Finance and Constitution Committee

Meeting date: Wednesday, September 13, 2017

Official Report 590KB pdf

---

Scottish Fiscal Commission (Forecast Evaluation Report)

Item 3 is evidence on the Scottish Fiscal Commission’s “Forecast Evaluation Report September 2017”. We are joined from the Scottish Fiscal Commission by Lady Susan Rice, who is its chair; Professor Alasdair Smith and David Wilson, who are commissioners; and John Ireland, who is the chief executive. I warmly welcome them to this morning’s meeting and invite Lady Rice to make a short opening statement.

Good morning, convener and committee. I should probably thank you three times over. First, I thank you for inviting us to give evidence today, and secondly, for engaging with us at various times earlier in the year, especially when the Organisation for Economic Co-operation and Development representatives were over and held a conference here in the Parliament. Thirdly, in preparing for the meeting, it suddenly dawned on me that we had not been in front of the committee officially and formally since the end of last year, so I thank you for the reprieve. That does not really make a difference, because you are always on our minds.

I gather that this is the committee’s first meeting in the new parliamentary year, and that this is your first public conversation after the recess. This is our first appearance since the Fiscal Commission became a statutory body. I am, for the first time, joined by my colleagues, whom you have just named.

As you know, there was another first this year on 1 April when the commission assumed responsibility for independently forecasting Scottish gross domestic product, devolved tax receipts and devolved demand-led social security expenditure. The transition from being a non-statutory body scrutinising the Scottish Government’s forecasts to becoming an independent non-ministerial department has been a fair piece of work. We will leave it at that.

However, I think that the committee will want to know that we have agreed a formal protocol with the Scottish Government, which is set out in a framework document. We are in the process of finalising similar arrangements with the Office for Budget Responsibility, Her Majesty’s Revenue and Customs, Revenue Scotland and other bodies.

Since April, in addition to induction for my colleagues here—I joined those sessions because it never hurts to be reminded—we have recruited 15 or so analytical staff from the Scottish civil service, the United Kingdom civil service, academia and the private sector. They have quite varied backgrounds: their experience includes fiscal forecasting, macroeconomic modelling, housing market analysis and public sector finances. They are a good team—we are very pleased with them, and we have been getting down to proper work.

We published our draft corporate plan yesterday, which members might have seen. If you have not, we hope that you will. Last week, we produced our first publications: the “Forecast Evaluation Report September 2017”, which we are here to discuss with the committee today, and a paper setting out how we propose to approach our forecasts at the time of the draft budget later this year.

We have also kicked off a programme of external engagement, because it is important that the people who care about what we do know us. That includes this committee, so we were pleased that a number of its members were able to attend the session that we held for parliamentarians in June. Over the summer, we also had some forecasting experts from around the UK engage with our teams in going over the forecasts, turning them inside out and commenting on the work that we are doing right now and what we will do in the future.

We will be meeting informally with a number of other economists next month to take them through our approaches and ideas. As with the journalists whom we met earlier in the year, we want to keep our stakeholders as well informed as we can—partly because it is the right thing to do, and partly because it reflects our adherence to the Organisation for Economic Co-operation and Development’s principles for independent fiscal institutions—IFIs—which is what we are.

That brings us full circle to this evidence session, in which we will be happy to answer questions about our “Forecast Evaluation Report September 2017”. I am sure that members want to get into the detail, so I will leave you with a couple of what I call keepers, or high-level thoughts. The first is that forecasting is a challenging but also inexact science, so at any point in time a range of valid and reasonable forecasts could be made. There is no single right forecast that we can pull out of the pot. Forecasting typically involves judgment, and judgments change over time. I know that committee members, wearing their other hats, have been grappling with the changing relationship between the UK and the European Union. That issue is also an example of where the commission will have to make broad judgments about the flight path, if you will, in order to produce our forecast, which will come later in the autumn.

The second thought is that forecasts benefit increasingly as data accumulate. You will see from the report how the approaches to our devolved taxes developed between 2015 and 2016 as we got more data over time. Under the new taxes, especially under a newly structured tax such as land and buildings transaction tax, the models will be further enhanced and developed. One example is the additional dwelling supplement—ADS—which was very challenging to forecast in the first instance because there were no data to start with. We therefore used the best approximation of the baseline numbers. We now have the first full year of data on how many properties fell under that category and we will have more over time, so that forecast will improve.

For us, the insights from the evaluation report have been really helpful as we develop our models for the forecasting that we will do later this year. We hope that the committee, too, found the report helpful. We are happy to try to answer your questions. Thank you for listening.

Thank you, Lady Rice. I remember that the whole issue of forecasting became clear in my mind—if forecasting can ever become clear in anyone’s mind—when Robert Chote said that it is a bit like spot the ball, if somebody was always moving the ball in the picture. That is for those of you who are old enough to remember spot the ball.

Paragraph 4 of your executive summary states:

“The residential LBTT forecast is sensitive to changes in house prices and the volume of … purchases”,

and that brings with it overall forecast errors. However, I am not sure that that fully explains the forecast changes between December 2015 and December 2016 that we see in table 1. Therefore, it would be helpful if you could explain to the committee the main reason why the Scottish Government forecast for residential LBTT for 2016-17 was £282 million in December 2015 and £181 million in December 2016.

The main difference between the two Scottish Government forecasts is the basis on which house prices were forecast. In the earlier forecast period—I think that the report says this somewhere—the Government forecasts considered a rather long period of experience with house prices, including before the global financial crisis of 2008, and expected that the economy and the housing market would revert to historical patterns, in due course. However, by 2015, we all understood that the effects of the global financial crisis on the housing market and elsewhere were pretty long lasting, so it was no longer reasonable to suppose that the housing market would go back to what it was in its pre-2008 period. There was therefore a substantial change in the view of what data about the housing market were relevant for the forecast. That is the big change that took place between the two Government forecasts. In the later forecast, no attention was paid to prices before 2008 because they were regarded—rightly, in our view—as ancient history by that point.

Okay. That sets the scene. Ash Denham has some questions on LBTT, as well.

Some of my question has been covered by the previous answer. However, I want a bit of clarity on the log-normal distribution model. Your report suggests that the model fitted the data, but then goes on to say that the

“forecast error for the top two tax brackets”

is

“partially due to the fit of the log-normal distribution”.

Will you explain that for me?

Yes—I will try. There is nothing magic about log-normal distribution. It is just a mathematical way of describing certain kinds of distribution, of which house prices—or, indeed, house transactions—are one. It happens to work pretty well in the Scottish housing market, as in other housing markets, except at the very top, because of the distribution. It is not that something extraordinary is happening up there; it is just that that little piece of mathematical kit does not fit so well up at the top, so it is necessary to make an adjustment. The adjustment that was needed in the second period was slightly larger than the one in the first period.

It is just a statistical way of describing the numbers that proves to be useful. Making the adjustment makes it more accurate. It is not perfectly accurate, but it is a very good working tool that helps the forecasters to make their forecasts.

So that is the tool that you would use to make the forecasts and you have no plans to move to anything else. Is anything better available, or is that best practice?

Our current plans are that the adjusted log-normal will still play a central role in the forecast, but we plan to develop the tools further by, for example, looking in greater depth at behavioural effects. Maybe we will come on to that. However, we envisage that the log-normal approach will still play a central role.

Alexander Burnett has a question about LBTT in the north-east.

One component of the errors is the number of property transactions. I think that the error accounted for £6 million last year, which is a £23 million swing on the errors for the previous year. What breakdown is there for the number of transactions in geographical regions? Were they in line with forecasts throughout Scotland? Specifically—if I understand the figures in the report correctly—you say in paragraph 2.34 that the north-east and Aberdeen housing market was in line with the forecast. If so, I wonder where that evidence comes from.

10:15  

Let me give you the first part of the answer, then I will turn to one of my colleagues to follow up. Last year, in our previous guise, when we were having challenge meetings with the Scottish Government forecasters, we saw variation in the numbers that were coming—that was all work in progress, probably in the summer of 2016—and we speculated that there might be an impact because of the economic constraints in the north-east and suggested that that be looked at specifically. We do not do regional analysis for 32 local authorities, or anything of that sort, but we felt that that was a significant area to explore, which is why it was teased out. Actually, what we found was that there was not a noticeable impact on house-price growth from what happened in the north-east—at least, at that point in time.

Paragraphs 2.34 and 2.35 set out what we did, and it is the case that house prices in Aberdeen and Aberdeenshire have had lower growth than prices the rest of Scotland. Figure 2.8 shows the effects on the 2015-16 forecasts of Aberdeen and Aberdeenshire having been different. Had Aberdeen and Aberdeenshire followed the Scottish average, there would have been much more LBTT revenue. As is set out in paragraph 2.35, there would have been 2.5 per cent and 7 per cent more revenue in 2015-16 had house prices in Aberdeenshire risen at the Scottish average, so there has been an effect.

Thank you. I would like to focus on the transactions element. You are saying that there is no link, and that when you do the modelling you are not collecting data from the 32 councils. If you are not taking in planning data and consents, how are you predicting how many transactions will take place?

That is done using national data that covers the whole of Scotland, rather than data that is broken down by region.

Potential transactions come through the planning process, though.

We have historical data on the number of transactions and we use that to predict the path of transactions in the future. A Scotland-wide transaction prediction equation is used, rather than finely detailed information about planning consents and things like that, because that would be an awful lot of work compared with the standard forecasting approach of using historical transaction data.

So, are you saying that there is no link between the model and what is physically being built on the ground?

No. There is a definitive link, because what is being built on the ground feeds into the transactions data. Remember that it is transactions data for the whole market, not just new build, so the links are there, but they are relatively weak.

I think that Ivan McKee’s question is on the same theme—or, at least, it is on LBTT.

Yes—my question is on tying up the forecasts. I have been looking at the charts bridging the forecast and the tax that was raised. In particular, I am interested in table 2.3, which shows the December 2015 forecasts for 2016-17. To clarify, and to put it on the record, are you saying that by far the lion’s share of the difference between the forecast and the outturn was down to average house prices, and that the forecast was out by £75 million? What you are saying about the distribution fit, if I understand it correctly, is that if there had been a change in the profile of the price bands, that is where it would have shown up, so that that had a minimal impact in the broad scheme of things. Is that correct?

That is correct.

The other point that I wanted to raise was about the report’s references to behavioural effects. Are those short-term effects, as in forestalling, or are there other longer-term behavioural effects that you are trying to pull in? I am not quite clear about how you can isolate those from what you would see in the normal movement of one of the other factors—be it average house prices, median or the log normal.

The behavioural effects include longer-run effects as well as forestalling.

Okay.

We will continue to look at both. In some ways, forestalling is easier to look at, because we can see it fairly easily in the data: at one point, transactions go up and then they immediately go down. Longer-run behavioural effects are harder work, because you have to look at whether, in particular areas of the market, transactions are falling off not as a result of a tax being introduced but because a tax is at particular level. The short answer to your question, however, is yes—both aspects are contained under the heading “Behavioural effects” in figure 2.3, and both will continue to form part of our forecasts.

Again, compared to the change in average house prices, it is a fairly small part of the overall error.

That is right.

Okay. That is fine.

I would add, though, that the £13 million in that particular forecast represents a behavioural adjustment that the Government made to its forecasts. It is therefore not the commission’s estimate of the behavioural effects—it is the adjustment that the Government made when it produced its forecasts.

Thank you.

On the same theme, you highlight a similar error in calculating non-residential LBTT, because of the use of averages and the way in which a high-value item skews the average price of a particular type of property. In general, is it safe to continually use averages in the methodology, when perhaps the median value of a transaction might be more accurate?

I am happy to have a go at that question, too.

I feel as though I am at one of my numerical methods lectures at university.

What is being described is an inescapable problem. Non-residential LBTT revenue is heavily influenced by a small number of transactions, and no statistical trick will get round that hard reality. As a result, looking at medians rather than means is probably not the way to go. The median would be looked at if the shape of the distribution in question was very off centre, but if the outcome depends on just a few big numbers that might vary a bit, the forecaster just has to do their best, look at experience and make a guess.

For forecasting, it is perhaps more important to form a good sense of how much uncertainty there is, instead of trying to be spuriously accurate. It is as important to tell a user of forecasts how much confidence you have in your central forecast as it is to tell them the central forecast. However good a forecast is and whatever fancy tricks have been used, the confidence interval will necessarily be wide in this area, because of the nature of the issue.

That deals with LBTT. We are keeping to the themes, so we will now deal with non-domestic rates.

On the forecasting for non-domestic rates, your report says that, although the forecast growth in total rateable value for 2016-17 was 1 per cent, the outturn was 0.33 per cent. I appreciate that that is not a huge cash sum in relative terms—it is a difference of about £10 million—but it is quite a major gap in percentage terms. Indeed, the outturn is a third of the forecast amount. Will you shed any light on why the outturn was so much lower than the forecast?

That gets to the question of the buoyancy of the changes in the increases in non-domestic rates from year to year, which we have to make an assumption about. The 1 per cent figure that we were evaluating against was the assumption that the Government rather than the Scottish Fiscal Commission made, and it seemed to us to be a reasonable assessment last year, based on what we knew about the trends in buoyancy.

However, we and, I think, the Government statisticians think that providing a statistical estimate of buoyancy based on some form of economic determinants is extremely difficult. In the past, the view was that the rate of change of non-domestic rate income somehow reflects economic circumstances or economic determinants. There is increasingly a view that, on the data, in one sense the economy must have some impact, but it seems to be so far removed from the year-on-year changes that we end up in a situation in which the estimate of buoyancy is a residual rather than something that is based on economic determinants.

The issue is perhaps similar to the previous LBTT issue. By definition, it is difficult to develop statistical measures to forecast such a thing when it falls out of a whole series of factors that are built into the system around appeals and changes in the overall system.

The main message is that the estimate was a reasonable one to make at the time about something that is extremely difficult to forecast. As you rightly said, the position has made some difference to the outturn, but we are not reading too much into it, as it reflects the need for a substantial change to how we make such forecasts in the future.

From what you have said, it sounds as if it is very difficult to forecast in this area. Are you really just guessing where the figure will end up?

The term “just guessing” is perhaps pejorative. A judgment needs to be made that is based on the best possible evidence that we can bring to bear. We are keen to get it across to members that the judgments are difficult to make, but we recognise that part of the reason for setting up the commission was that the overall budget process requires somebody to make a decision on the basis of the best estimate that we have, however difficult that is. That is the role that we need to play.

I have one follow-up question for clarity. Your report mentions that one possible reason why the level of growth was lower than the estimate is that there were

“several significant removals from the Valuation Roll”.

What does that relate to? Does it relate to properties being demolished? Does that explain it?

Again, I will draw a comparison with large transactions in LBTT. I will not go into the details of individual cases, but there are large properties and facilities that are no longer in use, and buildings are being replaced. I am not sure whether the St James centre can be seen from the Parliament, but it is an example of a large property that is changing its status and payments. A number of such developments are leaving and entering the register.

One measure that we could look at is a more detailed listing of major properties that are moving into and out of the system. Tracking that might help to indicate year-on-year changes. We have been looking at that issue. However, it is inevitable that the timings of precise changes, in the light of reviews, appeals and construction times, mean that the task is fairly demanding to undertake.

10:30  

I have a tiny point to add. To put it in simple terms, there is what is happening on the ground to buildings, as David Wilson just explained, but there has also been the court decision that moved properties into different categories, so there is an administrative impact, too, which makes the situation rather complicated.

As no one else has questions on non-domestic rates forecasting, we will move on to more general forecasting and the link between that and the fiscal framework, which Maree Todd wants to ask about.

We have talked about the fact that what is significant is not the differential between the forecast and the outturn but the difference between the outturn and the block grant adjustment. Given the complexity that we face as a devolved nation, there is an added layer of complication. Will you explain how your forecasting fits in with the fiscal framework?

I am sorry—I am trying to pinpoint what you are looking for. Are you asking how our forecasting affects the block grant negotiation or outcome?

My question relates to the fiscal framework. There is the amount of money that is forecast and there is the amount of money that we collect. On top of that, there is the amount of money that the UK Government is forecast to get and there is the amount of money that it collects. An adjustment is made between the two Governments. I am sorry to be simplistic, but will you explain that process a bit more?

I think that you have got it. The block grant adjustment depends on the difference between the revenue that is raised in Scotland and the revenue that is raised in the rest of the UK. We forecast the Scottish component of that and the Office for Budget Responsibility, by default, forecasts the rest-of-the-UK part of that. Those two things give the block grant adjustment, which is the Scottish Government’s additional money. You have got it exactly right.

In that case, we will move on to the commission’s readiness.

Good morning, panel. Lady Rice, in your opening statement you talked about the commission’s readiness to produce its first official forecast later in the year. As I see it, there are three strands to that. There is the resource that you need to do that—you mentioned that you have taken on 15 staff. There are also the models that you use and the methodologies that underpin those models. Will you give us an overview of how robust you feel that all that is as you move towards the production of your first official forecast?

I will give you an overview before handing over to John Ireland, who, as chief executive, has played a key role in pulling so much of the work together over recent months.

As I said, we have a really good team. A couple more people are due to join us, so the team is still forming, but the people we have are working together well and bouncing off one another well. We are pleased with the people we have. We think that we have the resources that we need for our remit and our responsibilities this year. If we find in a few years’ time that the remit grows, we will need to have more resource, as we do not have anybody spare at the moment. We think that we have the right people.

Bringing in the experts whom I mentioned in my opening comments helped us to develop our views as commissioners—we need to do that because, ultimately, we are responsible for the forecasts that will be produced in December. The experts’ job was to provide another independent view, at our behest. After going into the forecasts, they reported back to us, as commissioners, and to members of our senior team. They offered lots of thoughts and suggestions and made many interesting points, but overall, in all cases, they felt that we were on track. We take some comfort from that, as well as from our own judgment that we are on track.

Our report “Current Approach to Forecasting September 2017” summarises some of the points on modelling. We are building on models for the taxes that already existed in a devolved form. We judged them to be reasonable, with some challenges, and they are a good place to start. However, that does not mean that they will stay the same for ever, because we will enhance and change them over time.

For the new devolved taxes, we have to build the models ourselves. A lot of work is being done on the air departure tax, and we are building new models for the categories of social security that we are responsible for in the first instance.

John Ireland might want to say something about our overall resources.

Susan Rice has covered the resources that we have access to and the evolution of our models from those that we inherited from the Government and those that we have built in-house. However, I can add something on quality assurance.

We have three levels of quality assurance in the modelling work that we have been doing over the past year. First, there is an internal challenge process—the commissioners have been working closely with staff to go through the model. Secondly, we have worked closely with the Government analysts—we have taken back to them the models that we took from them, explained our changes and got their feedback. Finally, we have been working with external people, such as David Eiser at the Fraser of Allander institute, on things such as our income tax model. As Susan Rice explained in some detail, we employed three academics from the south to come up and, over three or four days, give us their insights into the modelling that we have been doing.

Having gone through that process of quality assurance, we are reasonably confident that we are in a good starting place for our forecasting.

Does the quality assurance check cover the process of the model? Have you done any testing that is based on numbers?

Yes. The experts and our quality assurance process have considered both the ability of the models to forecast and their methodological approach. One of the earlier questions was about whether it would be a state-of-the-art approach and whether there were alternatives. We have assessed that.

We have certainly been considering how well the models perform. The forecast evaluation report that we have just been discussing is an important part of that. We have also been carrying out similar exercises internally.

A new area of forecasting for us is to forecast the macro economy. We have constructed models and used the National Institute of Economic and Social Research Scottish Government global economic model—that framework was built for the Scottish Government. We have been producing dummy macro forecasts over the summer, and we are now in our third round of internal forecasting using that framework. We have been producing numbers and running our eyes over them.

So there is an iterative process in the sense that you have taken soundings from experts on the methodology, you have run tests that in some cases used live data and you continue to check that.

Yes—very much so.

I have a follow-up question on one aspect of the methodology that takes us a little bit beyond the evaluation report. It might be that you are not prepared on that point and would prefer to come back to us in writing, which is absolutely fine.

Lady Rice, you have mentioned a couple of times this morning—in your opening remarks and in your comments to Mr Kelly—that the Fiscal Commission has already taken on new responsibilities for making official forecasts for spending on devolved social security. What methodological challenges have you faced in developing models for that and how are you overcoming them?

If no one else wants to pick up on social security, I will touch on that and address air departure tax.

In the first instance, social security will be responsible for about £2.5 billion to £2.8 billion of a £40 billion spend. That represents a few programmes in social security, and part of what has to be done to begin with is understanding how those programmes will work. We anticipate that they may change as they are devolved and administered from Scotland. However, I will put that aside for a moment.

Air departure tax—or air passenger duty, as it was called—is an interesting space. We need to start at the beginning by understanding what numbers are included. For example, what is the base number of passengers departing, and what are the categories? Some of the information on passenger numbers comes from surveys. You may have been stopped in an airport and asked where you were going. If you said that you were going to London, for example, you would go into the count, but if you said that you were en route to somewhere else, you would not.

It is a case of understanding the baseline numbers, just as we had to have proxies for understanding those in relation to the additional dwelling supplement in its first year. I am not sure whether that answers the question, but that is where we have to start.

I guess that you are telling us that, at this stage, you do not know what the baseline numbers are in social security. I can see that John Ireland is beavering away trying to find an answer, but it would be better if you reflected on the question and let us know what early preparations you are making to begin dealing with the issues.

I can point to where that information is now, if that is helpful.

Okay.

Section 8 of the “Current Approach to Forecasting September 2017” method paper that we published, to which Susan Rice referred, outlines where we are on social security modelling. The Government has been working on that for some time and we have been working pretty closely with it. It has access to Department for Work and Pensions data, which gives us a fair amount of historical data.

There are particularly difficult methodological issues when it comes to take-up, for example. Our judgment and work will be really focused on that. We can get reasonable data—not perfect, but reasonable—on the health characteristics of the population. If the aim of the Scottish approach to social security is to operate a kinder system—to reflect some of the language in the programme for government—how we model that in terms of hard take-up data is quite a challenge. That is where our judgment will be most needed.

Okay. You have signposted us to where the material is available.

I hope that it has become clear that we take a different approach to forecasting the different areas of our responsibilities depending on what is needed. For example, the developing work on forecasting onshore GDP is a combination of statistical models and overall judgment about the development of the economy. That requires expertise in modelling and data but it also requires significant engagement with external commentators, business representatives and others who are also thinking actively about the issues. That will all feed into the work that we are doing. Therefore, it is not all about the modelling; it is also about external engagement through a variety of different processes.

Much of the work on social security is about engagement with the teams that are implementing the new approaches that the Scottish Government will take on and developing specific expertise in and understanding of the possible expenditure. It is partly about the development of statistical models, but a big part of our work at the moment is active engagement with the teams that are developing and implementing those systems so that we can work with them to ensure that there is an understanding of expenditure.

I clarify that it is not all about having a single model that will give us all the answers. There is a significant amount of engagement and collaborative working with various organisations to produce the estimates.

That is helpful.

I have a question that is not about the doing of the work—which is extremely complicated for most people, including us, to understand—but about the communicating of it.

We are all familiar with the political impact of the “Government Expenditure and Revenue Scotland” publication. Whatever we think about why it is what it is, it tends to further the polarisation of the stories that we tell one another about the Scottish economy rather than shedding light on things.

10:45  

I am hopeful that the summarising of your forecasting will not become a similarly polarising political event in the year and that publication of the block grant adjustment does not become an opportunity for all us politicians to say, “Well, that proves exactly what we have been saying all along.” That is not a very helpful dynamic, although it is something we are all guilty of participating in. There is an opportunity for the Scottish Fiscal Commission, as an independent body, to become a better source of publicly accessible, authoritative information.

To be fair, most people will not have hours to spend reading and wrapping their heads around the detailed reports that you publish. I am wondering how you intend to communicate in more accessible ways the key findings of your reports. Will you engage when, for example, the block grant adjustment is confirmed? Do you see informing the public about how such things have come about and what they really mean as part of your role? You have an active social media account, for example, but it mostly consists of links to your main publications rather than anything that is more digestible.

The fact that we have a social media account shows that we are moving on in that respect. You raise a really important issue and it is something that we are conscious of, care about and speak about among ourselves a lot. I will give you a couple of answers; I am sure that my colleagues can add to those.

Last year, we inserted an executive summary at the front of the report that we issued alongside the draft budget—those were not our forecasts, but we commented on them. When we read the content of the report, we thought that many people would not find it very accessible, so the purpose of the executive summary was to try to articulate in a way that might be more accessible what we were talking about in relation to the small number of devolved taxes. We think that that is important. We also intend to use our website to clarify things and speak in easy language, if you will. However, there are also those people who are highly technical and want to know about the technical side of our work, so we need to do both.

We will be able to comment on the work that we are responsible for because we understand it. We will not be commenting on other aspects of the fiscal infrastructure that are outside our purview because we will not be expert in that space. We will not be commentators in that sense. That would not be helpful.

John Ireland thinks a lot about communications.

In addition to the structure of the reports, with the non-technical executive summary, there is enormous value in charts. As today’s questions and the LBTT decomposition charts have illustrated, some carefully chosen charts can do an awful lot. We are spending a lot of time thinking about the charts that we produce and how we can use them in social media as well.

That is helpful. In December last year, you produced a series of infographics—charts would be a stage beyond that simplistic presentation of a single statistic.

There would be great value in ensuring that the commission is seen as a source not just of detailed information for the Government, academics and people who want to explore that detail, but of authoritative, clear information for people who might only read the headlines.

We endorse that view.

I welcome what you say about the role of the commission. To build on what Susan Rice has said, we are very conscious that there are some very major issues that will underpin the forecasts that we will make and in which there will be significant political interest. For example, on page 12 of the evaluation report there is a graph about the current position on productivity trends. Although it is often said that, when it comes to growth of wage incomes—the money in people’s pockets—and Government fiscal revenues, it is not all about productivity, most of it is. On the fundamentals of what is happening in the economy around productivity, behavioural responses to any tax changes and, as Susan Rice put it earlier, the flight path of how things will develop, we want to take on the role of communicating some of the major issues that underpin that.

I will add a slight caveat in relation to the wider issues around the block grant adjustment. We see our role as principally being about assessing the income side of the balance sheet—the income that the Scottish Government will receive from the devolved taxes—rather than dealing with the wider set of questions about the fiscal framework, which is a matter for the respective Governments to sort out for themselves. We play a key part in that overall framework, but it is for others to develop the more detailed understanding and precise minutiae of how the block grant adjustment will be taken forward.

That is helpful, thank you.

Yes, it was helpful and took us into some areas that I intended to cover.

All that remains is for me to thank the witnesses for coming along this morning. You have thrown some light on the commission’s forecasting and have enabled us to begin to understand the journey that you are on.

10:51 Meeting suspended.  

10:55 On resuming—