Producing economic statistics and advice for Scotland Highland Economic Forum October 2009 Sandy Stewart Senior Statistician Scottish Government Office of the Chief Economic Adviser
Dec 26, 2015
Producing economic statistics and advice for Scotland
Highland Economic ForumOctober 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 indicatorsGDP
Trade statisticsGCS/IME
SNAPNational Accounts/
GERS/Tax
Modelling/Budget advice
Input-Output/Impact studies and modelling
Business StatisticsMicro analysis/IDBR/ABI/R&D
Labour Market
Family SpendingHHFCE AEA Inventories by product LFSGlobal Connections AEA by Scottish industryInvestmentStock change
Regional AccountsSTES, ASHEABI1, ABI2PAYEProperty PricesTaxes & Subsidies
MPIMIDDSFSIONSHMTBERR
HMRCSPIFRSBERR
Labour Market Extensions
Scot : Government Expenditure and
Revenues Scotland
SNAP
Scot IME
Scot Global Connections Survey
Symmetric tables, IxI PxP, Leontief Inverses
CGE Models
HMT: COINS, PESA, CRA
ONS : PSA, QNA
UK GDP(P)
UK INPUT-OUTPUT
SCOT INPUT-OUTPUT
Scot GDP(E)
Scot GDP(I)
Scot GDP(P)
UK NATIONAL ACCOUNTS UK ENV ACCOUNTS
UK GDP(E)
UK GDP(I)Impact Assessments
Environmental 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
0
20
40
60
80
100
120
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
Year
UK
=10
0
Highlands and Islands
Caithness &Sutherland and Ross& Cromarty
Inverness & Nairnand Moray, Badenoch& Strathspey
Lochaber, Skye &Lochalsh and Argylland the Islands
Eilean Siar (WesternIsles)
Orkney Islands
Shetland Islands
Scotland
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
60.0
70.0
80.0
90.0
100.0
110.0
120.0
130.0
140.0
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
2003 2004 2005 2006
Year / Quarter
Ind
ex
(2
00
3=
10
0)
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:
– 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
• Estimation never repeated after initial trial– 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
GDPO and GDPI - Scotland
80.0
85.0
90.0
95.0
100.0
105.0
110.0
1998
1998
1999
1999
2000
2000
2001
2001
2002
2002
2003
2003
2004
2004
2005
2005
2006
2006
2007
2007
2008
2008
2009
Year quart
2005
=10
0
GDPO
GDPI
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