(kaldor) productivity and growth
TRANSCRIPT
-
8/10/2019 (Kaldor) Productivity and Growth
1/8
The Suntory and Toyota International entres for Economics and Related Disciplines
Productivity and Growth in Manufacturing Industry: A ReplyAuthor(s): Nicholas KaldorSource: Economica, New Series, Vol. 35, No. 140 (Nov., 1968), pp. 385-391Published by: Wileyon behalf of The London School of Economics and Political Scienceand The
Suntory and Toyota International Centres for Economics and Related DisciplinesStable URL: http://www.jstor.org/stable/2552347.
Accessed: 09/12/2014 19:56
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at.http://www.jstor.org/page/info/about/policies/terms.jsp
.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
.
Wiley, The London School of Economics and Political Science, The Suntory and Toyota International Centres
for Economics and Related Disciplinesare collaborating with JSTOR to digitize, preserve and extend access toEconomica.
http://www.jstor.org
This content downloaded from 132.248.180.231 on Tue, 9 Dec 2014 19:56:27 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/action/showPublisher?publisherCode=blackhttp://www.jstor.org/action/showPublisher?publisherCode=lonschoolhttp://www.jstor.org/action/showPublisher?publisherCode=suntoyhttp://www.jstor.org/action/showPublisher?publisherCode=suntoyhttp://www.jstor.org/stable/2552347?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2552347?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=suntoyhttp://www.jstor.org/action/showPublisher?publisherCode=suntoyhttp://www.jstor.org/action/showPublisher?publisherCode=lonschoolhttp://www.jstor.org/action/showPublisher?publisherCode=black -
8/10/2019 (Kaldor) Productivity and Growth
2/8
1968]
Productivity and Growth in Manufacturing
Industry: A Reply
By
NICHOLAS KALDOR
Professor Wolfe attacks
me
(in
the
May,
1968
issue of Economica) for
being
obscure
and
unconvincing
in the
inaugural
lecture
which I gave
at Cambridge.1In this lecture I attempted to put forward a complex
thesis concerning the causes
of
high
and
low rates of economic growth
under
capitalism,
and
this
was
necessarily somewhat
scanty;
there is a
limit
to the
material
one can
pack
into a
single lecture. Such a treatment
could
only
be successful
if
received
with
imagination
and some
goodwill;
Professor Wolfe
has
picked
on a
large
number
of
individual points
(many
of them trivial
in
substance) without attempting to come to
grips with
the
thesis as a whole.
I
propose
to
publish a more compre-
hensive
statement
of the
theory
in
due
course.2
Meanwhile rather than
follow the negative approach of answering each point in turn-which
would
be both tedious and unconstructive-I
shall concentrate on the
main points which have clearly
not been
understood.
Foremost
among
these is
my
notion
of
economic
maturity .
This is
not
some vague
notion which
can be defined in
terms
of
a
situation
in
which
there is
relatively
small
employment
in
agriculture .
In
my
lecture
I
defined
it
explicitly
as
a
state
of
affairs where real income
per
head
had reached
broadly
the
same level
in the
different
sectors of
the
economy .
I could have
added,
to
convey
its
significance better,
that
''economic maturity
could also
be defined
as the
end
of the
dual
economy ;
or a situation
in
which
surplus
labour
is
exhausted;
or one
in
which growth
with
unlimited
supplies
of
labour
(to
use Arthur
Lewis'
phrase)
is
no
longer possible.
The
neo-classical
framework
of
thought
cannot accommodate
no-
tions
like
disguised unemployment ,
the dual
economy ,
or the
distinction
between
capitalist
and
pre-capitalist enterprise.
For
neo-classical
theory
assumes that
the structure of
demand determines
the distribution of resources between different uses; that competition
and
mobility
assures that
factor
prices
tend to
equality
in
all
employ-
ments;
that
profit
maximization
ensures
equality
of
factor
prices
with
the
value
of
the
marginal products
of
factors;
subject
only
to
friction
1
J.
N. Wolfe, Productivity
and Growth
in
ManufacturingIndustry:
Some
Reflections
on Professor
Kaldor's Inaugural
Lecture , Economica, vol.
XXXV
(1968), pp.
117-26;
N.
Kaldor,
Causesof the
Slow Rate of EconomicGrowth
of the
UnitedKingdom,Cambridge
University
Press,
1966.
2 A
somewhat more
extended
exposition
was
given
in
three lectures
(subse-
quently published)
at Cornell
University: cf. Strategic
Factors in Economic
Development, thaca, New York, 1967.
385
This content downloaded from 132.248.180.231 on Tue, 9 Dec 2014 19:56:27 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 (Kaldor) Productivity and Growth
3/8
386
ECONOMICA
[NOVEMBER
etc.,
each factor
will tend to be used where
is makes the
greatest
contribution
to the
national
product.
It
would be generallyagreed
that these
assumptions
are
at their
most
inappropriate
n
the
case of an
under-developed
country,
or
a
country
in
the
earlier stages
of industrialization.
In such countries high
and
low
earnings
sectors
exist side
by
side;
there
are vast amounts
of
surplus
labour
or disguised unemployment
in
the
low-productivity
sectors,
so
that labour can
be withdrawn from
them without adverse
effects
on
the
output
of
those
sectors;
and the
supply
of labour in the
high-
productivity,
high-earnings
sector
is
continually
in
excess of the
demand,
so
that
the
rate
of
labour-transference
rom
the
low to the
high
produc-
tivity sectors is governed only by the rate of growth of the demand for
labour
in the latter.
In fact the
size
of the labour force in the
non-
industrial
sector
is
a
residual-entirely
determinedby
the total
supply
of
labour
on the
one
hand and the
requirements
or
labour
in
the
industrial
sector
on
the other hand.
The best definition
I could suggest for
the
existence
of labour
surplus
in
this
sense
is
one which is analogous
to
Keynes'
definition
of
involuntary unemployment :
a
situation
of
labour
surplus
exists when a
faster
rate
of increase
in
the
demand for
labour
in the
high-productivity
sectors
induces a faster rate
of
labour-
transferenceeven when it is attendedby a reduction,andnot an increase,
in
the
earnings-differential
etween
the
differentsectors.
For
reasons that
I
explained
in
my lecture,
the rate
of
growth
of
industrialization
fundamentallydepends
on
the
exogenous
components
of demand
(a
set
of
forces extending far beyond
the
income
elasticities
of demand
for manufactured
goods).
The higher the rate
of
growth
of
industrial
output
which
these demand conditions
permit,
the faster will
be
the
rate
at which labour
is
transferred
rom the surplus-sectors
to
the
high productivity sectors. It is my contention that it is the rate at which
this transfertakes place
which determines
the
growth
rateof productivity
of
the
economy
as a whole. The
mechanism by which this
happens
is
only
to
a
minor extent
dependent
on
the absolute differences
n
the levels
of
output per
head
between
the labour-absorbing
sectorsand the
surplus-
labour sectors.
The
major
part
of the mechanism
consists
of
the
fact that
the
growth
of
productivity
is accelerated as
a
result
of
the
transfer
at both
ends-both
at
the
gaining-end
and at the losing-end;
in the
first,because,
as a
result of
increasing
returns,
productivity
in industry will
increase
faster, the faster output expands; in the second because when the
surplus-sectors
lose
labour,
the
productivity
of the remainder
of the
working population
is
bound
to
rise.'
In
the
literature,
the
surplus
labour
sector is generally
thought
of
'Indeed,
the
existence f
disguised
nemployment
s
by itself
capable
f explain-
ing
theseresults, ven
in the absence
of increasing
eturns,
incethe
increase n
industrial utput,
brought
boutby labour
ransference,
ill be
a net addition
o
the GNP-there
will
be
no
compensating
reduction in output elsewhere.
As
a
matter
of historical fact,
I
am
convinced,
however,
that the growth
of productivity
resulting from increasingreturns (both internal and external) has been a very
important
part of
the picture.
(On this see below.)
This content downloaded from 132.248.180.231 on Tue, 9 Dec 2014 19:56:27 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 (Kaldor) Productivity and Growth
4/8
1968]
PRODUCTIVITY/GROWTH
N MANUFACTURING
INDUSTRY 387
as agriculture.
This
is
because
in the
early
stages
of
capitalist
develop-
ment
much
the
greater part
of the
population draws its living
from agri-
culture. However, disguised
unemployment
in services
had been just
as prevalent-in Victorian England (as in present-day India or Latin
America)
there
were vast numbers of
people
who eked out
a
living
in
urban
areas
as
hawkers, petty
tradesmen,
servants,
etc.
on
very low
earnings.'
In the field
of
services
however (unlike
in
agriculture)
there
are two
contrary
processes
at work: on the one hand
industrialization
absorbs labour
from services on
a
large
scale;
on the other hand, the
growth
of
industry
itself
gives
rise to the
growth
of
services
of
various
kinds
which
are
both
complementary
and
ancillary
to
industrial
activities (by ancillary I mean that the demand for these services,
e.g. transport,
distribution,
accountancy,
banking services, etc. are
derived
from,
but cannot
generate,
industrial
activities).
As
a result
the
total
employment
in
services tends
to rise
during
the
process
of
indust-
rialization though
less
(in
relation
to the
growth
of total
output) when
the growth
in total
output
is
relatively
fast.
While
it has
long
been known that labour
has
no
opportunity
cost
in
an
under-developed
country-the absorption
of
labour
through the
growth
of
industry involving
no
reduction
in
output
elsewhere-it has
not been generally recognized that the same applies to most of the so-
called
advanced countries with
relatively
high
incomes
per
head.
The
view that
growth
rates,
even
in advanced
countries,
are
dependent
on
the
rate at
which
labour is transferred nto
manufacturing
from
other
sectors
would
find
confirmation,
in the
first
place,
if
over-all
growth
rates
are
positively
associated
with rates of increase
in
employment
in
manufacturing.
This
is shown
for the
group
of
twelve
advanced
countries
given
in
Table 3 of my lecture for the period 1953/4-1963/4 by the following
:2
(1)
6
=2-665+
1-066,M
R2=.828
(015)
where
G
is
the rate
of
growth
of
GDP
and
PM
is the rate
of
growth
of
employment
in
manufacturing.
This
result
confirms
my general
hypothesis
unless
it
could
be shown
that
growth
rates
in
manufacturing employment
are
themselves
closely
relatedto growth rates of total employment so that the former could be
regarded
as a
proxy
for the latter.
Such
positive
association seems
1
This relates
to both
self-employed
and
employees
alike.
In the
population
Census
of
1891,
15-8
per
cent.
of
the
occupied population
of Britain were classified
as domestic
servants.
In
the
Census
of
1961
the
figure
was
14
per
cent. This
reduction
cannot
be
explained
n terms
of
a shift
in
consumer
preferences
or
by
the
assumption
that
domestic
service is an inferior
good
with a
negative
income-
elasticity
of
demand;
it
can
only
be
explainedby
the
growing absorption
of
surplus
labour
in
the
economy
which resulted
in a rise in
wages
in
domestic service which
was
much
in
excess
of
the
general
rise in
wages.
2
The sources for this and the following equations are given in the statistical
tables of
my
lecture.
This content downloaded from 132.248.180.231 on Tue, 9 Dec 2014 19:56:27 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 (Kaldor) Productivity and Growth
5/8
388
ECONOMICA
[NOVEMBER
ruled out,
however, by the fact
that
there is no association
at all between
rates
of
growth
of
GDP
and rates
of
growth
of total employment:
(2) G=4-421 +O043
Eg
R2=*018
(0.994)
whereEg
is the
rate
of
growth
of total
employment.
The
positive
correlation
in
Eq.
(1) could only be consistent
with
the
absence
of
any
correlation
in
Eq. (2)
if rates of
growth
in over-all
productivity
are
positively
associated with rates
of
growth
of
employ-
ment
in
manufacturing and negatively
associated with
rates
of
growth
of
employment
outside
manufacturing.
This is duly
confirmed
by
the
following three equations:
(3) Pg=
1868
+0991
EM
R2=
677
(0-216)
(4) Pg
=
4 924
-1
8OONM
R2-=
427
(0.660)
(5)
Pg=2 899+0821 EM-1-183
ENM
R2=842
(0.169)
(0.387)
where
Pgdenotes
growth
rates of
GDP per person
in
civilian
employ-
ment,
andEM
and
ENM
denote
growth
rates
of
employment
n
manufac-
turing and
non-manufacturing
respectively.
It
follows
from the above also that
in
a mature economy
it
would be
idle to
look
for evidence
of
a labour constraint
in
manufacturing
either
in terms of a differentialrise
in
wages
in
manufacturing
or
in
the
relative
incidence
of
unemployment
and unfilled vacancies
in the different
sectors.
Since
the
supply
of
labour
to
industry
does
not become
inelastic
until wages in the restof the economy have risen to levels comparableto
those
in
industry,
an
effective labour
constraint
in
manufacturing
would
manifest
itself
in
a
general
increase
in
wages
throughout
the
economy,
and
in
low levels of
unemployment
in
all sectors.
In
a
situation
of
approaching
abour
shortage
one
would
expect
the
unemployment
rate
to
be
falling,
and
wages
to rise faster
in the
non-manufacturing
sectors
than
in
manufacturing.1
One
would expect,
furthermore,
that
in
a
mature economy con-
strained
by
a labour
shortage,
the
average
level of
unemployment
would
be higherin the manufacturingsector than in the rest of the economy,
and
not lower. This is because
such an
economy
is
almost
inevitably
1
This has characterized
he recent situation in the
United States where,
pari
passu
with a
large
increase
in
employment
in the
manufacturing
ector
(of
2-8
per
cent. a year in the last five
years), wages rose fasterin the low-paid sectors (whole-
sale and retail distribution
and agriculture)than in manufacturing. Cf. Annual
Report of the Councilof EconomicAdvisers,
Washington,D. C., 1968,pp. 109-10.)
It should also be pointed
out that
in
the United States
(where he level of unemploy-
ment
has been much
greater
han
in
the United Kingdomthroughout
the
post-war
period) earnings
in
manufacturing
are much higher
in relation to the national
average (and particularlyin relation to agriculture and the distributive trades)
than
in the
United Kingdom.
This content downloaded from 132.248.180.231 on Tue, 9 Dec 2014 19:56:27 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 (Kaldor) Productivity and Growth
6/8
1968] PRODUCTIVITY/GROWTH
N MANUFACTURING
INDUSTRY
389
subject
to
a
stop-and-go
cycle;
variations
in the pressure
of
demand
induced by
fiscal
and
monetary policies
affect demand and
employment
in industry
far more
than
in the non-industrial
sectors.
There
is finally
the
question
of the existence
of
increasing
returns to
scale
in
manufacturing
ndustries.
I
emphasized
in my
lecture
that this is
a
.macro-phenomenon
which (in
the
words of
Allyn
Young)
cannot
be discerned adequately
by
observing
the effects
of variations
in the size
of
a
particular
firm
or of
a
particular
industry .
Studies relating
to
the
cost
and
size of individual
plants
are not thereforenecessarilyrelevant;
though
there
is a growing body
of
empirical
evidence relating
to
individual
industries
which tends
to confirm
its
importance.-
Nor do recent studies which fit production functions by means of
linear multiple regressions
lead
to
any
confirmation
that
the
parameters
of a Cobb-Douglas
function
conveniently
add
up
to
unity.
Indeed in all
recent
studies
the
Cobb-Douglas
function
was assumed
to
be a con-
stantly
shifting
one,
of
the form
AectKaLb
nd
not AKaLb,
as
indicated
by
Wolfe;
and
in
many
of these studies
the result
a
+
b
=1
is a
tautolo-
gical
one which
gives
no indication
whatever
of the
presence
or
absence
of
economies
of
scale.2
Wolfe
was
mistaken
in
suggesting
that
the constant in the
regression
equation of output on employment is an errorterm . It is a constant
term,
which
reflects
explanatory
variables
that
were
excluded,
one of
which
may
be an autonomous
time trend.
It does not
follow, therefore,
that
the
introduction
of further
explanatory
variables
(such
as
capital
investment)
would
necessarily
reduce
the value of the
regression
coeffi-
cient
on
employment-if
the two
variables are
not
inter-correlated,
the
introduction
of a
capital
term should reduce
the
constant
term,
and
not
the
regression
coefficient
on
employment.
Since
owing
to the accelera-
tion principlethere is always some inter-correlationbetween the rate of
growth
of
employment
and
investment
activity,
one would
expect
a
multiple
regression
of
output
on
employment
and
capital
investment
to
show
lower
coefficients on
employment
than
a
simple
regression:
but
there
is no reason
to
expect
that
the
sum
of the
two
coefficients
in
the
one
case
should be
lower than
the
single
coefficient
in the other case. I
have
indeed
computed multiple
regressions
of
output
on
employment
and
investment,
and
they
show
that even
when the role
of
capital
is
taken
into
account,
the
regression
coefficient
of
output
on
employment
remains significantlygreater than one.3
I
See for
example,
J.
S.
Bain,
Barriers to
New
Comnpetition;
. Pratten and
R.
M.
Dean,
The
Economies
of
Large
Scale
Produiction
n
British
Industry-an
Introductory
tudy;
J. W.
Kendrick,
Productivity
Trends n
the
United
States;
as
well
as
the
sources quoted
in E.
F. Denison, Why
Growth
Rates Differ, chapter
on
Economies
of Scale.
2
The
existenceof
a
linear-homogeneous
production
function
is derived
not
from
observation
but
from
a
priori
reasoning (i.e. profit
maximization
underconditions
of perfect
competition)
and the parameters
are
either
restricted o
as to add
up
to
unity,
or else are directly
estimated from
factor
shares
which
by
definition
add
up to unity.
3
These multiple regressions
were
unfortunatelynot
completed
in
time to be
This content downloaded from 132.248.180.231 on Tue, 9 Dec 2014 19:56:27 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 (Kaldor) Productivity and Growth
7/8
390
ECONOMICA
[NOVEMBER
It would
be
wrong
to
suppose
that the
regression between
output and
employment
showing
a coefficient that is
significantly greater than unity
is a reflection
of
short-term
or
cyclical
influences. The
figures
in
my
inter-country studies related to averages of ten-yearly periods, taking
two-year
averages
for both the
base-year
and the
end-year.
This
virtu-
ally
eliminates
short-period
or
cyclical
influences.
The
short-period
relationship
between
output
and
employment-showing
in
some
cases
a
5:1 relationship-largely
reflects
changes
in
the
degree
of
utilization of
capacity.
But the fact that there is short
period relationship
of
5:
1
does
not
exclude
a
long-period relationship
of
1'5: 1.
It
certainly provides
no
support
for
assuming
that the
long-period
relationship
must be
smaller
than 1:1.
Nor
is
it
correct
to
suppose
that more than a
small
part
of
the
Verdoorn
Law can be
explained by
embodied technical
progress
coupled with
the association
of a
relatively high
level of
investment in
fast-growing
industries,
in the
absence
of
increasing returns.
Unless
one
assumes
that the
rate of
technical
progress
on
successive
vintages is
itself
accelerated as
a
result of a
larger
volume
of
investment
(as
in
Arrow's
learning function ,
which comes to the same
thing
as
increas-
ing returns)
the mere
postulate
of
embodied technical
progress
does
not
entail that the average rate of growth of productivity should be signifi-
cantly
faster in the
faster-growing
industries.'
Differences
in
saving propensities
(as
measured
by
differences
in
the
investment/output ratio)
cannot
account for more
than a
small
proportion
of the
observed
differences
in
growth
rates.
Wolfe is
correct
in
saying
that
I do not
regard
the
supply
of
capital
as
a
serious
limita-
tion
on economic
growth .
This is because
savings
and
capital
accumula-
tion
in
a
capitalist economy
do not
represent
an
independent
variable
-a faster
rate of
growth
induces
a
higher
rate of
investment;
it
also
brings
about
a
higher
share of
savings
to
finance
that
investment,
through
its effect on the share
of
profits.
It is therefore more
correct to
say
that a fast rate of
capital
accumulation is
a
symptom
of
a
fast
rate
of
growth
than
a
cause
of
it.
I shall refrain
from
comment
on the last
sections
of
Professor
Wolfe's
paper
concerning policy
recommendations ,
where he
charges
me
with
several
sins of both omission and
commission.
My
lecture
was concerned
with
the
question
why growth
rates differ between different
countries.
The SelectiveEmployment Tax, public sector expenditure,immigration,
birth
control,
etc.
were
outside
its
scope
and none of
these was
mentioned.
included in the statistical notes
appended to my inaugural
ecture;
they are, how-
ever, included in the American
version, Strategic
Factors in Economic Develop-
ment, op. cit.,
pp. 81-3. (In
these equations
the
capital
term is measured by the
investment/outputratio in industry, which gives-under
the assumption that the
marginal capital/output ratio is greater than unity-a
lower limit to the capital
term
of a
Cobb-Douglas
function. Direct measurementof
the
rate of growth of
the capital
stock
would
have
been impossible statistically
and pretty meaningless
theoretically.)
'
The reasons for this cannot be gone into here; theyare set out in a forthcoming
paper by
W. A.
H.
Godley.
This content downloaded from 132.248.180.231 on Tue, 9 Dec 2014 19:56:27 PMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/10/2019 (Kaldor) Productivity and Growth
8/8
1968]
PRODUCTIVITY/GROWTH N MANUFACTURINGINDUSTRY
391
Finally, I would commend to Professor Wolfe
the
following lines
from Pope's Essay oz
Criticisnm:
A perfectJudgewill read each work of Wit
With the same spirit that its author
writ;
Survey he WHOLE,nor seek slight faults to find
Wherenaturemoves, and rapturewarms he
mind;
....
In
wit, as nature,what effects our hearts
Is not th'exactnessof peculiarparts;
'Tis
not a lip,
or
eye, we beautycall,
But
the
joint
force
and
full
result
of
all.
King's College, Cambridge.
This content downloaded from 132 248 180 231 on Tue 9 Dec 2014 19:56:27 PM
http://www.jstor.org/page/info/about/policies/terms.jsp