Finance and Public Administration Committee

Meeting date: Wednesday, December 20, 2023

Official Report 633KB pdf

---

Budget Scrutiny 2024-25 and Economic and Fiscal Forecasts

Good morning, and welcome to the 34th meeting in 2023 of the Finance and Public Administration Committee.

The first item on our agenda is an evidence session with the Scottish Fiscal Commission to discuss the Scottish Budget 2024-25 and the commission’s December 2023 economic and fiscal forecasts, which were both published yesterday. We are joined by witnesses from the Scottish Fiscal Commission: Professor Graeme Roy, who is the chair and Professor David Ulph, who is a commissioner. Claire Murdoch, who joins us online, is head of fiscal sustainability and public funding. She is joining us remotely, sadly, due to Covid. However, I am glad that she is able to participate in the meeting. Welcome, everyone. I invite Professor Roy to make a brief opening statement.

Our report, which was published yesterday, said that economic growth will remain fragile in the near term and, as indicated last year, we believe that living standards are set to fall between 2021-22 and 2023-24 and will not return to pre-cost of living crisis levels until 2026-27. Inflation is expected to remain above target for longer than we forecast last December.

Against that backdrop, the Scottish Government’s budget is set to increase next year by £1.3 billion from the latest figure for 2023-24, which is an increase of 2.6 per cent in cash terms, or 0.9 per cent after inflation. Most of that increase comes from income tax. An element of that increase follows from the new tax band for those who are earning over £75,000 and an additional penny on the top rate, but a larger part comes from the relatively strong earnings growth that we are seeing at the moment and the fiscal drag that that is leading to from the fixed key cash thresholds. We predict continued significant growth in the number of higher-rate taxpayers in Scotland, many of whom will be seeing an increase in the proportion of their income being subject to higher rates. However, despite those funding increases, there is little space for new spending once we take account of pressures such as the on-going consequences of changes to devolved social security payments and the linking of payment rates to inflation. Other policies, such as the council tax freeze and public sector pay, have further added to existing spending commitments.

I understand that a lot of the attention is focused, as mine has been so far, on the next financial year, so I would like to quickly extend the discussion to the remaining four years. Between 2023-24 and 2028-29, total funding for both resource and capital is expected to increase by 14 per cent in nominal terms and 4 per cent in real terms. However, within that, capital funding is expected to fall by 20 per cent in real terms over the same period. I think that it is useful to look at what is determining the outlook for growth in resource funding and some of the risks that are associated with that. Beyond 2024-25, the resource block grant from the United Kingdom Government is expected to grow by just 1 per cent a year in real terms on average, and the capital block grant is expected to fall in real terms, all in line with the UK Government’s medium-term spending plans. Crucially, the UK Government has not set departmental spending plans for 2025-26 onward, despite those overall totals. There are risks for the Scottish Government in terms of how those spending patterns might evolve, and the Office for Budget Responsibility and others have highlighted the risks that are associated with those totals.

The income tax net position is currently projected to contribute to real-terms growth in the Scottish budget. As we set out last year, the scale of the Government’s policies to raise revenues since 2018, through the creation of higher rates and new bands but, crucially, freezing the higher-rate threshold to below that in the UK, should lead to a positive net tax position in the years ahead, given the effects of fiscal drag. That net tax position will rise. From 2025-26 onwards, however, the exact scale of that positive net tax position becomes more uncertain. Part of the on-going growth is due to our slightly higher forecast for earnings growth in Scotland, compared with the equivalent forecast by the OBR for the UK. We therefore flag a downside risk that, if earnings growth in Scotland turns out to be more similar to that in the UK, the income tax net position from 2025-26 onwards could be lower than is currently projected.

Finally, the only part of the Scottish Government’s spending plans that we forecast is social security. We are expecting spending to rise from £6.3 billion next year to £8 billion in 2028-29. That is largely a result of payment rates being uprated by inflation and higher spending on disability payments, which reflects both a UK-wide increase in the number of people who are receiving disability payments and the effect of the Scottish Government’s policies on operational changes to disability payments.

Thank you for that helpful opening statement.

I will turn straight to your report, “Scotland’s Economic and Fiscal Forecasts”—my questions will be mostly based on that. On page 6 of the summary report, in the very first sentence of the introduction, you say:

“The Scottish Government’s budget next year is set to increase by £1.3 billion from the latest position for 2023-24. This is a rise of 2.6 per cent in cash terms or a 0.9 per cent rise after accounting for inflation.”

The point about the rate of inflation is always a bit of a bugbear for me, because it assumes a gross domestic product deflator for inflation of 1.7 per cent, but that does not bear any relationship to the real impact of inflation on the Scottish budget, given that more than half of the Scottish budget is salaries, which are increasing by significantly more than 1.7 per cent. Why does the assumption continue to be 1.7 per cent, given the differential between the GDP deflator and consumer prices index inflation?

There are a few things in that. It is a very good point. In normal times, the GDP deflator and the CPI inflation rate are very similar to one another, so that issue about the differential does not always come up. In most cases, the standard way of assessing the impact of inflation on Government spending is to use the GDP deflator because, in essence, it measures how much is being spent. That is standard practice. However, you are right that, particularly when pay awards are a significant driver of spending profiles, it is much more appropriate to use the CPI inflation rate to think about the impact on people’s living standards. That is why we have the differential between the GDP deflator and the CPI inflation rate.

The reason why we use the GDP deflator is that it is standard practice. It is the rate that the Scottish Government, the UK Government and the OBR use, so we largely do that for consistency. However, you are right to highlight that, at particular points, there will be differentials, which will mean that choices made on pay awards, for example, will look different from the overall inflation rate that is capturing total Government spend.

Yes, because we are talking about a real-terms increase in resource, but it is not a real-terms increase in the real world, if you are using a deflator that is so far below the increase in wage settlements, for example.

Yes, and you will recall that we had a similar issue last year about the differential between the GDP deflator and CPI because of the effect of energy prices and imports—that was one of the drivers of inflation, so the GDP deflator was lower than the CPI rate of inflation.

As I said, in time, the figures will start to move back into line with each other and there will be a better reflection of the different components of that. If we were to use CPI, for example, all the stuff that Government spends and purchases would be using too high an inflation rate—that is just one of the trade-offs that you have when you have this differential. The reason why we use the GDP deflator is largely because it is standard practice.

Obviously, I know that it is standard practice. It is frustrating that the figure is not based as much in reality as one would wish, certainly in these difficult times.

In the fiscal overview, on page 3, you talk about resource funding rising by 8 per cent by 2028-29. However, on page 5, unless my arithmetic is wrong, you say that the funding will increase in real terms from £52,449 million to £54,691 million, which is actually an increase of 4.2 per cent rather than 8 per cent.

Can you say that again, please?

Yes. You said that resource funding will rise by 8 per cent by 2028-29 but, in the figures on page 5, in the second line, there seems to be a 4.2 per cent rise rather than an 8 per cent rise. I just wonder why that is, or have I got it wrong?

Are you looking at figure 2.1?

This figure does not have a number. It is on page 5 of your summary report—it is right at the beginning.

The 8 per cent rise applies to resource funding, but the figure that you gave is for total funding. One of the big things that we are seeing is resource funding going up, but a significant fall in capital funding.

Okay. I assumed that that was resource only. I did not realise that it was capital and resource.

Yes. That top line is total funding. It is the resource funding that is going up. That is one of the really interesting things that we are seeing in the budget. We might come on to what that means for the net tax position, but there is a significant fall in capital funding. We talk about a fall in capital of about 20 per cent in real terms over the period. There is an 8 per cent increase for resource funding, but once you go to the total, the fall in capital funding drags it down.

It is interesting that, in her statement yesterday, the Deputy First Minister talked about a 10 per cent reduction over five years, rather than a 20 per cent reduction. That seems odd, given the figures that you have used. I did not see your figures until after the statement; if I had seen them, I would have asked the Deputy First Minister about that.

One of the issues that has been very much to the forefront has been the new additional rate of income tax, which starts at £75,000 a year. You have assumed that that will produce £82 million net, but the gross figure is £200 million. Basically, you are saying that the £200 million that, on paper, is being levied by that change will bring in only 41 per cent of that estimate, or £82 million. Would that be right?

Just to be clear, the figures that we are talking about are the impact of the introduction of that new tax rate and the increase in the top rate from 47 to 48 per cent. It is the combined impact of both of those changes that we are forecasting. We are saying that the gross impact is around £200 million and the net impact is about £82 million, once you net off behavioural effects.

How accurate are your predictions likely to be in that regard? What are the margins of error? I would have thought that they could be quite wide.

There are a couple of things that I would say on that. First, we have had a discussion before about the challenges in estimating tax elasticities. We have not changed our approach this time around. The numbers are based on all our work in the past and what we would expect the behavioural effects to be. The Government uses those sorts of things when it does its illustrative tax calculation to show the effects of such policies. In many ways, the numbers are not surprising, in the sense that they are exactly what we should be seeing, based on our assumptions. If you recall, back in March 2018, I think, we published a report on how we estimate behavioural change and tax elasticities. We then walk through the fact that, in order to come up with those tax elasticities, we have looked at a lot of academic evidence—a lot of studies that look at what happens, particularly within countries, when you start changing tax rates, but also what happens in the United Kingdom.

Of course, the reason why there is an element of uncertainty is that none of those studies are of Scotland. We are having to rely on looking at evidence from elsewhere. We then make judgment calls about which examples might be approximate and accurate in the Scottish context. We highlight that there is a level of uncertainty in that, and we are then, ex ante, estimating what the behavioural effects might be.

Clearly, over time, we would want to try to work out, ex post, whether those tax elasticities are in the margin that they should be in. His Majesty’s Revenue and Customs published a report in December 2021, in which it looked at 2018-19 and whether, from its evidence, there were tax elasticities. HMRC highlighted how difficult that is because, in order to estimate what people’s behaviour was and would have been, you have to think about what the counterfactual would be, which gets very complex and difficult. There are statistical methods that you can use. Long story short was that the estimated tax elasticities were not too dissimilar to what we use at the moment.

At this moment in time, based on the academic evidence and what we have got from the studies so far, we think that those tax elasticities are the best ones to use. However, it is something that we are very open about and happy to consult on and, as new evidence emerges, we will amend the figures as we go.

Just to add to that, we have not seen any evidence in past behaviour that says that something really important was missed. We have had significant changes in tax over the past few years and we have not seen anything in the data that says, “There’s a big surprise here. We’ve seen a change in tax revenue that lies way outside what we thought it would have been.” There is nothing standing out in the data that tells us that we are significantly off. I take your point that there are uncertainties, but we think that our judgments have been borne out.

12:15  

That is what we are really looking at: how accurate the forecast has been and how behavioural change has impacted in previous years.

Some of the forecasts seem to be pessimistic compared to others. For example, it now seems that we are going have a positive reconciliation in 2025-26 of £732 million. That seems to be very optimistic compared to what was predicted. What explanation do you have for that?

Box 4.2, on behavioural change, is an interesting walk through of the effects of behavioural responses and a discussion of the assumptions and the uncertainties.

I will make one final point about the tax policy that has just been introduced. We are estimating that it will raise £80 million rather than £200 million, but that is £200 million in a context of nearly £20 billion pounds worth of total revenues. Those are relatively small figures. You could change the assumptions significantly, but they would not actually have much of an impact on overall income tax revenues.

There is a point about expectations and assumptions about using different tax elasticities. These particular tax policies will not dramatically change income tax revenues, given the magnitudes that are involved.

To pick up on your point—

Sorry, but you are talking about 0.4 per cent of income tax revenues, which is £82 million.

Exactly. We will maybe come on to discuss what is driving the big growth in income tax. It is not the individual tax policies, but earnings growth and fiscal drag. That is what is generating the significant amount.

On your specific point about the £700 million revision, there are a few things that I would say. The easiest way to think about it is that we have revised up our income tax forecast on last year’s, by about £2.2 billion. However, £1.5 billion of that is essentially a result of trends that are similar across the UK. The OBR raised significantly its income tax forecast because earnings have been growing much more quickly than expected, which is in part because of inflation. All of that feeds through, and once you add in fiscal drag, you get really significant growth in income tax revenues. That is the first point.

That leaves the £700 million, which relates to that improvement in our forecast and therefore the improvement in the net position. There are a couple of things that I would say to help us understand that. The first is that you will remember that, back in July, we had better than expected outturn data for Scotland for 2020-21. When HMRC gave us the final outturn data of income tax receipts in Scotland, the figure was up by about £390 million.

We have dug into that and explained a bit where that that variation has come from. Part of it was an error in the way that HMRC was presenting the statistics for the previous year, which was over £100 million. That explains one of the reasons why, when we met the committee, I think in September, we were puzzled about the fact that there was such a gap between what we were thinking real time information would give us and what actually happened. Some of that was because the statistics were miscalculated for the previous year and HMRC have now corrected that, so we now have more accurate statistics in that context. That was part of it.

Another part was the emerging evidence that we were getting of Scottish earnings growth outperforming earnings growth in the rest of the UK. That has continued into this year and we have a discussion of earnings growth in Scotland in the economy section of our report. There is an interesting chart in figure 3.22, where you see the differential in earnings growth between Scotland and the UK, and Scotland moving ahead of the UK. We might come on to why we think that that could be the case, but we started to identify that differential in May. Once we calculate that in, it explains the remaining £300 million, which gives you your £700 million.

Some of it is the outturn data that we now have, which has lifted up the level. The second part is a judgment by us that the data on earnings that we are seeing in Scotland is actually real. That leads to us being able to be more optimistic about the earnings growth coming through.

In paragraph 3.47, you say that pay in the finance sector in Scotland is growing 3.5 percentage points faster than it is in the rest of the UK. In the next paragraph, you mention that, for a while, North Sea oil acted as a drag on earnings, which are coming back up to the average. Will you talk to us about that and the circumstances around that? Why is that happening? Is that likely to continue?

A couple of things are worth mentioning before I bring in David Ulph to pass on his reflections.

There is some really interesting data. The RTI data shows that average earnings growth for 2022-23 is 6.3 per cent for Scotland and 5.9 per cent for the UK. The annual survey of hours and earnings, which is the main Office for National Statistics publication on weekly earnings, shows growth of 9.7 per cent in Scotland and 6.2 per cent in the UK for last year.

Scotland’s earnings growth is the highest of any region or nation in the UK. We started to see that in May, which is one of the reasons why we uplifted our forecast. All the evidence that we have had since then backs up our assessment that earnings growth in Scotland from last year into the current year is running ahead of that for the UK.

In the report, we speculate about some of the reasons for that. One reason why we were more pessimistic before was that all the evidence had pointed to the north-east of Scotland lagging behind. If you look at the charts that I was talking about, you will see that the north-east is now back on track again; it has come back on stream.

In addition, there are differences in the financial services industry. Scotland has slightly different financial services from the UK, so, when interest rates rise in the UK, there is not the same huge growth in financial services.

The power of earnings growth to drive income tax revenues is really significant. David Ulph might talk a bit about some of the rules of thumb that we have. I note that, when you fix thresholds and relative performance to the UK, very small relative changes in earnings growth have a dramatic effect on income tax revenues relative to policy changes in that regard.

My final point is that £700 million sounds like a lot—it is a big number—but we think that income tax revenue will double by the end of this forecast period from where it was in 2016-17. We all have to uplift our numbers because of the effects of inflation. Something that might have been £300 million last time will now be £600 million simply because of the effects of inflation.

I will add a few points. We spent a lot of time digging into other evidence not just from data sources but from surveys that were being done. We saw a huge amount of evidence suggesting that labour markets in Scotland were tighter than labour markets in the UK as a whole. That supported our view that there was relative earnings growth in Scotland because of the tighter labour markets.

Another factor that we took into account was the difference between the financial sector in Scotland and the financial sector in the rest of the UK. In Scotland, the financial sector is more heavily dominated by work in pensions and insurance, where returns and earnings are fairly steady. In London and the south-east, the financial sector is more heavily represented in areas of investment, where returns are more volatile. Earnings are more volatile through, for example, being related to bonus pay. We could see steady growth in Scotland and a bad year for bonuses in the rest of the UK.

Those are some of the factors that we looked at when thinking through why earnings growth in Scotland is so strong relative to that in the rest of the UK.

I want to provide the committee with two stylised facts that might help you to think about why we have seen the growth in tax revenues and about the policies that you might want to consider to continue that growth in earnings and tax revenue.

One such fact is that, because Scotland’s tax system is now really quite progressive, for every 1 per cent by which earnings grow faster in Scotland than they do in the rest of the UK, net tax revenue will go up by £250 million. We calculated that figure by supposing that earnings growth was a certain percentage in Scotland and another in the rest of the UK and asking what that would do to net tax revenue. Every percentage point higher in Scotland would yield an additional net tax revenue of £250 million, because the Scottish tax system is already very progressive. However, because it is also relatively more progressive than the tax system in the rest of the UK, every percentage point of growth in earnings in Scotland generates £25 million more in net tax revenue than in the rest of the UK. That means that we get all the strong effects of both any earnings growth and any differential earnings growth in Scotland compared with the rest of the UK.

Promoting earnings growth is therefore a very powerful way of driving up tax revenue in Scotland. The more we can get earnings to grow faster in Scotland, the more we can get tax revenues to grow.

Fiscal drag and rising earnings—

It is not fiscal drag; those are just the effects—

But, to a large extent, fiscal drag is the reason for the forecast being £2,211 million higher now than it was a year ago.

Yes.

I want to come back to capital, briefly. I will ask questions for only another three or four minutes, because colleagues are keen to come in. What is the impact of the reduction in capital on medium to long-term growth? Even if we accept the GDP deflator—we know that the reality is completely different, but assuming that we accept that as a real figure—that will still mean a 20 per cent reduction in capital over five years. What will be the impact on growth and productivity of capital taking such a hit?

You are right. We talk a bit about that in our capital assessment. When we consider what such matters might mean for the economy overall, we highlight that aspect. The key factor that drives the fall in capital is the block grant. The Government might do some borrowing around the edges, but the big money comes mostly from the block grant.

About 90 per cent comes from the UK Government.

It could borrow more but, at the end of the day, the majority of the money comes through the block grant. That reflects the trend across the UK.

It is quite interesting that, under the UK Government’s plans, capital spending as a share of UK GDP will fall from 3.5 per cent to 3 per cent over the next five years. The OBR noted that, should those plans be delivered, about half of the rise in investment as a share of the British economy since 2019-20 would be reversed. If those projections hold true, that would represent a significant shift in the amount of capital investment going into the UK economy, because the UK Government is, in essence, flatlining capital borrowing in cash terms.

Speaking as economists, we feel that, if capital investment is spent wisely, it can have a huge impact on the economy. It connects communities and means that businesses can become more productive and that connectivity both within the UK and outside it can be better. Therefore, all the evidence points to capital investment being a really important part of economic growth in the long run.

Clearly, UK ministers are operating a balancing act in how they balance the books and how they can bring the deficit down. If they cannot do so, they must consider, on the one hand, the effects that that might have on the economy in the long run and, on the other, where they can put the balance of any adjustment. They have put it on capital spending, but that will have a significant impact on long-term infrastructure plans and potential long-term growth in Scotland.

I have a final question for the moment. In its autumn statement, the UK Government did not look beyond 2024-25. What impact has that had on your forecasting for Scotland?

12:30  

It has added an additional layer of uncertainty. We have broad high-level totals for what has been said about resource spending, but we do not have the departmental spending elements within that. We have indications from the current UK Government with regard to defence and overseas development, and that lets you do some calculations about what that might mean for individual spending portfolios. However, it means that things look exceptionally tight for the block grant over the five-year period.

As we have highlighted in our report, the OBR has said that, when you look at what has happened in the past, you get the sense that the spending plans will be exceptionally tight and that, on the balance of probabilities, it is more likely that future UK Governments will spend more than what is in those plans. We have highlighted that as a potential positive risk to the Scottish budget down the line, but it is still clearly a risk. The UK Government could choose not to change its plans, which would lead to more challenging settlements for the Scottish Government.

Thank you. I now open up the meeting to members’ questions.

Good morning. I will focus on just one point, because we have a lot of areas to cover and I know that everyone wants to come in.

On the issue of behavioural change, I fully accept what you said about the figures that the convener highlighted and your point that they are relatively small in relation to the overall income tax base. As I understand it, your modelling is based on those who might be impacted, but I am interested in exploring the second-round effects, if you like, against the challenge of counterfactuals.

I note that, this morning, Professor David Bell said:

“my concern is that it will be difficult to attract the workers and investment to Scotland that are needed to generate sufficient growth”.

The same sentiment has been expressed elsewhere. Instead of, as we would imagine it, people leaving the country, it might well be that people just do not come. That is one possible scenario among a multitude of them. I would like to hear a bit more on your thinking and reflections on that at a big-picture level.

We thought about both those dimensions—the impact of the differential tax situation between Scotland and the rest of the UK on people deciding to move their tax position away from Scotland and on deterring those who might otherwise have chosen to come to Scotland. I go back to my earlier point that, if we had been seriously wrong about this, we would, I think, already have seen something in the data. We would have seen some impacts on tax revenue, which we have not seen, because of the previous changes that have been made—

I am sorry to interrupt, but I suppose that you are looking at the data that is coming through and, as I heard Professor Roy say, past evidence. I am not talking about a Laffer curve tipping point, but I am trying to explore the perception of Scotland as a place to invest. Does this change anything at all in that respect?

We have thought about that quite a bit. There could be the cumulative effect of lots and lots of changes, with people saying, at some point, “Okay, I now see what it’s going to be like if I move to Scotland, and I see what it’s going to be like if I stay down south, in England. I see the picture more clearly, and my decision is not to move.” That sort of effect—the straw that breaks the camel’s back—kicks in after a period of time, and it will definitely change people’s minds about coming here.

Again, though, we have not seen a lot of that in the data. When we look at net migration, we see quite a lot of inward migration into Scotland, and a lot of that comes from people with relatively high incomes. These are people who might be retiring to Scotland, say, which would raise an issue about their contribution to GDP and the impact in that respect. The impact on tax revenue might not be so great if people have significant pension income that will be taxed at high rates. We have thought about that.

You bring out something else. People say that net migration from the rest of the UK into Scotland is higher than it is the other way round, but I suppose, if that is true, there is a question about the implications for tax and investment.

I appreciate that we do not have the answers here. I am just trying to get more of your reflections. The problem is that we do not know what will be true until after the event.

That is one of the real challenges. There is a really interesting research question about how you do this. What is the counterfactual? Who are the people who might come, and who will not come, or who are the people who come because they want to live in a country where there is additional investment in public services? You cannot experiment with that. We cannot test the counterfactual and see how people respond. As we have now collected data for a number of years, however, we can start to look at particular aspects of that question. Are people moving more money into pensions in Scotland relative to the situation in rest of the UK? Are higher earners in Scotland changing their hours of work relative to the situation in the rest of the UK? Can you start to pick up evidence about levels of investment and so on flowing in?

It is quite difficult for us to do that within our remit and resources, because we are very much focused on what happens next year. Everything that has happened before has been baked into our baseline, and we look at the change there. Our report says that the Government should look at this important question. It is the sort of thing that would really help us, in that we could prepare for our longer-term work and then show the effect of policy decisions not only on what happens next year, when someone’s behaviour might not be able to change that much, but on where investment decisions, vacation decisions and so on take place over five to 10 years.

I will add one other point. The way in which we analyse this when doing our calculations is that we look at what is called taxable income. We do not try to go down to the minutiae of whether there is an effect on decisions about people’s hours of work, how many weeks they work in the year or the amount that they put into pensions, or on the amount of tax avoidance or other types of tax behaviour that is engaged in. All that we look at is the effect on taxable income, which pulls together all the different ways in which people might respond to differential tax positions between Scotland and the rest of the UK.

We use elasticity for taxable income and look towards the high end of those who are out there in the data. We recognise that some high-income people in Scotland might have property in Edinburgh and in London, so they can change their tax residence relatively easily. We tend to use higher elasticities to reflect that. That is just a judgment; it is not based on, or supported by, detailed minute analysis.

We all agree that it is a problem that we do not know.

I would like to stick with the question on behavioural effects, but look at it from a different perspective.

I am struggling somewhat to square the circle with regard to the amount of airtime that we are spending and the amount of political debate that we are having on the risk of negative behavioural effects as a result of income tax changes and the data that we now have from five years of increasingly significant divergence. Despite increasing divergence in our more progressive system, we have seen growth in earnings and thus direct growth in income tax receipts. As Professor Ulph has pointed out, we still have net positive migration into Scotland from the rest of the UK, and we are doing very well in foreign direct investment compared with everywhere other than London, I believe.

Are we, therefore, spending a disproportionate amount of time discussing the potential negative behavioural effects of income tax divergence compared with other factors that affect the budget in a much greater way? As Professor Roy has pointed out, we are talking about relatively small numbers in the grand scheme of a budget of £60 billion or so.

First of all, I do not control the political discourse or what people do in that respect. That said, I made a point very similar to yours on one of the phone-in programmes this morning, saying that we have to look at both sides of the balance sheet, as that is what determines behaviour. We are projecting our net tax position next year as £1.4 billion. What we spend that money on matters to people’s decisions, to the economy and to the perception of Scotland, and it is about which effect balances itself out.

We do our best to come up with a behavioural estimate of what we think the tax policy will raise as people respond to an individual tax decision. If you are fortunate enough to be in the £100,000 to £125,000 bracket, and you are losing the personal allowance and facing a marginal tax rate of nearly 70 per cent, do you put more money into your pension or not? Those are the types of things that we are trying to capture. We are not looking, in that context, at the effect on the long-term structural growth rate of Scotland’s economy.

You are right. For the first few years of tax devolution, the net tax position was in quite a precarious state relative to what it was. That was because earnings growth in Scotland was lagging behind that of the UK. However, from the evidence that we saw at the time or have seen since, that had much more to do with what was happening in the north-east of Scotland and the performance of Scotland’s economy than with what was happening in terms of the relative tax differential.

That brings me back to David Ulph’s point, which is important. You are right that a debate is happening on these marginal changes in taxation, and I get that they are important, but the big thing that drives revenue is earnings. If you grow your earnings at a higher level, as the UK does, Scotland will get more revenues because of the progressive system and fiscal drag, and if you grow them more quickly than the rest of the UK, you will generate much more revenue than you would do from introducing a new additional rate at £75,000. Earnings are key to the economy.

When you think about a lot of these issues, I also ask you to keep in mind the stylised fact that I gave you—that is, that every 1 per cent by which earnings grow faster in Scotland and in the rest of the UK increases net tax revenue by £250 billion. It is an important number.

That draws us into the debate about the fiscal framework and whether relative tax growth is the best measurement from Scotland’s perspective. We have discussed that before, and I am sure that we will continue to discuss it for some time to come.

Box 4.2 of your report has an interesting reference to the behavioural effects and how you estimate, measure and mitigate them. It also references the HMRC report from 2021 on the behavioural effects of tax changes in Scotland. I remember that report, but I cannot remember why HMRC produced it. Does it do so on a cyclical basis? Should we expect another one, or was it a one-off?

It was an occasional paper that HMRC did for 2018-19. I note that HMRC is collecting longitudinal data, which will let people do that sort of work much more into the future. It means that the Government, working with HMRC, can actually make assessments of and undertake projects on what the behavioural effects will be.

For us, that is brilliant. If we can update our behavioural estimates with estimates based on Scottish data, and we can see the unique effects in Scotland, we will update our forecast accordingly. However, nothing in that study suggested that what we were doing was wrong.

If HMRC is, at some point in the short to medium term, producing more longitudinal data, it might be worth while for the committee to get in touch with it to ask about the timescale for that, because it would inform quite a lot of our work.

Box 4.2 also mentions the extent to which the USA and Switzerland are relied on, because there is such a rich evidence base in both countries. What types of evidence-gathering work or studies that have taken place in other jurisdictions are not taking place—or have not taken place—in Scotland? I am thinking about work that Government could commission or which independent organisations could be encouraged to undertake.

In our paper back in 2018, we reviewed the entire literature, which looked not just at places such as the USA and Switzerland for within-country changes in behaviour, but at behaviour at a national level. For example, the Institute for Fiscal Studies and others did a study on the introduction of the 50p rate in the UK. We can draw on those things. There is a question about how comparable the USA and Switzerland are to Scotland, so that is a challenge, and another question is how comparable a UK-wide system is to Scotland. That is where we have to use judgment.

The key is that we now have taxpayer data for Scottish taxpayers and taxpayer data for UK taxpayers. You can use sophisticated techniques to identify similar taxpayers in Scotland and the rest of the UK. You can control for everything else and then see whether there are differences in behaviours in people living in Scotland relative to those living in the rest of the UK that can be attributed to the changes, up or down, in tax policy. As we get more data, that will be the really positive bit, and we will be able to do that work. However, it is difficult for us to do that, so we really need Government and others to work on the matter.

12:45  

Going back to Michelle Thomson’s point, I would say that you need data on two different aspects. The first is how people respond to differences in taxes, conditional on where they choose to live. People living in a particular area might respond to the tax that they face in that area with decisions on savings, on work or on tax avoidance. By looking at people in different parts of the country who are facing different rates, because of local powers on taxes, you can pick up some of those effects.

The more complicated issue is how to pick up the migration effects—that is, the decisions that people make to move between one area and another, because of the tax differences. For that, you need data on movements of population; you cannot just study the people who happen to be in a particular region at a particular time. You need a combination of quite rich data sources to pick up those effects.

Absolutely. Thanks very much.

I will start with a very technical point. Your forecast for median earnings in Scotland for 2024-25 is £28,200, while according to Office for National Statistics data, the median salary at £29,675 for the current year. What is the difference there?

Is the ONS figure for Scotland and for the same year?

Yes.

I will need to get back to you on that.

It seemed a bit strange, because I would have thought that it would be relatively easy to work that out. If you could get back to us, that would be extremely helpful.

I also want to ask about the gap between private and public sector earnings, which seems to be widening, as far as I can work out. Can you give us a bit of background on that?

In terms of the differences between the public and private sectors more generally?

Yes. They seem to be moving apart quite a bit. Is that correct?

The gap tends to ebb and flow over time—and I should say that it is a UK effect, too. Part of that comes from impacts in relation to levels of unionisation and so on. During downturns, in particular, public sector wages tend to stick higher than private sector wages, and in high-inflation environments, public sector wages tend to be better. Typically, that gap closes over time.

On average—and this depends on using a median rather than a mean distribution—median public sector wages tend to be higher than median wages in the private sector, because the private sector has a larger number of very low-paid jobs. However, it also has a much greater number of very high-paid jobs. When you look at median public sector wages, you see that the average salary is typically higher than that in the private sector. However, as you have said, it widens and closes over time, depending on how the economy is doing.

Does that reflect how different sectors in the economy are doing, too?

Yes, exactly. Part of it is how quickly they respond to the economy. Public sector wages tend to be negotiated and set for the period ahead; indeed, the Government is now thinking about a wage policy for 2024-25. Those in the private sector are probably thinking about what their wage policy will be for the next few months, or what might happen to their employment, as it is much more volatile. What is typically seen in a downturn is that public sector wages take much longer to adjust and then lag behind any recovery, whereas private sector wages are much more volatile over the time period.

Given that the committee is trying to drill down into where the greatest potential for economic growth and productivity improvements might be, those gaps are quite important.

To broaden out the conversation, I should also highlight the important point in the report that about half the devolved resource budget is pay. If the whole devolved resource budget goes up by just over 3 per cent, any pay awards above 3 per cent will mean that the remaining half of the budget will have to go up by less than 3 per cent.

A table in the report shows that, if you strip out non-domestic rates income and social security, what will happen next year is that the resource budget will actually go down. That is a really difficult decision for the public sector. If inflation is 3 per cent, anything lower than 3 per cent means a real-terms cut in salary, but if half your budget goes on wages and everything else is under pressure, where do you find the additional funding?

You have very helpfully answered my next question, because what you have said leads on to the potential for public sector reform. I know that you cannot comment on that, because it is a political decision, but it will be predicated on the trends in the data that you have just highlighted.

Exactly. Back in May, in the medium-term financial strategy, the Government’s central scenario was that pay awards for this year would be 3.5 per cent; however, the figure has turned out to be 6.5 per cent. It is estimated that, in total, there will be £800 million more of spending pressures this year than the Government thought, because inflation is higher.

It is really difficult to strike a balance. If you have high inflation, how do you continue rewarding public servants and paying them a wage that is not being eroded by that inflation? If pay is half your budget and if the budget is not going up very much, where do you make the adjustments? Either you reprioritise spending or you change the head count in the public sector.

I will stick with the issue of the public sector workforce. I am interested in how Government policy informs what you have done and the numbers that you have produced. The resource spending review, which was a major piece of work back in May 2022, said that the Government would aim to return the total size of the devolved public sector workforce to around pre-Covid levels by 2026-27.

We were promised that there would be more detail about that in last year’s budget, but John Swinney did not provide that and said that it was up to public bodies to do so. On 16 May, I asked the permanent secretary about the status of that policy, but he said:

“I do not think that that has been publicly stated”—[Official Report, Finance and Public Administration Committee, 16 May 2023; c 36.]

He did not know what the status of that was. On 13 June, Shona Robison told this committee that it was

“A bit of a blunt tool”—[Official Report, Finance and Public Administration Committee, 13 June 2023; c 27.]

and said that she was abandoning RSR, but, in recent days, we have heard much more language about a big decrease in the public sector workforce. What is your understanding of Scottish Government policy and how has it informed your predictions and assumptions?

We use the current estimates of public sector employment and an assessment of public sector pay to make our tax and economy forecasts. The protocol that we have with the Government is that it should give us a final pay policy and a public sector employment policy. That does not mean that we need exact or specific detail; it can be relatively broad brush.

We did not get a final pay policy from Government. We have developed our own and I can explain what we have done, but we do not have a specific, final policy from the Government.

Was there a reason why the Government did not provide that?

The Government will be taking part in negotiations and thinking about plans. That does not really matter too much for our forecast. We do not have a final pay policy from the Government in our forecast; it is for the Government to finalise that in the coming months.

We have been open with the Government about what we are doing so that it knows. Our modelling includes a 3 per cent average pay award, which is based on the fact that we think that inflation will be 3 per cent next year. Adding in pay progression takes us to about 4.5 per cent. Therefore, our modelling includes an assumption of a 4.5 per cent pay award for the public sector. As I said to Ms Smith, that is greater than the overall resource budget projection for next year.

That leads to three possible implications. First, the pay award that we have in the modelling is too big relative to what will ultimately transpire. Secondly, it requires that the other half of the budget will have to take up more of the slack to pay for that award above the total resource increase in the budget or that, thirdly, public sector employment will need to fall.

What percentage of the overall budget is pay?

I think that it is just over half of the resource budget.

That is my understanding.

Perhaps Claire Murdoch can confirm.

Yes. The Scottish Government has estimated that pay accounts for £25 billion of the resource budget in the current financial year, so that is just over half.

You requested a pay policy from the Government. You said that that was one of your requirements but the Government did not provide you with a policy or an indication as to what it was doing on pay.

The protocol requires that the Government gives us a final pay policy.

By when?

It is part of the forecast. We would normally expect to have a final pay policy and public sector employment policy from the Government by our final round 3 forecast.

When would that be?

The timeline is set out in our document.

As part of the protocol, we have a deadline for economy-moving measures, which is just over two weeks before the budget, so the Government needs to provide us with that information by that deadline. That process is agreed with the Government when we get notification of the budget.

So the Government just did not do that. Were there any other documents that it had previously agreed to supply that it did not supply by the deadline?

There are a couple of things. The point about pay and public sector employment is that, in the absence of a definitive position from the Government, we come up with a judgment, which we share with the Government. It has an opportunity to object to it and say that it is wildly different from where we are.

In many ways, I am relatively relaxed that we have a broadly good assessment of what public sector pay and employment will look like. We are interested in the macro effects of that. The intricacies of the exact numbers are ultimately up to the Government. I understand that there will be careful negotiations with trade unions and the like.

You have worked back on the basis of that. You made an assumption and calculated a figure based on it.

What assumption have you made about the size of the public sector workforce as a result of the budget constraints and using that number of 3 per cent?

I do not know whether Claire Murdoch has that to hand. If not, we might be able to write to the committee with it.

I think that we estimate that it will fall by 1.5 per cent in 2024-25.

We will get the exact number for you.

Is that about 7,000?

That would be 3 per cent.

Would it be 3 per cent?

If we can get it to you, I would rather give you the exact number than give you one that I have calculated in my head.

It feels pretty major. You have not taken account of a policy objective in any of that. Previously, the Scottish Government set out a policy objective in the resource spending review. It is your understanding that that has been abandoned. In essence, you have had some media reports, so you have just gone with your assumptions on the numbers.

To be clear, the Government has not given us its pay policy and public sector employment policy for next year with exact numbers and asked us to use those numbers. Therefore, we have made a judgment based on a reasonable macro forecast for next year. We have said that inflation will be 3 per cent, and we added pay progression, which took us to 4.5 per cent, and used that throughout. We have not been given a particular policy in regard to that.

Okay. Thank you.

I am not sure that I got a specific answer to my question about whether there is anything else in the agreement with Government that it agreed to provide but did not provide by the deadline.

13:00  

We got everything that we needed to make our forecasts. I am entirely confident that we have everything that we need to make the forecasts.

There is a discussion about the process behind the forecasts on page 23 of the document, at the start of chapter 1. There is a table underneath that with headline dates. Members will see that the Government missed the deadline for providing us with final policy recommendations. Our final deadline for policy recommendations was 5 December. We agreed to extend that deadline to 8 December, because we did not have final decisions from the Government. There was a second extension for a small element of policy that took us through to 13 December. We highlighted that and were transparent about it, because it is important that we are transparent about the fact that we have a protocol and that the deadlines are there to be adhered to, but they were not.

I understand how difficult it is to set budgets. There is a condensed timeline, and there are really difficult decisions to make. That is why we, as the independent fiscal authority, were happy to be accommodating to the Government on that.

That has had impacts on us in respect of the amount of time that we have been able to spend on writing the documents, for example. At the end of the day, I am comfortable with the forecasts but, to be honest and transparent, the deadlines in the protocol have been missed.

Okay. Thank you.

I have a couple of other small issues to raise. We will move on, if that is okay. We are looking at fairly large increases in the social security spend, which you have already set out. Can you talk about the relative amount of the uplift that is the result of behaviour changes and policy in Scotland? There has been a lot of talk about a kinder and more generous policy. How much of that is based on assumptions about the intent of the policy, and how much of it is based on experience? Has Social Security Scotland provided you with figures? Are you basing what you say on modelling the increase or just on your broad assumptions about the policy intent?

There are a few things to say about that.

Many changes were made for things such as the adult disability payment, which was introduced only relatively recently. We have not been able to track through in any detail all the impacts of that on behaviour. All that we can say is that we have not seen anything in the data for the number of awards that are being granted that goes against the initial judgments that we made when we produced the forecasts. Although I cannot say that all our judgments have been fully borne out, equally we have not seen anything that significantly contradicts our initial judgments. It will be quite a while before we have enough data to be able to fully understand all the implications of the reforms that the Scottish Government has made to ADP.

There are a couple of interesting bits of data. Figure 5.9 in our document shows ADP and relative applications in Scotland, England and Wales. We can see spikes at the pilot and the national launch that took us ahead in the gap between us and England and Wales.

That is the type of data that we are tracking. We assumed that that would happen, and it has happened. The question is whether that will unwind or continue to be there. That is the type of information that Social Security Scotland gives us.

The support that we get from Social Security Scotland in providing such data has meant that we have been able to make tweaks to our forecasts. For example, we have changed child disability payment forecasts. We have seen a big spike in applications for that, so we have uplifted our forecast. However, Social Security Scotland has given us evidence on the average payment, which is slightly lower. Therefore, we can change that at the margins.

To answer your first question about the difference, roughly speaking, around half of the gap between spend and the block grant adjustment by the end of the period is from new payments, such as the Scottish child payment and the carers supplement. The other half is from what we think are the changes in behaviours and the different type of system that is running ahead of the equivalent system that exists in England and Wales.

A lot of the casework that I am getting in this area is about people who are just not getting decisions on adult disability payment. There are huge delays. I am doing some work on this and it looks as the delays in processing are longer than those in the Department for Work and Pensions. Are you getting sufficient data as a result of those long delays?

We see delays and then we see an acceleration. In the data that we track, there are backlogs at certain points. Social Security Scotland is very good about telling us, for example, that it has a high case load coming in and there is a backlog so we might see a spike and that is due to Social Security Scotland processing applications that have been delayed, rather than for some other reason. We track that and it is the sort of thing that we then use to help inform our modelling.

One interesting fact that we got from discussions with Social Security Scotland was that when people were told that they could provide a lot of information in support of their claim, they tended to assume that Social Security Scotland already had that information. They felt that, as it is part of the Government, it must have all those figures. People assumed that Social Security Scotland already had a lot of that information, but they had to then go away and get that information themselves from other sources. That was partly why the delays were so much longer. There was a misunderstanding about who actually held the evidence and an assumption was made by a lot of people that the Government already knew those numbers.

My final question is on a different area. I agree with colleagues that this is such a tiny amount of money—£8 million from the very top rate is being realised by this increase. Is it worth the risk? Is it worth all the risks that you have set out such as the reputation and attraction of the country? Is it worth doing?

I think that is for politicians and policy-makers to decide.

Okay—I thought that I would give it a try.

It is always worth flying these kites, but Graeme Roy is switched on, even when such questions are coming in his direction.

I will touch mainly on points that have been raised to some extent already. You have explained why Scotland’s earnings have been ahead of the UK’s in the past couple of years and in the current year. Are you assuming that that will continue to be the case? If so, is there a danger that that is a little bit optimistic?

We are looking at our forecast of earnings in Scotland compared with the OBR’s forecast of earnings in the rest of the UK.

We have a table in our report that shows that the forecasts that other people have been making for earnings in the rest of the UK are somewhat closer to our estimates of the forecast for earnings growth in Scotland. We have highlighted in the report that there is a risk that the OBR might revise its estimates for growth of earnings in the rest of the UK and that could come a lot closer to our estimates for the growth in Scotland.

The gap is not between our forecast of earnings in Scotland and our forecast of earnings in the rest of the UK; rather, it is between our forecast of earnings in Scotland and the OBR’s forecast of earnings in the rest of the UK. There are two different forecasts in that sense.

Okay, but there is certainly a risk that we do not continue as we have been in the past couple of years.

Yes, that is definitely a risk.

One point about that is that—just to get the magnitudes in line, and we have been playing around with numbers—if the earnings growth towards the end of the forecast horizon was to align between Scotland and the UK, that would reduce the net tax position by about £500 million, but the net tax position would still be positive.

That brings me back to the point that, because of the successive changes to the progressive system, the net tax position is moving and has become more and more positive. Therefore, even if there is alignment with earnings, there will still be a positive net tax position, because of higher earnings across Scotland and the UK.

In the past, almost the opposite happened, with quite large negative tax reconciliations being forecast that—thankfully—got a bit smaller. Do you think, then, that we are going in the opposite direction, with quite big positive reconciliations being forecast that might well reduce?

There are a couple of things to say about that. As you will recall, the previous net tax position was negative, because Scotland was lagging the UK economically—we have already talked about oil and gas et cetera. Currently, we have, instead of negative net tax positions, positive net tax positions that are very small relative to what they should have been.

We are saying that, if you look at last year and this year, you will see that Scotland’s potential relative economic performance—and, crucially, its earnings potential—is now better than that of the UK. That is why we are now saying that a positive reconciliation is kicking in. The fact that earnings are growing so much across Scotland and the UK—I am picking up on David Ulph’s rules of thumb here—is making the numbers much bigger than we were saying in May that they were likely to be. You will recall that, back in May, and even last December, we were saying that the net tax positions would shift up; however, because earnings are growing so much, they are going even higher. That is where we are this year, and we think that that will continue into next year.

The divergence—or the risk that David Ulph referred to—comes in towards the end of the forecast horizon, where there is a difference between ourselves and the OBR and, indeed, a difference between the OBR and other forecasters in the UK. If that narrows, the net tax position that we are projecting will be lower, and the uplift in the revenue budget will be lower than what we have set out in our current report. That is a risk to the budget that we have highlighted with regard to Government’s planning.

Okay. The figure of £668 million has been mentioned already. It is just 3.5 per cent of Scottish income tax revenues, which, as you have said in the report, is

“a relatively small shift in terms of income tax forecasts.”

I accept that, if we could forecast most things within 3.5 per cent, things would be fine. However, it is a very big figure when it comes to what we are allowed to have in reserves or to borrow. You seem to be quite relaxed about it, but do you understand that, for us and for the Government, the figure is quite worrying?

Yes, and that is one of the quirks of the framework. You are dealing with movements in two very large numbers. Such movements can lead to differences that might seem small but which actually turn out to be really quite significant when you manage the budget, and all the risk of managing that falls on the Scottish Government.

That has, in some ways, informed the fiscal framework review, with regard to lifting up the borrowing requirements on resource, giving greater flexibility in the Scotland reserve on what it can be used for et cetera. The numbers, therefore, have been increased. However, the really important point—it will be really important for both Governments to continue to keep an eye on this—is that those limits increase by inflation, which means that it will help if inflation continues to go up.

You will want to look at whether you have the management tools to manage the existing risk. I come back to the point that, in 2017-18, income tax revenues were £10.9 billion, and we are projecting that, by the end of the period, they will be £23 billion. These numbers are going to get bigger in the future, simply because we are dealing with bigger levels of income tax.

Before you ask another question, John, I see that Claire Murdoch wants to come in.

In the current period, we are seeing both a positive improvement in the income tax net position and forecasts of positive income tax reconciliations. However, it is not a given that those two things will happen at the same time, as the reconciliation is about the relative forecast error. That means that, even if we have a more positive income tax net position, we might still see negative income tax reconciliations. It is just that the current two income tax reconciliations that we are forecasting are basically picking up a period during which we did not quite capture the improvement in Scottish income tax revenues. It is almost as if the lag in that improvement is why we are seeing these positive income tax reconciliations.

13:15  

In the future, we might be better at forecasting Scottish revenues and the OBR might be accurately forecasting UK revenues, so the relative error that would happen might mean that we could have a negative income tax reconciliation at the same time as a large positive income tax net position.

That is helpful; thank you.

The issue of multiyear forecasting has also been raised. Is my understanding right that the UK Government and the OBR have not been very specific in their forecasting with departments, which has meant that neither you nor the Scottish Government have been able to make accurate multiyear forecasts?

We can make multiyear forecasts for revenues by using the UK Government’s broad-brush statements about the block grant. The Scottish Government’s argument would be that, while there is uncertainty about the net tax position and about what the UK Government might do, it can do a budget for only one year.

We say in our report that, as the independent fiscal institution, we find that disappointing and think that the Scottish Government should set out multiyear spending plans. In many ways, the fact that there is uncertainty means that the Scottish Government should do more to set out how it will manage potential variability in those elements.

I would always advocate Government making multiyear plans. We could do much more were that to happen. Take figure 1 on portfolio spending as an example. We can give figures for only one year because we do not have any information beyond that.

Is that approach almost inevitable, or could the Scottish Government do more forecasting, despite the lack of information?

The Scottish Government could do much more. Clearly, it would put the uncertainties and caveats around it. There is nothing to stop people doing multiyear forecasts.

The position is similar for the UK Government. There is nothing to stop it setting out spending review projections beyond what it has done at the moment.

A change of Government at UK level could change quite a lot.

Another risk that we highlight is that the figures are based on projections before a UK general election and that a future UK Government might have different opinions about how to manage UK public finances, which will, in turn, have an impact on the block grant.

We have published a five-year forecast of the Scottish Government funding position for resource and capital. Those are set out in nominal terms as growth rates in our report, and the supplementary tables have the cash figures, too. We published those alongside the budget.

The fact that the UK Government does not publish detailed departmental spending plans is not unusual. It publishes a spending review but nothing beyond that, which is not an unusual position. It is not ideal, and it makes it more difficult for the Scottish Government, but it is normal to have only the current spending review period of spending plans until you get to the next spending review. That is the context.

Michael Marra spoke about the demand for social security and the fact that the Scottish system is different from the UK one. What are your future assumptions? There is an increase of 6.7 per cent this year, which is in line with inflation. Do you assume that that will continue to be the case in future?

Are you asking about uprating?

Yes.

Our assumption is that most benefits will be uprated in line with the consumer prices index inflation figure in September of the preceding year.

The Government has not really changed its social security policy, so the drivers of the numbers are broadly similar to what they were last time. Those numbers are driven by changes and increases in inflation, and by trends in disability claims across the UK. That is the main reason why those numbers have increased.

Do you want to come in, Professor Ulph?

The increases in applications for disability benefits are UK wide, not just in Scotland. In our latest report, we said that those increases were driven by people suffering poor health after Covid, increasing waiting lists in the NHS and cost of living pressures. In its latest report, the OBR said that it has looked again at the increase in applications and it thinks that cost of living pressures are more dominant in explaining the increase in demand across the whole of the UK. Obviously, as cost of living pressures ease and inflation falls, that factor will fall away. That does not affect the immediate forecast for social security, but we say that we think that it will not grow so fast in the future because of the diminishing impact of the cost of living.

Looking at the figures quickly in the budget, I see a lot of areas that are being restricted and that social security is going up by about £1 billion. From an economic point of view, does it matter where the Scottish Government spends that £1 billion? Would it have a different impact on the economy and growth if it was spent on housing or education, or does it not really matter? Is that not something that you look at?

We do not get into the detail of alternative policy options. David Ulph might have a different view on what I am about to say. If you get into questions about different fiscal multipliers and the relative impact of different spend, in a short-term forecast, most of them will wash out to one another. It is more about the broader objectives that the Government might have.

In the long run, you get some really interesting questions about the potential benefits to the economy and economic growth over multiple years of tackling poverty, and that conversation is similar to the one that we were having about tax. However, forecasting what happens in the economy next year is not something that we do. To be honest, I do not think that it would have too much of an impact on our immediate forecast.

It is always worth bearing in mind that social security spending is essentially demand led. Once the Government has set all the policies for how people can apply for social security and the levels of remuneration that they get depending on their conditions, the amount that it ends up spending on social security is not under the Government’s control. That depends on the number of people who choose to apply for and get social security. You cannot change that number from one year to the next in the same way that you can decide to cut spending on housing or education. Once the Government has set its social security policies, then, essentially, the spending is out of its hands for quite a number of years until it introduces new reforms.

I understand that, but the Government could change the policy.

It could change the policy, but it is worth remembering that a lot of the increase in demand has been set in motion by decisions that were made a few years ago. You do not have the same luxury of tinkering at the margins.

I understand that. Thank you.

I am very conscious of time, so I will just ask about the tax issue. Professor Roy, this morning you were talking on the radio and highlighting the marginal rates. The marginal rate at £43,000 to £54,000 is 54 per cent, although I understand that that will come down. Obviously, we have talked earlier about the rate on £100,000 to £125,000 being almost up to 70 per cent. Would you comment or give us your concerns or thoughts on that?

We are talking here about what happens when Scottish devolved income tax policies interact with UK reserved tax policies and national insurance and what happens to the personal allowance. There is a quirk that means that national insurance goes from 12 per cent to 2 per cent when someone moves into the equivalent of the UK higher rate. Because the Scottish higher rate starts earlier, people will be paying the higher rate of income tax in Scotland as well as the core rate of national insurance. That is why you get that marginal tax rate of above 50 per cent.

The same thing happens when you get up to £100,000. At the UK level, you start to lose the personal allowance, so that increases the marginal tax rate for every pound of income that you earn. Because of the interactions with the new £75,000 rate in Scotland, you are also paying more on top of that, and that is why you get up to that nearly 70 per cent marginal tax rate.

It comes back to the conversation that we have had: if someone is thinking about whether to work overtime, work an extra day, take on an additional role or put more money into their pension, those decisions are determined by the marginal rate. If the marginal rate is higher, people might take different decisions: they might not work that extra day if they are going to get only half the pay that their gross pay would say that they are entitled to.

For those two rates, we are looking at very different salary bands. Would we be more likely to see behavioural change in people who are in the lower band because a larger part of their income would be affected?

It is an interesting point. It would be good to get David Ulph’s reflections.

Yes, it might be a larger part of your income but, because the tax system is more progressive, more ends up being taken out as you move upwards. Also, typically, it is the people at the top end of income distribution who have the greatest flexibility to change their behaviours. They have the ability to take some money and become incorporated; they have the opportunity to put more money into a private pension; and they have the opportunity not to take on an additional role.

A lot of the people who will be in the £43,500 to £50,000 band will be those on pay as you earn and work 35 hours a week. Their additional pay award at the end of this year will now be worth less because more of their income will be taxed at the higher rate.

There is less flexibility.

Just to bear out that point, we tend to think about individuals being able to choose lots of things in relation to how and where they work but, for many people, those things are set by their employer. They do not have those choices.

When I was on the NHS pay review body, we talked with nurses who were working phenomenal hours not because of any tax considerations but because they loved the job. That is what would draw them to work the hours that they worked. You have to bear in mind that quite complex factors drive people’s decisions.

We talked about the behavioural changes that people can or cannot make, depending on their circumstances. If, for example, people on the higher rate are choosing to work less, choosing to incorporate or putting more into pensions, that means that money is potentially not being spent in their local economies. Is there any way of modelling that, or is there any way that that can be identified or forecast?

You would need enormous data sources to be able to do that because the problem is that the way in which people can respond to tax is so complex—there are many dimensions to it.

I have referred before to the fact that we tend just to look at taxable income, but that can be the outcome of many, many decisions that people make. For higher earners, that could also involve decisions around complex avoidance schemes. Those people will be very well informed, they will often be working in the city and they may know all the schemes themselves, so there are lots of dimensions along which people can move.

The decisions that people make can affect the level of gross national product because people who are not working so much are not producing so much. However, people who are using tax avoidance schemes are still working hard, so a lot of activity is still taking place in the economy—it is just not getting translated into tax in the same way. The implications in relation to the impact of behaviour on the economy and the impact on taxable revenue can be somewhat different.

This morning, João Sousa was talking after you on the BBC and he recognised that we will not know some of the tax implications of the budget for about two and a half years. He said that he had sympathy with you having to make forecasts, based on the difficulties.

We now have a number of bands and we have some marginal tax rates. We are not going to know the information for a while—how does that impact on your ability to forecast, and what are the dangers of that in terms of accurate forecasting?

13:30  

The key thing for us when we do our forecast is the accuracy of earnings data and projections in that respect. We can put that through our modelling, which captures things relatively well.

In our “Forecast Evaluation Report—August 2023”, we talk about the increasing weight that we are giving to the RTI data. When people from the OBR were up here earlier this month—it might have been late last month—they, too, talked about how the RTI data is much improved. There have been real concerns about labour market data across the UK, so the RTI data is the key thing for us to look at; it is, I think, the most important bit.

There are always going to be uncertainties. For example, when we talked about our forecast evaluation last year, we said that our concern was about self-assessment, which is the one thing that we have no information on at all; we do not have it, and it is not something that we can get. You could argue that if you were to create a more complex tax system, you might see more money and income coming through the self-assessment route from those who, as David Ulph was saying, could be more flexible and would use self-assessment in a much more significant way than the standard pay-as-you-earn system. However, that is the dark bit that the forecast just cannot see into, and there is a potential risk for us in that respect.

All of our forecasts are essentially very aggregate macro-level forecasts—that is, they are forecasts of the economy, average earnings and so on—and we do not get down to the kind of micro-level forecast that would pick up the effects that you are talking about. Some of the details about precise marginal tax rates do not matter terribly much with regard to what drives our forecasts on tax revenues.

Thank you.

That concludes the committee’s questions, but there are a couple of other points that I want to ask about.

First, I want to try and pin the behavioural response issue down a wee bit. In paragraph 35 of your report, you talk about £200 million gross and £82 million net, but, as we know, £74 million of that net figure comes from the new £75,000 band, and £8 million is from the £125,000 to £140,000 band. Where is the gross in those two figures? Are we talking about £8 million out of £40 million, £60 million or whatever? I want to get a better idea of where the tipping point is. I remember being in the Basque Country many years ago and being told, “If you put income tax up by 2 or 3 per cent more than the Spanish average, it does not make much difference, but if you go to 4 or 5 percent, it suddenly goes zzhhh”—that is, it falls away in a kind of reverse Laffer curve, so to speak. That is what I am trying to say. How much of the gross £200 million comes from the £125,000 to £140,000 band and how much comes from the £75,000 band?

When we talk about the gross, perhaps the easiest thing to think about is what we call the static effect, which, essentially, is when there is no behavioural effect.

Indeed.

If you make it all the way to page 117 of our report, you will be doing well—

Well, I only got it at half past 5 yesterday. [Laughter.]

On page 117, we talk about the static effects and how the advanced rate will raise about £147 million before the behavioural effect is taken into account, while the change in the top rate—that is, the additional penny—will raise £53 million. That is probably the easiest way of thinking about the gross, if you want to use that term. It is essentially what revenues would be raised if there were no change in behaviour.

Basically, then, you are saying that, out of the £143 million, you will generate £74 million, or slightly over 50 per cent, but from the £57 million you will generate only £8 million, which is probably about 13 per cent. That is interesting.

Yes—again, that is consistent. It is at the top end of the distribution that you have the greatest ability to change your extensive and intensive margins and where the big behavioural bit comes through.

Okay. I got to page 96 of the report, which is on the land and buildings transaction tax. No one has touched on that yet. We are looking at quite a significant decline, from £813 million this year to a predicted £730 million. In other words, there will be an £83 million deficit, which, incidentally, is £1 million more than we will raise from the projected tax increases in the two rates that we have just talked about.

However, you then go on to look at residential tax, which, over the next four years, is forecast to go up by about 56 or 57 per cent—I am just doing the sums in my head. Can you briefly talk us through that? I see that, in paragraph 4.88, you say that you forecast that

“house prices would rise in 2022-23 by 6.0 per cent and transactions fall by 10.8 per cent”,

when, in actual fact, house prices rose by a wee bit more than that and transactions fell by a bit less. Can you talk us through the land and buildings transaction tax element, given that we are talking about a significant amount of money and a significant decrease, going forward into the next financial year?

You are right. We are still predicting that the total amount of revenue will decline—indeed, we were predicting as much last year—but we are now predicting a slightly smaller fall in 2024-25 than what we predicted last year. That is partly because of our updates to prices and transactions.

Broadly speaking, though, that does not change the overall profile. I highlight figure 4.24, which shows the path of house prices; our forecasts are now higher, after our having lifted things up slightly in May and again in December. Overall, we think that prices will decline over the next couple of years from 2022-23—the year in which, we think, prices will have peaked—in particular, as the effects of mortgages translate through.

Interestingly, although the Bank of England has been increasing interest rates quite significantly, the full effects of that have not come through yet, because lots of people are still on fixed mortgages. It is only when you come off a fixed mortgage and have to remortgage that you face the new higher mortgage costs. The legacy effect of higher interest rates is still feeding through, which is why we are predicting a dip in both prices and transactions.

Yes—and interest rates are predicted to stay higher for longer.

Finally, we will have the same 6.7 per cent increase in benefits as the UK is applying to its own portfolio of benefits. Obviously, that is good news for people who are in receipt of the Scottish child payment, which will be going up to £26.70. However, if there is a tipping point in tax, where is the tipping point in relation to social security and labour market participation? Has any work been done on that?

We have not done anything on that, but the one thing that I will say is that the social security payments that we would look at—that is, those that are devolved to Scotland—are quite heavily concentrated on people of all ages with disability and ill health. We look at fewer of the social security payments that one might think would have an immediate labour market effect, such as unemployment benefit—or jobseekers allowance, as it is known. We are not modelling this as an economy-moving measure, because we do not think that people’s incentives are changing. That is, in part, because of the nature of the benefits.

Thank you very much for responding to our questions so succinctly and directly; I really appreciate that. We will have a two-minute break so that we can bring in a weird and wonderful collection of individuals known as the Scottish Fiscal Commission, who will be giving evidence very shortly.

13:37 Meeting suspended.  

13:40 On resuming—