Contents
- Attendance
- Earnings
- Scottish VAT Assignment
Scottish VAT Assignment
Our next item is to take evidence on Scottish VAT assignment. We are joined by Derek Mackay, the Cabinet Secretary for Finance, Economy and Fair Work; Jonathan Sewell, who is the head of the Scottish Government’s income tax and tax strategy unit; and Iain Pearce, who is an economic adviser to the Scottish Government.
I welcome our witnesses to the meeting and invite the cabinet secretary to make an opening statement, if he so wishes.
Thank you, convener. I am mindful that the committee is exploring the Scottish VAT assignment, and is considering the evidence, as we continue to work it through with the UK Government. Officials are doing their best to reach a point at which we are confident about the methodology.
I should say, in that regard, that although discussions with the UK Government are on-going—I have recently raised the issue with the Chief Secretary to the Treasury—there are a lot of issues around volatility, and there are questions around Brexit and the margin of error that we are dealing with. For example, we will never have an outturn statement, so the figures are based on estimates. I have raised my concerns with the Chief Secretary to the Treasury because, in order to continue, we need the sign-off of the joint exchequer committee and the agreement of the UK Government and the Scottish Government.
As we work through those issues, I am becoming increasingly minded to postpone VAT assignment until VAT powers can be further discussed at the time of the fiscal framework review. That would give both Governments time to assess the robustness of the model and to understand the impact of European Union exit.
I am happy to explore with the committee the reasons for my concern about what VAT assignment might do to the Scottish budget.
It is a substantial move to say that you are considering postponing VAT assignment. However, we still have to go through the process today of ensuring that we understand the dynamics behind your decision—if that is the decision that you finally come to—and the challenges that have led you to that position. In our evidence taking this morning, we should be as robust as we can be to help you to come to a conclusion.
As you mentioned, one of the major challenges in assignment of VAT is that, unlike for other taxes, there will never be outturn data. As a result, assigned Scottish VAT will need to be estimated, which—as you said—introduces another layer of volatility. To what extent could that volatility be mitigated when VAT is finally assigned? Will you explain how that volatility is driving your thinking?
10:15
One of my concerns is to do with the fact that VAT assignment is based on survey data and estimates. We will never have outturn data; we will always be comparing one set of estimates with another set of estimates. Unlike what happens with income tax, there will be no reconciliation. The margin of error for both sets of estimates is such that we could be in a positive or a negative situation.
We will not have certainty. If we had to take money out of the budget or use our borrowing powers or reserves because there was less revenue, that would have a financial impact. I am concerned about the accuracy of the assignation approach, given that we are talking about almost £6 billion of revenue. There is a lack of certainty about its accuracy.
As far as the timescale for a resolution of the problem is concerned, we are in a transition period. If we had more time, we could look at the data, the surveys and the estimates. We would have more information to go on. We do not have that at the moment.
On top of the uncertainty, we face the potential impact of Brexit, which would hit the economy, consumption and VAT receipts. We do not know whether the impact on Scotland would be disproportionate. The lack of data is an issue, and because we are in a period of uncertainty, there is a risk that I am not convinced we should be carrying. Assignation is different from having a power on VAT: we cannot push and pull levers to make a difference on VAT income. I am concerned, given the lack of policy levers, about the level of risk and how we would address any negative impact through the block-grant adjustment, with which all members are familiar.
We want the relevant power to be devolved. Assignation is not a devolution of power, and it carries a disproportionate risk because of uncertainty and lack of data. Once we have boosted the data samples, we will have more data. We can discuss the merits of having a power, rather than assignation, more fully in the review of the fiscal framework, when we will have more data that will allow us to make an informed judgment. I have expressed that position to the Chief Secretary to the Treasury.
In effect, you are saying that the modelling will need to go on for longer so that you can find out whether the risks can be mitigated and so that you can see what the trends and flows might be.
Absolutely. Although assignation would still be based on survey data, at least we would have assessment over a period of time. At the moment, we face uncertainty with Brexit, which could have an impact on the economy. Given that we face volatility, added uncertainty and concerns about the level of accuracy of the assignation process, it does not feel as though we are in a position to sign off the proposal.
The issue is worth exploring in some detail, because we are talking about a lot of things at once. I would like to parse the issue and separate out the concerns. To what extent is your current position—wishing to postpone or delay assignation—associated with Brexit uncertainty? Would that have been the case regardless of Brexit?
I would like to put Brexit to one side for a moment. Notwithstanding Brexit, as I understand it—correct me if I am wrong—there can never be outturn data, so there can never be a reconciliation process. Was assignation therefore doomed to fail from the beginning?
It has been a difficult situation, because we have only ever been talking about assignation, and not devolution of the relevant power. Many people perceive assignation as devolution to Scotland of a power that will put Scotland in control of VAT. That is not the case. Assignation was always going to be challenging, and for it to proceed we need to have confidence in the accuracy of the estimates and their ability to truly reflect what is going on in the Scottish economy—
Indeed—I am sorry to interrupt—but is that possible? We all agree that we want confidence in the estimates. My question is whether that is possible. You and your officials, and the UK Government and its officials, have spent two years or more working on the methodology to try to nail down estimates so that they are not just guesswork. Notwithstanding Brexit uncertainty—I will come to that in a second—is that doable?
Our agreeing to boost the data does not address the issue of accuracy. I have two very well-educated officials here who might offer a view on whether we can fix the situation with even more data, but I fear that we cannot. I also fear that because we are talking about survey on survey and estimate on estimate, the accuracy issue is difficult to overcome.
Politicians and Parliament—this is about the Parliament’s powers; it is not just about the Government—make a decision on the level of risk that we wish to carry, and the level of risk that has been identified has not been overcome, even with boosted data, with the best will in the world and with the co-operation of civil servants. The answer is no—we have not found a way to address the risk issue, which is compounded by Brexit uncertainty.
To be absolutely clear, are you saying that the position is that the risks that are associated with assignment of VAT are so uncertain—because we are talking only about estimates and survey data, and there is no outturn data—that assignment could not be done to the satisfaction of the Scottish Government, irrespective of Brexit?
What we signed up to in the Smith agreement, as I understand it, is that there must be joint signing-off of assignment. We have not yet identified a system that gives us confidence in the accuracy of the assignment; civil servants have not yet identified a system that can deliver that level of accuracy. That is the position.
Of course, we still want to take on the power, while recognising that assignation is not a power. As things stand, the level of risk is such that we are not in a position to sign off.
The Smith commission unanimously recommended that there be no devolution of powers with regard to VAT rates and exemptions because that would be illegal—it would be contrary to European law. As I understand it, European law prohibits there being more than one rate of consumption sales or value added tax in a member state. Just to be clear, it is not that the United Kingdom Government is saying, “No—we are not going to devolve those powers”; it is that European Union law, by which we continue to be bound, makes that illegal. Is that a correct analysis?
Yes, that was correct at the time.
So, the Scottish Government’s position is that one of the advantages of Brexit is that it becomes possible to think about devolution of VAT, rather than mere assignment of VAT. Devolution of VAT is possible only if the United Kingdom leaves the European Union.
You are, at least, accurate in the proposition that you are putting forward, but it is not a compelling reason for the Scottish Government to support Brexit. It would not change my vote to remain in the European Union.
It is true, as you said, that the UK Government has choices—or might well have choices, if there were to be a different legal position. It could devolve the power if it so wished, which it could not have done under the previous arrangements. That is accurate, but—as I said—the position is not in itself a good reason to support Brexit.
Indeed. This will be my final question, for the time being. Earlier in your remarks, you said that Brexit uncertainty might have an impact on consumption and that it might have a disproportionate impact have an adverse effect on the revenue stream that would accrue to the Scottish Government under any methodology for calculating VAT assignment. What evidence supports the fear that a Brexit impact on consumption would be disproportionate in Scotland?
That is to do with the nature of the VAT that is paid in Scotland. One of the officials might want to offer a view.
There are two aspects to Brexit and VAT. First, there is the question of the disproportionate impact on Scotland. As members know, the majority of VAT is paid by the household sector. Based on, for example, the Scottish consumer sentiment indicator, which is published by the Scottish Government, there appears to be a larger impact on household confidence in Scotland, relative to the rest of the UK, and it appears to be associated with Brexit. There seems to have been a larger effect on Scottish households than on those in rest of the UK.
A second question around Brexit uncertainty and VAT is about the fact that VAT is a harmonised EU-wide tax, to a degree. Therefore, the collection of VAT and its paperwork systems are designed in such a way that they do not distinguish between import VAT from the EU and that for the rest of the world. At the moment, as far as I am aware, new paperwork to collect VAT is not in place; that will depend on how we leave the EU. There are risks about how collection of VAT would happen post-EU, which means that 2019-20 would become very uncertain as a base year.
I am still not sure that I understand how the impact of Brexit would be disproportionate in Scotland, compared with the rest of the United Kingdom, which is the issue.
Household spending is one of the largest components that drives VAT and the economy in general. If the effect of Brexit on household confidence is larger in Scotland—
Why would it be?
Households in Scotland have a slightly different preference on Brexit. They voted slightly differently from households in the rest of the UK in the referendum, so they might be more risk averse in relation to the impact that Brexit could have on their personal finances, which would lead them to save more and spend less. There are a number of channels—
Hang on. Would you therefore expect to see the same effect in London, which also voted heavily to remain? Are you seriously suggesting that the parts of the United Kingdom that voted to remain will lower their consumption after Brexit, compared with parts of the United Kingdom that voted to leave? Is that the Scottish Government’s position?
That is not the Scottish Government’s position. Evidence in the household survey data suggests that householders have already changed their behaviour in respect of how they respond to what they perceive as future risks to their incomes. They might change their spending patterns depending on how they view future risks to incomes. On average, Scottish households have a discernibly different approach to future risks from households in the rest of the UK. That approach to risk could, therefore, crystallise in different spending patterns in the future.
Convener, I can return to that issue, to be helpful. It is a fair question. Our concern is that uncertainty will be put upon uncertainty. It is a matter of accuracy while being about to engage in a fiscal system and fiscal change at a point of expected turbulence—if there is a Brexit. The economic consensus is that there would be an impact on various sectors and areas and, we would expect, on patterns of consumption. Because VAT is a consumption tax, there would be an impact.
There is also the potential difference between the two estimates. As we have discussed at great length, because of the different operations of the Office for Budget Responsibility and the Scottish Fiscal Commission, there are issues about the different analyses that will drive the number to which we would be beholden.
The issue is the level of risk. What is unknown is the potential divergence between the economies in Scotland and England in terms of consumption and VAT, and interpretation of that. There is uncertainty piled on uncertainty, so to make the shift at a point of such volatility feels ill-advised.
That said, who knows what the circumstances of Brexit will be or the outcome over the next few weeks? My point is that it is in our interests to have more time to collect more data and to do more comparative analysis. We could then see what the divergence is and make a decision on that basis, which would be better than what feels like going in blindfolded.
Cabinet secretary, you correctly say that, in order to have confidence in the VAT assignment process, you need to have estimates that are accurate and robust and in which you have confidence. It has been said that 70 per cent of VAT relates to household spending. How confident are you about the data sources in the current model?
10:30
That is my point—I am not hugely confident. We have gone as far as we can and even boosting the sample data does not give us numbers that fill me with confidence or make me feel that they absolutely truly reflect the state of the economy and the levels of VAT being paid in Scotland.
Bearing in mind that this is a £6 billion assignation, if the margin of error is tens of millions of pounds or even hundreds of millions of pounds, that is a risk. Fundamentally, we would have to pay for that through the budget, and the level of risk—the margin of error—is such that it could have a material impact on the resources that we have available.
I am happy for the officials to say more about the data sets, but I am not filled with confidence that they give us what we need to justify the level of risk that we are carrying.
It is also worth reflecting on the point that the volatility created by the data sources and surveys working together is not a risk that we should face. There is a performance risk around volatility, which goes with all tax. That risk is managed within the levers that we have. However, the way that the surveys work together—even if there can be confidence with some of the surveys—creates additional uncertainty that to us does not appear to be a robust or appropriate way to manage the public finances.
With the methodology as it stands, can you express a level of error or a level of confidence in the calculation that has been made? I am not asking you to state that level of error, but does the methodology contain that?
We cannot state that at the moment. It is being looked into and we intend to come up with that number, which is obviously important. We do not have it at the moment, because a large number of components go into the model, not just the household survey, and each of those has an associated individual uncertainty.
Although we might know the associated individual uncertainty, we are looking at different methods for adding them all together. That said, broadly, numbers that are already in the public domain suggest that an uncertainty level of around 2 per cent is quite standard for a household survey, so that is the sort of ballpark figure that you might expect any survey base measure to provide.
As it stands, you do not have an overall figure on what the level of error might be, plus or minus.
Not at the moment.
What would be the position on VAT-exempt businesses?
Before we move on to that question, I want to explore the issue of confidence intervals, on which our adviser and the Scottish Parliament information centre have provided information. Our background paper states that one of the main limitations of the paper that was produced by the Treasury and the Scottish Government is
“that it does not provide any assessment of the confidence intervals likely to be associated with the VAT estimate”.
For context, our paper adds that
“the GERS methodology paper states that the 95% confidence interval associated with the GERS estimate of Scottish VAT is +/- £223 million.”
If we do not even have that confidence interval information, the potential for volatility and a fluctuating position on VAT in the Scottish budget must be significant. We do not even know what that number is.
That is a fair point. The committee may be aware that Her Majesty’s Revenue and Customs will be publishing a statistical publication on VAT assignment at the end of this month, which I hope will provide more clarity on statistical confidence intervals around the VAT assignment model estimates.
I am sorry to interrupt you there, James, but I just wanted to make sure that that point was covered.
It is a valid point, convener.
We have seen a suggestion that VAT-exempt businesses have not been properly built into the methodology. Is that the case?
I am not sure that I fully understand the question. VAT-exempt businesses account for about 15 per cent of total VAT liabilities in the UK. That is reflected in the VAT assignment model, where it is estimated how much VAT liability is coming from those businesses and then we have an estimate of how much VAT they are generating in Scotland. In simple terms, VAT-exempt businesses have been built into the model, with Scotland-specific estimates.
You said that sample sizes have been boosted. What were their original sizes? What have they been boosted to in order to give you more data?
The boost in sample size relates only to the household sector of the VAT assignment methodology—it does not relate to other areas such as VAT-exempt businesses. The achieved sample size of the household sector in 2017-18 was, I think, 725 households.
That seems quite a small number. Is one of the problems that the sample sizes are very small?
One challenge with the VAT assignment household survey is that, because VAT is quite a complicated tax in the sense that different rates are charged depending on the product, detailed questions need to be asked about household expenditure to understand what products are being purchased. That contrasts with, for example, the Scottish household survey, which has a much larger sample size and does not ask detailed questions on what households are spending money on.
We have an unfortunate situation in that our need to collect detailed information becomes burdensome for the households involved and those collecting the data, which limits the sample size that can be achieved in such detailed surveys.
What is the calculation of the total number of households that contribute to VAT?
Do you mean how many households paying VAT would be involved in the survey?
No. You have obviously extrapolated a calculation from the sample of 725 that you mentioned. What does the 725 figure map on to?
There are slightly more than 2 million households in Scotland. I expect that nearly all of them pay VAT, because almost any form of household expenditure, other than very basic foodstuffs, includes some form of VAT—even heating or electricity attracts VAT at the 5 per cent rate. It would be difficult for a household not to spend income on VAT.
Right. I repeat that it seems to be a small sample size, given the size of the population that pays VAT.
Neil Bibby has a supplementary.
Is there any concern that the size of the cash-only economy is higher—or lower—in Scotland than it is in the rest of the UK?
That risk is difficult to quantify. There are two aspects to the cash-only economy. One aspect is traders who are below the VAT threshold and who are therefore not required to pay VAT. The model can attempt to adjust for that. Another aspect of the cash-only economy is people who are avoiding paying VAT. We have very little information about that on which to make a judgment as to whether that aspect is different in Scotland.
A number of members want to ask about technical issues, but I do not want the topic of the potential devolution of VAT to be lost in the discussion, so I will bring in Murdo Fraser, who has questions about that.
My colleague Adam Tomkins has touched on this issue. We know that VAT devolution is not an option now, so this is a hypothetical question about the situation post-Brexit. If VAT devolution was legally permissible, is it the Scottish Government’s view that that is practical and desirable?
For reasons that I think are well understood, the purpose of the Smith commission was essentially to enable the Parliament to have the powers to raise money and be responsible for how that money is spent. We used to be a spending Parliament; now we raise our own revenues. Having the power over VAT is more desirable than assignation. The issue with assignation is the level of risk that we carry, because we can never be sure that the analysis to assign an amount was fair. As I have said, the Brexit situation has added to the uncertainty.
The Scottish Government’s view is that we want as much devolution as possible, including the devolution of fiscal matters. However, that must be right, appropriate and based on the evidence. We would be supportive of the devolution of VAT. I was not involved in the Smith commission deliberations, but as a Scottish nationalist I naturally argue for the devolution of powers. It is more attractive to have the powers than for revenue to be assigned.
As I say, we have tried to be helpful here, and there is no lack of willingness. The committee is considering the risk—advisers to the committee have surely identified that the issue that we are being asked to consider is the level of risk.
The level of risk with the devolution of VAT would be exactly the same as it is with assignment, although I appreciate that there would be a greater degree of accuracy with devolution. However, the volatility of VAT receipts would be the same with devolution as it would be with any assessed survey model on assignment.
It is the volatility of Brexit right now that is adding to the general volatility. If we set that aside for the sake of the discussion, the important point about the devolution of powers is that we would be able to push and pull the levers to adapt to the policy choices. We would be able to be agile in responding to the needs of the economy and setting VAT rates that are appropriate to the circumstances. We would be able to stimulate the economy as appropriate and use the powers as an economic and social lever.
The difficulty is that the methodology is not particularly accountable or transparent, and the VAT that is raised in Scotland cannot be evidenced. Understanding the level of risk that we are carrying leads to material considerations for the budget, because we will have to provide for any shortfall if the block grant adjustment and estimates are not to our mutual satisfaction.
When the committee took round-table evidence on the issue a few weeks ago, I or one of my colleagues asked specifically about the devolution of VAT, and it is fair to say that there was a very negative reaction from the witnesses. They were concerned about the practicality of the devolution of VAT, first, because it would place a substantial additional burden on Scottish business to create a separate Scottish tax point and, secondly, because the nature of VAT is that it flows across boundaries, both within the UK and internationally. For example, a Scottish business might purchase goods or services from a business in Manchester and reclaim VAT that is paid on those goods or services. If we had a devolved system, how would we account for all of that? Do you appreciate that going down that route presents practical challenges and, potentially, a substantial additional cost to Scottish businesses?
That is why I propose to discuss the issue further with the Treasury and consider it as part of the fiscal framework review that we have agreed that we will conduct, after allowing a full session in which the Parliament has the powers in the Smith agreement. The risks have been fairly identified and we will have further discussions on them.
We will be better equipped and informed when we have a number of years of evidence on what the numbers produce and what the surveys show. Further data will be released by HMRC. If we have more information, at least we will understand the risks more fully. The alternative is to sign up to something now that might be to the detriment of the finances and the people of Scotland without being convinced about the levels of accuracy. I am minded to have further transition and further accountability around looking at those numbers and the analysis.
I hope that the current Brexit uncertainty will be resolved so that we will not be going through the process at a point of maximum volatility. I am trying to be constructive with the Treasury. It will not surprise Murdo Fraser to know that, as a Scottish nationalist, I support the devolution of fiscal powers to Scotland. However, I acknowledge the committee’s concerns about accuracy, what other commentators have said, and the consensus on the level of risk, and I propose to try to find a constructive way forward with Treasury. If we review the issue as part of the fiscal framework review, we will have more data and information. Although we recognise that the proposal was only ever for assignation and not for the power, a change in circumstances might precipitate an opportunity to revisit the discussion.
The words “level of risk” have come up a number of times. I think that Angela Constance was particularly interested in that.
Yes. I am keen to understand a bit more about what level of risk could be tolerated, given that risk can only ever be minimised and can never be eliminated.
Following on from Murdo Fraser’s questions, we have known about the potential difficulties for a while. The committee has been looking for a while at the layers of complexity, the problems with having no outturn data, and issues to do with sample size and fluctuations based on sampling as opposed to revenues. It seems to me that those are the practicalities of devolution or, rather, the complexities that we get into when we have partial devolution or partial assignment as opposed to the devolution of powers.
As well as wanting to understand what level of risk could be tolerated, I am interested in what you seek to discuss with the Treasury and what discussions you have had to date. Could you give us a flavour of that? Do the Treasury and the UK Government share some of your views about the difficulties? They are not manufactured difficulties. At their core, they are political difficulties, but some of what we are talking about is just technical stuff, of which one would hope there could be a shared understanding and a shared acceptance.
10:45
Those are appropriate questions. You asked about other views on the difficulties. Late last year, the Fraser of Allander institute said:
“Implementing a policy that exposes the Scottish budget with unnecessary risk, simply to increase the impression of accountability, is not a good way forward.”
On 6 December, the institute said:
“A key aim of the Smith Commission was to improve accountability and make Scotland’s politicians responsible for the money that they spent. Unfortunately rather than helping to deliver this aim, the current proposals for VAT assignment risk undermining that principle.”
That is the view of independent economists on what has been agreed.
Officials of the two Governments have been working together, but political sign-off at the joint exchequer committee, which involves the Treasury and the Scottish Government, is required. I have been open in saying to the committee that we engaged constructively in the work to build up the methodology to see what it would produce. In my view, it has produced a high level of risk, which Jonathan Sewell will be able to talk about. There is a high level of risk because of the lack of outturn data, the questions about accuracy and the volatility associated with the Brexit situation.
I have been briefed by civil servants on the work that has been going on in anticipation of potential sign-off at the joint exchequer committee. I have raised my concerns about the level of risk that the Scottish Government would carry and the material impact that that would have on the Scottish budget with Liz Truss, the Chief Secretary to the Treasury, in writing and in person. I did that at my most recent meeting with her, and I propose to do so again at the next finance ministers’ quadrilateral, which will be in Scotland; the date is yet to be confirmed. I have raised those issues with the Treasury as I have become concerned about them, but we have not been able to reach an agreement to take to the joint exchequer committee.
I ask Jonathan Sewell to cover the level of risk from a civil service point of view.
As the cabinet secretary said, there are areas of risk. Some are potentially more measurable than others. A lot of the work that has been done has been an attempt to identify and measure those areas, where we can. Unfortunately, there is quite a lag on some of the data.
As has been mentioned, HMRC will produce additional data at the end of the month, which should help us but, because of the various areas of risk and the differing abilities to measure them, it is hard to come up with an overall risk figure. Our way of thinking about the issue is that an appropriate level of risk should be proportionate to our ability to manage it. We are working to gain an understanding of that.
As has been said, it is part and parcel of taxation that there is volatility with any tax receipts; some taxes are more volatile and some are less volatile. VAT is not terrible in that respect. However, with the assignation process, we are exposed to the risk that is generated by the survey approach. That is more difficult to handle, because it is not based on real performance, which makes it hard to measure and address.
We are doing the best that we can to find out the information. Angela Constance is right that we have been looking at the issue for a while, but one problem pretty much across fiscal devolution is that lags tend to be associated with the data. VAT assignment involves £6 billion, which is a substantial figure. It is therefore best to take time to understand the position and not to expose the budget to significant risk.
Ms Constance asked what I am asking the Treasury to do about the issue. The transitional year was to be 2019-20, with implementation in 2020-21. Given all the issues, which I hope have been well aired this morning, I am saying that we should have further discussions and look at extending the transition period. If we understood the model more, we would know what the numbers told us. We also hope that we will have EU exit certainty—the Government’s position on Brexit and where Scotland lies in that is well understood.
We want to be in a much stronger position to understand all that before it impacts on the Scottish budget. It feels timely to discuss VAT powers at the time of the fiscal framework review, which is what I am asking the Treasury to do, rather than proceeding with the existing level of risk.
What have been the costs of the exercise to calculate the assignment? I appreciate that some costs are one-offs for the first couple of years, but how much will the exercise cost annually? More important, how much do the costs relate to the accuracy of the figures? Iain Pearce mentioned that only 725 households will be surveyed. How many households would need to be surveyed to improve the accuracy?
The total cost is £1 million, of which the Scottish Government pays half. The sum that we are assessing is about £6 billion, so you will understand why accuracy and risk are significant.
What is the relationship between the money that is being spent and the accuracy? How many more households would be needed to improve the accuracy? Has that been assessed?
We could survey more people, but the principal issue of placing estimates upon estimates in a model that is untested has never been resolved. We could pay for more surveys, but that would not address the issue of placing survey upon survey with no outturn or reconciliation, in a volatile period and without having tried the model. I do not think that my officials would say that there is any amount that could be spent to resolve those issues.
Marginal improvements could be made, but the overall accuracy would not be affected, as the cabinet secretary has said.
I will ask you to clarify a point about the 95 per cent confidence interval. Are you saying that you cannot work out a figure that gives you 95 per cent confidence? Is the level so far below that percentage that proceeding would be too risky?
The question is good but, when we are talking about assigning £6 billion, the accuracy level is really important. If there were a material impact of tens of millions of pounds on the Scottish budget, that would be significant—it would mean tens of millions of pounds that we could not spend or that we would have to find from elsewhere. What tolerance level for being wrong by would the committee accept and ask the people of Scotland to carry, perhaps unfairly? The assignation involves £6 billion. Even a tiny margin of error on either side of the calculation could lead to an outcome that was materially significant to the budget.
Would there be merit in getting agreement about a ceiling and a floor for variation, so that, no matter what scary figures you got, they could not go above or below certain amounts?
That is a good suggestion. If the committee shared the view of others about the level of risk, I would be minded to make such a suggestion to the Treasury. The Treasury might propose a mechanism that would resolve our concerns, although it has not done that so far. Perhaps it could create transitional arrangements that would mean that there was no financial detriment.
The UK Government might propose things that I would be open to, such as no financial detriment, recognising that we cannot push and pull levers here to fix the VAT issue. If the UK Government wants to respond with some sort of protection or mitigation, of course, I will engage with it constructively.
My point right now is that, short of that, not implementing VAT assignation until we have more data and a deeper understanding of what the model produces feels like the right thing to do. However, it is a good question and, if the Treasury wishes to introduce some form of mitigation as a transitional arrangement, that might be helpful and we could consider it. I would, of course, bring any such arrangement back to the committee.
The Fiscal Commission will have to build its VAT forecasts into future budgets, and I imagine that it is important that the figures that it uses are at least sort of reliable, so that they can be used.
Again, the Fiscal Commission’s forecasts are absolutely significant in relation to the whole block grant adjustment and reconciliation.
Will you be bound by that?
Yes, we are bound by those arrangements in terms of the Fiscal Commission’s forecasts.
Okay.
What discussions have you had with the Fiscal Commission? Trying to estimate this must be terrifying the people there, for want of a better description. What discussions have you had with them?
I am happy to ask officials to declare the level of terror within the Fiscal Commission. I think that it is enjoying its role, but it recognises, as the committee is well aware, that, at least where we have outturn data, there is a moment of truth—a point at which we can reconcile figures that were based on estimates. The difficulty is that you never quite get to that point. Jonathan Sewell can cover the engagement with the Fiscal Commission.
Yes, we have engaged with the Fiscal Commission—
I thought that you were going to say, “Yes, they are terrified.”
I could not possibly comment on the level of terror. The Fiscal Commission would have to offer its own response to that. We have engaged with it regularly on this matter, and the Fiscal Commission has engaged with HMRC on it. Again, it is for the Fiscal Commission to comment on this, but it agreed that it faces some of the same challenges as us in interpreting the data.
I am interested in accountability issues. The Fiscal Commission is responsible for forecasting receipts that are assigned to the Scottish budget; the OBR is in charge of producing VAT forecasts for the UK as a whole; the UK Government and the Scottish Government have jointly developed this methodology; and we have just heard lots of words about risk, data, boosted data and accuracy. Who is accountable?
I am accountable to the Parliament as Scotland’s finance secretary. That is partly why I am expressing the view that I am not convinced about the level of risk that we would be carrying. I am accountable for what we sign up to in the joint exchequer committee, and then I am accountable for Scotland’s budgets. The Fiscal Commission is also accountable to Parliament; it publishes its forecasts and it is accountable to this committee. You can challenge it on its evidence when it appears before the committee.
We work with the Fiscal Commission in a constructive way, as you would expect, not to change forecasts but to provide a constructive challenge. There are various strands of accountability in how all the different institutions work and how we operate. I think that I have been open and transparent with this committee in explaining that I share some of the concerns that you and others have raised about the level of risk. For the reasons that I have explained, that level of risk has not been resolved, and that is what is leading me to think that we should engage further with the Treasury.
To reiterate, is Brexit—and the uncertainty, instability and all the other issues around it—just another issue in the whole mix of things that is leading to a proposed delay in the process until we have better data?
Yes. This is the transitional year. We will have one year of data. The timing is absolutely not right to implement VAT assignation in 2020-21. For all the reasons that I have given to the committee, we need more time and more engagement with the Treasury to try to resolve this.
It is helpful that you have introduced the issue of transparency. Tom Arthur is also interested in transparency and public confidence in the process.
This is clearly a very complicated matter. Fundamental to the whole Smith Commission process and the Scotland Act 2016 was increasing the accountability of the Parliament. Given the challenges around wrestling with this issue and transparency, does it raise broader issues about democratic accountability and ensuring that the public understands the issue, considering the complexities of the fiscal framework?
11:00
The fiscal framework is incredibly complex. When we come to the published outturn figures on income tax, for example, at least we will know exactly what tax was paid, and we can then undertake a reconciliation with the forecasts that were made and that we were bound by. The difficulty is that we will never have that outturn or know, in the real world, what tax was collected in Scotland. My concern is about the level of risk, volatility and uncertainty and about not being convinced of the robustness of the model, although we could give the model more time to run before we make a decision on its appropriateness.
My point is about timing. If there was a major shock to the economy—if Brexit or a no-deal Brexit went ahead—that would be the worst possible time to engage in an estimation process that would never have a reconciliation. In fact, if the figures were not in our favour, it could mean that we would have to make decisions that would have a material impact on Scotland’s budget and the decisions that we would make as a Parliament. I do not think that that complexity would help with accountability or transparency.
I have a final question, to ensure maximum transparency. I noticed that, during our proceedings, a member of this committee tweeted a suggestion that VAT is a power. Just to avoid any doubt, can you confirm—absolutely, once and for all—that VAT is not a power that is being devolved to the Scottish Parliament but an assignation.
Convener, I have been paying so much attention to the questions from members that I have not engaged in Twitter while I have been giving evidence to the committee. In case someone has tweeted that VAT is a power and not an assignation, I can confirm that this would be an assignation of a tax that would be based on estimates, not a power—we have no power over VAT in Scotland.
Thank you for clarifying that.
I do not know what to say after that. We will have an investigation—that is what we will do. [Laughter.]
Have an inquiry. Find the Twitter tweeter.
Cabinet secretary, I want to understand the landscape of the discussions between yourself, the Treasury and HMRC. The fiscal framework document suggests that the discussions are held between HMRC and the Scottish Government and then go through the JEC. I do not want to go into the difference between the Treasury and HMRC, but is it HMRC or the Treasury that we are dealing with on this issue?
It is the Treasury, with regard to the politics and the agreement.
If we decide to take further evidence, that might help us to determine what would be more appropriate.
Naturally, HMRC would inform the Treasury in the same way as Revenue Scotland would inform the Scottish Government.
That is helpful.
Again, I thank our witnesses very much—particularly the cabinet secretary, who has been here for almost two hours. I also thank committee members for their questions.
11:03 Meeting continued in private until 11:12.Previous
Earnings