, I know that you did it, and not me!
It would be helpful if you could notify the committee of when the minutes will be published on the website, so that we have warning of that.
Again, they will be published in two locations because income tax is a partially devolved tax and we rely on HMRC for the administration. They will be on the Scottish Government website and the UK Government website. We will ensure that that happens simultaneously to respect both houses.
A question that arose previously was that we could not find the Scottish income tax board’s terms of reference published anywhere. Would you be able to supply those to us as well?
We could do that as part of the same publication, or we could send them to the committee.
Could you send them before we get the first set of minutes, so that we have a sense of what the object of the board is beforehand?
Of course.
Thank you. I am now going to move things on and invite Willie Coffey to put some questions to you.
I want to ask a couple of questions on data gathering, data integrity and data methodology. Alyson Stafford touched on those issues in her opening remarks, so I will ask both witnesses whether they are broadly satisfied that how we do what we do provides the accuracy that we expect and need?
During a previous session, Darren Stewart from the National Audit Office suggested that we might want to look at doing more frequent third-party data gathering exercises. Earlier, Alyson touched on the methodologies that we use. The Comptroller and Auditor General also said that he had identified limitations in the methodologies. He said that taking a proportion of UK-wide sampling of data as Scottish data did not accurately reflect the circumstances in Scotland because of the tax variations and bands. Could witnesses talk a little bit about how satisfied they are with third-party data, how accurate it is and whether there is scope to improve it?
I will ask Lorraine King to respond about the specific operational elements, because the processes that HMRC goes through to ensure that we have a high level of confidence in the taxpayer base being identified are all rooted in some specific data exercises.
The data clash exercise currently takes place every two years, following agreement between HMRC and the Scottish Government. That was previously judged to be appropriate, striking the right balance between resource and the additional assurance of identification that takes place.
The last data clash took place in 2021, and it showed that the proportion of correctly identified Scottish taxpayers was around 98 to 99 per cent when considering records that were active in self-assessment or PAYE systems. However, that does not mean that 1 to 2 per cent of cases were found to be wrong. The exercise compares a snapshot in time using HMRC data against a multitude of snapshots from third-party data sources. The 1 to 2 per cent is uncorroborated rather than incorrect, and in many cases it is due to ceased or inactive records.
The next data clash will take place this year. Phil Batchelor, from HMRC, might be able to provide more information on when exactly it will take place, but we expect to have results from it within the next couple of months, and once we have those results we will be able to assess whether undertaking the data clash annually might be more appropriate given the forthcoming divergence in the income tax systems.
We currently undertake our data clash at the same time as the Welsh Government, so if we moved to doing it annually, there would be some increased cost for the Scottish Government to bear. That would not necessarily be prohibitive, but we would need to bear it in mind and consider whether doing so offers value for money.
Obviously, wherever possible, we try to avoid UK-wide estimates for Scotland. For example, we look at Scotland-specific data for self-assessment debt. However, there are sometimes limitations when certain data for Scottish taxpayers is not available or is not available in a timely way. We constantly look at these areas, and they will come to the Scottish income tax board and be scrutinised. Simply because that is what we are doing at the moment does not mean that it will always be that way. Again, we just need to work out the most sensible area to look at next.
10:45HMRC is generally interested in third-party data. It allows us to double check much more than simply addresses; it allows us to double check income and other aspects. Third-party data will be of growing importance to us.
At a more strategic level, I completely agree with Lorraine King that we all need to keep under review how often we do the current matching exercise for Scotland and whether there are grounds for increasing the frequency of that. That obviously has a cost attached to it, so we would want to make certain that we are being proportionate to the risk that is being presented.
Will it ever be possible to get better performance than 98 or 99 per cent? Ultimately you are trying to get forecasts and outturn to be as close as they possibly can be, but perhaps that is an impossible task. Would you consider using Scottish data only to try to achieve that accuracy, or is it still appropriate to draw on UK data to give us that picture?
We would look at it on a case-by-case basis and ask whether the data is good enough. Part of the reason why we sometimes have discrepancies in those 98 or 99 per cent matching figures and others is that, for example, we end up with people not knowing their national insurance number and being given a temporary number, which cannot be matched very easily in our records.
There will always be a degree of churn in the labour market, with people moving in and out of jobs, which will create some of that noise, as will people moving between Scotland and the rest of the UK in either direction and the lags in that. There will always be a residual of cases that will be very difficult for us to match using our existing data, but we hope over time to be able to identify the sources of those things and tackle them.
As I alluded to in an earlier comment, I hope that employers, who we rely on for address data, will realise that having accurate address data for their employees is more important, because if they do not, one of Phil Batchelor’s colleagues will phone them up and ask, “Why are you not applying an S code in this particular case?” I hope that there will be some behavioural change over time as employers understand the importance of good address data. There are lots of reasons for thinking that the situation might improve, but there will always be people at the margins who do not have their national insurance number and people who move and take a little time to update their addresses.
Thank you. I hope that I can come back in on the S-code issue that my colleague might raise in a moment.
I will raise an issue that I think we raised with you last year. It is pointed out in the National Audit Office’s report in paragraph 2.32 that
“The tax gap is the difference between the amount of tax that should be paid and what is actually paid. HMRC does not currently produce a Scotland-specific tax gap.”
Why do we not know what the tax gap is?
Jonathan Athow can describe a lot more of what goes into calculating the tax gap overall for the UK, but it is broad—it covers all aspects of taxation. We are obviously looking only at Scottish income tax, but it covers other aspects of employment-type tax such as national insurance and broader taxes across the piece.
I will say a bit about that. As you say, the tax gap is a sort of level of non-compliance. We are one of the few jurisdictions that estimate a tax gap on an annual basis for pretty much the whole of the tax system. The challenge with providing more granularity for Scotland or any other aspect of geography is the way in which we undertake elements of that tax gap.
An example of what we do is something called a random inquiry. We would take a group of perhaps 1,000 taxpayers at random and investigate their tax affairs. Say that we find that 90 per cent of them are fully compliant, we would assume that 90 per cent of the entire population is fully compliant. That is the process that we go through. At the moment, we take that group of 1,000 people as being representative of the UK as a whole. If we were to try to do something for Scotland, we would have to have separate group of 1,000 Scottish taxpayers, which would be far more intensive for us to do. It would be expensive, because of the number of tax inspectors that we would need, and it would divert them from other work. Therefore, although it is possible in theory to do that, it would be challenging and costly to do it in practice.
We can look at other sources of data, such as our work with the Scottish income tax board, to find out whether there are particular risks in Scotland or whether particular groups of taxpayers might be more likely to be non-compliant. We can look at things like that.
We will address the underlying issue of risk, but we do not think that it is proportionate at the moment to do a separate tax gap calculation for Scotland. It is better to look at the underlying data and to use that to action the work that we do to ensure that we are tackling non-compliance.
If the fee attached to the service level agreement or other parts of the memorandum of understanding were revised, that is something that you could do, if you had the resources.
It is, but I stress that that is not a simple undertaking. For example, we have recently recruited a large number of extra compliance officers—I will use the term “tax inspectors”—but it takes many years to train a tax inspector. A case might be taken to a criminal prosecution, so inspectors need to be fully trained in many aspects of law enforcement. We cannot easily recruit and bring in extra people.
In theory, what you suggest could be done, but it would be a significant undertaking, with potentially significant costs. I can emphasise one element of that cost. If we make a random selection of 1,000 taxpayers, the tax inspectors working on that information are not working on the highest-risk cases, which means that we are also at risk of not collecting all the receipts that we need to collect, because we are doing those random inquiries. We do that because it is an important thing for us to do on a UK-wide basis and it informs a lot of our work on risk, but we do not see a case at the moment for making that new significant investment. It could be done, if the time, effort and money were put into it, but we do not think that it would provide sufficient benefit.
Alyson Stafford, would there not be a benefit to the Scottish Government in having that data? We are talking about £11 billion or £12 billion of revenue. We are the Public Audit Committee, but would it not be sensible, even at the basic housekeeping level, to know the difference between what should be paid and what is actually being paid in Scottish income tax?
Working jointly with HMRC, we look at the risk assessments for non-compliance and non-payment. That information is refreshed annually. The risk in that particular area is deemed to be low.
The area that has real merit and in which we are making an investment is the longitudinal data set. Jonathan Athow already said a little about that in his answer to Mr Hoy. That data will allow us to analyse and understand movements across the UK. We have also requested additional risk-based analysis of the behaviours that may be very specific to Scotland. We are targeting the areas that this committee often asks us about and the things that are likely to be more of a risk and will need to be addressed through the work of HMRC.
We have heard already, in response to some of the other questions, about the small relative margins that we are looking at. That extra work includes looking at taxpayers who own more than one property in the UK in order to see whether there is any risk of paper migration to avoid paying Scottish income tax.
We are targeting our investment in the areas that the risk assessments are informing. The particular one that you are asking about is not seen as having the same risk exposure. You would expect us to be making risk-based judgments and investments in what we are doing in continuing to achieve robustness and build on the work that HMRC has done.
As we have seen, the National Audit Office has taken a great deal of reassurance from HMRC’s processes and systems. I have to take some comfort from that, and I do. Equally, the Auditor General for Scotland has placed reliance on the systems and processes that the National Audit Office has gone through to reach its conclusion.
I accept that, but the quotation that I read out is from the National Audit Office’s report, so there is obviously an issue. It might be a low priority, or it may be a low risk in your estimation, but nonetheless the National Audit Office has drawn it to our attention as being worth commenting on. We understand that there are priorities and risks to be assessed, but to us it seems to be a pretty fundamental piece of data that it would be useful for the Scottish Government—and for the Scottish Parliament, because it sets the rates of income tax—to have in order to understand what the tax gap is.
We will move on now. I invite our deputy convener to put some questions to you.
Good morning. First, I have a quick question that follows on from Craig Hoy’s line of questioning. Do you currently have figures showing how many people work for Scotland-based companies but live in England, and vice versa? My question concerns hybrid working, which allows people to work in any country. Are there figures showing how many people would have worked in the UK before the pandemic but now live abroad while still being regarded as working in the UK?
I will take you back a bit. We know where people live—or we hope that we do, because that is why we have been doing all the exercises that we have been discussing. Often, though, we do not know their places of work. We will know where a company’s head office is, because that is where its payroll will be administered from, but that is not necessarily people’s place of work.
That arose as an issue during the pandemic, particularly when varying geographical restrictions were in place. For example, ministers asked us to consider whether there were ways in which we could target particular employers through the furlough scheme. At the time, because of the limitations that I have mentioned, we did not think that that would be robust, because we would only get information on employers’ head offices. For some large employers, such as the large supermarkets, all the data would be on where their head offices are and so it would not be particularly reliable. We have considered whether we could collect such data in future so that if, for example, in some future national emergency, we needed to put financial support in place, we would be able to do so.
Phil Batchelor was closer to that work, so he can talk about the challenges that it presented.
The first point to make is that, as Jonathan Athow said, a person’s liability to Scottish income tax is based on their residence as opposed to their place of work. Therefore we would not expect to see differences in the level of Scottish income tax collected as a result of changes affecting people’s places of work. It is definitely important to recognise that a person’s residence is the factor that determines their liability to Scottish income tax.
As Jonathan also said on the subject of data from employers, the way that we are organised is based on PAYE schemes, particularly for larger employers that might operate across borders. The data that is held on such schemes will not have the granularity to specify whether people are working north or south of the border or, indeed, whether their duties currently take them overseas.
So, you do not have those figures at the moment.
We would not currently have figures on that, but we are thinking about whether we can collect data in that area or whether employers even hold it. As committee members can imagine, for some employees—for example, those who repair lifts—determining where they work is an interesting idea, because they might be travelling around quite a lot. We want to ask employers whether it is meaningful to collect such data.
Obviously, for many employees, it is meaningful to collect it. If they are working at a supermarket in Edinburgh, that is pretty straightforward but, as I said, there will always be a category of employees for which that is difficult. We have asked employers whether they would be able to collect that data and provide it to us, to allow us, in future, to be more nimble in responding to events such as a pandemic or any other time when we would need to put in place additional financial support. The data is not available at the moment, but we are pursuing that for the future, in which case you would be able to understand exactly those sorts of changes.
11:00
I was wondering more about the behavioural impact. The First Minister has said that he wants to go further with progressive tax measures, so, my question was about whether, if there was more divergence, more people would cross the border.
As I said, the longitudinal data will look at residence, so we will be able to understand how patterns of residence might change over time but, at the moment, we do not have data on the dynamic between where people are resident and where they are working.
Do you have any figures to show how many people now live abroad but work here?
I will have to take that question away and look at how well that figure is estimated. There might well be some survey data or something else that allows us to understand that.
How would tax work in that situation? Where would those people be taxed?
Again, residence is the primary way in which tax—
Would tax be paid in the country in which they live?
Yes—indeed, and that is the basis both internationally and in the UK. The broad basis is one of residence, as I said, with some minor special rules for certain groups.
I will move on to questions on identifying Scottish taxpayers. Although it affects only a small percentage of cases, there are still employers who issue incorrect tax codes. The committee has been informed that this is usually the result of a software error. Could you tell us more about what the software error is and who the software belongs to? What, if anything, is being done at the moment to prevent it happening in the first place?
I will start and then I will hand over to Phil Batchelor to talk about that. Administering the PAYE system is the responsibility of employers. Obviously, we provide guidance, rules and codes for PAYE schemes to operate. Primarily, the issue is with the providers. The challenge is that a huge range of software is available for employers to run PAYE schemes, and we see that when we make changes to income tax or national insurance more generally. For some providers, it is very straightforward, because they use the most cutting-edge technology, which is cloud-based, so they can just change a parameter in a programme and everything flows through. However, there are also people who are still using more traditional payroll software, which might be quite old and might struggle sometimes to deal with changes or particular circumstances. Phil Batchelor can say a little bit more about that.
We have certainly seen an improvement in performance in that area. In 2019-20, about 4 per cent of employments might have had an S code applied incorrectly, and the figure is now down to something in the region of 1 per cent. That is still higher than we would like, and we continue to work with employers to improve it.
As Jonathan Athow said, there can be a range of situations. Particularly for a large employer, even a relatively minor glitch in applying the codes that we have put out can affect a large number of employees. As Jonathan said, if an employer has recently put in a new, complex payroll system and there is a glitch in how it apples the codes that we tell the employer to operate, that can flow through into a large number of employments in a year.
I will emphasise two things: first, in every case, we put those errors right. From the real-time information reports that they send back to us, we identify the employers that have failed to apply the correct S code and, subsequently, that will be put right. By the time that we get to an outturn figure for the Scottish Government, those issues will have been corrected.
Secondly, in the meantime, there is an impact on the employee who has been paying the wrong amount of tax in year, so we have quite an intensive programme of employer education, to get them to improve how that is going.
We are observing that it is relatively seldom that we see that happen again after we have contacted an employer—so, those education activities are successful. As I have said, we have quite a major reduction in the number of S codes that are being applied incorrectly.
How much does chasing employers about incorrect tax codes cost HMRC? Do you have that figure?
I do not; I think that that is part of the national activity.
Again, we are transparent about the cost of operating the Scottish income tax system. We normally do a huge amount of work to ensure that employers are compliant. In general, PAYE is a very good system, with low levels of non-compliance. The tax gap there is normally very low—around 1 per cent. We have to do a set of activities with employers to ensure that they are correctly operating the systems around income tax, national insurance and, sometimes, student loan payments, which we call “employer compliance”. A lot of it is about education and supporting people to get things right. Very few of the errors or problems that we find are deliberate; they are often glitches in the system or happen through people not fully understanding how they should operate it.
Various software packages are used, depending on the company. What are the characteristics of the employers that make those errors—are they big or small?
We see a range of activity. As I have said, we know people who are using older information technology systems can face difficulties and challenges. Such businesses might include, for example, fast-growing businesses that are scaling up and, therefore, have a payroll system or department that was designed for a small organisation. There are a number of factors beyond what one would expect. However, we are not seeing any particular areas of concern around that point. People operating the hundreds, if not thousands, of separate software products that are out there to get their tax system right is just an on-going challenge for us.
That situation has consequences when we want to make changes to the tax system in that we have to give people a lot of notice. Changing a system if you are working on an old one takes some time, and real challenges exist around that. For example, if the Scottish Government wished to make changes to tax rates or thresholds, we would communicate that as early as possible, because the earlier we can communicate it, the less chance of errors there is.
A number of factors drive that matter and nothing suggests that concerns exist around particular areas or practices.
So, you go back and inform employers of any issues. Is there any evidence of repeat offenders? If there are repeat offenders—who do so deliberately—is any action taken?
We can take action if people are not operating PAYE correctly, but that is very rare. Some people use avoidance schemes, in which we see people not operating PAYE correctly or operating it on only a portion of people’s incomes, so there are some issues there but, in general, PAYE works very efficiently.
Phil Batchelor might want to add colour to that.
I only note that we see no evidence of recidivism, as it were, in the application of S codes among employers. We concentrate on large employers, not because they are any less compliant, but because their system has an effect on a large number of employments and individuals. If we are prioritising our attention, that is where we do so.
I have one last question. Who is responsible for notifying HMRC if somebody moves house or changes address?
It depends on which income tax regime somebody is in. For PAYE, we rely primarily on employers—if you are recruiting someone new to your company, we expect you to have an address for that person. The challenge comes when people move and do not always tell, or take some time to tell, their employers.
For self-assessment, we primarily rely on the individual themselves because we have traditionally communicated with people by post. It varies between the tax regimes but employers are important in the process. As I said, through the operation of Scottish income tax, employers have become more aware of the importance of address data in getting their employees’ tax right. I hope that there has been a bit of a learning curve and that employers are now more sensitive to that than they have been in the past.
They are relying on employees to inform them. Is there nothing in place for estate agents or letting agents to notify changes of address or anyone buying properties?
No. We do not use that but we are looking at how we can use third-party data more generally. As we said, sometimes the clashing exercise that we do yields inconsistencies and we try to contact people to get things right, but that process is not automatic. The primary way in which we get addresses is through employers or the individuals themselves.
Willie Coffey wanted to come back in on this area.
I want to probe a bit further into the number of taxpayer records that do not have the proper S code attached. Something like 41,000 were identified in March 2022, and in March 2021 it was 39,000. The percentage might go down a wee bit but the actual numbers are going up of cases where S codes are knowingly not being applied correctly. Is it an offence for an employer not to apply an S code correctly when they know that they should?
Most of the time, we would not see operating the tax code incorrectly as an offence. It is primarily a question of education. At the end of the year, even if an employer has not operated a tax code correctly, we will correct it to make certain that it is right.
However, if we were to see an employer serially not meeting their obligations, that would be a concern for us and there are potential penalties that we could apply. As Phil Batchelor said, however, we do not see that. It is more a challenge caused by people not being aware or having slightly clunky IT systems that do not operate properly. In general, our approach is based around education and support rather than penalties or the threat of prosecution, particularly with employers that we know have shown very little deliberate non-compliance.
It says in the report that a small number of employers are involved, but it is 41,000 people. There must be a large number of people who are working for a single employer that is not paying its proper S code tax amounts. Is it fair to say that?
Exactly. As Phil Batchelor said, sometimes a large employer has a problem and, because they are a large employer, it will affect a large number of people. It is not necessarily the same employers who come up every year; it is simply a churn whereby a problem arises and something is not handled properly in their IT system and they need to correct it.
It is going to be interesting to see how much of that will get better over time as employers learn, and how much of it is natural because an employer makes a change and their IT system does not quite respond in the way that they expect. We will learn a bit more as we go on. I hope that we are seeing encouraging signs of things going in the right direction, but it remains a challenge.
The primary problem is that the employee is not paying the correct tax during the year. That either means that they end up with a little bit of a windfall at the end of the year or they have to pay a bit more than they expected. It does not affect the receipts being collected, but it does have an impact on the employee. That is the primary way in which the incorrect application of the S codes would affect the operation of the tax system.
11:15
Is information on the employers involved publicly available, or is it confidential and we will never know who they are?
The rules under which we operate give us very little discretion to name any taxpayer or organisation involved, so we would not be able to disclose that information.
We have powers to disclose sometimes, but that is only for the most egregious behaviour. The scenario that we are discussing involves inadvertent errors or problems with systems rather than anything deliberate, so it would be disproportionate to name the people involved.
So, employers can apply the incorrect tax code to their employees and remain anonymous?
If they were doing so systematically and repeatedly, we would look to issue penalties or prosecution. However, what we see at the moment are administrative problems and challenges, and those are very different behaviours from those that are harder edge. His Majesty’s Revenue and Customs sees a wide range of behaviours, from simple error and things going wrong through to hard-edged organised criminal gangs. The S codes issue covers all of that spectrum of error, so that is how we treat it.
My last query is whether there is an overall value of the loss attached to the 41,000 people who are not paying tax under the proper Scottish tax code.
As I said, I do not know whether we have that number. It would not be a loss overall, because it might apply to only one or two months. I would have to look at whether we have data on that. You might be interested to know whether, for example, 40,000 out of that 41,000 are out by only one month, because then the picture is different. Let us go away and see whether we can find some data on that 41,000—or whatever the number is—to give you a bit more insight.
Craig Hoy wanted to ask a quick question; I am not sure what it is on.
If you do not mind, I am going to crunch into reverse and go back to the issue of behavioural patterns, because there were some figures that I could not locate earlier, but I have found them now. A report that HMRC produced said that, for those in the upper-rate tax band—those who earn £150,000—a 1 per cent reduction in the percentage of income retained after tax leads to a reduction in income declared of between 0.52 and 0.77 per cent. The report made it quite clear that that was down not to non-compliance but to behavioural change. Does that not show that we already have empirical evidence that suggests that such tax increases will lead to behavioural change? Who should advise ministers about that—should it be HMRC, the Scottish Exchequer or the Scottish Fiscal Commission? Those figures seem pretty clear to me.
Those numbers were reflected in the Scottish Fiscal Commission’s estimate of how much additional income would be brought in—it applied its judgment to that.
As you said, there is quite a lot of research on those parameters now, and at the risk of going into jargon, what you described is called taxable income elasticity—TIE. There is considerable evidence out there.
The Scottish Fiscal Commission made a judgment on that but, normally, as operators of the tax system, HMRC does research and allows others to do research, and it is then for policy makers to determine how they take into account those judgments around those parameters. Our job is to make certain that that information is available, and it is then for policy makers to decide how they wish to weigh up the different factors that they take into account when setting policy.
I know that you have tax design in your job title; in terms of tax design, that is a pretty clear piece of the jigsaw that you would build into the tax system.
Indeed. To go back to a UK-wide policy announcement, that particular parameter was a key issue in deciding on the future of the 50 per cent tax rate that was introduced, which was then reduced to 45 per cent. All those sorts of decisions take into account those sorts of behavioural responses. It is then for ministers to weigh up the various competing issues that they wish to take account of. That is a matter for Scottish Government ministers to take into account.
Thank you for the clarification.
My final question is for Alyson Stafford. Could you tell us the Scottish Government’s position on the deployment of private debt collection agencies?
The administration and collection of Scottish income tax is actually part of what HMRC does. That is part of the arrangement to maximise fairly the tax collection that is due to Scotland.
I am aware that, because of the level of debt that built up through the pandemic, colleagues in HMRC have not only increased their own capacity but have used outsourced debt collection arrangements in particular situations. We would expect HMRC to commission that activity in a way that is compassionate and is a proportionate response to take in relation to the arrears. I am also aware that it may well be the case that, through those measures, HMRC seeks to collect not only tax that is fairly due for Scotland but other tax that may also be due.
We recognise that there is a challenge and that we do not have a direct link in. That is still part of the overall administration arrangements that HMRC has to apply across the board to the UK. It is fair to invite Jonathan Athow to talk about how HMRC administers those contracts, but we recognise that there has to be a mixed economy in how tax debt is followed up.
I am very much aware of the fact that tax debt is generated on income that has already been earned by individuals. Obviously, we want to make sure that, where the liability has been assessed, it is fairly collected, as I said, in a compassionate and proportionate way, because I am very much aware that the tax then pays for public services in Scotland and for public servants such as yourselves. It is about keeping the flow through.
Jonathan, would you like to add a little more on how you oversee that debt collection arrangement?
Before Jonathan Athow does that, I will ask you a question. You said that the Scottish Government is in favour of a mixed economy, but it does not have a mixed economy in relation to the recovery of overpayments of social security, for example. It has a very clear policy that any overpayments through Social Security Scotland are recovered by in-house teams. That is not outsourced to private debt agencies, so why is this different?
It is based on the fact that the administration of income tax is done by HMRC, and we rely on HMRC deploying its arrangements for Scottish income tax on behalf of Scotland. There is a distinction between something that is administered directly in Scotland by the Scottish Administration and a type of administration that relies on the systems, processes and services that are part of HMRC.
Individuals who owe tax have already had the income, and we want to make sure that the tax that is due on that income is fairly collected. As I have said, it is important to us that that is carried out in a compassionate and proportionate way.
We have a couple of minutes left. Jonathan Athow, if you want to add to that, please feel free to do so.
As Alyson Stafford said, we operate a debt collection service that is based on all debts owed to HMRC. In theory, a number of different taxes could have given rise to a debt, including Scottish income tax, other taxes or a mix of the two. We use debt collection agencies only for certain roles. They do desk-based work, such as making telephone calls, writing letters and sending text messages. They never go out to visit taxpayers or claimants, and they do not take enforcement action through the courts, so their role is limited. We oversee them, they are regulated by the Financial Conduct Authority and we also audit their behaviour to make certain that they are abiding by our rules and processes.
One of our organisation’s current challenges is that, at any one time, we have about £40 billion of outstanding debt. We want to avoid any of that becoming non-payment and therefore a loss to the Exchequer. For us, it is really important that we collect the money that is due, but we also need to make certain that we do that in a proportionate way. One challenge that we sometimes face is people not coming forward. If they come forward, we can offer them time-to-pay arrangements or take other steps to support them. We are mindful of the challenges here, especially given the wider economic circumstances that mean that many people feel financially squeezed at the moment.
I hope that I have given the committee a bit more context on how we operate with debt collection agencies.
We are really out of time, but I will make one final request. Could you supply to the committee in writing a list of the private debt collection agencies that operate in Scotland?
Yes.
That would be helpful.
This morning’s session has been very useful for the committee. I thank Alyson Stafford, Lorraine King, Jonathan Athow and Phil Batchelor for giving evidence. We will consider whether we would like to take any further steps but, for the time being, I thank you for your evidence and draw the public part of the meeting to a close.
11:27 Meeting continued in private until 11:38.