1. EU KLEMS project on Productivity in the European Union
Presentation for Working Group on National Accounts
Eurostat, Luxembourg, 27 June 2005
Bart van Ark (Groningen Growth and Development Centre, University of Groningen)
This project is funded by the European Commission, Research Directorate General as part of the 6th Framework Programme, Priority 8, "Policy Support and Anticipating Scientific and Technological Needs".

2. Main aspects
EU KLEMS project is 3-year statistical and analytical research project funded by 6th Framework Programme
Purpose is to create a database on growth accounts by industry (NACE 60+) for EU member states with a breakdown into contributions from capital (K), labour (L), energy (E), materials (M) and service inputs (S)
Full coverage of “old” EU-15 plus 5 new member states (PL, SK, HU, CZ and SI)
Limited coverage of other 5 new member states (CY, MT, LT, LV and EE)
Also comparisons with U.S., Canada and Japan
, with greatest detail for post-revision period
14 research institutes across Europe, led by GGDC and NIESR
In 2nd phase conduct a number of analytical research projects

3. Setup of project: Workpackages

4. Progress June 04 – June 05
Consortium meetings (October 2004, London; June 2005, Helsinki)
Co-ordination of database construction and analytical work
Presentation of project results
Data co-ordination group meetings (June 2004 & February 2005, A’dam)
Decisions about concepts
Identification of problem areas
Data collection:
Data availability questionnaire in January/February
1st round of data collection focused on pre-revision data
Data base management
Workshop Database Management (DM) group on 18 & 19 January
Acquisition and installation of SAS software & Prodsys
Contacts & meetings with NSI’s
Visited most NSI’s  three status positions
Data delivery schedules

5. NSI’s have shown major interest in project
Three status positions:
Subcontracting: Statistics Finland, ISTAT, Statistics Netherlands
Participatory status: Statistics Sweden, STATEC (Luxembourg), ONS
Observer status: INSEE, Statistics Denmark, Statistisches Bundesamt, Institute of National Statistics (Belgium), Statistics Austria, Statistics Ireland, INE, Statistical Office Slovakia, Statistical Office Poland, CSO Hungary
Contacts under development:
New member countries covered by project: Czech Republic, Slovenia
Contacts to be developed: Greece, Portugal, Lithuania, Latvia, Estonia, Cyprus, Malta
Also contacts with:
USA (Harvard University, BEA, BLS),Japan (RIETI, Hitotsubashi), Canada (Statistics Canada)

6. Main Issues on WP1-WP4 WP1:
Output and intermediate inputs will be based on revised Supply and Use tables (with FISIM adjustment and chain link)
Refinement to methodology to measure use table in basic prices (capture trade and transportation margins)
Industry classification (EUK-72) close to NACE Rev. 1 with some refinements
Consortium will integrate older pre-revision tables (where possible in co-operation with NSIs)
Working agreement with OECD STI (Science, Technology and Industry) on SUT work (in consultation with Eurostat)

7. Main Issues on WP1-WP4 WP2: WP3:
Distinction between labour quantity (hours and employment) and labour quantity (age, gender, education level)
Labour quantity uses integrated labour and national accounts as default; but alternative approaches are considered
Labour quality will be dependent on country sources
WP3:
Point of departure are capital formation series
Key issue is maximum breakdown of asset types (including ICT categories) …
… and method of allocation of assets to individual industries

8. Main Issues on WP1-WP4 WP4:
PPPs are required on industry-by-industry basis
Make partly use of Output PPPs based on unit value ratios (e.g. from Prodcom) …
… partly of Expenditure PPPs, allocated to individual industries and adjusted for relative transport and trade margins and relative net taxes
Multilateral weighting of industry PPPs on the basis of EKS weighting system

9. EU KLEMS Analytical Module is set up as a Relational Database
Central data-hub
… but can be mirrored by individual consortium partners or NSI’s

10. What is a Relational Database?
Separate data and meta-data
Meta-data (concordances) has number of dimensions, e.g.
Year: 1970, 1971,…..
Country: A, B,C,..
Industry: 1, 2, 3, ….
Unit: nominal, real, deflator
Type: investment in building, machinery,…,
If one wants to call up nominal investment in machinery in industry 3 in country A in 1970, dimensions are combined and link to a unique number
Relational database is different from a hierarchical system and more flexible

11. Core of the Database is Programmed in SAS which exploits Advantages of Relational Database Structure

12. Productivity and other computations are done in ProdSys©

13. What is ProdSys© and what can it do?
ProdSys is software program on top of SAS designed for productivity measurement and growth accounting (ESI-VU/ Federal Reserve Board)
Basic tools for EU KLEMS database:
Aggregation
Disaggregation
Balancing
Concording
Cross-classification
ProdSys will be ‘open source’:
Users can make modifications
EU KLEMS will maintain an ‘official’ EU KLEMS production version of ProdSys

14. The Analytical Module will be complemented with a Statistical Module

15. Statistical vs. analytical modules of database
Analytical module of the database
Core of the EU KLEMS database
Uses “best practice” techniques in area of growth accounting
Focuses on international consistency
Aim is full coverage (country * industry * variable) for revision period
Consider alternative or pioneering assumptions (for example, output and price measurement of ICT goods and non-market services, measurement of skill levels, construction of capital services, capitalization of intangible assets).
Statistical module of the database:
Developed parallel to the analytical module
Data consistent with those published by NSIs
Methods according to rules and conventions on national accounts, supply and use tables, commodity flow methods, etc. (SNA 1993, ESA 1995) or at least supported by NSI’s
Statistical module meets statistical standards of NSI's and Eurostat and can eventually be incorporated in their present statistical practices.

16. Issues in co-operation with NSI’s requires continuous contacts on bilateral basis
Access to unpublished data and sharing between consortium partners requires specific arrangements with NSI’s
Assistance and advice in using greater (unpublished) detail from Supply & Use tables as well as historical (pre-revision) SUTs.
In principle use integrated labour and national accounts for labour quantity data as default, but alternative sources (including micro data, where possible) are also important
Dilemma of not using “not so reliable” data on industry-by-asset tables or using estimation methods
Use of unpublished data for development of analytical and statistical modules of database needs to be carefully monitored
EU KLEMS aims to organize working schedules around individual NSI’s programmes, but flexibility is needed to make it work

17. Deliverables related to NSI’s
Statistical roadmap, including data base management system (draft version available, final version after summer 05)
Preliminary test version of EU KLEMS database in 1st quarter of 06
First public version of EU KLEMS dataset with EU KLEMS Manual in 4th quarter of 06
Statistical implementation plan (producer guide) by 3rd quarter 07
Statistical progress report (every 6 months)

18. Next steps 3rd and 4th quarter 2005 1st quarter 2006:
Feedback from NSIs on statistical Roadmap
15/16 September: expert workshop on WP1 (inter-industry accounts) (Groningen)
Possibility to plan additional meetings for exchanging information on methods, data, etc.
EU KLEMS participation in OECD workshop on productivity measurement (Madrid) (17-19 October Madrid)
November: expert workshop on WP2 (labour accounts) (London)
1st quarter 2006:
Release of preliminary version of EU KLEMS database
Early 2006: special meeting for NSI’s and research from new member states

19. Next steps 2nd & 3rd quarter 2006: 4th quarter 2006:
Feedback from NSI’s on preliminary version of EU KLEMS database
Additional data inquiries and collection
Start on discussion of statistical implementation plan
Session on productivity measurement at IARIW meeting (August 2006, Joensuu, Finland)
4th quarter 2006:
Release first public version of EU KLEMS database

20. Contact Details Bart van Ark (project director)
Groningen Growth and Development Centre,University of Groningen
PO Box 800, 9700 AV Groningen
Telephone:
Gerard Ypma (project administrator)
Telephone:
Website: