Transcript A. de Panizza and M. Visaggio

The Dynamics of Productivity and
Competitiveness in Major Advanced
Economies: ICT and Other
Determinants
Andrea de Panizza
ISTAT
Mauro Visaggio
University of Perugia
Productivity, Competitiveness and the New Information Economy
Business, Systemic and Measurement Issues
ISTAT – Statistics Finland Joint Workshop
Rome, June 26-27, 2003
Decomposition of labour
Productivity Acceleration
A) Macro Decomposition
I.
Capital deepening
The increase in investment provides
more capital per worker
II.
Labour quality growth
Change in the composition of labour force
with respect to workers’ productivity
III.
Total factor productivity (TFP)
growth\
Change in output that cannot be attributed
to both capital and labour (technological
2
and business organization effects)
Decomposition of labour
Productivity Acceleration
B) Sectoral Decomposition
d (ln yt )

dt
O
 K ICT d (ln( k ICT t ))
K O d (ln( k t )) 
 t
 
t
dt
dt


ICT
labour
Productivity
NON-ICT
Capital deepening
ICT
O


d
(ln(
A
))
d
(ln(
A
ICT
O

t
t ))
d (ln( H )) 
L d (ln( L ))


   
  t 

dt
dt



dt
dt


ICT
NON-ICT
labour quality
Total Factor Productivity
3
Three channels through which ICT
affects labour productivity
Innovations in ICT according to Moore’s Law implies
drastic decrease in price in ICT sector
3
1
Increase in TFP in
ICT sector
(production of ICT)
2
Increase in capital
deepening in ICT
(adoption of ICT)
Increase in TFP in
non-ICT sectors that
adopt ICT due to
business
reorganization
4
Components of US labour
productivity acceleration
(1996-00 relative to 1974-1995)
Gordon Olin-Sic. Jor-Ho-Stir
CEA
Lab. Prod. Accel.
1.04
1.03
0.93
1.58
I. Capital deepe.
0.37
0.48
0.52
0.39
- ICT
0.60
0.60
0.44
0.62
- non-ICT
-0.23
-0.13
0.08
-0.23
0.01
-0.02
-0.1
0.0
0.52
0.57
0.51
1.18
0.30
0.22
0.14
0.46
0.11
0.00
0.27
0.25
0.00
0.18
1.00
0.00
II. labour quality
III. TFP
- ICT
- non-ICT
IV. Other
5
Sources of Labour Productivity
Acceleration in the US
2
1.6
1.2
0.8
0.4
0
-0.4
100%
80%
60%
40%
20%
0%
-20%
Gordon
Other
TFP in ICT sector
Total (R.h.s.)
Olin-Sich Jorg-HoStir
Labor quality
Cap. Deep. non-ICT
CEA
Cap. Deep. ICT
TFP in non-ICT sector
6
Summary of decomposition
exercise results
I. Capital Deepening increased
- Adoption in ICT capital input
accelerated
- Investment in non-ICT capital input
decelerated or stalled
II. Labour quality stalled
III. TFP growth accelerated
-
Increase in ICT sectors productivity
- And, to a lower extent, in non-ICT sectors
7
The First Determinant of Acceleration:
Increase in ITC Investment
(change of ITC investment components in percent
of private fixed investements)
30
20
10
0
1960
1970
Computer
1980
Software
1990
Comm. Equip.
2000
8
Labour Productivity in the US:
Stylized facts
Data and further decomposition define a clear
and uncontroversial set of “stylized facts”
I.
The first determinant
(for relevance) of
productivity acceleration is the strong increase in
ICT investment (ICT capital deepening)
II. The
second
determinant
of
productivity
acceleration is the significant increase in TFP in
ICT production sector
III. The third and last determinant of productivity
acceleration is the weak increase in TFP in the
non-ICT sector
9
International Comparison of
Labour Productivity
International comparison on the 1995-2000 period with
respect to labour productivity points out the following
elements:
Increasing divergence in GDP growth rates
10
8
6
4
2
0
-2
-4
1971
1976
1981
JAP
1986
1991
USA
EU15
1996
2001
10
International Comparison of
Labour Productivity
Differences in labour productivity
have remained almost stable or
slightly increased.
Labour productivity can be
measured in two different ways
In terms of output per employed person
In terms of output per hour worked
11
Labour Productivity
output per person
employed
output per hour
worked
6
6
4
4
2
2
0
0
-2
-2
-4
1980
1990
USA
EU
2000
JAPAN
-4
1980
1990
USA
JAPAN
2000
EU
12
Different statistical sources in
measuring labor productivity lead to
different pictures US
Labour productivity relative to EU 15 (=100)
OECD
EUROSTAT
130
130
120
120
110
110
100
1990
100
1990
1995
2000
GDP per hour
GDP per person
1995
2000
GDP per person
GDP per hour
13
Open Issues
I. Contradictory pattern of labour
productivity
How is it possible that the US leadership
relative to EU both in GDP growth and ICT
investment is not mirrored in the dynamics of
labour productivity?
II. Determinants of labour productivity
acceleration.
Is the weak increase in TFP in the non-ICT
sector
due
to
the
narrow
industrial
reorganization in this sector or is it due to the
presence of other factors?
14
GDP growth unpacked
1992-1995
1996-2001
3
2
1
0
-1
EU15
Employment
GDP per hour
USA
EU15
USA
Hours per person
Avg GDP% growth
15
Structural change:
an explanation of the weird
productivity dynamics …
In all advanced economies
–
employment has moved towards lower productivity sectors
(personal services - Denison effect reversed);
–
productivity % growth in low productivity sectors tends to be
smaller (Baumol effect)
In the US, faster GDP growth and exchange rate
induced shifts in terms of trade (from tradables to
non-tradables), might have been associated with
a wider shift towards low productivity sectors, thus
offsetting productivity gains in ICT intensive
manufacturing and professional services.
16
Sectoral dynamics in the US
THE CASE OF Community, Social and Personal Services
(excluding P.A.-Defence-Compulsory soc. Security & housework)
•Relative weight in employment (1995-hours) : approx. 15%
•Employment growth (1996-2001) 18% : 73% higher than whole economy avg.
•Relative productivity (1995; $/Hour) : 35% lower than whole economy avg.
•Dynamics of productivity (1996-2001) : 20% lower than whole economy avg.
Resulting yearly percentage reduction in productivity growth: -4.5%
US VS. EU EMPLOYMENT SHARES (PERSONS) & DYNAMICS
US
share
2000 2000-1991
Euro-3
2000-1995
share
2000
2000-1991
2000-1995
Manufacturing
12.6
-0.2%
-0.6%
19.8
-16.6%
-2.3%
Services
78.3
21.2%
12.0%
69.1
13.8%
10.3%
- Comm.soc. & pers. Serv.
31.8
15.6%
8.6%
29.7
11.3%
5.9%
- Wholes. & ret. Tr; Repairs
23.3
16.9%
8.4%
15.1
4.9%
6.2%
Data availability and quality do not allow a robust comparison
17
… that calls for a choice on
indicators and improvements in
data production …
 Productivity (GDP growth)
• Divided by Hours worked vs. employment
Hours worked are the effective labour input
need for reliable data on hours worked
• PPP adjusted vs. real terms ?
This choice is more controversial; exchange rate
movements impact both on relative Gdp levels and,
in each country, on terms of trade between
tradables and non tradables;
 More attention to sectoral patterns
these would allow a more clear view on the impact
of ICT on the dynamics of productivity
Need for sectoral measures
18
… while posing other questions
GROWTH differentials remain to a
large extent unexplained
 we have to look at additional channels
by which ICT and other factors might
have influenced competitiveness and
relative performance
 ICT direct and indirect impacts
 Other (Political & Economic) country
specific determinants
19
ONE STEP BACK
The G7 as a group: a falling performance
6
GDP Growth 1970-2002
5
G7 3y MA
G7 BC AVG
World (PPA) 3y MA
World (PPA) BC AVG
4
3
2
1
and vs. the others
over time
0
1970
1974
1978
1982
1986
1990
1994
1998
2002
20
ADVANCED AND ESPECIALLY TRANSITION
ECONOMIES LOST SHARES IN WORLD GDP
The emergence of China as a NEW ACTOR
IN THE GLOBALISING ECONOMY
13
51
Shares in World GDP (PPP weigthed)
12
50
11
49
10
48
9
47
8
46
7
45
6
44
5
43
CHINA (Mainland)
Other emerging Asia
4
42
Ex-Comecon
G7 (rhs)
3
41
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002
21
Large performance differentials in the G7
4.5
1971-80
1981-90
1991-2000
2001-2002
US takes the lead, Japan falls behind
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
USA
EU15
UK
France
Italy
Ge rm any
Jap an
WINNERS: US & Canada, UK, France
LOSERS: Japan, Germany, Italy
22
THE US ECONOMY HAS SHOWN
TO BE MORE COMPETITIVE
Export dynamics (1992 =100)
180
NATIONAL CURR. 1995 Prices
260
170
ECU/EURO Crurrent Prices
240
160
220
150
200
140
130
180
120
160
110
140
100
Germany
Italy
USA
90
80
1992
1994
1996
France
UK
Japan
1998
2000
2002
120
100
1992
1994
1996
1998
2000
2002
23
THE US ECONOMY HAS SHOWN
TO BE MORE COMPETITIVE
12.5
Shares in world merchandise exports
11.0
US
D
9.5
JP
8.0
F
UK
6.5
IT
5.0
CA
3.5
1990
1995
2000
24
This is largely explained by
ICT direct effects…
• Market share of ICT products from 1992 to 2000
has passed from 11 to 17.5% of world exports
• The US has maintained a first rank share on
world ICT exports:
• Slight decline. from about 14 to 12%
• Summing up US & Mexico (where most exports origin from
subsidiaries of US companies), this percentage grows from
14.4 to 17.2%
• The weight of ICT products on US export has increased 6.3
percentage points, from 11.4 to 17.7%
• France and UK have also performed well, while
Germany, Italy and, especially Japan, have lost
shares
25
ICT shares in world exports and
countries shares
100%
18
Other Countries
MEX
80%
16
US
UK
60%
14
Nordic (FINS)
E-3
40%
12
20%
10
0%
8
1987
1992
1996
2000
JAPAN
NEW TIGERS
(ML+IN+TH+PH+SI)
OLD TIGERS
(TW+KO+HK)
CHINA
ICT in % of world X
26
… but we can also imagine indirect
(spill-over) effects on other sectors..
an INCREASE in market shares for ICT intensive products could
be a proxy for ICT induced process and product innovations
That could be the case of Machinery and equipment
(World export shares)
JP
US
2000
E-4
1996
1992
10
12
14
16
1830
32
34
36
38
40
42
THE US HAVE INCREASED THEIR SHARE NOTWITHSTANDING
LOWER THAN AVERAGE PRODUCTIVITY GAINS
27
…and also other elements related to
performance, though not explicitly to
productivity and competitiveness
Cons. coll eu15
0.7
Cons. coll us
The US vs. EU: Key differences in the
demand components of GDP
0.6
0.5
0.4
0.3
0.2
0.1
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
-0.1
3.5
Cons.priv. eu15
3.0
Cons.priv. us
2.5
2.0
1.5
1.0
0.5
0.0
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
-0.5
1992
2.5
Invest.F.L. eu15
2.0
Invest.F.L. us
1.5
1.0
0.5
0.0
-0.5
-1.0
2002
2001
2000
1999
1998
1997
1996
1995
1994
-1.5
1993
These last two were possible ‘burning’ some
resources thanks to the accumulated factors surplus
0.0
1992
• Earlier and fast fiscal adjustment
• stronger (and compensatory)
private consumption
• A sustained investment cycle
28
…and also other elements related to
performance, though not explicitly to
productivity and competitiveness
Amongst EU losers, instead...
– Italy: Public debt constraints
 reduced contribution of Collective Consumption
to GDP growth
– Germany: Unification costs
low growth content of public expenditure
+ many other features common to EU, from market
conditions, to high sub-regional disparities, and to
integration of CCs.
we ought to consider the role of cyclical and other
aspects, also on ICT diffusion profile and, subsequently,
on performance
29
Summing up
• ICT have had a positive but manifold impact on
recent performance differentials
• These do not (yet) appear clearly in productivity,
when we consider it in aggregate terms
• Instead, a clearcut competitive advantage can be
elicited for ICT producers vs. users
• Other (temporary?) factors have influenced
performance
• DATA REQUIREMENTS & INDICATORS:
– Sound, harmonised worked hours statistics (in the
pipeline) for relevant sectors -->[productivity]
– small areas estimates for ICT usage (survey design
improvements) --> [networking cum sectoral dimensions]
– Impact measures and indicators (that is partly w/in Nesis)
– accompanying sets of indicators that integrate ICT with
competitiveness patterns
30