Cambridge International AS and A Level Economics: Oligopoly and game theory (9708)

This guide follows the Cambridge International AS and A Level Economics 9708 syllabus for examination in 2026, 2027 and 2028. It covers part of syllabus section 7.6.4: contestable markets, price competition and non-price competition, collusion and the Prisoner’s Dilemma (with a two-player pay-off matrix), and efficiency in oligopoly. Topics 1 to 6 of the syllabus are AS Level content and topics 7 to 11 are A Level content. This unit is A Level only, so it is not examined at AS.

Market structures, barriers to entry and concentration ratios are taught elsewhere, as is the performance of perfect competition, monopoly and monopolistic competition. Here we apply the ideas to oligopoly. Related pages: revision notes, practice questions, the wider price system and microeconomy guide, the Cambridge Economics hub, the checklist and the free diagnostics.

All figures below are invented and illustrative. They are not real statistics.

What you must be able to do

Syllabus bullet	You must be able to
contestable markets: features and implications	Describe a contestable market, explain hit-and-run entry, and judge how it changes firms’ behaviour
price competition and non-price competition	Give examples of each, explain why oligopolists favour non-price competition, and evaluate both
collusion and the Prisoner’s Dilemma in oligopolistic markets	Explain why firms collude, why collusion breaks down, and read a two-player pay-off matrix
efficiency and X-inefficiency (applied to oligopoly)	Judge allocative, productive and X-efficiency in the short run and the long run

1. Contestable markets

A contestable market is one where the threat of entry disciplines the firms already in it. What matters is not how many firms are in the market now, but how easily new firms could get in and out.

Features

• Low or no barriers to entry and exit.
• No sunk costs, or only small ones. Sunk costs are spent and cannot be recovered on leaving, such as specific advertising or a purpose-built plant. Equipment that can be resold is not sunk.
• Potential entrants can use the same technology and face similar costs to incumbents.
• Entrants can see the incumbent’s prices and profits (good information).
• Hit-and-run entry is possible: an entrant slips in when profits are supernormal, undercuts, takes a share and leaves with its capital intact if the incumbent fights back.

Implications

1. A market with one or two firms can still behave competitively, because the incumbents fear entry. Market structure alone does not predict behaviour.
2. Incumbents keep prices near average cost (limit pricing) and earn only normal profit.
3. Incumbents have an incentive to keep costs low, because any slack invites entry. This pushes towards productive efficiency and against X-inefficiency.
4. Policy can aim at lowering barriers, rather than breaking up firms.

Worked example: a sustainable price

A single operator runs a small ferry route in Country A. Demand is P = 20 − 0.5Q (price in currency units, Q = crossings per day). Marginal cost is 8. A fixed cost of 60 per day can be fully recovered if the operator leaves, so there are no sunk costs.

Step 1: profit-maximising monopoly. MR = 20 − Q. Set MR = MC: 20 − Q = 8, so Q = 12 and P = 14. Step 2: profit. Revenue = 14 × 12 = 168. Total cost = 8 × 12 + 60 = 156. Supernormal profit = 12. Step 3: the entrant. An entrant charges 13. Demand at 13: Q = (20 − 13)/0.5 = 14. Revenue = 182, cost = 8 × 14 + 60 = 172, profit = 10. Entry pays, and if the incumbent retaliates the entrant leaves without loss. Step 4: the price that stops entry is where profit is zero, P = AC. Solve (20 − 0.5Q)Q = 8Q + 60. This gives Q² − 24Q + 120 = 0, so Q = 12 + √24 ≈ 16.9 and P ≈ 11.55. Step 5: check. At P ≈ 11.55, average cost = 8 + 60/16.9 ≈ 11.55. Profit is zero. At price 11, Q = 18 and profit = 198 − 204 = −6, so no entrant would go lower.

The threat of entry cuts price from 14 to about 11.55 and profit from 12 to zero.

If sunk costs exist. Suppose entry needs a one-off 45 of non-recoverable branding spend, and the incumbent can respond within 3 days. The entrant earns 3 × 10 = 30 but spent 45, losing 15. Entry is deterred. The larger the non-recoverable spend, the weaker the entry threat.

Evaluation

• Contestability is a matter of degree. Few real markets have zero sunk costs.
• Incumbents may deter entry by predatory pricing, exclusive contracts or heavy advertising, which raise the entrant’s risk.
• Perfect information is unlikely: entrants may not know the incumbent’s real costs.
• Where a market is nearly contestable, a low concentration of sellers matters less than the credibility of the entry threat.

2. Price competition and non-price competition

Price competition means rivals use price to win sales: price cuts, discounts, promotions, loyalty pricing, a price war, or predatory pricing (setting price below cost to push out rivals).

Non-price competition means rivals compete on something else: advertising and branding, product quality and design, after-sales service, free delivery, loyalty schemes, product innovation and wider choice.

Why oligopolists prefer non-price competition

Firms in an oligopoly are interdependent: each firm’s best choice depends on what rivals do. A price cut is easy to copy, so the gain does not last, while the lower margin does. Non-price competition builds loyalty and makes demand less price elastic.

Worked example: a price cut

A firm sells 1,000 units at 50 with a unit cost of 25. Profit = 25 × 1,000 = 25,000. It cuts price by 10% to 45.

Case 1, rivals do not respond. Sales rise to 1,300 (+30%). Price elasticity of demand = +30% ÷ −10% = −3. Revenue = 58,500. Profit = 20 × 1,300 = 26,000. The cut works.

Case 2, rivals match. Total market demand is inelastic, so sales rise only to 1,060 (+6%), a market elasticity of −0.6. Revenue = 47,700. Profit = 20 × 1,060 = 21,200, which is 3,800 lower than before.

Rivals will match, so the cut is likely to hurt everybody. This is why price wars are rare and short.

Worked example: advertising

A firm spends 2,000 on advertising and sales rise by 120 units, at a margin of 30 per unit. Extra profit = 3,600 − 2,000 = 1,600. If a rival advertises too and sales gains cancel out, each firm bears the 2,000 for no gain.

Evaluation

Non-price competition can raise quality and choice, and may support dynamic efficiency (7.3.4). It can also be wasteful, push up costs and prices, and create brand barriers to entry. Price competition benefits consumers but can destroy profit, and may be followed by exits and higher concentration. The firm’s mix depends on how easily its product can be differentiated, how elastic demand is and how rivals react.

3. Collusion and the Prisoner’s Dilemma

Collusion means rivals coordinate to restrict competition, usually to raise prices and profits towards the joint profit-maximising level. It may be overt (a formal agreement) or tacit (unspoken, for example following the price leader). Many countries have competition laws against overt agreements, so tacit collusion is more common. Collusion is more likely with few firms, similar products, high barriers to entry, stable demand and easy monitoring of rivals’ prices.

Reading a two-player pay-off matrix

Two firms, A and B, each choose a High or a Low price. The pay-offs show annual profit in millions (A’s first, B’s second). The data is illustrative.

	B: High	B: Low
A: High	A 40, B 40	A 10, B 60
A: Low	A 60, B 10	A 25, B 25

Method:

1. Take A’s view. If B chooses High, A gets 40 from High and 60 from Low, so Low is better. If B chooses Low, A gets 10 from High and 25 from Low, so Low is better.
2. Low is A’s best choice whatever B does: a dominant strategy.
3. The matrix is symmetric, so Low is also dominant for B.
4. Both choose Low: the outcome is (25, 25). This is a stable outcome because neither can gain by changing alone.
5. Compare: if they could trust each other, both would choose High and earn (40, 40), a joint profit of 80 instead of 50.

This is the Prisoner’s Dilemma. Each firm acts rationally, yet both end up worse off than if they had cooperated. Collusion fails because each firm gains most by cheating (60 against 40) while the other keeps to the deal.

Repeated play and punishment

Firms meet again and again, so cheating can be punished. Suppose a cheat earns 60 in year one, then the rival responds with Low forever, giving 25 a year after that. Without discounting:

Years	Keep to High	Cheat in year one
1	40	60
2	80	85
3	120	110
4	160	135

Over a short horizon cheating pays. From three years, keeping the deal pays better. Collusion is more stable when the future matters, cheating is spotted quickly and punishment is credible.

Evaluation

Collusion hurts consumers through higher prices and less output, but may fund research. It is hard to sustain with many firms, differentiated products, falling demand, secret discounts, new entry or legal action.

4. Efficiency in oligopoly

Allocative efficiency is where P = MC. Productive efficiency is production at the lowest possible average cost. X-inefficiency is when a firm’s costs are above the lowest possible because of weak cost control, which is more likely where competition is weak.

Worked example: the welfare cost of collusion

The market demand curve is P = 60 − Q (Q in thousands of units, P in currency units) and MC is constant at 12.

Outcome	Price	Output	Profit	Consumer surplus	Deadweight loss
Competitive (P = MC)	12	48	0	1,152	0
Full collusion (joint profit maximum)	36	24	576	288	288
Partial tacit collusion	28	32	512	512	128

Full collusion: MR = 60 − 2Q = 12, so Q = 24, P = 36. Deadweight loss = ½ × (36 − 12) × (48 − 24) = 288. At price 28: ½ × 16 × 16 = 128. Colluding firms are allocatively inefficient, and the further price is above marginal cost, the greater the loss.

X-inefficiency. A protected oligopolist sells 24 (thousand) at 36. Its average cost is 20 but 17 is achievable. Profit is 384 against a potential 456, and the slack is 3 × 24 = 72.

Short run and long run

In the short run, an oligopolist may earn supernormal profit, or may be pushed into a price war with temporary losses. In the long run, high barriers let supernormal profit continue, while a contestable market pushes profit back to normal and pressure on costs raises efficiency.

Evaluation

Large oligopolists may reach lower costs through economies of scale, so productive efficiency can be better than in a small-firm market. Supernormal profit can fund innovation, giving dynamic efficiency. But without rivals pressing on costs, X-inefficiency may persist. Efficiency depends on competition, not on the number of firms.

Common errors

• Writing that a contestable market must have many firms. It may have one.
• Treating sunk costs as any fixed cost. Only unrecoverable costs count.
• Calling Low the dominant strategy without checking both of the rival’s options.
• Saying collusion always lasts. Show why each firm is tempted to cheat.
• Forgetting that non-price competition costs money and can be matched.

Official syllabus

Cambridge International AS & A Level Economics 9708 syllabus for examination in 2026, 2027 and 2028, section 7.6.4, published by Cambridge University Press & Assessment.