Measuring Productivity: Conference On Next Steps For The Japanese SNA Tokyo, 25 March 2005
Measuring Productivity: Conference On Next Steps For The Japanese SNA Tokyo, 25 March 2005
Measuring Productivity: Conference On Next Steps For The Japanese SNA Tokyo, 25 March 2005
Measuring Productivity
Paul Schreyer, OECD
1
Contents of presentation
1. Why is productivity important?
2. Concepts: labour and multi-factor productivity
3. Measuring output
4. Measuring labour input
5. Measuring capital input
6. OECD Productivity database
7. OECD Factbook
8. Conclusions
2
“Productivity isn’t everything
but in the long run it is almost
everything”
Paul Krugman The Age of
Diminishing Expectations
3
1. Why is productivity important?
• Basis for improvements in real incomes and
economic well-being:
• Monetary policy (inflationary pressures)
• Fiscal policy (financing of health, education,
welfare)
• Slow productivity growth = conflicting demands
for distribution of income more likely
4
2. Concepts: labour productivity
Q
Labour productivity (LP) =
L
• Indicates how efficiently labour is used in production
• Not necessarily an indicator of effort per worker
5
Table of productivity measures
6
2. Concepts: multi-factor productivity
Qt Lt Kt At
ln t −1 = s L ln t −1 + s K ln t −1 + ln t −1
Q L K A
% change % change % change % change
in output in labour in capital in MFP
input input 7
MFP and labour productivity
Qt Lt Kt Lt At
ln t −1 − ln t −1 = (1 − s L ) ln t −1 − ln t −1 + ln t −1
Q L K L A
8
MFP and labour productivity
9
Decomposition of labour productivity, Canada
10
MFP – framework
11
MFP framework: measurement issues
Tricky areas:
-Implicit pricing
-Non-market
services
-Quality change
12
Measuring output: 3 tricky areas
1) Output measures for industries where there is
implicit pricing so that even nominal output
measures are debated:
• Banking (FISIM, how to deal with new
financial instruments and risk?)
• Insurance (how to deal with catastrophes?)
Î OECD Taskforce recommends to measure
insurance output as premiums minus expected
claims, not actual claims
Î Otherwise sharp drop in output when there is a
13
catastrophe
Measuring output: 3 tricky areas
2) Output measures for industries where technology
moves fast and where rapid quality change makes
price measurement difficult
• Example ICT: computer production,
communication services
3) Output measures for industries with non-market
producers
• Traditionally from input side
• But increasingly demand for output-based
measures Î Atkinson Report UK
14
Measuring output: 3 tricky areas
Chart 5. Real value added per employed person in the health and
social work industry
Indices, 1982=100
140
130
120
110
100
90
80
70
1982 1987 1992 1997
15
Source: OECD.
Multi-factor productivity framework:
measurement issues
Comparability of
GDP growth
rates:
-Hours actually
Comparability of worked
GDP growth
rates: -Hours worked
by
-Non-market -Measuring user occupation/skill
services costs
-High-tech -Capital
industries -Splitting mixed measures by
income industry
17
4. Measuring labour input
• Ideal measure: total number of hours worked,
broken down by type of labour input
• Number of persons employed or number of jobs
can generate biased measures of productivity if
hours per person change or if there are multiple
job holdings
• Total hours should comprise the hours of
employees, the hours of self-employed and the
hours of unpaid workers (e.g. family members in
agriculture)
18
Example: impact of different measures of labour
input on labour productivity, France
Industry (mining, manufacturing and construction) Market services
220.0 145.0
Value-added per Value added per
140.0
hour: 3.7% per hour: 1.8% per
200.0
year 135.0 year
180.0 130.0
Value-added per
employee 125.0
Value-added Value-added
160.0 (headcount): 3.3%
per full- 120.0 Value-added per per full-
per year
time person employed time
140.0 equivalent 115.0 (headcount): equivalent
person: 1.3% per year person:
Value-added per 110.0
120.0 3.4% per Value-added per employee 1.5% per
person employed
year 105.0 (headcount): 1.0% per year
(headcount): 3.4%
year
100.0 per year 100.0
78
80
82
84
86
88
90
92
94
96
98
78
80
82
84
86
88
90
92
94
96
98
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
*Output is me a s ure d a s a qua ntity inde x of va lue -a dde d.
Source: INSEE.
19
Labour input: statistical issues
22
Productive capital stock of a particular asset type
K t = I t + h1I t −1 + h 2 I t − 2 + ... + h T I t −T
23
Capital services of a particular asset type
Growth of capital services of a particular asset type =
Growth of productive stock of this asset type
St Kt
ln = ln because S t = λK t
St −1 K t −1
Normally assumed as
constant
24
Price of capital service = user costs
Rental price
or user cost Rate of
Required rate depreciation
or return
25
Price and quantity of capital services
User costs and the rate of capital service for each asset
type are brought together to form an index of overall
capital input = the rate of change of total capital
services
Kt K t
ln t −1 = ∑iw i ln t −1
t i
K Ki
Average share of each
asset in total value of
capital services
26
Capital services and net capital stock, Australia
27
Capital services and national accounts
29
6. OECD Productivity database
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
D Ita 5.0
en ly
m
ar
Ja k
p
B e an
lg
Sw ium
e
G de
N er m n
U eth an
ni
te erl y
d a
Ki nd
ng s
do
U Ca m
ni
1990-95
te nad
d
St a
at
F r es
an
1995-01 Po ce
rtu
A u ga
st l
ra
G lia
re
e
Average rates of MFP growth
Fi ce
nl
an
Ir e d
la
nd
31
7. OECD Factbook
33
Thank you for your attention!
34