Transcript Open - The Scottish Government

Producing economic statistics
and advice for Scotland
Highland Economic Forum
October 2009
Sandy Stewart
Senior Statistician
Scottish Government
Office of the Chief Economic Adviser
Contents
• Producing statistics and economic advice
for Scotland
• Scottish Issues
• Sub-Scotland Data and Analysis
• Modelling Sub-Scotland GVA – output and
income approaches
• Planning – looking to the future discussion
Producing economic statistics
and advice
Office of the Chief Economic
Adviser
Producing Statistics
Short term indicators
GDP
Trade statistics
GCS/IME
SNAP
National Accounts/
GERS/
Tax
Modelling/
Budget advice
Input-Output/
Impact studies
and modelling
Business Statistics
Micro analysis/
IDBR/ABI/R&D
Labour Market
UK NATIONAL ACCOUNTS
UK ENV ACCOUNTS
SNAP
UK GDP(E)
Family Spending
HHFCE
Global Connections
Investment
Stock change
Scot GDP(E)
AEA Inventories by product
AEA by Scottish industry
LFS
UK INPUT-OUTPUT
Environmental
Extensions
UK GDP(I)
Regional Accounts
STES, ASHE
ABI1, ABI2
PAYE
Property Prices
Taxes & Subsidies
Scot GDP(I)
Impact Assessments
SCOT INPUT-OUTPUT
UK GDP(P)
MPI
MIDDS
FSI
ONS
HMT
BERR
Scot GDP(P)
CGE Models
HMT: COINS, PESA, CRA
Scot IME
Scot Global
Connections Survey
Symmetric tables, IxI PxP,
Leontief Inverses
Scot : Government
Expenditure and
Revenues Scotland
ONS : PSA, QNA
HMRC
SPI
FRS
BERR
Labour Market
Extensions
Scottish Issues
Issue 1 – Meeting Users needs
– Ministers – routine publications and ad-hoc demands – want
more and earlier – statistical and non-statistical products
– Council of Economic Advisers – recommendation 21 “the quality
of economic statistics in Scotland does not yet meet the needs of
government. We welcome the many steps which are being
taken to improve the situation and have drawn special attention
in this chapter to some of the areas which we believe should
receive priority in these developments:
•
•
•
•
•
•
the financial services sector
the oil and gas sector
depreciation and environmental statistics
output prices, especially in the export sector
the public sector
alternatives measures to GDP”
– Media and public – critical, but strong underlying political interest
– Local knowledge – more extensive sub-Scotland analysis
required
Recent press comments
• “This tardiness in the publication of
Scottish Government (GDP) statistics is a
continuing scandal” – Scotsman 21/4/09
• “Scotland’s GDP numbers: An amazing
case of statistical mince” – Scotsman 22/4/09 –
criticism of chain-linking (Delphic
mysteries) and seasonal adjustment (need
peaks and troughs)
• But there were some anomalies in the
GDP figures ..” – Scotsman 23/4/09
Issue 2 – Dependency on UK data
– reporting units not determined by geography
– need for survey boosts (can be expensive!)
– programme dependent on UK programme
• Purchases inquiry
• National accounts re-engineering
– local considerations need to be balanced with
national considerations
– Allsopp developments welcomed – especially
GVA(P)
Issue 3 – SIC 2007
•
•
•
•
Implications for all our statistical products
Need for major (?) system re-writes
No real scope for delays to publications
Need to scope requirements and consider
resource implications
Issue 4 - Resources
• Already over-stretched
• Vacancies – recruitment problems for
qualified professional staff
• Tightening of budgets 
• Make better use of ONS methodology
directorate (e.g. SNAP review of deflators)
• Need to work with others to “pool”
information and peer review outputs.
Sub-Scotland data and
analysis
ONS – Regional GVA – 1.
• NUTS3 geographies
– Caithness & Sutherland, Ross & Cromarty
– Inverness & Nairn and Moray, Badenoch &
Strathspey
– Lochaber, Skye & Lochalsh and Argyll and the
Islands
– Eilean Siar
– Orkney Islands
– Shetland Islands
• Annual data – 1995 – 2006 latest available.
ONS – Regional GVA – 2.
• Industries
–
–
–
–
–
–
Agriculture, Forestry and Fishing
Production
Construction
Distribution, Transport and Communication
Business Services and Finance
Public administration, education, health and other
services.
– Total GVA
• Income components
– Compensation of employees
– Gross Operating Surpluses
Highlands & Islands GVA per head
120
Highlands and Islands
Caithness &
Sutherland and Ross
& Cromarty
100
Inverness & Nairn
and Moray, Badenoch
& Strathspey
UK=100
80
Lochaber, Skye &
Lochalsh and Argyll
and the Islands
60
Eilean Siar (Western
Isles)
Orkney Islands
40
Shetland Islands
20
Scotland
0
1995
1996
1997
1998
1999
2000
2001
Year
2002
2003
2004
2005
2006
ONS Regional Gross Disposable
Household Income
NOMIS – Labour Market
• Contents
– population, working age population
– employment, unemployment, economic
inactivity
– occupations, qualifications, earnings
– benefits & claimants
• Employment x Earnings can be used to
proxy Compensation of Employees –
largest income component of GDP.
Modelling Sub-Scotland GVA –
output and income approaches
Experimental GVA (output based) in
constant basic prices for Highland Local
Authority
140.0
130.0
Index (2003=100)
120.0
110.0
100.0
90.0
80.0
70.0
60.0
1
2
3
4
2003
1
2
3
4
1
2004
2
3
4
1
2005
2
3
4
2006
Year / Quarter
Agriculture, Forestry and Fishing
Construction
Production
Services
Total GDP
Experimental regional GVA (1)
• Initial research determined that a ‘bottom-up’
approach using the source data used in the
National GVA measure is not possible.
– The samples are drawn to provide UK-level statistics
– Taking a Scottish extract is not without its difficulties,
notably business units spanning the border
– Breaking this down further by L.A. leads to incredibly
small and inherently volatile samples (and estimates).
– The current panel-estimation approach used in
manufacturing is completely incapable of producing
regional (sub-Scotland) estimates.
Experimental regional GVA (2)
•
•
Hybrid ‘top-down’ approach trialled (during
2007).
The approach used 3 main data-sources:
1. Quarterly GVA (Scotland) – used as a National constraint
for regional estimates
2. Inter-Departmental Business Register (IDBR) – provided
employment growth by industry/L.A.
3. Annual Business Inquiry – provided benchmark levels of
GVA by local authority in addition to regional labour
productivity by industry.
•
In essence, relative movements in ‘2’ and ‘3’
were used to apportion overall growth in ‘1’ to
each region.
Experimental regional GVA (3)
•
In short:
–
–
–
–
•
Estimation never repeated after initial trial
–
–
•
The estimates are artificial; modelled using a range of indicators not
really intended to be used in this way.
No new primary data used in the estimates
Pragmatically, the best estimates that could be achieved with the
current data
Some plausible results achieved, some less so
Was clear that a vast amount of (IDBR) data-cleaning would be
required for initial development
Ongoing requirement to actively track changes in IDBR employment
against other non-statistical information (newspapers, news releases,
local knowledge) to ensure that regional trends were plausible
Would require a significant investment of resources with no
guarantee of success.
GVA (Income approach)
• SNAP project – builds on Regional Accounts
methodologies
• Currently produced as experimental statistics for
Scotland – see SNAP website
www.scotland.gov.uk/SNAP
• Proxy data for income components available
locally – employment, earnings
• Modelling
– combine output and income analysis where
appropriate
– augment with local knowledge
Year quart
2009
2008
2008
2007
2007
2006
2006
2005
2005
2004
2004
2003
2003
2002
2002
2001
2001
2000
2000
1999
1999
1998
1998
2005=100
GDPO and GDPI - Scotland
110.0
105.0
100.0
95.0
GDPO
GDPI
90.0
85.0
80.0
Planning – looking to the
future
Topics for discussion
• OCEA need to know how our data are
being used – or not used
• Need to understand how you use data
from ONS, HMRC, OGDs
• Sharing local knowledge – peer review of
outputs
• Need to collect supplementary data
• Need to develop a modelling framework