The Barnett Formula - Select Committee on the Barnett Formula Contents


Supplementary memorandum by Professor Iain McLean

  As I promised during my oral evidence, I have pleasure in sending you this supplementary note. Part A of this note has been agreed with my colleague on the (Muscatelli) Independent Expert Group reporting to the Calman Commission, Professor D N F Bell. The calculations in Part B were done by my coauthor Katie Schmuecker of ippr north.

A.  THE BARNETT FORMULA AND CONVERGENCE

  1.  Three members of your Committee, all with experience of government, expressed opposing views as to whether the Formula was designed to converge towards equal public expenditure per head in all four nations of the UK; and whether, regardless of intention, it actually has led to convergence. I hope the following may help.

2.  All researchers in the field[1] believe that the intention to bring about convergence (perhaps until "excess" spending per head in Scotland and maybe Northern Ireland had been brought down to the level implied by the 1979 Needs Assessment) was in the minds of some Treasury officials involved in the design and early implementation of the Formula. There is substantial oral evidence on this point, but little written evidence has surfaced.

  3.  However, the territorial departments in (at least) Scotland and Northern Ireland were very reluctant partners. The service-by-service details underlying the Needs Assessment were released to me in 2005 under a Freedom of Information request. They make it clear that the Scottish Office and HMT did not reach agreement as to the relative needs of Scotland.

  4.  The Treasury told me, in the letter accompanying that release, that there had been no subsequent "interdepartmental" needs assessment. However, a further FOI release at the insistence of the Information Commissioner in summer 2008 has shown that the Treasury conducted an in-house update of the Needs Assessment in 1984. Again, its conclusions were vigorously resisted by officials in the Scottish Office. It appears that pressure from the Secretary of State for Scotland led Treasury Ministers to agree to pursue the matter no further. That is my précis of the files released to the National Archives of Scotland at West Register House, Edinburgh, read in June 2008.[2]

  5.  Therefore I believe that I was correct to state in evidence that, whatever may have been the intentions of Treasury Ministers (of both parties) over the period in question, it was the intention of Treasury officials to bring about some convergence to what the Needs Assessment had persuaded them were relative needs.

  6.  As to whether the Formula actually (regardless of intentions) brings about convergence, there are two relevant factors. One is whether public expenditure is planned on a volume or a cash basis; the other is the change in relative population among the four nations of the UK.

  7.  In a world of constant relative populations, the mathematics of the formula entail that whenever expenditure in England increases, the Barnett block grant to the three Devolved Administrations (DAs) increases by a smaller amount per head. Some experts have described this as a differential equation system, a label I used in my oral evidence. However, others disagree, and there is no need to argue the point in this note, since there are two large and opposing effects: the convergence effect of the Formula; and the declining relative population of Scotland. These come close to cancelling out, as will be shown below.

  8.  Public expenditure was planned on a volume basis from the inception of the Formula until 1982. Barnett consequentials (ie changes in the block grant to each DA) are a function of increases in corresponding planned expenditure in England. If expenditure per head in volume terms was constant, then block grant per head, in volume terms, would also remain constant.

  9.  Since 1982, with public expenditure planned on a cash basis, there has always been an annual cash (ie nominal) increase in planned expenditure in England. In a world of constant relative populations, this should have led to convergence towards equal cash per head to spend on devolved functions in each of the four nations of the UK. The convergence would not have been linear (the downward slope would gradually flatten out); but one would expect it to have been close to complete by now.

  10.  The above ignores changes in relative population. However the relative population of Scotland has been declining since the inception of the Formula (and indeed for longer than that, back into the days of the Goschen Proportion). Until 1992 population relativities for England: Scotland: Wales were fixed at 85:10:5, ratios which were over-generous to Scotland and insufficiently generous to Wales. In 1992 Chief Secretary Portillo first rebased the consequentials on up-to-date relative populations, but did not change the then-current baseline. Since then population relativities have been updated from time to time. The current practice is that the relativities are restated, from the latest available ONS population data, with each new edition of HMT's Statement of Funding Policy, which is available to members of your Committee. These relativities then apply for the period of the ensuing Comprehensive Spending Review (CSR).

  11.  Scotland therefore gets favourable treatment per head of population for two reasons: (a) the historic baseline was not adjusted for the relative populations of Scotland and England in 1992; (b) if its relative population is declining through the period of a CSR, as has been the case for each CSR since 1998, this is not reflected in Barnett consequentials during the lifetime of that CSR.

  12.  The graphs from McLean et al, Fair Shares... (ippr 2008), which were discussed during my oral evidence, show that since 2001-02 the convergence effect and the population effect for Scotland have roughly cancelled out, so that resources per head available to the Scottish Government to spend on devolved services are roughly constant. Because of deficiencies in earlier versions of HMT's PESA database (Public Expenditure Statistical Analysis), the numbers are not reliable for years before 2002-03.

B.  THE EFFECT OF SOCIAL PROTECTION EXPENDITURE

  13.  Public expenditure may be grouped into three classes: non-identifiable; identifiable but not devolved; and identifiable and devolved.

  14.  Non-identifiable expenditure is to fund public goods such as national defence, foreign policy, UK debt interest, and the operation of central UK departments such as the Chancellor's departments. It is not apportioned to the population of the UK.

  15.  Some public expenditure is identifiable but not devolved. By far the largest component of this is what international statistical conventions group as "social protection". Until recently, agricultural support was also treated in the public expenditure system as if it were not devolved. This has recently changed, but for reasons of consistency over time the data in McLean et al, Fair shares..., on expenditure per head in the four nations of the UK excluded both of these functions of government. The purpose of this was to capture, as closely as the publicly available data permit, the "Barnett effect" of changes driven by the Formula on expenditure on devolved functions of government.

  16.  At my oral evidence session I was asked what would be the effect of adding social protection (and agriculture) back in to our data. This is done in the chart below. It is labelled "Figure 4.2 (Revised)" because it corresponds to Figure 4.2 in McLean et al., Fair Shares, with social protection and agriculture added back in.

  17.  On regional comparisons between public expenditure per head and GVA per head, our pamphlet already contained the data your Members requested. I apologise for failing to point that out to you at the time. The comparison with social protection and agriculture included is our Figure 5.2. The comparison with them excluded is our Figure 5.3. As your Members already have these charts I do not reproduce them again.

  18.  Comparing McLean et al Fig 4.2 with the attached Fig 4.2 (Revised), we observe:

  18.1.  Fig 4.2 (Revised) is somewhat more compressed. Regions with low relative spending (eg, Eastern and South East) and the region with highest relative spending (Northern Ireland) are closer to the UK average when social protection expenditure is included in the comparison than when it is excluded.

  18.2.  In Fig 4.2 (Revised), London does not overtake Northern Ireland, so it appears less of an anomaly than in McLean et al., Fig 4.2.

  18.3.  However, the ranking of the nations and regions stays roughly the same. In particular, Scotland has notably higher public expenditure per head than Wales on both measures, although it has substantially higher GVA per head.

  19.  Comparing McLean et al Fig 5.2 (including social protection and agriculture) with our Fig 5.3 (excluding them) we observe:

  19.1.  London is a less extreme outlier when social protection is included than when it is excluded. But on both measures it is by far the region with the highest public expenditure per head in comparison to GVA per head.

  19.2.  The relative position of Scotland is essentially the same on both measures.

  20.  The two sets of data serve different purposes. The data with social protection stripped out arise from an attempt to isolate the "Barnett effect" for the three devolved administrations. The "Barnett consequentials" described in the Statement of Funding Policy cannot be measured directly; identifiable public spending per head excluding social protection is a proxy for measuring "Barnett consequentials". For the regions of England, the same data series is a proxy for the effects of the distribution formulae used for health and local government services. Health and local government expenditure is devolved to subnational authorities in England; social protection expenditure, essentially, is not.

  21.  The data with social protection included may be used to measure the performance of the United Kingdom as a social citizenship network, in which poor people and poor regions are protected from adverse shocks.

  22.  On both sets of data, Members may feel that the position of London (and perhaps Scotland) appears anomalous.

Iain McLean FBA

Official Fellow in Politics

Professor of Politics, Oxford University

13 March 2009







1   Eg, I. McLean, the Fiscal Crisis of the United Kingdom (Basingstoke: Palgrave 2005); D Bell and A Christie, "Funding devolution: the power of money" in A Trench ed, Devolution and Power in the United Kingdom (Manchester: Manchester University Press 2007), 73-85; J Mitchell, "Spectators and Audiences: The Politics of UK Territorial Finance", Regional and Federal Studies, Vol 13 (2003) Winter, 7-21; D Heald and A Macleod, "Revenue-raising by UK Devolved Administrations in the Context of an Expenditure-based Financing System", Regional and Federal Studies, Vol 13 (2003) Winter, 67-90. Back

2   D Fraser, "Thatcher told her ministers to slash the Scottish budget", [Glasgow] Herald 06.06.2008, p 6; McLean et al., Fair Shares..., p 14; National Archives of Scotland, SOE6/1/1708-9, Sep-Oct 1984. Back


 
previous page contents

House of Lords home page Parliament home page House of Commons home page search page enquiries index

© Parliamentary copyright 2009