Y Pwyllgor Cyllid

Finance Committee


Aelodau'r Pwyllgor a oedd yn bresennol

Committee Members in Attendance

Llyr Gruffydd AC Cadeirydd y Pwyllgor
Committee Chair
Mark Reckless AC
Mike Hedges AC
Nick Ramsay AC
Rhianon Passmore AC
Sian Gwenllian AC

Y rhai eraill a oedd yn bresennol

Others in Attendance

Yr Athro James Foreman-Peck Athro Economeg, Prifysgol Caerdydd
Professor in Economics, Cardiff University

Swyddogion y Senedd a oedd yn bresennol

Senedd Officials in Attendance

Christian Tipples Ymchwilydd
Gemma Gifford Dirprwy Glerc
Deputy Clerk
Georgina Owen Ail Glerc
Second 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 09:32.

The meeting began at 09:32.

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

Bore da i chi i gyd. Croeso i gyfarfod Pwyllgor Cyllid y Cynulliad Cenedlaethol. Gaf i nodi, fel arfer, wrth gwrs, fod yna glustffonau ar gael ar gael ar gyfer y cyfieithu neu ar gyfer addasu lefel y sain? A gaf i atgoffa Aelodau i ddiffodd y sain ar unrhyw ddyfeisiau electronig sydd gennych chi? Ac a gaf i ofyn a oes gan yr Aelodau unrhyw fuddiannau i'w datgan? Dim byd. Iawn. Diolch yn fawr iawn.

Good morning, everyone, and welcome to the Finance Committee meeting of the National Assembly for Wales. May I note as per usual that there are headsets available for the interpretation and sound amplification? May I remind Members to switch the sound off any electronic devices that you may have? And may I ask if Members have any interests to declare? None. Right. Thank you very much.

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

Fe welwch chi fod yna un papur i'w nodi, sef ymateb gan Swyddfa Archwilio Cymru i adroddiad y Pwyllgor Cyllid ar y gwaith craffu blynyddol ar Swyddfa Archwilio Cymru ac Archwilydd Cyffredinol Cymru o 20 Chwefror 2020. Dwi'n cymryd ein bod ni'n hapus i nodi y papur hwnnw. Iawn. Diolch yn fawr iawn.

You will see that we have one paper to note, which was the response from the Wales Audit Office to the Finance Committee's annual scrutiny of the Wales Audit Office and the Auditor General for Wales's report from 20 February 2020. I take it we're happy to note that paper. Fine. Thank you very much.

3. Effaith amrywiadau yn y dreth incwm genedlaethol ac is-genedlaethol: Sesiwn dystiolaeth 3
3. Impact of variations in national and sub-national income tax: Evidence session 3

Ymlaen â ni, felly, at brif eitem y cyfarfod y bore yma, sef parhau â'r gwaith o edrych ar effaith amrywiadau yn y dreth incwm genedlaethol ac is-genedlaethol. A gaf i estyn croeso arbennig i'r Athro James Foreman-Peck atom ni bore yma, sydd yn athro economeg ym Mhrifysgol Caerdydd?

Mi awn ni'n syth i gwestiynau, os ydy hynny'n dderbyniol. Ac mi wnaf i gychwyn, os caf i, trwy ofyn: mi wnaethoch chi gynnal astudiaeth, wrth gwrs, fel rŷn ni wedi gweld, yn 2019, yn edrych ar ymfudo ac enillion drwy dreth mewn awdurdodaethau is-genedlaethol. Allwch chi efallai ymhelaethu ychydig ar y model y gwnaethoch chi ei ddefnyddio a'r rhagdybiaethau sy'n sail iddo fe?

We'll move on, therefore, to the main item of this morning's meeting, which is to continue to look at the impact of variation in national and sub-national income tax. May I extend a particular welcome to Professor James Foreman-Peck this morning, a professor in economics at Cardiff University?

We will dive straight into questions, if that's acceptable. And I will begin, if I may, by asking: you undertook a study, as we have seen, in 2019, looking at migration and tax yields in sub-national jurisdictions. Can you perhaps elaborate a little on the model that you used and the assumption that it's based on?

Hello. We began this work in, I think, 2016, and we put it to bed in 2019. It consists of two types of modelling: (1) an econometric one, to try and get some information about the extent of tax-induced migration; and the other (2) a computable general equilibrium model, to try and get the effects of migration on the Welsh economy. We distinguished three types of people that corresponded to three types of taxpayer, which made the exercise quite complicated, so there are about 30 equations in a CGE model, but to avoid it becoming any more complicated, we excluded capital from the model. We've got a lot of taxes in there, so apart from income tax, we include value added tax, national insurance, council tax, and then we have a composite for other taxes thrown in. What else can I say that would interest you? We have the public sector in there, with a public sector budget constraint, which is determined by the amount of tax revenue and the grant money from the central Government in Whitehall. An important point about the trix [correction: the model] is that there are no dynamics in it, so we can't actually say how long anything takes to work through. What we actually get is, 'If you do this, then it will bring this about, eventually.' We don't know whether it's going to happen instantaneously or not. We're just saying, 'Well, we've got two equilibria and we're moving between them.'  


Okay. So, is there enough data out there for us to have this robust research in relation to cross-border effects of tax variation, and what could the Welsh Government maybe do to facilitate more information and more data being available?

The data we used for the migration was movements between local authority areas, and we aggregated that. We tried to isolate the amount of that that was due to differences in local authority taxes. There's plenty of data on that, because there are a large number of local authorities in England and Wales, and a large number of moves. Even so, the moves will probably be understated because we don't have a police state, so you can't exactly control how many people move from one area to another. But the basis of that data is, ultimately, from the Office for National Statistics, but they use national health service data. So, if you register with a general practitioner you'll get picked up if you move GPs between local authority areas, and also higher education, because a big part of migration is people moving to and from higher education.

Sure, okay. So, in your view, how influential is migration across the Wales-England border when rates diverge and what other behavioural responses should be considered when modelling the impact of tax variations, do you think? 

Well, one of the characteristics of migration is that there's a lot, quite a lot, relative to the population going in both directions. So, the net migration is very small—more than 50,000 in each direction—[correction: But the net migration is very small—although more than 50,000 move in each direction—] and there is a net inflow of people coming to Wales. But a very small proportionate change in one of the gross outflows could cause quite a large flow in the net outflow, because it's the balance between two items.

The critical point about migration, I think—and this is important when you want to look to Scotland for information about what happens when you change rates—is the distribution of the population. It's [correction: Wales's is] very different from the distribution of the population in Scotland. I used to live in Newcastle upon Tyne, and you're very conscious that it's a long way to get to Edinburgh—I rarely did—or indeed York, whereas, if you're in Cardiff you know that actually getting to Bristol is pretty easy. We also know there's a lot of commuting; I was surprised at how much commuting there was to Chester from Wales, for example. So, it doesn't take much—I think it doesn't take much—to shift the balance of where you work, if it's worth while. And the people that it's going to affect are the higher income earners, because they pay the most tax, therefore they will have the most to gain or lose from tax changes. So, I wouldn't expect much to happen if you change the basic rate [correction: the basic rate for basic rate payers], but if you change the higher or additional rates, [correction: but if you change taxes on the higher or additional rate payers] then, eventually, there could be quite substantial changes. 

Okay. That certainly is a message that came through clearly last week as well, but I know Members will want to drill down a little bit into that later on. Can I just conclude my opening questions by asking how rate variations impact on the Welsh economy, particularly in areas such as gross value added and employment.


Well, the way we do it is our three types of people have different productivities according to the amount of money they earn and the amount of tax they generate. So, if you attract or repel higher earners, ultimately, this has a spillover effect to affect GVA per worker. The extent of the effect—we do find this in the model—. It's a relatively small effect because we don't have things that might boost it, partly because we really don't know. We would think that including capital in the model would probably make for bigger effects, but there are effects, given the assumption we've made about individual productivities. 

Jest ynglŷn â'r data roeddech chi'n ei ddefnyddio ar gyfer y gwaith—data roeddech chi'n ei gael yn ôl ardal awdurdod lleol, dwi'n cymryd, onid e? Oeddech chi'n cymharu awdurdodau lleol ar y ffin, ynteu a oeddech chi'n cymharu ar draws Cymru a Lloegr? Oeddech chi'n canolbwyntio jest ar un set o ddata yn ymwneud efo'r ffin?

Just on the data that you used for the work—that was data that you had according to local authority area, I take it. Did you compare local authorities on the border or across the Wales-England border, or did you focus on just one set of data relating to the border?

We didn't look at the border. We did an aggregation of—. First of all, we estimated an equation for the whole of England and Wales to see what the response rates in migration to differentials in local authority council tax were, and then we aggregated those [correction: aggregated Welsh local authorities] to get an effect for Wales. 

With regard to your very specific modelled study, is there anything that would be comparable to us in terms of our study around Welsh rates of income tax that you could deduce from your study that would have any, sort of, underscoring of what we're doing with this inquiry, or is it just stand-alone and there are no lessons to be learned from it as far as this committee is concerned?

You're asking whether there are other studies.

No, I'm asking: this particular study, with regard to the local authority study that you've mentioned—is there anything that would be of use in terms of its outcomes for this committee in terms of our inquiry?

Anything more? Well, I think we're telling you quite a lot about what happens if you change tax rates, which, I believe, is the central interest of the committee. 

Well, we can produce a table of 1p, 2p, 3p, 4p, 5p changes, and, indeed, it's on my website, because it wasn't published in the original article. So, we got quite a lot of useful information. There's a question of whether one believes it, but—

Well, we'll be picking up on some of those specifics in a minute now, I think. So, we'll come on to Mike, if we may.

You talk about the effects of tax, but it's almost as if tax is on its own. I mean, if you look at council tax, for example, there's about a £600 variation between Newport and Blaenau Gwent, but house prices are substantially higher in Newport than Blaenau Gwent. So, the type of property you get for a certain sum is entirely different. So, I think you have these other things coming in, don't you, in terms of house prices? We've seen not a taxation but a behaviour change following a reduction in costs with the removal of the tolls on the motorway bridge. So, you have all these other factors that come in. Isn't treating tax in isolation dangerous in the sense that you get a result but it seems to be based upon everything else being the same?

Well, we do consider lots of other things, and that's why the whole business has been so time-consuming. So, for example, in the migration model, we consider distance as an influence on migration—the distance between local authority areas. We did try including house prices, and it doesn't really make much difference, or didn't, but we've got individual effects for local authorities—all 348 of them, or whatever—so it's not as if we just considered the effects of tax changes. 


If you compare anywhere in Wales with Chelsea, for example, £1 million will buy you a one-bedroomed flat in Chelsea; £1 million in Wales will buy you some of the biggest houses available, so you do have this variation, don't you, in what a sum of money will buy you? A quarter of a million pounds will buy you a very nice property in most of Wales, but you're probably struggling to find a place somewhere in some of the more expensive parts of London, and even some of the medium-priced places in London, so you do have those variations.  

You do indeed have those variations, and we hope we've taken them into account. 

If I can move on to discussing Wrexham and Chester, because they effectively run into each other. The Deva Stadium where Chester FC plays football, the football pitch is in Wales but the offices are in England; it takes you to that sort of level of how they fit together. But with teachers' pay and NHS pensions being kept in line with England, I can see some argument that if house prices were the same, people might decide that they would be better off living on the English side of the border if taxes were higher, or the Welsh side of the border if taxes were lower. Do you think that's a reasonable assumption? 

It is, and I think it's not only a reasonable assumption, but it's borne out by the evidence of commuting flows. On balance, you'll find people commute from Wales to England much more than from England to Wales, and the likelihood is, along the lines that you've been indicating, that's because the houses are cheaper in Wales.  

I hold a view, which I'm not sure that you quite agree with, that for people in the higher rate but not in the very high rate of tax, house prices are incredibly important, far more than marginal income tax rates, but I'm not sure that's a view you would agree with.

Important to what exactly? 

To decisions on where to live. If we increased the tax by 1 percentage point here, compared to England, but if house prices were—[correction: were lower]. In Cheshire and Flintshire, you see very substantial price differentials, a lot caused by the fact that there's controlled planning in a lot of the smaller Cheshire villages, or even the larger Cheshire villages, so you've actually got people having to pay a premium to move one side of the border, which I would say would outweigh the 1 per cent cost. 

For most people. This is a fair point that, actually, people on the basic rate are not going to be particularly better off or worse off if you change the basic rate by a few pence, and that won't be decisive in influencing their location, but it only has to be of relevance to people who are thinking they might do this or they might do that. It tips the balance. That's one of the rules of economic analysis, that it's the marginal people that determine the change, not the average. The average person isn't going to be affected, but there are some people on the edge who are. 

Have you looked at Ontario and Quebec, which run into each other, or Luxembourg? You could live in Luxembourg and you could work in at least three different countries without commuting very far. So, the difference in taxation between some of those affects the movement of people. I know Monaco, which is almost zero tax but very small, is different, but it's a collection of sports stars. 

There are quite a few international studies that probably get a handle on the sort of thing that you're interested in. There's particularly one of Denmark, where they changed the upper tax levels. They got a lot of international movement in as a result of the change in the tax rates there, because, well, I don't know about the borders in Denmark, but high-income earners do seem to be relatively mobile when it comes to where they locate. 


Hence Monaco in very many cases. Your evidence notes that a 5p increase in the basic rate would generate additional WRIT but reduce total tax revenue generated in Wales. Why do you think that it would have that effect?

A reduction of 5p in the basic rate would increase tax revenue—sorry, it would reduce it because there'd be less tax paid per person. There would be some migration effect, but it would be, I would have thought, small enough to ignore. I mean, it's certainly small anyway because basic rate payers are not going to shift location for £200 or £300 a year. That's the bottom line. But once you get to higher and additional rates matters matter much more.

Can we concentrate on the higher rate taxpayers rather than the additional-rate taxpayers? Because this higher rate includes a lot of people who work in health, in education, in public sector bodies or work as accountants, architects or other jobs that they specifically need to be in the area they are. They need to live near to where they work; they haven't got that mobility. You talk about the additional rate, that's slightly different, that tends to move into a much smaller group of people, some of whom do have mobility—I don't disagree with that. But on the additional rate, am I wrong in saying a lot of them are in jobs where they are constrained by where they can move to?

Well, it didn't look like that from the exercises we did, that's all I can say. We didn't burrow down to find out why they appeared to be so mobile.

I mean, have you got a list of occupations for the additional taxpayers? Because occupation can quite often be a major determinant of whether you can move or not. If you happen to be the local solicitor in one of the Swansea suburbs, then you need to be in Swansea and around Swansea, if only to help develop your business.

Well, I agree that would be an interesting collection of data if we could get it.

And finally, on the additional rates, these have very few members in Wales but they disproportionately pay in. Some of them are in the public sector and are quite often constrained by the fact that this is where they live and work; others have greater mobility. Do you know what proportion are in there who have that greater mobility?

No, this is at a level of detail that we've not yet been able to get down to, unfortunately.

Ydych chi'n gallu dod i gasgliadau mawr cyffredinol yn sgil eich gwaith chi? Er enghraifft, sut fyddai gwahaniaethau treth rhwng Cymru a Lloegr yn effeithio ar allbwn fesul gweithiwr a sut fyddai hyn yn effeithio ar economi Cymru?

Can you come to any conclusions, major general conclusions, from your work? For example, how would tax differentials between Wales and England impact on output per worker, and what effect could that have on the Welsh economy?

As I said, we do pick up effects on output per worker, and the basic rule seems to be that if you can attract migration of higher earners, their higher productivity spills over to the economy as a whole and there is some effect. And conversely, if you tax people so you get higher earners moving out, then that tends to pull down the productivity, but it's a fairly small amount. But then, the extent to which it happens, I think, is probably not well captured in this model. I think just a qualitative movement is probably reliable but not the magnitude.

Ocê. Ac ydy'r darlun rydych chi'n ei ddisgrifio yn mynd i newid dros gyfnod? Ydych chi'n disgrifio'r darlun yn syth ar ôl i'r newidiadau ddigwydd ynteu ydych chi—? Deg mlynedd i lawr y lein, a fyddai'r un effaith yn bodoli?

Okay. And is the picture that you describe going to change over a period of time? Are you describing the picture immediately after the changes happen, or is it 10 years down the line? Would the same impact still exist 10 years down the line?


This is one of the weaknesses of the exercise. I would say, if you wanted a number, 10 years down the line would be more accurate than tomorrow, but, as I say, we don't have any time dimension in this, it's just today and sometime in the future.

Ac wedyn, rydyn ni wedi cael tystiolaeth ar gyfer yr ymchwiliad yma yn dweud bod angen amrywiad tymor hir o 5 y cant i 10 y cant mewn cyfraddau treth incwm i gael unrhyw effaith sylweddol ar ymfudo. Beth ydy eich barn chi am hynny?

And then, we've had evidence for this inquiry noting that a 5 per cent to 10 per cent long-term variation in income tax rates may be needed to have any significant impact on migration. What's your view on that? 

I suppose, to some extent, it depends on what you think is significant. My view is that a relatively small change in income tax would eventually have a significant effect. The point would have to be that it would have to be credible and sustained, precisely because we don't know how long it's going to take. So, it would involve people believing that this income tax change was going to persist for maybe 10 years or more, because of the time it takes for people to adjust their behaviour and to believe it's going to be worth while and won't be changed back to where it was before. So, I think the credibility of the change, or the credibility of the persistence of the change is probably the most important aspect of deciding on whether the effect is going to be significant or not.

Okay. So, 1 per cent could make a difference; it doesn't have to be 5 per cent. 

That's our reckoning, yes.

Can I just question the underlying hypothesis? If you have differences in council tax rates between pairs of local authorities, you could see migration on account of those. As a property tax, isn't it more likely that those differences in council tax would become embedded in house prices and rents, and that would be the mode of adjustment?

Well, if it was the mode of adjustment, and I take the point that you would expect some capitalisation, then you wouldn't get any migration in response to tax differentials. But I think that probably the answer is that the capitalisation is only partial. The capitalisation of the tax differential in house prices is not—well, there are papers that have come up with about 30 per cent, so there is still an effect. So, that's our argument as to why we think we find a migration effect for tax differentials, despite the capitalisation in house prices that you allude to.

You said twice 'tax differentials', isn't there a very important distinction between property tax differentials and income tax differentials?

There is indeed, and we had to jump through some technical hoops of some complexity to get them equivalent. Your point is that, I suppose, if you tax house prices, that affects the price of houses relative to other goods, whereas income tax affects the overall budget. Anyway, that's what we tried to take into account in our modelling exercise. So, we transformed the effect of one into the other.

When you say studies show a 30 per cent capitalisation, what is your interest rate, implicit or expressed, in those models that would link the annual council tax here to the capitalisation of house price value?

Well, I can't remember, because it was done quite a long time ago, but I take the point that the interest rate you use makes a difference to the result you get. The line we took was, 'Look, we're finding some migration as a result of this differential in tax prices, therefore capitalisation must only be partial.'


Have you considered the hypothesis that the council tax rates, rather than being an exogenous value, may reflect the standard and capability of local government in that area, such that council tax rates shoot up when the local government is having trouble in terms of the delivery and quality of public services?

We did think about it and, ultimately, this could be a weakness of the exercise, but the fact that we find this partial correlation makes us think we are picking up something. The method of local authority finance is, to some extent, to compensate local authorities that have got different financial performances, either because of their own competence or because of the tax bases they've got to work with. I believe that's been diminishing, but the period we did this was 2014, and I don't think it was as bad in that year as it has become subsequently. 

You referred to using GP registration data to at least assist in your modelling of migration. What consideration, if any, have you given to how accurate that data is? I wonder if it may significantly understate migration to the extent that a substantial number of people may not change their GP or update their GP data in response to moving, at least until they have an illness that requires them to go to the GP.

I think that's right. I think the migration is almost certainly understated, regrettably. Whether that would understate or overstate the effect we find I don't know, but it's the best data we've got. It's Office for National Statistics data.

Other things being equal, if the migration is higher than your data suggests, wouldn't that imply a higher tax impact than stated?

That's what I'd like to think, yes. [Laughter.]

Can I just ask you a bit about the higher versus the additional rate? I think, when we've had other evidence, we've perhaps struggled to stand up the hypothesis that you could increase your tax revenue by cutting higher or additional rates to attract more people to Wales. We've yet to find the evidence that would satisfy us that would be the case. I just wonder if you think that effect could be more powerful at the additional rate level, and would the impact primarily come through the 10p slice we'd then get of the higher rate and the basic rate of those people paying the additional rate who might be attracted in?

I have to say that I show a tendency to get confused between the additional and higher rates, and which is higher. Is the higher higher than the additional?

The additional rate people come out as being very mobile. That's the result. They pay a lot of basic rate tax, so that's one reason why they might be responsive to basic rate taxes. Changes in the additional rate itself are less important in our neck of the woods because, even if you're an additional rate payer, you're not paying that much more as a result of it, but you're paying a lot of basic rate tax or, indeed, higher rate tax.  

So, say we were to bring the additional rate down to the higher rate, halving our slice of the additional rate, could we expect that offset of 5p on those additional rate revenues, which are very small for Wales, to be offset by the people brought in who would be paying the higher rate on everything below £150,000 and then the basic rate on all that slice, and the productivity impact they might also have? Is that a credible proposition for public policy, do you believe?

I've got some figures here that I'm just looking at to make sure I'm getting it right. It looks as if changes in the additional rate don't make any significant difference to income tax generated in Wales. The important ones are the higher rate and the basic rate, and I think that's for the reason that I stated that the amount of money at stake at the top end is relatively small, whereas if you're paying basic rate over £150,000, 20 per cent of, say, £150,000 is quite a lot. But if you're getting a few thousand over £150,000, whether you pay an extra 5 per cent on that or not isn't going to make a great deal of difference. 


And you emphasise in your evidence that it could be perilous for us to seek to raise tax revenue by changing the rates, that we should perhaps focus more on expanding the tax base. What levers would you see as available for us to pull to do that?

Well, the policy that most impressed me was attracting Admiral to Wales. That seemed to be worthy of greater study as to how that was done, and I've made some casual enquiries about it, and the answer I get is that it was all due to personalities. But, Admiral is such a big company now, and it didn't exist before it came to Wales. And it seems to me if you could just do two or three more like that, you would be making a great deal of difference to the Welsh economy. 

Both Rhianon and Mike have been very patient, so Rhianon first, and then, Mike. 

Yes, thank you. It sort of goes back to where I started earlier. I've got your table in front of me. You've mentioned a few times that there's difficulty in aggregating data, and you've also mentioned, in terms of your study, the issue around duration and time as an impactor within that. So, I suppose it's really more—. I'm asking you really to extrapolate your model in terms of how you projected your hypothesis in terms of that study of local authority level data, and how that then can be extrapolated into this wider income tax agenda. Could you just give me a little bit more information as to how you did that, because it's obviously not an exact science, is it?

Well, because we use two models, and we have to go to and back between them, partly because the taxes are different involved, so—

So, we're looking at the effect of council tax, and saying—

That's my point, I suppose. 

Let's suppose it wasn't a council tax, it was an income tax raising the same amount of money, how would that affect people's behaviour, and put that in the model. So, that's what the computable general equilibrium model is about. 

Well, we all know the computer says 'no' often. So, in regard to that process, how robust do you feel that that is in terms of your conclusions based on that extrapolation? I suppose that's what I'm asking. 

I've no reason to believe there's anything wrong with it. [Laughter.] As yet. 

One comment, basically. Admiral was funded as a start-up, and as people heard me say yesterday, and I'll say again today, funding local start-ups, or start-ups in your own area, are far more economically beneficial than trying to attract branch factories that may stay only a short period of time.

But the question I was going talk about is migration. Within Wales, we've seen substantial migration into the Cardiff area, and that's been driven by employment. And migration also drives up house prices. And leaving aside those where we've got a substantial number of holiday homes, in general the migration coming in is the thing that is driving up house prices themselves. For what you're paying for a terraced house in Adamsdown, you can get a very nice property in some of the south Wales Valleys, because the movement is basically north-south within the Cardiff city region.

Yes. What do you want me to add to that analysis? [Laughter.

I would like you to agree, actually. But what I'm trying to say is: isn't that, over and above the council tax, the driving factor?


Well, jobs seem to be the driving factor behind the pattern of population and changing pattern of population in Wales, that's true. How do you take that into account? Well, the way we took it into account was we said that every local authority was going to have a different effect. So, some local authorities that have more employment will have a stronger effect than those that have weaker employment, and we let the data tell us which local authorities were in those categories. In the language of the game, we used dummy variables for each local authority—so, we had 348 dummy variables in the equation. 

Diolch, Chair. Good morning. In your article, 'A Welsh tax haven?', published in 2017, you found that any tax cut on the additional rate would always raise tax receipts, with a tax rise always leading to reduced tax revenue due to induced migration. Can you quantify how the degree of divergence between the additional rates in Wales and England would impact on the revenue?

This 2017 thing, which I would like to complain about the title of it. It made me very angry at the time, because I thought the Institute of Welsh Affairs people were either making fun of me or trying to discredit the whole exercise.

Yes, publicity, that's it—I forgot that—but we're retiring people, we academics. [Laughter.] So, the reason I'm flannelling about is that I think the model may well have changed since 2017 and I did submit a document to the committee that was more up to date. So, can you remind me what you think I said, what I did say or what—?

I often have to be reminded of what I've said in the past. You said that you found that any tax cut on the additional rate—so, that's the top-most rate, isn't it—would always raise tax receipts, with a tax rise always leading to reduced tax revenue due to induced migration.

Yes, I don't think that's true.

I'm sure I said it. Well, I'm not sure, but if you say I said it, I said it. But the point is that I mentioned that we started work on this in 2016 and we've been grinding away at it for three years, so I know that earlier versions of the model come up with different results from later versions, or at least our interpretation of the model comes up differently. According to the current version, the published version in 2019, basically, changes in the additional rate don't make any difference to income tax generated in Wales, because of the behavioural response, the migration response. Why is that?

Can I just clarify: income tax generation in Wales or income tax yield for Welsh Government? Which are we talking about?

Yes, good point. Either will do actually in this particular case—both seem to apply for—. Yes, so in the 2019 article, we distinguish between the Welsh Government's income tax and the total income tax generated in Wales, but in both cases, the additional rate doesn't make any significant difference. 

No, it's completely useless according to—. Well, actually, that's not quite true. It seems to make a small difference to GVA per worker, and to total tax revenue generated in Wales.


If I just ask you the second part of the question, which might help in terms of understanding the first bit, other evidence noted that, given the Welsh rate of income tax enables Welsh Government to benefit from the full mechanical effect of the tax rate change and only a proportion of the behavioural effect, then it is highly likely that increases in tax rates, even the additional rate, would increase Welsh Government tax revenues. Would you agree with that?

Well, increases in the basic rate increase Welsh Government tax revenue. Increases in the higher rate increase Welsh Government tax revenue.

That's what we think, except the additional rate doesn't seem to do anything.

It takes money away from the people who are paying it, but it doesn't produce any net revenue for Government, either in Wales or UK, are you saying?

That's the story, yes.

Thank you, and we've got the updated paper here as well. So, in terms of—. You've already partially covered this, in terms of a mention at the beginning around Scottish rates of income tax, but, in that arena, are you aware then of any behavioural effects associated with a divergence between English and Scottish income tax rates?

No. One of the things about Scotland that I suppose I'm moderately aware of compared to Wales is that there are two sources of high earners—they're the lawyers of Edinburgh and they're the oil men of Aberdeen. To go back to Mike's point, these people are not going to want to—. The propensity to migrate in response to tax differentials of these people I would think would be low. And I don't think we've got people like that in Wales, but I'm willing to stand corrected, if we can find the data.

We keep coming back to this, don't we, in terms of finding the data. It must be extremely difficult to model something if we don't have enough inputs into that machinery.

Well, the thing about modelling is it involves simplification, and the trick is to get the right simplification. And if you haven’t got as much detail as some people think, you just hope you've got enough detail and that this is the right simplification.

You said the lawyers of Edinburgh. You've also got the finance people in Edinburgh, haven't you? Edinburgh's got a very advanced finance industry, both in banking and in investment banking, and they could move, but the house price differential from living in Edinburgh to moving to central London is huge, isn't it?

Still, Edinburgh's pretty pricey, I'd expect.

But you can get more than a one-bedroomed flat in Edinburgh for £1 million.

Yes, yes.

Thank you. So, in regard to any behavioural effects—I understand your analogy and your anecdote there—what lessons could Welsh Government learn from the divergence between English and Scottish income tax? Or is that very difficult to encapsulate?

Well, my point was—is—that the distribution of population makes a big difference to what migration effects you're going to get, and there are millions of people that are scattered on either side of the Welsh-English border—i.e. within 25 miles or so—and you don't get that in Scotland. There's just this great empty space.

Okay. So, in terms of that comparability, you don't think it's a useful analogy.

I think it would be difficult to extrapolate from it, and that's the general problem with extrapolating from international studies.

Okay. I understand.

So, in regard to evidence sessions that we received last week, we discussed how the whole notion of incorporation in to regard self-employed individuals—how they could be potentially driven to incorporate their businesses in regard to non-dividends in response to tax rate variations, and business owners could minimise their personal income, obviously, by changing this form. So, in your view, how could any potential drivers to incorporation—? How could such responses impact on Welsh rates of income tax revenue, and have you been able to then model that in any of your two studies that you've mentioned in your model?


Well, we don't distinguish between employed and self employed, so that's just thrown into the whole melee.

Okay. Okay, thank you. Finally then, in regard to where we went in terms of means-tested benefits, for instance, universal credit and impacts on welfare benefits in regard to any differentials in terms of Welsh rates of income tax, what consideration have you given to the potential unintended consequences of Welsh Government tax policies, particularly how a variation in tax rates would impact on individuals receiving means-tested benefits?

I would say, if you're asking what consideration we've given, the answer would be none, frankly.

Fine. That's fine. I understand that from your previous answer, but I needed to ask it. Thank you.

But it is something that's come up in other evidence that that clearly might have an impact in the Welsh context. We're nearing the end of our allocated slot. Any further questions from Members? Are we content?

So, in regard to the extrapolation upward from your data that you've got around local authority differentials, your model in regard of how you do that, how can you reassure me that your model is significantly robust to be able to do that in regard to income tax—Welsh rates of?

Well, one way of testing robustness is to try changing the parameters of the model by a bit and seeing how significantly it affects the results, and we did a bit of that, but it's a fairly—. I mean, although it's supposed to be a simple model, it's got 30 equations in it, and a lot of variables. So, the number of things that can potentially be changed is enormous. So, one way you might get reassured is to try and get some other people to do the same thing and see whether—

So, in that regard, is there a lack of this type of study, this type of research, this type of modelling in Wales in particular, bearing in mind your point about it's very difficult to extrapolate from international studies, or even Scottish, English—?

Well, I discovered as a result of receiving an early draft of the minutes—no, the agenda—for this meeting that some of my colleagues were doing something like this. So, one might be able to give them a kick or two. 

And I think we'll be hearing from them soon as well. So, that's something that'll add to our deliberations, I'm sure. 

Okay. Well, can I thank you very much for your evidence this morning, Professor Foreman-Peck? We're very grateful to you for attending and for sharing the evidence with us. Diolch yn fawr iawn.

Thank you. Goodbye.

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


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


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.

We'll now move into private session. So, I propose, in accordance with Standing Order 17.42(vi), that the committee resolves to exclude the public from the remainder of the meeting. Are all Members content? Yes. 

Diolch yn fawr iawn. Ocê. Awn ni fewn i sesiwn breifat.

Thank you very much. We'll move into private session.

We'll move into private session.

Derbyniwyd y cynnig.

Daeth rhan gyhoeddus y cyfarfod i ben am 10:24.

Motion agreed.

The public part of the meeting ended at 10:24.