Chapter 10 — Monopoly and Competition
2. Cartels and Their Consequences
A. Cartels and “Monopoly Price”
But is not monopolizing action a restriction of production, and is not this restriction a demonstrably antisocial act? Let us first take what would seem to be the worst possible case of such action: the actual destruction of part of a product by a cartel. This is done to take advantage of an inelastic demand curve and to raise the price to gain a greater monetary income for the whole group. We can visualize, for example, the case of a coffee cartel burning great quantities of coffee.
In the first place, such actions will surely occur very seldom. Actual destruction of its product is clearly a highly wasteful act, even for the cartel; it is obvious that the factors of production which the growers had expended in producing the coffee have been spent in vain. Clearly, the production of the total quantity of coffee itself has proved to be an error, and the burning of coffee is only the aftermath and reflection of the error. Yet, because of the uncertainty of the future, errors are often made. Man could labor and invest for years in the production of a good which, it may turn out, consumers hardly want at all. If, for example, consumers’ tastes had changed so that coffee would not be demanded by anyone, regardless of price, it would again have to be destroyed, with or without a cartel.
Error is certainly unfortunate, but it cannot be considered immoral or antisocial; nobody aims deliberately at error.  If coffee were a durable good, it is obvious that the cartel would not destroy it, but would store it for gradual future sale to consumers, thus earning income on the “surplus” coffee. In an evenly rotating economy, where errors are barred by definition, there would be no destruction, since optimum stocks for the attainment of money income would be produced in advance. Less coffee would be produced from the beginning. The waste lies in the excessive production of coffee at the expense of other goods that could have been produced. The waste does not lie in the actual burning of the coffee. After the production of coffee is lowered, the other factors which would have gone into coffee production will not be wasted; the other land, labor, etc., will go into other and more profitable uses. It is true that excess specific factors will remain idle; but this is always the fate of specific factors when the realities of consumer demand do not sustain their use in production. For example, if there is a sudden dwindling of consumer demand for a good, so that it becomes unremunerative for labor to work with certain specialized machines, this “idle capacity” is not a social waste, but is rather socially useful. It is proved an error to have produced the machines; and now that the machines are produced, working on them turns out to be less profitable than working with other lands and machines to produce some other result. Therefore, the economical step is to leave them idle or perhaps to transform their material stuff into other uses. Of course, in an errorless economy, no excessive specific capital goods will be produced.
Suppose, for example, that before the coffee cartel went into operation, X amount of labor and Y amount of land co-operated to produce 100 million pounds of coffee a year. The coffee cartel determined, however, that the most remunerative production was 60 million pounds and therefore reduced annual output to this level. It would have been absurd, of course, to continue wasteful production of 100 million pounds and then to burn 40 millions. But what of the now surplus labor and land? These shift to the production, say, of 10 million pounds of rubber, 50,000 hours of service as jungle guides, etc. Who is to say that the second structure of production, the second allocation of factors, is less “just” than the first? In fact, we may say it is more just, since the new allocation of factors will be more profitable, and hence more value-productive, to consumers. In the value sense, then, overall production has now expanded, not contracted. It is clear we cannot say that production, overall, has been restricted, since output of goods other than coffee has increased, and the only comparison between the decline of one good and the increase in another must be made in these broad valuational terms. Indeed, the shifting of factors to rubber and jungle guidance no more restricts coffee production than a previous shift of factors to coffee restricted the production of the former goods.
It follows from our analysis that the producers’ original production of 100 million pounds was an unfortunate error, later corrected by them. Instead of being a vicious restriction of production to the detriment of the consumers, the cutback in coffee production was, on the contrary, a correction of the previous error. Since only the free market can allocate resources to serve the consumer, in accordance with monetary profitability, it follows that in the previous situation, “too much” coffee and “too little” rubber, jungle guide service, etc., were being produced. The cartel’s action, in reducing the production of coffee and causing an increase in the production of rubber, jungle guiding, etc., led to an increase in the power of the productive resources to satisfy consumer desires.
Furthermore, as we have stated above, an inelastic demand curve is purely the result of consumers’ choice. Thus, suppose that 100 million pounds of coffee have been produced and lie in stock, and a group of growers jointly decide that a burning of 40 million pounds of coffee will, say, double the price from one gold grain per pound to two gold grains per pound, thus giving them a higher total income acting jointly. This would be impossible if the growers knew that they would be confronted with an effective consumer boycott at the higher price. Further, consumers have another way, if they so desire, to prevent destruction of the good. Various consumers, acting either individually or jointly, could offer to purchase the existing coffee at higher than present prices. They could do this either because of their desire for coffee or because of their philanthropic dismay at the destruction of a useful good, or from a combination of both motives. At any rate, if they did so, they would prevent the producers’ cartel from decreasing the supply sold on the market. The boycott at a higher price and/or increased offers at the lower price would change the demand curve and render it elastic at the present stock level, thereby removing any incentive or need for the formation of a cartel.
To regard a cartel as immoral or as hampering some sort of consumers’ sovereignty is therefore completely unwarranted. And this is true even in the seemingly “worst” case of a cartel that we may assume is founded solely for “restrictive” purposes, and where, as a result of previous error and the perishability of product, actual destruction will occur. If consumers really wish to prevent this action, they need only change their demand schedules for the product, either by an actual change in their taste for coffee or by a combination of boycott and philanthropy. The fact that such a development does not take place in any given circumstance signifies that the producers are still maximizing their monetary income in the service of the consumers — by a cartel action, as well as by any other action. Some readers might object that, in offering higher demands for existing stock, the consumers would be bribing the producers, and that this constitutes an unwarranted extortion on the part of the producers. But this charge is untenable. Producers are guided by the goal of maximizing monetary income; they are not extorting, but simply producing where their gains are at a maximum, through exchanges concluded voluntarily by producers and consumers alike. This is no more nor less a case of “extortion” than when a laborer shifts from a lower-paying to a higher-paying job or when an entrepreneur invests in what he thinks will be a more rather than a less profitable project.
D. The Instability of the Cartel
Analysis demonstrates that a cartel is an inherently unstable form of operation. If the joint pooling of assets in a common cause proves in the long run to be profitable for each of the individual members of the cartel, then they will act formally to merge into one large firm. The cartel then disappears in the merger. On the other hand, if the joint action proves unprofitable for one or more members, the dissatisfied firm or firms will break away from the cartel, and, as we shall see, any such independent action almost always destroys the cartel. The cartel form, therefore, is bound to be highly evanescent and unstable.
If joint action is the most efficient and profitable course for each member, a merger will soon take place. The very fact that each member firm retains its potential independence in the cartel means that a breakup could take place at any time. The cartel will have to assign production totals and quotas to each of the member firms. This is likely to lead first to a good deal of bickering among the firms over the assignment of quotas, with each member attempting to gain a larger share of the assignment. Whatever basis quotas are assigned on will necessarily be arbitrary and will always be subject to challenge by one or more members.  In a merger, or in the formation of one corporation, the stockholders, by majority vote, form a decision-making organization. In the case of a cartel, however, disputes arise among independent owning entities.
Particularly likely to be restive under the imposed joint action will be the more efficient producers, who will be eager to expand their business rather than be fettered by shackles and quotas to provide shelter for their less efficient competitors. Clearly, the more efficient firms will be the ones to break up the cartel. This will be increasingly true as time goes on and conditions change from the time the cartel was first formed. The quotas, the jealously made agreements that formerly seemed plausible to all, now become intolerable restrictions for the more efficient firms, and the cartel soon breaks up; for once one firm breaks away, expands output and cuts prices, the others must follow.
If the cartel does not break up from within, it is even more likely to do so from without. To the extent that it has earned unusual monopoly profits, outside firms and outside producers will enter the same field of production. Outsiders, in short, rush in to take advantage of the higher profits. But once one strong competitor arises to challenge it, the cartel is doomed. For as the firms in the cartel are bound by production quotas, they must watch new competitors expand and take away sales from them at an accelerating rate. As a result, the cartel must break up under the pressure of the newcomers’ competition. 
E. Free Competition and Cartels
Some critics charge that there is no “real” free entry or free competition in a free market. For how can anyone compete or enter a field when an enormous amount of money is needed to invest in efficient plants and firms? It is easy to “enter” the pushcart peddling “industry” because so little capital is required, but it is almost impossible to establish a new automobile firm, with its heavy requirements of capital.
This argument is but another variant of the prevailing confusion between freedom and abundance. In this case, the abundance refers to the money capital which a man has been able to amass. Every man is perfectly free to become a baseball player; but this freedom does not imply that he will be as good a baseball player as the next man. A man’s range or power of action, dependent on his ability and the exchange-value of his property, is something completely distinct from his freedom. As we have said, a free society will in the long run lead to general abundance and is the necessary condition for that abundance. But the two must be kept conceptually distinct, and not confused by phrases such as “real freedom” or “true freedom.” Therefore, the fact that everyone is free to enter an industry does not mean that everyone is able, either in terms of personal qualities or monetary capital, to do so. In industries requiring more capital, fewer people will be able to take advantage of their freedom to set up a new firm than in those requiring less capital, just as fewer laborers will be able to take advantage of freedom of entry in a very highly skilled profession than in a menial position. There is no mystery about either situation.
In fact, the disability is much more relevant in the case of labor than in the case of business competition. What are modern devices such as corporations but means of pooling capital by many people of greater and lesser wealth? The “difficulty” of investing in a new automobile firm should be considered, not in terms of the hundreds of millions of dollars required for total investment, but in terms of the 50 or so dollars required to purchase one share of stock. But while capital can be pooled, beginning with the smallest units, labor ability cannot be pooled.
Sometimes the argument reaches absurd lengths. For example, it is often asserted that now, in this modern world, firms are so large that new people “cannot” compete or enter the industry because the capital cannot be raised. These critics do not seem to see that the aggregate capital and wealth of individuals have advanced along with the increase in wealth required to launch a new enterprise. In fact, these are two sides of the same coin. There is no reason to suppose that it was easier to raise the capital to launch a new retail shop many centuries ago than it is to raise capital for the automobile firm today. If there is enough capital to finance the large firms currently existing, there is enough to finance one more; in fact, capital could be withdrawn from existing large firms and shifted to new ones if there is a need for them. Of course, if the new enterprise would be unprofitable and therefore unserviceable to consumers, it is easy to see why there is reluctance in the free market to embark on the venture.
F. The Problem of One Big Cartel
The myth of the evil cartel has been greatly bolstered by the nightmare image of “one big cartel.” “This is all very well,” one may say, “but suppose that all the firms in the country amalgamated or cartelized into One Big Cartel. What of the horrors then?”
The answer can be obtained by referring to chapter 9, pp. 612ff above, where we saw that the free market placed definite limits on the size of the firm, i.e., the limits of calculability on the market. In order to calculate the profits and losses of each branch, a firm must be able to refer its internal operations to external markets for each of the various factors and intermediate products. When any of these external markets disappears, because all are absorbed within the province of a single firm, calculability disappears, and there is no way for the firm rationally to allocate factors to that specific area. The more these limits are encroached upon, the greater and greater will be the sphere of irrationality, and the more difficult it will be to avoid losses. One big cartel would not be able rationally to allocate producers’ goods at all and hence could not avoid severe losses. Consequently, it could never really be established, and, if tried, would quickly break asunder.
Let us assume for a moment that One Big Cartel could be established on the free market and that the calculability problem does not arise. What would the economic consequences be? Would the cartel be able to “exploit” anyone? In the first place, consumers could not be “exploited.” For consumers’ demand curves would still be elastic or inelastic, as the case may be. Since, as we shall see further below, consumers’ demand curves for a firm are always elastic above the free-market equilibrium price, it follows that the cartel will not be able to raise prices or earn more from consumers.
What about the factors? Could not their owners be exploited by the cartel? In the first place, the universal cartel, to be effective, would have to include owners of primary land; otherwise whatever gains they might have might be imputed to land. To put it in its strongest terms, then, could a universal cartel of all land and capital goods “exploit” laborers by systematically paying the latter less than their discounted marginal value products? Could not the members of the cartel agree to pay a very low sum to these workers? If that happened, however, there would be created great opportunities for entrepreneurs either to spring up outside the cartel or to break away from the cartel and profit by hiring workers for a higher wage. This competition would have the double effect of (a) breaking up the universal cartel and (b) tending again to yield to the laborers their marginal product. As long as competition is free, unhampered by governmental restrictions, no universal cartel could either exploit labor or remain universal for any length of time. 
3. The Illusion of Monopoly Price
A. Definitions of Monopoly
One common objection is that Ford is able to acquire “monopoly power” or “monopolistic power” because his product has a recognized brand name or trade-mark, which the wheat farmer has not. This, however, is surely a case of putting the cart before the horse. The brand name and the wide knowledge of the brand come from consumers’ desire for the product attached to that particular brand and are therefore a result of consumer demand rather than a pre-existing means for some sort of “monopolistic power” over the consumers. In fact, farmer Hiram Jones is perfectly free to stamp the brand name “Hiram Jones Wheat” on his product and attempt to sell it on the market. The fact that he has not done so signifies that it would not be a profitable step in the concrete market condition of his product. The chief point is that in some cases consumers and lower-order entrepreneurs consider each individual brand name as representing a unique product, while in other cases purchasers consider the output of one firm — one product-owner or set of product-owners operating jointly — as identical in use-value with products of other firms. Which situation will occur is entirely dependent on the buyers’ valuations in each concrete case.
 Professor Lawrence Abbott, in one of the outstanding theoretical works of recent years, demonstrates also that as civilization and the economy advance, products will become more and more differentiated and less and less homogeneous. For one thing, greater differentiation occurs at the consumer than at the producer level, and the expanding economy takes over an increasing proportion of goods once made by the consumer himself and therefore supplies more finished goods than raw materials to the consumer than formerly (bread rather than flour, sweaters rather than wool yarn, etc.). Thus, there is greater opportunity for differentiation.
Furthermore, to the familiar charge that business advertising tends to create differentiation in the consumer’s mind that is not “really” there, Abbott replies incisively that the reverse is more likely to be true and that advancing civilization increases the consumer’s perception and discrimination of differences of which he was previously ignorant. Writes Abbott:
as man becomes more civilized, he develops greater powers of perception with regard to quality differences. Subjective homogeneity may exist even when objective homogeneity does not, due to the inability or unwillingness of buyers to perceive differences between almost identical products and discriminate between them. . . . As a society matures and education improves, people learn to develop more acute powers of discrimination. Their wants become more detailed. They begin . . . to develop a preference, say, not simply for white wine, but for 1948 Chablis. . . . People generally tend to underestimate the significance of apparently trivial differences in fields in which they are not expert. An unmusical person may be unwilling to concede that there is any difference in tone between a Steinway and a Chickering piano, being unable himself to detect it. A nongolfer is more likely than a habitual player to believe that all brands of golf balls are virtually alike. (Lawrence Abbott, Quality and Competition [New York: Columbia University Press, 1955], pp. 18–19, and chap. I)
Also see ibid., pp. 45–46 and Edward H. Chamberlin, “Product Heterogeneity and Public Policy” in Towards a More General Theory of Value (New York: Oxford University Press, 1957), p. 96.
C. Consequences of Monopoly-Price Theory
(1) The Competitive Environment
Before engaging in a critical analysis of the monopoly-price theory itself, we might explore some of the consequences which do or do not follow from it. In this section we for the moment assume that the monopoly-price theory is valid.  In the first place, it is not true that the “monopolist” (used here in the sense of definition 3 — an obtainer of a monopoly price) is removed from the influence of competition or has the power to dictate to consumers at will. The best of the monopoly-price theorists admit that the monopolist is as subject to the forces of competition as are other firms. The monopolist cannot set prices as high as he would like, being limited by the configurations of consumer demand. By definition, in fact, the demand curve as presented to the monopolist becomes elastic above the monopoly-price point. There has been an unfortunate tendency of writers to refer to an “elastic demand curve” or an “inelastic demand curve” without pointing out that every curve has different ranges along which there will be varying degrees of elasticity or inelasticity. By definition, the monopoly-price point is that which maximizes the firm’s or the cartel’s income; above that price any further “restriction” of production and sales will lower the monopolist’s monetary income. This implies that the demand curve will become elastic above that point, just as it is also elastic above the competitive-price point when that is established on the market. Consumers make the curve elastic by their power of substituting purchases of other goods. Many other goods compete “directly” in their use-value to the consumer. If some firm or combination of firms should, for example, achieve a monopoly-price for cake soap, housewives can shift to detergents and thus limit the height of the monopoly price. But, in addition, all goods, without exception, compete for the consumer’s dollar or gold ounce. If the price of yachts becomes too high, the consumer can substitute expenditure on mansions, or he can substitute books for television sets, etc. 
Furthermore, as the market advances, as capital is invested and the market becomes more and more specialized, the demand curve for each product tends to become more and more elastic. As the market develops, the range of consumers’ goods available increases enormously. The more consumers’ goods are available, the more goods can be purchased by consumers, and the more elastic, ceteris paribus, the demand curve for each good will tend to be. As a result, the opportunities for the establishment of monopoly prices will tend to diminish as the market and “capitalist” methods develop.
(3) A World of Monopoly Prices?
Is it possible, within the framework of monopoly-price theory, to assert that all prices on the free market may be monopoly prices?  Can all selling prices be monopoly prices?
There are two ways in which we may analyze this problem. One is by turning our attention to the monopolized industry. As we have seen, the industry with a monopoly price restricts production in that industry (either by a cartel or a single firm), thereby releasing nonspecific factors to enter other fields of production. But it is evidently impossible to conceive of a world of monopoly prices, because this would imply a piling up of unused nonspecific factors. Since wants do not remain unfulfilled, labor and other nonspecific factors will be used somewhere, and the industries that acquire more factors and produce more cannot be monopoly-price industries. Their prices will be below the competitive price level.
We may also consider consumer demand. We have seen that a necessary condition for the establishment of monopoly price is a consumers’ demand schedule inelastic above the competitive-price point. Obviously, it is impossible for every industry to have such an inelastic demand schedule. For the definition of inelastic is that consumers will spend a greater total sum of money on the good when the price is higher. But consumers have a certain given total stock of money assets and money income, as well as a given amount, at any one time, which they may allocate to consumption spending. If they spend more on a certain good, they have less to spend on other goods. Therefore, they cannot spend more on every good, and not all prices can be monopoly prices.
(4) “Cutthroat” Competition
A popular theme in the literature is the alleged evil of “cutthroat competition.” Curiously, cutthroat, or “excessive,” competition, is linked by critics to the achievement of a monopoly price. The usual charge is that a “big” firm, for example, deliberately sells below the most profitable price, even to the extent of suffering losses. The firm acts so peculiarly in order to force another firm producing the same product to cut its price also. The “stronger” firm, with the capital resources to endure the losses, then drives the “weaker” firm out of business and establishes a monopoly of the field.
But, first, what is wrong with such a monopoly (definition 1)? What is wrong with the fact that the firm more efficient in serving the consumer remains in business, while consumers refuse to patronize the inefficient firm? A firm’s suffering losses signifies that it is not as successful as other firms in serving consumer desires. Factors then shift from the inefficient to the efficient firms. […] Thus, the elimination of inefficient firms cannot harm factor-owners or lead to their “unemployment,” since their failure was due precisely to the more attractive competing bids made by other firms (or, in some cases, to the alternatives of leisure or production outside the market). Their failure also helps consumers by transferring resources from wasteful to efficient producers. […]
The only conceivable problem is the one usually cited: that after the single firm has driven everyone else out of business through sustained selling at very low prices, then the final monopolist will restrict sales and raise its price to a monopoly price. Even granting for a moment the tenability of the monopoly-price concept, this does not seem a very likely occurrence. In the first place, it is time enough to complain after the monopoly price is established, especially since we have seen that we cannot consider “monopoly” per se (definition 1) as an evil.  Secondly, a firm will not always be able to achieve a monopoly price. In all such cases, including (a) where not all the other firms in the industry can be driven out, or (b) where the demand curve is such that the monopolist cannot achieve a monopoly price, the “cutthroat competition” is then a pure boon with no harmful effects.
Incidentally, it is by no means true that the large firms will always be the strongest in a “price-cutting war.” Often, depending on the concrete conditions, it is the smaller, more mobile firm, not burdened with heavy investments, that is able to “cut its costs” (particularly when its factors are more specific to it, such as the labor of its management) and outcompete the larger firm. In such cases, of course, there is no monopoly-price problem whatever. The fact that the lowly pushcart peddler for centuries has been set upon by governmental violence at the behest of his more lordly and heavily capitalized competitors bears witness to the practical possibilities of such a situation. 
 What of the allegedly vast “financial power” of a big firm, rendering it impervious to cost? In a brilliant article, Professor Wayne Leeman has pointed out that a larger firm will also have larger volume and will therefore suffer greater losses when selling below cost. Having a larger volume, it has more to lose. What is relevant, therefore, is not the absolute size of the financial resources of the competing firms, but the size of their resources in relation to their volume of sales and expenditures. And this changes the conventional picture drastically. Wayne A. Leeman, “The Limitations of Local Price-Cutting as a Barrier to Entry,” Journal of Political Economy, August, 1956, pp. 331–32.
Suppose, however, that after this lengthy and costly process, a firm has finally been able to achieve a monopoly price by the route of “cutthroat competition.” What is there to prevent this monopoly gain from attracting other entrepreneurs who will try to undercut the existing firm and achieve some of the gain for themselves? What is to prevent new firms from coming in and driving the price down to competitive levels again? Is the firm to resume “cutthroat competition” and the same deliberate losing process once more? In that case, we are likely to find that consumers of the good will be receiving gifts far more often than facing a monopoly price. 
 After investigating conditions in the retail gasoline industry (one particularly subject to allegedly “cutthroat” competition), an economist declared:
Some people think that leading marketers occasionally reduce prices to drive out competition so that they may later enjoy a monopoly. But, as one oil man has put it, “That is like trying to sweep back the ocean to get a dry place to sit down . . . .” [Competitors] . . . never scare, and never hesitate for long, and would move in immediately when prices were restored, offering little opportunity to a single marketer to recoup his losses. (Harold Fleming, Oil Prices and Competition [American Petroleum Institute, 1953], p. 54)
Professor Leeman has pointed out  that the smaller firm, driven out by “cutthroat competition,” may simply close down, wait for the larger firm to reap its expected gain of a higher “monopoly price,” and then reopen! More important, even if the small firm is driven into bankruptcy, its physical plant remains intact, and it may be bought by a new entrepreneur at bargain prices. As a result, the new firm will be able to produce at very low cost and damage the “victor” firm considerably. To avoid this threat, the big firm would have to delay raising its price for the very long time required for the small plant to wear out or become obsolete.
Leeman also demonstrates that the big firm could not keep new, small firms out by a mere threat of cutthroat competition. For (a) new firms will probably interpret the high price charged by the “monopolist” as a sign of inefficiency, providing a ripe opportunity for profits; and (b) the “monopolist” can demonstrate his power satisfactorily only by actually selling at low prices for long periods of time. Hence, only by keeping its costs down and its prices low, i.e., by not extracting a monopoly price, can the “victor” firm keep out potential rivals. But this means that the cutthroat competition, far from being a route to a monopoly price, was a pure gift to consumers and a pure loss to the “victor.” 
But what of a standard problem brought forward by critics of cutthroat competition”? Cannot the big firm check the entry of efficient small firms by simply buying up the new rival’s plant and putting it out of production? Perhaps a short period of cutthroat price-cutting will convince the new small firm of the advantage of selling out and will permit the monopolist to avoid the long periods of losses just mentioned.
No one seems to realize, however, the high costs such buying will entail. Leeman points out that the really efficient small firm can demand such a high price for its assets as to make the whole procedure prohibitively expensive. And, further, any later attempt by the large firm to recoup its losses by charging the monopoly price will only invite new entry by other firms and redouble the expensive buying-out process again and again. Buying out competitors, then, will be even more costly than simple cutthroat competition, which we have seen to be unprofitable.  
 Leeman points out, in a striking refutation of one of the myths of our age, that this is precisely what happened to John D. Rockefeller.
According to a widely accepted view, he softened up small competitors in the oil business by a period of intensive price competition, bought them out for a song, and then raised prices to consumers to make up his losses. Actually, the softening-up process did not work . . . for Rockefeller usually ended up paying . . . so handsomely that the sellers, often in violation of promises made, proceeded to build another plant for its nuisance value, hoping again to collect a reward from their benefactor. . . . Rockefeller after a time got tired of paying . . . “blackmail” and . . . decided that the best way to hold the dominant position he wanted was to keep profit margins small all the time. (Ibid., p. 332)
Also see Marian V. Sears, “The American Businessman at the Turn of the Century,” The Business History Review, December, 1956, p. 391. Moreover, Professor McGee has shown, after an intensive investigation, that in not one instance did Standard Oil attempt “predatory price-cutting,” thus destroying the Standard Oil myth once and for all. John S. McGee, “Predatory Price-Cutting: The Standard Oil (New Jersey) Case,” The Journal of Law and Economics, October, 1958, pp. 137–69.
 Leeman concludes, quite correctly, that large rather than small firms dominate many markets, not as a result of victorious cutthroat competition and monopolistic pricing, but by taking advantage of the low costs of much large-scale production and keeping prices low in fear of potential as well as actual rivals. Leeman, “The Limitations of Local Price-Cutting,” pp. 333–34.
[…] The cutting of price may just as well be due to inability to dispose of stock at any higher price as to “cutthroat” competition, and it is impossible for an observer to separate the two elements.
D. The Illusion of Monopoly Price on the Unhampered Market
Let us take a firm which is considering the production of a certain good. The firm can be a “monopolist” in the sense of producing a unique good, or it can be an “oligopolist” among a few firms. Whatever its position, it is irrelevant, because we are interested only in whether or not it can achieve a monopoly price as compared to a competitive price. This, in turn, depends on the elasticity of the demand curve as it is presented to the firm over a certain range. Let us say that the firm finds itself with a certain demand curve (Figure 68).
The producer must decide how much of the good to produce and sell in a future period, i.e., at the time when this demand curve will become relevant. He will set his output at whatever point is expected to maximize his monetary earnings (other psychic factors being equal), taking into consideration the necessary monetary expenses of production for each quantity, i.e., the amounts that can be produced for each amount of money invested. […]
On the basis of this logic of action, the producer sets his investment to produce a certain stock, or as a factor-owner to sell a certain amount of service, say 0S. Assuming that he has correctly estimated his demand curve, the intersection of the two will establish the market-equilibrium price, 0P or SA.
The critical question is this: Is the market price, 0P, a “competitive price” or a “monopoly price”? The answer is that there is no way of knowing. Contrary to the assumptions of the theory, there is no “competitive price” which is clearly established somewhere, and which we may compare 0P with. Neither does the elasticity of the demand curve establish any criterion. Even if all the difficulties of discovering and identifying the demand curve were waived (and this identifying can be done, of course, only by the producer himself — and only in a tentative fashion), we have seen that the price, if accurately estimated, will always be set by the seller so that the range above the market price will be elastic. How is anyone, including the producer himself, to know whether or not this market price is competitive or monopoly?
Suppose that, after having produced 0S, the producer decides that he will make more money if he produces less of the good in the next period. Is the higher price to be gained from such a cutback necessarily a “monopoly price”? Why could it not just as well be a movement from a subcompetitive price to a competitive price? In the real world, a demand curve is not simply “given” to a producer, but must be estimated and discovered. If a producer has produced too much in one period and, in order to earn more income, produces less in the next period, this is all that can be said about the action. For there is no criterion that will determine whether or not he is moving from a price below the alleged “competitive price” or moving above this price. Thus, we cannot use “restriction of production” as the test of monopoly vs. competitive price. A movement from a subcompetitive to a competitive price also involves a “restriction” of production of this good, coupled, of course, with an expansion of production in other lines by the released factors. There is no way whatever to distinguish such a “restriction” and corollary expansion from the alleged “monopoly-price” situation.
If the “restriction” is accompanied by increased leisure for the owner of a labor factor rather than increased production of some other good on the market, it is still an expansion of the yield of a consumers’ good — leisure. There is still no way of determining whether the “restriction” resulted in a “monopoly” or a “competitive” price or to what extent the motive of increased leisure was involved.
To define a monopoly price as a price attained by selling a smaller quantity of a product at a higher price is therefore meaningless, since the same definition applies to the “competitive price” as compared with a subcompetitive price. There is no way to define “monopoly price” because there is also no way of defining the “competitive price” to which the former must refer.
Many writers have attempted to establish some criterion for distinguishing a monopoly price from a competitive price. Some call the monopoly price that price achieving permanent, long-run “monopoly profits” for a firm. This is contrasted to the “competitive price,” at which, in the evenly rotating economy, profits disappear. Yet, as we have already seen, there are never permanent monopoly profits, but only monopoly gains to owners of land or labor factors. Money costs to the entrepreneur, who must buy factors of production, will tend to equal money revenues in the evenly rotating economy, whether the price is competitive or monopoly. The monopoly gains, however, are secured as income to labor or land factors. There is therefore never any identifiable element that could provide a criterion of the absence of monopoly gain. With a monopoly gain, the factor’s income is greater; without it, it is less. But where is the criterion for distinguishing this from a change in the income of a factor for “legitimate” demand and supply reasons? How to distinguish a “monopoly gain” from a simple increase in factor income?
Another theory attempts to define a monopoly gain as income to a factor greater than that received by another, similar factor. Thus, if Mickey Mantle receives a greater monetary income than another outfielder, that difference represents the “monopoly gain” resulting from his natural monopoly of unique ability. The crucial difficulty with this approach is that it implicitly adopts the old classical fallacy of treating all the various labor factors, as well as all the various land factors, as somehow homogeneous. If all the labor factors are somehow one good, then the variations in income accruing to each must be explained by reference to some sort of “monopolistic” or other mysterious element. Yet a good with a homogeneous supply is only a good if all its units are interchangeable, as we saw at the beginning of this work. But the very fact that Mantle and the other outfielder are treated differently in the market signifies that they are selling different, not the same, goods. Just as in tangible commodities, so in personal labor services (whether sold to other producers or to consumers directly): each seller may be selling a unique good, and yet he is “competing” with more or less close substitutability against all the other sellers for the purchases of consumers (or lower-order producers). But since each good or service is unique, we cannot state that the difference between the prices of any two represents any sort of “monopoly price”; monopoly price vis-à-vis competitive price can refer only to alternative prices of the same good. Mickey Mantle may indeed be a person of unique ability and a “monopolist” (as is everyone else) over the disposition of his own talents, but whether or not he is achieving a “monopoly price” (and therefore a monopoly gain) from his service can never be determined.
This analysis is equally applicable to land. It is just as illegitimate to dub the difference between the income of the site of the Empire State Building and that of a rural general store a “monopoly gain” as to apply the same concept to the additional income of Mickey Mantle. The fact that both areas are land makes them no more homogeneous on the market than the fact that Mickey Mantle and Joe Doakes are both baseball players or, in a broader category, both laborers. The fact that each is remunerated at a different price and income signifies that they are considered different on the market. To treat differential gains for different goods as instances of “monopoly gain” is to render the term completely devoid of significance.
Neither is the attempt to establish the existence of idle resources as a criterion of monopolistic “withholding” of factors any more valid. Idle labor resources will always mean increased leisure, and therefore the leisure motive will always be intertwined with any alleged “monopolistic” motive. It therefore becomes impossible to separate them. The existence of idle land may always be due to the fact of the relative scarcity of labor as compared with available land. This relative scarcity makes it more serviceable to consumers, and hence more remunerative, to invest labor in certain areas of land, and not in others. The land areas least productive of potential earnings will be forced to lie idle, the amount depending on how much labor supply is available. We must stress that all “land” (i.e., every nature-given resource) is involved here, including urban sites and natural resources as well as agricultural areas. The allocation of labor to land is comparable to Crusoe’s having to decide on which plot of ground to build his shelter or in which stream to fish. Because of the natural, as well as voluntary, limitations on his labor effort, that area of land on which he produces the highest utility will be cultivated, and the rest will be left idle. This element also cannot be separated from any alleged monopolistic element. For if someone objects that the “withheld” land is of the same quality as the land in use and therefore that monopolistic restriction is afoot, it may always be answered that the two pieces of land necessarily differ — in location if in no other attribute — and that the very fact that the two are treated differently on the market tends to confirm this difference. By what mystical criterion, then, does some outsider assert that the two lands are economically identical? In the case of capital goods it is also true that the limitations of available labor supply will often make idle those goods which are expected to yield a lesser return as compared with other capital that can be employed by labor. The difference here is that idle capital goods are always the result of previous error by producers, since no such idleness would be necessary if the present events — demands, prices, supplies — had all been forecast correctly by all the producers. But though error is always unfortunate, the keeping idle of unremunerative capital is the best course to follow; it is making the best of the existing situation, not of the situation that would have obtained if foresight had been perfect. In the evenly rotating economy, of course, there would never be idle capital goods; there would be only idle land and idle labor (to the extent that leisure is voluntarily preferred to money income). In no case is it possible to establish an identification of purely “monopolistic” withholding action.
A similar proposed criterion for distinguishing a monopoly price from a competitive price runs as follows: In the competitive case, the marginal factor produces no rent; in the monopoly-price case, however, use of the monopolized factor is restricted, so that its marginal use does yield a rent. We may answer, in the first place, that there is no reason to say that every factor will, in the competitive case, always be worked until it yields no rent. On the contrary, every factor is worked in a region of diminishing but positive marginal product, not zero product. Indeed, as we have shown above, if the value product of a unit of a factor is zero, it will not be used at all. […]
It is clear, further, that this criterion could never be applied to a monopolized labor factor. What labor factor earns a zero wage in a competitive market? Yet many monopolized (definition 1) factors are labor factors — such as brand names, unique services, decision-making ability in business, etc. Land is more abundant than labor, and therefore some lands will be idle and receive zero rent. Even here, however, it is only the submarginal lands that receive no rent; the marginal lands in use receive some rent, however small.
[…] But a criterion of monopoly or competitive price must apply, not to factors of different quality, but to homogeneous factors. The monopoly-price problem is one of a supply of units of one homogeneous factor, not of various different factors within the one broad category, land. In this case, as we have stated, every factor will earn some value product in a diminishing zone, and not zero. 
Since, in the “competitive” case, all factors in use will earn some rent, there is still no basis for distinguishing a “competitive” from a “monopoly” price.
Another very common attempt to distinguish between a competitive and a monopoly price rests on the alleged ideal of “marginal-cost pricing.” Failure to set prices equal to marginal cost is considered an example of “monopoly” behavior. There are several fatal errors in this analysis. In the first place, as we shall see further below, there can be no such thing as “pure competition,” that hypothetical state in which the demand curve for the output of a firm is infinitely elastic. Only in this never-never land does price equal marginal cost in equilibrium. Otherwise, marginal cost equals “marginal revenue” in the ERE, i.e., the revenue that a given increment of cost will yield to the firm. (Only if the demand curve were perfectly elastic would marginal revenue boil down to “average revenue,” or price.) There is now no way of distinguishing “competitive” from “monopolistic” situations, since marginal cost will in all cases tend to equal marginal revenue.
Secondly, this equality is only a tendency that results from competition; it is not a precondition of competition. It is a property of the equilibrium of the ERE that the market economy always tends toward, but never can reach. To uphold it as a “welfare ideal” for the real world, an ideal with which to gauge existing conditions, as so many economists have done, is to misconceive completely the nature of the market and of economics itself.
Thirdly, there is no reason why firms should ever deliberately balk at being guided by marginal-cost considerations. Their aiming at maximum net revenue will see to that. But there is no one simple, determinate “marginal cost,” because, as we have seen above, there is no one identifiable “short-run” period, such as is assumed by current theory. The firm faces a gamut of variable periods of time for the investment and use of factors, and its pricing and output decisions depend on the future period of time which it is considering. Is it buying a new machine, or is it selling old output piled up in inventory? The marginal cost considerations will differ in the two cases.
E. Some Problems in the Theory of the Illusion of Monopoly Price
(1) Location Monopoly
Let us, however, be generous to the location-monopoly theorists and grant that, in a sense (definition 1) this monopoly is enjoyed by all individual sellers of any good or service. This is due to the eternal law of human action, and indeed of all matter, that only one thing can be in one place at one time. The retail grocer on Fifth Street enjoys a monopoly of the sale of groceries for that street; the grocer on Fourth Street enjoys a monopoly of grocery service for his street, etc. In the case of stores which all cluster together in the same block, say radio stores, there are still a few feet of sidewalk over which each owner of a radio store exercises a location monopoly. Location is as specific to a firm or plant as ability is to a person.
Whether this element of location takes on any importance in the market depends on the configuration of consumer demand and on which policy is most profitable for each seller in the concrete case. In some cases a grocer, for example, can charge higher prices for his goods than another because of his monopoly of the block. In that case, his monopoly over the good “eggs available on Fifth Street” has taken on such a significance for the consumers in his block that he can charge them a higher price than the Fourth Street grocer and still retain their patronage. In other cases, he cannot do so because the bulk of his customers will desert him for the neighboring grocer if the latter’s prices are lower.
Now, a good is homogeneous if consumers evaluate its units in the same way. If that condition holds, its units will be sold for a uniform price on the market (or rapidly tend to be sold at a uniform price). If, now, various grocers must adhere to a uniform price, then there is no location monopoly.
But what of the case where the Fifth Street grocer can charge a higher price than his competitor? Do we not have here a clear case of an identifiable monopoly price? Can we not say that the Fifth Street grocer who can charge more than his competitor for the same goods has found that the demand curve for his products is inelastic for a certain range above the “competitive price,” the competitive price being taken as that equal to the price charged by his neighbor? Can we not say this even though we recognize that there is no “infringement on consumers’ sovereignty” in this action, since it is due to the specific tastes of his consuming customers? The answer is an emphatic No. The reason is that the economist can never equate a good with some physical substance. A good, we remember, is a quantity of a thing divisible into a supply of homogeneous units. And this homogeneity, we repeat, must be in the minds of the consuming public, not in its physical composition. If a malted milk consumed at a luncheonette is the same good in the minds of consumers as the malted at a fashionable restaurant, then the price of the malted will be the same in both places. On the other hand, we have seen that the consumer buys not only the physical good, but all attributes of a thing, including its name, the wrappings, and the atmosphere in which it is consumed. If most of the consumers differentiate sufficiently between food consumed in the restaurant and food consumed at the luncheonette, so that a higher price can be charged in one case than in the other, then the food is a different good in each case. A malted consumed in the restaurant becomes, for a significant body of consumers, a different good from a malted consumed at the luncheonette. The same situation obtains for brand names, even in those situations where a minority of the consumers do regard several brands as “actually” the same good. As long as the bulk of the consumers regard them as different goods, then they are different goods, and their prices will differ. Similarly, goods may differ physically, but as long as they are regarded by consumers as the same, they are the same good. 
The same analysis applies to the case of location. Where the Fifth Street consumers regard groceries at Fifth Street as a significantly better good than groceries at Fourth Street, so that they are willing to pay more rather than walk the extra distance, then the two will become different goods. In the case of location, there will always be a tendency for the two to be different goods, but very often this will not be significant on the market. For a consumer may and almost always will prefer groceries available on this block to groceries available on the next block, but often this preference will not be enough to overcome any higher price for the former goods. If the bulk of the consumers shift to the latter good at a higher price, the two, on the market, will be the same good. And it is action on the market, real action, that we are interested in, not the nonsignificant pure valuations by themselves. In praxeology we are interested only in preferences that result in, and are therefore demonstrated by, real choices, not in the preferences themselves.
A good cannot be independently established as such apart from consumer preference on the market. Groceries on Fifth Street may be higher in price than groceries on Fourth Street to the Fifth Street consumers. If so, it will be because the former is a different good to the consumers. In the same way, Rochester cement may cost more than Albany cement in Albany to Rochester consumers, but the two are different goods by virtue of their difference in location. And there is no way of determining whether or not the price in Rochester or on Fifth Street is a “monopoly price” or a “competitive price” or of determining what the “competitive price” might be. It certainly could not be the price charged by the other firm elsewhere, since these prices are really for two different goods. There is no theoretical criterion by which we can distinguish simple locational income to sites from alleged “monopoly” income to sites.
There is another reason for abandoning any theory of locational monopoly price. If all sites are purely specific in locational value, there is no sense to the statement that they earn a “monopoly rent.” For monopoly price, according to the theory, can be established only by selling less of a good and thus commanding a higher price. But all locational properties of a site differ in quality because they differ in location, and therefore there can be no restriction of sales to part of a site. Either a site is in production, or it is idle. But the idle sites necessarily differ in location from the sites in use and are therefore idle because their value productivity is inferior. They are idle because they are submarginal, not because they are “monopolistically” withheld parts of a certain homogeneous supply.
The locational-monopoly-price theorist, then, is refuted whichever way he turns. If he takes a limited view of locational monopoly (in the sense of definition 1) and confines it to such examples as Rochester vs. Albany, he can never establish a criterion for monopoly price, for another firm can enter Rochester, either actually or potentially, to bid away any locational profit that the first firm may earn. His prices cannot be compared with those of his competitors, because they are selling different goods. If the theorist takes an extensive view of locational monopoly — which would take into consideration the fact that every location necessarily differs from every other — and compares locations a few feet apart, then there is no sense at all in talking of “monopoly price,” for (a) the price of a product at one location cannot be precisely compared with another, because they are different goods, and (b) each site is different in locational quality, and therefore no site can be conceptually split up into different homogeneous units — some to be sold and some to be withheld from the market. Each site is a unit in itself. But such a splitting is essential for the establishment of a monopoly-price theory.
B. Some Arguments for Unions: A Critique
(3) Greater Efficiency and the “Ricardo Effect”
One common prounion argument is that unions benefit the economy through forcing higher wages on the employers. At these higher wages the workers will become more efficient, and their marginal productivity will rise as a result. If this were true, however, no unions would be needed. Employers, ever eager for greater profits, would see this and pay higher wages now to reap the benefits of the allegedly higher productivity in the future. As a matter of fact, employers often train workers, paying higher wages than their present marginal product justifies, in order to reap the benefits of their increased productivity in later years.
A more sophisticated variant of this thesis was advanced by Ricardo and has been revived by Hayek. This doctrine holds that union-induced higher wage rates encourage employers to substitute machinery for labor. This added machinery increases the capital per worker and raises the marginal productivity of labor, thereby paying for the higher wage rates. The fallacy here is that only increased saving can make more capital available. Capital investment is limited by saving. Union wage increases do not increase the total supply of capital available. Therefore, there can be no general rise in labor productivity. Instead, the potential supply of capital is shifted (not increased) from other industries to those industries with higher wage rates. And it is shifted to industries where it would have been less profitable under nonunion conditions. The fact that an induced higher wage rate shifts capital to the industry does not indicate economic progress, but rather an attempt, never fully successful, to offset an economic retrogression — a higher cost in the manufacture of the product. Hence, the shift is “uneconomic.”
A related thesis is that higher wage rates will spur employers to invent new technological methods to make labor more efficient. Here again, however, the supply of capital goods is limited by the savings available, and there is almost always a sheaf of technological opportunities awaiting more capital anyway. Furthermore, the spur of competition and the desire of the producer to keep and increase his custom is enough of an incentive to increase productivity in his firm, without the added burden of unionism. 
5. The Theory of Monopolistic or Imperfect Competition
A. Monopolistic Competitive Price
Essentially, the chief characteristic of the imperfect-competition theories is that they uphold as their “ideal” the state of “pure competition” rather than “competition” or “free competition.” Pure competition is defined as that state in which the demand curve for each firm in the economy is perfectly elastic, i.e., the demand curve as presented to the firm is completely horizontal. In this supposedly pristine state of affairs, no one firm can, through its actions, possibly have any influence over the price of its product. Its price is then “set” for it by the market. […]
The pure-competition theory, however, is an utterly fallacious one. It envisages an absurd state of affairs, never realizable in practice, and far from idyllic if it were. In the first place, there can be no such thing as a firm without influence on its price. The monopolistic-competition theorist contrasts this ideal firm with those firms that have some influence on the determination of price and are therefore in some degree “monopolistic.” Yet it is obvious that the demand curve to a firm cannot be perfectly elastic throughout. At some points, it must dip downward, since the increase in supply will tend to lower market price. As a matter of fact, it is clear from our construction of the demand curve that there can be no stretch of the demand curve, however small, that is horizontal, although there can be small vertical stretches. In aggregating the market demand curve, we saw that for each hypothetical price, the consumers will decide to purchase a certain amount. If the producers attempt to sell a larger amount, they will have to conclude their sale at a lower price in order to attract an increased demand. Even a very small increase in supply will lead to a perhaps very small lowering of price. The individual firm, no matter how small, always has a perceptible influence on the total supply. In an industry of small wheat farms (the implicit model for “pure competition”), each small farm contributes a part of the total supply, and there can be no total without a contribution from each farm. Therefore, each farm has a perceptible, even if very small, influence. No perfectly elastic demand curve can, then, be postulated even in such a case. The error in believing in “perfect elasticity” stems from the use of such mathematical concepts as “second order of smalls,” by which infinite negligibility of steps can be assumed. But economics analyzes real human action, and such real action must always be concerned with discrete, perceptible steps, and never with “infinitely small” steps.
Of course, the demand curve for each small wheat farm is likely to be very highly, almost perfectly, elastic. And yet the fact that it is not “perfect” destroys the entire concept of pure competition. For how does this situation differ from, say, the Hershey Chocolate Company if the demand curve for the latter firm is also elastic? Once it is conceded that all demand curves to firms must be falling, the monopolistic-competition theorist can make no further analytic distinctions.
C. Chamberlin and Selling Cost
Chamberlin falls into another error in implying that selling costs, such as advertising, “create” consumer demands. This is the determinist fallacy. Every man as a self-owner freely decides his own scale of valuations. On the free market no one can force another to choose his product. And no other individual can ever “create” someone’s values for him; he must adopt the value himself. 
 See Mises:
The consumer is, according to . . . legend, simply defenseless against “high-pressure” advertising. If this were true, success or failure in business would depend on the mode of advertising only. However, nobody believes that any kind of advertising would have succeeded in making the candlemakers hold the field against the electric bulb, the horsedrivers against the motorcars. . . . But this implies that the quality of the commodity advertised is instrumental in bringing about the success of an advertising campaign. . . . The tricks and artifices of advertising are available to the seller of the better product no less than to the seller of the poorer product. But only the former enjoys the advantages derived from the better quality of his product. (Mises, Human Action, pp. 317–18)
6. Multiform Prices and Monopoly
Up to this point we have always concluded that the market tends, at any given time, to establish one uniform market price for any good, under competitive or monopoly conditions. One phenomenon that sometimes appears, however, is persistent multiformity of prices. […]
The seller aims, as always, to maximize his earnings in voluntary exchange, and he certainly cannot be held responsible for the ignorance of the buyer. If buyers do not take the trouble to inform themselves of the state of the market, they must stand prepared to have some of their psychic surplus tapped by the bargaining of the seller. Neither is this action irrational on the part of the buyer. For we must deduce from the buyer’s action that he prefers to remain in ignorance rather than to make the effort or pay the money to inform himself of market conditions. To acquire knowledge of any field takes time, effort, and often money, and it is perfectly reasonable for an individual on any given market to prefer to take his chances on the price and use his scarce resources in other directions. This choice is crystal clear in the case of a tourist on holiday, but it is also possible in any other given market. Both the impatient tourist, who prefers to pay a higher price and not spend time and money on learning about the market, and a companion who spends days on an intensive study of the bazaar market are exercising their preferences, and praxeology cannot call one or the other more rational. Furthermore, there is no way to measure the consumer surpluses lost or gained in the case of the two tourists. […]
What if the good is not resalable? In that case, there is far greater room for multiform pricing, since ignorance is not required. A vendor can sell an intangible service at a higher price to A than to B without fear that B can undercut him by reselling to A. Hence, most actual cases of multiform pricing take place in the realm of intangible goods.
Suppose now that seller X has managed to establish multiform prices for his customers. He might be a lawyer, for example, who charges higher fees for the same service to a wealthy than to a poor client. Since there is still competition among sellers, why does another lawyer Y not enter the field and undercut X’s price to the wealthy clients? In fact, this is what will generally happen, and any attempt to establish “separate markets” among customers will lead to an invasion of the more profitable, higher-price field by other competitors, finally driving the price down, reducing revenues, and re-establishing uniform pricing. If a seller’s service is unusual and it is universally recognized that he has no effective competitors, then he might be able to sustain a multiform structure. […]
Multiform pricing has been accorded a curious reception by economists and laymen. In some cases it is deemed vicious exploitation of the consumers; in others (e.g., medicine and education) it is considered praiseworthy and humanitarian. In reality, it is neither. It is certainly not the rule in pricing that the most eager should pay in proportion to their eagerness (in practice, usually gauged by their wealth), for then everyone would pay in proportion to his wealth for everything, and the entire monetary and economic system would break down; money would no longer function. (See chapter 12 below.) If this is clear in general, it is difficult to see a priori why specific goods should be singled out for this treatment. On the other hand, the consumers are not being “exploited” if there is multiformity. It is clear that the marginal and submarginal buyers are not exploited: the latter obviously gain. What of the supramarginal buyers who are receiving less consumer surplus? In some cases, they gain, because without the greater revenues provided by “price discrimination” the good would not be supplied at all. Consider, for example, a country doctor who would leave the area if he had to subsist on the lower revenues provided by uniformity. And even if the good were still supplied, the fact that the supramarginal buyers continue to patronize the seller at all shows that they are content with the seemingly discriminatory arrangement. Otherwise, they would quickly boycott the seller, either individually or in concert, and patronize competitors. They would simply refuse to pay more than the submarginal buyers, and this would quickly induce the seller to lower his prices. The fact that they do not do so shows that they prefer multiformity to uniformity in the particular case. An example is private school education, which able but poor youths may often attend on scholarships — a principle that the wealthy parents who pay full tuition demonstrably do not consider unjust. If, however, the sellers have received grants of monopolistic privilege by the government, enabling them to restrict competition in the serving of the supramarginal buyers, then they may establish multiformity without enjoying the demonstrable preference of these buyers: for here governmental coercion has entered to inhibit the free expression of preferences.