Scottish Government

B EXPENDITURE METHODOLOGY

This appendix outlines the methodologies used to estimate public sector expenditure for Scotland and highlights where these methodologies differ from those used in previous editions of GERS.

Methodology

Figures for UK and Scottish public sector expenditure are taken directly from official UK Government sources.

Total Expenditure on Services ( TES) can be separated into two components:

• Identifiable expenditure: that is expenditure that can be clearly allocated to a country or region in terms of having been spent for the benefit of that country or region; and

• Non-identifiable expenditure; that is expenditure that cannot be identified as benefiting a particular country or region of the UK but is instead incurred on behalf of the UK as a whole.

Total Expenditure on Services covers most expenditure by the public sector. Adding an accounting adjustment reconciles it with Total Managed Expenditure ( TME). In 2006-07, TES accounted for 95 per cent of TME.

UK Expenditure Figures

The primary data source used to estimate Scottish public sector expenditure is the Public Expenditure Statistical Analyses ( PESA) database, published by HM Treasury ²². Within PESA, tables relating to a Country and Regional Analysis ( CRA) are available in which UK Government departments and devolved administrations have allocated expenditure programmes to Scotland, Wales, Northern Ireland and English regions.

In general, data on identifiable expenditure for Scotland are taken directly from PESACRA 2008. However, following a detailed scrutiny of the database, a number of amendments were made before the data were used in GERS. PESA 2008 is also the source for UK non-identifiable expenditure, a share of which was apportioned to Scotland. There are a number of possible ways in which to allocate to Scotland a share of UK non-identifiable expenditure: the Scottish share of UK population (8.4 per cent in 2006-07), UKGVA less extra-regio (8.1 per cent in 2006-07) or a more specific method if available. These are discussed in more detail below.

Summary of Changes

There are three key reasons why the expenditure data in GERS 2006-07 differ from those presented in previous editions of GERS. These can be categorised as:

1. revisions to PESACRA data

2. revised methodologies for apportioning non-identifiable expenditure

3. revised methodologies for estimating the accounting adjustment.

1. Revisions to PESACRA Data

One explanation for the differences to the expenditure data contained in this GERS report and previous publications are revisions to the data underlying the PESACRA analysis. In 2006 and 2007, progress was made by UK Government departments to better apportion public expenditure by UK country and region. These changes are now reflected in the more refined data contained in PESACRA 2008. Work is ongoing to continuously improve the quality of the PESACRA data.

While many anomalies in previous editions of PESACRA have been addressed and are now reflected in both PESACRA 2008 and this GERS report, a number of supplementary amendments to the PESACRA 2008 dataset were made. The aim of these refinements was to ensure that the public sector expenditure figure for Scotland captures as accurately as possible expenditure for the benefit of Scotland.

The revisions to PESACRA 2008 can be grouped under the following headings:

• re-classification of expenditure
• re-apportionment of expenditure
• amendment to Scottish Water capital expenditure
• elimination of double counting
• new methodology for estimating EU transactions
• new methodology for estimating debt interest

Many of these revisions are relatively minor and, if less than £500,000 in any one year, they will show as a '-' in the following tables. As a result, the figures in the following tables may not sum exactly, especially where there are a number of incidences where the figures involved are relatively small. The overall impact of these revisions is highlighted in Table B.1.

Table B.1: Revisions to Estimates of Total Public Sector Expenditure from PESA 2008: Scotland 2002-03 to 2006-07

Re-classification of expenditure in PESA 2008

In PESACRA 2008, some expenditure by UK Government departments with responsibility only in England and Wales was recorded as UK non-identifiable. This included:

• expenditure by the Office of the Third Sector,
• rural development payments,
• an element of central health administration expenditure
• expenditure by the Department for the Environment, Food and Rural Affairs (Defra) on protection of biodiversity and landscape,
• Department for Innovation, Universities and Skills expenditure on the Technology Strategy.
• and expenditure by the Department for Culture, Media and Sport ( DCMS) on libraries and administration.

In GERS, such expenditure has been re-classified as expenditure for England and Wales only. As a result, no share was apportioned to Scotland in this publication.

Similarly, the data entry for the Charity Commission had included identifiable expenditure for Scotland. However, this is a body which only operates in England and Wales and hence the expenditure entry for Scotland was removed.

Finally, approximately 21% of expenditure on London and Continental Railways, the company responsible for the Channel Tunnel Rail link was recorded in PESACRA 2008 as outside UK with Scotland allocated a per capita share.

Table B.2: Re-classification of Expenditure in PESA 2008: Scotland 2002-03 to 2006-07

- : less than £0.5 million

Re-apportionment of Expenditure

In PESACRA 2008, expenditure on nuclear decommissioning was classified as identifiable to the region where nuclear facilities are located. However, as discussed in Box 6.5 it is believed that this expenditure is best captured as a non-identifiable expenditure. This has been amended and nuclear decommissioning expenditure has been apportioned on a per capita basis.

In addition to this relatively major revision, a number of smaller amendments were made.

In PESACRA 2008, identifiable Post Office expenditure had been allocated on an exactly equal basis across UK countries and English regions. This was re-apportioned on a per capita basis.

In PESACRA 2008, identifiable expenditure on inward investment by the UK Trade and Investment body had been allocated on a per capita basis. This has been changed to a GVA basis.

Consistent with the interpretation that expenditure by the Office for National Statistics is non-identifiable, it has been apportioned across the UK in PESACRA 2008 on a per capita share. However, some of ONS activities do not relate to Scotland, such as the Population Census for England and Wales. Suitable adjustments have been made in GERS 2006-07 to allocate more accurately ONS expenditure to countries and regions.

Finally, in PESACRA 2008, expenditure by the Scottish Office Pension Agency was allocated only to Scotland. In contrast, expenditure by comparable pension agencies across the rest of the UK was allocated to the residence of the recipient of the pension. To correct for this asymmetry, information from the Scottish Office Pension Agency was used to re-apportion pension expenditure across the UK according to residence.

Table B.3: Re-apportionment of Expenditure in PESA 2008: Scotland 2002-03 to 2006-07

- : less than £0.5 million

Amendment to Scottish Water capital expenditure

Due to a reporting change, capital expenditure by Scottish Water in PESACRA 2008 was revised downwards from the expenditure reported in PESACRA 2007. While the figure for 2006-07 in PESACRA 2008 is accurate, amendments for earlier years were incorrect. Following a review of the actual capital expenditure undertaken by Scottish Water, GERS replaced the erroneous data in PESACRA 2008 with the correct data, entered in PESACRA 2007, for the years 2002-03 to 2005-06. These are consistent with Scottish Water's published annual accounts.

Table B.4: Amendment to Scottish Water Capital Expenditure in PESA 2008: Scotland 2002-03 to 2006-07

Elimination of Double Counting

Two elements of double counting of public sector expenditure for Scotland were identified in PESACRA 2008. These have been removed. The two concerned were -

• Duplicate record for expenditure in housing revenue account,
• Rail subsidy entries in 2006-7.

In 2006-07, this detection of double counting lowered total public sector expenditure by £326 million.

Table B.5: Elimination of Double Counting in PESA 2008: Scotland 2002-03 to 2006-07

New methodology for estimating EU transactions

EU transactions comprise two elements - receipts from, and payments to the EU. In GERS 2004-05, actual receipts from the EU to Scotland were used (as reported in the Scottish Government's Consolidated Accounts). UK Payments to the EU were apportioned to Scotland based on a GVA ratio.

In GERS 2006-07, the same methodology was used for receipts. However, for payments to the EU, a more sophisticated apportionment methodology was introduced. UK payments based on a GNI contribution were allocated to Scotland using a GVA ratio, aid to developing countries was apportioned on a per capita basis, and VAT-based payments were apportioned using the same shares as derived from the estimates of VAT revenue discussed in Chapter 4. Whilst the changes in the component parts were significant, the net impact was small.

Table B.6: New Methodology for Estimating EU Transactions in PESA 2008: Scotland 2002-03 to 2006-07

New methodology for estimating debt interest

GERS 2006-07 apportions Scotland a per capita share of UK central government debt interest (for further details see Box 6.7). For local government debt interest, per capita data are also now used. In GERS 2004-05, an alternative estimate based on a Scottish local authority debt interest model was used. However, this model is outdated and is not available for later years.

Table B.7: New Methodology for Estimating Debt Interest in PESA 2008: Scotland 2002-03 to 2006-07

2. Revised methodologies for apportioning non-identifiable expenditure

In GERS, the methodology to apportion non-identifiable expenditure and identifiable expenditure which occurs outside the UK to Scotland varies according to the particular expenditure estimated.

The methodologies used are listed in Table B.8. Each reflects the approach that is thought to best capture the 'who benefits' principle.

For the majority of spending programmes, the methods used to apportion non-identifiable expenditure to Scotland have remained unchanged from previous editions of GERS.

Table B.8: Apportionment Methodologies for Non-Identifiable Expenditure: Scotland 2002-03 to 2006-07 ^{1, 2}

1: Where there is no UK non-identifiable expenditure this is entered as not applicable (n/a).
2: Identifiable expenditure outside the UK, except EU transactions, is apportioned on a per capita basis.
3: In previous GERS publications, non-identifiable expenditure on agriculture, fisheries and forestry was allocated according to Scotland's share of UKGVA in agriculture, forestry and fishing. However following improvements in data identification, this expenditure category is now fully identifiable in PESACRA 2008.
4: All environment protection expenditure is apportioned on a GVA basis, except UKAEA and BNF expenditure on nuclear decommissioning which is apportioned on a per capita basis.

3. Revised methodologies for estimating the accounting adjustment

The public sector expenditure analysis in this report focuses on Total Expenditure on Services ( TES). In UK public finance documents, the principal measure of public sector expenditure is Total Managed Expenditure ( TME). The main difference between TES and TME is that TES does not include general government capital consumption and does not reverse the deduction of certain VAT refunds in the budget-based expenditure data. It also contains a number of items that are in budgets but not in TME, for example the grant equivalent element of student loans. TME is consistent with ONS National Accounts and the UK Public Sector Finances.

An accounting adjustment is introduced to align TES to TME.

The largest component of the UK accounting adjustment is general government capital consumption. This is a measure of the amount of fixed capital resources used up in the process of production.

The estimate of an accounting adjustment for Scotland is calculated using a variety of apportionment methodologies. Firstly, estimates of capital consumption from the ONS regional accounts, together with data from UK local government returns, have been used to estimate capital consumption for Scotland. In 2006-07, capital consumption for Scotland was estimated at £1.4 billion (11.3 per cent of the UK total). The published regional accounts data in December 2007 showed a significant downward revision to capital consumption for Scotland.

VAT refunds were allocated to Scotland on the basis of Scotland's share of UKTES. The student loans element and Nigerian debt cancellation were allocated according to Scotland's share of the UK population. The residual has been allocated by TES and disaggregated into current and capital components based on Scotland's TES ratios.

In sum, the accounting adjustment for Scotland was estimated at £2.7 billion in 2006-07, or 10.3 per cent of the total UK Accounting Adjustment.

Table B.9 summarises the effects of methodological changes and revisions to the estimates of the accounting adjustment for Scottish expenditure between 2002-03 and 2004-05, the final year for which such comparisons are possible.

In total, revisions to the underlying expenditure data and changes in apportionment methodologies have resulted in a reduction of £660 million to the estimate of the accounting adjustment for the year 2004-05 compared to the previous edition of GERS.

Table B.9: Revisions to Estimates of Accounting Adjustment: Scotland 2002-03 to 2004-05

Impact of revisions

The following tables highlight the impact of all data and methodological changes to identifiable, non-identifiable and total expenditure.

Revisions to total expenditure

The overall change between GERS 2004-05 and GERS 2006-07, following amendments to underlying data and in apportionment methodologies, are highlighted in Table B.10. This table displays the difference between total expenditure estimates in GERS 2004-05 and those contained in Chapter 6 of this publication. The revisions have been split into two groups, updates to the source data and revisions made in compiling GERS. Updates to the source data refers to revisions that can be traced to changes in the PESACRA data. Revisions made in compiling GERS reflect the methodology amendments detailed above.

As Table B.10 highlights, in addition to the reduction for Scotland, total expenditure for the UK has also been revised. This is largely the result of certain expenditure being reclassified back into the public sector.

Table B.10: Revisions to Estimates of Total Expenditure: Scotland 2002-03 to 2004-05

Revisions to identifiable expenditure

Table B.11 summarises the effects of methodological changes and revisions to the underlying PESACRA data on the estimates of Scottish identifiable expenditure between 2002-03 and 2004-05, the final year for which such comparisons are possible. For 2004-05, identifiable expenditure has been revised down by £557 million. Updates to the source data have reduced identifiable expenditure for Scotland by £706 million whilst revisions to the GERS methodology have increased expenditure for Scotland by £148 million. This revision equates to 1.4 per cent of identifiable expenditure in GERS 2004-05.

Updates to the source data are reflective of changes made by UK Government departments in allocating spending to UK countries and regions. The reduction of £706 million is the difference between the estimate of total identifiable expenditure on services for Scotland reported in Table 9.1 in PESA 2008 and that reported in the same table in PESA 2006.

Table B.11: Revisions to Estimates of Identifiable Expenditure: Scotland 2002-03 to 2004-05

Revisions to non-identifiable expenditure

Table B.12 summarises the effects of methodological changes and revisions to the underlying PESA data on the estimates of Scottish non-identifiable expenditure between 2002-03 and 2004-05, the final year for which such comparisons are possible.

In total, revisions to the underlying PESA data and changes in apportionment methodologies have resulted in a downward revision of £131 million to the estimate of total non-identifiable expenditure for Scotland in 2004-05. This revision equates to 2.1 per cent of total non-identifiable expenditure in that year.

Table B.12: Revisions to Estimates of Total Non-Identifiable Expenditure: Scotland 2002-03 to 2004-05

Total Revision to expenditure (including accounting adjustment)

The overall change in Total Managed Expenditure ( TME) is highlighted below. The revision is equal to the sum of revisions to Total Expenditure on Services ( TES) and the accounting adjustment.

Table B.13: Revisions to Estimates of Total Managed Expenditure: Scotland 2002-03 to 2004-05

Sensitivity Analysis

As the figures for public sector expenditure for Scotland contained in this report are estimates, it is important to obtain a sense of the impact on total expenditure of any possible variation in the estimates of individual expenditure components. Table B.14 illustrates the effect a 1 per cent variation in the estimate of each expenditure programme would have on overall expenditure for Scotland in 2006-07.

As the largest elements of public sector expenditure for Scotland, changes in the estimated level of social security, health and education will have the most significant impact on estimated total public sector expenditure for Scotland.

Table B.14: Effect of a 1% Change in Expenditure: Scotland 2006-07

- : less than £0.5 million

~ : less than 0.05%