Competition

views updated

Competition

Perfect competition

The theory of competitive prices

Alternative concepts

Empirical evidence of competition

Public policies

BIBLIOGRAPHY

OTHER WORKS CITED

Competition may be the spice of life, but in economics it has been more nearly the main dish. Competition has been a major force in the organization of production and the determination of prices and incomes: economic theory has accorded commensurate importance to the concept.

Competition enters all major areas of man’s life and generally connotes rivalry between two or more men or groups for a given prize. Competition is often an end in itself. Sporting events are clear illustrations: we should be shocked if two teams called off the event or arranged a tie and divided the prize. Indeed, the prize is a minor goal in a true sporting event.

In economic life competition is not a goal: it is a means of organizing economic activity to achieve a goal. The economic role of competition is to discipline the various participants in economic life to provide their goods and services skillfully and cheaply.

Perfect competition

Market competition

When one asks (as Cournot was the first to do in a precise way in 1838) whether the competition of three merchants will serve better than two, or why two (or three) do not combine into a monopoly, the answers prove to be elusive. But one can partially evade such questions by posing a very extreme degree of competition, which the economist calls perfect competition.

A main requirement of perfect competition is that the largest firm in an industry make a trifling fraction of the industry’s sales (or purchases) and therefore that there be many firms in the industry. No definite number has been found for the maximum share of one firm that is compatible with competition; presumably the permissible share can be larger the more elastic the industry demand and the easier the conditions of entry by new firms.

These many firms, no one or few of which account for an appreciable share of the industry’s output, are assumed to act independently. This can be viewed as a second condition for perfect competition, or as an inevitable corollary of the large numbers in the absence of legal controls over the industry. For it is a fact that there are insuperable difficulties in organizing an effective combination of many persons when it is profitable for each person secretly to depart from the agreement, as is generally the case in economic life.

Such large numbers suggest what is true: that (perfect) economic competition is impersonal. In the economic race there are 1,000 or 100,000 runners, and each gets a prize proportional to his efforts. The fortunes of any one firm are independent of what happens to any other firm: one farmer is not benefited if his neighbor’s crop is destroyed. The essence of perfect competition, therefore, is not strong rivalry but rather the utter dispersion of power to influence market behavior. The power, for example, to restrict quantities sold and raise prices is effectively annihilated when it is divided among a thousand men, just as a gallon of water is effectively annihilated if it is spread over a thousand acres.

A third condition of perfect competition is complete knowledge of offers to buy and sell by the participants in the market. This condition serves just the opposite purpose of the preceding condition. The assumption that traders act independently serves to keep them apart and, hence, numerous; the assumption that each seller knows what various buyers will pay, and vice versa, is necessary to keep the parties together—in the same market. If seller S and buyer B dealt only with one another in ignorance of all other traders, and similarly for every other pair of buyers and sellers, each transaction would represent an exchange under bilateral monopoly.

These conditions of perfect competition are enough to ensure that a single price will rule in a market (in fact, perfect knowledge is enough for this purpose) and that this price is affected only negligibly by the actions of any one or few buyers or sellers. (It is sometimes additionally assumed that the product of all sellers be homogeneous, but this can also be viewed as part of the definition of the market or industry.) The definition of perfect competition is therefore sometimes expressed in the equivalent form: the demand curve facing each seller is infinitely elastic; and the supply curve facing each buyer is infinitely elastic. (This definition also applies to the individual firm, which accordingly may be competitive even though the market in which it trades is not competitive.)

To these basic conditions of perfect competition—numerous traders on each side of the market, independence of action, and perfect knowledge—it is necessary to add divisibility of the commodity or service being traded. If the units are large, it is possible that minor discontinuities will emerge that allow some small market power to individuals. The point is sufficiently minor to be left to references (Edgeworth [1881] 1953, p. 46; Stigler 1957, pp. 8–9).

These conditions of perfect competition pertain to a single market, whether of shoes or bonds or carpenter’s services. So far as the presence or absence of monopoly power is concerned, it is not necessary to look at any other market. For this reason these conditions pertain to what may be called market competition.

It is traditional, however, to enlarge the conditions of competition, so they will ensure an optimal allocation of resources, by specifying the nature of the movement of resources among markets and industries. This enlarged concept, which may be termed industrial competition, is our next subject.

Industrial competition

If a productive resource is to be utilized efficiently, it must be equally productive in all of its uses—clearly if its (marginal) product is less in one use than another, output is not being maximized. Hence, two additional conditions commonly have been made a part of perfect competition: resources are mobile among uses; and their owners are informed as to yields in these various uses.

A vast galaxy of private and public barriers to the mobility of resources has been erected in various times and places: boycotts, certificates of convenience and necessity, patent licenses, settlements laws, franchises, licensing of occupations. Such barriers are all actually or potentially incompatible with competition. But it is not necessary for competition that the movement of resources be free: the retraining of a worker, or the transportation of a tool, may be costly without interfering with competition. We must enlarge our earlier condition of full information to include knowledge of the yields on resources in alternative employments. From another viewpoint, we may say that ignorance is a barrier to profitable movement of resources.

If these conditions are fulfilled, the maximum possible output (measured by value) will be obtained from a productive resource. If this be true for each resource, the output of the economy is at a maximum. This famous theorem (labeled “on maximum satisfaction” by Walras and Marshall) is subject to a qualification, as all interesting propositions are: the private marginal product of a productive resource (the amount its owner receives and hence what governs its allocation) must equal the social marginal product (private marginal product plus or minus the effects on others). Of course, the maximum value output is dependent upon the distribution of income, which affects the demands for goods and hence their prices.

Time and competition. What we have termed industrial competition—competition including mobility of resources—obviously has an implicit time dimension. It takes time to move resources out of unprofitable fields, especially if the resources are specialized and durable, so that only through disentangling depreciation funds can the resources be withdrawn. It takes time, too, to construct a new factory or shop when one wishes to enter an industry. Comparable statements can be made about the geographical and occupational mobility of labor. Similarly, time is a factor in the completeness of knowledge. It takes time to learn which industries or jobs are most remunerative, or to learn the prices quoted by various sellers, or the quality of service and product; and one’s knowledge is more complete and reliable, the more thorough the search for information and the larger the experience on which it is based.

Capital embodied in specialized and durable equipment will not be transferred to other uses in the short run except at extreme price differentials, even though in the long run the slightest differential in returns may be sufficient to move capital funds. Conversely, only under extreme incentives will new establishments be created virtually overnight, as we sometimes observe in wartime.

This fact that it is more expensive to do things very quickly than at a slower pace does not qualify the proposition that resources will tend to be put where they earn the most, but we are reminded of the implicit proviso: allowance must be made for the cost of moving the resources.

The differences in returns to a resource in various uses can be very great in the short run, but will decline in the long run to a minimum level set by the cost of the most efficient method of moving resources. There is implicit in economic literature the belief that these minimum costs of movement of resources are very small relative to their returns, so little imprecision arises from neglecting them entirely. This may be true but has not been demonstrated. The belief, nevertheless, led economists (for example, J. B. Clark) to postulate instantaneous and costless mobility as the pure case of perfect industrial competition. It seems preferable to say that minimum differentials in returns to resources are achieved only in the long run. Market competition is not so intimately related to time. One’s information about price bids and offers improves somewhat as one searches the market more thoroughly—itself a time-consuming process—but the changing conditions of supply and demand lead to changes in prices that make the old information obsolete.

The theory of competitive prices

The competitive structure of industry leads to the establishment of competitive prices. Competitive prices are characterized by two main properties. The property of clearing markets is that of distributing existing supplies efficiently; the property of equalizing returns to resources is that of directing production efficiently.

The clearing of markets

A competitive price is one that is not perceptibly influenced by any one buyer or seller. When we say that such prices are fixed by “supply and demand” we mean that the ensemble of all buyers and sellers determines price.

Since every buyer can purchase all he wishes of the good or service at the market price, there are no queues or unsatisfied demands, given the price. Since every seller can sell all he wishes at this market price, there are no undisposable stocks, other than inventories that are voluntarily held for future periods. The competitive price, then, clears the market—it equates the quantities offered by sellers and sought by buyers.

Whenever we find a persistent queue among buyers, we know that the price is being held below the level that clears the market, which we naturally call the equilibrium price. For example, when housing is unavailable under rent controls, we know that rents are below the equilibrium level. Whenever we find stocks held by sellers to be in excess of inventory needs, we know price is above the equilibrium level. The vast stocks of agricultural products held by the U.S. government are evidence that the prices of these products (more precisely, the amounts the government will lend on the products) are above the equilibrium level.

The importance of prices that clear markets is that they put goods and services in the hands of the people who most urgently wish them. If a price is held too low, some buyers who set a lower value on the commodity will get it while others in the queue who set a higher value get none. If the price is set too high, goods that buyers would be glad to purchase at a lower price go unsold even though (if a minimum price is imposed on a competitive industry) sellers would prefer to sell at this lower price.

The equalization of returns

It is part of the definition of industrial competition that every resource in an industry earn as much as it would earn in other industries, but no more. The self-interest of the owners of productive resources (including, of course, that most important resource, the laborer) leads them to apply their resources where they yield the most and thus to enter unusually attractive fields and abandon unattractive fields.

This equalization of returns, however, can be shown to imply that the prices of goods and services equal their (marginal) costs of production. The cost of a productive service to an industry is the amount that must be paid to attract it away from other uses—its foregone alternatives. (This most basic concept of cost is the essence of the alternative or opportunity cost theory.) If the amount the productive resource earns in an industry is in excess of this cost, clearly other units of the resource presently outside the industry could earn more if they entered. Conversely, if the productive resource is earning less than its cost or alternative product, it will leave the industry. Hence, if price exceeds cost, resources will flow into the industry and lower the price (and perhaps raise cost by raising the prices of the resources); if price is less than cost, resources will flow out and increase the price (and perhaps reduce cost).

The equality of the marginal products of a resource in all its uses is the condition for efficient production. The equality of average products has often been substituted, with a regrettable loss of logic: consider the catastrophic waste (of capital) in having equal output per worker in two industries when the capital equipment per worker is ten times as large in one industry as in the other. But if the marginal product of a resource is equal in its various uses, it follows that marginal cost must equal price. The resources necessary to produce one more unit of product A could produce an equal value of B, so the marginal cost of A—which is the foregone alternative of producing B—is equal to the value of A that it produces. Marginal cost, formally defined as an increment of cost divided by the increment of product associated with the increment of cost, and not the more easily measured average cost (total cost divided by output), is the economist’s fundamental criterion of competitive price—and of optimum price.

Marshall’s period analysis

The alternative uses open to a resource depend upon the time available for its redeployment (or more fundamentally, how much one is willing to spend on its movement). This principle, joined to an empirical observation that one can alter the rate of operation of a plant much sooner than one can build a new plant or wear out an existing one, provides the basis for the standard (Marshallian) theory of long-run and short-run competitive prices (Marshall 1890).

In the short run, defined as the period within which one cannot appreciably alter the number of plants (physical production units), the only method of varying output is to work a given plant more or less intensively. The so-called variable productive factors (labor, materials, fuel) are the only resources with effective alternative uses in this period and therefore the only services whose returns enter into marginal costs. The returns to the productive factors embodied in the plant are called quasi rents. So long as quasi rents are greater than zero it will be more profitable to operate a plant than to close it down.

The long run is defined as the period within which the entrepreneur can make any desired decision—including the decision to leave one industry and enter another. In this period all resources are variable in quantity, and therefore the returns to all factors enter into marginal cost.

The Marshallian apparatus permits very useful simplifications in price theory, but only if its underlying empirical assumption is fulfilled: the long-run adjustments of the firm are of negligible magnitude in the short run (and hence can be neglected), and the short-run adjustments do not appreciably affect the long-run costs. When these conditions are not met (they fail, for example, if discharge of workers in this period will lead to higher wage rates in the next period), the full analysis of the short run will still require explicit analysis of the long-run repercussions of the short-run decisions.

Alternative concepts

The austerity and abstractness of the concept of perfect competition have led many economists to seek a more “realistic” concept. This search has been reinforced by the need for a concept of competition usable in the enforcement of the U.S. antitrust statutes. A variety of concepts have accordingly been proposed, but because they were deliberately contrived to fit the infinitely varied circumstances of a vast economy they lack the analytical clarity of perfect competition.

Workable competition

The most popular of these variant concepts is that of J. M. Clark, which he labeled workable competition (1940). The philosophy of this concept is clear enough: actual industries will seldom have thousands of independent firms, and never will the entrepreneurs have complete knowledge. It is not useful to characterize all of these industries as imperfectly competitive, for some will be near-monopolies and others will have prices, outputs, and rates of progress that deviate in only minor respects from what perfectly competitive industries would experience. In particular, many industries do not depart sufficiently from perfect competition (which is, of course, unattainable) to create any need for antitrust actions or public regulation.

Workable competition has been a very popular concept since its formalization in 1940, but its serious ambiguity has not yet been reduced. How competitive an industry should be (using observable criteria we shall discuss below) to be workably competitive has never been settled. Indeed the criteria (prices, service, product innovation, rates of return) that deserve most weight in any application of the concept have not been agreed upon. Two competent persons who study a particular industry can disagree on its workable competitiveness, and there exists no analytical basis for eliminating the disagreement.

Monopolistic competition

The other leading concept, monopolistic competition, was formulated by E. H. Chamberlin (1933) and is directed to a different purpose. Chamberlin emphasized the diversity in the products of firms that are normally considered members of a single industry: they differ in details of quality, in repute, in locational convenience, in the religion of their producer, and a hundred other details that may influence their desirability to various buyers. He emphasized also the substitutability of products made by what are viewed as different industries: one may use aluminum or steel or wood to build a chair, and ostentatiously display one’s wealth with jewels, servants, or trips abroad. Each firm, in this view, has some elements of uniqueness (monopoly power) and yet many rivals, and the admixture gives rise to the title of the concept. The theory of monopolistic competition has led to a much more thorough examination of the problems of defining commodities and industries. It has not been found useful in the analysis of concrete economic problems.

Competitive equilibrium

The lack of conscious coordination of the behavior of individuals in a competitive market has led many writers to assert the impossibility of any stable equilibrium. Some have denied that any order is observable: the Continental cartel literature usually uses the word “chaotic” as a prefix to competition, and most proposals for an “orderly” policy assume that a competitive system is disorderly. Others have found cumulative tendencies in competition: for example, W. T. Thornton said that “if a single employer succeed in screwing down wages … his fellow-employers may have no alternative but to follow suit” ([1869] 1870, p. 105). Sidney and Beatrice Webb elaborated this view into their famous theory of “higgling in the market” ([1897] 1920, part 3, chapter 2).

Modern economic analysis, on the other hand, makes competitive equilibrium the central part of the theory of prices and allocation of resources. The presence of order and continuity in markets composed of many independently acting buyers and sellers has been established beyond serious question, on both theoretical and empirical grounds.

The main stumbling block in the layman’s acceptance of competitive equilibrium is the belief that many individuals acting independently will necessarily either undershoot or overshoot every appropriate change in output, prices, investment, and the like. If, for example, increasing demand calls for a 10 per cent increase in industry capacity, how can this precise total be achieved when a vast number of firms are individually and independently changing their plants in a hundred different proportions? In a sense this is a false question: no one can know that the following year’s demand will be exactly 10 per cent larger, and neither a public body nor a private monopolist can guarantee to have the “right” amount of capacity the next year. But let us put this complication aside.

The answer, then, is that there is much information available to guide the decisions of the numerous independent firms. In part this is current information: every trade is abreast of the investment decisions of its various firms, of the developments in products and production methods, and so on. This information comes from salesmen, trade journals, customers and suppliers, and a host of other sources. The firm is guided also by past behavior in the industry: if previous increases of output were supplied in some part by new firms, this becomes a factor in current decisions.

Empirical evidence of competition

A variety of statistical tests of the existence of competition have been proposed at various times, and at least three deserve some attention.

The presence of numerous firms, none dominant in size, is directly observable and is usually described by a low concentration ratio. The main difficulty with this structural test of competition is that the maximum concentration compatible with competition has not been determined, so the test is clear only when concentration is low. The problem is complicated by the fact that we have had no theoretical guide in summarizing the frequency distribution of firm sizes, which can, of course, be done in many ways.

Since a single price will rule under perfect competition, price homogeneity has often been proposed as a test of competition. We have already remarked that perfect knowledge is enough to ensure a single price, whether the market is competitive or monopolistic. Indeed, in a market of numerous sellers and buyers it is improbable that all prices in a given short interval of time will be uniform. It is improbable for two reinforcing reasons: the transactions will seldom be in completely homogeneous goods (quantity discounts, promptness of payment, and a dozen other characteristics vary almost infinitely among transactions); and the cost of learning market prices, given numerous traders, is such that complete information is not worth it. As a result, strict uniformity of prices has properly been viewed by the courts as a phenomenon more suggestive of collusion than of competition.

A related evidence of competition is more powerful: the absence of systematic price discrimination. If sellers are persistently obtaining higher net receipts (which need not be the same as prices) from some buyers than from others, we may be confident that they are acting in concert—a truly independent firm would concentrate its sales on the buyers who yielded higher net receipts.

A fourth, and perhaps the most traditional, test of the absence of competition is a high rate of return on investment. It has lost much popularity because of the difficulty of measuring profitability (in particular, the valuation of durable assets can conceal monopoly profits or create fictitiously high rates of return) and because an absence of high profits is compatible with various cartel arrangements. Yet it is true that unusually high or low rates of return will not persist for long periods in a competitive industry. More specifically, a recent study suggests that in unconcentrated manufacturing industries the rates of return of one year will provide no useful clue to the rates earned, say, five years hence (Stigler 1963, chapter 3).

Public policies

Laws, both statutory and common, have sought to protect competition for centuries. The Statute of Monopolies, which was passed in 1623 to restrain the crown’s use of grants of monopoly for revenue, was a famous example, as were the statutes (which Adam Smith compared in rationality to laws against witchcraft) against forestalling, engrossing, and regrating grain.

The Sherman Act of 1890 was pathbreaking, therefore, not in its prohibition of restraints of trade but in the implementation of this policy by an administrative force charged with ferreting out and prosecuting such acts. This most basic of all antimonopoly laws forbade not only conspiracies in restraint of trade but also attempts to monopolize—and in such broad terms as almost to defy conflicts of spirit and letter. Criminal penalties were supplemented by the incentive of triple damages to private parties who were injured by the forbidden acts.

The complaint that the Sherman Act came into force only after competitive markets had been destroyed (which was neither true nor wholly false), the belief that a group of specialists could deal with industrial problems more effectively than the judiciary, and the general impatience of reformers— all combined to bring about in 1914 the Clayton Act, which prohibited a set of practices that (it was believed) often led to monopoly, and the act creating the Federal Trade Commission to enforce the Clayton Act. With amendments—the most important being the Robinson-Patman Act of 1936 and the Celler-Kefauver Merger Act of 1950—the legislative basis of U.S. policy had been developed. This policy includes certain discordant anticompetitive elements (the Robinson-Patm an Act, with its goal of rigid uniformity of prices, and the legalization of resale price maintenance), as general policies have a habit of doing.

That this policy has contributed to the competitiveness of the U.S. economy is difficult to deny or to document. Yet international comparisons—in particular, of the same industry (often composed of the same firms) in Canada and the United States —suggest that the policy has had substantial effects. So, too, does the fact that the favorite practices of the formal cartel—a joint sales agency or division of customers—are quite uncommon in the United States.

The policy of restricting agreements among competitors (but not the policy of seeking to prevent monopolies) has spread to numerous other nations since its introduction in the United States. The most common form is to require registration of agreements among firms in an industry, and the subsequent approval or disapproval of the agreement by a specially constituted body. This is the practice of England, Germany, and several other nations, as well as of the European Common Market.

George J. Stigler

BIBLIOGRAPHY

Chamberlin, Edward H. (1933) 1956 The Theory of Monopolistic Competition: A Re-orientation of the Theory of Value. 7th ed. Harvard Economic Studies, Vol. 38. Cambridge, Mass.: Harvard Univ. Press.

Clark, John M. 1940 Toward a Concept of Workable Competition. American Economic Review 30:241–256.

Edgeworth, Francis Y. (1881) 1953 Mathematical Psychics: An Essay on the Application of Mathematics to the Moral Sciences. New York: Kelley.

Knight, Frank H. (1921) 1933 Risk, Uncertainty and Profit. London School of Economics and Political Science Series of Reprints of Scarce Tracts in Economic and Political Science, No. 16. London School of Economics; New York: Kelley.

Marshall, Alfred (1890) 1920 Principles of Economics. 8th ed. New York: Macmillan.

Stigler, George J. 1957 Perfect Competition, Historically Contemplated. Journal of Political Economy 65: 1–17.

Stigler, George J. 1963 Capital and Rates of Return in Manufacturing Industries. A study of the National Bureau of Economic Research. Princeton Univ. Press.