Front Page Titles (by Subject) Part II, Chapter VI: Minor Prerequisites for Perfect Competition - Risk, Uncertainty, and Profit
The Online Library of Liberty
A project of Liberty Fund, Inc.
Search this Title:
Also in the Library:
Part II, Chapter VI: Minor Prerequisites for Perfect Competition - Frank H. Knight, Risk, Uncertainty, and Profit 
Risk, Uncertainty, and Profit (Boston MA: Hart, Schaffner and Marx; Houghton Mifflin, 1921).
About Liberty Fund:
The text is in the public domain.
Fair use statement:
Part II, Chapter VI
Minor Prerequisites for Perfect Competition
In Part Two we have attempted an analytical construction of a perfectly competitive society, with a view to determining the precise meaning of the theoretical tendencies of a private property, free exchange organization of society, and especially the conditions necessary to the realization of those tendencies. The abstract conditions first enumerated in chapter III represented in part divergencies in degree only from real life, and were in part arbitrary abstractions from fundamental characteristics of the pecuniary organization made for the purpose of a separate study of the constituent elements. Those of the latter type have been dealt with in chapters IV and V, and the result, up to the present point, is an outline picture of the essentials of a perfect competitive system.105 The first, rather preliminary, objective of the study has thus been achieved, as far as the author is prepared or feels it advisable to go. The second and more fundamental purpose is to contrast this ideal, perfect competition with the facts of ordinary life, to examine the limitations of the general principles developed, and to inquire as to the directions in which they must be supplemented by detailed, empirical data before completely applicable conclusions can be drawn.
But it is not the intention to cover this field with any great degree of exhaustiveness. Only one of the theoretical simplifications is to be studied in detail, the assumption of perfect knowledge. Part Three of the essay will be devoted to a discussion of the meaning and consequences of uncertainty, the incompleteness and inaccuracy of the beliefs and opinions upon which economic conduct is based. But it is desirable to have as a background some brief notice of the other abstracted factors.106
It will readily be seen that many of the objections to the pure theory of distribution commented upon in chapter IV relate to these necessary scientific idealizations, and have real significance as limitations on the completeness and accuracy of the generalizations of theory. They are not, therefore, valid objections to the theory and have been advanced as such only because of the common failure to comprehend the nature of scientific reasoning, the meaning and use of general principles. This is especially applicable to the first point to be noticed, the assumption of continuous variability in the magnitude of all factors dealt with. The question of the size of the "marginal unit" is clearly relative to that of the flexibility of industrial organization, and the two must be considered together. When we give up the illicit procedure of funding productive agents into "factors" and deal with the actual competing units on their own account, this problem becomes of practical significance and constitutes an effective limitation on the application of the theory. In the case of labor especially, with which we are here particularly concerned, the human individual is a very effective unit; not only does he bargain as a unit, but he cannot practically be divided up between different establishments, and the range of occupations in which he can engage in any short interval of time is also very narrowly restricted. He may also be in a high and surprising degree unique; he does not always shade off by imperceptible gradations from one variety to another to the extent that perfect competitive imputation demands. His numbers (in proportion to the number of variants) are not nearly always so large as to make an individual a negligible fraction of a group of similars.107
As a consequence of the appreciable dimensions of the natural agent, the flexibility of the economic organization as a whole is restricted, and the criticism made by Mr. J. A. Hobson and Professor Wieser against the productivity theory is true to a considerable extent in many individual cases. There are many productive organizations consisting of small numbers of rather unique agents which very effectively supplement each other and are not so effectively demanded elsewhere. In such a case competition does not afford means of distributing the entire yield of the group among its members; an appreciable part of it resists automatic division and remains a joint product, dependent on the peculiar effectiveness of the particular organization. Many partnerships illustrate this point. Imputation goes as far as the group, giving that its proper income, but fails to distribute accurately within it. In case of a partnership this division between the members is usually made on ethical grounds or on the basis of "bargaining power," sheer personal force. In industry at large the special product of the organization above that competitively assigned to its components is likely to go, largely at least, to the entrepreneur, though bargaining power or the strategic situation always plays a large part in the proceedings.
The same factors give rise to a peculiar difficulty in dealing with the law of diminishing returns. When any agent is by its physical nature or any particular circumstances available only in relatively large blocks, so that only a few, perhaps only one, is used in a single competitive organization, the technological features of particular combinations may cause apparent exceptions to the "law" at some points; these may be apparent for certain sections of the curve for the simple reason that one element is not subject to decrease and the best proportions can be secured only by increasing the other elements. A conspicuous example is the case of railways, the principal crucial "agent" being the right of way. If the demand for transportation were large enough to require an indefinite number of tracks the curve would be smoothed out and would ultimately show increasing costs from the other elements in the equipment. So with gas or water mains, until a certain size has been reached, and many similar cases. The fact of limited divisibility is responsible for all differences in the economy of operation of establishments of different sizes. The amounts of certain agencies or elements in the operations not being continuously variable, other things have to be proportioned to them to get the best ratio, thus imposing restrictions on the size of the plant as a whole. Many, if not most, of these questions of size ultimately come back to the human being as a relatively indivisible unit.
Preliminary to a discussion of predatory activity, or acquisition which is not production, we must again refer to the question of the ethical implications of the productivity analysis. The purely causal meaning of productivity in a scientific explanation of economic phenomena is apt to be confused with social or moral issues which belong in an entirely different sphere. We have insisted that the word "produce" in the sense of the specific productivity theory of distribution, is used in precisely the same way as the word "cause" in scientific discourse in general. But the word "cause" itself is vague in ordinary speech, and it is natural that confusion should arise in regard to the economic synonym. For example, the socialists, with no lack of suggestion and justification from the loose usage of words by economists of non-socialistic schools, have insisted that all wealth is "produced" by labor. We need do no more than mention the names of Smith and Ricardo in this connection, while among contemporary writers Professor Taussig exemplifies the same practice, expressly stating that labor produces all wealth, but may not be entitled to all.108 We should say that the reverse is more correct, that labor does not "produce" all wealth, but may be entitled to all, on ideal grounds.
Inasmuch as any assertion of a cause and effect relation between particular events is always (as already pointed out) made on the ground of some special human interest or "bias," there is much justification for such usage, but this only makes the more imperative, a clear separation from the "scientific," use of causal terminology. Thus it is quite proper to say, in ordinary speech, that the cook "prepares" the meal, that the opening of the throttle of the locomotive by the engineer is the "cause" of the starting of the train, and that his failure to see the signal is the "cause" of the wreck and the deaths of the passengers. In an analogous way a small group of agents might for some purposes be credited with nearly the whole output of a large establishment; "other things equal," the product depends on their coöperation.
But it must be evident that scientific economics cannot use the word "produce" in this sense. The product of any productive service can for scientific purposes be only what we have defined it to be, that which is really dependent upon the service in question, that which can be produced by its aid and which cannot be produced without it, in the social situation as it is, allowing for the change in organization which would accompany its withdrawal from use. It follows that we cannot properly speak of the "product" of an economic "factor," even if we use the word "factor" in the possibly legitimate sense of a group of physically interchangeable things. The product of "labor," "land," or "capital," as aggregates, involves a still more illicit and meaningless use of terms. The only specific product which can be recognized is that of a single agent as such, an individual human being or machine, or such a parcel of land (or of liquid capital) as is actually bargained for and used in the production process (and for perfect competition to take place it must be negligible in size).
More important, however, is the error of attributing any sort of moral significance to economic productivity. It is a physical, mechanical attribute, attaching to inanimate objects quite as properly as to persons, and to non-moral or even immoral as well as virtuous activities of the latter. The confusion of causality with desert is an inexcusable blunder for which the bourgeois psychology of modern society is perhaps ultimately to blame, though productivity theorists are not guiltless.109 We must guard against thinking of the "natural" adjustment of the competitive system as having any moral import, though it is of course "ideal" in the scientific sense of being a condition of stability. To call it the "best possible" arrangement is merely to beg the question or to misuse words. The natural arrangement is only that under which, with the given conditions as to the demand and supply of goods, especially the existing distribution of productive power, no one is under any inducement to make any change. If we pass over the question of how far individual wants for specific things really dominate conduct, and neglect equally the whole category of wants for certain social relationships and interests in other individuals (not absolutely dependent), and assume in addition (we shall investigate the point presently) that no interests are involved in any exchange except those of the direct parties to it—then the result is a mere mechanical equilibrium of the pull and haul of interacting individual self-interests.
It is imperative that we bear in mind that the serpent's tail is always in the serpent's mouth, that what the competitive system tends to give back is just what is put into it in the way of human motives and human powers, natural, acquired, or conferred, and has in itself no moral attribute whatever. In real life the possession of property (or superior training) is supposed to represent saving or invention or some contribution to social progress. But it is clear that there is no technical (much less moral) equivalence between these services and the right to their entire fruits in perpetuity, and to confer it on one's heirs and assigns forever—particularly when we consider the enormous element of pure luck in all operations of this sort. The only sense and the only degree in which rewards for service are ethical is that of the necessity of paying the reward in order to get the service performed. From this point of view the only defense of most of the existing system is the difficulty of suggesting a workable alternative.
We must now turn again briefly to the point mentioned above, the extent to which outside interests not represented in agreements between individuals are affected by them (otherwise than through direct competition in the market). The mere mechanical effectiveness of competitive free contract in producing a reconciliation of individual interests under given conditions depends largely on the answer to this question. Obviously, outsiders may be affected either advantageously or disadvantageously. In the former case voluntary agreements will not be carried far enough to secure maximum social (total individual) advantage, while in the latter case they will be carried too far. These facts form the most important source of the need for social interference. Many services, such as communication and education, not to mention the administration of justice, confer a general benefit on the community in addition to the special benefit to the individual, and must be encouraged by bounties or actually taken over and performed by public agencies or they will not be developed to the point of maximum benefit. The most familiar illustrations of the opposite case in our society relate to the use of land for purposes which damage the neighborhood, or are thought to do so. It is perhaps of nearly equal importance that improvements on land and industrial developments generally may benefit neighboring property, and might be made much more readily and in ways involving less injustice if there were some practicable way of assessing these benefits. This is notably true of public and quasi-public works, which effect enormous uncompensated transfers of values. It may be doubted whether in fact any agreement between individuals is ever made which does not affect for good or ill many persons other than the immediate parties, and a large proportion have wide ramifications over "society."
In this brief sketch we can only mention and insist on the fundamental importance of the fact that a large part of what men want relates directly to other members of society. Man is, after all, zoön politikon and quite on a par with his personal needs are all sorts of interests in furthering the plans of people whom he likes and, always relatively and generally absolutely, obstructing those of others, in a wide scale of gradations down to Thackeray's " 'e's a furriner; 'eave a 'arf a brick at 'im!" or, "kill the nigger!" The relative importance of other-regarding motives and desires, directed not to material things, but to forms of social relationships, is sure to be underestimated by any one treating economic phenomena in a "scientific" way.
The extreme phase of the problem of the moral character of the economic system relates to positively predatory activity. Davenport, following Veblen, has stressed the contrast between (private) acquisition and (social) production, making much of the hiring of sluggers, assassins, and incendiaries as part of the demand for labor, the productivity of burglars and their implements, and the like. It is not really very difficult in most cases for one who is disposed to do so to distinguish between theft or brigandage and free contract, and perhaps all that is needful to say of them in treating the theory of contractual organization is that they are obviously outside of it. A large part of the critics' strictures on the existing system come down to protests against the individual wanting what he wants instead of what is good for him, of which the critic is to be the judge; and the critic does not feel himself called upon even to outline any standards other than his own preferences upon a basis of which judgment is to be passed. It would be well for the progress of science if we had less of this sort of thing and more serious effort to formulate standards and to determine the conditions under which free contract does or does not promote individual interests harmoniously and realize social ideals. In addition it is most desirable that some attempt be made to separate the evils for which the form of organization is more or less reasonably blamable from those which are inherent in nature and human nature, or in organization as such, irrespective of its form, and to keep the question in view, in criticizing the exchange system, of whether any other conceivable system would offer any possible chance for change or improvement.110
There is a close connection between the moral aspect of the economic order and the problem of monopoly. This subject is of especial importance in the theory of profit, since profit has often been ascribed wholly or in part to monopoly gain, as already noticed in the case of Macvane and the Clark School. "Monopoly" is a word used to cover things which for present purposes must be kept distinct, and its meaning must first be made clear. Monopoly is usually defined as the control of the supply of a commodity. A common but disastrous error is the confusion of control with natural limitation of supply. We need not pause longer than to characterize as a serious misuse of words the denomination of land rent, for example, as a monopoly income. Even J. S. Mill fell into the error of defining monopoly as limitation, and it is exemplified in its extreme form by Mr. F. B. Hawley, who virtually calls all income due to the "scarcity" of any productive resource a monopoly return. Now, as all income, from the distributive standpoint, is dependent on the scarcity of the agents which produce it, and all in exactly the same way, the meaninglessness of such a description is apparent. And of course the same applies to "scarcity income" in general, whether called monopoly gain or not. There is under free competition no other sort of income, qualitatively or quantitatively, and the designation neither distinguishes or in any significant way describes anything.
It is no part of our present purpose to go into an exhaustive discussion of monopoly, and we may pass over the ordinary type of the phenomenon very briefly. In its original meaning the word signified an exclusive right to produce or sell a certain commodity, and was essentially a legal concept. The "legitimate" representative of the type in modern industry is the patented article for consumption—not patented production process (including machines, etc.), which will be considered later. Monopoly may also be based on mere financial power, on the threat of local underselling, boycott, and other forms of "unfair competition"; this amounts in effect to a voice in the control of property owned by others or their persons as well; that is, to part ownership. Free competition, of course, involves the complete, separate ownership of every productive agent or natural unit, and the exploitation of every one in a way to secure its maximum value yield. Any sort of violent interference with competition manifestly contradicts this assumption and may be roughly designated monopoly.
In the same category of monopoly (control of a consumption good) we may place two other varieties significant in the modern economic world. The first is the "corner," in which only a temporary control is secured, amounting in reality to control over the time of marketing of an existing stock not subject to rapid increase at the moment by further production. The other is the use of trademarks, trade names, advertising slogans, etc., and we may include the services of professional men with established reputations (whatever their real foundation). The buyer being the judge of his own wants, if the name makes a difference to him it constitutes a peculiarity in the commodity, however similar it may be in physical properties to competing wares. And the difference from physically equivalent goods may be very real, in the way of confidence in what one is getting. Such goods are then commodities whose supply is controlled by the producer, and competition with other makes or brands is a case of substitution of more or less similar goods, such as a monopolist always has to take into account.
A monopoly, of the category described, is evidently "productive" in the economic or mechanical causality sense. It may be viewed either as a separate productive element, in which case it is property in perfectly good business standing, and may be exchanged for other property on an income basis. Allowance will be made for the security of the income, but this allowance is perhaps as likely to be in favor of the monopoly as against it. Or we may take the view that the monopoly of a consumption good confers superior productivity on the agencies producing it, above physically identical agencies in other uses. As long as these are debarred in any way from producing the monopolized good the effect is the same as that of a physical incapacity to do so, and they are, like the branded article, economically differentiated, however similar physically. If the monopoly is of the character of a patent, and freely salable separately from the plant producing the goods, it is better to treat it as a productive agency on its own account.
Again, monopoly may consist in the exclusive control of the supply of some productive agency, physically defined as a group of interchangeable units. The only incentive to obtain such a monopoly is the desire to secure one of the former type, the power to restrict the supply of some consumption good. The control of any type of productive agent, of course, gives control of the supply of commodities whose production is dependent on the use of that agent, through the power to withhold the agent from use altogether or restrict its use in the making of any particular commodity while leaving its employment in other uses free. Whether the monopolist produces these goods himself or leases his monopolized agency to others, he can secure the entire increase in the net revenue from the final commodity as a rent on the restricted and restricting agency. It is evident in this case also that the restriction on the use of the agency, whatever its basis, is equivalent in effect to a physical peculiarity, and that the causal productivity of the agency is increased by its limitation in the same way as if part of it had gone out of existence or undergone some incapacitating change. Nor should it be necessary to insist again on the separation of the causality aspect of the case from the question of social policy.
A somewhat different case is the exclusive control of a peculiarly effective method or system of organization of production. The question of the productivity of a special process protected by patent or kept secret is a difficult one. Treatment of it in economic literature varies from that of Lavergne,111 who insists that the idée productrice is an independent factor, always present along with land, labor, and capital, to that of A. S. Johnson, who contends that an idea or method cannot be regarded as productive because it is the nature of an idea to multiply itself indefinitely.112 Here, again, the crucial test can only be the facts in the case. Does the method or idea get product imputed to it? This is largely a question of whether it is salable and so takes on capital value. If so; it is productive in the sense of economic causality. If it is not salable it will represent an element in the productivity of its possessor and its yield will accrue to him in the form of a wage. The moral question, whether it "ought" to be a source of income, is of course another matter. It seems evident113 on the one hand that the highest social advantage would require the most rapid and general extension of the use of the best methods, and it is of significance that this can theoretically be done nearly without cost. On the other hand, it is equally evident that both justice and expediency demand a fair reward for the origination of better ways of doing things. It would seem to be a matter of political development to provide a better way of rewarding these services than even a temporary monopoly of their use; but this inquiry belongs in the theory of progress, and as a question of social policy is outside the scope of the present study.
We must again insist, however, that the method must be recognized as being productive, or as conferring superior productivity on the agencies employed in connection with it.114 An arbitrary restriction is again causally equivalent to physical limitation. The method or idea is merely less productive of goods (and more productive of exchange value) than it would be if its use were unrestricted. The same paradox holds for any productive good; if multiplied indefinitely it would yield more goods in physical units, but have no value at all. The only difference in the case of a method of production is that it can be multiplied indefinitely without much cost (after once worked out), an important distinction from the standpoint of social policy (perhaps), but not significant from the standpoint of a cause and effect explanation of things. And we must again insist that the danger of reasoning about social totals of exchange value, and still more the extreme treachery of all reasoning about human welfare in terms of any such concept as economic utility, be borne in mind in attempting to reach conclusions as to social policy.115
The position taken above, that monopoly is productive, is in opposition to the doctrine of Professor J. B. Clark and his followers that the monopolist merely appropriates product created by other agents. But when monopoly income is said to be "diverted from its real producers,"116 or is called "exploitative," in the sense that it "is not secured by the agent that creates it,"117 the words "create" and "produce" are not used in their correct (causal) meaning. Monopoly is impossible except on the basis of some control over an element essential in the production of a commodity, and the extra product is rightly imputed to this essential element, or to the condition which makes control possible, if separable from the rest of the situation.
Monopoly of productive agencies has hitherto been of restricted importance in actual affairs, for several reasons. Most productive resources are specialized only to a limited extent, and are subject to effective competition from a wide range of substitutes. And in the hitherto undeveloped and rapidly changing condition of the world, most agencies, even of the most specialized types, have been rapidly and irregularly increasing in supply through new discoveries, and open to deliberate increase through moderate expenditures in exploration and development work. Finally, the technique of the large-scale organization requisite to secure unified control has been crude and imperfect, while the opposition of public opinion has been increasing in force. It is of some interest to inquire into the implications of absolutely free competition in this regard.
With perfect intercommunication it would seem that the assumed absence of collusion is very improbable, as organization costs would naturally tend to a low level. Under static conditions (with the existing stocks of all agencies fixed and known), a great development of monopoly would apparently be inevitable. It is not unreasonable to suppose even that in the absence of organized social interference conditions would approach the result contended for by the Marxian socialists, monopoly universal, or at least prevalent to an extent involving the complete breakdown of the competitive system of organization.
A further consideration, which goes back to the requirement of negligible size in the marginal unit as a condition of effective competition, tends to reinforce this view. In the ordinary sense of monopoly, concentration of control is not profitable unless it is nearly complete. But with organization costs absent or small, there might be a continuous incentive to increase the size of the bargaining unit. It is true, as some objectors to the productivity theory of distribution contend, that as the bargaining unit is larger the product theoretically dependent upon it is larger in greater ratio, and this fact affords a small incentive to combine even on a very small scale, and to increase the size of the unit without limit. The extra remuneration of the block over what it could obtain if its constituent units bargained separately would come out of the shares of the other agents used in connection with the one affected, not out of increased payments extorted from consumers as in case of monopoly.
The argument may be shown graphically by recourse to the "dosing method" of explaining specific productivity, made familiar by Professor J. B. Clark. There is no fallacy in this analysis if by a "factor" of production we mean merely a group of physically interchangeable things, and not a sort of labor or capital pulp obtained by putting things of all degrees of heterogeneity through the mill of the competitive process itself and reducing them to value productivity units. We must also remember that the method is a logical device purely, and in no sense represents the process by which productive services actually get evaluated. If, then, we imagine a static society, and fix our attention upon such a group of competing agents, it will be seen that the different units or members composing it may be regarded as placed along the descending curve of diminishing productivity of the familiar diagram. The curve, like that of diminishing utility and diminishing demand price,118 is purely hypothetical; the ordinate of each point merely shows what would be the productivity of each unit in the series if the total number were reduced to that indicated by the corresponding abscissa and production reorganized along "natural" lines. It does not indicate differences in productivity, or anything else, at the moment. We also pass over the fact that it is impossible to construct such a curve except for a very limited range in the region of known conditions and that any considerable extension of it (for an important productive service) soon carries us into the realm of pure fantasy.
But ignoring the difficulties and imagining the curve drawn, it is obvious that under theoretical imputation each member of any such group of competing agents will get what is directly dependent upon that which occupies the least important position, which is all that is ultimately "dependent" upon any one. But if two or more such agents combine so as to compete as a unit instead of separately, they can get the total product of that number of units at the lower end of the series, which is more than their separate "marginal" products. Therefore, under perfect competition, they will combine and bargain as a unit; and the same incentive will urge them to keep on combining until a monopoly results.
The situation is easily understood from the conventional diagram. If the curve CD represents the relative importance of successive agents of a series, or units of some really fundable agent, then under perfect competition every unit will get the product DE, and a certain group E'E will get FDE'E. If now these EE' units combine so as to become marginal as a group, they can get instead D'DE'E, gaining D'DF over the former arrangement. The owner of the group can prevent the substitution of a (marginal) unit outside the group for any unit in it, and so cause a larger product to be dependent on the employment of the group than the aggregate marginal products of its members. Similar agencies outside the combination will only get the wage DE, and the surplus income received by our consolidated block will come out of the shares of the agencies with which it is combined, not out of an increase in the price of the product to consumers. The employers of the "block" use no more nor less of the agency than before and make no more nor less product; hence they must sell the same supply at the same price. But the other agencies are forced to take less for their services because the block cannot be replaced a unit at a time from the margin, but only by an equal number of marginal units at once, a transfer which will raise their price all along the line. Only "friction" (human limitations) prevents this in actual society, the "diminishing returns of entrepreneurship."
It need not be remarked that this process would not go far in fact until something would have to be done to stop it. There does seem to be a certain Hegelian self-contradiction in the idea of theoretically perfect competition after all. As to what the end would be, it is fruitless to speculate, but it would have to be some arbitrary system of distribution under some sort of social control, doubtless based on ethics or political power or brute force, according to the circumstances—providing that society or somebody in it had sufficient intelligence and power to prevent a reversion to the bellum omnium contra omnes. Competitive industry is or hitherto has been saved by the fact that the human individual has been found normally incapable of wielding to his own advantage much more industrial power than, aided by legal and moral restraints, society as a whole can safely permit him to possess. How long this beneficent limitation can be counted upon to play its saving rôle may in the light of current business development occasion some doubt. With this subject we are not here particularly concerned, but it has seemed worth while to point out, in connection with the discussion of an ideal system of perfect competition, that such a system is inherently self-defeating and could not exist in the real world. Perfect competition implies conditions, especially as to the presence of human limitations, which would at the same time facilitate monopoly, make organization through free contract impossible, and force an authoritarian system upon society.119
In connection with the meaning of productivity it is of interest to raise the question of the economic value of the State. What would be the effect upon our economic life if society as such, acting through the political organization, should assert itself as an economic individual and charge "what the traffic will bear" for its own service? Obviously the Government has a monopoly on an absolutely indispensable commodity. Business could not be carried on at all without the protection of property and enforcement of contract. Into this interesting, but intricate, question it is impossible to enter at length here, but it appears that what the Government could take, its economic product, is hardly limited.120 The writer is much more optimistic as to the possibilities of a drastic program of taxation for securing a greater degree of economic equality than over most proposals for social interference in contractual relations.
IMPERFECT COMPETITION THROUGH