THE CONTRIBUTION OF INVESTMENT TO GROWTH*
| DOI | http://doi.org/10.1111/j.1467-9485.1981.tb00087.x |
| Author | M. FG. Scott |
| Published date | 01 November 1981 |
| Date | 01 November 1981 |
Scorrish
Journal
of
Political
Economy,
Vol.
28,
No.
3,
November
1981
0
1981
Longman
Group
Limited
0036-9292/81/00180211
$02.00
THE CONTRIBUTION
OF
INVESTMENT
TO GROWTH*
M.
FG.
SCOTT
Nufield
College,
Oxford
At the end of their masterly survey of the applied economics of technical
progress, Kennedy and Thirlwall sum up as follows
:
Macro-studies of technical progress which seek to estimate the rate
of
technical progress as a residual component of the growth of output are
fraught with aggregation and identification difficulties but almost in-
variably find technical progress as the prime determinant of the rate
of
growth, regardless
of
the production function specified (Kennedy and
Thirlwall, 1972, p. 62).
They mention several early studies, “But what really set the field alight were
the findings of Fabricant (1954), Abramovitz (1956) and Solow (1957) that
between
80
per cent and 90 per cent of the growth of output per head in the
American economy over the previous decades could not be accounted for by
increases in capital per head and must therefore be due to some form
of
technical progress”
(op.
cii.
p. 17). Despite many subsequent criticisms and
qualifications, Kennedy and Thirlwall evidently feel that, even if
80
per cent or
90 per cent is an exaggeration, at all events technical progress has accounted
for the largest single fraction of the growth in output per worker, and
so,
by
implication, that increases in capital per worker resulting from investment
have accounted for a smaller fraction.
I
shall try to demonstrate that all this is highly misleading. It makes much
better sense to say that,
so
far from investment accounting for a minor part
of
the growth of output per worker, it accounts for all or virtually all of it. The
proponents of the orthodox view have erred in several ways, and, while some
ofmy disagreement with them may be called semantic, there is little doubt that,
whether semantic or not, there are important implications for economic
science and economic policy.
To
guard against misunderstanding
I
should add that the ensuing argumenl
is not at all the same as that of Jorgensen and Griliches (1967), who attempted
to show that, for the United States in the period 1945-1965, there was virtually
no “residual” productivity growth if inputs were properly measured. This
attempt was successfully challenged by Denison (1969), who showed that theiI
*Sixteenth Annual Lecture
of
the Scottish Economic Society delivered at the University
01
Glasgow on
6
May
1981.
The
Society acknowledges with gratitude the financial
support
provided
by Shell International
for
the Annual Lecture.
15
71
1
212
M.
FG.
SCOTT
own measurements were faulty in several respects. My argument is not
confined to a particular period in a particular country, and uses
a
different
concept of investment from that of Jorgensen and Griliches. It is thus both
more general and,
I
believe, more fundamental.
One can make
a
two-fold classification of investment
:
human investment
and material investment. The meaning of each will emerge more clearly in
what follows. For the most part
I
am concerned with the contribution
of
material investment to growth, but
I
return at the end to the contribution
of
human investment as well. In my view, both contributions have been seriousiy
underestimated.
The idea that technical progress is costless and exogenous to the economic
system, “manna from heaven”,
is
probably not seriously held by most
economists who have thought about the matter, even though it is the basic
assumption of a great deal of growth theory. If it is discarded, however, one
must ask what it is, exactly, that is meant by “technical progress”, and how it is
to be distinguished from investment. According to Kennedy and Thirlwall
:
Technical progress in all its aspects is impossible to measure precisely but its
essential quantitative characteristic is to shift the production function
(embodying all previously known techniques) enabling greater output to be
produced with the same volume of inputs, or the same output with lesser
inputs
(op.
cit.
p.
12).
However, they then go
on
to say that it is
not
manna from heaven, and that,
by and large, it “does not occur by accident but through the deliberate
diversion of resources to activities which generate progress in pursuit of fame,
profit or both. More likely than not, it also requires inputs for its embodiment.
Thus for technical progress to take place three main types of inputs may be
distinguished
:
first, research-type inputs
;
secondly, knowledge distribution
inputs (e.g. education) and thirdly, inputs required for changing over to
improved industrial methods. It is this last input, the cost of shifting from one
state of knowledge to another, that the theorists have picked on as the serious
omission in neoclassical theory”
(op.
cit.
p.
13).
Now this does seem to me to give the game away completely. If technical
progress requires inputs, and if, amongst those inputs are included
all
the costs
of “shifting from one state of knowledge to another”, then
I
cannot see the
difference between technical progress and investment. For what is investment?
The definition which
I
believe to be the most useful for analysing economic
growth is precisely the cost of change (Scott,
1976).
In
a static economy all
output would be consumed and investment would be zero. There would,
of
course, be expenditure on maintenance
(of
which more later), but that should
be regarded as part
of
the current costs
of
production. If output
is
to grow, the
static economy must be changed. It seems to me inevitable that this will
involve incurring expenditures which need not be incurred in a static economy,
for the building of new buildings, roads, vehicles, machines, etc., and the
improvement of existing ones, for moving labour from places where its
To continue reading
Request your trial