Producing economic statistics and advice for Scotland Highland Economic Forum October 2009 Sandy Stewart Senior Statistician Scottish Government Office.

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