Transcript EU KLEMS Tutorial

The Productivity Gap between Europe
and the US: Trends and Causes
Marcel P. Timmer
Groningen Growth and Development Centre
The EU KLEMS 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".
EU-15 Labour productivity convergence
to US has ended
100
(US=100)
GDP per hour
worked
90
80
GDP per capita
70
60
06
20
04
20
02
20
00
20
98
19
96
19
94
19
92
19
90
19
88
19
86
19
84
19
82
19
80
19
78
19
76
19
74
19
72
19
70
19
68
19
66
19
64
19
62
19
19
60
50
Alternative Explanations
•
•
•
Proximate causes: slower emergence of knowledge
economy
•
•
•
•
Small ICT-producing sector
Limited role of ICT-investment
Lower levels of skilled labour
Less innovation (product and process)
Ultimate causes: institutions
•
•
•
Role of labour markets (Blanchard, 2004)
Product market regulations (Nicoletti and Scarpetta, 2003)
End of catch-up (Aghion and Howitt, 2006): More R&D
investment, Higher education system reform
(Statistical myth)
Sector Contribution to Labour
Productivity Growth based on
Shift Share Analysis

VAi ,t H i ,t
H i ,t
VAt H t
Ht 
VA
VA

ln
  vi ln
   vi ln
 ln
VAt 1 H t 1
VAi ,t 1 H i ,t 1  i
H i , 1
H t 1 
i
Source: Stiroh (2002)
Sector contribution to market economy
labour productivity growth, 1980-2005
3.0
2.0
1.0
0.0
EU, 80-95
EU, 95-05
Market services
US, 80-95
Goods production
Source: Updated from van Ark, O’Mahony and Timmer,
Journal of Economic Perspectives, Winter 2008
US, 95-05
ICT production
Market services important source of
growth differences across Europe & US
4.0
Sector contribution to market economy
labour productivity growth, 1995-2004
3.0
2.0
1.0
0.0
FIN
US
GBR AUT FRA NLD BEL GER DNK ITA
-1.0
Market services
Goods production*
Source: Updated from van Ark, O’Mahony and Timmer,
Journal of Economic Perspectives, Winter 2008
ICT production
ESP
EU KLEMS Growth Accounts (1)
Gross output PF:
Y j  f j (K j , L j , X j ,T )
where Y is output, K is an index of capital service flows, L is an index
of labour service flows and X is an index of intermediate inputs,
Assumptions: competitive factor markets, full input utilization,
constant returns to scale and using the translog functional form
Growth accounting equation:
 ln Y jt  v  ln X jt  v  ln K jt  v  ln L jt   ln A
X
jt
v Xjt 
PjtX X jt
PjtY Y jt
K
jt
; v Ljt 
PjtL L jt
PjtY Y jt
L
jt
; v Kjt 
v L  v K  v X 1
PjtK K jt
PjtY Y jt
Y
jt
EU KLEMS Growth Accounts (2)
Labour services
based on hours
worked by 18 types
(education, age,
gender)
Capital services
based on 8
asset types
(ICT and nonICT)
 ln Lt   vl ,t  ln H l ,t
l
 ln Kt   vk ,t  ln S k ,t
k
Measurement of capital services

S i ,T    i ,t I i ,T t
Capital stocks for
each asset
t 0

S i ,T   (1   i ) t 1 I i ,T t  S i ,T 1 (1   i )  I i ,T
t 0
 ln Kt   vk ,t  ln S k ,t
Aggregate capital
services
k
and with contribution of capital services based on user cost equation
v k ,T  p k ,T S k ,T
p
k ,T
S k ,T
i
p
K
k ,t
p
r  k p [ p  p
I
k ,t 1 t
I
k ,t
I
k ,t
I
k ,t 1
]
What is new in EU KLEMS?
•
•
•
•
•
•
Data available for at least 30 industries and over 20 countries
Systematic data collection based on national accounts and
complementary official sources (LFS and other surveys)
Long time coverage 1970-2005, with greatest detail for post-1995,
harmonized methodologies on industry classification, deflation and
aggregations
Decomposition of inputs:
• Capital assets in 7 asset types
• Labour input in 18 categories (3 x skill; 3 x age and gender)
• Intermediate inputs: energy, materials and services input
Broad coverage of EU countries:
• Growth accounts coverage of 14 EU new member states
• Limited coverage of 11 other EU countries
• Also comparisons with U.S. Korea, Canada, Australia and Japan
www.euklems.net. Updated each year.
Explaining growth in value added based
labour productivity
VA
L
L
 ln
 w  ln
H
H
w
ICT
 ln
Labour
composition
K
ICT
H
N  ICT
K
N  ICT
w
 ln
H
  ln AVA
ICT-capital
per hour
Non-ICT-capital
per hour
Multi factor
productivity
ICT investment contributes to labour
productivity growth in market services growth in
all countries, 1995-2004
3.5
2.5
1.5
0.5
-0.5
-1.5
US
GBR
NLD
FIN
FRA
BEL
DNK
GER
ICT capital deepening contribution
AUT
ESP
ITA
PLUS Improvement in labour composition
3.5
2.5
1.5
0.5
-0.5
-1.5
US
GBR
NLD
FIN
FRA
ICT capital deepening contribution
BEL
DNK
GER
AUT
ESP
ITA
Labour composition change contribution
PLUS Non-ICT deepening
3.5
2.5
1.5
0.5
-0.5
-1.5
US
GBR
NLD
FIN
FRA
Non-ICT capital deepening contribution
Labour composition change contribution
BEL
DNK
GER
AUT
ESP
ICT capital deepening contribution
ITA
… but MFP contribution makes the big
difference between fast and slow growth
3.5
2.5
1.5
0.5
-0.5
-1.5
US
GBR
NLD
FIN
FRA
Non-ICT capital deepening contribution
Labour composition change contribution
BEL
DNK
GER
AUT
ESP
ICT capital deepening contribution
MFP growth
ITA
What is MFP growth?
•
•
Ratio of output growth over input growth
Disembodied technological change, under assumptions:
•
•
•
•
•
•
•
Constant returns to scale
Technical efficient
Allocative efficient
Competitive output and input markets
Outputs and inputs are measured without error
All outputs and inputs are included
Measure of our ignorance (Abramovitz)?
•
Deviation from assumptions above
What is MFP growth?
•
•
•
•
•
•
•
Constant returns to scale ? Perhaps at sector level justified
Technical efficient? Many firms/sectors are not on the frontier.
Allocative efficiency? ICT investments….
Competitive output and input markets? If not, prices do not reflect marginal
revenues and costs
•
•
Non-competitive product markets: Mark ups?
Regulation in capital and labour markets
Outputs and inputs are measured without error?
•
•
•
Capital (and labour) capacity utilization
Quality adjusted quantities of inputs
Quality adjusted quantities of outputs
All outputs and inputs are included?
•
•
•
Missing inputs (R&D, intangibles, management)
Measure of our ignorance (Abramovitz)?
Deviation from assumptions above
Good news: many of this are (or can be) tackled
How to explain differences in MFP growth?
Distance to the frontier-approach
1.2
1
MFP-level (US=1)
in Market Services
0.8
0.6
Source: Inklaar and Timmer (2008) GGDC Productivity level database
JA
P
HU
N
CZ
E
DN
K
NL
D
US
A
G
ER
IR
L
BE
L
FR
A
ES
P
SW
E
FI
N
UK
AU
S
IT
A
AU
T
SV
N
0.4
Distance to the frontier-approach
Unconditional convergence:

 ln MFPi   ln MFPF MFPi

Conditional convergence:



 ln MFPi   ln MFPF MFPi   Z   Z  ln MFPF MFPi
Imitation
Innovation
Interaction
Potential conditions include:
1.Skills, 2. ICT-capital and 3. Product market regulation
(Inklaar, Timmer and van Ark (2008), “Market Services
Productivity”, Economic Policy)

Regulation
Some evidence at detailed industry-level
Dep var: MFP growth
Technology gap
Barriers (average)
All industries
5
6
0.019*** 0.009
(0.004)
(0.007)
0.000
-0.004
(0.007)
(0.007)
Barriers (industry-level)
-0.010
(0.007)
Barriers*Technology gap
Number of observations
Selected set
7
8
0.009
0.010
(0.006)
(0.012)
2376
0.015
(0.009)
2376
715
-0.009
(0.010)
-0.002
(0.017)
715
Post and telecom
8
0.068***
(0.023)
-0.060***
(0.021)
0.037
(0.029)
264
Source: Inklaar, Timmer and van Ark (2008), “Market Services Productivity”,
Economic Policy, Table 10
Main findings
•
•
•
•
•
Importance of having detailed measures of productivity
at industry level, both growth and level estimates
End of convergence in productivity across Europe and
the US, mainly driven by differences in MFP growth in
market services
New productivity gap has opened up
Regulatory practice shows no clear impact on MFP, only
for post and telecommunications
Market regulation, competition and Scale and/or lagged
adjustment to ICT in Europe?
Some ideas for future research based on
methods teached in this course
•
•
•
•
•
•
•
•
•
•
Breakdown of TFP growth (DEA,SFA)?
Allow for increasing returns to scale?
Imperfect competition?
Localised innovation? (DEA)
Stochastic frontier estimation of MFP-gaps
Efficient use of ICT? (.Dot boom and bust)
Complementarity skills and ICT
Variable lags and interactions in effects of skills, ICT and
regulation
Innovation, trade and R&D (manufacturing)
Sectoral convergence in the world economy?