Memorandum by Economic Research Institute
of Northern Ireland
1. The Barnett Formula is named after Joel
Barnett (now Lord Barnett) who was Chief Secretary to the Treasury
when it was introduced in 1978.
The use of a formula for allocating at least some expenditure
among the territories (now called countries) of the UK goes back
to 1888 when Chancellor George Goschen in preparation for
Irish Home Rule introduced a set of proportions for allocating
resources between England and Wales, Scotland and Ireland in the
ratio of 80:11:9. This formula persisted in Scotland well into
the 1950s whereas arrangements for Ireland went a different way
with partition and devolved government in the North.
2. This short paper sets out the formalities
of the Barnett Formula as they currently apply, with particular
reference to Northern Ireland. It also gives some insight to how
the formula has worked in practice. The detailed questions posed
in the call for evidence are addressed in an annex.
3. The most common misconception about the
Barnett Formula is that it determines the total allocation of
public expenditure to Scotland, Wales and Northern Ireland. This
is not the case. The formula is a mechanism that adjusts the public
expenditure allocations at the margins. Moreover, it applies only
to parts of public expenditure. It does not, for instance, apply
to demand-led expenditure such as social security benefits which
are funded on a need or claimant basis.
4. The three key elements of the Formula
(a) changes in expenditure on services in England,
England and Wales or Great Britain, depending on the coverage
of the expenditure considered;
(b) the degree to which the English et al
services have counterparts in the devolved administrations. This
is called the "comparability proportions"; and
(c) each country's population as a proportion
of the population of England, England and Wales or Great Britain
depending on the coverage of the expenditure being considered.
The outcome of the formula is the product of
(a), (b) and (c) and is know as the "consequential".
This is the amount of additional spending made available to the
EXAMPLE: Suppose expenditure on an English service
increases by £100 million. The service is 100 per
cent comparable in Northern Ireland and Northern Ireland's population
relative to England is 3.4 per cent. Then Northern Ireland's
consequential is £100m x 1.0 x 0.034 = £3.4 million.
As a rough rule of thumb a 1 per cent increase
in a comparable service in England would provide enough extra
resources to fund a ¾ per cent increase in Northern Ireland.
5. Population proportions are one of the
more objective elements of the formula. There has been an erratic
history of updating these percentages. Up until 1992 the
mid-1976 population levels were used. In 1992 there
was a one-off adjustment but from 1997 onwards the latest
mid-year population estimates have been used. The failure to use
up-to-date estimates was an advantage to Scotland where in the
relevant period population proportion was actually declining.
In the 2007 Spending Review the population proportions for
Northern Ireland were 3.43 per cent relative to England,
3.24 per cent relative to England and Wales and 2.96 per
cent relative to Great Britain.
6. Comparabilities are a key element of
the formula and in some instances are open to considerable interpretation.
Comparabilities are calculated as a weighted average of expenditures
by the relevant Whitehall department. Up to CSR 2007 the
basis of expenditures were the sub-programmes operated by the
department. In CSR 2007 these were replaced by "programme
objects". Comparability proportions for each devolved administration
are estimated for each of these programme objectives and then
a weighted average is constructed for the entire department using
the baseline expenditures for each programmes objective in the
year immediately preceding the CSR.
7. Two problems arise with this approach.
First, a weighted average of a Whitehall department's expenditure
may not be a good guide to actual expenditures in a devolved administration.
Sometimes the latter will gain and sometimes lose from this procedure.
This is known as "taking the rough with the smooth".
8. Second, where administrative arrangements
for delivering services differ substantially, estimating a consequential
can be problematic. The classic example is local authority delivery
of services in England that are delivered by central government
in Northern Ireland. In England these services are part financed
by central government grants and part by the Council Tax and the
authority's share of the uniform business rate. Calculating a
consequential on the total spend would give Northern Ireland an
advantage by relieving local ratepayers of having to make a contribution.
On the other hand a consequential based on aggregate external
finance (mainstream grants to local authorities in England) may
fall well short of actual expenditure in Northern Ireland.
9. A further problem is "departmental
unallocated provisions" or in simple terms the reserves UK
departments are encouraged to create against unexpected expenditure
demands. The convention is that consequentials for these provisions
are calculated on the assumption that they mirror the weighted
average comparability of the department. In principle this is
fine so long as when these resources actually are spent they follow
this pattern. Otherwise the reserve may or may not end up in areas
where comparability is significantly higher or lower than the
10. A mathematical feature of the Barnett
Formula is that it should, other things being equal, tend over
time to converge per capita spending on comparable services in
the devolved administration towards the English per capita figure.
This is known as the "Barnett Squeeze". The phenomenon
arises because the formula gives Scotland, Wales and Northern
Ireland additions equal in per capita terms to those in England
(this is another way of saying these administrations get their
population proportion relative to England of any increase). But
it is generally the case that existing per capita expenditure
on such services is greater than these marginal additions so that
the average per capita spends will converge.
11. Theoretically the convergence phenomenon
should be faster:
(a) the greater the initial per capita lead in
the devolved administration; and
(b) the greater the increase in expenditure on
comparable services in England.
Since the formula is entirely symmetric falls
in English comparable expenditures should widen the per capital
12. Empirical evidence for the "Barnett
Squeeze" is limited partly for data reasons and partly because
the convergence is likely to be slow, so that other changes to
expenditure not dependent on the formula can cloud the issue.
The usual data source used is "identifiable public expenditure"
which is published in the annual Public Expenditure Statistical
Analysis (PESA) which accompanies the Budget. Identifiable public
expenditure is expenditure identified from administrative records
as being in or on behalf of the devolved territory (country).
Settling just what is identifiable expenditure in practice is
an issue that at the margins provokes considerable debate, particularly
The current identifiable expenditure ratios for the devolved administration
relative to the UK set at 100 are:
So over this period at least Northern Ireland and Wales appear
to have been squeezed slightly but Scotland has extended its lead.
However, not too much weight should be put on small movements
over a short run of years particularly when classification changes
and amendments to methodology are taking place.
13. One reason why theoretical Barnett Squeezes do not
materialise is that additional allocations are made to the devolved
administration outside the working of the formula. These are commonly
referred to as "bypasses".
14. Since the Treasury has invested heavily in the Barnett
Formula they are generally resistant to bypassing it and significant
departures are usually associated with either technically unavoidable
changes or highly political issues. All the devolved administrations
have benefited from bypasses at one time or another. Wales, for
example, got over £200 million additional cover for
Objective 1 EU Structure Funds programmes in the 1990s while
Northern Ireland was given additional funding to support the privatisation
of aircraft production and shipbuilding in the regions. Northern
Ireland also secured additional funding to cover the series of
Peace and Reconciliation Programmes launched by the EU in the
mid 1990s and some costs associated with implementing the Good
15. In recent years the Treasury has tried hard to keep
the devolved administration on a strict Barnett Formula diet and
has largely succeeded. Despite announcements of new packages of
support for devolution these usually turn out, on closer inspection,
to be rescheduling of expenditure or movements in non-cash items
16. The Barnett Formula is the centre piece of the rules
governing the funding of devolved administrations.
However, it is the interplay of the formula with these rules that
gives the process texture and allows interesting possibilities
17. Of particular interest is the rule that says "if
the UK Government makes a general cut to the budgets of UK departments
it is entitled to impose the same adjustments to the budgets of
the devolved administrations".
18. Alternatively the reductions in UK departments could
be fed through the Barnett Formula to give negative consequentials
(reductions) to the devolved administrations.
19. An interesting combination is to apply across the
board cuts to baselines including those in the devolved administrations
and then give Barnett consequentials on any allocations restored
to UK departments. Since the latter are based on population proportions
while baseline proportions are usually higher this is an indirect
way of cutting budgets for devolved administrations within the
20. Baselines do not enter into the Barnett Formula except
as weights in the calculation of the departmental weighted average
for comparabilities. However, baselines are very important. As
baselines stand at the moment they are historical constructs reflecting
a myriad of past changes, including changes from previous applications
of the Barnett Formula. A systematic revision of baselines requires
some form of Needs Assessment.
21. The basic idea of a Needs Analysis is to start with
a benchmark for expenditure in some policy area which in the UK
is usually expenditure in England. This expenditure is then associated
with a number of "objective factors" such as total population
or population structure for those receiving the services and this
gives an idea of the unit cost of the service "Objective"
in this sense means factors that cannot readily be adjusted by
policymakers. The pattern of objective factors in the devolved
authorities is then compared to the same factors in England to
give an idea of how much more or less it would take to deliver
the same service as in England in the circumstances of the devolved
administrations. Comparing this to actual expenditure shows whether
the devolved administration is over or under provided for that
22. This is the barest outline of the technique and in
practice Need Assessments are data heavy exercises fraught with
difficulty in matching expenditure data and properly identify
relevant factors. They work best where services are clearly linked
to population such as in education or health programmes but are
much less successful in areas such as economic programmes.
23. The only official Needs Assessment in the UK was
carried out in 1976 in preparation for devolution to Scotland.
Only a summary report was published in 1979. Since devolution
did not occur at that time this work faded from view, although
there were periodic updates carried out internally by the Treasury.
In 2001 the Northern Ireland Executive initiated a unilateral
update of Needs Assessment but that work was abandoned when Direct
24. Needs Assessment is often presented as an alternative
to the Barnett Formula but that can not be the case. The exercise
is too resource-intensive to be repeated annually and there are
concerns that as time goes on fundamental changes in the character
of services in one area of the UK as opposed to another progressively
render the needs assessment technique invalid.
25. Some commentators have argued that the approach adopted
by the Commonwealth Grant Commission in Australia for allocating
monies to the States could be adopted in the UK even though the
constitutional situation is rather different. However, this again
is an elaborate exercise and certainly not immune from political
26. The Barnett Formula has been operating in its basic
form for 30 years and has been incorporated into a comprehensive
set of funding rules for the devolved administrations. It is a
uniquely British approach to devolved financing with nothing similar
elsewhere in the world.
27. Critics of the formula generally focus on two issues.
The first is a fear that repeated use of the formula as the main
means of adjusting devolved budgets will lead to steady convergence
to English per capita expenditure on services. The other, which
is basically the same point, is that using a formula that only
takes account of population proportions is a poor way of capturing
28. One approach would be to replace the formula by one
which was the reciprocal of the relative outputs of the devolved
administration and the UK (or England). In Northern Ireland's
case the relative output (GVA) ratio is 80 per cent so this
formula would give Northern Ireland approximately 125 per
cent of comparable spending the in the UK. Since regional productivity
figures are notoriously unreliable this would not seem much of
29. Combining other factors with population in the formula
raises the question of what these should be and what relative
weight they should have. This in turn can lead down the path to
what statisticians call an index number problem.
30. Convergence is an inherent characteristic of the
formula but in practice has not been a critical issue. When devolution
began some commentators feared that the new administrations would
quickly run out of money as the formula took its toll. In fact
the reverse could be argued, that the devolved administrations,
at least initially, had rather too much money and made some unfortunate
spending decisions as a result.
31. Another aspect of the formula is not its impact on
the devolved administration but on England or at least the perception
among some English commentators that it gives the devolved administrations
too much. This, of course, is a misconception arising from confusing
the baseline expenditure in the devolved administrations with
changes in these baselines. Whether the devolved administrations
have too much of a share of public expenditure or too little is
not a question the Barnett Formula can answer.
32. Should the formula be abandoned or replaced with
some other funding mechanism? That is essentially a political
rather than a technical question but some points that need to
be kept in mind are:
The Barnett Formula does remove the need for detailed
negotiations with the Treasury on the minutiae of budgets in Spending
Reviewsa very big plus;
It offers some protection to any existing expenditure
advantage enjoyed by a devolved administration; and
Its workings alongside the other funding rules are
reasonably well understood so it offers a degree of stability
that a replacement might take a long time to deliver.
The formula was devised by Sir Leo Pliatsky and has always officially
been known as the funding formula. The term "Barnett Formula"
is attributable to Professor David Heald, a long time student
of devolved finance in the UK. Back
For example, J Cuthbert and M Cuthbert, "A Constructive Critique
of the Treasury's Country and Regional Analysis of Public Expenditure",
2005, is an interesting example of the debate. Back
Funding the Scottish Parliament, National Assembly for Wales and
Northern Ireland Assembly: Statement of Funding Policy, October
2007, HM Treasury. Back
See http://www.scotlandoffice.gov.uk/freedom-of-information/document.php?release=78&doc=179 ,for
a description of the technique used for an update of the 1979 study
for Scotland. Back