Finance and Public Administration Committee

Meeting date: Tuesday, September 5, 2023

Official Report 691KB pdf

---

Scottish Fiscal Commission (Publications)

The next item on our agenda is an evidence session with the Scottish Fiscal Commission on its “Forecast Evaluation Report” and its paper on “Productivity and Fiscal Sustainability”, both of which were published on 29 August 2023.

I welcome from the Scottish Fiscal Commission Professor Graeme Roy, chair, Professor Francis Breedon, commissioner, and Claire Murdoch, head of fiscal sustainability and public funding.

I intend to allow up to 75 minutes for this session. Before we open to questions from the committee, I invite Professor Roy to make a short opening statement.

Good morning, and thank you for inviting us to give evidence on our latest “Forecast Evaluation Report” and our paper on “Productivity and Fiscal Sustainability”.

Forecast evaluation is an important part of how we work as an organisation. Our forecasts are an essential part of the budget. By evaluating them, we aim to improve transparency and confidence in the process.

I will start by talking about income tax, for which we now have outturn data for 2021-22. Looking at our budget setting forecast for that year, made in January 2021, we see a significant underestimate of the outturn data—by around 12 per cent, or £1.5 billion. Importantly for the Scottish budget, there was a similar scale and direction of error in the block grant adjustment, based on forecasts by the Office for Budget Responsibility.

What caused that underestimate? It is partly a good-news story in that the early 2021 lockdown was not as severe or as long as many had feared that it would be, and the economy and labour market recovered more rapidly than many people had anticipated. Because of the rapid global recovery in 2021, inflation also started to rise, pushing up nominal incomes. Another factor that we have seen is strong growth in tax revenues from the top end of the income distribution in Scotland, particularly among those paying tax via self-assessment.

As the financial year 2021-22 progressed, we published updated estimates of Scottish income tax revenues in that year. For that, rather than relying on revising our economy forecasts, we use real-time information—or RTI—on pay-as-you-earn collections from His Majesty’s Revenue and Customs to incrementally update our forecasts. By December 2022, our estimates for that year were based on complete RTI tax data.

That RTI data helped to inform our upward-revised estimates of Scottish income tax as the year progressed and as the economy opened back up. However, the committee will recall that, when we met in June, we were discussing a potential reconciliation of £712 million, based on a comparison of those latest estimates and those of the OBR. We now have a near-final reconciliation figure for 2021-22 of £390 million based on the outturn data. The difference between those two figures is explained by two factors. First, the outturn data was stronger than what was suggested by the RTI data. Secondly, there were higher than expected Scottish income tax revenues at the top end of the income distribution, particularly through self-assessment. We continue to see RTI as a helpful tool for monitoring income tax revenues in-year. We continue to look at what options there are to improve accuracy and to communicate the range of uncertainties that we face.

I will turn briefly to some of our other forecasts. Land and buildings transaction tax revenue of £848 million was £99 million higher than was forecast for 2022-23—an error of 13 per cent. Residential properties saw faster price growth, and there were more high-value transactions than was forecast.

Spending on devolved social security payments in 2022-23 was £4.2 billion, which was £127 million—3 per cent—higher than was forecast. That was a result of higher spending on disability payments and post-budget policy announcements by the Scottish Government. We might come on to this in discussion, but a higher part of that spending being on disability payments appears to be a United Kingdom-wide trend.

The Scottish and UK Governments recently published a joint communiqué following the fiscal framework review. Both Governments support this committee’s recommendation that we publish frequent full and supplementary fiscal sustainability reports, and I am pleased to say that we will develop that work over the coming months.

Last week, we published a paper exploring how changes to productivity growth would affect the projections made in our fiscal sustainability report. In that paper we conclude that, although higher productivity growth leads to faster economic growth, higher wage growth and likely better public services for the people of Scotland, on its own it does not necessarily translate into a more sustainable fiscal position. Improvements in fiscal sustainability are likely to require changes not only in the economy but also in public spending and tax policy.

Thank you very much for that, Professor Roy.

I will kick off the questions and then open up the discussion to members of the committee.

I will start with the issue that you have just touched on. I will not quote the full paragraph in your letter, but you have said:

“While higher productivity growth has a clear positive effect on the economy, the net effects on the public finances are complicated and to improve fiscal sustainability will require changes in public spending or tax policy.”

Professor Spowage talked about that at our away day last week. I would be interested if you could explain the reasoning for the fact that increased productivity might not reduce the sustainability gap. That is crucial to our deliberations as we move forward, and it is important to get that on the record.

The second issue is the changes that would be required to close that gap, even in a situation in which we have increasing productivity.

I will kick off. My colleagues will probably want to come in on that subject, too.

The first important point to make is that faster productivity growth improves the economy of Scotland overall and therefore potentially leads to better public services. However, if we think about the issue in a fiscal and public finance context, we see that certain different factors are going on.

Faster productivity growth leads to higher wages, which lead to higher revenues. That is a positive effect on the public finances. On the other hand, faster productivity growth leads to a faster-growing and bigger economy, and what we assume about what happens to public services in that context is crucial. The assumption that we make, which is the same as the OBR makes, is that, essentially, public services maintain the same share of the economy as the economy continues to grow. Therefore, as the economy gets bigger, the relative public services share will increase in line with that.

There are good reasons for that. Let us think about wages, for example. If wages in the economy grow at 3, 4 or 5 per cent, we would expect public service wages to grow at the same rate, to keep pace without having a gap between the two.

There are two effects. There is faster revenue growth and faster spending growth. In a devolved Scottish context, a proportion of revenues is from income tax, non-domestic rates and council tax. That is a smaller proportion in the overall spending envelope. Two elements grow. One is the larger public spending element, which grows with the faster growth in productivity, and revenue grows at the same rate, but one is smaller than the other. That is why the gap opens up. Growth in productivity does not necessarily lead to an improvement in the fiscal position.

What is crucial is what is done about the benefits of faster productivity growth. Do you choose to increase spending by the same amount or by less? That is ultimately a policy choice.

That goes back to our conclusion, which is that you would need to make a conscious decision to change the relative share of public spending in the economy, make a decision on taxation, or make a decision on how much growth you are translating through to the public finances over time.

You are absolutely right. That is, of course, a policy choice, and I am not asking you to make such policy choices, because you would demur if I did so. However, it is interesting that you have touched on the situation with the current devolution arrangements and have said that we would have to change public spending or tax policy. What would we have to change to bring the finances into long-term sustainability?

That goes back to what we said in our fiscal sustainability report. In the context that we have looked at so far, spending is running ahead of the funding that is likely to be available and is being driven by demographic pressures and the rising costs of key services.

Long-term sustainability will come from a combination of everything. Improvements in the economy will lead to better quality public services, which might mean that you can make savings on public services over time. Secondly, there will be choices about prioritising some areas over others, which does not necessarily mean cutting expenditure or reducing services but could, as was discussed before, mean a shift to prevention in order to reduce the long-term spend on healthcare during someone’s lifetime. Ultimately, there will have to be a conversation about what to do with taxation. Where might you look to increase the amount that you raise from the economy in order to spend that on public services?

Another point that is really important, as we have said before, is that the challenges that we are talking about here are not unique to Scotland but interact with the fiscal framework and with decisions made at the UK level. If the UK had faster productivity growth, we would get more money coming through the block grant. If Scotland and the UK were to have faster productivity growth, we would get more funding as a result of that, but, at the same time, there would be pressure at the UK level for public spending to rise in line with the economy.

There are lots of different moving parts, and we conclude in our report that there cannot be a purely economy solution or a purely public spending or taxation solution. There must be a combination of all three elements.

Obviously, if the economy improved, one would expect fewer people to be dependent on public services, and there would be the same impact if people’s health improved.

Exactly. We are speaking from a purely public finance perspective and on a numerical basis. If we have faster productivity growth and invest in public services, those public services will be better. There will be higher-quality public services and improved outcomes. We are saying that doing that means that spending will increase, so there will not be an automatic reduction in fiscal sustainability. I do not want to create the impression that faster productivity growth is not a good thing for public services, because it is. However, unless you take a conscious decision not to pass the benefits of that on to public services, you are not going to close the fiscal sustainability gap.

I think that that is pretty clear. Colleagues can ask further questions about that if they wish.

Let us move on to look in more detail at the forecast evaluation report. Page 19 shows that the number of top-rate taxpayers has increased from 14,700 to 18,000. Yet, on page 46, you say:

“the tax revenues paid by top rate taxpayers have actually fallen in two years.”

Why is that? If there is a significant increase in the number of such taxpayers, and given that we are in an inflationary situation, why have those revenues fallen?

I will make a general point first. We were surprised by the growth in the number of top-rate taxpayers in Scotland. The number of people at the really high end of the tax distribution was higher than was being reported through the pay-as-you-earn system and higher than the number we thought we would get. That offset a fall that we had seen in previous years, when the rate of growth in that number was less. There was an increase for just one year, 2021-22.

We have spoken in the past about there being 13,000 or 14,000 top-rate taxpayers in Scotland. That number is now up to 18,000. A chart on page 47 shows that the number of people paying the top rate of tax had actually fallen in the two preceding years and that there was then an increase this year. So, there is an increase this year, but that comes on the back of relatively weak growth in the preceding couple of years.

09:45  

Right—so, the revenue from that group is not decreasing. I asked the question because there seems to be a contradiction between what is said on page 19 and what is said on page 46. One talks about—

On page 46, we are saying that the tax revenues paid by the top band have fallen over two years, not in the last year. The tax revenues paid by the top band fell from the previous year in both 2019-20 and 2020-21, but in the year that we are now evaluating—2021-22—the revenues paid by the top band increased significantly.

I touch on that issue because about one sixth of income tax comes from 0.7 per cent of taxpayers. Throughout the document, you make a point of talking about the volatility of that specific group. I therefore think that we have to keep an eye on that issue.

Another point that comes through quite strongly in the report on a number of occasions relates to self-assessment. Looking at the graph on page 21, I was quite astonished at the jump in the outturn figures for self-assessment—they far exceed any others. Can you, for the record, explain that a wee bit?

Before I get on to that issue, I want to come back to your general point about the importance of the higher and top-rate taxpayers. The nature of the income tax system is that those taxpayers are absolutely crucial to the revenues that are raised, because of the progressivity of the system, and small changes in the number of people in the top rate or their earnings will have significant impacts on tax revenues. That is why that forecast is in there. We do our best to estimate the number of taxpayers and the changes that might arise, but the fact is that very small percentage variations can lead to significant swings in income tax. On page 46, I think, we say that the 18,000 taxpayers in the top rate are equivalent to about £2.5 billion. As a result, if you change the numbers and percentages, you can get large swings.

That brings me on to your question about self-assessment. All top-rate taxpayers—that is, those who earn above £100,000 and have complex tax affairs—pay through the self-assessment system. Even at the UK level, it is the most volatile element of the tax system, and it is subject to the greatest change, because it reflects a whole manner of different factors, including adjustments for previous years, sources of incomes, classifications and the like.

It is the hardest bit to forecast at the UK level, but it is even harder at the Scottish level, because of the data that we have. We do not get any self-assessment data until the January after the end of the tax year. We can look at, say, PAYE and track that through the year, but we will not get any idea of self-assessment until the end of the January of the following year. HMRC has said that many people pay their self-assessment returns in January, so we really have to wait until the very end to get to that data.

Secondly, the quality of the data and the information that we have about Scottish self-assessment is really quite weak. That is probably the biggest gap that we have in the income tax system. We do not have, for example, regional self-assessment data across the UK, but, again, that reflects some of the complexities with regard to taxing individuals. PAYE is quite obvious, because it is quite a simple mechanical transaction involving an employer and an employee, while self-assessment is much more difficult to pin down.

Self-assessment, therefore, has always been the greatest unknown in our forecast. When there is a 20 per cent growth in self-assessment, as the graph on page 21 that you mentioned shows, it becomes very difficult for us to forecast, and that is why the margin of error becomes that much greater overall.

You say in paragraph 26 in annex A of the report:

“The variation in tax revenue generated by the highest earners is likely to continue to be a source of significant uncertainty and forecast error, with very limited data available on this group. In the future, HMRC’s MTD project may improve the situation.”

You then say:

“To continue to improve our forecasts, we will focus on better understanding what determines changes in tax revenues of the highest earners.”

That brings us back to page 3, where you mention “points for improvement”. I take it that that is one of the areas that you are talking about in that respect. What specifically will you be able to do, given the paucity of data with which to improve forecasting in this volatile area?

I do not underestimate the difficulty. We are speaking to HMRC about what more information we can get on self-assessment, on the PAYE data and on how they relate.

I caution that the nature of self-assessment means that it is quite difficult to get the data in a Scottish context and that, even in what we can get, small variations lead to big swings.

Thirdly, I am keen that we be as transparent as possible and clear about the range of uncertainty and variability that could exist in the forecasts. We now have a number of years of data points that show how volatile self-assessment could be. We can factor that into our forecast and say that there is good potential for variability.

To give an example of the potential variability, PAYE real-time information—administrative data that is tracked by HMRC to show PAYE income in Scotland—picked up around 12 per cent growth in incomes, while self-assessment showed 20 per cent growth. It is therefore difficult for us to think about whether growth could be 20 per cent or 2 per cent next year. We need to get a better handle on that, doing our best not just to estimate it but to communicate that volatility and the fact that it has implications for how the Government plans its budget and the potential scale of the reconciliations that could come from the difference between forecast and outturn.

Even in retrospect, we do not really understand why that big surge in self-assessment incomes has happened. At the edges, we can improve our forecast in that area, but we will always have to manage its volatility, as Graeme Roy has said. It is a very difficult area of tax to get a handle on. Indeed, the OBR, which has a longer history of doing it, with more data, finds it extremely challenging. We have to live with the double problem with self-assessment: it is very volatile and it is pretty much the last data that we get, so there is always the potential for surprise right at the end of the process.

There is just one more question from me, and then I will open up to colleagues round the table. In paragraphs 8 and 9 of annexe B to your report, you say that

“Scotland’s lagging earnings growth since 2016-17 has been exacerbated by much stronger earnings growth in the financial services sector in London and the South East”,

and that

“Scottish employment linked to activity in the North Sea has ... fallen, lowering the average participation rate in Scotland. These jobs ... were generally high paying, which has likely contributed to the divergence in average earnings between Scotland and the UK.”

What kind of pace are we talking about? What decline in overall earnings is coming from that sector, relative to the rest of Scotland? Obviously, that is fundamental—including for the green transition process that we are trying to undergo.

We can get some updated numbers. We had a box—not in last December’s forecast, I think, but maybe in the May or December before that—with a bit of analysis about trends in employment and income in the north-east, and we saw that divergence. We can send an update on that to the committee, but it shows far less employment growth and far slower earnings growth than in the rest of the country.

That comes back to the point about the importance to the forecasts of the progressive nature of income tax. People in oil and gas tend to be on high earnings, because those are difficult places to work and involve high profits and risks. If that sector does not grow, or if jobs are lost there, that has a disproportionate impact on the tax revenues. That is one of the factors in the opening of the gap between Scotland and the rest of the UK.

From a policy perspective, the key is to think about how we transition those high-paying jobs, which we know will disappear in the next few decades as oil and gas transitions, into other high-paying jobs in other sectors, including renewables but also in broader sectors of the economy, in order to offset the differences in earnings. In the fiscal framework, what matters is the relative growth in earnings.

Given that we are coming out of a pandemic and that we have the Ukraine war, the cost of living crisis and high inflation, a GDP forecast error of 0.2 per cent is remarkable. I know that there have been one or two areas in which the figure has been quite high, but the Scottish Fiscal Commission has still done an excellent job.

I open up the meeting to colleagues.

For clarification, you mentioned that there is not enough self-assessment data for Scotland. Is that unique to Scottish self-assessment data, or is it a UK-wide issue but one that matters more to us because of scale and the way that our public finances work?

There are a couple of reasons. Self-assessment is always difficult to forecast, even at the UK level. The OBR would tell you that it is one of the most difficult elements for it to forecast. However, it does not matter as much to the OBR, because it is one element of income tax among a multitude of other taxes. In the Scottish fiscal framework, it is a key element of the one big tax that we have. If the OBR had forecast errors of 15 or 20 per cent in self-assessment tax, it would just get washed up with everything else. However, it matters more in the Scottish context.

There is also an extra element. At the UK level, identifying the location of self-assessment taxpayers is more difficult than is the case with the PAYE system, where there is an employer and an employee in residence, which makes it very simple. Self-assessment is much more difficult, and we do not get the information until later.

A combination of those factors makes it difficult to forecast on self-assessment.

In Scotland, we are forecasting based on when tax is due—when people earned their self-assessment income—but at the UK level they are slightly more worried about when the cash is coming in. That difference means that the UK gets real-time information on how much self-assessment is paid during the financial year. Ultimately, that is what the UK Government cares about when managing its budget and borrowing, but the way that the Scottish Government has to manage its budget means that we are a lot more concerned about when the tax was liable and not when it was actually paid to HMRC.

You mentioned that you have been speaking to the Treasury about that. We can ask ministers about this when they come to give evidence, but are you aware of whether the Scottish Government has engaged with the Treasury on those points?

I do not know. We do it from a purely technical point of view. We are dealing with analysts and officials to try to get their insight, because they have been doing this for much longer than we have. We try to get their insight about the quality of the information and the quality of the data.

Our forecast in May was a surprise on the downside—the outturn data was higher—but so was the OBR’s; the OBR forecast for Scottish income tax was similar to ours. We are all trying to wrestle with this to get a better understanding of the drivers of the flows.

It is fair to say that not much is going to change with regard to the self-assessment data. We are looking for other indicators that will give us a clue as to what the self-assessment data will be. That is the sort of thing on which more data is coming, because the range of stuff that we are getting from RTI is expanding.

Secondly, we struggle across all our forecasts—this is true of the forecast evaluation—because we still have a very short history of data to work with. When we try to do statistical analysis to find out what the issues are, five observations are simply not enough to give us a lot of that, so the passage of time will help improve the way that we do this.

On a somewhat similar point, about the challenges of the short period in which we are doing this, I am going to ask the same question that I ask every year at this point about the work that you do to look back at behaviour change estimations that are related specifically to changes to income tax. We started making significant variations to the UK from 2018. Given that we are now getting somewhat further away from that point, and recognising that it is hard to disaggregate that from all the other changes that might result in a change in the revenue that is eventually raised, do you have any further observations about whether your estimations on behaviour change related to income tax rises have borne out?

One thing that we are talking about and looking to do, as we now have more data points, is whether we can start to look at some of the data that is being collected to see whether there are patterns of behaviour and change. However, it is still quite early. I know that you will probably be fed up of me answering that question in that way, but we will start to get information that enables us to see changes in categories in the PAYE system, such as earnings and residence, although it will be subtle.

10:00  

We have talked about behavioural change. A lot of the time, people think that the obvious behavioural change is that, if you put up tax, people will leave or, if you cut it, they will come in. The behavioural change that we are talking about is really subtle. It is about people changing their earnings, changing how much they work at the margin or changing their classification between non-savings, non-dividend income and dividend income. It is very marginal, so it means that we will need data for a longer period to see what happens.

Thanks very much.

Good morning. Professor Roy, I come back to an answer that you gave to the convener about the important issue of the increase in public spending and tax revenues. You set out that increasing productivity is all very well but there are issues about the rate of change in the public spending commitment as well as the rate of change in tax. That is what will be crucial.

I relate that to a comment that David Bell made:

“Nevertheless, given the potential consequences of a widening fiscal gap on the ability to provide public services, it is important that as full an understanding of the causes of changes in demand for public services and changes in tax revenues be available to”

the Scottish Government as possible. How easy is it to get the necessary data to understand what is causing the changes in demand for public services? That is critical to policy decision making.

To be clear, in our report, we take the existing demand for, and use of, public services and push it forward with the changes in population. We do not make a judgment about changes in drivers of demand within that. We just say what the entitlement is, how much we spend on healthcare at this moment, age the population and ask what the additional increases in public expenditure will be. We do not make a judgment call about demand.

The quotation that you gave from Professor Bell starts to get to the heart of what you do about that. You can continue to have the same amount of public expenditure but shift the balance within it so that you change the drivers of the demand.

An example is the high amount of healthcare expenditure that we have. We just push that forward but, at the moment, a lot of that expenditure is on corrective surgeries or addressing ill health. If we can change that to be more about preventative spend and reducing demands over time, that will change how much money you put into the healthcare budget over the longer term. That would allow you to make choices about how much you spend.

The question is how you change what you choose to spend that money on. That could come through discretionary choices to shift balances of expenditure or by changing the demand on those services, which is the point that David Bell makes.

I am interested in the issue because, as you rightly said, it gets to the heart of policy making. I fully understand the Scottish Fiscal Commission’s role in the matter and that you are very much involved in the quantitative statistical analysis, such as projecting demographic trends and how the population is changing. However, what matters to policy makers is whether there are trends within those demographic changes that lead to changes in demand. That is the crucial point for the Scottish Government. From the work that you do over years, would it be possible to isolate some trends in that change in demand or would the commission not do that?

We have done it to an extent, in the sense that, in individual portfolio components, we change the demand. For example, in our projections, we reduce the share of expenditure on education because we have fewer young people, and increase it on health because we have more older people.

The question is whether we could go a level below that. Within the components, we make an assumption that the average spend on health increases as the population gets older. Could we potentially consider whether, within that, there are different elements that we could use—for example, if there are higher proportions of degenerative diseases, could we change the margin? That is one potential area we could look at.

In our fiscal sustainability consultation, we spoke about looking at what could be driving health inequalities, which lead to higher demands over time. We considered what the effects could be if that changed. Drilling into those elements is one thing that we could do.

I ask the question because, before the summer recess, we held a number of sessions with witnesses from the public sector. As a committee, we have been asking what we can do to progress public sector reform. The answer depends on some of the stuff that you are saying, because knowing what the future demand for some public services will be, and whether demand will increase or fall, is absolutely critical for the Scottish Government to be able to make sensible policy announcements. I am anxious to know whether we think that we have the right data to enable us to get to that set of information, so that the Scottish Government, and the committee, are able to make sensible policy suggestions.

We can definitely have a look at that. We can see what we would do with our projections and individual components of them. A key point is that our analysis shows that, if you try to make any adjustments on the spending side, essentially, you need to reduce spending per head over the lifecycle of someone demanding healthcare services. If you do not do that, and if we all demand at the same level, as we become older, expenditure will increase. You would need to consider how to shift into things such as prevention, which gets you into challenges about how to prioritise.

We could consider doing a deeper dive into an area such as healthcare and the different components in the health budget, looking at what might happen if we project forward for particular elements. We would want to be quite careful that we were not making choices about health policy. However, we could consider whether we could decompose the elements of health into some that are particularly tied to demographics or to inequalities.

I understand the point that you are making—it is not your job to suggest the policies. However, to be effective in policy making, we need the right data. In your sustainability report, you have projected increases in health spending, social care spending and social security spending over 50 years. Those are all big asks. It is helpful to understand what the demand is composed of and whether there are other areas of public sector spending in which there might be a little more scope for efficiencies. That is a dilemma that faces the committee.

We could potentially write a bit more about the individual components that are underneath the portfolios when we do our update. In areas such as education, it is relatively simple, which makes it obvious to split—it is high schools, universities, colleges, primary schools and nurseries. However, healthcare gets a wee bit more tricky, because it does not go into sets of boxes. In health, we could look at different elements and consider whether there is data available that could help to provide an idea about the different pressures that may be felt as a result of demographics and age over time.

My final question is about preventative spend. I find it incredibly difficult to measure, because it is almost an unknown. However, it matters as an opportunity cost as well as in considering how much money we could save because of it. Do you have any guidance as to what methodology we should use to look at preventative spend? It will matter for policy.

Perhaps less can be done on that within the context of our work on fiscal sustainability, because the aggregates that we use are so high. For example, a key aspect of reducing health demand over the next 10, 15 or 20 years could be housing, or it could be education or employability. It would be much more difficult to pin that down.

That takes us to the point where the evidence base must become better. If we invest in a particular area, we must know what the long-term outcomes are. We should be better at that and there should be more tracking and evaluation of programmes. We have had expansions in childcare and changes in employability policies and in different elements of healthcare and education. If those are preventative policies that are designed to create long-term outcomes, the key question is whether the evidence shows that they have actually had an impact. Once you have that evidence, you can update projections to see how you might tackle issues over time. That is the sort of thing that we will cover in our work on health inequalities.

There are very good microstudies—done not by us but by others—of some of the work on preventative measures and on the cost of benefits. Those studies evaluate the impacts of individual measures. As Graeme Roy said, we do not go into that level of detail, but those studies are a good source if you are thinking about particular areas.

I am going back to some of the predictions and forecasts that were made, both by you and by the OBR. I accept that the war in Ukraine was probably not predictable—although some experts in the field might say that it was—but neither the Fiscal Commission, the OBR nor wider society saw the rise in inflation coming. In retrospect, should we have been able to predict that inflation was coming?

The projection was always that inflation would increase after the pandemic. The significant expansion in monetary policy at the time was designed to prevent a slowdown in the economy, so there was always going to be an increase in inflation. The real surprise that we and economists did not fully anticipate was that the dislocated start-up across the global economy immediately after the pandemic meant that supply chains across the economy became gummed up, which led to a significant spike in manufacturing prices, particularly because of what was happening in China.

If you look at the track record in forecasting, there was a forecast increase in inflation, but that grew as it became more obvious that global supply chains would act as a brake. On top of that, there was the war in Ukraine and the spike in energy prices, which led to the really significant increase in inflation that we have seen.

There is still a debate about how quickly inflation will come down. Francis Breedon knows the area far better than I do and may want to come in. Every time we get monthly data about inflation, everyone jumps in to comment on whether the change is below expectations or is a faster than expected fall. We know that inflation is coming down, but the question is how sticky it will be, which takes us into questions about wages and so on. Inflation was always going to increase, but I do not think that anyone expected it to increase as much as it did.

I have had quite a long career in forecasting. If the conditions that we had prior to the Ukraine war, with tight labour markets and trade shock, had happened in the early part of my career, we would have been forecasting very high inflation. What we have seen since the era of low inflation is that inflation has been incredibly anchored. There have been all sorts of shocks for so many years and we all got used to the idea that you could shock anything and inflation would stay anchored. That was the mistake that we all made. Perhaps old hands like me should have come in earlier. It is not that I thought that this was going to happen, but it was what we were used to in the 1970s and 1980s and we have not been used to it for a very long time.

I remember 15 per cent inflation. We just lived with that, but people are not used to it now. Do we have to accept that the same thing is likely to happen in the future and that we might see a bit of inflation coming but that it will be hard to predict how much?

The idea of incredibly anchored inflation is gone now; the anchor has been lost. We are not back in the 1970s and 1980s, but we are in a period when the economy cannot be shocked in any direction while inflation stays at roughly 2 per cent. We are in a world where we will have to get used to a bit more volatility.

I just want to touch on some of the minor taxes, although I accept that their impact is much less. Someone commented that house prices are unpredictable and had risen more than was anticipated. Was that just purely because of the inflation factor?

10:15  

It is a combination of things. Higher prices and higher-value transactions were the two drivers on the upside. Overall, there were fewer transactions than we were expecting, so there was a smaller turnover of houses, but the houses that were being turned over had a higher value and increase in price. The situation is quite volatile and we have underestimated and overestimated in the past, but we are getting a better handle on the genuine volatility that exists at the moment.

On landfill tax, the comment was made that incineration capacity was limited and uncertain. I do not know whether that was about machinery breaking down or what it was, but I assume that it meant that more had to be put into landfill when it should have been incinerated.

Yes. We make assumptions about what can be incinerated, and we forecast that more would be incinerated than was incinerated. In a report, we talk about various aspects, including capacity, new things coming on stream, faults in incineration and so on, that vary that forecast.

The bigger question is about social security. We now understand that the Scottish Government is generally nicer than the UK Government and so is more generous in handing out social security. Is that built into the system now? You seem to have slightly underestimated how much was going to be paid out each year.

There are a couple of things going on there. We were forecasting 2022-23, when the adult disability payment, which is the really big one, was only just coming on. It is therefore less about what the Scottish Government has been doing with the new system relative to what has gone before.

Our error on social security is essentially driven by two elements. One, which is a smaller element of around a third, is a change in policy by the Government since we made our forecast, and that was the introduction of the Scottish child payment earlier than we had anticipated.

The main change and therefore the main reason for the forecast error was the growth in adult disability payments, but we do not think that that is due to the Scottish Government changing policy or your interpretation of the two Governments’ niceness. It is more to do with trends in disability payments across the UK.

You have probably heard that a lot of discussion is going on at the moment about the increase in the number of applications for disability benefits across the UK and the rise in inactivity among people who are in ill health, perhaps as a hangover from Covid or perhaps because of broader concerning trends in things such as mental ill health. Across the UK, we are seeing an inflow of people on to disability payments that is much higher than it was pre-Covid, and that affects their number in Scotland. Our error therefore mirrors the error that we see at the UK level, and that is probably more of a reason than a differentiation between Scotland and the rest of the UK at this stage.

I will move back to income tax to carry on with some of the points that have already been raised. I get the point about higher earners and self-assessment. If people are self-employed, they might not know what their earnings will be for a year, let alone anyone else knowing. However, the earnings of higher earners such as the chief executive of a council or someone like that, who, I presume, will be in the self-assessment category, would be quite predictable. Are they just swamped by the ones that are unpredictable?

Yes. Some people who do self-assessment pay through PAYE and we get the returns in that way, but other people have much more volatile and complex tax affairs. Self-assessment does not just cover high earners; it covers people who have complex tax affairs and that is the part of self-assessment that is much more volatile.

In our report, we also talk about the difference between the RTI PAYE data that we can track and the outturn data on PAYE, and again, we see differences in that. This is about people perhaps changing tax codes at the end of the year through being identified as a Scottish taxpayer rather than a UK taxpayer, for example. That just means that the PAYE data that is collected through the year is different from the PAYE outturn data at the end of the year. People having complex tax affairs mean that those numbers move around a lot.

That is a little bit disappointing, because, when you were at the committee previously, I got the impression that the RTI information would be the big answer and would give you a lot more clarity. Has it disappointed you a bit?

It is surprising just how different it is this time around. It is still a really helpful indicator, because it meant that we were uplifting our forecast all the way through. If we can see from administrative data what has been collected by PAYE through the year in Scotland, that is still really useful. The bit that we do not know at the end is how it looks once people have reported, once all the changes have been made to people’s tax codes and once various adjustments have been made. It might just be because it is the end of Covid and it has been a volatile tax year with lots of things going on, but that gap—the error—between the PAYE outturn data and the PAYE RTI data is bigger than we had hoped it would be.

The other surprise is that Scotland was stronger than the rest of the UK. We worked on the assumption that self-assessment would be roughly the same across the UK, so the divergence between the two parts was a bit of surprise.

Is the self-assessment money in Scotland more volatile than that of the rest of the UK, or are we not sure about that?

The potential for it to be more volatile is higher in Scotland because of the numbers, particularly if we are talking about people at the really high end, because there is a small number of them. Percentage variations are likely to be higher, as we see with top-rate taxpayers, because there is a smaller number of them and it is a smaller sample size; variability is likely to be higher.

We have too little data to say anything much more than that.

There is an interesting chart on page 50 of the report that shows the self-assessment revenue growth for Scotland and for the rest of the UK. You can see that the numbers for Scotland are more volatile than those of the UK.

I would keep the warning that, with that many observations, I am nervous about making a strong statement about volatility.

That does not necessarily reflect the economy or anything like that; it is much more to do with just individuals.

As I said, on the latest information, there are 18,000 taxpayers in Scotland who pay the top rate, which amounts to about £2.5 billion. If you have a few hundred of those moving around, that number becomes that much bigger. It is just a scale effect. It is also because, due to the nature of the devolution settlement, we have this one tax, which is the most important. In the UK context, it would be matched by errors in VAT or corporation tax that would offset it all, but it becomes so much more important in our context.

The final thing that I want to touch on is the fact that the total number of taxpayers has increased by 5.1 per cent, if I understood that correctly. Perhaps linked to that, the immigration numbers were higher than expected, with immigration from the rest of the world rather than from the European Union. Can you say anything about that?

We revised up our migration forecasts last time, in part through the international element. The growth in the number of taxpayers can be due to things such as people moving into the system. It does not have to be about there being more people in Scotland but, if there is a higher number of people in Scotland, that will clearly lead to a growth in the number of taxpayers. That growth could also be due to people moving into tax bands. Measures such as freezing the personal allowance and not increasing the personal allowance by inflation mean that, as people’s earnings go up, more people move into the tax system.

There is an interesting table on page 19 of the report in which we talk about the growth in the number of taxpayers. What we see is people moving into higher tax bands, which is quite interesting; it potentially affects the fiscal drag and where earnings growth is concentrated in the labour market. Unpicking that and understanding it will be really important for our updated forecasts. Are we seeing something of a trend in higher earners’ earnings growth contributing to higher income taxes?

I suppose that I expected people to move up the tax bands as the limits were frozen, but I was a bit more surprised that the total number of taxpayers had increased.

Yes.

Fair enough. Thanks, convener.

A quick wee calculation off the top of my head tells me that, if there are 18,000 top-rate taxpayers, they each pay an average of just under £140,000 a year. That is a very interesting section of the tax-paying public indeed.

I will stay with that area for a couple of minutes, if that is okay, convener. It is a subject of significant political debate, so the panel will understand why we are particularly interested in it. Given the level of volatility—from the sounds of things, that is partly tied to how few of such people there are and to small variations—how much confidence should we as a Parliament and the committee have in your work and predictions? Do Government predictions about the contribution of the top rate of tax to cover the deficit come with a strong health warning?

I think so. As we get more data points, we can offer much more clarity about the potential for variability, change and uncertainty that exists. Given the nature of what we do, we must make a central forecast. Around that, we can do more to talk about the potential variability.

We have a really good data point—the RTI data and all the information that we had suggested a reconciliation of about £700 million, but it has turned out to be about £400 million. Next year, if I tell the committee that the reconciliation is £200 million positive, I will be able to ask the committee to remember the potential variability that we saw when we were here last year. That really matters for top-rate taxpayers, too.

We are talking about one data point. We do not know whether this is a really good year for Scotland. Does it represent a bounce back after Covid, or is it part of a trend over time? It will be important to be clear with the committee about the changes that could happen and how to interpret that.

The Fraser of Allander Institute has told the committee that it could take up to a decade to provide sound modelling of behaviour effects. Does that sound realistic? You have said that having more data points is a good thing.

Exactly—we will be able to see more over the next few years. HMRC has been working on pulling together data sets that can be used not just by us but by academics to look at the effects and see whether changes are happening.

We need a combination of data points, so that we are not dealing with individual years, and we need change. You want Governments to change policy, because you can see behaviours moving after that. In many ways, Governments announcing and doing things is quite good from a statistical point of view, and we can see what happens over time.

It is slightly amusing that we would like you to roll the dice and set some policies so that we can see what happens. I make the point every time that, although we have errors around our forecast, the errors are on both sides. If there is a really big number in our forecast, the wrong way to interpret it would be to say, “They make mistakes, so we can just ignore it.” If we have forecast a big number, you have to say that, in practice, it could be even bigger or could be smaller—we do not know—and you have to work with that number and not say that it is not worth anything because it has an error around it.

It strikes me that, five years down the line, there could be such significant externalities from the behaviour of such a small group of people that the data points that we have talked about might be moot.

That could be the case. There are various techniques for reflecting on that. You can match taxpayers across the UK and control for different effects. In relation to LBTT, stuff has been done to look at changes in tax policy and at people shifting forward—you can see behavioural effects there. The UK has decades of income tax data, so there are lots of detailed studies about how people shift and change behaviours to move figures up or down. We just need such data sets.

You talked about the increasing level of applications for adult disability payment. Is that matched by an increasing level of awards? Is there a differential between awards in the UK and in Scotland?

Across the UK, we are seeing an inflow of applications and an inflow of people claiming personal independence payment in the rest of the UK and adult disability payment in Scotland. We are seeing increasing applications, but I caution that we do not yet have definitive evidence of a disconnect between Scotland and the UK. We think that that will happen, because of the nature of the policies, but at this stage all that we are seeing is an inflow. We need to see more data points before coming to a conclusion.

10:30  

There are more applications, but is there a higher proportion of awards? Are more being approved?

We do not have that information yet. Social Security Scotland has said that there are processing delays with the applications, so we do not yet have information about whether the large number of applications that we expected to see—and have seen—for the adult disability payment is translating into a different success rate in terms of whether people receive the award and what the payment amount is.

What Graeme Roy is saying is that a lot of the spending that we are classifying as being on the adult disability payment is actually still spending on the personal independence payment, because people have not yet moved over to the Scottish system. Across the UK, we have seen that increase. At the moment, in most of the forecast area where we are seeing higher disability spending, it is to do with that UK-wide trend. We are not yet seeing evidence that there is no additional cost in Scotland. It is as we would expect. It just takes us a bit longer to get the actual data.

It would be useful if, as you do more of that analysis, you could apply a time series to it as well. It would be interesting to see, for example, whether there is a spike in awards versus applications against a particular timeframe. Some anecdotal evidence that I have heard suggests that part of the issue about delays in awards might be driving an increase in awards. It would be useful for the committee to have a way of understanding that, given how significant the area is in terms of spend.

In our forecasts, we are clear about what we are forecasting in terms of spend, average payment, applications, case load and so on. That allows us to have a discussion about what that shows us.

You will remember that, when we came to the committee last year, we spoke about data needs and so on. We were keen on having much greater dialogue with Social Security Scotland about exactly those qualitative indicators, so that we could get a sense of what is happening with the case load, what the trends are, what its expectations are and so on, and then make judgments based on that information. We are more than happy to set that out in our forecast in December.

Is Social Security Scotland open to those conversations?

Yes, it has been very helpful.

Excellent.

My last question concerns modelling around productivity and childcare. In recent years, the country has invested a significant amount in the uplift in childcare to 1,140 hours. What can you tell us about the relationship between that investment and the productivity of the economy?

There are good theoretical arguments to suggest that investing in childcare leads to improvements in productivity in terms of better access to the labour market, improved skill levels, jobs and so on. I have not seen any evaluation of whether the policy has had the intended effects in a significant way or what the indications have been in terms of dead weight, so that has not informed our forecasts and we have not been able to say what the impact of the policy has been in terms of productivity. We cannot say anything specific about it in a Scottish context.

It is a significant area of public investment. Who would you expect to be doing that analysis? If increased productivity is one of the aims of the Government’s investment of taxpayers’ money in that area, we should be seeing those numbers.

That comes back to an evaluation of policy. When significant investments are made, you would expect to see the evaluation of policy. If we can see evaluations by Government or academics who have been commissioned to do them, we can use that to inform our assessments of what might happen in terms of productivity or changes in policy over time. However, I have not seen anything that we could use in the way that we have used data in our forecasts in the past.

Am I correct in thinking that the broad indicator is that there has been little growth in the productivity of the economy in recent years?

Yes. That is part of a much bigger conversation about what has been happening to productivity in the UK context and issues around measurement, management, changes in sectors and so on. The two key ways in which you would expect changes in childcare to come through are in labour market participation and, crucially, in productivity. In theory, we could expect to see positive economic improvements from the policy but, as I say, we have not seen any explicit evaluations that would enable us to say that an investment of X amount has resulted in outcome Y.

I think that about 12,000 more people are employed in the sector. However, the productivity outcome is not something that we looked at when we were doing post-legislative scrutiny in this area a year or so ago.

Good morning. I just want to follow up on a couple of points that have already been made. On the point about productivity, what would be the impact if there were high UK productivity but low Scottish productivity?

In terms of our fiscal sustainability?

Yes.

This is where I have to think off the top of my head.

In that instance, you would get faster growth in the block grant. Spending at UK level could go up, which would then feed through to a positive position in the Scottish spending component; however, it would also mean a faster growing block grant adjustment relative to tax revenues, which would lead to a squeeze. Whether the net effect of that would still be positive or—

I think that it would still be positive. Faster productivity growth in the UK is positive for Scottish finances.

I was interested in the discussion that was had with regard to Michael Marra’s question about top-rate taxpayers, of which there are 18,000. You have said that we do not yet have the data from the Scottish context; however, we obviously have historical UK data. What analysis can be done on the tax mobility of those individuals, admittedly within an international context? Of course, such mobility would probably be even more beneficial within a UK context.

In our forecasts, we look at what are called tax elasticities, which are about how people potentially respond to changes in tax, and we look at, for example, average and marginal effective tax rates. The marginal rate is all about how people change their behaviour at—as it would suggest—the margins, for example, by choosing to work more hours or shifting to different types of income, while the average rate relates to people making much bigger strategic decisions about whether to work or not or whether to move, relocate and so on. We build in assumptions about such effects, based on international and UK evidence, and it means that every time that we make a forecast, we have both a gross and a net estimate. The net estimate is the effective one that accounts for these factors.

Coming back to the discussion that we have just been having, I would point out that we do not yet have a specific Scottish elasticity that captures people’s behaviours, because we just do not have the data or the time series for that. We therefore use other people’s evidence of what has happened elsewhere to help inform the Scottish context.

When you look at the data from the UK—or, indeed, from other countries, depending on how you do it—you must find additional flexibility within, say, the UK market. Is that what you expect to see?

When we have looked at international evidence, we have very much focused on countries with similar situations, such as Spain, for example, where there are different complexities across different regions. For a period in the United States, there was a thing called the millionaire tax, as a result of which very-high-income people paid higher tax in some states but not in others. We have drawn on those kinds of specific examples, because, as you have said, the tax situation is different with internal rather than external migration, and that has informed our current numbers.

However, as Graeme Roy has said, we would ideally like to do something specific on Scotland to find out the answer in that context. As I have said several times today, though, I am afraid that it is still too early to say anything about that, but a Scotland-specific study of such effects is certainly very much on our agenda. In any case, the international evidence from all of those countries is relatively similar: there is an effect, but it is not very powerful.

Because there could be various reasons for people choosing to relocate. It might not be because of tax.

Yes. Perhaps the US example is the most surprising, because people stuck to whatever state they were brought up in or had chosen to live in. Even though the US has a very mobile economy, you will find that the tax difference did not generate a very big amount of movement.

I just want to ask briefly about a couple of other things. We have talked about fiscal sustainability over the next 50 years, but I wonder whether that can be broken down on a regional basis. In the Highlands and Islands, which I represent, healthcare and other services are very difficult and more expensive to deliver, and one would therefore expect the fiscal sustainability of those services to come under even more pressure. How much can the data forecast or reflect that?

That is a really good question. There is a bit in the fiscal sustainability report where we looked at population projections across the regions of Scotland; I think that we looked at local authorities. We saw huge variations, particularly between places where there are population declines, such as rural areas and old industrial parts such as Inverclyde, and the east of the country, where there is increasing expenditure.

We did not go into the detail of that in the report, because we look at the national level, but that variation at the local level means there is a really interesting subtlety around how fiscal sustainability can be managed on a more local basis. On the one hand, there are places such as Edinburgh, Midlothian and East Lothian that have rapid growth, and there is pressure on spending as the growth increases. On the other hand, in rural areas, there are potential increases in the costs of delivering the same quality of service to fewer people, and that has the potential to become really expensive as well. If we look at that level below, we see those interesting subtleties around what the fiscal sustainability questions could be at a regional level, and therefore at a national level. It could potentially become more costly overall because of the pressures of delivering the same level of service in areas where populations are growing and, equally, where they are declining.

Although you do not look at that aspect, the data is there to be enable those projections to be made. In rural areas and island communities, such as the one that I live in, there is already concern about how expensive services can be, and I will not even go into issues such as ferries. Lots of decisions to do with spend there will need to be made in the future. The concern is that that pressure is only going to grow, so how does the Government prioritise that spending, or at least ensure that that spending is there?

I hope that the principles that we have used in the fiscal sustainability report will give people who are interested in the local context ideas about doing something similar to what we have done: looking in a very transparent way at population and costs of delivery and then pushing them forward. These findings are not forecasts; they do not say “This is exactly what will happen”. They are just raising awareness that if you do the arithmetic on what happens around the cost of service delivery and population demographics, these are the pressures that you will see.

My last question is about house prices. There was an underestimate of the revenue from the land and buildings transaction tax. Given that there now seems to be a slowdown and houses are taking longer to sell, would you anticipate that there might be an overestimate for next year?

We lowered our forecast. We thought that LBTT would be under more pressure next year because of falling prices and fewer transactions because of rising interest rates, the economy and so on. That had an effect on our forecast. When we made our forecast in December last year, we were more cautious about the growth in this year, because we thought that the slowing nature of the economy would have an impact on it. We will not know until next year whether that has had an impact.

We can make intuitive judgments, recognising that interest rates have gone up and the economy is slower—for instance, that house prices could fall and revenues might be under pressure. However, if there are a small number of high-value transactions in some key areas, or if house prices are resilient to changed interest rates, we could be surprised on the upside. That is just the nature of forecasting.

Thank you.

It has been an interesting discussion. I have three quick questions to round things off. Professor Roy, you quite rightly said that forecasts play a central role in setting the budget, but you have highlighted forecasting errors in revenue from income tax at 12 per cent and from land and buildings transaction tax at 13 per cent. What impact did that have overall on the Scottish Government’s budget in terms of what was available to it?

In broad terms, one thing to remember is that the block grant adjustment is changing as well. The income tax forecast error of 12 per cent should be couched in the fact the Government also had to deal with the BGA change. What really matters there is the reconciliation, which was ultimately £390 million. That, essentially, is the consequence of those forecast errors on income tax. That gives you the scale or the size of the issue. It comes back to the discussions that we have had before about the fiscal framework and the ability to manage that risk in future years.

You estimated that the reconciliation would be £745 million.

That is where it gets quite complex. Members may remember that, while the nature of our forecast back in January 2021 was that we were forecasting higher revenues than in the rest of the UK, we were cautioning to say, “Hold on a minute—we think this is a timing issue. The UK BGA is going to be revised up.”

We then updated that forecast over time, and we knew that there was going to be a big reconciliation. We did not know how big it would be—I cannot remember what we were saying at the start; I think that we were talking about £300 million or £400 million.

10:45  

As we were getting the real-time information data, we were saying, “Look—it looks like it could potentially be even more significant”, and that is when we pushed it up to £700 million. It has come in at £400 million, and that is ultimately the forecast error between ourselves and the block grant adjustment.

The difference between the higher and the lower forecast element could have been resource that would have been available to the Scottish Government.

With the £390 million, the reconciliation there is essentially money that the Government has had and has spent, and it is now essentially having to manage that back—

With the timing differences.

Exactly. That is the fiscal framework.

It is not money that the Government has lost—it just has to manage how it then deals with that forecast error.

It has to manage a lower amount than was previously estimated.

Yes. The £390 million is now going to come out of the budget next year and that, ultimately, is what we have potentially to remove.

Have the forecast errors led to changes in the model going forward, so that you can reduce the level of forecast errors?

Again, there are a couple of points to make. I would not say that we are going to change much on our forecast—the budget that we push out. That is the one that really matters; that is the £390 million. That comprised our judgments on the economy, employment, earnings and so on, and they were pretty solid and accurate.

What we have been discussing today is how we update the information as we go to give the committee and Government a potential real-time indication of how accurate these forecasts are going to be. That is the discussion that we have had about the RTI data, which is essentially tracking administratively pay-as-you-earn data for Scotland, so it should be really good. This year, we see that, instead of predicting 14 per cent growth, it was predicting only 12 per cent growth, and that is where the error is.

I come to my final point. I accept that you may not have seen these figures, but we were talking about the volatility around self-assessment taxpayers. What impact do you think the acceleration in the move to a cashless society has had? A report last year said that 23 million people in the UK no longer use cash, and that within a decade, the proportion of cash payments would be close to 6 per cent.

It is an interesting question. I have not thought too much about it, in the sense that self-assessment and PAYE are much more about the earnings that you take and your track record in terms of the income that you have received.

The question, I guess, is that there is a suggestion that if we are moving to more of a cashless society, there is more tracking, which means that there is potentially more reporting going into it. That is interesting. Whether that means that there will be more official stuff coming through PAYE and self-assessment because there will be less stuff coming through some particular activities is an interesting point to consider.

That concludes questions from the committee. I have one further question, which is about David Bell’s paper. John Mason touched on it, so I thought that I should do likewise.

In the exchange, we discussed the issue of potential spending changes and tax changes that will be necessary—not the specifics, but the fact that they might be necessary. Professor Bell talks about loss aversion, which is something that I have raised before in the committee. That is the issue whereby people, if you give them additional funding for whatever, simply shrug their shoulders and say, “Thanks”, but if you take something away, they are extremely hostile to that, and it causes much more of a political backlash than the gain that you would get from doing something to give them the same financial sum.

What implications does that have for long-term fiscal sustainability and the ability of Governments to take decisions that may require to be taken?

We have not covered that in the productivity paper, but it is clear that, at a time when incomes are rising, changing tax policy is easier because there tend to be fewer total losers—there are net losers, but you can make more changes with less resistance. In a period like the one that we have been going through, in which incomes have risen very little, tax policy becomes very sensitive, because you are imposing genuine losses on people rather than just taking away some of the gains.

Your question gets to the heart of the question that we have been debating in Scotland for 10 years since the Christie commission. We know that we need to move into prevention and shift expenditure into certain areas to reduce demand and improve outcomes in the long run, and improve people’s wellbeing, but how do you do that when budgets are under pressure? That means that you are going to have to take money from one area to another, and that is the bit that is really difficult for any Government to do.

Our fiscal sustainability stuff is, I hope, about trying to say, “Look—you need to do this, because if you don’t, it becomes unsustainable.” You cannot afford to keep spending in these key areas without discussions about the economy, taxation and which areas of service you prioritise.

I wanted to end on that because fiscal sustainability is going to be the key issue in our budget scrutiny this year.

I thank you very much for your evidence today, Professor Roy and Professor Breedon, and I also thank Claire Murdoch. It is very much appreciated, as always.

It has been a reasonably long session so far, so I call a break until 11 o’clock.

10:51 Meeting suspended.  

10:59 On resuming—