Progress report on EU KLEMS
project on Growth and
Productivity in the European
Union
Presentation for OECD Workshop on Productivity Analysis and
Measurement
17 October 2006, Bern
Bart van Ark, Groningen Growth of 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".
Main characteristics of EU KLEMS
•
•
•
•
EU KLEMS project is 3-year statistical and analytical research
project funded by 6th Framework Programme
Create database on growth and productivity 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) from 1970 (revision and pre-revision)
15 research institutes across Europe and co-operation with
institutes in U.S., Canada, Japan, Korea, China and Australia
Strong involvement of many national statistical institutes (NSI’s):
•
•
•
•
Provide basic data (in particular national accounts) for EU KLEMS
database
Provide feedback on data use and methodologies
Validate statistical module of the database with matches official statistics
(there will also be an analytical module)
Work with Eurostat and European Commission on implementation of
2
growth
and productivity accounts in national statistical programmes
•
•
NSI’s and other third parties actively contribute
to EU KLEMS
Four status positions:
•
•
Subcontracting: Statistics Finland, ISTAT, Statistics Netherlands
•
Observer status: INSEE, Statistics Denmark, Statistisches
Bundesamt, Institute of National Statistics (Belgium), Statistics
Austria, Statistics Ireland, INE (Spain), Statistical Office Slovakia,
Statistical Office Poland, CSO Hungary, Czech Statistical Office,
Statistical Office Slovenia, INE (Portugal), COS Malta
•
In contact with: NSS Greece, Lithuania, Latvia, Estonia, Cyprus
Participatory status: Statistics Sweden, STATEC (Luxembourg),
ONS (UK)
Also contacts with:
•
USA (Harvard University, BEA, BLS),Japan (RIETI, Hitotsubashi),
Korea (SNU), China (Univ. of Beihang) Canada (Statistics
Canada, University of Alberta), Australia (ABS, Productivity
Commission)
•
Eurostat, OECD (Statistics Dept., DSTI)
3
Output from EU KLEMS database
•
•
•
•
•
Series on nominal values of output (gross output &
value added), intermediate inputs (energy, materials,
service inputs) and factor inputs (labour by
age/skill/gender and capital by asset type) by industry
& country, compensation of factor inputs
Volume series (1995=100 and growth rates)
Growth accounting series on output (gross output and
value added-based) and contributions of intermediate,
factor inputs, and total factor productivity
Additional variables:, technology variables (R&D &
patents), import shares in intermediate inputs
Also comparative levels with industry specific output
and input PPPs
4
Preliminary version of analytical module of
database available in March 2006
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 (e.g., output and price
measurement of ICT goods and non-market services, measurement of skill
levels, construction of capital services).
Analytical database is a research data base and are not official statistics
Statistical module of the database:
•
•
•
•
To be developed parallel to the analytical module in co-operation with NSIs
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
can5eventually be incorporated in their present statistical practices and in
New Cronos.
Time schedule final EU KLEMS year
from Sep 06 - Dec 07
Sep-06
Database
activities
workshop NSI's
CEE countries
Vienna
Analytical
research
work on
research
projects
Communication
6
Oct-06
Nov-06
Dec-06
work on draft
1 November:
1 December:
public version of Completion of draft
Release of
EU KLEMS
public version of database to NSI's
database
database
for checking
work on
research
projects
work on
research
projects
12 December:
WP9 workshop
Helsinki
10-11 September: 1 October: close of development of EU
feedback on
workshop NSI's feedback round on KLEMS manual
database from
CEE countries preliminary version
NSI's, Eurostat,
Vienna
of EU KLEMS
Commission, etc.
Time schedule final EU KLEMS year
from Sep 06 - Dec 07
Jan-07
Database
activities
Feb-07
Mar-07
15 February:
15 March:
11 January:
Data Coordination Completion of 1st Launch of 1st
version of EU
version of EU
Group meeting
KLEMS database KLEMS database
Amsterdam
Analytical
research
Communication
7
Apr-07 to Nov-07
Dec-07
development of
2nd version of
EU KLEMS
database
completion of 2nd
version of EU
KLEMS database
work on
research
projects
work on
research
projects
16-17 March:
consortium
meeting on
research
workshops on
EU KLEMS
research projects
final external
conference
on EU KLEMS
project
20 January: close
of feedback round
on draft version of
database
development
of statistical
implementation
plan
publication of
1st EU KLEMS
Productivity
Report
development
of statistical
implementation
plan with NSI's
completion of
2nd EU KLEMS
Productivity
Report
“Mind the Gap”-paper based on pre-EU KLEMS
data - indicative of what EU KLEMS will deliver
•
•
•
•
Compare performance of continental European
countries (France, Germany, Netherlands) with AngloSaxon countries (Australia, Canada, UK, U.S.)
Take a bottom-up, industry approach after extensive
data construction and harmonization
Analyze sources of productivity growth (1987-2003) and
productivity levels (1997, extrapolated to 2003), in
particular in market services
Analyze sources of gaps in productivity differentials
Continental European countries fall behind AngloSaxon due to lower uptake in ICT intensity and
slowdown in TFP growth
3.5
Sources of growth contributions to market economy LP growth (87-03)
Anglo-Saxon Economies
Continental European Economies
3.0
19871995
2.5
2.0
1.5
1.0
0.5
0.0
AUS
CAN
UK
USA
FRA
DEU
NLD
-0.5
ICT capital deepening
Non-ICT capital deepening
Labour composition and reallocation
TFP
19952003
Differences mainly due to differences in TFP growth in
market services
Industry contributions to market economy TFP growth (1987-2003)
1.5
Anglo-Saxon Economies
Continental European Economies
1.3
19871995
1.1
0.9
0.7
0.5
0.3
0.1
-0.1
-0.3
-0.5
AUS
CAN
UK
ICT production
USA
FRA
Goods-producing industries
DEU
Market services
NLD
19952003
Market services in continental European countries fail to
transform ICT intensity in TFP growth
3.5
Sources of growth contributions to market services LP growth (87-03)
Anglo-Saxon Economies
Continental European Economies
3.0
2.5
19871995
2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
AUS
CAN
UK
USA
FRA
DEU
NLD
-1.5
ICT capital deepening
Non-ICT capital deepening
Labour composition
TFP
Reallocation of hours
19952003
Methodology for level accounting
(more tomorrow in final session on PPPs)
•
•
•
Dual productivity decomposition due to index number
problems with primal
Decompose relative sectoral output price (PY) between
countries x and y into relative input price (PX) and
productivity (A):
PYx
PXx
Ax
ln y  ln y  ln y
PY
PX
A
Use CCD (Caves/Christensen/Diewert) index to
aggregate across outputs (Y) and inputs (X) and
industries (i)
Level accounting requires PPPs at industry level
(more tomorrow in final session on PPPs)
•
•
•
•
•
A mix of production PPPs (unit value ratios) and specified
expenditure PPPs for industry (gross) output (Timmer,
Ypma and van Ark, 2006)
Supply and Use tables to move from gross output to
sectoral output PPPs (Inklaar and Timmer, 2006)
Intermediate input PPPs: combine industry output PPPs
for domestic inputs and exchange rates for imports with
Supply tables to get commodity PPPs
Labour PPPs: relative wages of university & nonuniversity labour
Capital PPPs: investment PPPs for 6 assets combined
with annualization factor (relative gross return on capital)
Low productivity levels in market services in AngloSaxon economies, while high productivity levels in
European service industries are eroding
Relative Levels of Labour Productivity in Market Services,
1987 and 2003, US=100
Anglo-Saxon Economies
120
Continental European Economies
110
100
90
1997
80
2003
70
60
50
40
AUS
CAN
UK
USA
FRA
DEU
NLD
Labor productivity gap in Anglo-Saxon market services
driven by TFP, but TFP gap small or other way in Europe
Breakdown of Labour Productivity Gap relative to US (% gap to US, 1997)
50
Anglo-Saxon Economies
Continental European Economies
40
30
20
10
0
-10
-20
-30
AUS
CAN
ICT capital deepening
UK
Non-ICT capital deepening
FR
DE
Labour composition and reallocation
NL
TFP
Next steps
•
•
•
•
•
•
Are differences in productivity growth and levels in
services real or are there elements of measurement
error?
Understanding of high productivity levels in European
market services (‘perversely’ high?)
Explaining differences in input use and productivity
across industries & countries, e.g. impact of differences
in outsourcing
Evaluate index number alternatives for productivity level
comparisons
Improve methodology for updating to other years on
basis of harmonized SUTs (in EU KLEMS)
…more tomorrow in final session on PPPs: WELCOME !