Our next item of business is to take evidence on the block grant adjustment. I welcome Professor David Bell and David Eiser of the University of Stirling—my alma mater, of course. Members have received copies of the executive summary of a recently published report examining the issue. I invite one of our witnesses to make a short opening statement before we move to questions from the committee.
Thank you. This seems a bit like dejà vu—it takes me back to 2012, when I wrote a report for the committee that mentioned the block grant adjustment quite extensively and covered the various issues that seem to be coming to the fore once again.
The approach that I took was the same sort of approach that the Holtham commission took, which was to look at issues of risk and incentives for the Scottish Parliament once new tax powers had been devolved. It is difficult to see the same kind of narrative as one goes through the Smith commission report. The block grant adjustment had been left somewhat as an afterthought, but it is now widely acknowledged that it is a critical part of the overall settlement. In our paper, we go through three possible mechanisms that might be used for the block grant adjustment, which I am afraid are rather nerdy. Also, there is an error—which we will correct—in one of the equations that we use to describe the different mechanisms. I am sure that the committee has noticed it. Essentially—
Excuse me for interrupting, but I am going to point randomly to a member of the committee and ask them to describe the error in the equation. They can all sit there looking nervous until I have fingered one of them for that.
We can go through the different issues in different ways, but we are finding it difficult to determine one clear and simple mechanism for arranging the block grant adjustment, while holding on to the Barnett formula, that seems simultaneously to satisfy all the so-called principles in the Smith commission report, such as taxpayer fairness and the no-detriment principle, which we may well end up discussing. Those are the general issues that we see, some of which our paper attempted to elucidate. One of our main conclusions was that it was very difficult to satisfy all the principles that Smith apparently set down.
It is an issue of risk and reward. Scotland will not necessarily do significantly worse as a result of particular mechanisms. Our paper shows that Scotland would have done relatively well with one of the mechanisms. Had the block grant adjustment started in 1999, Scotland would be more than £1 billion better off, rather than worse off. We have seen some stories in recent weeks about how Scotland might be worse off. It is a question of what risks Scotland faces and how the Scottish economy performs, particularly in relation to the amount of tax revenues that it generates.
I thank you and David Eiser for an excellent paper. Most of the questions that I would have asked are answered in the paper but, for elucidation, I will ask some of them anyway. In the second paragraph of the executive summary, you say that
“the precise way in which the”
block grant adjustments
“are indexed over time could mean differences of over a billion pounds a year in the Scottish Government’s budget after a relatively short period of time.”
I take it that the short period of time that you are talking about is 16 years. Thank you for your elucidation of that point in your introduction, because I was going to ask over what period you had assessed it. You also say that
“analysis suggests that the Smith Commission’s principles may not be workable and are not mutually compatible.”
David Eiser can add to what I am about to say. I was at the committee about a month ago discussing behavioural responses to—
—the Scottish rate of income tax.
Yes. On whether the evaluation of behavioural responses, which may form part of the imposition of the no-detriment clause, can be consistent with simple, mechanistic, year-on-year adjustments to the block grant, which will not require a great deal of negotiation between the parties concerned, I would say that that does not seem to add up.
A number of different principles are set out in the Smith commission agreement, some of which are not particularly compatible in a number of ways. David Bell spoke about the principle of no detriment in relation to policy decisions, which recognises the interdependencies between the policy decisions of one Government and the revenues of another and implies that there should be some sort of compensating payments between the Governments where a policy decision by one Government affects the revenues of the other. Working out what those compensating payments might be would be subject to a great deal of contention and uncertainty. Satisfying that principle at the same time as satisfying the principle that the fiscal framework should be simple, fair and largely rules based will be a challenge.
There are other sets of tensions between some of the no-detriment principles. For example, there is the principle that there should be no detriment from the decision to devolve. There is likely to be tension between the block grant indexation mechanism by which we might choose to achieve that principle and the no-detriment principle that relates to taxpayer fairness and how a tax rate change by the UK Government might affect the revenues through the block grant that is available to Scotland.
Thank you for that. You have said that, in future years, the block grant adjustments will have to be indexed to account for inflation and likely economic growth and that they should be indexed appropriately.
In your submission, you refer to the three methodologies that could be used—indexed deduction, per capita indexed deduction and levels deduction. You go into some detail on the pros and, in many cases, the cons of those basic approaches.
I take it from the analysis in your submission that although we do not know whether the per capita indexed deduction method would be the best method going forward, it would seem on balance—certainly based on figures for the past 15 or 16 years—to be the best methodology. Is that a reasonable suggestion?
That method has some attractions. Again, to use the language of risk and reward, it takes out the risk of differential population growth because it works on a per capita basis. One could make the case that Scotland does not control immigration policy and therefore does not have the political levers to fully affect the size of Scotland’s population; a case could therefore be made for making per capita adjustments on that basis.
The third method is the so-called levels method, which is the most symmetric as far as the Barnett formula is concerned. As you know, under the Barnett formula, Scotland gets its population share of spending increases in Westminster.
According to the levels method, the block grant adjustment would increase by Scotland’s population share of the increase in tax revenues at Westminster. That is symmetric with the Barnett formula in the following way. If there was an increase in comparable spending of say £100 million south of the border, which was financed by an increase in income tax of £100 million south of the border, the net effect on the Scottish budget would be zero. Scotland would get its population share of the spending increase, which would be offset by its population share of the tax increase. They would be the same—it would be £100 million in each case. There is symmetry there.
The reason why Scotland might well do worse under that method is that Scotland’s per-head level of tax receipts is lower than levels in the rest of the UK. We keep getting adjustments to the block grant adjustment that do not compensate in the same way as the per capita indexed deduction, which is measured over the whole grant, as opposed to the change in the grant.
One way of thinking of it is that the levels method increases the rate of convergence, which we talk about in relation to the Barnett formula. The levels method would increase that rate of convergence towards level spending per head in Scotland and in the rest of the UK, as opposed to the current situation where per capita spending per head is somewhat higher in Scotland than it is in the rest of the UK.
You say in your paper:
“There is a lack of consensus, or even debate, about the type of fiscal risks and incentives Scotland ... should face.”
That is why we are discussing these issues now. You continue:
“Should Scotland face risks associated with relative population change? Should it face risks associated with differential demographic or economic change, over which the Scottish Government might have only limited control. Without such a debate it is difficult to recommend a particular form of BGA as these are intimately linked to the type of fiscal risks and incentives a Scottish Government will face.”
However, you are suggesting that the per capita indexed deduction method is the least worst option.
If you were to assume that, for the foreseeable future, population growth is likely to be slower in Scotland than in the rest of the UK, you would probably go for per capita indexed deduction as the mechanism for indexing your block grant adjustment, because it protects you from the risk that your population grows relatively slowly.
I suppose that one of the arguments against that is to do with how the Barnett formula works on the spending side. It effectively rewards Scotland for relatively slow population growth, because it only adjusts for that relatively slow population growth on the margin of the new spend. The rate of convergence that we would expect to get through the Barnett formula is offset on the spending side if Scotland’s population grows more slowly. Why, then, should we have per capita indexed deduction protecting Scotland from relatively slow population growth on the revenue side?
Indeed. As you say in your paper, that would perhaps mean that there would be no incentive for Scotland to grow its population. Even if Scotland tried to grow its population, London would still remain a magnet for people across the world, and it would be difficult to compete with it as an entrepot. That is not so much about the rest of the UK; it is about London in particular.
You have also said:
“Scotland’s Barnett-determined block grant will clearly need to be adjusted to reflect both the new tax-raising powers and new expenditure responsibilities being devolved.”
I want to press you a wee bit on how that should be adjusted.
12:00
Given that £2.5 billion will be added to Scotland’s budget to cover the new responsibilities in relation to welfare, there is a question about how that should be adjusted. There are different potential approaches. Do we adjust on the basis of population or do we make some adjustment for need? The latter would be pretty difficult, given that a variety of different needs are covered by the different welfare powers that are being transferred, such as the carers allowance, attendance allowance, disability living allowance and so on. You could try to fine tune the adjustment for the new welfare powers in relation to the particular benefits that are being covered, but if Scotland decides to have a different benefit system, it will be increasingly difficult to figure out what we call the counterfactual—what would have been spent in Scotland on the powers that are being transferred if they are no longer being used by the Scottish Parliament.
I know that you are being cautious and putting out various options but, if you were making the decisions, given what is going to be further devolved, what would your approach be? When I say “you”, I am referring to the whole panel.
It is unfortunate that David Phillips is not here. Do you have to be called David to be a member of your group of authors?
Yes.
I thought so.
Again, it is partly a question of trade-offs between the different principles in the Smith commission agreement and which principles are thought to be the most important. Another question is consistency in adjusting the block grant in relation to tax.
If the block grant adjustment for devolved taxes is to be indexed on the basis of indexed deduction—that is, the percentage rate of growth of tax revenues in the rest of the UK—there is an argument for using the same approach to index the block grant adjustment in respect of the devolved welfare powers. That means that the block grant adjustment for the devolved welfare powers would be indexed on the basis of the percentage change in total spending on those welfare benefits in the rest of the UK.
There is an argument for symmetry in how the block grant is adjusted for tax and how it is adjusted for welfare, but there is also a question of trade-offs between the principles in the Smith commission agreement that are thought to be the most important. That influences which indexation method is thought to be the most appropriate.
On the spending side, Scotland’s block grant would be adjusted on the basis of changes in spending down south. If you go for the per capita indexed deduction for the block grant adjustment, you are adjusting on the whole of tax revenues south of the border. As David Eiser said, there is already asymmetry in those approaches. If something different again was introduced for welfare powers, you would end up negotiating over different mechanisms for different bits of the budget, which would undoubtedly slow the agreement process.
As if the process was not slow enough the last time a block grant adjustment was made.
You said:
“The Smith Commission says that the UK government should bear ‘economic responsibility’ for its own policies, and the risk of any shocks that affect the whole of the UK, and the Scottish Government should bear the ‘economic responsibility’ for its policies”,
but that the Smith commission
“says nothing about who should bear the underlying revenue or spending risks in Scotland.”
This is taking a step back, but it is probably worth remembering that the block grant adjustment in year 1 is simply the level of taxes that are raised in Scotland in that year. We want to index the block grant adjustment over time to some measure of how taxes change in the rest of the UK.
We want to index to the changes in those tax revenues in the rest of the UK to protect Scotland from fiscal shocks that hit the whole UK. If revenues fall in the rest of the UK, there will be an offsetting reduction in the block grant adjustment. The argument is that, over time, if Scotland’s revenues grow relative to the counterfactual—that is, what we would have expected to happen to Scotland’s revenues had they not been devolved—the Scottish budget will benefit from the difference between the block grant adjustment and the actual growth in revenues in Scotland.
The paper makes the point that it will not be possible in reality to ever know to what extent differential growth in Scotland’s devolved revenues is due to specific policy decisions of the Scottish Government that have caused revenues to grow above the block grant adjustment and to what extent it is due to factors that are beyond the control of the Scottish Government or the UK Government. There is an area that falls between risks that are insured by the UK Government and rewards that are directly due to Scottish Government policy.
I will touch on one final area—borrowing—before opening up the session to other committee members. David Phillips previously pointed out to the committee that the method for indexing the adjustment of the block grant
“will have major implications for the scale of current borrowing powers required.”
Will you talk to us about the relationship of the three potential indexing methods to borrowing?
The issue as far as borrowing is concerned relates to uncertainty about the forecasts for tax revenue growth in the rest of the UK. If that is forecast badly, the block grant adjustment will be forecast badly. The consequence might be that the Scottish budget is out of line with what the reality turns out to be.
The question then is whether it is more difficult to forecast the change in taxation or to forecast the overall level of taxation. You would think that they must be the same, but uncertainty is associated with tax revenues in the UK as a whole, even a year and a half after they were supposed to have been collected.
Under the indexed deduction or per capita indexed deduction method, the size of errors could be greater. The consequence is that greater borrowing powers would be needed. We would need to consult a spreadsheet and make assumptions about the kinds of errors that might turn up before we could commit to a number. I worked out an example in a paper for the committee when we discussed borrowing powers earlier this year. It showed that the proposed powers to borrow £300 million to cover current revenues, which is essentially what we are talking about, would not have protected Scotland at the time of the financial crisis. The implication is that the change in the block grant adjustment would have been substantially bigger than that.
I open up the session to colleagues around the table.
Gentlemen, I ask you to imagine that you are speaking to a layperson—somebody who might not have immediately identified your inaccurate equation. It seems that the concerns that we have seen in the media about the impact on the Scottish budget have related to the indexed deduction, as you describe it in your paper, and that the dangers and risks of that are substantially, if not entirely, compensated for by the per capita indexed deduction. Whatever the merits or otherwise of that approach, there is a lot of protection for the Scottish budget in it.
As David Eiser said, it is a question of which risks Scotland should take. I went back to Holtham’s paper on the initial discussion about the block grant adjustment for Wales. It clearly considers different risks and the implications for the block grant adjustment. It is regrettable that the Smith commission did not have sufficient time to examine risks and incentives as far as the Scottish Parliament and Government are concerned.
In the indexed deduction, we take the risk that Scotland’s tax revenues will grow more slowly than those in the rest of the UK. They could do that if productivity growth per capita in Scotland was, say, less quick. We do not know that that is the case, but we might surmise that it is; it is a possible forecast. We also take the risk that the number of taxpayers in Scotland could grow more slowly than in the rest of the UK.
With the indexed deduction method, we have both those risks. If we do the per capita indexed deduction, we take out the population part of that and are left with only the question whether revenues per head will grow more slowly or more rapidly in Scotland than they do in the rest of the UK. Therefore, it removes part of the risk.
The population projections show Scotland’s population growing more slowly than the rest of the UK’s up to 2060. Population projections are not always right, but they are indicative. As I said earlier, we could certainly argue that Scotland does not have full policy control over the size of its population.
We are debating future tax levels in Scotland with the new powers, but there has been little indication from the Scottish Government or others of whether, with those powers, there will be any big move to cut tax rates to attract more people into Scotland. The committee has heard evidence that it would be difficult to guarantee such a population flow into Scotland by cutting tax, because the international evidence seems to be sketchy about whether such changes make people behave in that way.
I go back to the evidence that I gave about a month ago on the behavioural response to changes in tax rates. It is not clear that the powers under the Scotland Bill would allow Scotland to follow the kind of policies that seem to have an effect. It is also difficult to translate lessons that are learned in other cultures, where there are different attitudes to taxes, into something that allows one to make a strong prediction about what would happen in Scotland. On that basis, the Scottish Government should be cautious in the first instance.
Given that, and given that the Barnett formula currently protects Scotland for its lack of population growth relative to the rest of the UK, what are your further thoughts on why the per capita indexed deduction system does not pass the test of taxpayer fairness?
Further on in your paper, you discuss reforming the Barnett formula with a formula akin to the PCID approach. You also state that none of the systems
“fully satisfy ... ‘taxpayer fairness’ ... in all circumstances.”
Why is that true specifically for PCID?
12:15
David Eiser can speak about taxpayer fairness, and then I will come back in.
I will try to. The issue of taxpayer fairness that we discuss in our paper was raised in the Smith commission report, which argued that, if the UK Government were to change income tax rates once they had been devolved to Scotland, the change in rates in the rest of the UK should have no bearing on spending in Scotland.
As we have touched on, the Barnett formula operates by calculating the change to Scotland’s grant as a population share of changes in levels, whereas the indexed deduction and the per capita indexed deduction both work on percentage changes. Ultimately, that means that there is not a balance. If the UK Government were to increase income tax rates and spend that money on comparable services in England, the Scottish Government would benefit through the Barnett consequentials by more than the amount by which its block grant adjustment would be increased, because of the increase in tax revenues in England. That is because of the difference between a levels approach on the spending side and a percentage approach on the revenue side. There is an issue with the taxpayer fairness principle in that respect.
Would that approach be unfair to the rest of the UK, in effect?
Yes.
Fine. That may not be the—
I am sorry—I will just clarify that. If tax revenues were increasing in the rest of the UK, the situation would be unfair to taxpayers in the rest of the UK, but things would work the other way round if there was a cut in income tax rates in the rest of the UK.
Right.
I will make one point in relation to the Barnett formula. The adjustment methods that are being proposed are contingent on the existence of the Barnett formula. There must be a block grant that is determined in the way that it currently is in order to have a potential set of adjustments to that grant.
As I think that Jim Cuthbert and Anton Muscatelli have argued, there is a possibility of a big divergence between where Scotland might have been if the Barnett formula had just continued as is and the situation that would arise if the adjustments were to take place. We have shown that, in recent history, Scotland has done relatively well, but I am not necessarily arguing that that predicts where it might go in the future. It all depends on relative economic performance. The argument is that, potentially, there is no end point to the adjustment. Jim Cuthbert made the point fairly clearly that Scotland’s block grant could, in certain circumstances, wither away entirely.
None of the modelling that has been done so far has captured the following point. Let us suppose that tax revenues in Scotland were growing substantially and consistently less rapidly than those in the rest of the UK. Scotland’s grant would be falling but, in the Scottish economy at large, real wages would not be keeping up with those in the rest of the UK. There would come a point when investors in the rest of the UK noted that, as Scotland was a relatively cheap place to invest in, they could increase their investment here. To some extent, that would offset the difficult scenarios that other commentators have portrayed.
So we are in a world of multiple hypotheticals.
Yes.
Indeed. We are now a small but perfectly formed committee of five, given that Jean Urquhart has abandoned ship.
You have set out three options—ID, PCID and LD. To what extent have the three of them been tried anywhere else, at any point? If they have been tried, to what level of success or otherwise have they been utilised in other scenarios and countries?
David Eiser can chip in on this, but my understanding is that, because no one else has anything that is equivalent to the Barnett formula, there is no equivalent to the block grant adjustment.
Most other jurisdictions work on some measure of ensuring that parts of the state that are relatively poor can provide a given level of public service. They do that in two ways. First, they make transfers to increase the spending power of the poorer areas. Secondly, they adjust tax revenues where the taxable capacity of an area falls below that of the state as a whole.
The Barnett formula is different from those mechanisms. However, in recent months, a potential floor on the Barnett formula has been introduced in relation to Wales. That says that, because the Barnett formula ultimately means convergence to equal spending per head across the UK, convergence has been very slow. However, as far as Wales is concerned, convergence will not go below a given level. That says implicitly that Wales has a certain level of need relative to England, Scotland and so on. In that instance, the Barnett formula is being adjusted for need; at the minute, it does not do that at all.
That was a very long-winded answer, which could have been made much more simple by saying that nobody else has the Barnett formula, and there is nothing else like the indexation methods that we are talking about.
I would describe that answer as interesting, as opposed to long winded. Does Mr Eiser have anything to add?
No—I would make the same points.
That is helpful. So, basically, there is no precedent and we are in uncharted territory.
An agreement will be reached at some point. We are told that that will happen after the Scottish budget process. I do not know whether that will be in January, February or thereafter, but an agreement will be reached. Presumably, the preference is for the agreement that is reached to be of at least medium-term duration so that we are not revisiting it every year.
Once an agreement has been reached, would there be merit in having a system in place whereby every 10, 15 or 20 years—whatever the number of years ought to be—there is a formal review of the mechanism? Obviously, doing that every year would defeat the point of devolution, but would there be merit in having a longer-term system of formal review?
There is certainly a case for periodic review, although quite how that would be written into legislation I am not sure—that is not what I am an expert in.
It seems to me that intergovernmental arrangements will have to be ramped up to make the system work. There are cogent arguments around whether the Treasury should be both judge and jury in determining what the block grant adjustment will be. It sets the statement of funding policy, which has had a significant impact—David Eiser has written about this—on Scotland’s DEL allocation for the next four years, as was announced in the spending review last week. There have been changes in non-domestic rates. It is a Treasury-only document that determines the block grant, and that has knock-on implications for the block grant adjustment.
Aside from getting an agreement on the indexation mechanism, there has to be an agreement about how it will be administered and what level of intergovernmental co-operation is necessary to make it work effectively.
On page 5 of your executive summary, you say:
“The ‘compensation principle’ set out by the Smith Commission will be impractical to fully implement.”
That is perhaps something of an understatement. Are we kidding ourselves that it can actually be implemented?
You speak about
“isolating the effect of government policies”
and trying to work out and consider the knock-on effects and so on. To some extent, I do not think that the compensation principle has been discussed as widely and deeply as it might have been across the UK. Do you think that there is some way of implementing it, or are we kidding ourselves?
For a given policy change by one Government, there could be any number of knock-on effects. Were the Scottish Government to increase income tax rates, that may affect RUK revenues on non-devolved taxes in Scotland, or it could affect spending on non-devolved benefits in Scotland. The real question is about how far we want to take the compensation principle.
We cannot imagine that we can estimate and calculate the effects of every knock-on effect imaginable, no matter how small, although there is possibly a case for saying that particularly large effects could be calculated. There will always be contention around such effects, even if there are some pretty major ones.
Given the trade-off between that no-detriment principle and some others, there is a decision about how important the no-detriment principle is. It is particularly unusual in the context of other devolved systems. There is an argument for saying that it undermines some of the point of devolution if we take that principle very far, so perhaps we should not take it very far and we should prioritise some of the other principles.
The House of Lords Economic Affairs Committee took evidence from Mr Cottarelli from the IMF. He studies fiscal federalism across the world, and he was asked whether the no-detriment principle had a parallel elsewhere. He could not readily find an example of a similar approach.
As David Eiser said, it seems almost impossible to follow through all the ramifications for the two jurisdictions of one party changing taxation or spending, along with the implications for other taxes and the knock-on effects on benefits and so on throughout the UK. It would be well beyond the existing expertise in the OBR, for instance, to follow the money through that.
I take it that you are not planning to dedicate a paper to that single issue?
Not immediately, no.
Thank you.
12:30
This is quite a complex area. You mentioned that the Barnett formula was intended to bring convergence to expenditure per head throughout the UK, although that has not happened. Is that because the Scottish population has not grown in line with that of the UK, which was not foreseen at the time?
No, I do not think so. When the Barnett formula was first thought of in 1979, it was felt that it roughly represented levels of need in different parts of the UK. I am not sure whether it was a rough rule of thumb that seemed convenient at the time or if people had thought through its convergent properties. Nevertheless, it is true that, particularly as a result of Scotland’s slower population growth, the convergence has been very slow. The same is true in Northern Ireland, but perhaps less so in Wales, which is now closer to the UK average.
When we say, as the Smith commission did, that Barnett is being continued, are we referring to the literal mathematical side of it or also to the principle of Scotland being rewarded for its population not growing?
By retaining the Barnett formula, we are keeping both the mathematical property, with everything that that entails about relative population growth and convergence, and the existing starting point. That is what I interpreted the Smith commission to mean in saying that we will retain the Barnett formula as the starting point and adjust from there.
The same calculation, which is Barnett determined, will be done each year. That implies convergence over time and the calculation is also adjusted by the issues that we have talked about today. The principle of Barnett remains.
Towards the end of the executive summary to your paper, you refer to
“what the UK’s fiscal union is for”
and to having
“a proper debate about these issues.”
Is that in the longer term, beyond where we are at the moment and the block grant adjustment?
Yes. There is no time to get those wider issues debated sufficiently to influence the current tide of events. If the arrangements in the Smith commission report, which then fall into the Scotland Bill, receive the consent of the Scottish Parliament, they will somehow have to be made to work.
Would I also be right in saying that, whatever method we choose, depending on what happens to the relative populations, in one scenario that method would benefit Scotland and in another scenario it would benefit the rest of the UK? Is that almost inevitable?
Yes; absolutely. In the short to medium term, the changes in relative population take place at a relatively slow pace. It is by no means as variable as, say, economic growth.
Therefore, depending on relative economic growth, one or other part of the UK will benefit.
Yes. That is so in the short run, unless we have massive short-run changes in population, which I do not think that we envisage.
You talked about the tax per head in Scotland being lower than in the UK and higher rate taxpayers tending to live in the south-east of England and so on. If either the UK or Scotland becomes flatter or more equal—whatever the term might be—would that have an impact on how Barnett works? One of the arguments that we have been given is that if we make tax fairer, we will get less tax, because the rich people pay the most and if we cut their incomes, we will get less tax—even if it were shared out more equally. Is that part of the equation? If we were to make Scotland fairer, would that inevitably hurt our tax base and therefore affect the Barnett formula?
I do not think that it is inevitable. That is part of the responsibility that comes from the devolution of the taxes. Once income tax has been devolved to Scotland, it will be up to the Scottish Government to have whatever tax rate structure it wants. If, by changing the tax rate structure, it increases its revenues relative to the block grant adjustment, its budget will be in a better position than it would have been without the devolved revenues. If, on the other hand, it changes the tax rate structure and as a result, the revenues are lower than the block grant adjustment, its budget will be worse off.
There is no guarantee that any particular tax rate change will result in any given revenue change. That is not what we have modelled in the report. The point is that the power—the responsibility—is devolved, so that you get the rewards for whatever tax policy you choose to implement. There may well be a trade-off between the tax policy that maximises revenue and the tax policy that achieves the level of redistribution that you want. That decision will have been devolved.
Thank you.
The convener quoted from your executive summary, in which you say:
“The Smith Commission says that the UK government should bear ‘economic responsibility’ for its own policies”,
and that the same should apply to the Scottish Government. Does that make a difference between an assigned tax, such as VAT, and a partially devolved tax, such as income tax? Is it harder to measure? We could vary income tax differently from the UK, which I assume would make it harder to measure, but could we say that almost everything that happens to VAT is the responsibility of the UK Government, or is that not clear?
VAT is an assigned tax, so Scotland gets its share of 10p-worth of the VAT revenues from Scotland, but it has no power to vary the structure of that tax—that power rests entirely with the UK Government. The UK Government will have to bear the consequences of any changes as far as VAT is concerned.
As you suggest, income tax is more complex, because Scotland can take decisions over which it has to assume responsibility and the UK Government can make its changes and it accepts responsibility in relation to those. You must then think about the interaction between changes that are made at the different levels. That is where the no-detriment principle comes in and trying to calculate whether the other Government has been made better or worse off as a consequence of any changes.
So VAT is more closely linked to what the UK Government does than income tax. Ultimately, if the UK Government mismanaged the economy and VAT fell, for various reasons, should the UK Government compensate us for the reduction in VAT revenue, or do we just get exactly half of what the UK Government collects?
There is an added level of complexity with VAT. When we have been talking about the block grant adjustment in relation to income tax, we have been talking about tax revenues, whereas when you talk about the adjustment for VAT, you are probably talking about a tax base, for the reason that Mr Mason has described.
For example, if the UK Government were to cut the rate of VAT, you would not want there to be an impact on the Scottish block grant, because the rate of VAT would also have been cut in Scotland—Scottish taxpayers would also be experiencing the lower rate of VAT. Therefore you would not want the adjustment to be based on revenues. By making the adjustment on the tax base, rather than on revenues, you would avoid that linkage, because the base does not change when the VAT rate changes.
That is helpful, thank you. My final point is that we are struggling at this committee to really understand all this—
Speak for yourself.
—apart from the convener. You folk are experts. It seems to me that the public are very unlikely to really get into the nuances; I do not mean to insult them. Must we choose some fairly simple method, accept that it will not be perfect and then, as Gavin Brown suggested, review and revise it every five or 10 years? That is not as pure an answer to the issue as I would like, but in practice do you think that that is where we have to be?
I certainly suspect that that is the case. I think that by making the fiscal framework a sort of addendum to the Smith commission report—a relatively minor part of the overall report—you are not exposing the Scottish public to any real debate about the issue that, it seems to me, this should all be about. That is the question of the balance of risk and incentives. Should Scotland be protected from an economic shock that hits the whole of the UK? Should it take the risks and the rewards that are associated with the policy decisions of the Scottish Government?
If the debate had been framed in that way from the outset, rather than it just being a question of going through a list of taxes and saying, “You have this, you have that,” and doing the same for welfare benefits, people would have had a better understanding of what is going on. That is not to say that it would have been simple.
I agree that you will probably have to make do for a certain period of time and then review the situation a bit down the line; I do not know how far down the line—maybe five years.
That is helpful; thank you.
I hope that we will end up with the best method rather than the simplest one.
Most areas have been touched on. You have spoken about the second no-detriment principle. I think that most people have now accepted that that principle is extremely complicated to fathom, if not virtually impossible, given how far into the future we can project behavioural responses to a tax change, and how much we can directly link it to that decision versus other factors.
However, your paper seems also to cast doubt on the first no-detriment principle—the principle of no detriment at the point of devolution—particularly where it says that we cannot, in essence, have that and the element of taxpayer fairness. Do you want to expand on that?
The two principles are no detriment from the decision to devolve, and taxpayer fairness.
To address the no detriment from the decision to devolve principle, we have an indexation method that, in effect, tries to ask what the UK Government would have raised from the taxes that have been devolved had they not been devolved—like a counterfactual. It seems to us that the most sensible way of producing that counterfactual is to base it on some measure of the percentage growth in revenues that are devolved to Scotland. For example, if the income tax revenues grew by 10 per cent or by 10 per cent per capita in the rest of the UK, there would be an argument for saying that we would expect those revenues in Scotland to have grown by 10 per cent or by 10 per cent per capita—whichever of those methods we choose.
12:45Take a simple example: assume that there is one taxpayer in England and one taxpayer in Scotland. The taxpayer in England pays £100 and the one in Scotland pays £90. That is the kind of situation from which we start. If revenues in England were to increase by 10 per cent, the figure would go from £100 to £110. If we were using one of the indexed deduction methods—either the per capita indexed deduction or the total indexed deduction—that would imply that we assume that the revenues in Scotland would also grow by 10 per cent, so we would go from £90 to £99. The adjustment would be £9. However, if we used the levels deduction method and the increase in England was from £100 to £110, the population share of that increase would be £10 for Scotland. We get a difference between what the adjustment would be if we used the percentage approach, which would be £9 in Scotland, and if we used the levels approach, which would be £10 in Scotland. That is what we mean when we talk about the tension between percentages and levels and why we get a trade-off between the two principles.
What is better to use as an assumption about how Scotland’s revenues will grow over time? It seems that the percentage principle is the fairer. The levels approach seems to be a bit unfair to Scotland because it assumes that, in percentage terms, revenues in Scotland have to grow by even more than they do in England just to keep up. However, the levels approach gives us symmetry if there is a tax-rate change in England that funds spending on comparable services.
The passage that leaps out from your submission—it is in bold, so I suspect that it is designed to do so—is the one that says:
“it is impossible to design a block grant adjustment system that satisfies the spirit of the ‘no detriment from the decision to devolve’ principle at the same time as fully achieving the ‘taxpayer fairness’ principle”.
You have described the tension that exists. We are all fumbling in the dark until we see the detail of the fiscal framework, but do you get the impression that the tension that you highlighted is on the radar in the discussions that have been taking place?
I have no doubt that it is on the radar.
We have seen the messages from the likes of Professor Muscatelli and the Scottish Trades Union Congress about the potential pitfalls, depending on how the fiscal framework is drafted. Do you get the impression that work has been done, or is being done, in the fiscal framework discussions to offset pitfalls? I am particularly interested in work that is being done from the Treasury perspective, as most of the commentary that I have seen was about potential detrimental impact on Scotland, rather than on the Treasury.
Obviously, we are not party to the discussions. I have no doubt that plenty of analysis of the implications for the Scottish budget and the Treasury under various different scenarios is going on on both sides. It gets complicated quite quickly because, depending on what we assume about how population changes relative to how revenues and spending change, we get different answers about which adjustment mechanism might favour the Scottish budget or the Treasury’s budget over a period. Quite how all that analysis will come together to arrive at a chosen method, I do not know.
The exercises that we have done show that the outcome may depend quite a lot on the year when we start the calculation. There is an argument for considering an average over a two-year or three-year period, initially. If the no-detriment principle applies at the point of devolving the tax, the decision could be taken to average that out over two or three years. Then, I think that—
I am sorry to interject. You are quite right that the received wisdom has used the point of devolution. Using the point at which the decision to devolve is made sets things slightly differently. In your paper you say that there could be, depending on the indexation that is used, an attritional factor over the longer period, which could be construed as part of the decision or the point of devolution if there is an acceptance, at that stage, that the indexation that is used will lead to that attritional factor. Could you trace that back and say that that violated—for want of a better term—the no-detriment principle at the point of devolution if you knew that that attritional factor was built in?
If it is possible to do a counterfactual from the year when devolution takes place, it will always be possible to do the counterfactual from alternative years round about then, or from whatever point in the future. That will cause one party or the other to regret that the decision to devolve was not made at a point in time that was more favourable to their particular case.
It is difficult to know how the negotiations will go—obviously, they are closed. It is not clear that there is any agreement about what should be the underlying principles on which the decisions that govern the design of the block-grant adjustment are based. I guess that, as David Eiser said, both parties are trying to make the best guess about which index would give the best outcome for their particular budget—the Treasury or the Scottish Government—over the medium to long term.
That reinforces the point that, without a set of principles guiding that decision and if—horror of horrors!—a compromise were to be reached between the levels method and the per capita index deduction method, the matter would probably need to be revisited in the medium term, say after five years, to ascertain where the mechanism is leading as far as public spending in Scotland and the growth of the Scottish economy are concerned. We would all, however, struggle to understand what such a compromise might be.
I entirely take your point about the counterfactual. You were all blessed with 20:20 hindsight on these matters.
You have comprehensively outlined the risks that are attached to certain types of indexation and the likelihood of detriment, depending on which indexation is applied. It could therefore be reasonably assumed that, if the type of indexation that was being applied was known at the point of devolution, and if there was a comprehensive analysis saying that, as a result of that indexation, there would be an attritional detrimental element to the revenues, the no-detriment principle was, arguably, not being complied with at the point of devolution. I realise that there are some caveats built into that. One could always factor in future counterbalances, such as improved economic performance or unexpected population growth, which could spike that. If it were known at the beginning that there would be potential detriment if nothing changed and the indexation were applied as a rule, that could be construed as the point of devolution.
That is an interpretation of no detriment and it is a line that could be pursued. There is nothing against that argument, in principle. That interpretation of what the second no detriment principle means does seem—
I am referring to the first principle. We have worked out that it is virtually impossible to come to come to a conclusion on the second principle.
If, hypothetically, we get the income tax powers, according to your paper,
“Scotland does lose out somewhat under the ID approach”
but
“under the LD approach it loses out even more”.
Your paper suggests that Scotland will lose out if those methods are applied.
That will depend on how circumstances evolve. Under the levels deduction method, tax revenues fall over time. That is likely to be the best method for Scotland. We are trying to select a method up front and then allow Scotland to retain whatever revenues it raises over the baseline that is determined by that method. It would be difficult to look backwards and say that, because Scotland’s revenues have not done as well as they would have done under a counterfactual, we are going to try to recoup some of that through a no-detriment argument. That would defeat the purpose, which is to decide on an indexation method, and agree that, if Scotland’s revenues grow by more they can be retained, and if they grow by less its budget suffers the consequences.
My interpretation of what you said is that, if in retrospect the adjustment that was made in year 1 was incorrect because of what has happened to the Scottish economy or to Scottish tax revenues, we should have had a different adjustment and should be compensated for that over time. That argument would be symmetric: the Treasury might say that the method had not worked out in the way that it had expected, that it was falling short on revenue and therefore wanted to be compensated for the start point. That would be very difficult to evaluate and would have the potential for further contention.
Thank you. You are saying that we cannot have our cake and eat it.
Professor Bell has said a number of times that Scotland has to decide what risk it should take. We are not, however, in a bilateral situation like the Basque Country in Spain is, as we found out when we were there. Ultimately, the Treasury will take the decision. Although the Scottish Government will negotiate with the UK Government, we do not have the whip hand. We hope to get a system that has been negotiated, but that will not necessarily be the case.
The case has certainly been made in the Finance Committee and other committees that there ought to be a third party involved that does not have an iron in the fire and that can look at the issues dispassionately. Its role should be accepted as being objective.
If you have no further points to make, we will wind up. It is now five seconds from one o’clock, so it is probably a good time to finish our meeting: that is the end of our public deliberations. Thank you to both our witnesses.
13:00 Meeting continued in private until 13:01.Previous
Subordinate Legislation