Y Pwyllgor Cyllid - Y Bumed Senedd

Finance Committee - Fifth Senedd

17/10/2018

Aelodau'r Pwyllgor a oedd yn bresennol

Committee Members in Attendance

Jane Hutt
Llyr Gruffydd Yn dirprwyo ar ran Steffan Lewis a Chadeirydd y Pwyllgor dros dro
Substitute for Steffan Lewis and temporary Committee Chair
Mike Hedges
Nick Ramsay

Y rhai eraill a oedd yn bresennol

Others in Attendance

Dr Edward Jones Prifysgol Bangor
Bangor University
Dr Helen Rogers Prifysgol Bangor
Bangor University
Helen Mary Jones Aelod Cynulliad, Plaid Cymru
Assembly Member, Plaid Cymru

Swyddogion y Senedd a oedd yn bresennol

Senedd Officials in Attendance

Bethan Davies Clerc
Clerk
Christian Tipples Ymchwilydd
Researcher
Dr Ed Poole Cynghorwr Arbenigol
Expert Adviser
Georgina Owen Dirprwy Glerc
Deputy Clerk

Cofnodir y trafodion yn yr iaith y llefarwyd hwy ynddi yn y pwyllgor. Yn ogystal, cynhwysir trawsgrifiad o’r cyfieithu ar y pryd. Lle mae cyfranwyr wedi darparu cywiriadau i’w tystiolaeth, nodir y rheini yn y trawsgrifiad.

The proceedings are reported in the language in which they were spoken in the committee. In addition, a transcription of the simultaneous interpretation is included. Where contributors have supplied corrections to their evidence, these are noted in the transcript.

Dechreuodd y cyfarfod am 09:02.

The meeting began at 09:02.

1. Cyflwyniad, ymddiheuriadau, dirprwyon a datgan buddiannau
1. Introductions, apologies, substitutions and declarations of interest

Bore da. Croeso i chi i gyfarfod Pwyllgor Cyllid y Cynulliad Cenedlaethol y bore yma. A gaf i nodi bod clustffonau ar gael ar gyfer offer cyfieithu? Gallwch chi gael y cyfieithiad ar sianel 1, a gallwch chi amrywio lefel y sain ar sianel 0. A gaf fi hefyd atgoffa Aelodau i ddiffodd y sain ar unrhyw ddyfeisiadau electronig sydd gennych chi? A gaf i ofyn, hefyd, i Aelodau a oes ganddyn nhw unrhyw fuddiannau i'w datgan? Na. Diolch yn fawr.

Rŷm ni wedi cael ymddiheuriadau oddi wrth Steffan Lewis, David Rees a Neil Hamilton, ac mae Helen Mary Jones yn ymuno â ni bore yma ar ran Plaid Cymru.

Good morning. Welcome to this meeting of the Finance Committee at the National Assembly this morning. Could I note that headsets are available for interpretation? You can have the interpretation on channel 1 and you can have amplification on channel 0. Could I also remind Members to turn off any electronic devices, or put them on mute? And can I ask Members whether they have any interests to declare? No. Thank you very much.

We've had apologies from Steffan Lewis, David Rees and Neil Hamilton, and Helen Mary Jones is joining us this morning on behalf of Plaid Cymru.

2. Papurau i'w nodi
2. Papers to note

Mi symudwn ni, felly, at yr ail eitem, sef papurau i'w nodi. A gaf i wahodd Aelodau i gytuno ar gofnodion y cyfarfod a gynhaliwyd ar 11 Hydref? Pawb yn hapus â hynny? Ydych. Diolch yn fawr.

We'll move on, therefore, to the second item, which is the papers to note. Could I invite Members to agree on the minutes of the meeting held on 11 October? Is everyone content with that? Yes. Thank you very much.

3. Cyllideb Ddrafft Llywodraeth Cymru 2019-20: Sesiwn dystiolaeth 3
3. Welsh Government Draft Budget 2019-20: Evidence Session 3

Yr eitem nesaf, felly, yw'r sesiwn dystiolaeth y bore yma ar gyllideb ddrafft Llywodraeth Cymru, ac a gaf i groesawu Dr Edward Jones a Dr Helen Rogers o Brifysgol Bangor? Croeso i chi atom ni. Mi gychwynnwn ni'n syth gyda'r cwestiynau, ac mi gychwynnaf i, os caf i, drwy ofyn i chi efallai a allech chi esbonio eich rôl chi yn craffu yn annibynnol ar ragolygon treth Llywodraeth Cymru.

The next item, therefore, is the evidence session this morning on the Welsh Government's draft budget, and could I welcome Dr Edward Jones and Dr Helen Rogers from Bangor University? Welcome to both of you. We will start immediately with the questions, and I will start, if I may, by asking you to explain your role in independently scrutinising the Welsh Government's tax forecasts.

Diolch yn fawr iawn. So, the Welsh Government, under the terms of the fiscal framework, are committed to put in place arrangements for the independent scrutiny of its devolved tax revenue forecast. We entered into a competitive procurement exercise in March 2017 and were successful for that year. So, we completed the work for the previous year, and we've now been asked to do the work again for this year, which we've completed. So, Bangor Business School was appointed to undertake this work, and the aim of the work was the provision of independent scrutiny and assurance of the Welsh Government's forecasts for Welsh taxes, so the land transaction tax, landfill disposals tax, non-domestic rates and Welsh rates of income tax, for inclusion in the 2019-20 budget and potential future years, but also we had the role of providing advice on improving methodologies for future years' forecasts.

09:05

A ydy'r rôl wedi newid o'r llynedd i eleni?

Has the role changed from last year to this year?

No, the role itself hasn't changed. The only difference between this year and the last is the inclusion of the Welsh rates of income tax, so everything else is the same.

Sut, wedyn, ydych chi'n bwydo'r gwaith rydych chi'n ei wneud i mewn i'r Llywodraeth?

How, then, do you feed the work that you do into the Government?

The process itself is an iterative process, in that we have meetings with the Welsh Government. They will explain to us the work that they're doing on their models and so forth. We'll then analyse their work, feed back to the Welsh Government, they'll take the advice on board, they'll give us another presentation and so forth, so it's that iterative process that we have there.

What we've also done is had meetings with the Office for Budget Responsibility and the Scottish Fiscal Commission, both of which have been very helpful in providing support. So, we've been able to see where the Welsh Government is, in terms of their performance relative to the Scottish Fiscal Commission and the OBR.

I think one of the things about modelling is that it's very easy to model the past. It's the future that becomes much more difficult. Have you made any significant changes to your forecast model since the 2018-19 forecasts were made?

Yes, I agree with your point in terms of modelling. What we have benefited from with this year is the in-year data from the Welsh Revenue Authority. So, what we've been able to do, following a lengthy discussion, is integrate that data from the WRA into our forecast. Therefore, the models themselves—they're not based purely on historical data, but they do take into account any information we know at this point in time.

And have you had discussions with the OBR regarding your modelling as opposed to theirs?

Yes, we've had a meeting with the OBR. The modelling philosophy would be strongly aligned to the OBR. The modelling philosophy is well aligned to the Scottish Fiscal Commission. There is a slight difference in terms of the LTT model that relates to the distributions that they use. However, having said that, the Scottish Fiscal Commission do recognise that they're having difficulties because they're using a set distribution, where the Welsh Government aren't. So, that gives us much more flexibility in the modelling.

Good morning. You recommend that further consideration should be given to undertaking further work to provide Welsh-specific data. What progress has been made on developing Welsh data for forecasting devolved taxes since the 2018-19 budget? 

In terms of the Welsh data, the biggest improvement has come from the WRA. There is, from what we see, a very good working relationship between the Welsh Government's Treasury team and the WRA, and therefore it makes sharing the data a lot easier.

In terms of the macro data that would feed into the model, the Welsh Government at the moment are looking at the potential of developing a Welsh gross domestic product, but that is something very long term and there is a lot of work to be done there. But they are currently looking at, 'Okay, what needs to happen to get that GDP number?', so it's at a very early stage.

Sorry, to answer your question, all of the data development would be related from the WRA.

How long would we have to have Welsh-specific data for it to be considered fit for purpose, or before you would look at it and think, 'Okay, so we can calculate something meaningful from that'?

First of all, the data that we do have is of a good quality. The only problem is that we are looking at one data point—maybe by the end of the year we'll have two data points. So, whatever data we do have, there needs to be discussion to understand what exactly that data does tell us. Now, I would expect to see a full-year set of data before we can use it with confidence, but that's not to say that the data that we have at the moment, the in-year data—there has been plenty of discussion around that, so everybody is comfortable on its use in the models.

09:10

I remember the Finance Secretary said it appears to be raising four times what it should, but he said that's only one quarter.

Yes, it's—. The LDT, it's—. It is a learning process with these models—you know, we're not talking—. They haven't been in use for a long time, so we do expect the performance of the models to improve over time. That's not due to changes in the modelling philosophy, but just the data becoming better and what we've seen with LDT is that that relationship between the WRA and the Welsh Government Treasury team is working, because there's that exchange of information between both that has picked up on this data, not to call it an 'issue', but—

So, it could be that the UK data had been woefully underestimating, possibly. I mean—. 

Yes. We've identified what's happened to the data. I think it's too early to say why that's happened. We can make assumptions—whether that's the WRA working better or better compliance. We're not too sure. We can assume, at the moment, but not make a definite—.

We'll wait for the full-year data, but it's looking positive at the moment.

And I've got to mention the 'Brexit' word: what consideration has been given to assessing the risks associated with leaving the EU, and how have these been applied to forecast models?

So, the forecasts themselves, they rely—

By the way, sorry, I've just realised, turning over the page, I think I've just trodden on your question. Sorry. Anyway.

Yes, Brexit itself. The models—they do rely on forecasts from the OBR. So, the forecasts currently used are based on the OBR's forecast from March of this year. They do incorporate a Brexit scenario into their forecast. So, Brexit is incorporated into the Welsh Government's forecast.

To be honest, the OBR's forecast does imply a soft Brexit—a strong, continued trading relationship with the EU. When we do get an update from the OBR on these forecasts at the end of this month, I do expect that scenario to have changed somewhat, given what's happened between March and the end of this month.

Of course, if you look at the taxes, you split them up into pairs. The land transaction tax and income tax would be highly dependent on what happens in terms of Brexit, and they could collapse, as they did in 2008-09. But rates and landfill disposals tax are much less affected by external activity, aren't they? So, you have taxes that are highly cyclical, and I don't think you can get a more cyclical tax than land transaction tax, can you? So, do you see that there are the two sides—one that is highly dependent and one that is much less dependent on what happens with Brexit?

Yes. With landfill tax, the same macros, to that sense, wouldn't be feeding into that model. So, I wouldn't expect Brexit to have that much of an impact on the forecast there. With land transaction tax, unfortunately, the key drivers there are property price growth and property growth transactions. And, of course, Brexit is going to have a big impact on them. So, we would expect to see that in the forecast as well.

So, in an ideal world, you'd have had the devolution of income tax—all right, you've got stamp duty land tax now—and the other taxes separate from the Brexit process, because it's going to skew the test of it, isn't it? You're going to have to really look into what's causing—

There's a lot we could say about an ideal world and Brexit, I think, but, yes, I think you make a fair point. Can I just ask one question about how suitable are UK economic determinants for forecasting the future performance of the Welsh economy and what more can we do to improve that, because we're heavily dependent on that, aren't we, really?

We currently are. What we've done is that we've looked at the historical relationship between UK data and Welsh data. So, in Wales at the moment, we are fortunate that we do have a quite rich collection of data, historical data. We've used a correlation test—it's a standard statistical technique to look at the strength of the relationship between two historical time series. What we found is that the correlation between Welsh data and English data specific to this exercise is quite strong. I think we've got correlations well above 90—so, a correlation of 100 would indicate perfect correlation, a correlation of 0 would indicate no relationship at all—so we're fairly comfortable from a statistical point of view that there is a strong relationship in the historical data. We can therefore assume that the forecast coming from the OBR—there will be that strong relationship again with any Welsh forecasts.

09:15

Yes, I think I'll just pick up where Nick Ramsay left off on LDT, exploring a bit more about forecasting, the forecasting issues with LDT. You say that the Welsh Revenue Authority data that they collected has influenced LDT forecasts, but we need to understand fully why there's been this significant upward revision in LDT forecasts since your updated statement. It would be helpful to tease that out a bit more. 

Absolutely. What we've seen is that the Welsh Revenue Authority have collected more standardised waste than what was anticipated. Again, nobody's too sure why there's been more collection of  standardised waste. It's certainly something the WRA is looking into. Now, it could be that the Welsh Revenue Authority have a better relationship with the stakeholders, it could be better compliance—so that part is still being investigated. But what we know is what has driven LDT, that forecast, is that standardised waste has been much greater than what we anticipated.

Now, the problem itself—from the meeting we had with the OBR, I don't think it's unique to Wales, in that the OBR representative also expressed that it is difficult to get the split between standardised and the other wastes correctly—. So, what I expect to happen is that, over the time this model is used, the split between standard waste and the other waste will be improved over time.

Yes. Obviously, there are dynamic policy changes driving this as well in terms of the local authorities' and NRW's roles in this. Also, you have no historical information in terms of those tax revenues. Also, we've touched on the issue of Welsh-specific data and the lack of it and how we can move forward, but particularly relating to LDT. How can we develop this to ensure that we've got that specific data integrated into long-term forecasting?

At the moment, what I'm seeing is that the WRA is improving the Welsh data that we have, and, due to that relationship, then—the good relationship between the Treasury team and the WRA—the models will benefit from that data collected from the WRA. So, that's the improvement we're seeing on the LDT side.

Diolch, Llyr. I'd like to take us through the land transaction tax and modelling around that, if that's okay. So, can you elaborate on why the forecasts have been revised downwards for land transaction tax since your updated statement?

Sure. There are two reasons for this: first of all, the worsening outlook for the property market since the last report we produced—so, the economic scenario we see in section 2 of the report, it is worse than the economic scenario in the previous report—but also the residential in-year data that we have had is a lot softer than we anticipated. For the non-residential, the in-year data was more or less aligned to the anticipated revenue we expected. However, with the residential, there was such a difference that we thought that that has to be taken into account in the model. So, it's between those two factors. The worsening outlook for the property market, plus the in-year data, then, for the residential, have pushed the revenues down for the LTT. The other thing to remember is that the residential—the revenue coming from the residential taxes do account for, roughly, 50 per cent of the total revenues raised in LTT, so anything that happens to the residential will have a big impact on the overall revenues raised.

09:20

Okay. That's helpful. So, the current model for land transaction tax is based on HMRC and OBR data to determine forecasts. Is it the intention to develop more appropriate Welsh data in order to develop a Welsh-specific model? I just have in my mind that the English data will include London and the south-east, which hugely skews the data for England, never mind hugely skewing the data for Wales.

What's happened with the property data is that there have been discussions with property consultants, or estate agents, large estate agents, just make sure that what we have as a forecast for Wales—sorry, for the UK as a whole—is relevant to Wales. So, we take comfort from those discussions. Historically, again, we've seen that there is this strong relationship in the data between Welsh property prices and the UK as a whole. I understand what you're saying about London, and I was quite surprised myself when we saw that strong relationship. So, between the two of them, that does give us comfort to still use the OBR data for this exercise.

There is a challenge with forecasting property prices, and, if I'm honest, I'd much prefer to see efforts being put towards modelling GDP—I think that would be much more beneficial than trying to focus on property prices. So, there is a trade-off there in developing property forecasts.

Two questions. Do you see an end-of-the-year effect: solicitors decide on finishing properties off, where there was some movement to try and finish off at the end of the year under a system they knew rather than going into the next year into a system they didn't know?

Yes. That's one thing that came up in the discussions with the property consultants and so forth. There has been a little, but nothing that would cause us to change anything in the model or the data that we've seen. It does happen, but not to any significant effect.

And land transaction tax is highly dependent on confidence, that people—a belief that they can upgrade their houses and still afford the new payments. Whatever your view on Brexit, it does put uncertainty into the lives of many people. Do you think that might have had some effect?

Again, I don't think it's had a significant effect on the forecast. It certainly does happen, but we're not seeing it in any other statistics; it hasn't been told to us by the consultants either.

So, if you just take house sales—tell me if I'm absolutely wrong, but isn't house buying another part of consumer expenditure, as is car buying? We've seen a drop-off in car buying, we've seen a drop-off in consumer expenditure on items they don't have to buy, so, discretionary expenditure. Why wouldn't a drop-off in the sale and purchase of houses occur at the same time?

Sorry, just to clarify, I'm sure it does happen, but I don't think it's happened to an extent where we'd have to change something significantly in the model to capture that. What we have seen already is the softening of the residential property market, so we've taken into account that in-year data. I think there's a drop of £10 million that we have in the model. So, that kind of reflects that slowdown in the consumer market. Other than that, I don't think it'd be appropriate to make any further adjustments to the model.

Absolutely. Absolutely. And one of the key recommendations that we have for this time is that there should be ongoing reviews of the performance of the models.

09:25

Thank you. You previously recommended that Welsh-specific elasticity should be calculated when suitable data becomes available. How's that piece of work progressing?

It's progressing very well. Just to point out, elasticities are not needed for this exercise. However, the Welsh Government is continuing to develop their capabilities in calculating elasticities. They have recently completed an exercise in looking at historical data, which could be used if needed. However, what they are trying to do is wait until the end of the year. Therefore, we'll have full one-year data, and then from that we can calculate, or they can calculate, elasticities. The benefit of doing that is that we'll be able to calculate the elasticities for Wales, but we'll have the rest of the UK as a control group. So, from a statistical point of view, it's quite good to have that control group because you can see what the impact was in Wales and what the impact was outside of Wales or what happened outside of Wales at that time.

That's useful. Can you tell us a bit more about the Welsh Government's approach to monitoring the accuracy of the landfill disposals tax and the land transaction tax forecasts?

Yes. The performance of the models will only be really known by the end of the year, when they have a full year's data available to them, due to the seasonality effects that we know happen for these revenues. At the moment, they are working on—once data becomes available, there is a discussion about: should this data be included in the models so that they do take into account in-year data and consider that in-year data against the performance of the model? What we've currently started discussing with them—and I'm sure we'll progress these ideas further before the end of the year—is having this annual review of the performance of the models. So, that would be a structured approach to reviewing the performance of the models and to decide whether the models are still fit for purpose. Now, that's not a simple case of looking at whether the revenue forecasted is the same as that collected. There needs to be a look at each component of the model to see whether it performed as expected. So, a good example would be with LTT: was the property price growth in line with expectation? Because what you could have with these models is the revenue forecast matching the revenue collected, but what you had in the model is that you overestimated the property price growth and you underestimated another part, and then on average—

—the model still works. So, what we're working on is developing a process that the Welsh Government can follow where they look at each individual performance of the model to say, 'Yes, it's still fit for purpose' or 'The model needs to be changed.'

That's really interesting. That might be something that the committee might want to take a look at with Welsh Government and see how that work develops.

Can I take us now to non-domestic rates, please? Your current report recommends that the availability of information on appeals could be improved. Has there been any development in this area since the issue was highlighted in October 2017?

Yes, we did make that recommendation. Appeals are probably the key uncertainty in NDR forecasting, and some progress has been made. The Welsh Treasury team have spoken with the Valuation Office Agency about what can be done and how to improve communication, and they're also in contact with Scottish officials who raised a similar issue, and they're working on it. The VOA have now got a specified individual who's the Welsh Treasury team contact, and they're communicating very frequently on how to get better information. The VOA are going to provide a list of appeals in progress in the system and the rateable value of each of those appeals. That is really helpful, but there still will be uncertainty, because the most useful information would be whether the appeal is going to succeed or not and, without that, there's still uncertainty. But that is really useful information that they can use in the forecasting process, yes.

Is it possible to use historic data to show what kind of percentage of appeals have been successful in the past, and is that useful?

Yes, it is, and that's what they're doing at the moment in the absence of anything else, but there's potential for just one or two large appeals to have a real impact, and so getting more intelligence on what's going on would be really useful.

09:30

Data collection. It's almost the opposite of corporation tax, isn't it, because larger multinationals have the ability to move the corporation tax amounts that they have to pay by using abroad as an internal movement of costs, whereas they can't do anything about the buildings they've got; but smaller companies, which are incapable of avoiding corporation tax because they're based here—of legally avoiding it or reducing it—are quite often able to, by closing and opening, avoid paying non-domestic rates. By the time their first rate bill comes in, they've closed down and moved to a place elsewhere. Do you think more ought to be done in order to ensure that the amount of rates coming in that you're modelling is actually close to the number of properties that exist?

Well, there was a consultation on this in the summer, on potential avoidance of non-domestic rates. That was only published yesterday, so I'm not fully familiar with the findings. The Barclay review, as well, in Scotland has made some suggestions. We'll see what happens. There are reliefs for smaller businesses, but I think the one you're referring to about the empty properties one, there have been instances of abuse, yes, and clearly, that's not fair to the other taxpayers who don't use that.

If I can move to the Welsh rate of income tax. The Scottish rate of income tax, they changed their rates and bands, have you looked at the behavioural effect in Scotland of that? 

Yes, I would be careful with looking at the behavioural effects up in Scotland. Certainly, there are lessons to be learnt there for Wales, but what I don't want to see happening is that behavioural effects are calculated directly from the Scottish case and plugged straight into the model. I think what's important is that analysis is done on what happened in Scotland, there's a discussion around that and then there's an agreement if that should feed into the models. At the moment, the Welsh rate model doesn't need the behavioural effects, but it does have the capacity to incorporate behavioural effects if needed, but there needs to be discussion about what those behavioural effects should be.

And, as you know, looking at Scotland, they had great difficulty identifying Scottish taxpayers, and there were a number of high-profile mistakes, but probably thousands of non-disclosed mistakes. It's not going to matter very much in terms of the first year, is it, because whenever the model comes out, we're not going to be affected? But are you confident that the model is going to pick up the vast majority/all Welsh income tax payers?

I think you had a comment on that one.

Well, I think the model will, the issue is what happens in practice with them being picked up. HMRC were before the Scottish finance committee a couple of weeks ago and they said in Scotland that they were confident that they had 99 per cent of the taxpayers identified. So, they clearly have steps in place that they're confident with to identify the taxpayers.

Yes. If I can take us on slightly further, the problem, again, certainly in parts of north-east Wales, is that people move across the border during the year. I'm sure if they work for Airbus, they update the place where they're living as they move, because they've got a human resources department that collects this data, et cetera, and payroll wants to know these things, but the small garage employing four or five people who know Dai or Jane may not be that bothered about where they live because it's never mattered up until now. Do you see a problem with that? And, do you see a problem with that fitting into the model, especially if there's a variation?

What will happen there, we'll find—. If there's a difference between the forecast and the in-year data, if the Welsh Government identifies a big difference between them, then certainly that would be something that has to be investigated with the WRA. So the model will be able to flag up if there's a big issue. Then, it would be a case of working with the WRA to identify what can be done to resolve that issue.

09:35

And, of course, we'll have HMRC before us next week, so I'm sure that's a question that we can pursue.

Sure. I want to pursue that, but I was just talking about the fact that you have a model, which is based on numbers; if the numbers are wrong, the model inherently gives you a wrong answer. However clever the model might look, and even if it takes everything into account, if 5 per cent of Welsh income tax payers end up not being identified as Welsh income tax payers, then the model will distort. And also, although you've got very few very-high-rate or top-rate income tax payers here, there's a tendency—and, again, tell me if I've got this wrong—there's a tendency for some of those very high income tax payers to own multiple properties. They are supposed to fill in where they spend the most time, but that may well be Monaco, or maybe somewhere like that. It does create difficulties, and there's the possibility of people flipping and that, again, not feeding into your data. Now, at the same amount of income tax, to an individual, it makes no difference whatsoever where they pay because they're paying the same amount, but if there's any variation then do you see the model having to take account of that in terms of flipping?

I wouldn't expect the modelling philosophy to change; I would still expect that to remain as it is. And, if there was a decision to change the modelling philosophy, that would be a concern because we'd then be looking at a different modelling methodology to other administrations that we know of. So, that would cause concern. What I anticipate is that the data feeding into the model will improve over time, and that will improve the forecast generated by the model. 

Okay. And, in doing the model checking, does anybody take what will be the English taxpayer—? Will somebody take what is the English taxpayer, the Welsh taxpayer and the Scottish taxpayer and make sure and work out if they add up to the total number of taxpayers, to make sure that there aren't people who are being missed, or even double counted?

I'm sure that is something that the Welsh Government treasury team and the WRA will work between them. The benefit with the model at the moment is that it will identify if there is an issue there with the data. Then, once that's been flagged up, it can be investigated further. 

Which takes me on to my last question: what data should the Welsh Government collect in order to refine the model?

The Welsh Government, HMRC and the Scottish Fiscal Commission are currently working together to develop real-time information that could be used by the model. There is currently a rich data set from HMRC; however, that data is somewhat out of date. It's still useable, but real time will certainly be better. So, I think progress in that area will certainly benefit the model. 

What also needs to happen is to develop our understanding of the tax base itself, which brings me back to the point about GDP. I think it's important for us to look at: should we be developing a Welsh-specific GDP, and the work involved? I think it's a long-term aim, but it could be a worthwhile aim, because it will help us understand the tax base. 

Okay. On that, my understanding of GDP—again, correct me if I'm wrong—is GDP is collected where it's earned, not where people are living. Now, in the north-east, movements from north Wales to Ellesmere Port and from England into Airbus are well known, but there are people moving across to much smaller organisations—and I have a view that the GDP of Cardiff is overestimated, because of the people moving in who are earning, and the GDP of the south Wales Valleys is underestimated. That's not to say the south Wales Valleys aren't poor and Cardiff is relatively affluent in Welsh terms, but it does tend to do that. Don't you see a problem of doing it only in Wales that you may be counting income earned in Wales for people who are not taxpayers in Wales? 

09:40

And I think you're right, and that's why it's important to initially look at the challenges involved in calculating Welsh GDP. So, a discussion can be had, 'Is there a benefit to have this number, given the costs or the work involved?' I think that part is important. That's why I'm saying calculating Welsh GDP is a long-term aim, because we need to do this work at the start to see what challenges are involved. 

Again, you can disagree with me if you like, but I think collecting—household income is a much more useful amount to tell how affluent an area is, and more helpful in that, than actually collecting GDP, which is earned in one place and spent in another. As far as the GDP is concerned, I help Cardiff GDP and Nick helps Cardiff GDP, though we actually live, one in Monmouth and one in Swansea, which are well outside of Cardiff.  

I think, again, you're right, and that has to be part of the discussion to see, well, is there a better measure that we could derive for Wales, rather than GDP itself, but I do think it's important that we have those initial steps to understand what challenges there are in calculating Welsh GDP. At that time, we can decide, given the challenges and the work involved: do we want to proceed with that, or would it be better to look at other measures? 

Thank you very much. Do Members have any further questions? No. Okay. 

Iawn. A gaf i ddiolch o galon ichi am ddod atom ni'r bore yma? Diolch yn fawr ichi am eich tystiolaeth. Mi fyddwch chi'n derbyn drafft o'r trawsgrifiad i gadarnhau o safbwynt cywirdeb ond, gyda hynny, a gaf i ddiolch yn fawr i'r ddau ohonoch chi am eich tystiolaeth? Diolch yn fawr iawn.

Fine. I'd like to thank you very much for being here this morning. Thank you for your evidence. You will receive a draft of the transcript to confirm factual accuracy, but, with that, I'd like to thank you both for your evidence. Thank you very much. 

4. Cynnig o dan Reol Sefydlog 17.42 i benderfynu gwahardd y cyhoedd o weddill y cyfarfod
4. Motion under Standing Order 17.42 to resolve to exclude the public from the remainder of the meeting

Cynnig:

bod y pwyllgor yn penderfynu gwahardd y cyhoedd o weddill y cyfarfod yn unol â Rheol Sefydlog 17.42(vi).

Motion:

that the committee resolves to exclude the public from the remainder of the meeting in accordance with Standing Order 17.42(vi).

Cynigiwyd y cynnig.

Motion moved.

A gaf i hefyd symud at yr eitem nesaf, sef fy mod i'n cynnig yn unol â Rheol Sefydlog 17.42(vi) fod y pwyllgor yn gwahardd y cyhoedd am weddill y cyfarfod? A ydy Aelodau yn fodlon â hynny? Iawn, diolch yn fawr iawn. 

I would also like to move to the next item, namely that in accordance with Standing Order 17.42(vi) the committee resolves to exclude the public from the remainder of the meeting. Are you all content with that? Okay, thank you very much. 

Derbyniwyd y cynnig.

Daeth rhan gyhoeddus y cyfarfod i ben am 09:42.

Motion agreed.

The public part of the meeting ended at 09:42.