Download Haskel_BoE_roundtable_capacity_10Dec10

Inflation and capacity utilisation:
some issues of measurement
and an intangibles perspective
Jonathan Haskel
Imperial College Business School and CEPR
[email protected]
Bank of England presentation, December 2010
1
Capacity utilisation: the puzzle
Output per hour indicates spare capacity……but surveys do not
Output per hour
Capacity utilisation surveys
Source: Bank of England inflation report
(a) Output per hour.
(b) Pre-recession trends are calculated by projecting forward labour productivity from 2008 Q2 using the
average quarterly growth rate between 1996 Q1 and 2008 Q1.
2
Is trend productivity growth lower than we think?
Extrapolation above includes late 1990s. Productivity growth
in the late 90s:
1. Seems unusually high
2. Revised up strongly with FISIM in BBook09
Labour
Productivity
Growth (% p.a)
Contribution of
Human Capital
Deepening (%
p.a)
Contribution of
Tangible Capital
Deepening (%
p.a)
Without intangibles
1990-1995
1995-2000
2000-2007
2.94%
3.25%
2.65%
0.20%
0.29%
0.19%
1.09%
0.89%
0.86%
With intangibles
1990-1995
1995-2000
2000-2007
2.94%
3.53%
2.69%
0.17%
0.25%
0.17%
0.95%
0.74%
0.68%
Source: Haskel et al (2010), work for NESTA innovation index
Contribution of
TFP Growth (%
Intangible Capital p.a)
Deepening (%
p.a)
1.66%
2.07%
1.60%
0.64%
0.67%
0.55%
1.19%
1.87%
1.30%
3
Is current productivity mismeasured?
• 1. Hours
– Labour hoarding suggested by
• Productivity fall and
• limited hours fall relatively large output fall
– Seems counter to popular discussion of flexible labour market
– Possibility: are hours overstated?
• Levels of hours overstated in LFS especially for non-manual
professions
– 30% of LFS hours responses are imputed
– comparison with employer-based ASHE shows over-reporting
– (note US data are employer-based)
• No evidence on cyclical variation, but if downsizing
disproportionately affect manual occupations, overstatement might
rise
4
Is current productivity mismeasured?
(Contd)
• 2. Output
– For around 17% of services value added deflators are not based
on prices but costs
– =(% ch earnings- %ch service sector productivity)
– Measured service sector productivity
• Rises in booms
• Falls in recessions
– Thus via deflators: measured real service sector output
• Rises in booms
• Falls in recessions
– Thus, some tendency for
• Current measured service productivity is too low
• Past measured service productivity was too high
• So service sector is closer to true productivity trend than measured:
output gap is lower
5
Is current productivity mismeasured?
(contd)
•
3. Output: financial services
–
FISIM has cushioned the GDP fall
•
•
Directly measured financial services volume has fallen
But volume of deposits/loans hardly changed, margins have risen, so share of real FISIM
has risen
1.50%
1.00%
0.50%
0.00%
-0.50%
2007
Q2
2007
Q3
2007
Q4
2008
Q1
2008
Q2
2008
Q3
2008
Q4
2009
Q1
2009
Q2
2009
Q3
2009
Q4
2010
Q1
2010
Q2
-1.00%
-1.50%
-2.00%
-2.50%
-3.00%
GDP without FISIM
GDP with FISIM
6
Future productivity trends
• Long run productivity trends drive by (intangibles view)
•
•
•
•
Physical capital
Intangible capital (knowledge = software, R&D, design, training etc.)
Labour
Technical progress
– Years 2000-07 dln(Y/L) driven by
•
• Intangibles = 20%
• Computers/tangible = 25%
• Labour composition = 5%
• TFP = 50%
Additional questions raised by intangibles
– Manufacturing v services. Intangible investment in manufacturing > tangible
investment. So manufacturing becoming more like services. Doubts over
interpretation of capital utilisation in services spill over into manufacturing?
– If future increase in uncertainty/cost of capital, intangible investment falls?
– What has happened to intangible K/L in the recession: does this help explain and
predict future Y/L?
7
Intangible v tangible investment when tangible
falls: 1991-2 recession
100
90
S/ware
80
70
R&D
60
50
40
Adv
30
20
Tang
10
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
1975
1974
1973
1972
1971
1970
0
Intangible investment holds up. But, intangibles depreciate faster than tangibles,
so net effect on stocks and hence contributions to dln(Y/H) ends up similar…
8
Contributions of intangibles and computers
to dln(Y/H)
0.014
0.012
0.01
0.008
cont comp
0.006
cont intan
0.004
0.002
0
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Contributions show similar paths post-recession even though investment in
intangibles is less cyclical.
9
Summary
•
Service sector productivity
– True output might be
• Higher due to deflator effect
• Lower due to FISIM
– True hours might be less
•
Long run productivity trend may be slightly lower than we think
– because 1990s overstated
– if higher cost of capital in the future
•
But what does inflationary pressure via capacity utilisation mean in a service
economy?
– Y* in a car company : capital => at Y* factory is full
– Y* in a design company?
• Capital
– = a MacBook per designer
– = Ideas from designer teams/experience
– So even if Y close to Y*, less inflationary pressure if
•
•
•
•
Supply of MacBooks is elastic
Teams have been hoarded
Immigration of high-skilled labour is relatively easy
Service tasks can be out-sourced
10