Man, Economy, and State with Power and Market

Perhaps the most important conclusion of the theory of monopolistic or imperfect competition is that the real world of monopolistic competition (where the demand curve to each firm is necessarily falling) is inferior to the ideal world of pure competition (where no firm can affect its price). This conclusion was expressed simply and effectively by comparing two final equilibrium states: under conditions of pure and monopolistic competition (Figure 70).

AC is a firm’s average total-cost curve—its alternative dollar costs per unit—with output on the horizontal axis and prices (including costs) on the vertical axis. The only assumption we need in drawing the average-cost curve is that, for any plant in any branch of production, there will be some optimum point of production, i.e., some level of output at which average unit cost is at a minimum. All levels of production lower or higher than the optimum have a higher average cost. In pure competition, where the demand curve for any firm is perfectly elastic, Dp, each firm will eventually adjust so that its AC curve will be tangent to Dp, in equilibrium; in this case, at point E. For if average revenue (price) is greater than average cost, then competition will draw in other firms, until the curves are tangent; if the cost curve is irretrievably higher than demand, the firm will go out of business. Tangency is at point E, price at 0G, and output at 0K. As in any definition of final equilibrium, total costs equal total revenues for each firm, and profits are zero.

Now contrast this picture with that of monopolistic competition. Since the demand curve (Dmf) is now sloping downward to the right, it must, given the same AC curve, be tangent at some point (F), where the price is higher (JF) and the production lower (0J) than under pure competition. In short, monopolistic competition yields higher prices and less production—i.e., a lower standard of living—than pure competition. Furthermore, output will not take place at the point of minimum average cost—clearly a social “optimum,” and each plant will produce at a lower than optimum level, i.e., it will have “excess capacity.” This was the “welfare” case of the monopolistic-competition theorists.

By a process of revision in recent years, some of it by the originators of the doctrine themselves, this theory has been effectively riddled beyond repair. As we have seen, Chamberlin and others have shown that this analysis does not apply if we are to take consumer desire for diversity as a good to be satisfied.75 Many other effective and sound attacks have been made from different directions. One basic argument is that the situations of pure and of monopolistic competition cannot be compared because the AC curves would not, in fact, be the same. Chamberlin has pursued his revisionism in this realm also, declaring that the comparisons are wholly illegitimate, that to apply the concept of pure competition to existing firms would mean, for example, assuming a very large number of similar firms producing the identical product. If this were done, say, with General Motors, it would mean that either GM must conceptually be divided up into numerous fragments, or else that it be multiplied. If divided, then unit costs would undoubtedly be higher, and then the “competitive firm” would suffer higher costs and have to subsist on higher prices. This would clearly injure consumers and the standard of living; thus, Chamberlin follows Schumpeter’s criticism that the “monopolistic” firm may well have and probably will have lower costs than its “purely competitive” counterpart. If, on the other hand, we conceive of the multiplication of a very large number of General Motors corporations at existing size, we cannot possibly relate it to the present world, and the whole comparison becomes absurd.76

In addition, Schumpeter has stressed the superiority of the “monopolistic” firm for innovation and progress, and Clark has shown the inapplicability, in various ways, of this static theory to the dynamic real world. He has recently shown its fallacious asymmetry of argument with respect to price and quality. Hayek and Lachmann have also pointed out the distortion of dynamic reality, as we have indicated above.77

A second major line of attack has shown that the comparisons are much less important than they seem from conventional diagrams, because cost curves are empirically much flatter than they appear in the textbooks. Clark has emphasized that firms deal in long-run considerations, and that long-run cost and demand curves are both more elastic than short-run; hence the differences between E and F points will be negligible and may be nonexistent. Clark and others have stressed the vital importance of potential competition to any would-be reaper of monopoly price, from firms both within and without the industry, and also the competition of substitutes between industries. A further argument has been that the cost curves, empirically, are flat within the relevant range, even aside from the long- vs. short-run problems.78

All these arguments, added to our own analysis given above, have effectively demolished the theory of monopolistic competition, and yet more remains to be said. There is something very peculiar about the entire construction, even on its own terms, aside from the fallacious “cost-curve” approach, and practically no one has pointed out these other grave defects in the theory. In an economy that is almost altogether “monopolistically competitive,” how can every firm produce too little and charge too much? What happens to the surplus factors? What are they doing? The failure to raise this question stems from the modern neglect of Austrian general analysis and from undue concentration on an isolated firm or industry.79 The excess factors must go somewhere, and in that case must they not go to other monopolistically competitive firms? In which case, the thesis breaks down as self-contradictory. But the proponents have prepared a way out. They take, first, the case of pure competition, with equilibrium at point E. Then, they assume a sudden shift to conditions of monopolistic competition, with the demand curve for the firm now sloping downward. The demand curve now shifts from Dp to Dmo. Then the firm restricts production and raises its price accordingly, reaps profits, attracts new firms entering the industry, the new competition reduces the output salable by each firm, and the demand curve shifts downward and to the left until it is tangent to the AC curve at point F. Hence, say the monopolistic-competition theorists, not only does monopolistic competition suffer from too little production in each firm and excessive costs and prices; it also suffers from too many firms in each industry. Here is what has happened to the excess factors: they are trapped in too many uneconomic firms.

This seems plausible, until we realize that the whole example has been constructed as a trick. If we isolate a firm or an industry, as does the example, we may just as well start from a position of monopolistic competition, at point F, and then suddenly shift to conditions of pure competition. This is certainly just as legitimate, or rather illegitimate, a base for comparison. What then? As we see in Figure 71, the demand curve for each firm is now shifted from Dmf to Dpo. It will now be profitable for each firm to expand its output, and it will then make profits. New firms will then be attracted into the industry, and the demand curve will fall vertically, until it again reaches tangency with the AC curve at point E. Are we now “proving” that there are more firms in an industry under pure than under monopolistic competition?80 The fundamental error here is failure to see that, under the conditions established by the assumptions, any change opening up profits will bring new firms into an industry. Yet the theorists are supposed to be comparing two different static equilibria, of pure and of monopolistic competition, and not discussing paths from one to the other. Thus, the monopolistic-competition theorists have by no means solved their problem of surplus factors.

But, aside from this point, there are more difficulties in the theory, and Sir Roy Harrod, himself one of its originators, is the only one to have seized the essence of the remaining central difficulty. As Harrod says:

If the entrepreneur foresees the trend of events, which will in due course limit his profitable output to x – y units, why not plan to have a plant that will produce x – y units most cheaply, rather than encumber himself with excess capacity? To plan a plant for producing x units, while knowing that it will only be possible to maintain an output of x – y units, is surely to suffer from schizophrenia.

And yet, asserts Harrod puzzledly, the “accepted doctrine” apparently deems it “impossible to be an entrepreneur and not suffer from schizophrenia!”81 In short, the theory assumes that, in the long run, a firm having to produce at F will yet construct a plant with minimum costs at point E. Clearly, here is a patent contradiction with reality. What is wrong? Harrod’s own answer is an excellent and novel discussion of the difference between long-run and short-run demand curves, with the “long run” always being a factor in entrepreneurial planning, but he does not precisely answer this question.

The paradox becomes “curiouser and curiouser” when we fully realize that it all hinges on a mathematical technicality. The reason why a firm can never produce at an optimum cost point is that (a) it must produce at a tangent of demand and average-cost curves in equilibrium, and (b) if the demand curve is falling, it follows that it can be tangent to a U-shaped cost curve only at some point higher than, and to the left of, the trough point. There are two considerations that we may now add. First, there is no reason why the cost “curve” should, in fact, be curved. In an older day, textbook demand curves used to be curves, and now they are often straight lines; there is even more reason for believing that cost curves are a series of angular lines. It is of course (a) more convenient for diagrams, and (b) essential to mathematical representation, for there to be continuous curves, but we must never let reality be falsified in order to fit the niceties of mathematics. In fact, production is a series of discrete alternatives, as all human action is discrete, and cannot be smoothly continuous, i.e., move in infinitely small steps from one production level to another. But once we recognize the discrete, angular nature of the cost curve, the “problem” of excess capacity immediately disappears (Figure 72). Thus the falling demand curve to the “monopolistic” firm, Dm, can now be “tangent” to the AC curve at E, the minimum-cost point, and will be so in final equilibrium.

There is another way for this pseudo problem to disappear, and that is to call into question the entire assumption of tangency. The tangency of average cost and demand at equilibrium has appeared to follow from the property of equilibrium: that total costs and total revenues of the firm will be equal, since profits as well as losses will be zero. But a key question has been either overlooked or wrongly handled. Why should the firm produce anything, after all, if it earns nothing from doing so? But it will earn something, in equilibrium, and that will be interest return. Modern orthodoxy has fallen into this error, for one reason: because it does not realize that entrepreneurs are also capitalists and that even if, in an evenly rotating economy, the strictly entrepreneurial function were no longer to be required, the capital-advancing function would still be emphatically necessary.

Modern theory also tends to view interest return as a cost to the firm. Naturally, if this is done, then the presence of interest does not change matters. But (and here we refer the reader to foregoing chapters) interest is not a cost to the firm; it is an earning by a firm. The contrary belief rests on a superficial concentration on loan interest and on an unwarranted separation between entrepreneurs and capitalists. Actually, loans are unimportant and are only another legal form of entrepreneurial-capitalist investment. In short, in the evenly rotating economy, the firm earns a “natural” interest return, dictated by social time preference. Hence, Figure 72 must be altered to look like the diagram in Figure 73 (setting aside the problem of curves vs. angles). The firm will produce 0K, its optimum production level, at minimum average cost, KE. Its demand curve and cost curve will not be tangent to each other, but will allow room for equilibrium interest return, represented by the area EFGH. (Neither, as some may object, will the price be higher in this corrected version of monopolistic competition; for this AC curve is lower all around than the previous ones, which had included interest return in costs. If they did not include interest, and instead assumed that interest would be zero in the ERE, then they were wrong, as we have pointed out above.)82 And so the paradox of the monopolistic-competition theory is finally and fully interred.83

• 75And the product differentiation associated with the falling demand curve may well lower costs of distribution and of inspection (as well as improve consumer knowledge) to more than offset the supposed rise in production costs. In short, the AC curve above is really a production-cost, rather than a total-cost, curve, neglecting distribution costs. Cf. Goldman, “Product Differentiation and Advertising.” Furthermore, a genuine total-cost curve would then not be independent of the firm’s demand curve, thus vitiating the usual “cost-curve” analysis. See Dewey, Monopoly in Economics and Law, p. 87. Also see section C below.
• 76See Chamberlin, “Measuring the Degree of Monopoly and Competition” and “Monopolistic Competition Revisited” in Towards a More General Theory of Value, pp. 45–83.
• 77See J.M. Clark, “Competition and the Objectives of Government Policy” in E.H. Chamberlin, ed., Monopoly and Competition and Their Regulation (London: Macmillan & Co., 1954), pp. 317–27; Clark, “Toward a Concept of Workable Competition” in Readings in the Social Control of Industry (Philadelphia: Blakiston, 1942), pp. 452–76; Clark, “Discussion”; Abbott, Quality and Competition, passim; Joseph A. Schumpeter, Capitalism, Socialism and Democrac (New York: Harper & Bros., 1942); Hayek, “Meaning of Competition”; Lachmann, “Some Notes on Economic Thought, 1933–53.”
• 78See the above citations by Clark; and Richard B. Heflebower, “Toward a Theory of Industrial Markets and Prices” in R.B. Heflebower and G.W. Stocking, eds., Readings on Industrial Organization and Public Policy (Homewood, Ill.: Richard D. Irwin, 1958), pp. 297–315. A more dubious argument—the flatness of the firm’s demand curve in the relevant range—has been stressed by other economists, notably A.J. Nichol, “The Influence of Marginal Buyers on Monopolistic Competition,” Quarterly Journal of Economics, November, 1934, pp. 121–34; Alfred Nicols, “The Rehabilitation of Pure Competition,” Quarterly Journal of Economics, November, 1947, pp. 31–63; and Nutter, “Plateau Demand Curve and Utility Theory.”
• 79But cf. Abbott, Quality and Competition, pp. 180–81.
• 80The author first learned this particular piece of analysis from the classroom lectures of Professor Arthur F. Burns, and, to our knowledge, it has never seen print.
• 81Harrod, Economic Essays, p. 149.
• 82After arriving at this conclusion, the author came across a brilliant but neglected article pointing out that interest is a return and not a cost, and showing the devastating implications of this fact for cost-curve theory. The article does not, however, apply the theory satisfactorily to the problem of monopolistic competition. See Gabor and Pearce, “A New Approach to the Theory of the Firm,” and idem, “The Place of Money Capital.” While there are a few similarities, Professor Dewey’s critique of the “excess capacity” doctrine is essentially very different from ours and based on far more “orthodox” considerations. Dewey, Monopoly and Economics in Law, pp. 96 ff.
• 83Since the erroneous but popular theory of “countervailing power,” propounded by J.K. Galbraith, falls with the monopolistic-competition theory, it is unnecessary to discuss it here. For a more detailed critique of its numerous fallacies, see Simon N. Whitney, “Errors in the Concept of Countervailing Power,” Journal of Business, October, 1953, pp. 238–53; George J. Stigler, “The Economist Plays with Blocs,” American Economic Review, Papers and Proceedings, May, 1954, pp. 8–14; and David McCord Wright, “Discussion,” ibid., pp. 26–30.