ON THE THEORY OF DIVERSIFICATION: A REPLY
| Date | 01 February 1974 |
| Author | C. J. Sutton |
| Published date | 01 February 1974 |
| DOI | http://doi.org/10.1111/j.1467-9485.1974.tb00179.x |
Scottish
Journal
of
Political Economy
Vol.
XXI,
No.
1,
February
1974
ON
THE THEORY
OF
DIVERSIFICATION:
A
REPLY
C.
J
SUTTON
R.
M.
Grant criticizes my approach to diversification (Sutton, 1973) partly
because
of
its specific assumptions, but primarily bemuse of a general distaste
for
behavioural theories. He presents an alternative neo-classical model,
together with some empirical results.
There are several points at whioh it is possible to question the details of
Grant’s analysis. These include, for example, (i) whether the objectives of the
neo-classical
firm
under conditions
of
uncertainty are properly defined as
‘
profit maximization and
risk
minimization
’
(Grant, p. 79), and
(ii)
whether
a
behavioural assumption supported by explicit argument (Sutton, 1973, pp.
35/6) should be rejected
on
the basis
of
an assertion that a different assump-
tion
is
plausible (Grant, p.
77).
However, this note will concentrate
on
two
more general topics: a) the empirical tests, and b) the general oase for a
behavioural model
of
diversification. In each case
I
believe that Grant’s
analysis is incomplete and may be misleading.
A.
The
Empirical
Tests
a) Although recognizing some of the practical shortcomings
of
the data he
uses, Grant maintains that diversification is measured correctly by any
change in the number of industries in which a firm operates, but not
by
the
change in the ratio of non-primary output. It is surely more correct
to
argue
that each is only an imperfect measure
of
one aspect
of
diversification.
Diversification is not an unambiguous concept. Even
if
we can agree on
Amey’s definition in terms of the
‘
spreading
of
its operations by a business
over dissimilar economic activities’ (Amey, 1964,
p.
252),
and
can
accept
(resignedly) that economic definitions
of
‘
similar
’
activities are not synony-
mous with the definitions of Census industries, it is still true that a firm may
spread its activities in different ways. For example, a firm may diversify
by devoting
a
very small share
of
its resources
to
eaoh of several new
activities,
or
it may commit a much larger volume
of
resources to a single
new activity. It is not possible to say that either development necessarily
involves more diversification than the other. Given that diversification is a
multi-dimensional concept, no uni-dimensional measure can be superior in all
contexts.
b) Despite his comment that the behavioural theory
is
unnecessarily complex,
Grant uses a simpler model for the behavioural than for the neo-classical
theory in his empirical tests. His apparent reason for doing
so
is not that the
behavioural model excludes such variables as research expenditure, but
rather that it assumes that their impact will depend
upon
local conditions
85
To continue reading
Request your trial