Economy, Jobs and Fair Work Committee

Meeting date: Tuesday, October 31, 2017

Official Report 466KB pdf

---

Economic Data

Welcome back to this morning’s session. I welcome the witnesses from the Scottish Fiscal Commission: David Wilson, chief commissioner; John Ireland, chief executive; and Mairi Spowage, deputy chief executive.

I remind members to keep questions succinct and to the point. Our witnesses need not answer every question; there will be an opportunity to submit further information in writing should there be aspects that they would like to cover following the meeting.

We move to an opening statement from David Wilson.

Thank you for the opportunity to talk to the committee this morning.

We are particularly pleased to be able to assist in any way we can with this important inquiry. We share with the committee a strong interest in the quality and accessibility of Scottish economic statistics, and we are very keen to support you in whatever way we can.

Before introducing my colleagues, I would like to make some introductory remarks; first, about the Scottish Fiscal Commission; and, secondly, to frame the further advice that we will give over the course of the morning.

The committee will know that the Scottish Fiscal Commission has been in place since the middle of 2014. When it was initially established, it was a non-statutory organisation, the function of which was to scrutinise the forecasts that the Scottish Government produces of the likely revenues from devolved taxes under the new arrangements under the Scotland Act 2012 and following the Smith commission. Over time we have evolved as an organisation and, since April this year, we are slightly different. The key point of continuity is Lady Susan Rice, the chair of the commission, but, since April, we have been different in an important way and I want to ensure that that is covered for the record.

The first change is that we are now a statutory organisation. There are three commissioners—Lady Susan Rice, me and Professor Alasdair Smith—and we have a growing organisation that is designed to support the work that we do, led very ably by John Ireland and Mairi Spowage.

Our role now is to produce the forecasts and to be responsible for them, rather than to scrutinise Scottish Government forecasts. Under our remit, we will produce forecasts of Scottish onshore gross domestic product and estimates of the likely revenues from the devolved taxes over a five-year period, and we will also make estimates of the likely expenditure on a number of specific social security responsibilities that are being devolved to the Scottish Government. I want to emphasise that shift in our role.

In taking forward the work that we do, last month we published a report, “Current approach to forecasting September 2017”, which sets out how we intend to go about undertaking our function. We will be a significant and major user of official and national statistics, which are largely produced by the Scottish Government and the Office for National Statistics. We see the commission as a user, not a producer, of the range of different statistics that are available—we produce forecasts that are not, in the jargon of this world, official statistics; they are not even statistics in the way that we understand them. We will produce forecasts building on the information that is available to us, using a range of datasets, models and information.

Finally, I will comment briefly on some of the issues that have been raised so far in the inquiry. Clearly, the committee’s inquiry is significantly wider than the commission’s specific responsibilities, so there are a whole range of very important issues that need to be looked at that probably go beyond the direct remit of the commission. However, particularly given the importance of estimating or forecasting the future of GDP, the commission’s interest in available statistics actually ranges quite widely. I emphasise that although the committee’s review probably goes wider than us, we are interested in a very significant range of the different data sources that you are looking at.

I also wanted to mention that the responses in our submission focus on economic statistics issues, rather than some wider and detailed issues around the fiscal forecasts that we will produce. For example, we have not given you lots of information about how we will develop forecasts of revenues from landfill tax and some of the other more detailed taxes. We will be happy to give you that information, but in the main our submission focuses on economic statistics, which feed into overall economic determinants and the estimates of economic growth.

That is all I wanted to say by way of introduction, except for a bit more about the team that is here. As the convener said, I am David Wilson, a commissioner on the Scottish Fiscal Commission. My day job is as director of the international public policy institute at the University of Strathclyde and my career has included a number of years working in the Scottish Government on a variety of economic issues, including as director of business and initially—many years ago—as an economic assistant working on GDP and labour market statistics.

On my left is Mairi Spowage, deputy chief executive of the Scottish Fiscal Commission. Mairi is a professional statistician by training and experience. She was previously head of national accounts in the Scottish Government and can speak authoritatively on those issues. On my right is John Ireland, chief executive of the Scottish Fiscal Commission, who, in a previous life, was also a director of the Fraser of Allander institute and worked in some detail on issues of this kind.

That is my introduction. I look forward to the rest of the morning.

We have heard from a number of witnesses about gaps and weaknesses in Scottish data. In your evidence, you highlight a number of challenges, including construction, lack of real-terms GDP data for expenditure and income methods, and the lack of Scottish-specific price data. How certain can you be that the figures and statistics that you produce and rely upon are robust and accurate?

Perhaps I should make some further comments before answering that question. The first thing is to reiterate many of the points that have already been put to you by a number of witnesses about the quality and accessibility of economic statistics in Scotland. John McLaren made a good point when he said that the first thing that we need to finish is a full set of national accounts. I am not sure that we will ever be finished developing a full set of national accounts, but over time there has been a significant improvement in the range of economic statistics available and the range and understanding of economic activity in Scotland.

Just to be clear, we are in a much better position now than we were 10 or 20 years ago; we are probably in a better position than other areas of the UK in terms of the information that is available in Scotland; and we are also in a strong position in that there are a number of things that will contribute to further improvement in the quality both of statistics and the analysis that underpins work on the Scottish economy. I mention specifically the work that the Scottish Government is doing to improve its range of statistics. I understand that you may be hearing evidence from Scottish Government officials in future weeks. They have set out some clear areas in which they intend to improve. If I may say so, I think that the work of the Fraser Allander institute is also contributing to greater analysis and scrutiny, and I would also like to draw attention to our own activities. Our feeling is that, now that the Scottish Fiscal Commission is established as a statutory organisation with a clear purpose to develop forecasts, there will be an on-going process of increased demand for statistics and increased scrutiny and attention, which will in itself lead to further improvements.

That is a slight caveat before I come back to your question—we are in a better place than we were. That said, there are many gaps. Witnesses have already identified gaps around imports and exports, and trade data is a key area. There are issues around the availability of information on investment, which is a key factor, and David Bell made a good point about weaknesses in the data on earnings and inequalities. There are areas that need to be touched on and improved and, overall, I have been struck by the consistency of the advice that you have been given about areas for improvement.

We wanted to identify some specific issues that could be taken forward quickly and at reasonable cost that would directly help the work that we do. Those are the four areas that we have identified in the submission, and I am happy to go into those in a bit more detail. Alongside those, there are a number of other areas, including improved trade statistics and improved information investment, that would greatly assist our work, but we wanted to emphasise and focus on four specific areas that other people might not have raised.

On the direct question about our confidence in our forecasts, which we will publish initially in the middle of December, we will make the best use of the information that is available to us and will develop our analysis and judgments utilising all the available information. We will set out clearly in our reports our relative confidence in what we are able to do, the upsides and downsides and any risks and concerns that we might have about the robustness of the data. We are confident that we will make the best estimates that we can in relation to our duties, but in doing so we will have to take account of the strength and weaknesses of the available information.

10:00  

Thank you for your written submission, a small portion of which I will read out:

“Given the information that is available, and accepting the uncertainty inherent in it, the Commission strongly encourages the Scottish Government to produce more information in certain instances which would support the Commission’s activities.”

Will you explain a little bit further what you mean by that? Which areas does the commission need more information on in order to carry out its work?

I suggest that we go through the four areas that we have identified. They may appear to be fairly technical—they are fairly technical—but I would not say that they are in any sense technical and unimportant. They go to the heart of the need for continuing improvement in the quality of Scottish economic statistics more generally and of those that specifically support the development of our GDP estimates and, in turn, our estimates in relation to the devolved taxes, many of which are influenced by our forecasts of GDP and economic activity.

The first area is the availability of specific information about inflation in Scotland. Some information is available about how inflation in consumer and producer prices might differ in Scotland. It is relatively limited in focus, but specific in certain areas.

That issue has long been recognised. There has been some evolution, but there is scope for further improvement in the robustness of our baseline estimates of where we think the economy currently is in our forecasts.

In order to go into the detail a bit more, I suggest that Mairi gives you bit more background information about the specifics of price deflators.

Before I go into that, I will respond to Ash Denham’s question about our statement on accepting the “uncertainty inherent” in information. That referred to the fact that because information is collected from businesses at the Great Britain level, much of the data for Scotland needs to be derived through regionalisation in some way, which is normally done through employment shares. We were trying to separate out whether the collection of that data should be changed or improved from what is done with the data, given where we are and the data that we have. How could the collection of the data be improved, given that making changes to the collection of business data would need much bigger investment?

On the issue of price data, some sources give a feel for the price differences in Scotland and the UK as a whole, but those sources are very limited. In the past, the ONS has produced information about different price levels in different parts of the UK, but you do not get a feel for any change over time—it really is not a useful source of information for national accountants who are trying to estimate differences over time. The national accountants in the Scottish Government build up their view of the economy at a very detailed level, which allows them to take account of different industrial structures in Scotland.

Overall, on aggregate, there will be different price movements in Scotland from those in the rest of the UK. However, on the whole, UK-level price information is used. Service sector information is broad brush and not very detailed even at the UK level—that is one problem that the Bean review identified—so the different industrial structures are not being taken into account in a very granular way.

It is difficult to know how different the data on consumer or producer prices in Scotland would be, because we do not have the data.

Do you suspect that there is a difference, though?

It is reasonable to say that there is a difference, particularly in producer prices, because of the different industrial make-up and because Scottish companies potentially have different import propensities and other things—there are different sorts of companies, so it is a reasonable hypothesis that there is a different experience. That is particularly so for businesses; I am not sure whether there is a different consumer experience in Scotland compared with the experience in the UK as a whole. It is just a gut feeling; I am not sure. The issue should definitely be looked into.

One source might be the surveys that the ONS uses to collect information and construct the consumer prices index and the producer price indices. The Scottish Government could work with the ONS to consider that data to see whether there is potential to produce Scotland-specific indices, maybe not monthly but perhaps annually to begin with, and any prospect of using that data. That would be the place to start. I suspect that the ONS will probably use the powers in the Digital Economy Act 2017 to get more administrative data on prices; it might compel companies to give it information so that it can build better consumer price indices.

The fact that the Scottish Government has boosted the living costs and food survey for use for VAT assignment provides other opportunities. That survey is the source of the weighting for price indices, and there will be better data on that for Scotland. There is quite a lot of scope to investigate what is already collected and to see whether it can be enhanced to produce something for Scotland that would give us a feel for how consumer prices, for example, evolve, even if that is done on only an annual basis for the past few years. The analyst in me would certainly be interested in looking into that.

You have mentioned the price indices, but are there any further areas?

I will finish a point that Mairi Spowage started. We would not say that there is a gigantic hole or gap, or a profound problem, in relation to deflators, but as we gradually improve the information that is available—as it evolves—we think that we should consider that area more. The review that Professor Sir Charles Bean has undertaken of overall statistics refers to the fact that developing that area might not pass the cost benefit test—there have been a number of discussions about whether it might be too costly, because there would be insufficient benefit. To an extent, we are not well placed to judge either the costs or the benefits, but we would like the issue to be looked into more. There is certainly a demand from us that the Scottish Government continue to look into it. We would support that.

That discussion leads directly to a second one, which is about using the deflators that could be produced in a bit greater detail and granularity to develop a more accurate picture of the expenditure version of GDP. GDP is talked about in many different ways and there are a number of different ways of developing it. Developing the expenditure version of GDP using Scottish deflators would greatly assist our work, particularly in understanding how spending activity in Scotland might evolve and how it impacts on a range of wider issues such as the GDP and income tax forecasts.

That would be valuable to us even if it used the current UK-level deflators, and it would be fairly straightforward for the Scottish Government to publish alongside its current price information. An official statistics source of that information would greatly assist us, rather than having to deflate it ourselves after the fact, which is what we do at the moment. That could be a quick win while the issue of Scottish deflators is investigated.

If that was done centrally by the Scottish Government, it would mean that everybody who uses the data would have access to the same data. Rather than the Fraser of Allander institute doing it one way and us doing it another way, it would produce standardisation, which adds to the power of a quick win on that.

Panellists have explained where there are gaps in the data. I was interested and slightly worried by Mr Wilson’s phrase that you will “make the best estimates” that you can make. I suppose that you would do that even if you had no data. What are the risks to your forecasts from the lack of data? Should we be worried?

We will obviously base the forecasts on the best information we can get. I want to emphasise that there is a significant amount of information out there. We have described how we try to complete the picture that we paint of what is happening in the Scottish economy. A number of areas in which we seek further information can be described—I can go into those. There are issues about earnings and wages that we would be particularly interested in going into.

On the risks to the forecasts, I want to distinguish between uncertainty that arises from lack of information about what is happening in the current economy or the past economy—in other words, gaps in data and information—and risks and uncertainties in the future course of the economy, which are inherently unknown: they have not happened yet, so we have to make judgments. We obviously use the best possible information on past and current activity to try to understand how it might influence the future, and to make our best judgment. That is what we all do.

Inevitably, SFC forecasts will be compared with those from down south. Even if you have a better ability to judge what will happen in the future, you are still dependent on what has happened in the past. It seems to me that if folk in London have much better data than you have, that will put you at a huge disadvantage when the public or the media make comparisons.

I will draw a distinction on your point about the responsible authorities in London having better information than we do. They undoubtedly have better and more complete information about the UK economy than we do about the Scottish economy, but my feeling is that we have information and analysis about the Scottish economy that are at least as good, if not better, than the what the authorities in London have about the Scottish economy.

We are in a strong position to develop ever-improving analysis and information about the Scottish economy, but at a fundamental level—I think that all the witnesses have said this—the range of available official national statistics about the Scottish economy is inevitably less than the equivalent for the UK economy. That is in the nature of—in national accounts terms—subnational data.

In respect of many data sources that are key to producing economic forecasts and forecasts of the labour market—which are key for our income tax forecasts—it is more challenging for Scotland because data are more volatile because Scotland is a smaller economy: for example, the labour force survey on which such forecasts are based uses a smaller sample. Producing forecasts for a smaller economy with slightly more volatile data might mean that our forecasts are subject more to data revisions and to other things that are outwith our control.

10:15  

One way that we deal with that is by taking a more aggregated view: rather than putting so much emphasis on month-to-month movements, we can look at movements over a quarter or annually. That may have its downside, but I think that it is a sensible approach to take in dealing with more volatile information.

It will be important for us to present areas in which we have less certainty or in which we have to make broad judgments, which is part of producing economic forecasts, in particular. It is necessary to make broad judgments about the long-term path for the economy and for things such as productivity and population growth, which influence economic growth in the long term. Volatility in labour market data, for example, makes it challenging to produce forecasts for the near term, in particular.

On labour market data, the surveys are clearly limited, but Her Majesty’s Revenue and Customs is basically sitting on all the data, is it not? Is there a problem getting the data from it, or is it that the big change in the whole system that you have referred to means that it is just too much to consider doing at the moment?

I will develop that point, because labour market data—specifically, income and earnings data—is a really good example. In order to develop our overall modelling for gross domestic product and, more specifically, our income tax forecasts, we need to develop a fairly detailed understanding of how much people in Scotland earn. One might think that that information would be readily available; it is very important information that we need to collect from a range of sources.

For example—I will not go into particular detail—we use what is known as the survey of personal incomes, which is a very detailed sample survey of income tax records that are held by HMRC, which John Mason mentioned. That survey is the fundamental basis for our modelling of income tax forecasts, in particular. In one sense, it is very comprehensive—there are more than 40,000 records in Scotland and the survey produces a wealth of detail about the structure and composition of incomes in Scotland.

Inevitably, though, that information takes time to be developed, produced and analysed. I do not want to say that the information that we are using is out of date, but I think that we are currently using information from 2014-15 as the basis for our modelling. That is the most comprehensive assessment. The data that we currently use is what is known in the jargon as the public use tape—the published data that HMRC makes available.

As John Mason rightly said, HMRC has much more detailed information and, as all the recent reports about economic statistics have been recommending, there is a big push to use such administrative data and to analyse it in a bit more detail. We would love to move further in that direction.

A range of other data can be used to give a wider and more detailed picture of what is happening in incomes in Scotland, such as the annual survey of hours and earnings, for which the latest figures have come out very recently. That is quite a different survey; it is an annual survey and it is more up to date, so it can add some information and colour on what has happened more recently. Scottish information is included in that. There is, inevitably, a set of issues about linking together the annual survey with the survey of personal incomes. That linking is a challenge in much of what we do.

Finally, just to make the link back to our four asks in terms of more information, we also mentioned a set of data that the UK Government develops. There is, for example, a survey of average weekly earnings, which we feel would greatly assist our understanding of what is happening—not quite in real time, but close to it—in the labour market.

We would very much support the development of more information and more recent and contemporaneous information. At the moment, we do not have access to Scotland-specific information, so more surveys would probably be required, but that would greatly assist our work. That is the third ask that we identified.

To summarise, the task that we must undertake is to piece together the available information, which is stronger in some areas than it is in others.

Some of this is quite difficult. I appreciate your answer.

Is there any way of measuring the risk? When you mentioned HMRC data from 2014-15, I immediately got worried, because that seems like a long time ago: many things have happened since then. When you make forecasts, only part of the process is about looking at the past. Regardless of how up to date your past information is, you still have to make judgments and forecasts. Is there any way of measuring the risk in that area?

It would be very difficult to apply a statistical measure to what we see as the ex ante risk in the forecasts; I do not think that we will be able to do that. Forecasters in other areas try to provide upper and lower estimates for their forecasts. Members will have seen what are known as a fan diagrams in various organisations’ documents.

In our forecasts, we will set out and elaborate on what we see as the risks and uncertainties on the upside and on the downside, but the role that is set out for us in statute is highly specific: it is to produce a single set of estimates for use in the Government’s budget documents. Our emphasis will be on producing the best possible single set of estimates, while setting out and elaborating on what the risks on either side might be.

We are also required by the Scottish Fiscal Commission Act 2016 to produce a forecast evaluation. In September, we produced an evaluation for some of our fiscal forecasts and some of the Government’s fiscal forecasts. Next September, we will look at how the forecast that we will produce in December has performed. Once we have a number of those forecast evaluations, it will be possible to get a much better understanding of how accurate the forecasts are. Unfortunately, it will take a bit of time to build up that track record. Other forecasting institutions that have been forecasting for longer than we have do that already, although they publish the information in a slightly shorter form. Such information is available.

Where we see key risks to forecasts or key areas of uncertainty, it will be important for us to set out sensitivity analyses of those areas so that users of our documents can understand how much difference various assumptions on key factors might make to our forecasts. It is important to be transparent about the uncertainties and risks that we are dealing with when we produce forecasts.

I could go on all day on this subject. I am sure that you could, too, but the convener wants to move on.

Yes, I do. Andy Wightman is next.

You refer in your submission to time lags in data being published. What concerns do you have about time lags in publication of the data that you need to make your forecasts? How could they be overcome?

I will not go over the wages and incomes story again. Perhaps the most obvious and important example of an area in which progress is being made is the GDP estimates that are produced by the Scottish Government. A number of other witnesses have mentioned that the time lag in availability of estimates of quarterly Scottish GDP has reduced over time. From memory, I think that it has reduced from 110 days after the end of the reference period—that was the case before 2016—to, I think, about 87 days after the end of the reference period. We would like to have access to the GDP estimates that are produced by the Scottish Government as soon as it is able to publish estimates in which it has confidence.

However, I sound a note of caution, which also goes back to Mr Mason’s questions. We do not want—it would be of great concern to us—a push to accelerate publication of statistics, which might then be published before there is confidence in them, and would therefore be revised in the future to a greater degree. The last thing that we want is early publication of estimates but with reduced reliability. A balance needs to be struck.

Access to GDP estimates is the principal example to which I would draw attention. Mairi Spowage and John Ireland might want to comment.

There are a number of sources on the fiscal side that would assist us in our forecasts. We have mentioned the survey of personal incomes and the fact that we are dealing with 2014-15 data at the moment. We have access only to the public-use version of that; we do not have more information than anyone else has.

However, the SFC has just been approved as a researcher at the HMRC datalab. That means that we have to go to London, but it also means that we will be able to look at the more detailed survey of personal incomes data that sits in the background, which will give us a better feel for the years for which HMRC has not published a public-use version of the survey, and for the areas of the survey that are anonymised to prevent us from identifying anyone. That will give us a better idea of the taxpayers, particularly at the top end. That is good progress, but we would like to be able to access that information without having actually to go to the datalab, so we are in discussions with HMRC about that. Obviously, its principal concern is taxpayer confidentiality, which is fair enough.

Other information that would be useful includes information about how much income tax is being collected as we go along, such as there is at UK level in the monthly receipts. HMRC is developing real-time information on that; it is too early to say what its quality will be or how much it will chime with the receipts that are being collected. That is in development, so in future years it will be a more up-to-date source of information on actual receipts collected.

We must not forget about the Revenue Scotland information, which we use for our devolved taxes. Revenue Scotland has significantly speeded up its production of monthly statistics on land and buildings transaction tax, and it has massively expanded the analysis that it produces each month. That has happened partly due to the requests that we have been making of Revenue Scotland, but it has made all of that extra analysis available, which is really useful. It has also significantly brought forward its publication of quarterly landfill tax revenues. The more recent the information that we have, particularly for the very new taxes, the easier it is to forecast revenues.

I can clarify a point that I made earlier about the time lag in the publication of the GDP estimates; I have found my notes. The current gap between the end of the reference period and publication of the quarterly GDP estimates is 97 days. If we compare that with the equivalent estimates at UK level, the ONS produces the preliminary estimates after 25 days and what are, in effect, final estimates after 84 days. That is another example of UK information being ahead of the information for Scotland. As I think everybody knows, any reduction in that gap would be helpful to us, subject to the caveats that I mentioned earlier.

I have a couple more brief questions. Do you have to pay for any of the data that you require or is it all made available under protocols or openly available?

10:30  

We have to pay for some of the data. For example, we are in the process of paying Registers of Scotland for data on house prices and property transactions.

There are issues around that. Some data is pretty expensive. Outside Scotland, the Civil Aviation Authority produces some good data on passengers, which is especially important in relation to air passenger duty. It is, however, so expensive that we cannot afford to buy the full dataset, so we buy only a small portion of it. There are issues about public authorities, be they in Scotland or the UK, charging for their data. Of course, it has to be said that they do not do that only to make money; it costs money to employ statisticians and there is a resource cost of manipulating data, in particular administrative data, and producing it in a form that people like us can use. I suppose that the question is this: how can that resource cost be met in a way that benefits the public sector, academics and other people who are interested in using the data?

HMRC also buys that data from the CAA. My understanding is that the Scottish Government and Revenue Scotland buy the ROS data, too. We are no different from them. Obviously, we have a statutory right of access to information but, where there is a charging model in place, our statutory right of access does not override it.

Thank you. Perhaps we might combine some of those questions in a later inquiry.

Can you confirm that you have no statutory role in estimating the quantum of future assignation of VAT? I think that, according to the fiscal framework, that will not have an impact on the Scottish budget anyway, until the end of the operation of the current fiscal framework. Do you envisage having a role in estimation of VAT receipts from the end of the current fiscal framework?

My understanding is that it has not been decided who will be responsible for forecasting VAT receipts in Scotland. It is not currently in our remit.

I make it clear that we are not doing any work on that at the moment. That will be a decision for Parliament and Governments in the future.

In the fiscal framework, it is for the two Governments to decide what body will be responsible for forecasting VAT receipts. That decision has not been made, as far as I know.

You have talked a lot about the income tax powers that are coming to Scotland. Obviously, they are one of the major areas of your forecasting. In your submission, you say that one key area in which improvements are needed concerns the statistics on earnings and on non-savings, non-dividend income tax. A lot of things are missed out from the assessment of household income. You talk about wages and salaries, but there are some forms of household income that do not fall into that category. I would like to know your feelings on that issue and how you are going to address those gaps as you make forecasts. It is an important issue.

That is a very good question. There is a set of issues about what might be seen as household income rather than personal income. Some of the questions that you have previously asked about using gross national income rather than gross domestic product have touched on extremely important issues—they are certainly important to us. However, our starting point involves making the best estimates of the specific revenues that we are tasked to deal with, and our specific task is to forecast income tax revenues. We develop all of the issues that we have already described to put that together.

We will try to understand to a level of depth what is happening in terms of hours worked, wider forms of income and activities within the labour market—clearly, a lot of complex changes are happening in the labour market that are impacting on personal incomes and household incomes. That is very much the context within which we will develop our income tax forecasts. However, to be clear—without being too heavy-handed about it—our immediate task is to develop the income tax forecasts. The wider context is crucially important, but we will inevitably focus on that specific task.

I completely understand that, but one worry that we politicians and policy makers have is that there might be a knock-on effect of people incorporating themselves in order to avoid income tax. How could we address that in our report and in our recommendations?

Tax-motivated incorporation is clearly a factor that we will need to take into account when we consider the behavioural impact of income tax changes. It is obviously relevant to UK authorities as well as Scottish authorities. We have been working quite closely with HMRC on its estimates of how incorporation might impact on forecasts and income. If there are different levels of income tax, there might a differential motivation for incorporation, and we will need to take views on that. At the moment we are working with HMRC to explicitly have that issue as part of the forecasts that we will publish in December. I welcome any comments from my colleagues on whether there are specific things that the committee could recommend. I suspect that there might be, although I cannot think of any off the top of my head.

HMRC has done a lot of detailed work on that, and it has been very open and shared a lot of the work with us. There has been quite a good collaboration between us and HMRC. The Office for Budget Responsibility obviously has a big interest in that at the UK level as well, and it has done a lot of modelling not just at the UK level but at the level of parts of the UK.

The issue is quite new, so perhaps we can have a think about it away from the committee and see whether we can come up with any specific things.

One thing that has come out of the evidence is that, although data is being collected by lots of agencies, such as enterprise agencies and business support agencies, it may not be being collected in a way that is useful to you. Some of the recommendations that I am talking about could be on that.

There are a number of areas in which further improvements could be made on the accessibility and quality of information. Opening up access and having greater publication of information by a number of public bodies would greatly help. I am happy to provide a bit more detail on that general issue and on incorporation specifically.

Our immediate interests would probably be in the information that public bodies hold in Scotland that might assist us with our fiscal forecasts. For example, John Ireland mentioned the Civil Aviation Authority and air passenger duty, and, with regard to issues around housing and property markets, information is held by Registers of Scotland and others. We have strong links with a number of those bodies. There may be further information and we would be happy to provide a note on that.

I know that my colleagues have more questions on that area.

As you mentioned, income tax receipts will be increasingly important. Approximately one third of Scotland’s budget will depend on them. At the moment, much of the data that you use is UK-wide data on earnings and wages. Given that the earnings profile in Scotland is different from that of the rest of the UK, in terms of average salary and basic-rate, higher-rate and additional-rate taxpayers, is there a concern about a slight mismatch if you are extrapolating from UK-wide data?

Inevitably we have to make judgments in this area. Generally, we have quite good information about the distinct composition and breakdown of incomes and earnings in Scotland. We have that data, but, as I touched on earlier, it is not always immediately up to date, and it can always be improved on. Although we have relatively good information, we must make sure that it is fed into our various modelling activities and that the trajectory of the forecasts takes into account recent events and activities. We need to bring in the more recent information on how things might be changing.

It is a much wider discussion, but how the labour market has developed since the financial crash has turned a number of our expectations on income pattern and the composition of the workforce on their head. There is some evidence that there are differences in how those have evolved between Scotland and the rest of the UK.

We are keen to make sure that we have up-to-date information. Generally, we have good information for our modelling work on income tax from the survey of personal incomes, but it is not always up to date. That is the nature of those data sets.

To follow up on that point, there seems to be a range of different figures—from around 12,000 to 19,000—for the number of additional-rate taxpayers in Scotland. Why is there such a large discrepancy in people’s estimates for the number of additional-rate taxpayers in Scotland?

I am not aware of the different estimates that Mr Lockhart has described. We use information based on 2014-15. It is a sample survey and, inevitably in a sample survey of upper-rate taxpayers, there is a degree of uncertainty. We take into account different sources of information. We consider that we have the best information available for the overall forecast that we make.

Does Mairi Spowage have anything to add?

I do not have the numbers off the top of my head. We could provide follow-up information on our understanding of the position. That is something that we would make available through our publications, because we know that there is interest in the distribution of different categories of taxpayers.

We have heard about leakage from people moving from income tax to corporation tax through dividends. What about leakage from additional and higher-rate taxpayers in Scotland reregistering elsewhere in the UK if there is a tax differential to encourage that? Would the commission be able to model or forecast that? Would you talk us through the impact on behaviour of differential tax regimes and what you would look at to anticipate what leakage might occur?

Yes. Building on work that the Scottish Government did some years ago, we have developed thinking on what, in economic jargon, are called the elasticities of the behavioural impact—our estimates of how much people may change their declarations and change where they are registered for tax. Those are factors that we need to take into account, and we are working with HMRC to develop our understanding of those elasticities.

It is important to take into account the academic evidence that is available, and we have looked at that. Inevitably, there is no established evidence base on the impact of differential income tax in a UK context, because there have been no differences so far, so there is no real information. There are a number of sources of information on how elasticities might play out from movements between countries and particularly between states in America. That information has been useful for us in developing initial estimates of what the impact might be.

When we have to make the estimates, as we expect we will have to do, we will set out information on the elasticities that we are using and make the basis for our estimates clear. We will also draw out the impact on the overall numbers.

The key point that I want to emphasise is that the estimates will be a judgment based on the best information that is available to us. We will set the basis out, building on the work that the Scottish Government did previously and the work that we are doing with HMRC.

10:45  

I have a final point on that. Under the fiscal framework, has it been settled how someone would be judged to be an English taxpayer as opposed to a taxpayer in Scotland? Is a day count fraction or number involved, or is there another test?

There is an established process that is led by Revenue Scotland rather than us to classify somebody as a Scottish—

I am sorry to interrupt, but it is led by HMRC.

I apologise. There is an established process that is led by HMRC to classify people as Scottish rate taxpayers rather than rest-of-the-UK taxpayers. We will utilise that information, but that decision is for HMRC.

I have a brief follow-up question on income tax. Obviously, Scottish taxpayers are now meant to be identified with an S code so, in your forecasting, the data that you will get in relation to Scottish taxpayers will be accurate. I think that the National Audit Office was concerned that there are still around 420,000 potential taxpayers who are excluded. Will you say something about the reliability of the S code? It is no doubt better than the survey of personal incomes.

My second question is about self-employment. I have yet to make my tax return for the 2016 calendar year, and it will be well into 2018 before my self-employed earnings are reflected in any national statistics. Self-employment is rising in Scotland. Given that there is a time lag for a certain category of taxpayers who do not declare their income for a considerable time after it has been earned, I presume that you will reflect that in your confidence about your forecasting when you present your forecasts.

I will touch briefly on the second question. I ask Mairi Spowage to pick up on the first question and to make any further comments on the second one.

When it comes to detailed tax information, it is inevitable that there is a time lag for people in self-employment as well as for people in employment, in terms of declarations. That is the nature of taxation, and that is why there are different time periods for the availability of information, which I have already been through in response to Mr Mason’s question.

The sharp increase in self-employed workers in the period since the financial crash is one of the features of the changing labour market that we have to take into account in making the forecasts, which are based on the information that is available to us about changing employment and self-employment patterns. We take that into account with as much information as is available to us.

As David Wilson said, we use the survey of personal incomes to forecast income tax liabilities. It is too early for us to be able to assess how the population of Scottish taxpayers in the SPI relates to the population of Scottish taxpayers who have been identified with S codes by HMRC. We cannot do that until there is more real-time information such as the monthly receipts and the ultimate outturn information that is published.

The self-assessment issue that Andy Wightman raised means that we will not have the full year’s picture for 2017-18 until the summer of 2019. That is one of the challenges, because we are trying to forecast liabilities and receipts. Currently, we use the SPI data as the best available data source, and we need to keep a watch on that. As the real-time information develops and we have the three sources together, we will be able to say how well the SPI has done, what the real-time information can tell us, and how that relates to the outturn liabilities at the end of the period. We will know more in future years but, basically, we have to base our forecasts on the SPI at the moment. Therefore, I cannot comment on how well the identification of Scottish taxpayers has gone.

My question is about the availability of localised or regional data in Scotland. Obviously, Scotland’s regions are very diverse. I want to know whether there is a shortfall in that regional and localised data and how that impacts on your forecasting.

Again, that is a crucial and important issue. I was struck by the evidence from Highlands and Islands Enterprise about the challenges that it faces in understanding what is happening in the economy in its area. At the risk of sounding slightly unhelpful, I want to re-emphasise our national remit. Our task is to produce a forecast of overall GDP and therefore income tax across the board. In one sense, our emphasis is on Scotland-wide information, but we are very alive to and interested in the pattern of economic activity around Scotland, because that greatly impacts on and influences our assessment of what is happening across Scotland as a whole.

An example of that is the work that the commission did a year or so ago on what is happening in the Aberdeen economy—for example the impact on the housing market—as a consequence of changes in the oil and gas sector and the falling price of oil. Regional issues such as oil and gas have an impact on the national picture. We touched on construction earlier, which may have a regional dimension. In that sense, we are very interested in local data for the purposes of developing our overall understanding.

It might be slightly different for other users of statistics, whether it is Highlands and Islands Enterprise or North Ayrshire Council. Others have given perceptive and useful evidence on the need for a greater degree of information, but such information is probably used for purposes that go beyond our activities.

As you say, that information is of interest. Is it available to you at the moment? If not, what could be done to make it available? We talked about whether some of the agencies could provide more access to their information.

There are two things to mention. Inevitably, if information available at a Scottish level is a subset of information at the UK level, and is probably slightly further removed in time than information at the UK level, the same applies to sub-regional data in Scotland. Concerns have been raised about information on gross value added and the more detailed employment estimates. I understand the concerns that are often raised about access to information.

What would potentially be of great interest is if local authorities and enterprise agencies were to subject the administrative data that they hold to analysis and data assessment rather than just publication. I think that Highlands and Islands Enterprise and Scottish Enterprise are willing to do that. That may bring out more information at the local level, which could assist the work of other agencies. It is something that we would be very interested in, particularly in relation to the property market.

As David Wilson touched on, any issues with using business information at a Scottish level are only amplified when we go down to a lower level. The ONS produces data on regional GVA, gross disposable household income and other indicators. That disaggregation is expanding quite a lot, but it is all the same data that is being modelled down to those areas. It does not use new or different data; it just cuts the data in a different way. That is just a note of caution about some of the quite low-level estimates of things such as GVA.

As David Wilson said, we would be particularly interested in data that can help us to evaluate our forecasts. If we forecast LBTT and there is a shortfall, is there a regional dimension that helps us to explain where the economy was driving some of that difference? That is why the data can be particularly useful to help us understand why our forecasts are different from the outturn. It is inevitable that there will be a difference, but the data helps us to explain and analyse the difference. For example, regional LBTT receipts would be of interest to us, if Revenue Scotland were able to produce those in future.

Out of interest, if you looked at very localised data, I take it that there would have to be consistency in the methodology used in capturing that data?

Yes. To be useful at a national level, there would have to be consistency for us to be able to draw conclusions from it.

In your written evidence, you suggest a greater role for HMRC and the Office for National Statistics in helping to improve the quality and coverage of statistics for Scotland. You have spoken quite a bit about HMRC but not as much about the Office for National Statistics. Could you fill in any blanks about HMRC and how you would characterise your relationship with it, and say whether the ONS does things that could be improved or whether it could help you to fill gaps? It will be useful for us to get that on the record.

Our relationship with both organisations is very strong. We are building up various relationships with a number of agencies; one that we have not mentioned much is the Office for Budget Responsibility. A key factor is that, as an organisational development activity, we are developing memoranda of understanding with each of those organisations to make sure that they understand our needs and that we understand what might be seen as the wider challenges that they face as organisations. There are positive developments in relationships and links.

Inevitably, as an organisation with a focus on Scotland that liaises with organisations that have a UK-wide remit, there are things that are of great importance to us. It is not that those organisations see those things as being of lesser importance to them, but they have to take into account a wider perspective. There is constant engagement to further evolve and develop what they may term regional breakdowns in the rest of the UK. That often comes down to a judgment about the perceived need for regional breakdowns and their perceived costs, and a more fine-tuned estimate of whether the sample size of a particular survey or piece of work is robust enough to enable us to use local information, which I look to our professional statisticians to judge.

That is the dialogue and negotiation that we have with both the organisations that Richard Leonard mentioned. Generally, my experience is of positive evolution. More could happen, and we would like greater access to and publication of some of their datasets, but that is part of the evolving process.

There is a large role for the Digital Economy Act 2017 and the new powers that it gives the ONS in onward transmitting information.

I know that in the committee’s previous evidence sessions there has been discussion of access to VAT records. That has the potential to transform data and information about small businesses in Scotland and the UK, which are not in every business survey. There is a fairly large sampling fraction but universal coverage of those one-site small businesses would really improve the information that is available to the Scottish Government so that it can produce its national statistics. There will always be a role for business surveys for large entities and multi-site companies. VAT returns will never give the detail that we need to model them because they create the biggest VAT groups that they can to minimise the number of returns that they need to send in. There is a large potential there.

If the ONS uses that information in its national accounts and its GDP estimates, the 2017 act allows it to onward transmit the Scottish slice of that information to the Scottish Government, so that it can improve the Scottish estimates. We would greatly encourage the ONS to do that with the Scottish Government, and we will benefit from the fruits of their labour with better quality national accounts.

Such administrative data sets can be used to improve the quality of information, but there will always be an element of regionalisation of large companies and other such things because they still report at GB level. There are quite a lot of opportunities for the ONS and the Scottish Government to do things together, which would improve the data that we use. We are encouraging all of that. A lot of it is about the improvement of the national accounts, which we would then utilise.

11:00  

Gillian Martin asked about the possible transfer between people’s self-categorised earned and unearned income. Do you have any broader conversations with HMRC about tax avoidance and tax evasion?

That is part of our discussions. We are not a tax collection agency and in that sense we are not actively engaged in measurements and assessments. We are developing estimates of where tax may be motivating incorporation and wider tax issues and we have to take tax avoidance and evasion into account in the forecasts. However, I would not say that that has been a major issue or that we have discussed or debated it in detail.

Through its production forecast, the OBR considers the measures that HMRC proposes to reduce avoidance and therefore increase tax yield. The OBR’s assessment of the sorts of revenues that such measures will raise has changed a bit over the years. It released a report in the past couple of months that suggests that it may be more conservative about its estimates of the sums that such measures may raise.

It is important for us to be aware of the measures if they are going to impact on the Scottish tax take. We give quite detailed consideration to UK policy measures that may impact on the Scottish tax take. As David Wilson says, we have been working very closely with the OBR—it has been tremendously helpful and has given us advice as a fledgling independent fiscal institution. It has also raised awareness of our role and statutory function with UK Government departments, which has been very helpful in opening doors. We work closely with the OBR and listen to its views on those UK policy measures in order to inform our own forecasts.

One of the other areas that the committee has been interested in is the extent of the independence from the Government of any statistical body. Do you have any views on that? As Mr Wilson mentioned at the very beginning, the Scottish Fiscal Commission is a statutory body and is therefore at arm’s length from the Government. From your experience, do you have a view on the debate about the extent to which there should be a statistical body that is independent of the Government?

Our formal position is that we do not have a corporate view—it is not really for us to take a view on a potential statistics institute or any recommendation about such a body.

There are a couple of points to consider. One should not underestimate the degree of proper processes that underpin the quality of statistics and the independence of their production and release. It is a long-standing issue and there are many regulatory processes that, for me, provide a great deal of reassurance about the independence of national and official statistics. However, I understand that there is a debate to be had about the issue.

The point that my colleague Stuart McIntyre made about the Northern Ireland Statistics and Research Agency is that one needs to look carefully at the regulatory processes to see quite what independence means. The statute is very clear that the Fiscal Commission is a non-ministerial department and is operationally independent of the Scottish Government. That situation is set out very clearly in the statute for good reason. However, there are other parts of Government activities that have different processes for guaranteeing quality and independence. It is for others to judge the role of statistics, but I emphasise that we should not underestimate the proper care and professionalism that already go into the production of statistics.

I have a final question. I understand that you will produce a tax and GDP forecast alongside the Scottish draft budget for 2018-19. Do you know when that will be published?

Our document will be published at the same time as the Scottish budget, on 14 December. The timing is very deliberate as part of the overall fiscal framework arrangements—we publish at the same time, on the same day.

Thank you for the clarification.

Just to be clear, for the five-year forecasts, we finish the current year and then cover the budget year plus the next four years.

Are those revised annually?

They are revised twice a year—at least.

So it is twice a year or perhaps more.

Thank you for attending the committee meeting.

11:06 Meeting continued in private until 12:20.