Oligopolistic Interdepence

1. Introduction

An industry is oligopolistic when so large a share of its total output is in the hands of so few relatively large firms that a change in the output of any one of these firms will discernably affect the market price”. In 1516, Sir Thomas Moore, who first coined the term “oligopoly” in his Utopia already noted that prices need not fall to competitive levels simply due to the presence of more than one supplier. Market power, defined as “the ability of a firm or group of firms to raise prices, through the restriction of output, and maintain them for a significant period of time above the level that would prevail under the competitive conditions and thereby to enjoy increased profits from the action”, can be exerted anti- competitively by dominant firms or by multiple firms which, although not individually dominant, form a cartel. Modern western antitrust laws however forbid either way. The main problem of oligopolies is that they can, apparently, collectively exert market power without having to turn to an explicit agreement. This phenomenon has intrigued economists and lawyers in the past, and still does intrigue them nowadays. To get an insight into the “oligopoly problem”, in this paper the main economic views concerning oligopolistic markets will be set out. 

Oligopoly is a market structure where there are a few firms, large in size, accounting for most of the production. All of these firms, generally, can enjoy substantial profit in the long-run due to entry barriers. That means, due to large size of existing firm or heavy initial investment, entry of new firms, in the oligopolistic markets, is restricted. Examples of oligopolistic industries may include automobiles, heavy machinery & equipment, power generation, Steel, petrochemicals, etc.

In any other type of market, each firm could take either price or market demand as given and largely ignored its competitions. In oligopolistic market however, it is necessary to consider the behaviour of competition while determining the output or price. Similarly, the competition will also base his decision on the behaviour of the first firm. In other words, all the firms in oligopoly are interdependent. This makes their pricing strategies to be different from other types of market structure. Following is a brief analysis of some of those models.

1.1 Neo-classical economic theory

According to neo-classical economic theory, consumer welfare is maximised when a market is perfectly competitive. Already in 1776 Adam Smith, in his secular work “Wealth of Nations”, proclaimed that the forces of competition, driven by private, self-interested behaviour, referred to by him as the “The Invisible Hand”, in general generate a socially desirable result. The concept of “perfect competition”, which was introduced by Marshall in 1890 in his work “Principles of Economics”, can be seen as a modern version of Smith’s Invisible Hand.

1.2 Paradigm of perfect competition

In a perfectly competitive market - i.e. a market with a large number of sellers and buyers, who each sell or buy a relatively small quantity of homogeneous products, free entry into and exit out of the market and perfect information - no one competitor possesses any market power, and prices will equal marginal costs. The effects of a market being perfectly competitive are that both productive efficiency and allocative efficiency are maximised. The result is that both consumer welfare and producer welfare are maximised and could not be improved upon even by an omniscient regulator. If the assumptions hold true, competition is held to regulate itself. It is also being claimed by economists that perfect competition will enhance dynamic efficiency by stimulating innovation, as competitors strive to produce new and better products for consumers.

1.3 Paradigm of monopoly

In the other extreme benchmark model, i.e. the monopolistic market, the monopolist possesses maximum market power, having full control over the price and output level. A monopolist will set his output, in order to maximise his profits, at the level where the marginal revenue he draws from selling an extra product will equal its marginal costs, resulting in a a suppressed output and prices raised above the competitive level. One effect of a monopolistic market is that it causes allocative inefficiency, as the resources of both consumers and producers could have been applied more efficiently, leading both to consumer and producer welfare loss. The loss in value to consumers who at the competitive price would buy the product, but who at the monopoly price are deflected to ‘inferior’ substitutes is referred to by economists as “deadweight loss”. Consumer and producer welfare loss taken together constitutes the social welfare loss. Posner has pointed out that welfare losses from monopoly is even larger than just the consumer and producer welfare loss taken together, as it does not take into account the costs resulting from rent-seeking activities, i.e. the costs of competition in order to obtain a monopoly position. As to the productive efficiency, the absence of competitive pressure can give rise to slack and inefficiency while enjoying a ‘quiet life’, known as X- inefficiency or managerial inefficiency, which leads to further welfare loss.

Another question, however, is whether perfect competition does in fact maximise consumer welfare in the long run. Various economists have argued that another type of efficiency, i.e. dynamic efficiency, cannot be maximised under perfect competition. Dynamic efficiency embraces activities of innovation, research and development, which can, in the long run, lead to better products and possibly lower prices, thereby increasing consumer welfare. Schumpeter, for example, argued that short-term monopolies are the best market form for dynamic efficiency, as they have easiest access to outside financing. The “perennial gale of creative destruction” will, however, protect the public interest. Some authors even argue that an oligopolistic market form optimises the possibility for innovation.

2. Theories of oligopolies

In conventional economic theory, therefore, perfect competition is held to maximise consumer welfare. The assumptions upon which the models rest cannot, however, be found in the real world. Several economists questioned and criticised the model of perfect competition  as a realistic benchmark. Chamberlin and Robinson have contemporarily, albeit independently, studied the imperfectly competitive markets, developing upon an earlier study by Sraffa. Clark in his work “Towards a Concept of Workable Competition”, observed that perfect competition “does not and cannot exist and has presumably never existed”. In his article he coined the term “workable competition”, being the “the most desirable form of competition, selected from those that are practically possible, within the limits set by conditions we cannot escape”. He argues that if one of the conditions essential to perfect competition is lacking, then it no longer follows that we are necessarily better off for the presence of the rest of the conditions. In Clark’s concept of workable competition, welfare may increase by the addition of  imperfect conditions to an imperfect market situation. If perfect competition is not achieved in many sectors of the industry, the one possible response is the attempt to identify “workable competition”, being the “second best” solution. Another theory worth mentioning is that of the “contestable markets”. According to the developers of the theory, Baumol, Panzar and Willig, firms will be forced to ensure optimal allocation of resources provided that the market is perfectly contestable, i.e. entry and exit are costless and immediate, which will have as a result that prices will be set at the competitive level. Such a market is subject to “hit-and-run entry”, i.e. firms enter the market, take short- term advantage of the profit opportunities, and leave without costs.

3. Economics of oligopoly

As stated above, the danger of oligopoly lies in the fact that they can apparently collectively exert market power without having to turn to a formal cartel agreement. The difference between the two benchmark models of perfect competition and monopoly, set out above, and oligopolies is that in an oligopolistic market competitors individually interact with each other, whereas in a perfectly competitive market competitors interact with the market as a whole, while monopolists, by definition, are not subject to any interaction. Oligopolists, in their choice of behaviour in the market, will necessarily have to take into consideration the behaviour of their competitors. Although oligopoly fits conceptually between the extremes of perfect competition and monopoly, it cannot be studied by these static models. The study of oligopoly requires instruments that take into account the interaction between the competitors. This is provided for by game theory, developed by the economic current of New Industrial Organisation. Nevertheless, it is useful to first get an insight into static models of oligopolistic competition, as they remain influential, the contemporary view on oligopolies (and are in any case consistent with game theory).

3.1  Static models of Oligopoly

3.1.1 Cournot

The most famous model, and up to date most used as the benchmark model of oligopoly, stems from the French mathematician Augustin Cournot, in 1838. He assumes that oligopolists compete by means of setting output. In his model, only two suppliers are present in the market, firm A and firm B, who can set their quantities only once. Both firms assume their rival’s output to be fixed. For every level of output, set by firm B, it is assumed that there is a unique output to be set by firm A to maximise its profits, and vice versa; the price level is then determined by the point where the two output curves meet, called the ‘Cournot equilibrium’. The outcome of the application of the model is that the price set for a product is higher than the one which would be set in a perfectly competitive market. However, the higher the number of firms present in the market, the more the overall output will increase, and prices will thus move towards the competitive level, i.e. marginal costs. The price level, however, will always remain higher than the level in a perfect competitive market, where competitors are not influenced by interaction between them.

The model of pricing put forward by Augustin Cournot in 1838 is a duopoly model (existence of only two firms in the market). Cournot’s model assumes that there are only two firms in the market – A and B – each one producing mineral water at zero cost (This is because each of the firms is assumed to be owning a spring of mineral water). In other words, the model is based on following assumptions.

1) There are two firms – A and B.

2) They are operating with zero cost.

3) They are producing identical product.

4) They decide their own output on the assumption that the competition will not change his output level.

5) Firm A starts producing first.

Based on these assumption, the duopoly firms will operate in the market as shown in the following diagram.

As per the diagram 1.1, DD₁ demand curve MR_A market and MR_B – marginal revenue curves of firm A & B, respectively. Since the operating cost of firm A is zero, the diagram does not have cost curve. Firm A will produce at that point where the profits are maximized. (The students can recollect the equilibrium condition – MC = MR)

A – Level of output for firm A where MC (which is equal to zero in this model) = MR_A. OP – equilibrium price for Firm A. CE₁ – Firm B’s demand curve (under the assumption that the competition will not change his output level. In this case firm A is assumed to keep its output fixed at OA). B- equilibrium point where MC (which is equal to zero) = MRB AB – profit maximizing output for firm B OP₁ - profit maximizing price for firm B As can be seen from the diagram, Firm B produces half of market not supplied by Firm B – that means it produces ¼ th of the market. 

In the next period, Firm A will produce ½ of the market not supplied by B (under the assumption that the competition will not change the output level). So ‘A’ will produce ? th of total market.

Now Firm B will react by producing ½ of the remaining market which is 5 /16

According to Cournot, this kind of action – reaction pattern of firms will continue till they cover 2/3 rd of the total market. Equilibrium of Cournot’s model is explained below.

The output of firm A goes on declining and by solving this geometric series what is obtained is 1/3 of the market is supplied by firm A. In the similar way, firm B also supplies 1/3^rd  of the market. Both the firms together supply

2/3^rd  of the market. 

Important observations from the Cournot Model.

1) Cournot solution is stable.

2) More the number of firms in the market, more will be the quantity supplied and less will be the price. (It can be shown that if 3 firms exist in the market, 3/4 the of the market demand will be supplied).

3) Since the firms do not recognize the interdependence, they can not act as monopolist.

4) Each firm maximizes its profit in each period but industry’s profits are not maximized.

Criticism of Cournot’s Model : Cournot’s duopoly model, expressing the limiting case of Oligopoly, is criticized on many grounds.

Following are some of the important ones : -

a) Assumption of costless production is highly unrealistic.

b) It is a ‘closed’ model where there is no entry for new firms.

c) In each successive period, price is brought down by the action-reaction pattern of two firms.

Conclusion :

Cournot and Stackelberg models are two different approaches to oligopolistic market. for those industries where all the firms have more or less similar market share and none of then has leadership position, Cournot’s model may be applicable. Whereas, for those industries like mainframe computers (where IBM is the leading firm) Stackelberg’s model may be more appropriate.

3.1.2 Bertrand

The assumption that firms compete by setting their outputs instead of setting prices does not, however, seem to reflect reality. 45 years later, in 1883, Joseph Betrand reviewed Cournot’s model, assuming that oligopolists compete by means of setting their prices. Again, firms have only one opportunity to set their prices. Suppose again that there are only two firms present in the market. For any price set by firm A, firm B can slightly undercut this price, detracting the whole market demand, and again vice versa. The limit is however reached if one of the firms charges prices that equal marginal costs, meaning that undercutting this price is not profitable. The final outcome of this model is thus that the price in a market with oligopolistic competition is similar to that in a market with perfect competition, i.e. equal to marginal costs. This result remains unchanged when the number of firms present in the market increases. While this model seems to be based on a more realistic assumption, i.e. that oligopolists compete on prices, it is however subject to many assumptions, which in reality do not hold true: it is, for example, implied that goods are perfectly substitutable, marginal costs are constant and equal for both firms and that firms are not subject to capacity restraints, with the result that for now Cournot’s model remains the workhorse oligopoly model.

Bertrand developed the model in 1883. The model is applicable for the firms which produce homogenous product and make their pricing & output decisions at the same time. The model differs from that by Cournot on the ground that the firms compete on the basis of price and not on the basis of quantities (as was the case in Cournot model). Another important assumption on which the model is based is that there are only two firms competing in the market.

By using the same tool & demand equations introduced in the earlier section, we will show, how the two firms determine their equilibrium, by choosing prices instead of quantities. Suppose the market demand curve is P = 30 - Q.

As we are aware, Q = Q_A + Q_B i.e. quantity produced by firm A and firm B.

Suppose the marginal cost of production for both the fimrs is Rs. 3 (instead of zero in the earlier models), MC_A= MC_B = 3. By solving this equation for Cournot’s equilibrium, (as per the earlier section) we get following results.

QA = QB = 9

Price = 12

Profit = 8 for each firm.

As per Bertrand, the firms will compete on the basis of price. Since they produce homogenor’s products, the consumer will buy from the firm selling the product at the lowest price.

So there will be 3 possibilities in the market.

1) A firm which charges higher price than its competition will have no market share.

2) A Lower-price firm will capture entire market.

3) If both the firms charge same price, the consumers will be indifferent.

That means both the firms will supply half of the market each.

Now, if the P = MC_A = MC_B = 3, then Q = 27 and each firm will supply 13.5 units of the product. (Substitute P = 3 in the above equation and equally divide Q into both the firms).

Since P = MC, each firms makes only normal profit (or zero economic profit). Even though both the firms make zero profit, they will have no tendency to increase the price. This is because, if one of the firms raises the price, it will loose entire market. On the other hand, if one of the firms reduces the price below Rs. 3/- it will capture the market, but its profits will become negative (as it is already facing zero profit situations). Suppose both the firms decide to raise the price to say Rs. 6. In this case, there may not be a stable situation. Either of the firms may use the strategy of slightly reducing the price and capturing the entire market. Each competition may under cut the price till both reach back to P = Rs. 3. We can conclude this section by commenting that a change in strategic variable firm quantity to price, a drastic change is evident in the market outcomes. 

3.1.3 Stackelberg

Stackelberg has developed a model in response to Cournot’s oligopoly model. While in the Cournot model each firm takes the other’s action as given, in the Stackelberg model, one firm takes the other firm’s reactions as given. In his model he analyses whether it pays more for a firm to either be a follower or a leader in the setting of output. If one firm decides to be a follower and the other a leader, the outcome is called a ‘Stackelberg equilibrium’. If each chooses to be a follower, and each firm expects the other firm to be a follower, the Cournot equilibrium will be the outcome. If both firms, however, choose to be a leader, the outcome is indeterminate; this situation is called Stackelberg warfare.

This model was developed by a German economist Stackelberg. In the Cournot’s model it was assumed that both the competitions make their output and price decisions at the same time. Situation will be different if one of them moves first. Stackelberg presented a duopoly model in which one of the two firms sets its output before the other firm.

Let us first understand the concept of linear demand curve with the help of numerical example.

Suppose duopolists face following market dd curve.

P = 30 - Q.

Q = Total output (Q = Q_A + Q_B)

A & B - two duopoly firms.

Suppose both the firms have zero marginal cost. MC_A = MC_B= 0.

To maximize profit, firm A sets MR = MC. So its total revenue TR_A is given by 

TR_A = P. Q_A = (30 - Q) Q_A

 (Recall that total revenue is obtained by multiplying price by Quantity). = 30Q_A - (Q_A + Q_B)

Q_A (because Q = Q_A + Q_B) 

= 30Q_A - (Q_A + Q_B) Q_A (because Q = Q_A + Q_B)

= 30Q_A -Q_A² -Q_B Q_A

Marginal Revenue of firm a is additional revenue resulting from additional output so

The equilibrium condition is setting MR_A = MC_A which is equal to zero. So Firm A’s output is 

Doing similar calculations for firm B would give B’s output curve as

Equilibrium will be that levels of output Q_A & Q_D at which the two output curves intersect. (That is, that level of output which one gets by solving equations (1) and (2)).Q_A = Q_B = 10 Total quantity produced in the market is 20. So equilibrium price is P = 30 - Q = 10. Each firm’s profit is P × Q =10 × 10 = 100.

Using the above-explained numerical example, we will try to understand which firm benefits more in a situation analysed by stackelberg’s model and how the output levels of each firm will be determined. Suppose Firm A sets the output first and in setting its output it has to consider the reaction of firm B. Firm B decides its output level after firm B and it takes Firm A’s output as fixed 

Firm 1 will choose its output at that level where MR = MC (which is equal to zero).

TR_A = PQ_A

30Q_A - Q_A² - Q_BQ_A 

 Equation (4) Because the revenue earned by firm A depends upon the output level of firm B (Q_B), firm A must anticipate what firm B would produce Frim B, on the other hand, will produce by taking firm A’s output as fixed. So by substituting equation (3) for Q_B in equation (4), we will get.

Its marginal revenue is - 

By soling equation (5) using MR_A = 0; we get the output level. 

Firm A = 15

Firm B = 7.5

That means firm A produces twice the level of firm B and hence enjoys the profit twice as much Firm B. stackelberg calls it as the advantage of first mover.

Important implications of Stackelberg’s model.

1) The first mover will announce higher output and to maximize the profit, the other producer has to acknowledge it and produce a lower level of output. Otherwise price will come down and both the firms will suffer.

2) Stackelberg model brings about the need for collusive agreement between the duopolists as they are mutually interdependent.

3.1.4 Edgeworth

Edgeworth argues that the oligopoly outcome is indeterminate, and that any equilibrium reached is not necessarily stable. Since each firm has an incentive to undercut its rival’s price when that price is high, but raise prices when the rival is pricing at marginal cost, Edgeworth suggested that the market would fail to settle down, and rather that prices would cycle between high and low values. This theoretical pricing pattern is known as an “Edgeworth cycle”.

3.1.5 Conjectural variations

Many studies have been done on oligopolistic markets and their outcome. The most common method of studying reactions in static homogeneous goods models is by the use of the concept of “conjectural variations”, as developed by Bowley in 1924.  A firm’s conjectural variation is defined as the response it ‘conjectures’ about rivals’ outputs if the initial firm alters its own output. A conjectural variation value of –1 reflects the fact that the first moving firm expects, when introducing and increase of output, that its rival will react actively and also increase its output. The first firm will thus find it unattractive to increase its output. A conjectural variation of +1, on the other hand, indicates that the first firm expects that its rival will passively accomodate, i.e., decrease its output as a reaction to an output increase of the first firm. An output increase is then expected to be profitable. A conjectural variation of 0 can be found in the Cournot model, i.e. the output-increasing firm expects his rival’s output to remain unchanged. 

3.1.6 Kinked demand curve

Another approach to oligopolistic behaviour is the well-known ‘kinked demand curve’ theory of oligopoly developed by Sweezy and Hall and Hitch. They suggest that oligopolists expect their rivals to match any price decrease, but not price increases. The ‘kink’ in the demand curve at the level of the prevailing price reflects the more elastic demand curve for price increases than for price decreases. Therefore, price reductions do not result in an increase of market share, and are thus unprofitable, while a price increase, on the other hand, results in a large loss of customers. Prices are, therefore, held to be ‘sticky’ on the upward. 

The origin of kinked demand curve can be traced in Chamberlin’s analysis. But he did not explicitly used this tool in his analysis. Hall and Hitch, in their article ‘Price Theory and Business Behaviour’ used the term kinked demand curve for explaining the price-stickiness in oligopolistic markets. It was Paul Sweezy, who for the first time, used kinked demand-curve as a tool for explaining equilibrium in the oligopolic market. In this part of the unit, we will try to focus on how oligopoly firms will attain equilibrium when the prices are sticky. It has been observed that under oligopolistic markets, price and quantity tend to remain inflexible. Kinked demand curve hypothesis is used to explain such a rigidity of prices. Under oligopoly without product differentiation, if a firm raises the price, it will loose all its customers. So this firm will have no tendency to change its price. Alternatively, firms without product differentiations may enter into formal or informal agreement and maintain price rigidity.

Following diagram explains kinked demand curve hypothesis.

DKD, is the demand curve showing a kink at point k which corresponds to existing price-level. DK is more elastic portion of demand curve. KD₁ is more inelastic portion of demand curve. Demand curve is said to be having a kink of point ‘K’ because each oligopolistic firm believes that though its rivals will not increase the price, if this firm increases the same, but they will certainly reduce the price, if this firm decides to do so. If a firm decides to reduce the price below prevailing price level (OP), the competitors will fear that their customers will start buying from that firm and their market share will go down, so the competitors follow price cut policy and hence the firm will not gain much in terms of market share.

As per the diagram (2.4), firm will reach inelastic part KD₁ of the demand curve. In case a firm decides to increase the price, the competitors may not follow the firm and hence the firm may loose a large part of its customers. In other words, if firm raises the price, it will reach at the DK part of demand curve which is highly elastic in nature. It is obvious from the above discussion that whether a firm reduces the price or increases the price, it will be a loser. So this firm has no inclination of changing the price. Price remains sticky or rigid at “OP”.

3.1.7 Different schools of thought

On how such a supra-competitive result is reached and maintained there are various schools of thought differ in opinion. Although the crude division these various schools of thoughts suggest have largely broken down under the influence of more modern economic analysis, there remains some value in the distinction, and it is still common to find them applied to either personalities or approaches. 

3.1.8 Structuralists

The Structuralist or Harvard school, basing itself on the Structure-Conduct- Performance paradigm, retains that when the Structure of the market is concentrated, the Conduct, i.e. way of competing, of the few competitors present in the market is interdependent, which leads to a Performance of decreased output and supra-competitive prices, leading to supra-competitive profits. The structuralists in their studies thus retain that concentration almost inevitably leads to supra-competitive pricing due to the adaptation of oligopolists to each other’s behaviour. However, the adequacy of many studies has, subsequently, been disputed by scholars of the Behaviouralists’ school and the soundness of  the results of empirical research confirming the link between market structure and profit levels have been questioned.

The Structuralists base themselves on the economics of oligopoly as set out by Edward Chamberlin in 1929, who argued that “[s]ince the result of a cut by any one is inevitably to decrease his own profits, no one will cut, and although the sellers are entirely independent, the equilibrium result is the same as though there were a monopolistic agreement between them”. Relying on the fact that each oligopolist is aware of this fact, price-cutting may not occur, and the market as a whole can maintain supra-competitive prices without the need for a formal or tacit agreement. This “theory of interdependence”, however, gives no explanation on how the supra-competitive pricing is reached.

3.1.9 Behaviouralists

The Behaviouralist or Chicago school, on the other hand, challenged the structural school’s view, that supra-competitive pricing in oligopolistic markets is virtually inevitable, as they blame supra-competitive pricing on behavioural factors. Their ideas are based on the economics of neo-classical price theory. Building upon Stigler's  seminal article “Theory of Oligopoly” of 1964, behaviouralists sustain that, in a concentrated market, as in an unconcentrated market, a consensus level of pricing needs to be reached, the adherence thereto monitored, and secret price cutting needs to be prevented by detecting punishing deviations of the consensus-pricing. Stigler suggested that the greatest obstacle to collusion, in absence of entry, would be what he characterised as “secret price cutting”. Although they recognise that concentration is an important, and perhaps even necessary condition for oligopolists to collude, they argue that other factors in the market, such as the existence of countervailing buyer’s and presence of barriers to entry can inhibit collusion from coming into existence. 

Furthermore, the threat of punishment must be such as to constitute a stabilising factor for collusion. The losses to be incurred from punishment must be severe and credible enough as to outweigh the short-term profits the cheating firm can draw from undercutting the collusive price. Stigler’s insight thus led to the recognition that it is far from inevitable that oligopolists behave non- competitively, as firms under certain market conditions will have a powerful incentive to cheat on any co-operative understanding. Stigler’s work influenced, and still continues to influence economic thinking concerning oligopolies.

Furthermore, behavioural scholars point out that market structure should not be taken as exclusively exogeneously determined, but at least partially also influenced endogeneously. Large companies should not be condemned per se as they can be the result of superior efficiency. They thus maintain that the SCP model of the Structuralists also works the other way around: market structure is determined by the performance of firms, and should not be attacked because firms have succeeded. 

 

4.  Game theory

With the arrival of the New Industrial Organisation Theory, a new instrument for analysing and predicting oligopolistic, interdependent, behaviour has been developed, i.e. “game theory”. Although Neumann and Morgenstern developed the first seminal work in the field of Game Theory in 1944, it did not receive full attention until the 1970’s. Since then, game theory has rapidly developed into a useful instrument which can be used to gain insight into the dynamics of oligopolistic interdependence. Game theorists see the market-play between competitors as a “game”, the competitors being rational players in the market, trying to maximise their profits. Everything revolves around the search for a possible equilibrium or equilibria, i.e. a combination of the strategies that represent the best strategy for every player.

A game theoretic perspective differs from earlier economic understandings in part because it assumes that each manager adopts his or her best strategy, based on assessment of competitors’ best strategies, not simply that managers in oligopolistic markets attempt to take competitors’ likely reactions into account.⁵⁶ It furthermore recognises that the oligopoly game in the market is not a zero- sum game, i.e. somebody’s gain is not necessarily somebody else's loss.

Two main branches of game theory exist: co-operative and non co-operative game theory. In co-operative game theory models competitors are allowed to make binding agreements that restrict their feasible strategies. As the particular problem of oligopoly is that competitors behave independently on the market, co-operative game theory is not generally used to analyse oligopoly games.

4.1 Non co-operative game theory is the currently accepted economic mode of analysing oligopoly interactions. This branch of game theory begins with a non- co-operative viewpoint, i.e. it assumes that each firm’s independent choice of its best strategy will result in equilibria (of strategies), that are non-co-operatively optimal given the others’ similarly calculated optimal strategies. It is presumed that firms cannot communicate, or they cannot, at least, rely on the contents of the communication.

4.2 Static games The one-shot game models constitute the simplest form of non-cooperative game theory. In such a static game, every player can move only once, without having knowledge of the move of the other player. It is presumed that both players move contemporarily, so that there is no possibility of knowing beforehand the move the other player makes. Cournot’s and Bertrand’s analyses of oligopolies can be seen as a one-shot game, depending on the variable the competitors have chosen to determine their strategies. The choices oligopolists face in the market-game can be illustrated by the so-called “prisoner’s dilemma”. 

4.3 Prisoner’s dilemma

Suppose two villains are caught in a stolen car after having robbed a bank. Both are being put into separate cells in prison; there is no possibility that they can communicate with each other. The District Attorney lacks the required proof in order to charge the villains with bank-robbery, and would need a confession of one of the crooks. She does however have sufficient proof to charge them with car stealing. The villains, being separately interrogated, are being given the following choice: confess or deny. Do both confess, they will be charged with six years of imprisonment each. If only one of them confesses, he will walk a free man, while the other, who denied, will have to face 10 years in jail. Do both deny, they can only be charged with car stealing, the punishment for which is one year of imprisonment each. The choices can be reproduced in the following “pay-off matrix”, indicating the years in prison:

A/B	Confess	Deny
Confess	6,6	0,10
Deny	10,0	1,1

 

Choosing purely out of self-interest, A is better of confessing, whatever choice B makes. The same holds true for B. the outcome then will be that both confess and both will be imprisoned for six years.

Such a “dilemma” can be transposed to oligopolistic market situations, where firms have to choose between setting a high or a low price (or output, depending on the variable chosen). The outcome of the dilemma is that both firms will prefer not to run the risk of losing demand by being the only one charging the high price, and both firms will set a low price, earning lower profits than would have been possible by both setting the high price. This, i.e. the best strategies of each firm to maximise its profits, given the strategies chosen by the other competitors, is known as the Nash equilibrium. In this equilibrium, neither of  the firms has an incentive to change its strategy, given the strategy of the other firm. This outcome is however “Pareto  inferior”, i.e. it is not collectively the best outcome. The “Pareto optimum” combination would be of both firms charging high prices, leading to the most profitable outcome for both.

4.4. Repeated games

Oligopolists do not play the market game only once, and firms know that they will “play the game” repeatedly with the same competitors. Firms can monitor the strategies of the other firms in the market played in previous games and base their present decisions on the behaviour of their rivals in the past. Past actions do however not affect current feasible actions and pay-offs: any effects arise purely because the oligopolists remember what has happened and condition their actions on that history. Repeated game theory does not take into account the possibility that firms might make lasting investments or commitments.

In a non-co-operative repeated game with a finite horizon, i.e. the number of periods in the game is known and limited, collusion will not occur if the players know which period will be the last. This result can be explained by backward induction: in a one period prisoner’s dilemma non-co-operative game, we have seen that the best strategy for each player will be to not co-operate. This means that it is rational for both players not to co-operate in the last period. Given the certainty that there will be no co-operation in the last period, then it is not rational to co-operate in the pre-last period, as there will be no rewards of co- operation in the last period, and so on. In reality however the market game is played infinitely, or at least, competitors have no knowledge of which is the last game. Firms can either choose to price low, competing vigorously with their competitors, to the detriment of all, or choose to reduce competition between the competitors. In non-co-operative games with an unknown or infinite number of periods, termed super-games, a plethora of different equilibria, amongst which collusive ones, is possible.

Contemporary economics however suggest that it is typically the difficulties of policing collusion rather than the difficulties of reaching the terms of collusion  that inhibit co-ordination in most oligopolies. It can be seen from the “prisoners’ dilemma” mentioned above that an incentive exists for both firms to “cheat” on the collusion by charging the low price, thus undercutting the other firms’ price. The dynamics of continuous reciprocal price-undercutting, leading to low price level, is unprofitable for both firms. Whether oligopolists will choose to either stick to the collusion or to cheat depends on whether the short-run returns to cheating outweigh the medium and long-term profits to be drawn from the collusive level of prices. It is necessary for the stability of collusion that deviation will be detected within a small enough time lapse that the punishment will outweigh the profits to be drawn from deviating, so that the “collusive-equilibrium” can be artificially upheld. Profitability of deviation and detection thereof are dependent on the transparency in the market. Of further importance in the decision of a firm either to stick to collusion or to defect is the severity of the punishment the cheating firm will receive. While a threat of a fierce price war, resulting in a return to the competitive Betrand equilibrium might be severe enough to make firms stick to collusion, a mere return to the higher prices of the Cournot-equilibrium might not constitute a sufficiently severe threat to offset the unprofitability of cheating and consequently being caught.

Another prerequisite is that the threat of punishment has to be credible. As punishment is costly also for the punishing firms, the punishing firms need to possess or be able to produce low cost excess capacity in order to lower the prices. While below cost pricing might not constitute a credible threat for punishment, as the punishing firms also have to suffer losses, the threat to return to competition instead might be convincing. This strategy is however only credible if it is more profitable for the punishing firms to return to competition than to passively accomodate, i.e. allowing the cheating firm to deviate.

To enhance the credibility of punishment, firms can tie their hands by making irreversible (sunk) investments, i.e. in order to acquire excess capacity, or commit themselves to most favoured customer clauses or meeting competition clauses in sales contracts. Another way of enhancing credibility of punishment is when firms agree upon a trigger price level, studied by Green and Porter in their quantity-setting super-game model. If the price drops below this level, all firms will revert to a static Cournot equilibrium for some period of time.

4.5 Dynamic games

Brief mention should also be made of the concept of dynamic games. Unlike in normal repeated games, in dynamic games, strategic interactions take into account the history of the game, as past actions affect current feasible actions or pay-offs. Prediction of the strategy of a firm is slightly more complicated than in a one-shot game. In a dynamic game, moreover, the Nash equilibrium concept cannot deal with sequential moves in a satisfactory way. Stigler’s core insights also remain important in repeated-game theory: in order to co-ordinate in repeated game theory models firms firstly need to determine the terms of their co-ordination, then monitor adherence and punish deviations. As to reaching the terms of agreement, game theory cannot (yet?) explain how oligopolists can, without explicitly colluding, reach coordinative outcomes. Game theory, moreover, does not make a distinction between explicit and tacit collusion; it would suggest that both be maintained because it is in the individual competitor’s self-interest to do so. The critical issue is to analyse whether market characteristics are such as to prevent, or enhance such co-operative behaviour, resulting in a high price rather than in a low price.

5. Market characteristics

Many authors before have drawn up lists of market characteristics facilitating and stabilising collusion, affecting incentives for firms in choosing their strategies. A paradox exists in that the characteristics which lead to perfect (excessive) competition frequently are the same as those that increase the likelihood of cheating on any collusive agreement that evolves. Shapiro refers to it as the “topsy-turvy” principle of collusion: “anything … that makes more competitive behaviour feasible or credible actually promotes collusion”. Neither game theory nor conventional economic theory can provide a detailed qualitative answer that encompasses all characteristics. What is usually done in game-theoretic models is that the influence of one or two factors is analysed while applying rather rigid assumptions concerning the other factors. A list of the main market characteristics enhancing and/or stabilising (tacit) collusion will be given here. 

5.1 Number of competitors and concentration

5.1.1 Number of competitors The first factor influencing the probability of collusion is how the supply side in the market is structured, i.e. the number of suppliers in the market.  Adam Smith already observed that independent action might occur with only two sellers, but was more likely to occur with 20 or more sellers. It is obvious that it is easier for a few suppliers to co-ordinate their behaviour than if there were dozens of competitors present in the market. The number of paired relationships rapidly grows with the increase of the number of firms present in the market, according to the formula N(N-1)/2.  Among a limited number of suppliers interdependence is more acutely felt: the larger the number of suppliers, the lower oligopolistic interdependence generally becomes, as the awareness of collective interest is lowered. In a market with a large number of suppliers, apart from making co-ordination of behaviour more complex, the possibility of detection of price undercutting is also reduced, as cheating will give rise to relatively smaller quantity changes in the demand each firm faces. Moreover, as the number of firms in an industry increases, generally, the price elasticity each individual firm faces increases.

In assessing the absolute number of suppliers in a market, several economists have laid down indicative numbers beneath which oligopolistic dependence with a danger of collusion is presumed to exist. Hay and Kelley in their empirical survey consider 10 or fewer in 79% of the cases they studied. They found that the average number of firms involved in conspiracy was 7.25. Scherer and Ross hold that as a very crude general rule, if evenly matched firms supply homogeneous product in a well-defined market, they are likely to begin ignoring their influence on price when their number exceeds ten or twelve. Empirical studies employing US data suggest that co-ordinated behaviour is unlikely if co- ordination of more than four firms is required to control 50% of the market. More recently, Bresnahan and Reiss suggested that three firms were enough for effective competition. Discussion exists on whether, in some industries, even the co-existence of only two firms in a market is enough to ensure effective competition. 

Game theoretic research by Selten, built upon by Phlips, has led them to conclude that “4 are few and 6 are many”. The outcome of their models is that when four or fewer firms operate in a market the likelihood of collusion will be one, while it drops to zero with six or more firms. 

5.1.2  Concentration of competitors

The number of firms, however, does not always provide a reliable indication of how the market is structured: if there are a large number of firms operating in the industry, but only a few firms account for a large proportion of the industry sales, then these firms may be able to co-ordinate their behaviour without taking into account the numerous other smaller firms. A more sophisticated way of appraising the market structure on the supply side is by taking into consideration the size and the size distribution of the suppliers, measured in terms of market shares, either by the value or by the volume of the output. 

Several instruments for measuring concentration, such as concentration ratios and indexes, have been developed. A concentration ratio takes into account the market share of a certain number of the largest firms in the market. An example of concentration ratios can be found in German competition law: If two or three firms have a market share of 50% or over, or five or less firms have a combined market share of two-thirds or over, they are assumed to be market dominating. Concentration ratios, however, do not give any indication on the relative size of the companies. Another shortcoming in the use of concentration ratios is that the total number of firms present in the market is not taken into account. Concentration indexes might therefore constitute a more reliable instrument of measuring concentration. The Herfindahl-Hirschmann Index (HHI), calculated as the sum of the squares of the market shares of the firms in the market, is  being used in US merger control, providing indicative levels of concentration above which serious doubts that competition is threatened are raised. Firms can purposefully manipulate the number and concentration of firms in a market. This can for example be achieved through acquiring a competitor or forming a perfect cartel, in such a way that the members, in adopting the same conduct on the market, will act as one entity. 

 

5.3 Homogeneity of suppliers

The homogeneity of suppliers is a factor that can often provide a first indication of equal market prominence and a commonality of interests, while reducing the degree of uncertainty each competitor faces concerning the costs and demand conditions of the others. One can look at homogeneity of size, market share, costs, capacities, degree of vertical integration etc. When producers are homogeneous, there is an adequate balance in gains to be had from collusive practices. Discussion however exists on whether symmetry in market shares is a prerequisite for collusion.

5.4 Cost structure

An important factor determining the homogeneity of suppliers is the structure of the costs of the different suppliers. Different cost structures might be a factor  reducing the probability of collusion, as companies facing different cost situations could have a very different view on the prices that they would like to prevail in the market. High cost firms will be reluctant to agree upon substantially decreased output. In such a situation it obviously becomes more difficult to co-ordinate the behaviour of the suppliers.

It is furthermore important to look at how the costs of the suppliers are structured. It is in general easier to sustain collusion where the marginal costs are relatively inelastic or when marginal costs are high in relation to the total costs, as an increase of output causes a large increase of marginal costs, reducing the gains of cheating. In contrast, when the marginal costs of a firm are low in relation to the fixed costs, gains from cheating are high, which makes collusion unstable. Paradoxically, low marginal costs in an industry can also have a stabilising effect on collusion, as the threat of punishment in such cases is credible, since it is for other firms less costly to increase output in order to punish a deviator.

5.5 Capacity

Homogeneity of suppliers depends also on their capacities, as equal capacities will enhance collusion. Differences in capacity utilisation which lead to different average costs instead make collusion more difficult to attain and maintain. An ambiguous factor affecting the likelihood of collusion is excess capacity: under-utilisation of capacity provides an incentive to cheat, as competitors are eager to exploit more of their capacity. The danger in cheating on collusion however is that, when more competitors possess excess capacity, it can lead to deep price cuts. High excess capacity of competitors in the market can, furthermore, constitute a barrier to entry, deterring new firms from entering the market. 

5.6 Countervailing competitive power

Real competition

Profitability, and thus attractiveness of collusion, can be limited in the exertion of market power by the existence of other competitors in the market. These “fringe” competitors can, in response to a price raise by the firm(s) exerting market power, increase their outputs, thus impeding the price increase from leading to a profit increase for the former. A prerequisite is that they possess excess capacity or that it is not too costly for them to significantly increase their output. The factor of competitive countervailing power can also be manipulated, either by absorbing a (maverick) competitor through acquisition, or by forming a cartel, forcing competitors to abstain from competing.

5.7 Potential competition

Firms that threaten to enter the market, attracted by the supra-competitive profits that can be earned furthermore constitute a competitive constraint. Two different types of entry can be distinguished: short-term and long-term entry. The first type of entry, the so-called hit-and-run or uncommitted entry, can occur in contestable markets, as no, or only low, sunk costs are required. Normative suggestions can be taken from the ‘theory of contestable markets’, developed by Baumol, Panzar and Willig mentioned in paragraph . Baker refers to this kind of entry as “trump”. Entry of this type constitutes a credible threat or remedy against supra-competitive pricing, and a disincentive for collusion. The possibility of hit-and-run is however rare, as in reality some kind of investments will always be required. Hit-and-run entry might be possible for firms already operating in neighbouring markets, which can readily divert their production to producing the product in the market where supra-competitive profits can be earned. This phenomenon is referred to as supply-side substitution. The prerequisite however is that the opportunity costs of substituting out of the current product is low.

The second type of entry is constituted by entry requiring significant sunk costs, which will only be undertaken by firms that plan to remain in the market in the long term. In order for such entry to form a countervailing restraint, entry must occur timely, likely and sufficient in its magnitude, character and scope. The US 1992 Merger Guidelines refer to this type of new competition that requires expenditure of significant sunk costs of entry and exit as “committed entrants”.

5.8 Barriers to entry

Whether the possibility of entry constitutes a competitive restraint depends mainly on the presence or absence of barriers to entry. Academic discussion existed and still exists as to what needs to be understood by the notion “barrier to entry”. According to Bishop and Walker the key question is “what factors enable an incumbent or incumbents to earn profits in excess of normal profit levels while other equally or more efficient firms are excluded”. Many categorisations of barriers to entry have been laid down. Kühn, Seabright and Smith make divide barriers to entry into three main categories: artificial, natural and strategic barriers. The first type of barriers to entry, i.e. the artificial ones, are barriers posed by formal restrictions, such as legislation or trade association rules. Such barriers are easily recognisable. The second type of barriers, the natural barriers to entry are constituted by technology of production, e.g. high economies of scale and/or high fixed costs requiring high investments, including high sunk costs, i.e. costs which the firm will be unable to recoup.

One can distinguish between exogeneous and endogeneous sunk costs. The former are incurred by any firm entering a market, e.g. acquisition of a plant and economies of scale. Endogeneous sunk costs, on the other hand, are determined by the entering firm itself, e.g. advertising expenditure. Incumbents can manipulate the latter type of entry-costs, e.g. by heavily investing in advertisement, creating consumer loyalty. The entering firm will face higher sunk costs, as it will now also need to advertise heavily in order to be able to penetrate the market.

This leads us to the third type of barriers to entry, the strategic ones. These barriers are being set up by firms in order to deter entrants, e.g. acquiring low cost excess capacity, enabling low pricing, perhaps even predatorily low pricing. In order to make the investment worthwhile, the prospect for the entering firm has to be that supra-competitive profits can be earned: a firm might not be willing to enter a market where fierce competition exists  

5.9 Countervailing buyers’ power

Another factor influencing the possibility of exerting market power is constituted by the countervailing power buyers can exert in a certain market. High buyer concentration and buyers’ sophistication increases the dependence  of sellers on the buyers, which puts the latter in a better bargaining position. The use of long-term contracts likewise creates dependency of the supplier on the buyer. A further consequence of long-term supply contracts is that market transparency is lowered, because of the confidential contractual relations, reducing the likelihood of collusion. Furthermore, the ability and/or possibility for buyers to switch to alternative resources of supply or to develop production upstream and provide for their own in-house supply in response to a price raise will influence the potential of charging anti-competitive prices. 

5.10 Homogeneity of products

Another factor enhancing collusion is that the market consists of homogeneous commodity products, appeasing agreeing upon a price or a level of output comparing them and detecting any undercutting. The homogeneity of products increases substitutability between competitors, increasing the demand elasticity each firm faces. Collusion is then enhanced so as to increase profits for every competitor within the oligopoly without losing a part of their market share. Furthermore, monitoring is facilitated in a market with homogeneous products, as changes in market shares are more likely to provide an indication of price-cutting when products are homogeneous than when they are heterogeneous.

A further role is played by the importance of technology in a product market. If the relevant product is not a commodity product, but relatively high technology is needed to manufacture a product, the complexity of products and production processes increases, reducing the likelihood of collusion. High technology will also increase the proportion of fixed costs. Furthermore, in a product market where research and development play a major role, product innovation leads to product differentiation, and collusion will become, as shown above, more improbable. A mature technology product market instead appeases collusion.

Firms can try to artificially standardise their products. Conversely, firms can try to purposefully differentiate their products, for example, by investing in advertising, in order to lower the demand elasticity they face, which can enhance a different kind of collusion, namely, market area division.

5.11 Transparency

Another important factor is the level of transparency in the market. High transparency as to prices or market strategies clearly enhances collusion, as it facilitates the acquisition of information on the behaviour of competitors, reducing uncertainty concerning their strategies. A further consequence of such high transparency is that it makes it easier to detect a competitor “cheating” on the collusion, enabling the other competitors to take immediate retaliatory action, so that the cheater will thus not be able to draw much profit from his actions. The threat of such instant actions works as a stabilising factor of collusion. Links, either structural or other, created between undertakings, can constitute a channel through which information can be exchanged. An indicative factor of the existence of collusion can be found when competitors have developed facilitating practices in order to increase the transparency, and thereby reduce uncertainty, in the market.

5.12 Multi-market contacts

It is necessary, when assessing likelihood of collusion, to consider interdependencies which reach beyond the boundaries of the market being examined.  As the product and geographic areas where oligopolists encounter each other widen, their interdependence becomes greater, so each has more to lose from any move, which in turn might make any price initiative increasingly unlikely. Through these multi-market contacts the competitors are in a better position to develop an impression of interest and characteristics of their competitors, thus reducing uncertainty as to the future market behaviour of competitors, as the contacts may reduce information gaps. Multi-market contacts between oligopolists will thus facilitate, as well as stabilise collusion, as they make it easier to detect cheating and increase the possibilities of retaliation, by creating a “network of vulnerability”. 

 

5.13 Variations of demand over time

Growth of demand in the market is another factor to be taken into account. There is evidence that collusion occurs more frequently in a market where demand is stagnant or slowly growing with below-average profits. In a stagnant or slowly growing market growth can only be achieved at the expense of market shares of competitors. A mature, saturated market can be seen as indicative of an aggravated threat of collusion, while a considerably growing market instead decreases the tendency towards collusion, as competitors will try to capture a share of the new demand. A growing demand in a market furthermore renders the detection of deviations more difficult. When an industry is subject to cyclical fluctuations in demand, it becomes more difficult for firms to determine whether changes in demand for their products reflect cheating or whether they are simply caused by an overall change of industry demand. Moreover, volatility in demand adds noise to the data firms monitor in order to try to detect cheating.

 

5.14 Characteristics of transactions

A prerequisite for the awareness of interdependence is that competitors meet each other frequently on the market, or, in game-theoretical language, play the game repeatedly. The possibility of detecting deviations is dependent on the frequency and size of transactions. The incentive to cheat increases, and the possibility of detection decreases the more lumpy and infrequent transactions are in a market. Furthermore, large orders increase the attractiveness of cheating, by granting buyers secret discounts, earning high short-term profits. Stigler has pointed out that secret price cuts would not be made to “…buyers whose purchase fall below a certain size relative to his aggregrate sales.” Such large orders furthermore increase countervailing buyers’ power, as they can play the firms that will compete for the order off against each other.

According to game theorists, competition authorities need not be concerned when a market is characterised by a limited number of large contracts, as rational operating companies will not collude in such markets. In “bidding markets”, i.e. a market where sales are made through a tender process, however, the risk exists that competitors adhere to a bidding-cartel, so-called bid rigging, dividing amongst themselves the bids to protect themselves from the possibly detrimental effects of competition.

 

5.15. Facilitating practices

“[U]ncertainty is an oligopoly’s greatest enemy”

Firms have developed facilitating practices that enable them to reach desirable terms for collusion and/or support collusive outcomes, limiting the influence of factors that destabilise co-operative outcomes. The different facilitating practices have in common that they increase transparency, reducing uncertainty on the behaviour of competitors. Facilitating practices can affect the values of the pay-off matrix and alter a firm’s incentives to arrive at and stick to collusion. Salop refers to the adoption of facilitating practices that directly affect the pay-off matrix as “ incentive management”.  As facilitating practices can serve both anti-competitive and pro-competitive ends they have become a frustration for competition authorities.

As we have seen above, the theory of repeated games adopts the perspective that, in achieving a collusive equilibrium, many outcomes are typically possible, and the choice of the outcome turns on one of the outcomes standing out as self- evident. Such an outcome in game-theory terminology is called “focal”. Thomas Schelling, the originator of the idea of “focal points” and its application to oligopoly co-ordination, held that oligopoly behaviour can probably be explained by “tacit communication and its dependence on qualitatively identifiable and fairly unambiguous signals that can be read in the situation itself”. Focal points might be enough to signal to the other competitors the preferred outcome, and for all the competitors to select it as being the self- evident way to behave. Firms can also identify “focal rules”, such as basing point pricing, preserving existing price differences or existing market shares. Once a focal rule of this sort is identified, identifying then a focal point merely turns on choosing a single parameter, such as the common percentage price increase or common percentage output reductions. 

A list of facilitating practices facilitating collusion has become a staple of both economic texts and antitrust monographs. 

5.16 Price leadership

Price leadership is an industry practice in which one firm’s pricing movements are followed by its rivals. Price leadership is a phenomenon that is difficult to judge. It can constitute a facilitating practice, as the price leader indicates the preferred ‘focal’ consensus price or output level, but can, on the other hand, also be merely the natural working of the market. Economic studies have shown that price leadership of some sort occurs in many oligopolistic industries. Structural scholar Joe Bain’s, in analysing price leadership systems, found that “in an oligopolistic market any independent price change by a single oligopolist tends to be read as an ‘offer’ by his rivals, and an acceptable reaction to the price change may be interpreted as an acceptance of the offer of the first firm. Thus, negotiation can perhaps take place through a series of public announcements rather than through a meeting of persons, and the meaning of true consensual action becomes vague”.

Three different forms of price-leadership can be distinguished. The first form is dominant price-leadership, i.e. one dominant competitor, in other words, the only firm large enough to significantly affect the market, imposes its prices upon the industry, and the other competitors follow, as they will have little to gain from diverging from the dominant firm’s prices. The second form is barometric price-leadership, a non-collusive form of price leadership whereby the price leader merely acts as a barometer of market conditions: In Stigler’s words: the barometric firm “commands adherence of rivals to this price only because, and to the extent that, its price reflects market conditions with tolerable promptness”. Such behaviour cannot, however, be condemned, as it is entirely legitimate to raise prices as costs increase. The third form is collusive price- leadership, either explicit or tacit, where the competitors commit themselves to adapt to price increases initiated by one of them, being the price-leader.

5.17 Exchange of information

We have seen that, in a competitively structured market, full information is a requirement for competitive outcomes. However, in an oligopolistic market, the increased transparency makes it easier to detect a deviating firm, shortens detection lags, and increases the possibility of punishing such a firm adequately. By decreasing the transitional losses from price rises and the transitional gains from price discounts, incentives are altered in such a way as to make the joint profit outcome easier to achieve and to attain. Game theoretical studies into the subject of information exchanges have led to useful insights for competition policies. 

Information can be exchanged directly between competitors or indirectly, via an intermediary, e.g. a trade association. Such trade associations engage in much activity that is both legal and useful; routine activities include publications containing beneficial general information about the relevant industry. Some information collection and dissemination by trade associations is competitively helpful or neutral, but can, however, contemporarily have anti-competitive effects, as it can enhance cartellisation or interdependent pricing. Information dissemination brings particular risks with it if it is done in oligopolistically structured markets.

Information exchange can furthermore take place in public, through, for example, public advanced price announcements, or run via customers. However, such practices often escape competition law control, as it is very difficult to distinguish between rational business behaviour and behaviour that consciously enhances collusion.

The harm of information exchanges to competition depends on, besides the market structure, the content of the information being exchanged, the way it is exchanged, and the frequency by which it is exchanged. The exchange of aggregate data concerning the whole market or industry is less likely to enhance collusion than information concerning individual companies, enabling the monitoring of compliance with collusion, so that the incentives to defect are diminished, as the threat of punishment is increased. Green and Porter have pointed out, however, that the exchange of aggregate data cannot always be considered harmless: In order to activate a punishment mechanism, firms can agree on a critical market price level, the so-called trigger-price, below which  the oligopolists will automatically assume that someone has cheated and trigger of a price war.

It is furthermore the information concerning current information that, from a game theoretical perspective, may alter incentives for firms and increase the likelihood of collusion. The exchange of such information reveals who is defecting, and consequently allows for immediate punishment. The frequency of information exchanges determines, furthermore, the lapse of time that can evolve between cheating and punishment thereof, and therewith the profitability of cheating. According to game theorists, competition policy should focus on frequent exchange of detailed, current information.

The exchange of historic information, as a general rule, is not expected to increase the likelihood of collusion to any serious extent, nor does the exchange of information regarding future behaviour, i.e. communication between competitors on how they are going to act, constitute a real threat to competition. From a game theoretical perspective, communication, carrying the promise to adhere to an agreement, while this is not a firm’s best choice, is just “cheap talk”, and will not directly affect the pay-offs. Only when the pay-off matrix is such that, with the help of a credible punishment strategy, parallel behaviour is the logical outcome anyhow, i.e. is an equilibrium amongst the possible equilibria, then communication just helps to determine and choose between collusive outcomes.

However, when communication concerning future behaviour carries a commitment, e.g. by publicly informing the consumers, the pay-off matrix is affected, and the likelihood of collusion might increase.  Firms can commit themselves to contractual obligations, such as the “most-favoured-customer” clause and the “meeting competition” clause, increasing the probability of detection of deviations. Cheap talk then turns into “less cheap talk”, with a higher potential of distorting competition.

Furthermore, is the content of the information exchanged of importance. From a game theoretical perspective, data can be divided into three different categories. The first category is constituted by data concerning prices, other sales conditions, sales and output. The exchange of such data helps colluding firms to identify deviation. The exchange of the second category of information, i.e. data on current costs, demand and capacity, helps to identify possible collusive equilibria, and to determine availability of effective punishment strategies. Exchange of the third category of information, namely data on investment, R&D, future costs and demand, helps to reduce the long-term uncertainty that may threaten a collusive equilibrium. According to game theorists the first two categories of information being exchanged should receive the most attention from competition authorities, as they are most directly related to the success of collusive behaviour.

5.18 Product standardisation

There can be very significant benefits from product standardisation, e.g. when the standards deal with quality, used experience or safety. Standardisation can, however, also yield competitive harm. Standardisation may facilitate collusion by increasing transparency thus making it easier to agree on price and to detect deviations. Standardisation might, furthermore, slow down innovation.¹⁷²

5.19 Geographical Pricing Systems

Another facilitating practice, frequently used by competitors in order to standardise, or at least facilitate calculation of prices are geographical pricing systems, such as delivered pricing systems, which can consist of basing point pricing or zone pricing.

The belief that delivered pricing systems can facilitate collusive pricing has an intellectual tradition in antitrust economics. According to Dennis Carlton’s cogent summary of the relevant theory, delivered pricing systems may facilitate collusion in two ways. First, he points out that they will remove price discretion on shipping charges. Secondly, using the theory developed by Stigler,176 he retains that adherence to delivered pricing schedules might facilitate collusion in settings where shocks to demand in different geographic areas are imperfectly correlated. A delivered pricing system allows firms in the low demand area to penetrate the high demand area without costly and potentially destabilising revisions to the price schedule.

However, non-collusive rationales exist for the use of delivered pricing systems. Business justifications for the use of delivered pricing can counter the finding of anti-competitiveness, but will have to be all the more convincing when the market structure is more prone to collusion. Relevant characteristics to be assessed are the homogeneity of products, transportation costs constituting substantial portion of price, large scale of production, geographically scattered producers, and, most importantly, the concentration within the relevant industry.

6. Most Favoured Customer Clauses and Meeting Competition Clauses Long-term sales contracts often include most-favoured-customers (MFC) clauses, which provide the buyer with the insurance protection against the contingency that the seller may offer a lower price to another customer. The MFC requires the seller to pay a monetary penalty if he reduces price. Retroactive MFC clauses prevent price discrimination when the seller offers a discounted price to another buyer in the future, while contemporaneous MFC clauses are a non-discrimination guarantee between current customers.

The advantage of the inclusion of such a contract for an individual buyer is that he is guaranteed the lowest price that a seller charges to any of its other customers. MFC clauses in contracts furthermore reduce buyers’ search costs by easing the process of securing a favoured price without an extensive search for the best available price. The MFC device can also reduce negotiation costs by guaranteeing to buyers that they will not be disadvantaged in the future vis-a-vis rivals who buy from the same seller.

Theoretical studies have, however, shown that inclusion of MFC clauses in contracts may facilitate collusion in oligopolistic settings and lead to supra- competitive pricing. MFC clauses can be regarded as constituting information exchange by means of customers policing price levels. Furthermore, inclusion of MFC clauses provides a means of dis-incentivating deviations from a supra- competitive consensus price. Incentives for a seller to discount, and thus deviate, are reduced, as MFC clauses state that such deviations must be granted to all of its buyers. The use of MFC clauses thus raises the marginal costs of deviations, and decreases the short-term profitability thereof. The theory is that, when sellers understand that competitors face the same disincentives,  confidence is raised that each will respond similarly, and will stick to supra- competitive pricing.¹⁸⁶

Thomas Cooper has also analysed the contribution of MFC clauses to the achievement and maintenance of a tacitly collusive pricing structure. Cooper showed that, for firms to choose to adopt MFC clauses unilaterally, there need not be any explicit agreement among the rival producers to include these provisions in their sales contracts. Additionally, this analysis showed that all firms could profit from higher prices due to the use of MFC clauses even if not all firms introduce them into their contracts.

A meeting-competition clause (MC) in a long-term supply contract provides the buyer with insurance protection against a lost opportunity of being offered a lower price by another seller, as the original seller commits himself to match the lower price offered by any other seller.  A common variant is the meet-or- release clause, giving a seller the opportunity to meet a lower price, offered to the buyer by a rival seller, or to release the buyer from the contract.

As with MFC clauses the inclusion of a MC clause in a contract serves as an information exchange device, buyers controlling adherence to consensus pricing by oligopolists. Inclusion of the clause also diminishes or eliminates detection lags of chiselers, facilitating the selective matching of otherwise secret discounts. Assuming that the seller wished to match the discount, the rival’s strategy is countered. As a result, the co-operative joint profit outcome is made relatively more credible.

In the case of the inclusion of a no-release meeting-competition clauses in sales contracts, the seller binds himself to match any lower price offered by a rival seller. Economists argue that, with inclusion of MC clauses in contracts, like the inclusion of MFC clauses, supra-competitive consensus pricing is stabilised, as no oligopolist will wish to deviate from the joint profit outcome. Salop argues, moreover, that the inclusion in a contract of a MC clause also facilitates the successful achievement of the supra-competitive pricing outcome. When a no- release MC clause is provided jointly with an MFC clause, oligopoly co- ordination is held to be further facilitated.

 

7. Conclusion on facilitating practices

The effect on competition of the use of facilitating practices is difficult to analyse. While the use of such facilitating practices can have considerable competitive benefits, they can, and do often contemporaneously, have anti- competitive objects and effects, as they can facilitate the achievement of, and the monitoring and detection of deviations from oligopolistic pricing, thus helping to prevent a breakdown in oligopolistic discipline. For this reason, competition authorities encounter great difficulties in assessing the impact of such practices on competition in oligopolistically structured markets. 

8. Conclusion

Many studies on oligopolistic behaviour have been completed, but unlike perfect competition of pure monopoly, there exists no single “theory of monopoly”. Game theory in its present state does not seem to offer much guidance to competition policy concerning the number of firms that will enhance collusion. However, the only absolute conclusion that can be drawn from economic studies, both empirical and game theoretic, is that formulating general rules, which can be applied across all industries, appears to be problematic, if not impossible.¹⁹¹ However, useful suggestions can be drawn from the economic analysis for the application of competition laws to oligopolistic markets, especially those which resulted from game theoretic studies. One should, nevertheless, bear in mind that economic theories are necessarily based on simplifying assumptions often obtained in the context of stylised economic models that cannot take into account all the complexities of real life cases.

Furthermore, economics typically look at the result of collusion, without distinguishing between explicit collusion and tacit collusion: both have as an outcome supra-competitive pricing which is detrimental to consumer welfare. Lawyers, on the other hand, distinguish between explicit collusion, for which proof of prior concertation exists, and tacit collusion. For the latter, lawyers feel uncomfortable condemning it, relying merely on economic evidence.