Our second item of business is to take evidence from the chair and members of the Scottish Fiscal Commission in relation to the commission’s report on forecasts for the devolved taxes and for non-domestic rates. I welcome to the meeting Lady Susan Rice CBE, Professor Campbell Leith and Professor Andrew Hughes Hallett.
Committee members have the report, so we will go straight to questions. As is the custom on the committee, I will ask questions first. Members can come in subsequently if there are any questions left to ask—I try not to steal all the good ones at the start.
The commission describes its approach to consideration of the Government’s forecast as one of
“enquiry and challenge, followed by response, followed by further enquiry and suggested improvements.”
Do you intend to publish details of the inquiry and challenge, as recommended by the committee?
Let me respond in the first instance and then I will turn to my colleagues. Are you happy for the three of us to have a conversation with you?
Yes. I should have made that clear at the start. One or more people can comment; the questions are not directed at you alone. You can answer, and if your colleagues also wish to answer, that is great. You may not wish to answer certain questions at all if they can be more appropriately referred to the professors on either side of you.
That is very helpful. I will at least start off in response to the question.
The words that you quoted are mine. The three of us drafted the report, but we took on different bits and sections and then made it into—we hope—one voice. I was simply trying to reflect the nature of how we do things. Some fiscal bodies in other countries and places look at forecasts after the fact, but our job is to look at them before the fact. Those words were meant to point out the nature of what we are doing.
We started our work once the forecasters were ready, in the latter part of the summer, by having a presentation from them on their approach to forecasting—their models and so forth. Inquiry and challenge was a matter of asking a lot of questions, such as, “What do you mean by this?”, or, “Have you considered this or that?” Out of all the to-ing and fro-ing, we ultimately drew our conclusions. For instance, one conclusion that is very clear in the report is that these are new taxes without decades of historical data. That is a challenge, and data is an issue. Out of the inquiry and the challenge came some of our conclusions. In summary, that is what is encompassed in the report.
If you are asking us to publish something more specific, the answer is that if you tell us what you would like, we can do it. If you would like a summary of meetings or conversations that we had with the economic forecasters, for instance, we can prepare something of that sort.
However, we felt that we had reported overall on the nature of how we worked, which was to challenge beforehand. The results of all those conversations were the results and conclusions that we put into our report.
Did your forecast change as a result of that inquiry and challenge?
Well, we are not forecasting—
I should have said the assessment of the forecasts. Quite clearly, there are two different forecasts: the Scottish Government’s forecast on new taxes, for example, and the Office for Budget Responsibility’s forecast. As you know, there is a £51 million differential between those forecasts. My question is whether your assessment of those forecasts changed.
I will ask my two colleagues to respond, because to begin with they each took a tax, although we shared all of this—at the end of the day, it was very much a committee effort.
Our understanding of what was happening changed absolutely as we had the conversations. Some of the economists came back to us to answer more questions and give us more information as we delved into the subject.
In terms of our assessment, we started by finding things out. We did not start by saying, “This is good or bad”; we started by saying, “Help us understand what you’ve done.” You cannot start anywhere else.
Absolutely.
I would not answer your question in exactly the terms in which you posed it. Let me turn to one of my colleagues for an answer.
The nature of the inquiry took several forms. In addition to written submissions from the Government to us, and the queries that we wrote in response, we had meetings. Those essentially took the form of an academic seminar, in which we put points of clarification to people during the presentation of their forecasting methods. That was followed by a long series of questions.
We also received spreadsheets containing the models from the Scottish Government. In a sense, that is all part of the inquiry process, even though we were not specifically questioning the Scottish Government economists at that point but were just looking through the models.
A wide range of activities took place in the inquiry. The initial part of that involved not so much a review of the forecasts, because they were not finalised at that point, as an exploration of the methods. A large chunk of the initial part of our inquiry involved looking at how the Scottish Government proposed to forecast, in a general sense. What were the underlying methods? What techniques were the forecasters using, and what models had they built to enable them to do that? The bulk of the inquiry was focused on that.
The forecasts were finalised relatively late on in the process, and we looked at those in light of the discussion about methods that had taken place.
Lady Rice said that how things developed changed quite considerably during that process. Can you give us a bit more detail about how that came together?
There was not a big window of opportunity for any of us to come together, understand the issues and produce our report. It was a very intense period—it did not happen over a period of many months. We were formed only in July, when we had our letters of appointment.
I am trying to remember the specific words that you used—I am trying to find that in the depths of my short-term memory. You said that the way in which you explored the issues in the inquiry changed or evolved a wee bit over the period.
Yes. Sorry. Let me make that clear if I have not done so.
We started out by trying to understand how the forecasters were coming at it—what models they were using and how they were approaching the forecasts. We could not really make a judgment about how reasonable their forecasts were unless we understood their approach. That was the early part of our engagement. Then, as Campbell Leith has just said, when the forecasters had more or less concluded their forecasts, we looked at those and needed to make a judgment about whether they were reasonable.
You will have seen our conclusions in the report. In one area, we pointed out that the forecasts were within the bounds of reasonableness but quite optimistic, and that was responded to. That was the outcome of the second phase, when we applied to the forecasts what we knew about the forecasters’ models and their approach. Is that clear? I am sorry if I am not making it clear.
I think that it is me who is not so clear, rather than you. However, I will move on a wee bit.
Throughout the report, you talk about data. Obviously, there is frustration—which the committee has shared for a long period—about the sources and quality of data that is available specifically to allow us to look at and drill down into Scottish figures. Was all the information that you considered necessary for your report made available in good time by the Scottish Government?
The point about data is a very important one, as I indicated in my answer to your first question. We are talking about taxes with a new shape, so there is no long history of data that could feed into the models that the forecasters were using. As my two colleagues will tell you, economists want a long series of data. That makes them more comfortable with the forecasts that they come out with.
Given that fact, we and the economists had a lot of conversation about what data they had, what their sources were, whether there was anything better, whether people co-operated with them when they went to acquire data, whether it was the best data that they had and so forth.
I will make two other comments before I turn to Andrew Hughes Hallett to add a little more. First, I have prompted a series of meetings with some of the agencies that have started to provide data to the Government economists. Those conversations are continuing. I have had a couple of meetings with the Scottish Environment Protection Agency, and I have a further meeting coming up. Revenue Scotland and Registers of Scotland are also involved. It is for us to make a judgment about how willing, happy and responsive they have been. There has certainly been a willingness on all sides when it comes to the sources of data.
In the past—for reasons that we can understand—the OBR has, in some instances, gathered data, made a proportional cut and applied that to Scotland. That is not the same as having Scotland-only data, which is where we all want to be. That could not be the case in the first round.
My second point is that we had a great deal of debate among ourselves, sometimes well into the night, about what forecasters and economic modellers should do when they do not have a perfect data set—that is often the case, because life is rarely perfect—and what the best approaches to that, and ways of dealing with it, are. I ask Andrew Hughes Hallett whether he would like to say something about that.
To go back to the initial question, we want long data series if we can get them because we are trying to look for some regularity in the data. In a short period, it is not possible to tell whether the variation is just noise or whether there is something irregular at play. Therefore, it is important to go back as far as possible.
It sometimes takes a while to discover whether the data goes back in time—we might not know whether it does, or it might be hidden somewhere. For example, we have discovered that the housing data from Registers of Scotland could go back to 1617—it goes a long way back, but it is very patchy.
It is very patchy before 2003.
Apparently, it is pretty good as far back as 2003. Beyond that, we have to ask how much regularity we can get out of the data.
In other cases, such as housing, it is obvious that some financial data is desirable. However, such data might not exist or, as Susan Rice said, it might exist only at the United Kingdom level. Up until now—until we can collect some of our own data or find some other way of getting it—the best that we could do was take a Scottish cut of the data, and such data reflects conditions in the rest of the UK rather more than it reflects Scottish conditions. Therefore, with housing, the numbers might get blown apart because of the London effect, and that leaves us asking whether the cut that we have taken is good enough or whether we can make an assumption that we believe takes out the London effect, although it will not be possible to be sure about that until we can hold a post mortem.
That is a very long way of saying that it is inevitable that there will be compromise. We know how, ideally, we would like to do things. When we discover that that is not possible, the question is how far we are prepared to compromise. I think that the methods that the Scottish Government people have used up until now are reasonable, but that is not to say that they could not be improved. Our job for the coming year is to see what we can do to improve them. Assessing how we can sensibly go forward in a reliable way will be a bit of a—excuse the phrase—touchy-feely process.
I mentioned that the difference between the forecasts for receipts from the devolved taxes is of the order of £51 million—the Scottish Government forecasts a figure of £558 million, whereas the OBR forecasts one of £609 million. Which of those figures do you think is the most accurate?
09:45
I will try to answer that. Essentially, there is a continuum of ways of forecasting, and the Scottish Government and the OBR are at opposite extremes of that continuum. The OBR has a large macroeconometric model with hundreds of equations describing all aspects of the economy, largely because the OBR has to forecast all macro variables plus the whole range of fiscal variables that it is interested in. The bulk of its work is focused on that. In forecasting the devolved taxes, given the limited data that there is, it tends to apportion some kind of share of UK projections to Scotland, which may be just a historical average but may have some slight drift in it.
On the other hand, the Scottish Government has looked at each tax on a case-by-case basis and has used the data that it has for that tax to build small, simple statistical models. Sometimes, it has simply extrapolated historical data to project those variables forward.
Those are completely different approaches and it is not obvious which one will produce the best forecast. Horse races between those forecasting techniques are conducted in the literature and it is not automatically the case that a big, 100-equation model will dominate simple statistical analysis. No one approach is better than another; essentially, they are fundamentally different ways of doing forecasts.
Turning to specific taxes, the OBR’s forecast for the land and buildings transaction tax is slightly higher than the Scottish Government’s forecast. However, the OBR is using UK-wide projections, and the UK as a whole has a more buoyant housing market than Scotland does. The OBR is shading those projections down a little bit for Scotland’s share but not to the extent that it appears the Scottish Government is doing implicitly, by building up from Scottish data. Those are the fundamental differences in how they are doing things.
I know that there are fundamental differences. To be honest, your answer was almost “How long is a piece of string?”
We are quite keen to know which forecast you think would be more accurate. There is a difference of some 15 per cent between the residential transactions revenue forecast from the OBR and the forecast from the Scottish Government, and there is a 9 per cent difference between the landfill tax forecasts. Those are quite substantial differences in terms of proportion. These are fairly small taxes but as we go forward with other devolved taxes, such differences will be very significant. That is why we are quite keen to get the most accurate forecasting and to get your assessment of which forecast is able to deliver the most accurate prediction of where we will be.
In order to be accurate, what are called standard errors need to be attached to the forecasts, which neither the Scottish Government nor the OBR does at present. It is very difficult to do that with large-scale macroeconometric models, and the Scottish Government has insufficient data to build statistical models that can assess the statistical accuracy of its forecasts. The Bank of England can attach errors to its forecasts when it has sufficient data in a small-scale model. That is how it produces its fan charts, which forecast the upper and lower bounds of inflation.
My conjecture would be that if you drew similar fan charts and did proper statistical analysis, the fan charts would get wide very quickly and would encompass both sets of forecasts. They are not statistically different. They look largely different but, statistically, they are not different.
I will add a very small point. In one instance where there is not a lot of Scottish data in terms of bulk, the Scottish economists took a three-year average, which is a very sensible approach. They were quite conscious of the limitations of the restricted data that they had.
Yes, I saw that in the report.
I want colleagues round the table to have an opportunity to ask questions, so I will touch on just one other thing—non-domestic rates. The commission’s report talks about buoyancy and suggests that the increase is on the optimistic side. Indeed, Mr Swinney has reduced his forecast by £83.5 million. The report also says:
“NDRI revenues are five to six times larger than the LBTT and SLfT taxes combined. It would therefore pay to make them as reliable as possible, as quickly as possible.”
I am not sure why you suggest that because, as Mr Swinney said on 9 October when he presented the draft budget, some £13.1 billion has been collected in non-domestic rates income since 2008, yet the cumulative variance has been only £40 million, which is 0.3 per cent. Professor Leith talked about margins of error and we have talked about a difference of 15 per cent between the OBR and Scottish Government LBTT predictions, yet on NDRI we are talking about only a 0.3 per cent variance. Why are you as a group so concerned about that issue?
In essence, it is because the taxes are on different activities. Land and buildings transaction tax and landfill tax are taxes on transactions, whereas non-domestic rates are a tax on the stock of rateable floor space, so what changes in the forecasts is the rateable floor space, which is to do with new business premises coming online and so on. The forecasters have to forecast the change in domestic rates rather than the level of domestic rates. Perhaps our report is slightly misleading in suggesting that there is volatility in the aggregate level of that. In looking at the accuracy of forecasts, we should have compared the change in non-domestic rates tax revenues with the tax revenues from the other taxes.
The variance might be 0.3 per cent, but it is quite a lot of pounds. The tax is much larger, which is why we are concerned about it. I am slightly behind the times on this, but I think that you said that the figure involved is about £40 billion, convener.
No, it is £40 million out of a £13.1 billion sum over six years. In that time, we have had a recession and all the rest of it, yet the accuracy of the figures is remarkable. It is remarkable to predict such figures within 0.3 per cent.
Yes, but that is over a number of years. I do not have the figures for individual years, but they might cancel each other out. There might be quite a lot of variability year by year.
Of course, we are talking about forecasting for one particular budget. We can get lucky, or in this case perhaps unlucky—I am not sure which it is. That is why we are concerned about the issue. The number of pounds could be quite large.
Have the predictions been way out in any year? Has Mr Swinney predicted significantly higher income than was delivered? I am not aware that that has been the case.
I do not know the numbers.
We do not have those numbers, but the issue will bear looking at. As we work on our work plan for the coming 12 months, we will definitely pick up on that point.
We certainly need to go back into it. As I said, we have to do a post mortem on all three taxes to see how well they performed. The report took quite a long time to produce but we had to do it in a month. If you wanted to come and join us at midnight, you would be welcome.
We all appreciate the work that you are doing—I am not trying to criticise.
You have put your finger on the right issue. We need to look back and consider how well the taxes performed. The phrase that you picked out is mine. I am concerned that the impact on the budget could be a lot larger.
Thank you. I will let my colleagues come in.
I thank the commission very much for its work. You have been able to “endorse as reasonable” the Scottish Government’s forecasts. On that basis, I will return briefly to the difference between the Scottish Government and OBR forecasts. The convener explored that fairly thoroughly but, for absolute clarity, I will go back to the points that Professor Leith made about the difference in forecast methodology. Is the difference explained simply by the fact that the OBR comes to a UK assessment and then apportions a proportionate share of that to Scotland?
Largely, yes. That is the fundamental difference between the two approaches to forecasting. There are slight differences in using a three-year moving average as the base year for forecasting non-residential transactions—the OBR does not do that—and there are other subtle differences. The landfill tax forecast is based on a Department for Environment, Food and Rural Affairs model for the OBR, but it is built up from SEPA data for the Scottish Government. There are other small differences, but the fundamental difference is that the OBR uses a big, macroeconometric model and scales a portion of that for Scotland, whereas the Scottish Government economists start with a small set of data relating just to the devolved taxes and roll that forward in forecasting the taxes.
Does the OBR apportion Scotland’s share on the basis of Scotland’s population equating to just under 9 per cent of the UK population?
No. It has historical data for the tax revenues raised in Scotland from a particular tax relative to what was raised in the rest of the UK and it uses those historical averages, although in recent forecasts it has tweaked the figures slightly. For example, the UK housing market is now more buoyant, so the OBR has reduced the share of stamp duty land tax apportioned to Scotland on the basis that the Scottish housing market has not been as buoyant as the housing market in the rest of the UK has been because of the London effect.
The OBR still appears to think that the Scottish market will be more buoyant than the Scottish Government thinks it will be.
Yes—the OBR has not scaled the figure back.
The convener touched on the frustration that is felt at not having access to data, which is a major theme of your report. You have suggested that the situation should improve over time, but how long will that time be?
That is the question. How long is a piece of string? Thirty years from now, there will be a series of data that covers 30 years, but none of us might be sitting around this table then. We need to go in and do a drains-up review to see whether anything more could have been done for this round to get more data, to get it differently or whatever. How does one accommodate, in a forecasting sense, the fact that we do not have the ideal set of data? Are there ways to do that? We cannot predict the future—we just take our best shot with what we have in order to come out with a picture of what the future might look like.
The data will improve. I have had conversations with SEPA and I will go out to test this on the ground, as it were, in a couple of weeks. SEPA will be much closer to the landfill sites than the OBR has been able to be and it will see the relationship between the tax and what goes into those sites. We believe that it will have a truly on-the-ground ability to reflect what is actually the case. That is not to fault either side for the differences; that is just the way that it has been.
The data will improve over time, but the improvement will be gradual and we will not get the data for those 30 years until the 30 years are up.
Maybe it is unfair to ask you the next question—feel free to dodge it.
I did not know that we could do that.
Feel free to answer the question if you want. You have not heard it yet, but it is not that difficult.
It has been suggested that there should be a dedicated statistical agency for Scotland to look at data not just in the area that you are looking at but more widely. Would that be useful for your commission, for other public bodies, for the public at large and for parliamentarians?
Forecasters need data. It does not matter where it comes from. If a new statistical agency generated that data, that would be a good thing.
There is already Registers of Scotland, whose job is to hold, create and analyse data. I am avoiding giving a yes or no answer. People would have to ask whether there would be an overlap with what is already done, whether what is already done is good enough or whether another agency would be helpful.
10:00
Over time, the new taxes will generate revenues, and new data will be generated. However, before all this happened, transactions were taking place in the Scottish housing market, landfill was being sent to landfill sites and taxes were being paid on all that. I presume that those transactions have been recorded somewhere; they might not be easily accessible but, if that data exists in some raw form in existing bodies, it would be good to encourage those bodies to dig it out and put it into some usable form.
My experience is that often more data is available—somewhere—than we might think. Of course, it is not until people get into one of these forecasting exercises—I should point out that this is the first time that it has been done—that we know the questions that we want to ask and realise that we want this rather than that data.
There will be improvements in that sense, and there will be even more improvements when the process has been gone through two or three times—I should say that that will be three years, not 30 years, down the line—and what I might call the ad hocery in some of the variables has been refined. All those things will happen more quickly than other things; there will not be the perfect answer.
As for establishing a statistical agency, whether people want to pay the money for a whole other agency is one question, but it would be useful in putting an explicit focus on Scottish data instead of data that has been derived from elsewhere, perhaps from the UK, and just carved up.
You say in your report that you
“would welcome feedback from any reader”.
I hope that I have got the terminology right, but I believe that Professor Leith previously talked about peer review. Have you had any such feedback?
We have carried out an exercise to look at reviews of other fiscal commissions or similar bodies in other countries. They are all slightly different and there is not one that is absolutely like ours, but the OBR, which is closest to our world, has been peer reviewed by a Canadian with experience. We have all had a look at that review; it was carried out five years in, which gave the OBR the chance to establish itself and work out its methods and provided an opportunity to find out whether it had got things right or wrong, whether it had changed or adapted and so forth. We agree that such a peer review is absolutely appropriate but, as we have been operating for only two or three months, I guess that it will not happen in the next two or three months.
When we published our report, we did not send it out widely; instead, we put it on our website and sent out a little press notice to ensure that anyone who was interested could find it. However, I sent it to the OBR for review. Over time, as we continue our conversations with the other bodies and agencies with which we have started to have a relationship, we will request feedback and input. It would be good to get that sooner, but a formal peer review—with capital letters—will be some period down the line. I do not know exactly when, though.
I was not necessarily suggesting a formal peer review. In any case, you essentially answered my question when you said that you had sent the report to the OBR. People are looking at this and you are getting feedback.
Indeed. We also received feedback in our recent conversations with Revenue Scotland and Registers of Scotland, and we welcome that. After all, as I have said, this is a first for all of us, and the more people come into the conversation and tell us what they think, what works and what is or is not clear, the better we will do this the next time.
There have been a lot of questions about the amount and availability of data, and I suspect that there will be more. Continuing with that theme, I note that in the recommendations in your report on the non-residential model for LBTT, you say that
“creating a new Scotland-specific data set may be the only reliable strategy.”
However, you then say that
“That is neither a short-term nor cheap undertaking”,
which is the bit that worries me. There must come a time when having more data does not result in a better forecast, in which case spending money on getting that data would be a waste. How do you decide how much data you need? It has been pointed out that some data might already exist, although it is not available. How do we balance such things?
There are many ways of answering that question but, before I turn to my expert colleagues, I should say that part of the answer concerns the quality of the data. The amount of data is important—after all, models require a certain volume of data, which is very helpful, as Andrew Hughes Hallett explained—but the issue is also about how good and specific the data is. We are beginning to see a focus on Scotland-specific data. That is important, because it means that, even if the data that we get is limited in bulk and volume, it will perhaps be better.
Perhaps Andrew Hughes Hallett or Campbell Leith wants to respond to your question about how much data is enough and where the cost benefit comes in.
I suppose that I could say a few things. John Mason is right to pick on that element, because it was one of the weakest parts of the current forecast. If we did this properly, we would go back and say, “What kind of model do we want? How would we, in principle, forecast this?” We would see what data was required and then no doubt discover that it did not exist and that we would have to dig for it. I am not sure that we would do the digging, but we can tell people what to expect.
It is a bit difficult to give you an answer ex ante, because neither we nor the Scottish Government people have carried out that exercise. At that point, however, we might well find that we desperately need Scotland-specific data, and we would need to set up a project on that, which we might commission. That is how I would go about this; I would start at the model end instead of trying to find all the data that we can imagine on what is affecting the non-residential sector.
As the convener pointed out, the forecasts for non-domestic rates income have been quite accurate over six years. Professor Hughes Hallett suggested that there might be many fluctuations in that, but is that important? I realise that, if the fluctuations were too great, they would affect the borrowing limit, but should we not be more interested in ensuring that we have a balanced budget over six years or whatever? Is one year so important?
We are interested in having a balanced budget over time, but I am not sure that the Scottish Government is in such a position. It has to balance the budget as best it can every year because, in effect, it cannot borrow. The amount of borrowing that is allowed under the current arrangements for spending is tiny; indeed, at 0.5 per cent of gross domestic product, it is scarcely visible to the naked eye. The Government is constrained, which is why I made the comment that you have highlighted.
If the Government were asked to follow convention and went to a model in which it balanced the budget over the cycle—let us forget the last financial crisis, which was a bit of an extreme event—things would be a whole lot easier, and we would be interested in performance in exactly the way that you have described. I need to go back and check how much fluctuation there was year by year.
It was just a general question.
That was a general answer to it.
One or two things about landfill tax jumped out at me from your report. First, I was concerned about your comment that SEPA is reporting data with a two-year time lag. Can you comment on that?
That came out of our explorations of data sources. As you know, SEPA has been changing the way in which it gathers landfill data. Something that I think we will all agree is very good—I certainly think so—is that, unlike the OBR, which looks only at landfill as reported and says that the tax is such and such, SEPA looks at illegal landfill and cases of people breaking the rules. As I understand it, there will be a bigger penalty—or tax payment—for such activity. SEPA is therefore in the process of changing its approach. It has been developing an approach in which it will go out, visit and view sites, and that will lead to some difference.
That was a comment about what has happened. The approach is already being refined, and it will continue to be refined.
So it looks as though the data will come out more quickly in future.
I cannot tell you about the time lag, but there is a very important and more general question about the timing of the availability of data. Our specific point was about SEPA, which is, I believe, thinking about the timing. However, data is also held by the Registers of Scotland, which has a quarterly reporting pattern. Does that pattern suit? In other words, what are the times when we all need data, and what are the times when we need data? We are starting conversations and are in the middle of conversations about when it would be best to have data and timetables.
Is there a general problem that the data exists, but it is simply coming out too late?
Yes, there is, or there could be. That is why we want to have a discussion with all the data providers and come up with a view on when the data would be most useful and effective for future forecasts. That is an issue, but that is not to say that there is a problem that led us to say that any of the forecasts was totally unreasonable. As you know, we have not said that, but there is an issue that we would like to get under, and perhaps we could influence change.
I am sure that the committee would be interested to know whether you are getting data quickly from one area and not from another. We would be happy to make comments on that.
That was about the data as it is normally reported from those areas. I hear what you are saying and thank you very much. Leave it with us to have conversations about the regularity and periodicity of the data that comes out. If we need some muscle, I thank you for the offer.
Thank you.
Another specific figure is the 16 per cent. If I am reading things correctly, it appears to me that the landfill waste should raise a certain amount of tax, and 16 per cent less is arriving at Her Majesty’s Revenue and Customs. That also concerns me a wee bit. Are we clear about what that means?
That is our understanding of the matter. Stuff is being dumped and the tax levels should be higher, but the HMRC data is in conflict with that.
I—
That makes forecasting almost impossible.
I am sorry; I did not mean to interrupt. I understand where you are coming from.
There is a certain amount of waste tourism—that is the jargon term. Purveyors of waste sometimes cross borders and things become slightly complicated for the OBR, because it deals with firms that create and put waste to landfill. If those firms cross the border, things are a little harder to distinguish. That probably accounts for some of the difference. However, as I said, I have confidence that the changes that SEPA is bringing in will be much more specific and will more accurately reflect what is happening on the ground in Scotland. That is the place that we want to get to.
My final point is also about landfill tax. If I am correct, the point has been made that the environmental folk within Government are really keen to push down the amount of landfill, so there is quite an aggressive target, and the finance folk—John Swinney and so on—basically reflect that in the budget. The two approaches are therefore exactly the same. I accept that that gives consistent government, but should the two be exactly the same or should the environmental folk be aggressive and the finance folk cautious?
We have had some conversations about that. That is a very good question, as well.
I think that Andrew Hughes Hallett prompted some of the early discussion about the fact that the model took a straight-line approach that showed that, by year X, we should have only a certain amount going to landfill and where we are now. There has been some progress because of the escalation of the taxes, which has had a behavioural impact. What else does one do but show a straight line? That is really what you are reflecting.
10:15That brings us back to the question about challenge and inquiry, which is really discussion. We had a bit of discussion about whether raising the tax significantly would have a different impact. Would firms do something else with their waste? It might or might not go to landfill, it might go down south or it might go somewhere else. We explored some of those issues. However, you asked whether the current approach is a good way to do it; it is an acceptable way to do it.
Yes. The only thing that I would add is that we need to do some backtracking on that, too. If the trend is as simple as a straight line we ought to be able to look at it and ask, “Did they keep to it or are they slipping?” I think that the word “slippage” appears somewhere in the report. At the moment, we do not really know what is happening.
The difficulty, of course, is that we do not know that those really are the targets. We know what the end target is in 25 years’ time or so, so we know two points on the line, but in principle anything might happen in between. If we put a bit of pressure on and ask, “Are you are keeping up with this?” things will eventually become clearer.
Susan Rice is right about the difficulties of knowing exactly what we are talking about, because OBR data and SEPA data are somewhat different. SEPA data is Scottish, so we presume that it is better. Therefore we hope that we are in the right place.
There is also a twist, which we do not fully understand, in that there is a difference in the mix of what is being dumped in the ground, and the tax rates differ in that regard. We do not have specific data on that sort of thing. Numerically it probably does not make a huge difference, but the effect will be a little bit more where the differences are coming.
If I was asked for my private opinion I would reckon that the SEPA approach is more appropriate for the forecasts, but I cannot prove that—yet. We might be able to prove it later.
The issue is very much on our agenda and has been discussed.
Thank you.
Michael McMahon is next.
My questions have been asked—by John Mason—and answered.
Okay.
What happens next for the Scottish Fiscal Commission? You have produced your report to set alongside the draft budget. What is next in your workstream?
A number of things. We have already had a bit of a drains-up session about what we did—there was a very intensive effort to get the report out. From talking to the fiscal unit in the Government, my understanding is that the unit will shortly have its own drains-up about the process, which I think was intensive for that team, too. We have agreed that we will then come together and talk about what worked, what did not work and how we can work together more effectively, if that is needed. Our primary assumption is that we want to start much earlier for next year’s budget. We will be able to do that next year; we could not start early this year.
We are continuing to create and develop what I call relationships with our relevant organisations and bodies—I mentioned some of the Scottish agencies. We spoke to the chairman of the OBR at the beginning of the process in August, and we have been in touch and will talk to him more fully.
We have had a conversation with the Parliament’s budget unit—that is the wrong title, but I am talking about Simon Wakefield’s unit, which looks at the expenditure side of the budget. We want to make sure that we know what people are doing, to see whether anyone has ways in which they can help us or feed into our work.
We have been in touch with the interparliamentary finance network, which is the network of the UK nations’ fiscal and budgetary bodies. IPFIN has a get-together in November, to which I believe that I will be going—IPFIN invited me—so that we can go, meet, talk and find out.
We have been in touch with the Organisation for Economic Co-operation and Development, which has a get-together in spring. Again, we have been invited to that, and we think that it will be useful, because we can learn from fiscal commissions from a number of countries. When the Finance Committee was doing work on whether to set up a fiscal commission, I think in February, you took evidence from the Swedish and Irish commissions. We have been in touch with both commissions. In fact, Andrew Hughes Hallett has spoken to the Irish commission and we have some paperwork from the Swedish commission. Part of our work is to find out more about how others do it. It is about building relationships, reaching more widely around the networks, understanding what we need and want, and having the right timetable to go into next year’s budget round.
I want to pick up on some of the questions that you are raising with us this morning, some of which came out of our report. The core of what we need to do is to explore those questions and find some answers. How much data is enough? What do we do when we do not have it? These are important questions for the future. We will set out to do all of that.
The other part of our work programme might seem like an aside but it is important. As I think the committee knows, we are being hosted at the University of Glasgow, which has been very co-operative and is in the process of giving us a little office and helping us to find a couple of research assistants. We need to continue that process until we are actually functioning there. We have a website, which, as you might have noticed, does not have very much on it. However, we got the report out at the right time, which was what really mattered, and our three names are on it. We want to do a little bit more with the website.
If anyone has any suggestions about what they would like us to do, please feed them in.
Thank you.
A couple of members have asked about the differences between projections from the OBR and the Scottish Government projections for devolved taxes, and theories have been suggested for those differences. I am looking at the tables that I have in front of me. The OBR table is from March of this year and I presume that the Scottish Government table is from September or October. The OBR table is on stamp duty and land tax, and the Scottish Government table is on land and buildings transaction tax. Could either of those factors go some way towards explaining the differences?
The simple answer is absolutely yes. The Scottish data will not all come from one point in time, whether it be 1 August or 1 September or whenever. The organisations will have done their best to get the best and most up-to-date data that they could, but there will be a timing difference and you are absolutely right that that makes a difference.
Towards the bottom of page 4 of your report, you talk about the residential model for LBTT. Paragraph iii) says:
“the relatively high tax rates applicable to the upper band of the new LBTT may also induce an additional behavioural response which has not been factored into the forecast.”
The committee has talked about behavioural response and quizzed others on it. Were you able to see whether any work was being done on the behavioural response to LBTT and landfill tax? Are you saying that none is being done but you think that it ought to be done in future?
The formal techniques that the Scottish Government is using do not include any behavioural responses. It is not that kind of modelling.
On the housing market side of things, the Government has looked at various models that are more behavioural and structural and which would include the kind of effect you are asking about. However, it has not been particularly successful in finding workable versions of those models—such models are typically difficult to develop successfully. At the moment, the Scottish Government does not have such models and we are encouraging it to look into the possibility of developing them.
One of the reasons for this not being at the top of the agenda is that the value of the transactions in the top band is rather small, so the focus has been on other things first. If there was a behavioural response, I presume that it would be driven by the relatively small difference between the tax rates in the top band and those in the next band down. The difference between those bands and the one below that is rather larger.
Your question is the right one, but the answer might be because of a different bit of the distribution. That does not help you very much.
You said that you wanted to see enhanced forecasting methods. One of the main areas that has been talked about is data, so I will not ask about that. Other than in relation to data, which it will obviously take a degree of time to get, are there any other specific obvious areas where you think that the current forecasting methods could be improved for next year?
We have discussed data. If there are ways to bring in any of the behavioural aspects, that would be helpful. That would apply to each of the taxes—there are potentially different behavioural responses to each of them. I turn again to my colleagues to give you a fuller answer.
Without knowing exactly how far it has gone, we are aware that the Scottish Government people are working on getting some behavioural responses into the non-domestic rates bit. We are watching that and we will perhaps go through the process with the Scottish Government step by step to see what we can say before we get to the next report. The same should apply elsewhere, but the work is probably not as advanced at the moment.
I will move on to non-domestic rates. I think that your view as a group was that the buoyancy increase for business rates growth was on the optimistic side—I think that you used that term. The Government has reduced its forecast slightly on the back of that. Can you talk me through what happened there? Did you say, “We think it’s optimistic by X per cent and we think you should do this” or did you simply use the term “optimistic side” and the Government then did the work? It would be useful to know what happened.
We looked at the Government’s approach to forecasting. I think it has eight years of data showing the historical average for growth and buoyancy over that period, as well as a range of macroeconomic and microeconomic indicators, which it uses to try to suggest whether the economy will grow faster or slower than in previous years. It looked at those indicators, took a reasonably optimistic view of them and then went to the upper level of the historical bands that we had observed for buoyancy. Perhaps it was correct to take an optimistic view of the economic conditions going forward, but, in relation to the impact on buoyancy, it was going to the upper limit of what had been observed in the past. We pointed out that that seemed to be at the upper limit of anything that we had seen in the limited data that we have and said that it appeared optimistic. We said that the view might be reasonable, but it seemed on the optimistic side. The Scottish Government adjusted its forecasts on the basis of that, but we did not tell it what adjustment to make.
It is not our job to do that.
My last question is on business rates, which ties in with the previous question. There are previous years’ data to look at. For next year could you look at the estimates and projections for business rates? We have 10 or 15 years’ worth of data for that. Could you look at the original estimates versus the outturn and what the estimates were based on? Could work be done on that to see how we can improve the model going forward?
Yes. For all taxes we would seek to look at forecasts and outcomes. The more information we have on that, the more we can inform improvements in modelling.
It is a difficult thing to do, because you want to look at the differences and then ask why they happened. It is a cheap point to say that what happened in this case tells you that the inquiry and challenge approach is effective—that has happened on several points. Considering how much of an allowance you would want to make in the future is much more judgmental.
Yes, but part of our job is to make a judgment and we do that better and better the more data and evidence we have, so Gavin Brown has made a good suggestion.
I am grateful. Thank you.
Does the Scottish Fiscal Commission have any input into or involvement with the Smith commission?
10:30
I will explain exactly the involvement. I wrote to Lord Smith a little while after his appointment was announced and wished him well in a challenging task; I know him, so it was fairly easy to say it in these words. I said, “I just want to remind you that the new Scottish Fiscal Commission exists.”
We do not intend to make a submission to the Smith commission because that is not our role. On the other hand, if the commission comes to us at any point to share an idea, try something out or ask us questions, we would respond, but only within our remit. We are very conscious of our remit and we are conscious not to stray. That seemed to be the appropriate way to engage with the Smith commission. I would appreciate any guidance on that.
In your other roles, will some of you make submissions to the Smith commission?
There is a non-incorporated body called the 2020 climate group, which is a group of volunteers from business, Government and others who are trying to influence the momentum on the climate change agenda. It may possibly make a statement to the Smith commission, but it has not done so yet. My thinking was that I could be excluded, but it is not even a real body, in the sense that it is a non-organisation—it is a group of volunteers.
Looking at my other engagements, I have a very big involvement in the festivals forum, as I chair the board that looks after the strategic positioning of the city of Edinburgh and all the festivals. The board has not made a submission. If that was discussed, I would think about any relationship to the Fiscal Commission and would not necessarily put my name to something. I am not aware of any other inputs that would cross over my worlds.
My answer is exactly the same. The Fiscal Commission does not do policy advocacy, so it could not make a submission. If the Smith commission came to us, we could say something, although it is a bit difficult to say what that might be.
If the Smith commission approached us, we would see on what basis it did so. It may not approach us; why would it, in a sense, given what we are responsible for? However, if it did, we would discuss whether and how we could respond.
I have colleagues at the University of Glasgow who are making written submissions to the Smith commission. I was asked to be involved but declined, given my position on the Fiscal Commission.
Okay. Thank you.
I return to the data—the missing data, if you like. When you looked at the systems that the OBR employs, did you find them really impressive? Did you get excited about the detail in the data collection that it has to make its forecast?
The OBR has access to the full set of Office for National Statistics and Treasury data that is needed to run multihundred equation models, so it has a large set of data on which to base its forecast.
Has its forecasting with such data always been impressive? My experience on this committee is not great, but the OBR has often been fairly spectacularly wrong in its forecasting.
The OBR’s approach is to build a very large macroeconometric model. The academic literature performs horse races between alternative approaches to forecasting and puts that style of macroeconomic model against relatively simple statistical models. Over a short-term horizon, short-term statistical models often outperform large-scale macroeconometric models, but the OBR needs to produce a set of forecasts for a huge number of variables and it needs to be coherent and consistent. It needs a big model to perform that task, so there are trade-offs.
It might be worth noting that we have exactly that kind of debate among ourselves about whether one type of approach—the big model or the simpler approach—is better. The simpler approach that is being taken here out of necessity also felt quite comfortable and helped us to come to the conclusion that the results were reasonable.
On the collection of that data, do you foresee yourself engaging with other institutions such as universities, which often do quite detailed research in some of the areas that would be relevant to you?
Yes. We spoke about that at the beginning and we probably know some of the universities that gather up a fair amount of data. Again, we have reached out a little bit, and part of our follow-up programme is considering the bodies that we might have relationships with. We might also want to commission some research at some point, for which we would be likely to turn to a university because its work would be impartial and independent and would have the integrity of academic research.
I can think of several universities or institutes that do such research, but it is usually done on one bit of the economy. I can think of one example that focused on the labour market, so we could look at the important factors in that area to see whether they affect the forecasts that we are looking at. However, I am not sure that we will find very much research on the Scottish economy per se—perhaps such work should be done. I can think of one example but, having been into it, I would not use that model. The research is patchy.
In general, you are saying that the focus of the work in some of the areas that you are looking at is almost microscopic compared with that of the OBR, but the chances are that it could be much more exact.
Do you mean the smaller models?
Yes.
There is a chance that that could be the case, but part of that could be about the data collection and the integrity of the data as it is stored and reported. The biggest issue lies in getting genuinely Scottish-specific data. That is important and it would help hugely when it comes to quality and the assurance that we feel towards the forecasts. We see the way forward, what is intended and what has started here. That is an important place to be.
Meeting fiscal commissions from elsewhere would allow us to chat about the kind of question that Jean Urquhart raised with somebody completely outside our ambit in order to get a little bit of their experience and learn what they have found useful. It is a good question.
Thank you.
That concludes questions from the committee. Would the witnesses like to make any further points?
Do you have anything, Andrew?
No, I think that we have covered everything.
I just want to thank you, because you have obviously gone into the report in some detail. I am a note-taker and I have noted a number of the committee’s questions. I will make sure that they come up on our agendas for discussion. We appreciate that.
Thank you very much for the evidence that you have given to the committee.
I suspend the meeting until 10.45 to allow a changeover of witnesses and to give members a natural break.
10:38 Meeting suspended.