Professional Documents
Culture Documents
Royal Economic Society, Wiley The Economic Journal
Royal Economic Society, Wiley The Economic Journal
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 support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
Royal Economic Society, Wiley are collaborating with JSTOR to digitize, preserve and
extend access to The Economic Journal
This content downloaded from 14.139.240.145 on Mon, 09 Apr 2018 08:39:50 UTC
All use subject to http://about.jstor.org/terms
The Economic Journal, go (September I980), 520-528
Printed in Great Britain
GROWTH OF FIRMS*
[ 520 1
This content downloaded from 14.139.240.145 on Mon, 09 Apr 2018 08:39:50 UTC
All use subject to http://about.jstor.org/terms
[SEPTEMBER i980] MANAGERIAL LIMITATION TO GROWTH 521
rates. Similarly current costs are not independent of the rate of installation of new
plant. One important conclusion from these analyses is that where adjustment
costs are not separable, the results from dynamic optimisation analyses do not
always produce optimal decision rules that are qualitatively similar to those
derived from comparative static analyses.
Now the 'costs of growth' are a particular type of adjustment cost, and I shall
argue that they are unlikely to be separable. When one reads in the literature of
firms that have grown too fast, one does not get the impression that the extra
costs of rapid growth are entirely an addition to overheads which do not affect
the efficiency of the operating divisions of the firm. The story is more that too-
rapid growth causes loss of coordination throughout the enterprise, with both
overhead and current costs being higher than they might otherwise have been.
To establish the case more firmly, a very simple model of how costs of growth
might arise will be constructed and analysed. Edith Penrose (1959) and many
other writers (e.g. Richardson (I964)) have all stressed that the main limitation
on a firm's growth is a managerial one. There are other problems of course, but
generally they are of a secondary nature: the firm's growth is ultimately restrained
by its inability to find, train and absorb new management of sufficient quality
faster than a given rate. The increasing costs of growth are due to the deterior-
ation of effective control by management in the face of too rapid change.
However, a significant problem arises when one attempts to incorporate this
phenomenon into a formal model of the growth of firms: in all the models of
firms usually employed by economists, there is no explicit recognition of manage-
ment having any role to play which bears upon the firm's performance. Thus we
have 'Hamlet' without the Prince - management is recognised to be the crucial
factor that limits the growth of firms, but our models do not normally include a
management variable.
In the absence of such a management variable, the 'Penrose Effect' has had to
be imposed on models from outside as it were, in various ad hoc ways, using its
impact either on total costs, or on profitability.' Although almost all writers
justify the inclusion of growth costs by reference to Mrs Penrose's work, it is my
belief that these ad hoc procedures have not in fact captured the essence of the
Penrose Effect, which relies on the special nature of management as a factor of
production. To overcome this problem, the model presented here does give
management an explicit role to play. The Penrose Effect will be seen to arise
naturally from within the model as the firm attempts to grow, and it will be
found that the resulting costs of growth will not be simply separable from the
current costs of the firm.
Assume that a firm produces an output (Q) using two inputs, Labour (L) and
Management (M). The role of management is to organise the labour force to
1 E.g. Baumol (I962), p. io8o, Williamson (I966), p. 5, Solow (I97I), p. 32I, use the cost formulatio
Marris (I964), pp. 249 ff., (I97I) p. i8, Heal and Silberston (I972) p. I38, use the profitability method
To my knowledge, only Lesourne (I972) uses a formulation similar to the one used here.
This content downloaded from 14.139.240.145 on Mon, 09 Apr 2018 08:39:50 UTC
All use subject to http://about.jstor.org/terms
522 THE ECONOMIC JOURNAL [SEPTEMBER
work in the most effective way. However, the crucial distinction between these
two factors of production is the way in which the firm may acquire them. Labour
is readily obtainable in the market at a constant wage w, and may be hired and
fired instantaneously and costlessly. Management, however, cannot be obtained
instantaneously and costlessly: raw management 'material' can be so obtained
at wage s, but this material must undergo a training and assimilation process
before it can be used effectively by the firm, and this process requires the
services of some of the existing management.
It follows that if the firm wants to expand it cannot use all of its existing
managerial services in the organisation of its current production activities; some
managerial services must be diverted to the training of new managers. This
diversion of management effort from production activities will mean that these
activities are less well controlled than they might otherwise be, and that therefore
the costs of current production are higher than they might be. It is these higher
current costs that are the opportunity costs of expansion to the firm, and consti-
tute Mrs Penrose's increasing costs of growth.
Let M1 be the managerial services engaged in current productive activity, and
M2 be the managerial services engaged in the training of new managers. Then the
firm's production function is assumed to be Q = Q(L, M1), with constant returns
to scale, and diminishing returns to each factor taken singly.
However, the creation of new members of the management team is not simply
a question of management training. Edith Penrose continually emphasises that
management must work as a team, and not as a collection of individuals skilled in
management. Teamwork can only be developed by giving individuals ex-
perience in working together. Training pure and simple is necessary, but not
sufficient; the same goes for getting to know your colleagues and for getting to
know the quirks of the business. All these are necessary, but if they were to be
acquired independently of each other, they would still not be enough. What
really counts is getting to know how to work-with your colleagues on the problems o
firm, and how they work with you, and this experience can only be gained by
working on actual problems.
The experience of working together is not just a requirement for the new
managers; the existing management must also find out how to work with the
newcomers, how far they can rely on them, how much they can delegate to their
subordinates, etc. Unless they do this their own efficiency will suffer. Again, the
only way in which this experience can be gained is by observing the newcomers
in action. If there are few possibilities for gaining this experience, the rate at which
the effective managerial team can be augmented will be severely limited.
The conclusion is therefore that existing management limits the rate of intake
of new managers in two ways. Firstly some of the existing management team
(A2) must be allocated to hiring and training the newcomers, and secondly the
rate at which the newcomers can gain relevant experience is limited by the
amount of productive work that the remainder of the existing management team
have in hand.' A good measure of this amount of productive managerial work
going on is simply M1.
1 This is most clearly set out in Penrose (1959), pp. 46, 47.
This content downloaded from 14.139.240.145 on Mon, 09 Apr 2018 08:39:50 UTC
All use subject to http://about.jstor.org/terms
1980] MANAGERIAL LIMITATION TO GROWTH 523
We shall assume that this function has constant returns to scale, with diminish-
ing returns to either factor taken on its own. If we also write
M2 = cM (o < a I) (2)
then
M~ J M1 M2\
M \ M ~7'MJ'
M f [(I -oc),oc],
M
M - g (a) (3)
That is, the rate of growth of the stock of effective managerial services is a
function of a alone. The assumptions made about f imply that g is a concave
function of ac, d2g/dc2 < o. It seems also reasonable to assume that g(o) = o, and
that dg/dac(o) > o. g may be an increasing function of ac throughout the range
O < a <- I, or it may have an internal maximum within that range. However,
no profit- or growth-maximising firm would ever consider operating where
dg/dac < o, so we can confine attention to the positively-sloped section.
PROFIT MAXIMISATION
Having formulated the managerial constraint on the growth of the firm, we can
now investigate the optimal decisions of a profit-maximising firm subject to such
a constraint. We take profit-maximisation to mean the maximisation of the
present value of the flow of distributed profits through time. For simplicity we
assume that the firm is able to expand its sales at a constant price. This is not
altogether an unreasonable assumption, and has been used by several previous
writers who justify it by the firm's ability to diversify into new markets instead
expanding down the demand curve of its initial product against the competition
of the other firms in the market. However, it is used here purely to simplify the
analysis - more realistic treatments of demand could be employed, but they
complicate matters without adding anything to the general conclusions to be
drawn here.
The profit-maximising firm's problem may now be written as:
Maximise
T (pQ-wL-sMf)e-rrt dt (4)
subject to
Q = Q (L, Ml)) (5)
M=Mg((a) O < a < I, (6)
M(o) = M, (7)
L,M>o, (8)
This content downloaded from 14.139.240.145 on Mon, 09 Apr 2018 08:39:50 UTC
All use subject to http://about.jstor.org/terms
524 THE ECONOMIC JOURNAL [SEPTEMBER
where
Q = output at time t,
L = labour employed at time t,
M = managerial services available to the firm at time t,
Ml = (I - aC) M = managerial services employed in production at time t,
M2= cM = managerial services employed in training at time t,
M = initial stock of managerial services available at t = o,
p = price of output,
w = wage of labour,
s = wage of management,
r = rate of interest.
Set up in this way, the problem is very suitable for analysis using optimal
control methods. However, in this simple model only steady-state growth paths
are optimal,' so that for simplicity of exposition here we can derive the opti-
mality properties from the much simpler analysis of steady-state paths. For
steady-state growth the firm will choose an initial value LO for labour input, and
select a value for ac which will be held constant for all time. This defines both
M1f and the steady-state growth rate g, and henceforward Q, L, M1, M2 all gro
exponentially at that rate.
The present value of the profits accruing along this path is:
00
This content downloaded from 14.139.240.145 on Mon, 09 Apr 2018 08:39:50 UTC
All use subject to http://about.jstor.org/terms
1980] MANAGERIAL LIMITATION TO GROWTH 525
[(I-OC) Q sQ dg
This flow would grow at rate g through time and be discounted at rate r. If
we equate its present value to the opportunity cost of the extra managers,
GROWTH MAXIMISATION
This content downloaded from 14.139.240.145 on Mon, 09 Apr 2018 08:39:50 UTC
All use subject to http://about.jstor.org/terms
526 THE ECONOMIC JOURNAL [SEPTEMBER
- a..~~-aQ -d
@grdg
dg [-(r-g)pM
+ A,+ (pQ?_-wL
L i d- sM)-](0
0 (20)
aL (r-(g) a )
Assuming that the profitability constraint is binding, the labour productivity
condition is the same as in the profit-maximising case. The management con-
dition may be derived as before from (20), using Euler's Theorem and (2I):
(r - g)2dg P AQ dg] dg
AM = p (g- r) +(i -ad)g - dci
therefore
[ (r-g)2] dg
CONCLUSIONS
This content downloaded from 14.139.240.145 on Mon, 09 Apr 2018 08:39:50 UTC
All use subject to http://about.jstor.org/terms
I980] MANAGERIAL LIMITATION TO GROWTH 527
of his current operations (as evidenced by the efficiency condition for labour
usage), if he is in fact growing faster than the equivalent profit-maximiser he will
be trying to maximise his short-run profit with reference to a higher marginal
cost curve, and therefore will choose a lower output than the profit-maximiser. In
the Williamson model, where marginal cost is unaffected by the growth rate,
both firms choose the same output.
The model presented here has been kept as simple as possible for purposes of
clarity, and there are several obvious extensions that might be considered de-
sirable: the incorporation of fixed capital and a less restrictive treatment of the
demand side are two that spring quickly to mind. However, these do present
quite serious problems of analysis without affecting in principle the point
at which this paper has been directed: that the correct formulation of the
managerial limit to growth will in general imply a total cost function which is
nonseparable in Q and g.
Nor is the general result entirely dependent on the particular assumptions
employed here. If we combine equations (5) and (6) to eliminate ax, we may
rewrite the production function as
Q = 0(L,M,M)
and in this form the similarity with the original Treadway (I970) analysis is
obvious. The results derived here therefore hold true for any production-cum-
growth process for which 92Q/9L aM < o, whereas in the Williamson analysis,
the implicit assumption is that 92Q/9L AM = o.
The advantage of the particular formulation used here is that it points directly
to the opportunity costs of growth: a fast-growing firm uses its managerial resources
(and in the real world other resources as well) in coping with the problems of
growth itself, when they might otherwise be employed in solving problems and
reducing costs within the current operations of the firm. In the light of the grow-
ing literature on the unprofitability of mergers, and the evident control problems
of such merger-created firms as British Leyland, this is perhaps an important
practical point that deserves more attention. On the theoretical side, the im-
portant implication is that price and output decisions cannot be dissociated
from the growth policy of the firm.
REFERENCES
Baumol, W. J. (I962). 'On the theory of expansion of the firm', American Economic Review, vol. 52,
pp. I078-87.
Brechling F. (I975). Investment and Employment Decisions. Manchester; Manchester University Press.
Eisner, R. and Strotz, R. (i963). 'Determinants of business investment'. Research Study No. 2 in
Impacts of Monetary Policy, prepared for the Commission on Money and Credit. Englewood Cliffs,
N.J.; Prentice-Hall.
Gould, J. P. (I968). 'Adjustment costs in the theory of investment of the firm'. Review of Economic
Studies, XXXV, no. I, pp. 47-55.
This content downloaded from 14.139.240.145 on Mon, 09 Apr 2018 08:39:50 UTC
All use subject to http://about.jstor.org/terms
528 THE ECONOMIC JOURNAL [SEPTEMBER 1980]
This content downloaded from 14.139.240.145 on Mon, 09 Apr 2018 08:39:50 UTC
All use subject to http://about.jstor.org/terms