A number of markets seem to be dominated by relatively few firms – think of motor vehicle manufacturing or commercial banking in the UK, or the newspaper industry. A market with few sellers is known as an oligopoly market. An important characteristic of such markets is that when making economic decisions, each must take into account its rivals’ behaviour and reactions. The firms are therefore interdependent.

An important characteristic of oligopoly is that each firm has to act strategically, both in reacting to rival firms’ decisions and in trying to anticipate their future actions.

There are many different ways in which a firm may take strategic decisions, and this means that there are many ways in which an oligopoly market can be modelled, depending on how the firms are behaving. This chapter reviews just a few models.

Oligopolies may come about for many reasons, but perhaps the most convincing concerns economies of scale. An oligopoly is likely to develop in a market where there are modest economies of scale – economies that are not substantial enough to require a natural monopoly, but are large enough to make it difficult for too many firms to operate at minimum efficient scale.

Within an oligopoly market, firms may adopt rivalrous behaviour or they  may choose to cooperate with each other. The two attitudes have implications for how markets operate. Cooperation will tend the market to the monopoly end of the spectrum, while rivalry will tend it towards the competitive end. In either scenario, it is likely the market will be somewhere between the two extremes.

The kinked demand curve model

One such model revolves around how a firm perceives its demand curve. This is called the kinked demand curve model, and was developed by Paul Sweezy in the USA in the 1930s

The model relates to an oligopoly in which firms try to anticipate the reactions of their rivals to their actions. One problem that arises is that a firm cannot readily observe its demand curve with any degree of certainty, so it must form expectations about how consumers will react to a price change.

The diagram above shows how this works. Suppose price is currently set at P; the firm is selling at Q, and are trying to decide whether to alter price. The problem is that it knows for sure about only this one point in the demand curve.

However, the firm is aware that the degree of sensitivity to its price change will depend on whether or not the other firms in the market will follow its lead. In other words, if its rivals ignore the firm’s price change, there will be more sensitivity to this change than if they all follow suit.

If the firm raises the price of the product, it is likely that the other firms in the market will ignore this raise, as it puts them in a better position. This is a non-threatening move that gives market share to other firms. This means that the quantity demanded will decrease by a much higher amount than the price decrease, as customers can go to the cheaper competitors, so price is elastic above P.

On the other hand, if the firm lowers the price of its product, then it is likely that firms will follow suit and also lower their prices. This is a threatening move and firms will lower prices to preserve their market positions. This means that a lowering of price is unlikely to amount to a large gain in quantity demanded, as competitors will be charging the same price, so price is inelastic below P.

Putting these together, the firm perceives that it faces a kinked demand curve. Furthermore, if the marginal revenue curve is added to the picture, it is seen to have a discontinuity at that kink. It thus transpires that Q is the profit maximising output for a wide range of cost conditions from MC1 to MC3; so even in the face of a change in marginal costs, a firm is unlikely to alter its behaviour.

Thus, the model predicts that, if a firm perceives its demand curve to be of this shape, it has a strong incentive to do nothing, even in the face of changes in costs. However, it all depends upon the firm’s perceptions. If there is a general increase in costs that affect all producers, this may alter the firm’s perception of rival reaction, and thus encourage it to raise price. If other firms are reading the market in the same way, they are likely to follow suit.

It is important to note that this model does not explain how price reaches P in the first place.

Game theory

A more recent development in the economic theory of the firm has been in the application of game theory. This began as a branch of mathematics, but it has become apparent that it has wide applications in explaining the behaviour of firms in an oligopoly.

Game theory itself has a long history, with some writers tracing it back to a correspondence between Pascal and Fermat in the mid seventeenth century. Early applications in economics were by Antoine Augustin Caournot in 1838, Francis Edgeworth in 1881, and J. Bertrand in 1883, but the key publication was the book by John von Neumann and Oskar Morgenstern (Theory of Games and Economic Behaviour) in 1944. Other famous names in game theory include John Nash (played by Russel Crowe in the film A Beautiful Mind), John Harasnyi and Reinhard Selton, who shared the 1994 Nobel Prize for work in this area.

Almost certainly, the most famous game is the Prisoner’s dilemma, introduced in a lecture by Albert Tucker (who taught John Nash at Princeton) in 1950. This simple example of game theory turns out to have a multitude of helpful applications in economics.

Two prisoners, Al Fresco and Des Jardin, are being interrogated about a major crime, and the police know at least one of the prisoners is guilty. The two are held in separate cells and cannot communicate with each other. The police have enough evidence to convict them of the minor offence, but not enough to convict them of the major one.

Each prisoner is offered a deal. If he turns state’s evidence and provides evidence to convict the other prisoner, he will get off – unless the other prisoner also confesses. If both refuse a deal, they will just be charged with a minor offence.

In each case, Al’s sentence (in years) is shown in orange, and Des’s in blue. In terms of the entries on the table, if both Des and Al refuse a deal they will each go down for the minor offence for a year. However, if Al confesses and Des refuses a deal, then Al will go free and Des takes the full rap of 20 years. If Des confesses and Al doesn’t, the reverse will happen. However, if they both confess they will each get 5 years.

Think about this from Al’s point of view, remembering that the prisoners cannot communicate, so Al does not know what Des will choose, and vice versa. You can see that whatever Des chooses to do, Al will be better off confessing. This is known as the dominant strategy – coined by John Nash.

The dilemma is, of course, symmetric, so for Des too the dominant strategy is to confess. The inevitable result is that if both prisoners are selfish they will both confess and get 5 years in jail. If they had remained silent they would have both been better off; but this is too risky a strategy for either of them to adopt. A refusal to do a deal might have led to 20 years in jail.

What has this to do with economics. Suppose there are two firms, Diamond Tools and Better Spades, operating in a duopoly market, Each firm has a choice of producing high output or low output. The profit made by the firm depends on two things: its own output and the output of the other firm.

If one company produces high, and the other low, then high gets £3m, the other £0. If they both produce high, they each get £1m. If they both produce low, they each get £2m.

Therefore, the dominant strategy is to produce high, as you will always profit maximise no matter what the other player chooses.

When players have made their choices and seen what the other player has done, each player will feel justified in its actions, and thinks that it took the right decision given the rival’s move. This is known as the Nash equilibrium, which has the characteristic that neither firm needs to amend its behaviour at any future point in time. This model can be used to investigate the wide range of decisions the firms need to take strategically.

Cooperative games and cartels

Look back at the Prisoners Dilemma. It is clear that the requirement that the players be unable to communicate is  serious impediment from the firms point of view. If both players could agree both remain silent, they would minimise their joint losses, but they will not risk this strategy is they cannot communicate.

If they could join together in a cartel, the two players could come to an agreement to adopt the silent-silent strategy. However, if they were to agree to this, each firm would have a strong incentive to cheat,  because if each now knew the other firm was going to be silent, they would also know that they could confess and dominate the game – being successful whatever the other players choice.

This is a common feature of cartels. Collusion can bring high joint profits, but there is also a temptation for each member firm to cheat and to sneak some additional market share at the expense of the other firms in the market.

There is another downside to the formation of a cartel. In most countries, apart from Hong Kong and a few others, they are illegal. For example, in the UK the operation of a cartel is illegal under the UK Competition Act, under which the Office of Fair Trading is empowered to fine firms up to 10% of their turnover for each year that the cartel is found to have been in operation.

This means that overt collusion is rare. The most famous example is not between firms but between nations, in the form of the Organisation of Petroleum Producing Countries (OPEC), which for a long period of time has operated a cartel to control the price of oil.

Conditions favouring collusion

Some conditions may favour the formation of cartels – or at least, some form of collusion between firms.The most important of these is the ability for each of the firms involved to monitor the actions of the other firms, to ensure that they are keeping to the agreement.

In this context, it helps if there is a relatively small number of firms; otherwise it will be difficult to monitor the market. It also helps if they are producing similar goods; otherwise one firm could try to steal an advantage by varying the quantity of the product. When the economy is booming it may be more difficult to monitor market shares, because all firms are likely to be expanding. If firms have excess capacity, this may increase the temptation to cheat by increasing output and stealing market share; on the other hand, it also makes it possible for the other firms to retaliate quickly. The degree of secrecy about market shares and market conditions is also important.

Collusion in practice

Some firms may join in loose strategic alliances, in which they may work together in part of their business, perhaps in undertaking joint research and development or technology swaps.

Firms may also look at stronger alliances, such as the ones found with airlines, such as the Star Alliance and the One World Alliance, who carve up the long-haul routes between them.  Such alliances offer benefits to passengers, who can get access to a wider range of destinations and business class lounges and frequent flyer rewards, and to the airlines, who can economies on airport facilities by pooling their resources. However, the net effect is to reduce competition.

Alternatively, firms may look at tacit collusion, in which the firms in a market observe their behaviour very closely and refrain from competing on price, even if they do not actually communicate with each other. Such collusion may emerge gradually over time in a market, as the firms become accustomed to market conditions and each other’s behaviour.

One way in which this may happen is through some form of price leadership. If one firm is a dominant producer in a market, then it may take the lead by setting the price, with the other firms following its example. It has been suggested that the OPEC cartel operated according to this model in some periods, with Saudi Arabia acting as the dominant firm.

An alternative is barometric price leadership, in which one firm tries a price increase and then waits to see how other firms follow. If they do, a new, higher price has been reacted without the need for overt discussions between the firms. On the other hand, if the other firms do not feel the time is right for the change they will keep their prices steady, and the first firm will drop back into line or else loose market share. The initiating firm need not be the same one each round. It had been argued that the domestic air travel market in the USA has been operated in this way on some internal routes. The practice is facilitated by the ease with which the prices can be checked via the computerised ticketing system, so that each firm knows what the other firms are doing.

The frequency of anti-cartel cases brought by regulators in recent years suggests that firms continue to be tempted by the gains from collusion. The operation of a cartel is now a criminal act in the UK, as it has been in the USA for some time. I will discuss this further in a future article.


  • An oligopoly is a market with a few sellers, each of which takes strategic decisions base on likely rival actions and reactions.
  • Because there are many ways in which firms may interact, there is no single way of modelling an oligopoly market.
  • One example is the kinked demand curve model, which argues that the firm’s perceptions of the demand curve for their products is based on their views about whether or not rival firms will react to their actions.
  • This suggests that price is likely to remain stable over a wide range of market conditions
  • Game theory is a more recent and more flexible way of modelling interactions between firms
  • The Prisoner’s Dilemma can demonstrate the potential benefits of collusion, but also shows that in some market situations each firm may have a dominant strategy to move the market away from the joint profit maximising position.
  • If firms could join together in a cartel, they could indeed maximise profits, but there would still be a temptation for firms to cheat, and try to steal market share. Such action would break up the cartel, and move the market away from the joint profit maximising position.
  • However, cartels are illegal in most societies.
  • Firms may thus look for covert ways of colluding in a market, for example through some form of price leadership.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s