As we noted in Chapter 15, there are two acid-test criteria for distinguishing the applicability of monopolistic and oligopolistic competition models. If competitors are essentially oblivious of each other's identities, and if profits tend to be dissipated due to entry of new firms into the market, then monopolistic competition is almost certainly the appropriate model to apply. On the other hand, if the competitors are conscious of each other's identities to the point of devising market strategies oriented toward specific competitors, this is a sure sign of oligopolistic competition.
The number of competitors was not stressed in Chapter 15 as a critical distinction between monopolistic competition and oligopoly due to the importance of the relevant geographic market. We noted that in any Standard Metropolitan Statistical Area (SMSA) there are literally hundreds of retail gasoline stations that are differentiated from each other by brand name and locale. If the owner or manager of each station sets prices vis-a-vis the prices of those other stations within sight up and down the street, the relevant geographic market likely consists of a half-dozen or fewer stations in the near neighborhood. This makes retail gasoline distribution essentially oligopolistic. The critical criterion is not number of sellers, but rather consciousness of identities of competitors. Experience indicates that the majority of real-world competitive contexts are more appropriately analyzed with oligopolistic models than with monopolistically-competitive models.
(1) The oligopolistic market consists of a relatively small number of firms, but as we noted in the previous section, the number of firms in the market is not the most critical criterion for distinguishing oligopoly from monopolistic competition. The lower limit of oligopoly is duopoly which is a market populated by only two firms.
(2) The products (goods or services) sold by the firms in the oligopolistic market may be homogeneous (examples might be asphalt once laid, concrete once poured, a grade of gasoline once in the tank) or differentiated (locale, billing practices, and levels of service distinguish even pavers, ready-mix concrete suppliers, and retail gasoline stations). But if differentiated, the products must be close-enough substitutes in use or function that prospective consumers give serious consideration to the alternative products before purchasing.
(3) Unlike monopolistic competition, in oligopoly there may occur significant differences in managerial abilities and organizational structures. Also, widely divergent technologies may be employed by different oligopolistic competitors to produce identical or close substitute products.
(4) Unlike monopolistic competition, firms in an oligopolistic market have differing levels of access to information, some of which may have been acquired by deliberate research. Some information is specific to each firm, particularly its own prices, production plans, and marketing strategies. This information is likely to be kept secret and protected so that it is not readily available to other firms. There are likely to be few enough firms competing in the market that the manager of each one can know the particular identities of each of the other competitors in the same market.
These descriptive characteristics lead to the following behavioral characteristics of oligopolsitic competition:
(a) As with pure monopoly, scale of operation, technological complexity, or governmental grant of exclusive position (by certification, franchise, patent, or trademark) may constitute effective barriers to entry of new competitors into the market. It may even be possible for existing competitors, by some form of concerted action, to erect entry barriers. However, the incentive for a single oligopolist to unilaterally erect entry barriers is likely to be slight because competitors can be free riders in enjoying the benefit of another firm's efforts.
(b) As in monopoly, exit from the market is always possible by failure and dissolution, but oligopolists wishing to exit a market may have another option not available to the monopolist, i.e., to combine with or dispose of assets to a competitor.
(c) Unlike monopolistic competition and pure competition, oligopolistic competition is hardly anonymous. Because there are few enough sellers for each to know the identities of the others, virtually every market-oriented and productive decision needs to take into account the range of possible reactions by competitors as well as the most likely reaction. And even if no deliberate decisions are under consideration by the management of the oligopolistically competitive firm, its manager needs to monitor the activities of the competition and be prepared to react accordingly.
(d) Oligopolists, like monopolists, have a great deal more pricing discretion than do monopolistic competitors, but the pricing discretion may be severely constrained by the pricing practices of competitors and their likely pricing reactions.
(e) Because of the inherent or contrived barriers to entry into an oligopolistic market, supernormal profits may persist. Yet, price competition among oligopolists, as among monopolistic competitors, may tend to dissipate any supernormal profits even if no new entry occurs. Managers of oligopolistic firms may even deliberately initiate a price war that is intended to take profits to subnormal levels long enough to induce competitors to exit the market.
(f) As in both pure and monopolistic competition, there may tend to be a market-wide convergence upon a common price, but for different reasons in the oligopolistically competitive market. These may range from an effort to limit price to a profit submaximizing level to forestall entry into the market, through various forms of price leadership/followership behavior, to overt collusion among the oligopolistically competitive firms in the market. The oligopolist's demand curve likely will be of somewhat steeper downward slope than that of the monopolistic competitor, and it may account for some fraction of the total market demand that is greater or less than 1/n if there are n firms in the market. The steeper downward slope of the oligopolist's demand curve suggests greater pricing discretion, but this may be rather illusory for all of the firms in the market except the price leader.
(g) Difficulties with pricing strategies and price leadership/followership relationships in the oligopolistically competitive market may lead the managers to prefer non-price forms of competition that are similar to the efforts to differentiate and promote the product in monopolistic competition. But practically everything that can be said about price leadership/followership in oligopoly can be repeated for non-price competition ladership/followership. That is, there may occur design, service, or promotion competition, and a market leader may emerge in any of these areas. The leader then initiates aggressive decisions while the followers make defensive decisions of the "catch-up" or "me-too" variety.
The object of competition in oligopoly often becomes market share rather than profit or any other behavioral goal. A common problem for oligopolists engaging in non-price competition is to incur design, service, or promotion costs (fixed costs) that simply cancel out each other's efforts, leaving their market shares unchanged (the "smoking-more, enjoying-it-less" syndrome). To the extent that non-price competition has self-cancelling effects, profits are diminished. The market leader may have the best hope of gaining market share, and that may be only temporary.
For each of the other market structure types (i.e., pure competition, pure monopoly, and monopolistic competition), we were able to describe one fairly standard model generally accepted by economists. Unfortunately, this is not possible for oligopolistic competition because the circumstances and the potential for competitor reaction render each oligopolistic situation unique. Since no two oligopolistically competitive situations are alike, it is necessary to model each one to fit the specifics. There is no single oligopoly model as there is a single monopoly model. Oligopoly models are legion.
We can, however, identify a limited number of oligopolistic behavioral patterns into which nearly any oligopolistic situation can be classified. A general type of model can be specified for some of the patterns. We shall describe seven broad behavioral patterns:
(1) Oligopolistic competitors may choose to ignore each other in a "live-and-let-live" attitude, each pursuing its own goals. This may in fact be the reality for most of the trivial and day-to-day decisions made by the manager of a firm in oligopolistic competition. However, as we shall show in our discussion of price leadership/followership, choosing to ignore the competition's market-related decisions can have serious market-share or growth consequences.
(2) Managers of oligopolistically competitive firms may engage in aggressive competitive behavior, occasionally manifested by open price- (or design-, service-, promotion-) warfare or other predatory or even criminal behavior to the end of eliminating competitors so that monopoly (or more-limited oligopoly) position can be achieved.
(3) Oligopolistic managers may be so fearful of the possible deleterious effects of a ruinous price (or other kind of) war that they enter into extreme decision rigidity (e.g., price rigidity) in a "don't-rock-the-boat" or "don't-make-waves" attitude.
(4) Managers of firms in oligopolistic competition, to avoid both price rigidity and price warfare, may engage in overt (open, public) collusion (e.g., by forming a cartel). Collusion may be only occasional consultation, or it may be conducted on a continuing basis. Such overt collusion, if formalized by contract or treaty, constitutes a cartel. The monopoly model described in Chapter 13 is adequate to the analysis of the behavior and identification of decision criteria for an effective cartel. However, collusive relationships and cartel arrangements are inherently unstable because of the difficulty of maintaining discipline among the members as to agreed-upon decisions or policies. They are prone to cheating by members who engage in "under-the-table" transactions. It has been said that if a near-universal cartel exists, the best position to be in is outside the cartel in order to be able to subvert the cartel's prices, promotions, etc.
(5) If overt collusion among separate firms is frowned upon by the authorities, the same end may be accomplished by combination (acquisition, merger, forming "trusts") among the competitors to achieve monopoly position.
(6) If combination to achieve monopoly is frowned upon as well by the authorities, then competitors may try to engage in implicit or covert collusion with no apparent agreement or even contact among themselves. Such so-called "conscious parallelism of action" may be effected through the use of common rate schedules, transportation basing points, catalogs or price books, or trade-association reporting schemes. No one talks to anyone else, but everyone knows that each of the others is following a common policy. In the latter case, although the information reported is historic, there may emerge a consensus about how near-future prices may be predicated upon recent past prices.
(7) And finally, even if none of the above occurs, some type of price leadership/followership is likely to emerge. This author is of the opinion that the existence and effective administration of antitrust law inevitably causes price leadership/followership relations to become the most common patterns of oligopolistically-competitive interaction in Western industrialized nations. This is not to imply that price leadership/followership is the pattern of choice among oligopolistic competitors. Left to their own devices (i.e., without constraint or interference from governmental authority), oligopolistic competitors would choose not to compete, but rather would collude, cartelize, or combine. If these avenues to monopolistic position are foreclosed by effective antitrust administration, price leadership/followership will emerge by default. Then there remains the residual question for government authority of whether leadership/followership relationships are deleterious to the public interest, and if so, what is to be done about them. This and other questions should be of extreme interest for managers of oligopolistically competitive firms.
Price Warfare. The earliest of what might be considered oligopoly models were described by Augustin Cournot, Joseph Bertrand, and F. Y. Edgeworth during the nineteenth century. The Edgeworth variant described a simple duopoly whose managers are incredibly naive, but the model can serve to illustrate the effects of price warfare. A model similar to Edgeworth's is illustrated in Figure 16-1. The manager of Firm A, the first to set up shop, naturally assumes himself to be a monopolist who can exploit the entire market demand. He builds a plant that by assumption has constant variable costs represented by MC=ATC, and with plant capacity Q1, the output at which monopoly profits would be maximized when sold at price P1.
The second seller, Firm B, enters the market. Its manager realizes that Firm A is already in operation, and naively assumes that Firm A will continue to sell at price P1. Firm B builds a plant also with capacity Q1; the manager takes as its demand curve D1, which is half of the total market demand. The Db demand curve is represented in Figure 16-1 as a Janus curve facing left (i.e., output increases from right to left), which backs on the Dm demand curve at its vertical axis. The manager of Firm B, however, realizes that if he will undercut price P1 by a small margin, say at price P2, he can sell his whole output, invading the market of Firm A by line segment rs.
Figure 16-1. A Duopoly Model of Price Warfare.
Meanwhile, the manager of Firm A has become aware of the presence of Firm B, accepts it as a fellow duopolist, and now takes Firm A's demand curve to be Da, which also is half of the market demand. The manager of Firm A, who now makes the naive assumption that Firm B will continue to sell at price P2, cuts price to P3 and dumps Firm A's entire capacity output on the market, invading Firm B's market by amount tu.
It is easy to see that as long as each manager naively makes the assumption that the other firm will continue to sell at its current price, the competitive price cutting will continue through several more iterations until the price charged by one of the firms, B in Figure 16-1, reaches the level of its average total cost. The manager of Firm A notices (with sudden and unexplained astuteness) that Firm B has done its worst; it is dumping its entire capacity output on the market at a price just equal to ATC. It dare not go to a lower price, else it will suffer losses. So the manager of Firm A reasons that there is nothing to lose by returning to its profit-maximizing price, P1, and to produce and sell Q2. And the whole price war starts all over from P1 back down to P6.
The moral of this story is that if the managers of oligopolistically competitive firms are so naive as to assume that competitors will continue to charge the same price forever more, and if they are incapable of learning from experience, they will get into price wars where price oscillates between the preferred profit maximizing price and the level of ATC.
Real-world business firm managers are neither so naive nor incapable of learning from experience. As intelligent and perceptive decision makers, they are unlikely to lapse into such mindless competition. However, as noted earlier in this chapter, if the manager of one oligopolistically competitive firm has a cost advantage or a greater financial capacity relative to competitors, a price war may be initiated with the intention of driving price down below the AVC of the competitors, thereby to induce them to shut down in the short run, and to exit the market if the low price continues long enough.
Oligopolistic Price Warfare. In Figure 16-2, suppose that two duopolists equally share the market demand so that their demand curves are coincident and each equal to 1/2 of market demand, Dm. Firm B, however, has lower costs, represented by ATCb, AVCb, and MCb. Firm A's costs are higher at ATCa, AVCa, and MCa. Firm A would like to maximize profit at P2, while Firm B prefers the lower price, P1, for maximization of its profits. But if the manager of Firm B knows enough about both the firm's own costs and those of Firm A, he or she will realize that price can be taken as low as P4 before incurring a loss greater than average fixed costs. Price P3 is at the minimum point of Firm A's AVC curve. Firm A's options now are either to meet price P4 and go out of business in the long run (because P4 is less than its AVC), or to stay at Price P2 and lose market share as its demand curve shifts leftward far enough for it to go out of business because of declining market demand for its product. Thus, by pursuing a deliberate profit-nonmaximizing strategy in the short run, Firm B may be able to achieve monopoly position that will allow it to maximize profits in the long run. But this sort of aggressive price cutting behavior is likely to be regarded as predatory by antitrust authority.
Figure 16-2. Price warfare in a duopoly market.
Extreme Price Rigidity. During the late 1930s the prices in certain oligopolistic industries, notably tobacco products, were observed to be constant for years at a time. During the late 1930s economist Paul Sweezy proposed an oligopolistic model to explain extreme price rigidity among competitors ("Demand under Conditions of Oligopoly," Journal of Political Economy, vol. XLVII (1939), pp. 568-73). Sweezy reasoned that competitors are likely to react asymmetrically with respect to price increases and price decreases initiated by one of the firms in the market. Sweezy implicitly assumed recession economic conditions (the effects of assuming an expanding economy are explored in appendix to this chapter). Given these conditions, competitors are far less likely to follow a price increase than a price decrease. Sweezy then reasoned that demand would be relatively more elastic above the current price, but relatively more inelastic below the current price. This implied that the oligopolistic competitor's demand curve is bent or kinked at the price as illustrated in Figure 16-3.
Figure 16-3. Sweezy's Kinked Demand Curve.
If the demand curve for the oligopolist's product really is kinked at point A in Figure 16-3, then a price increase to P2 will result in a larger percentage decrease of quantity demanded (to Q2) than the percentage increase (to Q3) that will result if price is cut to P3. Assuming that demand is elastic above the kink, but inelastic below the kink, this alone will provide a revenue disincentive for the manager to change price from P1. It may be recalled from Chapter 7 that a price increase when demand is elastic will reduce total revenue; and a price cut when demand is inelastic will also reduce total revenue. The manager is in a lose-lose situation. There is no price above or below P1 at which the firm can increase its revenue. Hence, price remains rigidly at P1. A second possible explanation of extreme price rigidity lies in what is reputed to be a gap or vertical segment in the marginal revenue curve below the kink. This hypothesis is also examined in the appendix to this chapter.
There are several problems with the so-called kinked-demand hypothesis. One is that it provides no explanation of how any actual price, like P1, is established in the first place. Second, when price is increased but competitors do not follow, customers can be expected to defect from the firm's product to those of its competitors. This is an example of an unsuccessful effort at price leadership, and it constitutes a change of one of the determinants of demand, i.e., the cohort of the consumers purchasing the firm's product. The consequence of a change of such a non-price determinant of demand is a demand curve shift (in this example, to the left). The movement from point A to point B is really not along the same demand curve; rather, the firm's genus demand curve (a concept borrowed from Chamberlin's monopolistic competition model) shifts to the left as illustrated in Figure 16-4. Point B is on a relocated genus demand curve, D'.
Figure 16-4. Asymmetrical Price Leadership for Price Increases and Decreases.
The path from A to B is in this example a contraction path. The demand curve is not kinked; rather, it has simply shifted to create an identification problem (as described in Chapter 8). Point C is on the original demand curve, D, and is reached when the firm cuts its price to P3, and all of its competitors follow suit (in this case, the firm is a successful price leader). All of the firms in the market sell some more of the product at the lower price, but there is no realignment of customers with sellers, so the demand curves do not shift.
Leland Yeager has suggested that there is only one actual point, like A, on any demand curve ("Methodenstreit over Demand Curves," Journal of Political Economy, vol. LXVIII (February 1960), pp. 53-64); all others are "virtual" in the sense that they may be revealed under other circumstances (different prices). So it is not possible to be confident that a demand curve has any particular shape, whether straight, curved, or kinked, away from the single existing point.
Even if a demand curve is not truly kinked as reputed in the Sweezy model, asymmetrical responses of competitors to a firm's increases and decreases of price could result in the theorized rigidity. If the firm cuts price when demand is inelastic and other firms follow the price cut, revenues will decrease because of the demand inelasticity. If the firm raises price and other firms do not follow, the firm's revenues will decrease because its demand curve shifts left (but not because its demand is elastic). Although economists have looked very hard for empirical evidence of kinks in demand curves, virtually none has been reported in the professional literature over the past half century. This implies either that the kinked demand curve is a relatively rare circumstance, or that economists have been looking for the wrong thing (kinks rather than shifts).
Price-Leadership/Followership. We have made a case that oligopolistic competition is the most prevalent form of commercial organization in Western society; and we have asserted the most common pattern of oligopolistic interaction where antitrust laws are vigorously enforced to be price leadership/-followership. But we recognize that each price leadership circumstance is unique and demands its own model for analysis. All that we can do in this section is to select a few of the more prevalent types of price leadership to model as guides for the reader to use in encounters with price leadership circumstances.
Economists have identified four broad categories of price leadership:
(a) Asymmetrical price leadership occurs if the firm is successful in one direction of price change, but not in the other; the kinked demand model is reputed to be an example.
(b) Barometric price leadership is where the manager of one of the oligopolistic firms establishes a reputation for perceptiveness and sensitivity to changing market conditions, and a record of making timely and successful adjustments to those perceived changes. Managements of other firms then watch the price leader's activities and attempt to emulate his decisions.
(c) Dominant firm price leadership is where the market consists of a dominant firm surrounded by a competitive fringe of smaller firms. The dominant firm behaves as a benevolent monopolist, tolerating the existence of the smaller firms and allowing them to sell any amount of the product that they wish at the price that the dominant firm prefers. The dominant firm then takes its demand as the residual of the market demand not met by the competitive fringe firms, and proceeds to behave as a pure monopolist in maximizing profits. There is no doubt that dominant-firm price leadership examples (in the automotive and computer industries, to name but two) can be found, but we venture the guess that they come into being only as a consequence of vigorous antitrust enforcement that constrains the predatory tendencies of the dominant firm. In Figure 16-5, the market demand is Dm. The competitive supply, Sc, is the sum of the marginal cost curves of the competitive fringe firms. The locus of the dominant firm's demand curve, Dd, is found by subtracting the competitive supply from Dm at each possible price. The dominant firm then sets price at P1 to maximize its profits by selling output Q1, while the competitive firms behave as purely competitive price takers to sell quantity (Q2-Q1). This form of oligopolistic market organization is quite workable, and can persist as long as antitrust law is vigorously enforced and the dominant firm behaves itself. The dominance of the dominant firm may break down when one or more of the competitive firms begins price experimentation or product differentiation/promotion in the effort to capture a larger share of the market.
Figure 16-5. A Model of Dominant-Firm Price Leadership.
(d) Differential characteristics price leadership may be based on three aspects of the constituent firms' characteristics:
Combinations of these differences may also be bases for price leadership.
(1) differences in per-unit costs;
(2) differences in sizes of plant; or
(3) differences in market shares.
Price Leadership based on Cost Differences. Suppose, in Figure 16-6, that there are two firms in a market, and that, as illustrated in panel (a), they initially share the market demand equally, i.e., the demand curves are coincident at Df, each of which is one-half of market demand. Firm B has a cost advantage (it hires labor or buys material inputs, components, or energy in lower-cost resource markets) than does Firm A. In order to maximize its profits, Firm B would prefer the lower price, PB, at which it sells quantity QB, than that preferred by Firm A, PA, at which it sells the smaller quantity QA in order to maximize its profits. Which firm has the potential for exercising price leadership?
Figure 16-6. Price Leadership based on Cost Differences.
If Firm A chooses to charge its preferred price, PA, ignoring Firm B's preferred lower price, PB, some of Firm A's customers will defect to purchase from Firm B. This constitutes a change of a non-price determinant of demand for both firms, i.e., the population of consumers purchasing from each firm. Firm A's demand curve will shift to the left toward position DA as illustrated in panel (b), carrying with it its marginal revenue curve toward position MRA, with the consequence that Firm A will prefer an ever-lower price. Firm B's demand curve will shift to the right toward position DB in panel (b), carrying with it its marginal revenue curve toward position MRB, with the consequence that it will prefer an ever-higher price. Theoretically, these shifts will continue until the preferred prices converge to a common price, PC, but with a significant difference: the two firms now have divergent market shares, Firm B now with a larger share than Firm A. Firm A will sell an even smaller quantity, QA', and Firm B will sell an even larger quantity, QB'.
Alternately, had the manager of Firm A been willing to meet Firm B's preferred lower price (a deliberate profit sub-maximizing strategy in the short run), it could have preserved its share of the market. Thus, the firm naturally preferring the lower price (in this case, Firm B) has the potential to be the price leader when demand or cost circumstances change. The other firm(s) may choose to follow or not; they can either go ahead and meet the leader's preferred lower price (and thereby preserve market share), or they can lose market share and end up preferring the same price as the leader.
Price Leadership based on Plant Size Differences. The same phenomenon can be seen where there is no cost or demand difference between firms, but there is a difference in plant sizes. In Figure 16-7, the two firms again have equal initial market shares. They also use the same technology and have access to the same labor and materials markets, or different markets with the same market prices. The evidence of this is that their ATC curves have the same shapes and reach bottom at the same per-unit cost levels. Firm B, however, has a slightly larger plant evidenced by the overlap of its ATC curve to the right of Firm A's ATC curve. Again, Firm B has the potential for price leadership because it naturally prefers a lower price in order to maximize profits. But in this case, the basis for price leadership lies in Firm B's larger plant.
Figure 16-7. Price Leadership based on Plant Size Differences.
Price Leadership based on Unequal Market Shares. Finally, we suppose in Figure 16-8 that the two firms have identical plants and hire labor and purchase materials from the same resource markets. These conditions are evidenced by the fact their ATC curves are coincident. Firm B, however, initially has a slightly larger market share than does Firm A. In this case, Firm A prefers the lower price in order to maximize its profits, and thereby has the potential for price leadership. In this example, the basis for price leadership lies in the smaller initial market share (we can only imagine that Firm A's sales force is pushed to be more aggressive, "We're Number 2, we try harder!"). If Firm B now does not meet the lower price preferred by Firm A, it (Firm B) will lose market share to Firm A. This process theoretically could continue until they have the same market shares and thus prefer the same market price.
Figure 16-8. Price Leadership based on Unequal Market Shares.
While the potential for price leadership can readily be discerned in each of these models, it does not follow that the actual price leader will coincide with the potential price leader. The theoretical follower could "take the bull by the horns" and undercut the price preferred by the theoretical leader in each example. But by so-doing he risks the initiation of a price war as described earlier.
Competition, if not subverted by collusion, predatory behavior, the erection of barriers to entry, or the grant of exclusive position by government, can serve as a mechanism for the social control of industry and commerce. Competition doesn't have to be pure in order to exert its controlling function. Even the competition present in the oligopolistically-competitive market in the form of price leadership/followership can be an effective vehicle of social control because the threat of market share loss will induce competitors to meet the lowest price preferred by any firm in the market. Society's job in formulating public policy with respect to oligopoly then is to suppress collusive behavior that exploits consumers, prevent predatory behavior that destroys existing competitors, and disallow the creation of entry barriers that diverts potential competition.