American Economic Association

Monopoly and Competition: A Classification of Market Positions

Author(s): Fritz Machlup
Source: The American Economic Review, Vol. 27, No. 3 (Sep., 1937), pp. 445-451
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/1802042
Accessed: 29-11-2015 16:06 UTC

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 support@jstor.org.

American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The American Economic
Review.

http://www.jstor.org

This content downloaded from 137.189.170.231 on Sun, 29 Nov 2015 16:06:23 UTC
All use subject to JSTOR Terms and Conditions
 The
American Economic Review
Vol. XXVII SEPTEMBER, 1937 No. 3

MONOPOLY AND COMPETITION: A CLASSIFICATION

OF MARKET POSITIONS
The concepts of monopoly, imperfect competition, monopolistic competition, oligopoly,
etc., as used in recent writings are in need of re-definition. The classification of market
positions suggested here distinguishes, first of all, "sellers conscious of their rivals' reactions"
from "sellers heedless of rivals' reactions." The first group is divided into duopoly and
oligopoly with differentiated and with standardized products; the second group is divided
into monopoly, monopolistic competition and pure competition. The terms "perfect" and
"imperfect" are used to denote immobility of factors, degree of friction, speed of adaptation;
they refer, therefore, to potential changes in demand for the product of individual sellers
due to a flow of factors into or out of an industry. Competition may be pure and perfect,
pure and imperfect, monopolistic and perfect, monopolistic and imperfect; the first in each
of these pairs of adjectives refers to the shape of the demand curve, the second to its
position in relation to the cost curves. Subtle distinctions such as those between "potential
competition of the imperfect monopolist," "monopolistic imperfect competition," and
"oligopoly with differentiated products" can be made more easily on the basis of the
classification suggested here.

More discussion of the terminological and conceptual disorder is needed

before the theory of the blendings of monopoly and competition can
progress further after its great advance in the last decade. Attempts at
clarification, at clearer definitions, and at a more uniform use of words have
been undertaken recently.' A classification of the various forms of competi-
tion or the various positions on the selling market will be attempted here.
As in any classification, the lines drawn between the different types are
artificial, but such dividing lines are necessary for clear reasoning.
As criteria of classification we may choose the types of considerations
and reflections in the mind of the seller when he pictures the quantities he
could sell at various prices. The main division is that between sellers whose
calculations or deliberations include expected reactions of rival sellers and,
on the other side, sellers who do not engage in such complicated reflections.
The latters' thought is concerned with the buyers' reactions only. No account
is taken of what rival sellers may do in consequence of his action. For want
of a shorter expression we call this type of seller the seller heedless of
rivals' reactions, and contrast it with the seller conscious of rivals' reactions.
The latter anticipates not only the buyers' reactions on his higher or lower
1 Horace G. White, "A Review of Monopolistic and Imperfect Competition Theories,"
Am. Econ. Rev. vol. xxvi, 1936, pp. 636-649; Paul M. Sweezy, "On the Definition of
Monopoly," Quart. Jour. of Econ., vol. 51, 1936, pp. 362-363.

This content downloaded from 137.189.170.231 on Sun, 29 Nov 2015 16:06:23 UTC
All use subject to JSTOR Terms and Conditions
 446 Fritz Machlup [September

selling prices, but also, and sometimes primarily, his rivals' reactions. When
he raises the selling price, when he lowers it, will the rivals (or the rival)
follow suit or not; if so, fully or partly, immediately or with some delay?
The customers' reactions and thus the effects of a price change, cannot be
estimated or imagined by the seller if he fails to take into account the most
probable of the rivals' reactions.
If the seller expects the reactions of his rivals to take place without any
delay, then the estimate of these reactions often precedes the estimate of
buyers' reactions. If he expects some delay in his rivals' reactions, then
he may first make an estimate of buyers' reactions for the period between
his price change and his rivals' reactions (i.e., a short-period demand curve);
second, an estimate of rivals' reactions; and finally, one of buyers' reac-
tions subsequent to his rivals' adjustments (i.e., the long-period demand
curve). How many customers will he be able to take from his rivals; how
many will he lose to them? This question becomes of foremost importance
to the rival-conscious seller. And, the demand curve, i.e., the schedule of
quantities which the seller believes2 to be salable at various prices, acquires
a still greater degree of vagueness, because each point may refer to a differ-
ent set of anticipated reactions of rivals.
It will be clear from what we have said about the rival-conscious seller
that duopoly and oligopoly are the cases in point. It may be wortlhwhile to
distinguish between oligopolists with standardized products and oligopolists
with differentiated products. The differentiation of the product in this case
where sellers are few does not, however, change essentially the type of sellers'
reflections; it will merely diminish the ease with which in their belief
customers can be detached from their rivals or lost to their rivals. In other
words, the differentiation of the product will make for a smaller elasticity
of demand in the minds of the individual sellers; but these elasticities are
in any case smaller than infinity because of the fewness of sellers.
With few sellers in the market the sellers must be rival-conscious, that is
to say, they have to make guesses about their rivals' reactions together with
guesses about the buyers' reactions. This is not so in the case where there
is only one seller, or in the case where there are very many sellers. Both
the monopolist and the competitor in a market of very many sellers are
unconcerned with rivals' reactions; the one because he has not any, the
other because he has too many.
The monopolist's product has of course (distant) substitutes and is, there-
fore, in rivalry with many other products. The monopolist "competes" with
all sellers of all different products for the consumer's dollar. This, however,
is too wide a concept of competition for the purposes of our classification.
The competition of commodities outside the industry-beyond the "gap
2 Consistent methodological individualism calls for an interpretation of demand curves

as pictures of typical sellers' estimates. What the "actual facts" do is to influence and
reshape the estimates of the sellers who learn by experience.

This content downloaded from 137.189.170.231 on Sun, 29 Nov 2015 16:06:23 UTC
All use subject to JSTOR Terms and Conditions
1937] Monopoly and Competition: A Classification 447

in the chain of substitutes- "3 is so anonymous and so nebulous that expec-

tations as to the policy of their sellers are excluded. The perfect monopolist
has no rival whose reactions to his own price policy concern him.
One seller in a market of very many sellers has, of course, many direct
competitors, but he has so many that he does not expect they would care
about what he does. He is conscious of his small share in the market and he
knows or thinks that none of his competitors would feel any tangible effect
of his actions. Thus he will not be expectant of possible reactions of rivals.
The distinction between the competition of homogeneous or standardized
products offered by many sellers, and the competition of differentiated
products of many sellers is important. In the first case, that of pure competi-
tion, every single seller knows or thinks that he could not sell a single
unit of his product should he ask a price higher than the market price, and
that he could sell any quantity he cares to sell without charging less than the
market price. In other words, the demand conceived by any such single seller
is infinitely elastic. In the second case, that of monopolistic competition,
each seller knows or thinks that he could keep some of his customers even
if he raised his price, and that he could sell some more, but still very limited
quantities of his product, if he lowered his price. In other words, the
demand conceived by the monopolistic competitor is less than infinitely
elastic because of the attachment of customers to the slightly differentiated
product or service they get from the individual sellers.
I put the seller under monopolistic competition (i.e., the seller in a mar-
ket of many sellers offering differentiated products) and the seller under
oligopoly with differentiated products (i.e., the seller in a market of few
sellers offering differentiated products) in different groups, and venture to
hold that the case of the latter is much more adequately treated as an
oligopoly case and that little is gained by treating it as a case of monopolistic
competition. What the "small" or "great" number of sellers in the market
mainly does is to affect the policy of the individual seller. "Sellers are few"
means that the single seller is afraid his actions will be felt by, and cause
reactions of, his rivals. "Sellers are many" means that the single seller is
not conscious of any reactions of rivals. And it will be agreed that the
difference in the sellers' reflections in the two cases is essential.
These are the types which we have distinguished so far:
A. Sellersconsciousof their rivals'reactions
(1) Duopoly
(2) Oligopoly with differentiatedproducts
(3) Oligopoly with standardizedproducts
B. Sellers heedless of rivals' reactions
(1) Monopoly
(2) Competitionwith differentiatedproducts (monopolisticcompetition)
(3) Competitionwith standardizedproducts (pure competition)
'Joan Robinson, "What Is Perfect Competition?" Quart. Jour. of Econ. vol. 49, 1935,
pp. 104-120.

This content downloaded from 137.189.170.231 on Sun, 29 Nov 2015 16:06:23 UTC
All use subject to JSTOR Terms and Conditions
448 Fritz Machlup [September

The seller under monopolistic competition has sometimes been regarded

as a monopolist in his particular product; since neither the monopolist nor
the monopolistic competitor care for rivals' reactions, it may seem hard to
see any distinguishing characteristic between the two. One might be satis-
fied with a difference in the degree of substitutability of products and hence
of elasticity of demand. Monopolistic competition would then comprise the
cases of closer substitutes and more elastic demand curves, while monopoly
would comprise those of remote substitutes and steeper demand curves. Or
one might make a mere physical comparison of products and call the
monopolistic sellers of physically very similar products monopolistic com-
petitors.
Another aspect of these market positions, however, yields distinguishing
characteristicsof much greater significance. It is the chance of supernormal
profits and the ease of entry into the industry which have to be examined.
And such examination will also suggest the proper place in our classification
for the concepts of imperfect competition and imperfect monopoly.
The distinction between pure competition and perfect competition, sug-
gested by Professor Chamberlin4 was felicitous. It can be carried through
further than he himself may have done. The two pairs of antonyms can be
shown in four different combinations; competition may be pure and perfect,
pure and imperfect, monopolistic and perfect, monopolistic and imperfect.
Pure competition, i.e., the competition of very many sellers of a homo-
geneous product, implies infinitely elastic demand curves for each individual
seller. Perfect competition means uniform prices for homogeneous goods
and uniform earnings for equal services, hence the absence of supernormal
or subnormal earnings. In geometrical language pure competition means a
demand curve parallel to the x-axis, that is, a demand curve coinciding with
the corresponding marginal revenue curve. (Average revenue equal to mar-
ginal revenue.) While pure competition refers, thus, to the shape of the
demand curve, perfect competition refers to its position in relation to the cost
curves. If under perfect competition abnormal returns are excluded, selling
price and average cost must be equal for the output chosen (i.e., for the
output for which marginal revenue and marginal cost are equal) and this
implies that the demand curve must be tangent to the average cost curve. It
can be easily seen that the demand curve may be parallel to the x-axis with-
out being tangent to the average cost curve, or that it may be tangent to the
average cost curve without being parallel to the axis, or that it may be both
at the same time, or that it may be neither.
If competition is pure but imperfect,5 the demand curve to the individual
'Theory of Monopolistic Competition, pp. 6-7 and 25-26. The suggested "distinction
of terms not does seem to have impressed recent writers on the subject," as Mr. White
(op. cit. p. 641) states with regret. Is it hoping against hope that writers of the Cambridge
school may occasionally adopt foreign-coined terms?
' Our use of the concept "imperfect" competition must not be confused with Mrs. Robin-
son's. See op. cit. and Economics of Imperfect Competition.

This content downloaded from 137.189.170.231 on Sun, 29 Nov 2015 16:06:23 UTC
All use subject to JSTOR Terms and Conditions
1937] Monopoly and Competition. A Classification 449

seller will, due to the great number of sellers and to the homogeneity of the
product, be infinitely elastic, but not tangent to the respective average cost
curve, thus leaving profits under or above normal. This will be due to fric-
tions, which stand in the way of quick transference of productive factors
(including entrepreneurial capacity) between different industries, and to
the costs of transference, which prevent such adjustment until price differ-
entials or profit margins are large enough to pay for the switching of factors
between localities and occupations.6
If competition is perfect but monopolistic, the demand curve conceived
by the individual seller will be less than infinitely elastic because of the
differentiated service which he offers, but it will be tangent to the average
cost curve. If the demand for products of this and similar kinds rises, and
the demand curves conceived by the individual sellers have a tendency to be
higher than tangent to the respective average cost curves, the prospect of
supernormal returns in this industry would instantly bring about an influx
of new firms7 and thus push the demand curves of the single firms back to
tangency to the respective average cost curves. In other words, the free and
easy entry of new firms into the industry (which turns out products not
homogenenous but similar) does not allow abnormal returns to be reaped.
With any prospect of such returns new firms offering a close substitute
would be attracted and the demand for the products of the existing firms
thus pressed down. This pressure stops only when returns are reduced to
normal so that no more firms are attracted. Geometrically expressed, the
demand curve is tangent to the average cost curve.8
If competition is monopolistic and imperfect, so that the influx of new
firms into the industry (or the withdrawal of old firms) is delayed, abnormal
returns can appear.9 Whether or not the profits of these sellers under
8
Space does not permit discussion here of the notion of "normal profits." Mrs. Robinson
(op. cit., p. 106-107) objects to linking "normal profits" with the concept of competition,
chiefly because "entrepreneurship" is not homogeneous. If an industry is licensed by the
state, are entrepreneurs with license and those without license heterogeneous factors? Should
physical, psychical and legal inabilities, natural and institutional obstacles to perfect sub-
stitution, all be treated alike? It is the economist's view as to control and flexibility of
institutions that is back of his thought that certain inabilities give rise to a "normal" differ-
ence, others to an "abnormal" difference in earnings. Yet, there is a meaning in the application
of these terms and I cannot share Mrs. Robinson's verdict that they cease to be of use.
7 Perfection of competition does not depend on perfect fluidity of all units of all factors.

The fluidity of a certain portion of the total amount of each factor would suffice to produce
the assumed result, i.e., equal earnings of equal factors.
s The nature of the average cost curve can be conceived in either of two ways. If
"differential rents" are included in long-run average cost, then the above statement is
adequate. If, on the other hand, "differential rents" are excluded from cost, the statement
is true only of the "marginal firm."
9 When the product differentiation is considerable, monopolistic competition will
most likely be imperfect as well. This is because the uncertainty involved in the establish-
ment of new firms will be much greater than if products are less differentiated or even
homogeneous. Firms with established good-will or well known trade-marks are often given
as illustrations of monopolistic and imperfect competition. Most of these cases, however,
should more appropriately be treated as oligopolies with differentiated products.

This content downloaded from 137.189.170.231 on Sun, 29 Nov 2015 16:06:23 UTC
All use subject to JSTOR Terms and Conditions
450 Fritz Machlup [September

monopolistic imperfect competition should be considered monopoly profits

is not an urgent question. That they appear at all is due to the imperfection
of competition, but they are possibly higher than if competition were pure,
because of the sloping demand curve and the output restriction of each
monopolistic competitor who keeps the selling price above marginal cost.
The absence of supernormal returns is a distinctive feature of monopolis-
tic perfect competition10 as against perfect monopoly. This distinguishing
characteristic, however, fails if monopolistic competition is imperfect. The
delayed emergence of new firms allows the existing monopolistic competitors
in this industry to reap supernormal ("monopoly") profits. The only cri-
terion that then distinguishes monopolistic imperfect competition from
monopoly is the time element. Competition is imperfect if it works slowly,
being hampered and delayed because of frictions and costs of moving. But if
enough time is allowed, the results approach those of perfect competition.
Monopoly, on the other hand, is not based on mere frictions, time-lags, and
costs of moving. It is based on more reliable and more enduring conditions.
From a short-run point of view, monopoly and monopolistic imperfect
competition are the same. It is from a long-run point of view that they are
different. In the long run the "imperfections of adaptation" will be over-
come, while the conditions constituting a monopoly position may survive.
Monopoly may be perfect or imperfect. It is imperfect if the factors
making for the monopoly position are such that the seller can depend upon
their endurance only within limits.' There may be serious limits for his
price policy, limits set by the seller's fear of government interference, and
limits set by the seller's fear of "potential competition." This last possibility
interests us most in our classification because it bears on the distinction
between the rival-conscious seller and the non-rival-conscious seller. The
perfect monopolist has no rivals who are of concern to him, but the imper-
fect monopolist may have "potential rivals." The latter may be afraid that
above certain prices his monopoly position could be terminated through
the appearance of rivals offering the same product or close substitutes. Yet
this type of rival-consciousness is different from the one observed with the
duopolist or oligopolist. The imperfect monopolist has a range of prices
where he feels himself entirely free from possible reactions of rivals. He
shies merely from going above a limit where, even though only in the long
run, rivals may come into existence.
The potential competition of the imperfect monopolist finds its geometri-
cal expression in the shape of the long-run demand curve conceived by the
seller: above a certain price range the demand curve would rapidly flatten
out toward the left. That is to say, at certain high prices the monopolist
would expect rivals to appear on the market so that from then on a slight
price increase would cause a relatively great loss of sales. It should be noted
"0'"Perfect monopolistic competition" would sound better but it might be read as
"perfectly monopolistic" while "perfect" ought to refer to competition.
" One may question whether cases of perfect monopoly exist in reality.

This content downloaded from 137.189.170.231 on Sun, 29 Nov 2015 16:06:23 UTC
All use subject to JSTOR Terms and Conditions
1937] Monopoly and Competition: A Classification 451

how differently the "potential competition of the imperfect monopolist" and

"monopolistic imperfect competition" are to be represented in geometry
and as types of conduct. In the former case there is the sudden turn to high
elasticity in the upper part of the long-period demand curve; in the latter
case there is a sloping demand curve which as a whole will be reshaped and
shifted more and more to the left and downward as time goes on and the
imperfections of the market are gradually overcome. In the former case the
shape of the long-period curve is conceived by the seller who avoids the
high price ranges because of the potential competition. In the latter case the
movement of the demand curve in the long run is a succession of revised
demand anticipations of the seller; it need not be subject to any specific
measures in his price policy; it is simply the realization of pressure on the
market due to the gradual emergence of suppliers of close substitutes.'2
This pressure, realized only gradually, may not be foreseen by the seller, but
if it is foreseen, then it may lead him to a policy contrary to that of the
monopolist. He may, for instance, pursue a consciously short-run policy
because he knows that after some time the market will be spoiled by the
unfailing appearance of more sellers.
The price policy of the seller under monopolistic imperfect competition
who is aware how short-lived is his favorable market position borders again
rather closely on the policy of sellers conscious of their rivals' reactions.
There is indeed no great difference between the type of policy which an
oligopolistic seller of differentiated products may pursue when he believes
that his rivals will take only delayed reactions on his actions, and the case
of monopolistic imperfect competition which we have just described. But the
classification here offered does not attempt to create water-tight compart-
ments in which to force the specimens of all cases observed in reality.
We have seen that a subdivision of the groups and sub-groups of our
classification seems appropriate, distinguishing perfect from imperfect
monopoly,. perfect from imperfect monopolistic competition, perfect from
imperfect pure competition. But we have still to bear in mind that the kinds
of imperfections are so manifold and different that it is hardly permissible
to label the various market positions satisfactorily with the few tags now
currently in use.
FRITZ MACHLUP
University of Buffalo
12 Paul M. Sweezy,
op. cit., p. 362, includes in his monopoly definition the condition
that the demand curve for the product is independent of the profit made. This formulation
allows of at least two interpretations. It may mean, first, that the monopolist is not afraid
that the profits he makes will attract others to supply substitutes, and thus the long-run
demand curve conceived by the monopolist will not show in its shape any expected loss
of sales to other suppliers. Second, it may mean that the profits of the monopolist will
''actually" not attract others to supply substitutes so that the "actual" demand curve for
his product will not be shifted to the left and downward. This difference between events
expected by the seller, and events "actually" in store for the seller is, I hope, clearly brought
out in the text.

This content downloaded from 137.189.170.231 on Sun, 29 Nov 2015 16:06:23 UTC
All use subject to JSTOR Terms and Conditions