National Assembly for Wales

A warm welcome to you, Robert, in joining us today. I think we'll go straight into questions, if that's okay, because we've got quite a lot of ground to cover in the hour or so that we have. So, maybe just to start, I'll ask you to give us a brief explanation of your role in forecasting Welsh taxes and how your tax forecasts are used to adjust the Welsh block grant as part of the fiscal framework, as a bit of a scene setter.

Certainly. Well, thank you very much indeed, Chair. Thank you very much indeed for the invitation. You are my last official gig before I step down at the end of the week, so it's a great pleasure to be back with you again. I should therefore make a couple of preparatory points, as, basically, my successor may, of course, disown anything that I say over the course of the next hour. So, that's one thing to worry about. Also, in terms of if we do stray into the current economic and fiscal situation, I am deliberately keeping my hands off the early stages of the process leading up to the November forecast, as I won't be there to take the blame for it when it happens. So, my answers should be couched in those terms.

Turning to your question, one should obviously preface this by saying that the OBR's initial role is to be producing a comprehensive picture of the public finances for the UK as a whole, which necessarily would involve producing a forecast for taxes and spending at every level of government: UK, devolved administrations and Governments, local government, public enterprises, et cetera. So, specifically in this context, we are charged with preparing and publishing independent forecasts for the devolved taxes—so, the Welsh rates of income tax, land transaction tax and landfill disposals tax, and that role we took on formally in April 2019. The contrast there in the other obvious context with Scotland is where that function is undertaken by the Scottish Fiscal Commission. We work very closely with HMRC, Welsh Government officials as well, to develop and run the models that we use to do that in very much the same spirit and approach that we use for forecasting across the whole of the UK, which is that we are a relatively small body and we rely on others to crank the handles of particular models, but very much the forecasts that are produced at the end are ones for which we have ownership, and the methodologies that underpin them, the judgments in particular forecasts, are ours, and we take responsibility for those.

In the Welsh Government context, as with the UK and as with Scotland as well, part of our role in producing a forecast is obviously to look at the costings of particular policy announcements by the Welsh Government and by the other Governments, and again we apply the same principles there that we do in the forecasts and the policy costings as a whole, which is that we only certify and include in our forecasts what we regard as sufficiently firm policies, rather than aspirations, if I might put it that way. As you know, we're publishing a forecast twice yearly, once in the Welsh taxes outlook, and the once other, and the aim, obviously, is to do this alongside the Welsh Government draft in concert, so that it feeds into, appropriately, the Welsh Government budget process.

The irony, of course, in terms of the block grant adjustment, is it's not the forecast we do of Welsh devolved receipts that matters, it's the forecast of the rest of UK, which, in this context, actually is England and Northern Ireland forecasts for the equivalent receipts to those that are devolved, and it's those forecasts that are then used to make decisions about the block grants in accordance with the fiscal framework that's been agreed between the Governments. It's very clear—all we do here is to provide the numbers that we have been asked to provide that allow those decisions to be taken. It's not our responsibility to say whether the framework is sensible or fair or working appropriately, or indeed to judge those adjustments specifically. Our aim is to provide the numbers that the Governments need to make that set of adjustments. So, I hope that serves as an overview.

That's useful, thank you very much. I know you've said you don't want to encroach on your successor's territory, but I'm just wondering: is there anything you could tell us about how the pandemic is going to affect this process? Because, clearly, it gives a huge amount of uncertainty, doesn't it, really.

Indeed. It's throwing an enormous amount of uncertainty into both the economic and fiscal outlook throughout the UK. One point, I think, maybe to start with, which will be pertinent, presumably, to how you're thinking about where the Welsh budget process fits in, is that we've essentially moved, for entirely understandable reasons, given the unfolding nature of the public health crisis, from a business-as-usual model, which is reflected in our legislation and our memorandums of understanding with, for example, the UK Government, in which, essentially speaking, fiscal decisions take place in two discrete events during the year—the budget and the spring or autumn statement, depending on where the dice have fallen in any particular year.

What we've seen, essentially, since March is that we're moving to almost a rolling fiscal policy. So, the Government is making really quantitatively significant spending and revenue decisions almost in real time—and we saw, obviously, the last lot with the winter economic plan—and not accompanied by the forecast and the same relatively lengthy process of policy scrutiny and forecast preparation that we would normally go through. That clearly makes life difficult. It makes it difficult for us to be able to present the public with a fully coherent, internally consistent view of the economy and the public finances that incorporates all policy decisions that we know of, because you don't know when the policy decisions are coming along.

So, the position we're now in is that the Government, as you know, a short while ago said that it no longer intended to hold a budget this year, but it had previously asked the OBR to produce a forecast for the public finances and for the economy in mid to late November. And my colleagues will be continuing to work on that timetable. The Government has not withdrawn that request, or that requirement, and so the aim will be to produce a forecast on that timetable, which, hopefully, would still fit in with, as I understand it, the timetable for the Welsh budget process, in terms of having a new forecast there. 

But I think the other point to bear in mind, of course, as with any forecast at the moment—. I always appear before you and say, 'These things are hugely uncertain; there are many things that can throw this off course.' This is obviously doubly, triply, true at the moment. The economic outlook depends, crucially, on the evolution of the pandemic, the nature of the policy response in terms of restrictions on activity and that feeding through to economic activity. And then, of course, overhanging all of this as well, we have the fact that we don't know yet whether we will be in a 'deal' or 'no deal' Brexit situation at the end of the year. So, the notion of coming up with a well-defined central forecast is unusually challenging in this case, and my successor has already said that, at this time, you really have to put more of the presentational effort into thinking of a variety of scenarios, including different outcomes for the coronavirus process and also for Brexit, rather than saying that we can meaningfully say what a central forecast is at the moment.

The role of the central forecast in the UK context is generally key, because you use it to judge whether the Government has a greater than 50 per cent chance of meeting the fiscal rules that it's set itself. Of course, another difference at the moment is that we don't really have fiscal rules that are—well, we have some in legislation, we have some that were announced on an interim basis in March, but I think most people are expecting that there will need to be a fresh set of those at the next opportunity that the Government thinks it's appropriate to put that. So, the relationship as well between the central forecast and that assessment of the fiscal rules has been broken as well. So, this is clearly generating a considerable degree of uncertainty, and knowing in particular when you're preparing a forecast whether you'll suddenly get, for entirely good reasons, some late-breaking policy developments that aren't possible to integrate easily or properly, is a challenge for everybody. 

And the impact on the block grant adjustment calculations, of course, is huge really, isn't it, potentially, depending on where we are. Could you tell us a bit about how you engage with the UK Government, then, on those discussions, particularly around the block grant adjustment—your tax forecast for the purpose of the block grant adjustment—and whether the Welsh Government actually plays a role in that as well?

Well, we don't engage specifically on the block grant adjustment. As I said, essentially speaking, we have requirements for numbers that need to be produced. Now, obviously, we are producing the forecast for the devolved receipts, so the Welsh rate of income tax forecast— which, again, HMRC will be turning the handles for that—LTT and LDT—Welsh Government, again, us, looking closely at that. Obviously, if these taxes had not been devolved, we would not be producing England and Northern Ireland forecasts for income tax on a WRIT basis, for example. That just wouldn't be material. So, there has been, obviously, engagement in the sense of it being clear to us what we need to provide for you, the Welsh Government and the UK Government to make those sorts of decisions, but we don't engage with them specifically on what they do with those numbers. 

In terms of producing the forecasts—and this was true before we took on the official role in 2019—when we are preparing the forecasts, we, as you know, invite the Welsh Government officials. Normally we do this across the devolved receipts, for example, as whole, so you'd also have Scottish officials involved, looking at what we've come up with in terms of a draft pre-measures forecast for the various receipts. And that's a very useful process of engagement there, but it's on the mechanics of, in particular, are the local government officials—sorry, the Welsh and Scottish Government officials—aware of data, new developments, that we wouldn't be that could usefully be fed into the process. Also, of course, there's often useful—. Something that's happening in Wales may shed unexpected light on what's happening in Scotland, or vice versa. 

Yes. Okay. Thank you so much. Nick.

You need to unmute. 

It's so easy to forget that, isn't it? Afternoon. Do you feel the mechanism for feeding into the block grant adjustment process is appropriate, and how is it reviewed to ensure it continues to be fit for purpose? 

Well, I think, from our point of view, it seems fit for purpose. Obviously, again, the block grant adjustments are a matter for the Welsh Government and for the Treasury. We've not heard from either of those that the process is failing to provide them with the information that they need and the associated explanation or discussion of it that they think that they need for that purpose. So, purely in terms of our inputs producing and publishing those forecasts, I don't think we have any particular concerns. Whether you, obviously, have a broader concern about how well the block adjustment operates is a separate policy question. 

But in terms of forecasting, there are no issues. How will—

No. As you say—. The other half of your question is obviously about how this is reviewed. We had, over the course of the summer, an opportunity for Welsh Government, and for us as well, to look at how things were going: for example, whether the Welsh taxes outlook was providing the sort of information that would be useful more broadly to stakeholders, as well as just to you and the Welsh Government. As I understand it, certainly, the feedback we had was pretty positive there. So, it doesn't look as though there are any particular lacunae in all of that, although, obviously, if there are we'd be very keen to hear about it. 

And as a follow-up to that, how will you monitor—or how will your successor, I should say, monitor—the accuracy of OBR tax forecasts for Wales and the impact of these adjustments on the block grant once reconciled to outturns? 

Well, the Welsh taxes outlook, I think, probably—. In terms of the UK forecast as a whole, as you know, we have a specific forecast evaluation report, which allows us to look back at both the economic and fiscal forecasts, essentially picking time horizons that look to be particularly relevant, and we go through a very thorough process on that and try to draw out of it lessons for modelling approaches, for example. I think in practice the Welsh taxes outlook is likely to be the document in which we would end up explaining where we think there have been successes and difficulties and whether there are possibilities that you might need to change modelling approaches. What we've tried to do, certainly under my watch, is, when we think there are ways that you need to change things, to try to give people a sense of that before you use it first in anger, which I think is the right thing to do. 

In terms of how, again, that feeds through to the adjustments to the block grant adjustment, I think the question there, again, is whether we are providing you with the information that you feel you need in order to assess how well that is working. We had a similar discussion, for example, with the Scottish Government earlier in the process about were we providing the right data on the rest of UK forecast, because, as I say, it's a slight oddity that our most important role in the block grant is nothing to do with what we're forecasting for Wales, and ditto for Scotland—it's for England and Northern Ireland, which seems quite strange. 

So, again, if there are areas where you think that that data could be presented differently or broken down in ways that would be more helpful, then we'd certainly be keen to look at that. Obviously, one of the issues with, as you say, reconciling and looking versus to outturn is the lags that are involved in this, and particularly the lag on income tax. So, a proper assessment of how well the forecast went for 2019-20 income tax is not going to be possible for some while until you have a firm outturn number from that, and we've seen from Scotland how that process of the lag towards final numbers can throw up some surprises here and there. 

So, there are different levels of risk within in-year reconciliations for LTT and LDT. I think that means that the Welsh Government bears more of the risk of short-term falls in revenue on that score; less so for income tax, because the outturn isn't available for longer, but then there's more of a risk when the reconciliations could be larger in later years than they otherwise would be. 

Thanks, and good luck for the future. 

Thank you very much; that's very kind. 

Thank you, Nick. Alun Davies. 

I was just reflecting on the conversation you just had with Nick, and this might be the worst possible time to ever ask you this question, but I'm interested in the structure of public finances and the structure of decision making, decision taking. In an earlier answer, you said we were in an almost real-time fiscal event since March, and I think that's a very good way of putting it. But if I think back to the last few years, one of the things that I find really striking is that the framework only works if you're sitting in the Treasury, in many ways, and you could say that about any aspect of public finances in the UK, I accept that. But because the UK budget and the UK fiscal event, shall we say, are moveable feasts through the year, in many ways the decisions that we need to take here are dependent on sometimes political considerations taken in Westminster, and that has, then, an impact for yourselves and everybody else in the system. To what extent is it your experience—? And you're coming to the end of your role here, so I'm not asking you to make a valedictory remark here, but is it that this structure does not provide the certainty that we require for effective decision making and planning for all of us who operate outside of the Treasury, in many ways? 

I have some sympathy with that view and it's clearly got more difficult in recent years. Obviously it's not for me to give policy advice, but I note that the OECD's independent review of the OBR, which was published last week, which also looked at the way in which we work in the broader context, made a recommendation that it would be helpful to all concerned if the dates of fiscal events were fixed in legislation as the first Monday in X and that's where it happens every year. When I've left this job, you might find me expressing some sympathy with that view. 

In terms of producing the outputs that we do—and the same is true as part of this for the Welsh forecast—it's a long process and it's an iterative process. So, we go through several rounds of the forecast in the weeks running up to the UK budget, and for that to work well and to work properly and to be able to deliver the outcomes that we want to in terms of transparency for the public and for Parliaments, you need plenty of notice of when the forecast is being asked for, and you need a clear timetable and you need everybody involved to observe those timetables, in particular when information is exchanged. So, the Treasury, obviously, in the preparation of the UK budget needs to get a forecast of what we think is going to happen to the public finances if they don't do anything, as a starting basis, in good enough time for them to make some decisions about what the structure will look like—of their package. Similarly, if we're going to end up with a robust forecast at the end of the process, they need to give us decisions about what policies they are going to pursue and exactly what the content of the package is going to be, not at the last minute, but in time to incorporate that.

It has to be said that, even in the difficult periods under my 10 years, we're a way away from the old days pre OBR, in which measures would be inserted in budget documents while the document was at the printers. There has been a greater degree of discipline with that process in. Early on in our 10 years, this process of getting forecasts and policy scrutiny meshed together was helped by the fact that we were newly created, and there was obviously a sense of commitment to this idea of the importance of good process and good timetabling. You also, coincidentally, had a coalition Government that needed to reach agreement on key policy decisions within the coalition, so there was an additional timetabling discipline, rather by accident.

What has changed, of course, in the more recent years is first the ups and downs of the Brexit negotiations. So, it's become much harder for the UK Government to decide well in advance, and to stick to it, when it wants to have a particular budget or fiscal event so that we can set out the timetable for that—most obviously, when the November 2019 budget went the way of all flesh because of the election. So, you've moved from Brexit already—the Brexit timetable, not the substance of it—providing difficulty there, and then now, as we were discussing a moment ago, moving into this period of what is essentially real-time, rolling fiscal policy, with the challenge of having a period with enough clear water to be able to say, 'Well, actually, we can take stock, we can give a proper view of the economy, we can look at all the policies that have been announced up to now, we can ensure they've all had the proper scrutiny and present a coherent picture of that'. I think that's very important.

Now, of course, the counter argument would be to say, 'Forecasting, now more than ever, is a mug's game, so what value would that have?' I still think, nonetheless, that even the process and the transparency and the scrutiny and the clarity we can give to Parliaments and the public is important, even if everybody would admit that the chance of forecasting next year's gross domestic product growth number, or next year's budget deficit, is a factor more difficult than it would be even normally.

I accept these are difficult times at the moment, and it does feel like an impossible dream to be able to consider a structured year at the moment, but I think it's important that we do take the time, even at the moment, in order to look at the sort of structures that would enable everybody to do their jobs effectively in the future. But you've also had 18 months now, I think it is, since the memorandum of understanding was agreed, which determines how you will operate with the revenue authority and with the Welsh Government. Are you happy with the way it's operated? Has it operated in a way you anticipated and expected it to operate? Are there any potential changes that, perhaps, your successor may wish to consider?

I don't think so. I think, from our point of view, obviously, we went into the MOU process having had interaction before and, obviously, a sense of how it was working and how it wasn't. So, it wasn't a blank sheet of paper and a completely experimental approach—it was building on a good set of working relationships already, which has certainly helped that. You should always keep these things under review. We certainly, at the UK level, had some serious thinking about changes that could be made when the first MOU was reviewed four or five years after the event. In particular over the exchange of information, the publication of information et cetera, there were changes that everybody felt were sensible to make. But at the moment, I think that the MOU—. We've certainly not had to use it in anger or apply dispute mechanisms or anything as dramatic as that. It's essentially embodied what was already a good working relationship with Welsh Government, the Welsh Revenue Authority and the OBR. One should always keep these things under review. I'm not conscious of anything that's gone particularly wrong with it, but we're obviously ready to look at it if anybody else is.

Thank you. I have no further questions.

There are two issues here. One would be the devolution of some existing taxes, and then there are, obviously, already the possibilities of new taxes that the Welsh Government might decide to levy that don't exist already. I think we've learnt important lessons from the way that we approached the task when devolution started off in the first place, and that's to try to think whether, when we know that there is an intention to progress devolution further, you can produce indicative forecasts of what that would look like as a starting point, before you get to the point at which the tax actually exists and you have outturn data to go on.

So, I think, as has been the case throughout this process, and in Scotland as well as in Wales, initially you inevitably start off with indicative forecasts that are based largely on a share of the relevant UK tax, and obviously there's been plenty of discussion over time about whether there are particular factors that might lead you to expect that the Welsh share of overall income tax or the Scottish share of overall income tax would move up or down, for example because there may be policy measures that affect higher or lower earners more and because there's a different distribution of higher and lower earners across the nations. Those things would differ as well. So, there's a good opportunity to get to grips with those sorts of issues at a relatively broad level while you have that share-based approach.

Once you have the tax under way, then obviously over time you get outturn data that you are able to draw upon, but both in terms of the indicative share-based approach and with that more bottom-up approach, from tax to tax, the lags before you get firm outturn data (a) differ and (b) can be quite long, and that's always going to cause you difficulty in any transitional period. If there was a decision to devolve further powers, the first starting point is that we would get the Treasury to submit us a costing note, as we would get them to submit a costing note on any other tax or welfare policy change, and I would expect that HMRC would prepare that—depending on what sort of tax it was—and that would probably also start off with a share of UK bases.

If you have a new tax altogether that comes in, then as with the same thing happening in the UK context, how confident you are of producing a forecast of that depends an awful lot on how much you know about the tax base—i.e. is there good data, is there a good run of data for the thing it is that you have decided to tax for the first time. You need that as a starting point, because otherwise if you're going to tax it by 10 per cent and you don't know whether it's 10 per cent of £10 million or £5 million or whatever, then you're in a difficult situation already.

And the other point, of course, with both new taxes—well, it happens inevitably with new taxes, but also with the devolution of existing ones—it's obviously the case that the greater the Welsh Government were to decide that it wanted, in the case of a newly devolved tax, to have very different rates from the ones that the UK Government has, the more likely it is that you're going to need to worry about the behavioural changes that's going to cause for the individuals or the businesses that are paying the tax. If you were devolving something and you were going to double the tax rate, then there's a whole question about how are people going to react to that, will they change their genuine economic behaviour, will they decide to pretend that they live in England or Wales depending on which way the advantage goes, is there an opportunity to shift income or spending and the way in which it is viewed as well. So, that's the other issue. If you devolve and you want to do something very different, it's a much more complicated task than if you devolve and you actually don't want to do anything terribly different but you just want the power over it.

For new taxes, I think the process would be robust. One thing that we try to do is to be clear—and obviously this has been true for a much longer period with HMRC and with the Department for Work and Pensions—when a new tax or a new policy comes along, we try to help the officials there with a list of—basically, 'This is the model note we need you to give us in order to reach a robust judgment on this.' So, here are the five questions that, if you're proposing something new, we'd want you to be answering: what is it exactly that you're going to tax? How well defined is this? Do we have good background data on this? Is there any evidence on the behavioural impact of similar taxes, because maybe Scotland's done one, or maybe there's an overseas example that gives you an impact on this? Is there any academic evidence on exactly how responsive behaviour is in those particular areas? Is this something that's likely to generate avoidance opportunities or evasion? What are the implementation challenges? You say you want to do this; if you do it next year, have you got a cat in hell's chance of collecting all the money that you think you ought to be collecting in a steady state in year 1 or year 2? So, I think by giving the Governments—and, as I say, as we do with HMRC or DWP—a clear sense of what we need from them in order to be able to stand up, or sit down, and come before you subsequently and say, 'Well, we've looked at this; it's very uncertain'—we always say that—'but this is the basis upon which we have taken this decision.'

The other thing that we do do, obviously, ex post at the UK context is, when a new policy measure comes along, we give it a what we call an 'uncertainty rating'. So, we look at the new measure on a variety of criteria: the robustness of the data; do you expect big or small behavioural effects; how easy is the modelling; are you just basically child benefit: multiply a number of children by a number of pounds; or is it a change in insurance CT treatment, in which case there might be seven different judgments you have to make in order to get there. And so that again—we can provide people with a sense of, you know, 'This may be our best guess, but this is something you will be particularly worried about, their being bigger as in one direction or another.'

Well, the latest my colleagues gave me: the latest information on the 2019-20 was actual cost, £99,727. So, I'm not quite sure who gets the remaining £273. Maybe we could just—particularly as two members of the committee aren't here, we could share it and I'd get a decent drink out of that, but it's clearly pretty close. As it's turned out, we've ended up—and this is not unusual in budget planning, spending—staff costs have been somewhat higher, non-staff costs have been somewhat lower, but it's clearly come down to broadly the same level. I don't know—in terms of what would happen for the following year, whether the £100,000 should be uprated in line with inflation or public-sector pay increases, I'm not entirely clear what the level of agreement is on that. I'm not sure COVID specifically would necessarily affect the cost a great deal; although, again, there, my successor may be quietly shrieking in a room somewhere that I've said that. It's not obvious to me why that would be the case, but at the moment, it's come in broadly as one would expect. Clearly, you would expect—the bulk, 77 per cent, 78 per cent, of the £100,000 is staff costs and if staff costs are rising more, then that's obviously going to put pressure on it.

Okay. On we go, then, to Mike Hedges.

Thank you. You previously announced the OBR would increase its presence in Wales. Can you tell us what progress you've made on that?

Well, there have been a variety of elements to that. As you know, we offered a technical briefing to Members some while back. My colleague Andy King presented at the Welsh Government tax conference, and I think there is a Welsh Government tax conference in virtual form, is there, in a week or two's time, and I think he's appearing at that again. We found that the Wales Governance Centre's conference in July 2019 was a good opportunity—we had a couple of people presenting at that, which was a good opportunity to meet a wider range of stakeholders. There are obviously appearances before you; I guess this must be my third since we took on the role of the official forecaster. We've had, obviously, sessions discussing and meeting with Welsh Treasury officials and a very interesting talk from the chief economist some while back. And we tried to hold a sort of summer challenge meeting outside of the forecast, so an opportunity to look at the forecast judgments and the process and the methodologies a little bit away from the heat of battle.

I guess the other point to make is, by coincidence, our head of devolved work is based in Cardiff. That's not part of the agreement, but it just so happens that she was working at Newport for the Office for National Statistics and is based there. So, we even have a physically manifested presence. But the lesson I would take from all of that is that you will know better than we will, and the Welsh Government will know better than we will what are the most efficient meetings and opportunities to get the right stakeholders together to input on that, as with the Wales Governance Centre conference et cetera. So, I'm sure that my successor will be very keen to know whether there are gatherings of people, virtually or physically, that you think would be particularly helpful ways of increasing that presence further.

Okay, thank you. On the income tax, you are highly dependent on HMRC providing you with information. How accurate do you think that is, and is there more information you need from them than you're actually getting?

There's no specific information that we've asked for that we need. Clearly, as we've discussed before, one of the major challenges on the income tax side is the lag in getting the outturn data. So, I think the outturn data for the Welsh rates of income tax receipts for 2019-20 we would expect next summer; in the absence of that, you draw on the survey of personal incomes, which is an annual survey of 750,000 people who are in contact with HMRC through the tax system—that's also published with a lag. I guess the change, the most important change, again at the UK level, that we've seen in the last couple of years is the ability to make greater use of real, or so-called real-time information—RTI—and that gives us a bit of flavour for the latest movements in earnings, which can help give you some sense of divergence between Wales and the UK as a whole. So, the data I think is what we need for the models that we're using. The publication of the outturn data next summer is going to be very useful and important, but, again, there is a lag there. But, of course, taking a step back, the uncertainties about what is going on with income tax over the next—well, now and over the next year are hugely uncertain at the UK level because of the health crisis. I suspect between—[Inaudible.]—'Do you know materially what's going on?' It's different in Wales than the rest of the country.

The final question I've got on this is that people will move between tax bands—they'll be on the basic and then they'll be on the high and on the very high tax. That can have an effect, can't it? Have you got any way in your model that would affect people moving between those different tax bands?

You're right. Obviously, in the Welsh context, there's a greater focus on what's going on in the individual bands than there is in the equivalent adjustment process in Scotland, for example. You take me beyond my technical knowledge here, but I think, in terms of the modelling process, you think about the way in which the distribution of incomes is likely to move over time and therefore change the number of people in those tax bands. There's clearly the possibility that changes in the tax rates themselves could encourage more of that. And, obviously, if you end up in the situation where the tax rates differ between the different bands in Wales, from a different way in which they differ in the rest of the UK, that too could affect behaviour. The key issue here, obviously, given the relative income distributions, is changes to the personal allowance that are affecting the basic rate are going to be more important in Wales than changes that are affecting the higher end, and more people, essentially, in London with relatively high incomes as well. So, you will see movements between that. Whether one can do anything very much more sophisticated than look at the existing pattern of incomes and assume that people migrate through that and make some adjustments for changes in the relative tax rates—the science probably can't get much more precise than that. 

Thank you.

Thank you, Mike. Just a couple more questions from me, then: HMRC acknowledged a while ago, of course, that some Welsh taxpayers were paying Scottish rates of income tax, if you recall. I'm just wondering, really—that reminds us of how misidentification of taxpayers can impact on your forecasts—if maybe you can tell us a little bit about what that can mean to your work and obviously what it ultimately would mean to the calculation of the block grant adjustment. 

In that specific case, as I understand it, the error took place and HMRC managed to correct it before we produced our first forecast for WRIT. So, given that one of the data sources that we used to produce the forecast is the RTI, the real-time information earnings data, given that that error has been rectified, you wouldn't expect that to have an impact on the forecast that we produced this year. Now, clearly, if the problem or a similar problem surfaces somewhere else, it's a question of how quick it would be to do that.

What I'm not clear about—this may be known, but I just don't know it—is how quantitatively significant that would have been in that instance. My guess would be that it's unlikely to be hugely quantitatively significant against the overall number, but I'm saying that on the basis of no concrete knowledge.

Yes, that's correct, but there's always that risk, isn't there, that taxpayers are misidentified? And I suppose you're going to say that you can only work with what you're given.

Exactly. We have to work with the codings people are given, and there's always, as we've discussed, I think, many times when I've appeared with you before, this whole question about how you decide how people are flagged—whether that corresponds to the survey data and whether you think that a Welsh or a Scottish taxpayer in the survey of personal incomes ends up being the way they're flagged, even if they're flagged correctly—is a separate source of difference, potentially.

Sure, and we know that Welsh rates of income tax will represent around £2 billion, I think it is, isn't it, of the Welsh Government budget, around 12 per cent of the budget? What's the likelihood of those revenues being overestimated in the future do you think, and what impact that would then have on the block grant adjustment and the Welsh budget?

Well, the conventional answer I would give you is that, when we produce the forecast, the chance of any forecast being over or undershot is 50 per cent by definition because it's a central medium forecast. Obviously, we can't yet evaluate the 2019-20 WRIT forecast because we don't have the outturn data from HMRC yet, but that should be doable in the Welsh taxes outlook next year.

If you think about the sort of errors at the UK level for the overall income tax forecast for 2018-19, if I just take an example here, the error was 5 per cent in the March 2017 forecast, and then went down to 4 per cent, 2 per cent, 1 per cent as progressive forecasts came closer. There's no necessary guarantee that you would see similar numbers there.

The challenge, obviously, looking forward, is that forecasting income tax receipts at the moment, given the disruption to the labour market, uncertainties around unemployment, what's happened through the furloughing scheme et cetera, is extremely difficult.

If you look at the way in which the numbers are coming in at the moment relative to the central scenario we put out in our fiscal sustainability report back in July, both the revenue numbers and, presumably not coincidentally, the economy have been performing better than was anticipated there. There has been a sharper recovery in the economy and receipts have performed better, partly maybe because of that, partly also because the furloughing scheme, for example, was more lower paid—part-time workers were affected by that, so it was less of a hit to receipts than we had originally anticipated some while back.

Obviously, the uncertainty we have looking forward now is, although you had good news over the year to date, if we are having a second wave of some description, certainly a second round of public health restrictions, how much impact will that have? How quickly will that show up? And I think if you look out over the longer term, the fact that we've had a sharper V initially and the possibility of an unexpected retrenchment or a retrenchment that wasn't assumed in the next few months, we're really none the wiser as to what the long-term scarring effect of this crisis on the potential of the economy and therefore the potential tax receipts across a whole range of taxes is going to look like in the medium term. And that's going to be something, as with the financial crisis—hopefully not as bad as the financial crisis—that we're only going to learn and be able to grope at in terms of the outturn numbers over the next few years.

So, whether the economy ends up being the same size or 3 per cent or 6 per cent smaller in five years' time from what we anticipated it would be just before the crisis hit, who knows? It will be difficult enough judging that just with coronavirus and then you also have where we end up on Brexit.

Plenty in the in-tray for your successor there, as you articulated. Okay. Any further questions from Members? No. There we are.

Well, can I thank you, Robert, again, for joining us? As you mentioned at the start, this is your final appearance before this committee, so we're very grateful for the invaluable contributions that you've made when you have appeared before us; it has really enriched the work of this committee and we appreciate it very much. We wish you very, very well for the future. Diolch yn fawr iawn.

Thank you very much indeed. I have very much enjoyed my appearances. It's been very engaging, both literally and a pleasure to be in touch. So, I hope we will remain in touch in some way.

I'm sure we will and thank you for that. So, Members, we will now take a short 10-minute break and we'll reconvene for 15:35. Diolch yn fawr. Thank you.

Thank you, Chair. My name's Richard Harries, I'm an audit director with Audit Wales and I'm responsible for the financial audit of the Welsh Revenue Authority.

I'm Mark Jeffs, I'm an audit manager with Audit Wales.

Certainly. Thank you, Chair. Well, if I just give a little background to start. Clearly, as auditor general, I'm responsible for the statutory audit of the financial statements of the WRA and also have the power to undertake value-for-money studies into their economy, effectiveness and efficiency. In respect of wider Welsh tax matters, you'll be aware that the administration of the WRIT—the Welsh rate of income tax—that's a matter administered and collected by HM Revenue and Customs. From an audit perspective, that falls to the remit of the UK Comptroller and Auditor General, and I believe the C&AG is giving evidence separately to you as part of your inquiry.

There we are. Diolch yn fawr. So, could you summarise, then, just to start, how you felt the transition to devolved taxes has been administered and what you think maybe the strengths and the weaknesses of the approach have been to implementing those Welsh taxes?

Certainly. Well, I remember I appeared with Richard before you a couple of years ago, I think. We presented the third of what had been three reports that we'd undertaken into the development and implementation of the Act. In summary, all three of those reports were very positive, so our conclusion was that the WRA had operated effectively in terms of administration of the devolved taxes and that, also, the Welsh Treasury had appropriate arrangements in place to gain assurance over HMRC's implementation of the WRIT.

From our perspective, it was a very positive period and process, so we were engaged very constructively and fully by WRA and the Welsh Government throughout the process. Two of our staff—. One of my senior directors sat as an independent observer on one of the key oversight committees establishing the WRA, and one of our IT managers sat in a similar specific IT and digital group. So, a very positive process overall from our perspective.

I think in terms of the strengths of that process, as I said, we observed very effective continuous engagement with a range of stakeholders, ourselves included, and also yourselves as a committee, I would say, and a culture in the WRA and Welsh Treasury that appeared to us open to challenge and open to learning from others, most notably, for instance, from Scotland. The quality of individuals that were attracted to those roles was high, so at the leadership level and throughout the organisation, and from our perspective, we observed strong project management and governance arrangements overall. So, it's not often that I appear before committee being quite as fulsome as this, but from our perspective, a very positive process.

There we are. Okay, thank you very much for that. We'll move on then to Nick.

Thanks, Chair. Good afternoon, I'm seeing you in a different setting. Can you explain what value-for-money exercises you've undertaken regarding the Welsh Revenue Authority and what's been the outcome?

As I said at the top of the meeting, I have the powers to undertake value-for-money studies in respect of WRA. To date, the three pieces of work that I outlined are the only areas where we've exercised those powers. So, I am minded, I think, to do some more VFM work in this area, but only once the WRA has had time to bed in and settle down. So, they had quite a lot of Audit Wales engagement and time in those early years. I think it's right that I step back and give them a little space to get up to speed and really settle down as an organisation, and I'll be very interested to see the conclusions of the committee's inquiry, as that may shed some light on areas where it would be most valuable for me to focus our attention in maybe a couple of years.

That said—Rich may want to speak to this in more detail—we do have ongoing engagement with the organisation, because we are there to audit their accounts each year, and that process gives us some insight into any organisation. I think it's fair to say that that positive impression that I outlined from those three early studies has been reinforced by our exposure to the organisation through the annual accounts process. So, we see an organisation that, generally speaking, is well managed and is delivering against its targets.

That was going to be my next question. It's good to hear that you've had a positive picture from them, auditor general. Are there any areas of weakness that, even in these early days of the organisation, you would flag up with them and say are areas that they can improve in?

Not so much areas of weakness—I think in our previous studies we highlighted a few areas that they needed to keep an eye on. I said at the start that they'd been successful in attracting very capable staff, and I suspect part of that was because it was a very exciting project to be involved in. Retaining those staff, building up permanent staffing capability, was certainly an area I think they need to keep an eye on, but, so far as I'm aware, that's not manifested itself as an area of further concern. Rich, is there anything further from your perspective?

Thanks, Adrian. Probably just to echo that, Adrian. I think the ongoing work we've done in the last couple of years on the accounts has been very positive. They've continued to engage in a very constructive way with us, and I think they're still trying to look at how they develop their own value-for-money conclusions in terms of the way they're presenting to the wider public and how they're demonstrating that. I'm conscious at the moment that I think they've written to the committee; they're finalising their annual performance report at the moment for October. Again, that's going to set out the targets that they set themselves and how they're performing against those targets. But I think a really good example this year is, I suppose, the issues they've faced, like other bodies, in terms of the floods. They had to evacuate the offices earlier in the year when their Treforest office flooded, and then with COVID hitting them as well, and they've adapted excellently to that. We had to work virtually with them throughout the audit process, but they've managed to do that, given the IT systems they've adopted from day one and just the way they want to engage and learn and develop some of the things that they do going forward—so, again, a positive message from the annual accounts process as well.

Diolch.

Diolch yn fawr. Okay. That's very heartening. We move on to Alun Davies, then.

This could be a very short session, couldn't it?

That's fine. [Laughter.]

You described the report as being quite fulsome earlier, and certainly the written evidence that you've provided and the response that we're having in this discussion this afternoon—. And that's very good news—it's very good news to see, and it's good to see the way that the Government in its wider sense has been able to accrue new powers, create structures to operate those new powers, and then for that to bed down in such a way. It's not always been the case, I think it's fair to say. So, that's certainly good news.

In terms of the joint governance between the WRA and Welsh Treasury, do you see that working well together? Do you think the joint governance arrangements are in place and working well, and do you see the culture, if you like, of a necessary tension between different parts of the structures working well together? Do you think the culture is in place that there needs to be to deliver excellence in the future and to embed excellence in going forward?

I'm going to sound like a stuck record, because my answer will be 'yes', I believe we do see those things—an appropriate understanding of the different, more arm's-length arrangement that is important for WRA, with the Welsh Treasury being, clearly, a central function of Welsh Government itself. Richard will have more direct, hands-on experience of having observed some of those joint governance arrangements in place, but my observation would be that it starts from fundamentally a very positive place. But, Richard, perhaps you could elaborate on that.

Yes, thanks, Adrian. Again, a very positive message, which is pleasing to give, really, I think. When we did the work in 2018, we looked at how those arrangements were being set up, and I think the culture was very strong between the two separate organisations, in a way, to work together. I think there was a joint partnership group set up between the Treasury, which the Permanent Secretary and the director of the Treasury were on, and the chair and chief exec of the WRA, and that allowed a lot of the ongoing discussions to work through some of the challenges they had. It was also a constructive process. Those reported openly and transparently, and, as far as I'm aware—so, from the ongoing discussions that we have with the Welsh Government—those things are carrying on. But I think it's a very positive process.

I think it was one of the areas that we flagged probably in our last report to keep in view, keep developing. We haven't gone back and formally assessed it, but, from what we're picking up, there are no concerns, Adrian.

In that case, I'm very happy, and I'll leave it at that.

Rich, can you do that, please?

Yes, I can, yes. When we did the report in 2018, I think the corporate plan was still being developed, and I think that, in the report that we produced then, we flagged up seven or eight areas for them to consider in making sure that plan identified where they were trying to go as an organisation and set some of those performance targets that would allow them to be held accountable and scrutinised. I think the corporate plan was published, then, in May 2019 as a document that came through Finance Committee, and they're in the process now of producing their annual performance plan, setting out the performance against those targets. So, again, it probably wasn't critical of the corporate plan process, it was just the timing of when the work was being done and where they were in formulating that plan. We've seen the corporate plan, we've commented on the annual accounts and what they are doing—again, in a positive manner—and I think it's good when the performance report will now be published so you can see exactly how they perform as an organisation as well, then. 

My understanding is that, when the organisation was first created, quite a large proportion of staff was seconded in or brought in on a temporary basis, so there was a risk at that time that there would be an exodus of those staff and, as a relatively small organisation, it would be quite vulnerable to that. So far as we see, that is not a risk that has manifested itself in any way that causes us concern. But, Rich, anything more specific from your perspective?

Only to say I think one of the challenges that we flagged up in the last report was making sure the training and development of those staff was a vital part of the process, because I think retaining them could have been an issue, given the nature of the organisation and, I suppose, the newness and the culture and the changes they're adapting to, and the evaluation of what that training was actually trying to achieve and whether it was successful or not. And, again, from what we've seen in terms of how they're reporting the accounts and the annual report, they've put a positive message in that, and we haven't seen a deterioration in the calibre of staff that we've been dealing with. Where people have left, they've been quickly changed and quickly recruited to the posts that were left behind. So, again, I think it was that retaining them was going to be a challenge, given the nature, but training and development around that was hopefully something that could address those challenges and, from what we've seen, I think that is the case.   

It's not a routine part of our audit of accounts. We will pick up on issues like that just through the general observation of the governance of any organisation. I mentioned at the start of the meeting that it is my intention, within the next two or three years, to do a more focused piece of value-for-money work on the WRA, once it's up and running and fully established, and maintenance and training and development of its workforce might well be an element of that piece of work. But it's not something that we're routinely monitoring at this stage, no.

Yes. So, I have powers to match data for some parts of the public sector and some data sets under something called the NFI, the national fraud initiative. I mentioned in my written submission to you that, for some years now, myself and my predecessor have been calling for those powers to be updated and strengthened. This is one area where I think there could be value from doing exactly that. So, my powers in respect of data matching in respect of the taxation, the fiscal system and the WRA itself are far more limited than they are in local government or the NHS. Where this would be of particular value would be if we could match large data sets held by the WRA and other parts of the public sector to not only look for examples of fraudulent activity, but also irrecoverable debts and other aspects of activity that would be of direct relevance to improving value for money for the WRA. 

So, it's not specific to the WRA—this is a series of powers where there would be value for the whole of the public sector if my data-matching powers were enhanced. But if the committee felt minded to add its voice the calls that I've made and my predecessor have made, that would be very well received. 

Yes, I'm afraid there would need to be legislation, yes. 

Thank you, Chair. You previously mentioned that the Welsh Treasury and HMRC would need to develop a governance structure for Welsh rates of income tax before it began to be collected in April 2019. Can you provide the committee with an update of how these arrangements were implemented?  

Rich. 

Yes, I think, again, at the time of writing the 2018 report, the tax was going to be collected from 1 April 2019, so we hadn't had sight of exactly all those governance arrangements. I think what we did see was the fact that there was a tax devolution programme board, which was established by HMRC, which Welsh Government played an active part in. So, they were allowed to, I suppose, challenge some of the thinking, work together and develop that framework. And I think the work that the National Audit Office have done since—I think that report came to the Finance Committee in January this year—setting out the assessment of those arrangements, particularly in terms of the, I suppose, identification of taxpayers, the information that's being used, that was put in place and the assurances that we saw at the time of writing the report from the Welsh Government allayed some of those fears as such that it would go wrong like it did in Scotland. But it didn't happen in Wales, given some of the developments that they put in place. So, I think, again, at the time of writing the report, those arrangements were being developed. I think since, when the NAO have looked at the HMRC arrangements, they've come back with positive conclusions to the Welsh Treasury. 

Thank you. Are you able to comment on the progress the Welsh Treasury has made in acquiring more up-to-date income tax data? 

Not from our own perspective, in a way. We haven't gone back and looked at that but, again, the work that the Comptroller and Auditor General does with the NAO—they look at that data regularly. I think the comment we had in the last report was that they were using the most up-to-date data, but that was going back to 2015-16. So, there's a bit of a time lag in that work coming through, and I think that's something HMRC and Welsh Treasury have been looking at and working on across the UK, not just for Wales, to make sure the most accurate data is being used for the forecast and the budgeting. 

Okay. Thank you, Mike. Your report noted that correct identification of Welsh taxpayers, of course, was one of the Welsh Government's priorities for the successful implementation of WRIT. Do you feel that appropriate assurances have been obtained from HMRC, given that there were, of course, issues that we know of now, where some Welsh taxpayers ended up paying Scottish rates of income tax? 

I think, again, coming back to what we saw in terms of doing the work, those assurances were coming through and they were being triangulated within Welsh Government in terms of some of the stats teams there to make sure the right information was being used. Some things got through the net, as you say there, but I think, looking at what the NAO reported last time, that's an ongoing process. It's not at a point in time; things will change. I think the assurances are coming through the system, but I think it's got to keep evolving and keep being challenged by all parties to make sure that things like that don't materialise again. 

Okay. That's fair enough. Finally from me, you said in your consultation response that the committee may wish to explore how the new needs-based formula will work if we return to a period of austerity in future. I don't know whether you want to elaborate on that. 

There's a good reason why Mark is here, because he's the man who is able to elaborate more eloquently than I can on that. 

Okay, thank you. Mark. 

Yes, basically—. I'll take a deep breath before getting into trying to explain the Barnett formula and the needs-based factor. But, basically, it goes back to—. What the Barnett formula does is it deals with year-on-year changes—and the changes side of it is important—to the block grant. So, in normal times what changes means is how much the block grant goes up by each year, but it also determines how much the block grant goes down by when there are times of austerity. So, what the needs-based factor does is it adds a little bit of multiplication, basically, onto the end of the Barnett formula. It's worth just briefly recapping that the needs-based element came about as a response to the issues of convergence, where funding per head in Wales was slowly converging with levels in England, and that was identified by the Holtham commission about a decade ago now. 

What the needs-based factor is supposed to do is to permanently ensure that Wales always gets around 115 per cent of the funding compared to England on an equivalent set of devolved services. It does that by multiplying that change that I mentioned at the beginning. For example, it's currently set at around about 105 per cent—that multiplier at the end—but, as spending starts to converge again, it'll move to 115 per cent. That's a lot of numbers but, in practice, what that means is that if there's, for example, a £100 million increase through the Barnett formula, that will go through the multiplication and, at the moment, while the factor's set at 105 per cent, it'll become a £105 million increase. In future, at some point, it will become a £115 million increase, which is great for Wales when the funding is going up.

But, as I said, it also applies to reductions in times of austerity. So you would apply that same multiplication number to a cut. So, a £100 million cut becomes a £105 million cut. In other words, the needs factor means that cuts would be bigger in Wales in times of austerity than they would otherwise be if the needs factor wasn't there. Now, there's a temporary protection against that set out in the fiscal framework, and what it says is that from 2018-19 and for the remainder of the current spending review period, all uplifts will attract the 10 per cent factor and any reductions below the 2017-18 level will be applied without the additional factor. So, it's spelt out that, for the current spending review period, cuts won't have that additional multiplication applied to them. But we've had two general elections since then. For the past week, there's been a lot of speculation as to when the next spending review will happen. 

Given the current state of public finances and the possibility of a fiscal squeeze at some point, we just thought it was something that the committee may want to get clarity on, in terms of the Welsh Government's current understanding of that protection and whether or not it expects it to apply to future spending review periods. 

Okay. Fine. That's certainly something we'll need to pursue with Welsh Government and indeed the UK Government as well. Thank you for that. Any further questions from Members? No. Adrian, you want to make a point before we finish.

If I could just take you back—. I'm conscious that in answering Siân's question I wasn't particularly clear in giving you a concrete example of how improved data matching powers might be exercised. If I can leave you with just one concrete example: for instance, we could look at the WRA's effectiveness in ensuring the accuracy of their LTT assessment for property developer transactions by matching WRA information with local authority information on things like planning permission, council tax, non-domestic rate status and so on. So, matching those big data sets would be an extremely valuable exercise for us to undertake.