APPENDIX 1
THE STRENGTHS
AND WEAKNESSES
OF THE
BARNETT FORMULA
(1) It is in the context of the principles set
out in paragraph 11 of our evidence that we will comment
on the Barnett formula.
(2) The Barnett formula suffers from a number
of significant weaknesses as a mechanism for allocating resources
to Wales and should be replaced. The origins of the Barnett formula
are well understood: the formula was intended as a short term
expedient introduced in the late 1970s as a means of allocating
certain monies related to expenditure on functions devolved to
Scotland, Northern Ireland and Wales. Lord Barnett himself has
on numerous occasions in the House of Lords drawn attention to
the fact that the formula was intended to be a short term, stop-gap
measure. In practice it has remained in place with little alteration
for over 30 years.
WEAKNESSES OF
THE BARNETT
FORMULA.
(3) The Barnett formula has a number of weaknesses
as well as one or two possible advantages. The first three weaknesses
noted in this appendix (paragraphs 4 to 6) arise from faulty
implementation whilst the remainder (paragraphs 7 to 14)
are fundamental criticisms of the formula.
Arbitrary setting of base line expenditures in
1978
(4) Base line expenditures to which consequential
increases were cumulatively added were themselves based on the
expenditure patterns in existence prior to 1978: indeed in the
case of Scotland allocations were partly related to the Goschen
formula established in the late nineteenth century. Furthermore
although the Treasury undertook a needs assessment exercise in
the late 1970s the outcome was not used to alter the base line
allocations used in the Barnett formula. This is in spite of the
fact that the Treasury's assessment indicating that Wales was
receiving less than its needs based share of funds and Scotland
was receiving more.[3]
The conclusion of the Treasury study was that identifiable expenditure
on devolved services in Wales was 106 per cent of the corresponding
level for England whilst the needs estimate was 109 per cent.
This shortfall of 2.8 per cent was effectively locked into
Barnett increases and has persisted for the last 30 years. An
attempt was made by the Treasury in the early 1980s to resurrect
the idea of a needs assessment in order to reduce the allocation
of funds to Scotland and Northern Ireland (but not significantly
to Wales) but was abandoned.
Inaccurate population ratios used from 1978 to
1997
(5) The population ratios used in the application
of the Barnett formula did not accurately reflect the respective
populations of the four countries of the UK and in practice favoured
Scotland and disadvantaged Wales. The "consequentials"
were allocated to England, Scotland and Wales in the proportion
85: 10: 5 whereas the actual population resident in the three
countries would have suggested 85.3: 9.6: 5.1. Thus the initial
distribution of funding when the Barnett formula was implemented
and the population basis for allocating increases were flawed
from the outset. The difference may appear small but the cumulative
effect over the years prior to 1997 meant that annual increases
for Wales were approximately 2 per cent lower than justified on
a relative population basis. Until 1997 the population ratios
were not, with one exception in 1992, adjusted to reflect changes
in actual populations and this arguably further favoured Scotland,
where the population was declining, and penalised Wales where
the population was increasing broadly in line with that of England.
(Between 1976 and 2008 the population of England grew
by 11.0 per cent, Wales by 8.1 per cent and Scotland
declined by 0.9 per cent). Population ratios have been adjusted
annually since 1997 but the way that Wales lost out in the
intervening period of almost 20 years has been locked in to the
consequent baseline to which subsequent annual Barnett Formula
increases have been added.
Application of the Formula is not subject to independent
audit
(6) The application of the formula is not subject
to independent audit despite there appearing to be instances of
incorrect and inconsistent application of the formula. Examples
of this range from relatively small areas of expenditure to more
material ones and examples are set out in Appendix 3.
A mechanistic rather than needs based formula
(7) A central weakness of the formula is that
it is a mechanistic formula based on relative population levels
and does not take account of relative need except to the extent
that such needs were reflected in the baseline expenditures in
existence 30 years ago. (The authors of the Treasury needs assessment
in 1979 were careful not to claim that the expenditure levels
in existence then were related to need). There is significant
evidence that relative needs have changed during the intervening
period if measures such as relative GDP per capita or relative
expenditure per capita on social protection are used as proxies
for relative need. In the case of relative GVA per capita this
has, in the case of Wales, declined from 88 per cent of the
UK average in the late 1970s to 75 per cent in 2007 but
this has had no influence on the Barnett formula determined increases
to the block grant.
An arbitrary convergence formula
(8) Another objection to the formula is that
it is a convergence formula: it is not passive. If the formula
is applied rigorously, as it has been since 1997-8, then devolved,
identifiable expenditure per capita will converge to the same
level across the four countries of the UK. Expenditure within
the four countries of the UK would not converge to a uniform
level because different needs based formulae are used to allocate
such intra-country monies. Changes to other identifiable public
expenditure programmes and in particular Social Protection are
not subject to the Barnett formula and do not therefore suffer
such convergence.
(9) In the absence of needs assessment there
is no cogent case for converging per capita public expenditure
across the four countries of the UK. In the case of Wales the
deterioration over the last 30 years in relative GVA per capita
would suggest that identifiable expenditure should have been increased
in relative terms.
(10) As has been noted the Barnett Formula has,
with the exception of EU funding of Objective 1, been applied
with rigour to Wales since 1998. To test whether or not convergence
is happening in the case of Wales, data was analysed from the
Treasury's Public Expenditure Statistical Analyses reports (PESA
2008, Table 9.4; PESA 2005 Table 8.4) which set out growth
in total identifiable public expenditure per capita by country
in real terms between 1999-2000 and 2006-07. Growth
for England, Wales and Scotland was 33.03 per cent, 29.96 and
32.86 per cent, respectively. The lower growth in Wales cannot
be attributed to expenditure on social protection which represents
over 83 per cent of identifiable expenditure on non-devolved
services (PESA 2008, Table 9.17) which remained constant in relative
terms (~114 per cent of UK average) over the period. If identifiable
public expenditure per capita in Wales had grown at the same rate
as in England then it would have been £700 million higher
in 2006-07 after allowing for the above Barnett EU funding
(£128 million in 2006-07 compared with zero in
1999-2000). The impact of convergence on Wales will build up cumulatively
if the Barnett Formula continues to be employed in determining
changes to the block grant.
Lack of Transparency
(11) The absence of convergence in practice in
the first 18 years of the formula is a perverse example of the
lack of transparency in the operation of the formula. Students
of the formula have found it extremely difficult to explain why
the convergence inherent in the formula had not been noted in
practice. This is particularly striking because for most of the
period during which the formula has been in operation the changes
in public expenditure used to determine changes in the block grant
have been nominal changes which should have accelerated
convergence. It is assumed that there have been many examples
of the so-called by-passing of Barnett in the period 1979 to
1997 although academics and others struggle to find definitive
evidence. It may be fair to surmise that the Barnett formula did
not receive much government or public attention between 1979 and
1997 although recently released papers from the then Scottish
Office show that the Treasury attempted unsuccessfully in the
early 1980s to reduce the block grant to Scotland and Northern
Ireland but not, significantly, to Wales. It was thus possible
for governments of the day to deviate from the formula without
much public attention being paid to such practices.
(12) This lack of convergence was, of course,
prior to the establishment of the Scottish Parliament and the
National Assembly for Wales in 1999. It is also the case that
in 1997 the incoming UK Government stated that it would apply
the Barnett formula rigorously and this appears to be the case
with a few, well documented exceptions. These include the provision
of most of the EU sourced Objective 1 and Convergence funds
to Wales being allocated over and above changes to the block grant
determined by the Barnett formula. The UK Government also publishes
a Statement of Funding Policy to coincide with each spending review
and this casts additional light on the operation of the formula.
However there appears to be no adequate means for challenging
what appear often to be arbitrary decisions regarding comparability
factors used in the formula (see Appendix 3).
(13) It is inevitable that many adjustments will
be made to planned and actual expenditure over the life of the
Spending Review cycle. These changes will have an impact on the
Barnett consequentials but such changes are made by the Treasury
and it appears that there are cases where that department is in
practice "judge and jury" without reference to the Assembly
Government (see paragraph 18 of this appendix for a material
example).
A formula driven by the public expenditure priorities
of England
(14) The Treasury remains the driving force and
key influencer in all cases of budgets including public expenditure
plans. The allocation of funds to the devolved administrations
is a consequence of spending decisions agreed between the
Treasury and the various departments of state in London. In particular,
changes to expenditure on education and health in England
represent approximately 70 per cent of the change to the
block grant but in the case of Wales these programmes enjoy a
high degree of devolution (100 per cent and 99.3 per
cent, respectively). Thus spending decisions for England made
by the Department of Children, Schools and Families and the Department
of Health have a major influence on changes to the block grant
to Wales despite the fact that policy-making in these portfolios
is wholly devolved. The Assembly Government has minimal influence
on policy making and, in particular, budgeting, by these Whitehall
departments. Thus changes to the block grant are determined by
departments of state which have little or no financial remit in
the territories of the devolved parliaments and assemblies. Indeed
with the exception of spending on social protection, there is
a surprisingly small proportion of expenditure by UK government
departments in Wales which have some responsibility for expenditure
in Wales on non-devolved, identifiable expenditure. With the exception
of social protection, £1.5 billion only out of total
identifiable expenditure of £24.2 billion in 2007-08 corresponded
to identifiable, non-devolved expenditure. To a large extent a
number of UK departments of state are, from a financial viewpoint,
departments for England only and yet changes to their budgets
have a major impact on changes to the block grant directly impacting
Wales and its citizens with little democratic input. Such a "rough
and ready" approach may have made sense as a short term expedient
but it is not a satisfactory basis for the longer term which in
the case of the Barnett Formula has now been in force for 30 years.
A formula that takes no account of the massive
structural changes in Wales since 1979
(15) Since the Barnett formula was introduced
there have been massive structural changes in the Welsh economy
such as the closure of the coal mining industry which employed
40,000 in 1979 and the run down of the steel industry
where employment has declined from 51,800 in 1979 to
approximately 10,000 today. Because these industries were
concentrated in small areas of Wales they have had a devastating,
knock-on social effect. These changes have proportionately been
far greater in Wales than in England but the Barnett formula takes
no account of this.
An outdated formula whose operation does not reflect
devolution and the existence of the National Assembly
(16) The Barnett formula pre-dates the establishment
of the National Assembly by 20 years but the approach remains
the same and does not reflect the changes constitutional position
of Wales. The Assembly Government has minimal influence on the
funding of its activities and on the operation of the formula.
The Joint Ministerial Committee has no formal role in the resolution
of any disputes and the Secretary of State for Wales is in a weak
negotiating position as a member of the UK cabinet which is dominated
by the interests of England.
ADVANTAGES OF
THE BARNETT
FORMULA
(17) There are a number of arguments advanced
in favour of the Barnett formula:
The formula reduces the need for negotiation with
the Wales Office and the Assembly Government
(18) It is true that the formula reduces the
need for the Treasury to negotiate with departments representing
the devolved administrations. The counter argument is that there
is a corresponding democratic deficit under the current arrangements.
The Treasury is "judge and jury" in its own case with
little scope in reality for challenge. If the Assembly Government
and its members are to be empowered to represent the electorate
of Wales they must be able to negotiate with the UK Government
in such a crucial area as funding.
The block grant permits the Assembly Government
to have a high level of spending discretion
(19) An advantage of a block grant (as distinct
from the Barnett formula per se) is that once the size
of the block grant is determined it is to a significant extent
a matter for the Assembly Government as to how the money is spent.
It is possible that a detailed needs assessment approach to determining
the block grant could undermine that freedom.
The formula leads to predictable outcomes
(20) Defenders of the formula argue that outcomes
on a year to year basis are predictable which facilitates the
annual budget planning of the Assembly Government. Recent evidence
tends to undermine this claim. In the run up to the 2007 Comprehensive
Spending Review (CSR) the UK Government made a number of forward
spending commitments in England including health and education.
From those commitments it was possible to estimate the Barnett
consequential for Wales for the three year CSR period: this indicated
an average annual increase in real terms of approximately 2.7 per
cent.[4]
When the CSR was published this gave an annual increase of 2.4 per
cent. However the baseline expenditure on which the growth was
based had been reduced due to postponement of certain health expenditures
in England during the prior spending period. As a result the year
on year real increase in the block grant was 1.8 per cent.
It was clear from the comments of the Assembly's Finance Minister
that this was unexpected as well as unwelcome.
Operation of the formula offsets Wales's weak
negotiating position
(21) Past evidence suggests that Wales has suffered
compared with Scotland in the allocation of funds due to its weak
negotiating position. One advantage of the current arrangement
is that the Department of Children, Schools and Families and the
Department of Health are powerful departments within Whitehall
and Wales may indirectly benefit from such strength given the
correlation between increases in spending in those departments
and changes to the block grant. This may be true but it is hardly
satisfactory that a democratically elected Assembly Government
is not in a position to negotiate the funding settlement for Wales.
3 HM Treasury Needs Assessment 1979. Back
4
Institute of Welsh Affairs, Agenda. Winter 2007. Back
|