Examination of Witnesses (Questions 540
FRIDAY 20 MARCH 2009
Dr Eurfyl ap Gwilym, Professor James Foreman-Peck
and Dr Gillian Bristow
Thank you all for coming. You know what we are about. You have
heard that we have this inquiry by the Select Committee of the
House of Lords, and I think you have seen the terms of reference
of the Committee, so you know what we can look at and what we
are not supposed to look at. Can I say two housekeeping things?
First of all, this is a public hearing. As I understand it, the
BBC is taking a full sound video of the whole thing. Whether they
use it or not, I do not know. A full transcript is obviously going
to be taken. When the transcript is out, you will have an opportunity
to look at it and see whether you want to make any alterations
to it. If you want to make small alterations, fine; but if you
want to change the sense of what it is you said, I am afraid,
like Hansard, the answer will be "no". I wonder
if I can ask you some basic questions to start us off! What do
you understand was the object of the Barnett Formula? What was
the object of the exercise when it was first introduced? Do you
think its purpose has changed over time? It is a long time now
since Joel Barnettwell, he did not introduce his formulait
was not even called a formula for about a decade, but it is one
that now bears his name. Do you think it was designed to reduce
tensions arising from disparities in public spending per head
of population, or do you think that is the way it has developed,
and do you think it has been successfully responding to any of
Dr ap Gwilym: I imagine that the purpose is
rather clear. Bear in mind the timing of this as well; it was
in the run-up to the devolution referenda in 1979. There was already
funding, with funds going to Wales, Scotland and Northern Ireland
for devolved services. Clearly there needed to be some sort of
funding mechanism. Why those three departments did not have bilateral
negotiations with the Treasury the way other departments of state
have, I am not clear. I assume one of the advantages to the Treasury
was that there were three fewer sets of negotiationsremember
the old days, the Star Chamber and all thatthough that
is speculation! I do not think it is reasonable, by the way, but
it probably seemed reasonable from a London viewpoint that you
said, "Let us look at the changes in expenditure when planning
our programmes in England, and in the case of those programmes
where the spending on the programmes in Wales and Scotland is
devolved, let us change those by the same amount as in England
on a per capita basis." The interesting thing about
that, of course, is that they did not do the obvious thing, which
is that if spending is going up on a particular programme in England
by X per cent, that you do the same for Wales and for Scotland.
Instead of that, they said, "Let us take a monetary amount
per capita"and we will come, presumably, later
on to the question of convergence, which is, to my mind, intrinsically
built in to the way the Barnett Formula works. Nevertheless, it
was a way of saying, "Right, we are going to spend so much
extra on a programme in England and where it is devolved in those
two countries, and we will take the lead from the change in England
and that is how we will change the moneys coming to Wales and
to Scotland." That was the purpose. I think the "reducing
tensions" consideration, if it was there at all in those
days, comes back to the next question you have got on your list,
which is the convergence issue. I do not think it does reduce
tensions because the key issue there, is that if you look at the
disparity in identifiable public expenditure per capita
within England, you have a very wide range within England, and
that has nothing to do directly with Barnett at all. Barnett is
designed to allocate moneys to three of the four countries of
the United Kingdom and has nothing to do with intra-England allocations.
Indeed, if you level the identifiable public expenditure per
capita across the UK, England gets about 3 per cent morethat
is allso that would not sort out, for example, the north-east
of England problem. If you look at the disparities, the IPPR report
last year again, from Nuffield College, states that the huge additional
spend is in London, where it is about 28 per cent above the UK
average. Broadly, Barnett does what it says on the tin, as it
were, which is to allocate these moneys in a fairly easy way,
using a rather simplepeople might argue crudeformula
on a population basis. That is what it has done the whole time.
We may come back, Chairman, to questions about convergence and
why there has not been so much until the last decade, in a moment.
Dr Bristow: I would just echo that and say it
is probably quite well acknowledged now that the Barnett Formula
was a temporary, probably political expedient that became permanent.
It is something that was designed principally, as Eurfyl said,
in relation to the first proposals for devolution in the 1970s
and used to determine how public expenditure should be allocated
to Scotland principally, in the context for those preparations
for devolution which subsequently did not occur. I think Lord
Barnett himself has stated that he did not expect the formula
to last as long as it has, but clearly it has been something that
has worked. It has been practical and effective in that sense.
He told us on a number of occasions that it was a convenient way
of avoiding three different rows between parts of the UK and the
centre; and the convergence as far as he was concerned did not
enter into it at all. Interestingly, too, on convergence, we had
giving evidence to us this week Lord MacGregor, who was Chief
Secretary to the Treasury in the mid-eighties, and he said exactly
the same thing; that convergence was not part of the exercise
as far as certainly the ministers of the Treasury were concerned.
Whether the civil servants of the Treasury took a slightly different
view as to how this worked is another matter, but certainly as
far as the ministers were concerned, two or three of them said
it did not enter into their consciousness in the operation of
Professor Foreman-Peck: I do not have a great
deal to add about the intentions of ministers, which were obviously
opaque. One can infer what people intend from what they do, but
this is a pretty risky strategy. Certainly on the basis of what
you told us, civil servants are quite smart sometimes at getting
policies through despite ministers, and that is certainly consistent
with what I know in this case.
Q542 Lord Sewel:
Jim Callaghan always said the real purpose of Barnett was to enable
the Cabinet to have lunch! There might be an element of truth
in that! Let us talk about convergence. It is a property of the
Formulathere is no doubt about thatand we could
have debates about whether ministers realised it or did not realise
it. I am pretty sure that some people in the Treasury did realise
it, because it sticks out straight away as soon as you look at
it, that it has an inevitable convergence property, and you can
almost leave it at that. The interesting point is, though, as
I think was said earlier; if you look at Scotland the convergence
has not taken place. There has been a series of Formula by-passes,
partly a delay factor because of the Scottish falling population,
which delays the impact from one comprehensive spending review
to the other, so you are always a bit behind; but, clearly, there
were lots of wadges of public expenditure that were shifted to
Scotland without going through the Formulaand I am afraid
some of us were involved in those dirty deeds! It was basically
national negotiated wage settlements that were the cause of major
disruption in the progress in Scotland. That does not seem to
have happened in Wales. Wales is almost the example of how the
Barnett Formula is intended to operate where you do have a convergence
effect. Any explanation for that divergent experience?
Dr ap Gwilym: I think it is quite fascinating,
and we are going to come back later on maybe to talk about some
data and information where you can have lots of distinguished
academics, like these on my left, delving into this material for
a lot of time, trying to work out what exactly happened. I take
your point about national wage settlements and so on, the so-called
Barnett by-pass, though sometimes I think Barnett by-pass is a
convenient way of trying to explain away lack of convergence.
Even in the case of Wales, there was not that much convergence
until about 1997. Why I say 1997bear in mind that soon
after the Formula came into being, of course, there was a change
of government at the UK level, and then the Conservatives had
a long run until 1997. I get the impressionand it is only
an impressionyou will be better placed than myself to judgethat
Barnett was often more honoured in the breach. Even in the case
of Wales for a long period there was quite a skilful Secretary
of State, Nicholas Edwards, Lord Crickhowell, who was quite influential
and probably punched above his weight a little bit in Cabinet
as well. Since 1999, thoughfirst of all, bear in mind the
incoming Labour Government did say, "We are going to use
Barnett with some rigour." In the case of Wales there has
only been one exception to my knowledge, which is the EU funding
of Objective 1, and that effectively took the removal of the very
first First Minister in Wales to achieve. However, if you look
at the growth in total identifiable public expenditure in WalesI
say total not just the devolved, because again we have got problems
with datasince 1999 it has grown in real terms by 2006-07
by 30 per cent, whereas England grew by 33 per cent and Scotland
by 32.9 per cent; so Scotland is tracking England again but Wales
has gone the other way. I think I put in my written evidence that
in 2006-07, the last year for which we have reasonable public
expenditure data, the shortfall for Wales was £700 million.
If you take away the other elements of expenditure in Wales, you
come back to the conclusion that it is due to the operation of
the Barnett Formula, not necessarily only convergence of course,
because there are substantial sums where the consequential (if
I go to the jargon of Barnett) is zero, and where that money is
spent of courseit might be spent in Wales, in Scotland
or in England, the theory being that it is for the benefit of
the UK as a whole. I suspect most of that is spent in England,
not on a population basis, but a disproportionate amount, just
because the UK is a highly centralised state and many British
spending programmes take place in the south-east of England. The
£700 million has happened, and therefore we have experienced
convergence in Wales since 1999; but I think that you can partly
at least speculatively relate that to the incoming government
saying, "we are going to use Barnett with rigour".
Q543 Lord Sewel:
Is it also that actually getting round Barnett becomes more difficult
in the context of devolution?
Dr ap Gwilym: I think that is a very good point,
because it was much easier, to put it crudely, to do backroom
deals in the old days, whereas I think it is more difficult now
to do that. This, of course, does not explain Scotland, because
Scotland has not experienced that reduction compared with England
and Wales has. I think Northern Ireland has as well, by the way.
Professor Foreman-Peck: I think we can all agreedown
here anyway we all agree that Wales has been squeezed and Scotland
has not. We think that Wales has been squeezed because the Barnett
Formula has been applied in Wales's case. We do not really understand
what goes on in Scotland's case.
Q544 Lord Sewel:
Mr Swinney, the nationalist Finance Minister in Scotland, is very
much in favour of the Barnett Formula.
Dr ap Gwilym: I can well understand that he
To be fair, he did not actually say that!
Dr ap Gwilym: I suspectthey always relate,
if you ask them about it"give us fiscal independence
and then we can talk about Barnett", which is their way of
addressingor not addressing the question
Could not care less about Barnett; it did not really matter: fiscal
autonomy was what they wanted.
Dr ap Gwilym: That is right. Scotland hasI
hesitate to say this but it appears to have done rather well,
going back to the Goshen Formula in the 1890s. I leave the Scots
to argue their case. I do not feel obliged at all to defend them.
Dr Bristow: I would agree again with what was
said. Wales has certainly experienced the Barnett squeeze since
1999 in particular because the Formula has been applied more rigidly.
It is interesting in the case of the Barnett Plus Settlement that
Wales secured, in as much as it shows how, with devolution, the
whole issue around the settlements given to the devolved administrations
becomes more transparent and has become a bit more subject to
political scrutiny and debate. That perhaps raises issues around
how comfortably the Barnett Formula now sits and works within
the system where we have devolved governments.
Q547 Lord Sewel:
For the record, is convergence a good thing or a bad thing?
Dr ap Gwilym: It is a bad thing. It is very
clear. In any sensible societyand I think James has done
some work on this and maybe presented it in his evidence herethe
amount of public expenditure a citizen makes use of, public services,
is partly dependent on their own situation, whether they are young,
old, poor, sick, healthy, in work or out of work. To the extent
that it is unrealistic to expect to have a homogeneous distribution
of those characteristics across the United Kingdom, you would
not expect to have a level settlement of public expenditure per
capita. Quite clearly, again, if you look at all the statistics,
the UK is one of the most regionally unequal states within the
European Union. It is most unequal in fact if you use measures
such as GVA per capita. Therefore, you would expect logically
that you have got some poorer regions which need a greater consumption
of public services.
Professor Foreman-Peck: May I take issue with
Dr ap Gwilym's contention of convergence not being a good thing?
It is undoubtedly a good thing where Scotland and Northern Ireland
are concerned. What the question begs is how we know what the
appropriate devolved budget is for these authorities. That really
seems to be the nub of the problem. There are two ways you can
look at it, roughly. You can take an aggregate measure, because
it is an aggregate budget; or you can try and get some more sophisticated
needs analysis, paying particular attention to education and health.
If you take an aggregate measure, which is what I favour, the
obvious measure is disposable household income. This is not widely
focused on, compared to GVA per head, but it is much more appropriate
because income is a reflection of needs, in a way that gross value-added
is not. The disparities between regions and devolved administrations
are much less when you look at household disposable income; and
they become even less when you divide through by regional prices.
One of the reasons that Wales is such a pleasant place to live
is that it is considerably cheaper to live than anywhere else
in the UK, certainly once you have adjusted your budget or your
spending pattern to take into account of what is cheaper here
relative to elsewhere. If you take household disposable income,
the gap, as last measured, is about 10 per cent in nominal terms
between the UK average and Wales, maybe 11 per cent with a fair
wind; but there is about a 7 per cent lower price level in Wales
if you include housing6-7 per cent or 7-8 per cent. Once
you have taken that into account the disposable income difference
between Wales and the UK average is about 4 per cent. So if you
take 4 per cent as the inequality level, it is not a great deal.
That is why I submit that the inequities of the Barnett Formula
are not that great for Wales. But when you apply this to Scotland
and Northern Ireland and compare it to their receipts of public
expenditure, they clearly do far better on this aggregate needs
measure than England or Wales. For Northern Ireland and Scotland,
it seems to me that if you have got to get some form of equity
between England and Wales and these two administrations, convergence
is a good idea.
That is if you lump England and Wales together.
Professor Foreman-Peck: Wales is pretty close
to England in a number of respects, but the critical one is in
the level of public expenditure per head relative to income, and
it is particularly close to the north-east of England in that
respectnot perfectly close!
Dr Bristow: I do not think convergence sits
comfortably with a notion of territorial justice or equity, which
is an important principle. Public spending levels ought to be
such that there is opportunity provided for the different regions
to offer equal levels of public service provision if they wish.
That is a fundamental principle.
Dr ap Gwilym: I do not want to get into too
lengthy a discussion with my good friend James nowthere
will be another opportunitybut one thing is the provision
of public services, but of course I do have a concern about things
like gross value-added. One of the reasons why the cost of housing
in Wales is on average lower than other parts of UK is because
Wales is performing very weakly economically. The fact that relative
GVA per capita is down to about 75 per cent of the UK average
is indicative of this. Therefore, the danger is that you have
Wales with a low cost of living, low housing, low incomes, and
you say that they are not doing too badly overall. I must say
that, having spent a lot of my career working in the City of London,
I was very happy to pay high house prices in Central London, because
that was a measure also of the high GVAs we were generating. I
think gross household income and looking at the cost of delivering
public services is a rather narrow view of looking at this. We
would like to see Wales being in a position where it is going
to generate more wealth for itself, which is why we talk about
remediation of weakness, not just delivery of public services.
That is quite a large field we could explore maybe some other
If you want to produce a fair and equitable allocation system,
to use a fairly neutral phrase, how do you do it? You would have
to have some kind of needs assessment involvement in it. I am
not absolutely clear in my own mind as to precisely where it comes
in and how it comes in, but to have one that is certainly divorced
from any kind of assessment of what the needs of a particular
country are seems to me to be very difficult to support.
Dr ap Gwilym: It is, and if you focus on spendingI
think roughly 70 per cent of the block grant that comes to the
Welsh Assembly Government is focused on education and health.
Even if Wales pursued a somewhat different set of policies say
from England or from Scotland, that will be the broad figure,
so with slightly more than two-thirds being spent on education
and health, one would hope one could come up with some measures
in those areas of need and relative need. In fact, as you know,
at various stages attempts have been made to do that. The Treasury
did attempt this not just in the well-publicised case in about
1976-77 but we have seen from the Freedom of Information Act papers
that came out last year that they had another shot at it in about
1984 and retired bruised from their confrontation with the Scottish
Office. There are mechanisms for assessing needs. It is complex,
and one needs to recognise that. That is the argument for having
expert advice, but also having it done in a very transparent way,
so that when people see these issues you can have an intelligent
debate about them. At the moment what we have is essentially a
very crude formula. The virtue of it is that it is quick and simple,
but it has many drawbacks as well.
Professor Foreman-Peck: In my submission to
the Committee I suggested a way of taking into account needs at
the aggregate level by looking at the allocation or consumption
of state benefits by income group across the United Kingdom. You
can find that there is a very close inverse statistical relationship
between income group and consumption of state benefits. That partly
depends on the composition of the income group. But if you wanted
a simple way of allocating money, you could just apply that inverse
relationship that applies across income groups in the United Kingdom.
You could apply that to the devolved administrations using average
income groups of those administrations. That is the calculation
that I did in my paper.
Q550 Lord Moser:
I found your paper extremely interesting, as a statistician. Once
we decide or are advised to go beyond the Barnett Formula towards
needs assessment in a serious way, some people then picture an
enormously complicated approach with hundreds of indicators. It
can be done, as you knowand then there are weighting issues,
et cetera. Obviously, it is rather off-putting. Some people
favour getting it down to three or four indicators measuring maybe
something very economic, something environmental, et cetera,
and there are still some weighting problems but more manageable.
Some people, including yourself in your paper, Professor Foreman-Peck,
look for a proxynot just population but, say, GNP, GVP
inverse, or one of the papers we had suggested a poverty measurement.
There are social security measurements. If I understand you right,
you favour disposable household income. We have to think of the
politics of all this and on the assumption that it is done by
somebody other than the Treasury, because the Treasury probably
would never do it. Could you talk a bit about the pros and cons
of the multiple approach versus the proxy approach and your favoured
examples for each? I would like to hear a little bit more whyI
have not quite understood why in your paper disposable household
income is superior to the conventional GNP measure. Any help on
this would be appreciated.
Professor Foreman-Peck: To take the last point
first, GVA per head is the critical indicator that many people
favour. The reason why you get such wide discrepancies between
regions in the UK is because participation rates vary very substantially.
Wales has about a 4.5 per cent lower participation rate in the
labour force than the UK average. That means, for various reasons,
a smaller proportion of the population is working. GVA per head
of population is not just a measure of productivity; you have
to multiply participation in the work force times productivity
of the work force to get GVA per head of population. That really
is the critical issue.
Q551 Lord Moser:
Could you do that?
Professor Foreman-Peck: Certainly you could
do that, yes; you could control for whether people work or not.
The second point is: what exactly has productivity got to do with
what we are talking about here? Productivity is to do with industry
and making things. Here, we are talking about, as I understand
it, the equitable provision of public services. I would have thought
most people would have said that those with low incomes are more
likely to have unsatisfied needs than people with high income.
That is the basis of my proposal to use household income as a
Q552 Lord Moser:
Disposable household income.
Professor Foreman-Peck: Yes, after tax and after
support. Arguably, the weakness of the numbers I come up with
derive from the cross-UK consumption of state benefits in kind
with income, because that is the only thing I could put my hands
on at short notice. I think you can probably get a more precise
indicator than that, given more time. I think social security
is not a good aggregate index because that is already covered.
In a sense, households receive social security, or social protection
or family income supplement effectively from central government.
It is largely annually managed expenditure. So you would be counting
the outlays twice in a sense. I am not convinced that that is
a good idea. My objectionand it is not a particularly strong
objectionI can see there are arguments for doing it, for
looking at disaggregated measures of education and health and
so onis that when you devolve authority you allow authorities,
administrations, to decide on how they are going to spend on education
and health. So if you give them money based on your calculation
of how much it is reasonable to spend on education and health,
you are in a way double-guessing them. I did a little experiment
in something that the IWA published a couple of years ago, called
Time to Deliver where I compared health spending across
regions and education spending and used standardised mortality
ratios for health spending and free school meals for education
spending; and you get very, very close fits. Wales looks fairly
similar to the English regions in that respect. In fact it spends
slightly less on education, especially about three or four years
ago, and spends rather more on health. Scotland spends more on
education and health than England on both those criteria.
Q553 Lord Moser:
How would you feel aboutChairman, if I can go on for a
momentthe other sort of approach, 4, 5 or 6 indicators,
with suitable weighting? The weighting would probably vary between
regions. Well, it would certainly vary between regions because
of your 70 per cent that does not fit the other regions.
Professor Foreman-Peck: I have no fundamental
objection to it. My understanding is that when you look at the
way these formulas operate in practice, particularly in local
government, they become areas for bargaining. There is a very
strong pressure to increase the number of indicators and change
the weights for special circumstances. In other words, it is a
Pandora's Box and extremely boring as well! I favour simplicity
on the grounds of sanity.
Q554 Lord Moser:
On the assumption that Barnett is free of bargaining?
Professor Foreman-Peck: Barnett is not going
to last for ever, and I think that a simple needs-based formula
is better than a complicated needs-based formula.
You could go half-way between the two, could you not? The suggestion
Lord Moser is making is for four to six variables, which should
not be too difficult to categorise, and on that basis you could
arrive at a better assessment of needs and fairness than by just
Dr ap Gwilym: Without being facetious, I am
not sure if you are familiar with the Lottery Fund, the way they
allocate moneys. They have a formula that is related to GDP and
to a deprivation index. I am not an expert on that at all, but
it might be well worth a look at.
Q556 Lord Sewel:
Should you put in the cost of providing services as factors? I
have an interest here because I live north of the Highland line.
Professor Foreman-Peck: The cost of providing
services is arguably determined by the system, and so if you have
a formula that takes into account the cost of providing services,
then it is likely to affect the cost of providing the servicesand
the cost-plus contracts in defence are the classic example. If
you do want to take into account costs, you need to have something
Q557 Lord Sewel:
There is a problem, is there not, in, say, education provision
in the Islands of Scotland? There is a cost driver there.
Professor Foreman-Peck: Yes, and this is why
many people think that one of the indicators, which Lord Moser
would favour, would be population density as a contributor.
Except density has problems as well as advantages, does it not?
Dr Bristow: I would make two points. One is
that there is never going to be a system that is perfect and entirely
free of political bargaining, some degree of trade-offs and compromise,
and that is the reality in a sense. Secondly the key then is finding
a situation, or a settled compromise that minimises potential
trade-offs and conflicts that might arise. What is critical is
devising a system that is transparent and accountable; and we
therefore have to understand what we are trying to achieve first
and foremost and then design a needs-based system around that.
It is important to have a debate about the key objectives you
are trying to achieve, and then determine the indicators on the
basis of that. There are a number of different options you might
choose. One might be to have a fairly comprehensive annual rigorous
assessment of needs undertaken, and therefore the desire that
you try to achieve a unanimous verdict on the settlement as a
result of that, and if you cannot reach some sort of unanimous
settlement, then you go to a default option where you use simpler,
cruder indicators as your default mechanism for allocated spending.
That is certainly something others have suggested as a practical
solution and way forward.
Q559 Earl of Mar and Kellie:
I would like to ask about the absolute need for reform. From what
I have seen and heard, the workings of the Barnett Formula are
probably best demonstrated in Wales. I suspect they are distorted
by other factors in Scotland and Northern Ireland. Therefore,
I wonder whether a replacement will in fact ultimately be any
better; and would any of these suggestions so far be fairer?
Dr ap Gwilym: I am not sure I would agree with
you when you say "best works in Wales" because what
do you mean by "best" and what are your criteria? As
you appreciate, we would contend that it is very unfair to Wales,
and therefore it does not work well for Wales; and therefore that
is why we want it replaced. In my submission, I gave a whole series
of criticisms of the Formula. One recognises of course that replacing
it will be difficult and complex, but I think we need to replace
it. The other issue is that whatever system we have, it needs
to be open and transparent, and then you will have the debate.
You will have disagreements but that is at the heart of democracy,
that people debate and disagree and then try to reach a compromise.
At the moment we do not get that. The other thing of courseI
do not want to be party-political, but since 1999 when we have
had the Assembly here in Wales, you have had the Government in
London and the Government in Wales, even allowing for the couple
of coalitions in Wales, being led by the same party. It might
be somewhat different if you have a different party in London
and in Cardiff; that might open up the differences a little bit
more. At the moment you have not had a rigorous debate between
those two organisations as far as I can see, about these issues.
For obvious reasons, it is rather difficult, being of the same
party. I suspect if you had a Conservative government in London
and a Labour-led government in Wales there could be more friction,
and hopefully out of that friction you might get a little bit
more light on these mechanisms. I am afraid the whole working
of the Formula is probably exemplified most clearly in Wales,
and I think it exemplifies the weaknesses of the Barnett Formula