Oligopoly Explained: Interdependence, Cartels, and Game Theory

An oligopoly is a market structure in which a small number of large firms supply most of the output, and each firm is big enough that its own pricing and output decisions visibly affect its rivals. That last clause is the whole subject. In an oligopoly firms are interdependent, so no firm can set a price without first working out how the other firms will respond. Every other feature of oligopoly follows from that: the barriers that keep the group small, the tendency for prices to sit still, the pull toward collusion, and the fact that the model is usually analyzed with a payoff matrix rather than a single cost-and-revenue diagram.

Oligopoly is taught in near-identical form in AP Microeconomics, IB Economics, A level economics and college introductory courses. The vocabulary shifts slightly between syllabuses, but the model does not. This guide covers what counts as an oligopoly, how concentration is measured with worked numbers, the kinked demand curve and its weaknesses, cartels and collusion, the prisoner's dilemma with an explicit payoff matrix, price leadership, real industries, and the efficiency verdict.

What counts as an oligopoly

There is no fixed number of firms that turns a market into an oligopoly. The working definition is that a few firms account for a large share of industry sales, so the standard phrasing is "a few large firms" rather than "exactly four" or "fewer than ten". A market with two dominant firms is a special case called a duopoly, and most of the oligopoly reasoning applies to it unchanged.

Three conditions define the structure:

• A few large sellers. A handful of firms hold most of the market. Small fringe firms can exist alongside them; what matters is that the leading firms are large enough to move the market.
• High barriers to entry. New rivals cannot easily arrive to compete away profit, which is why the number of firms stays small and why economic profit can persist into the long run.
• Interdependence. Each firm's best decision depends on what the others do, and each firm knows the others are reasoning the same way.

The product itself can go either way. In some oligopolies the good is close to homogeneous, as with cement, steel, aluminum and crude oil. In others it is heavily differentiated, as with cars, smartphones, soft drinks and breakfast cereals. Differentiation changes the marketing tactics but not the structure. What sorts markets into oligopoly is the number of firms and the height of the entry barriers, not whether the products look alike.

Interdependence, the defining feature

Interdependence is what separates oligopoly from every other market structure, and it is the point examiners test hardest.

A firm in perfect competition is one of thousands. If it doubles its output, the market price does not move, and no rival notices. A monopolist has no rivals to notice. A monopolistically competitive restaurant that cuts prices takes a sliver of business from hundreds of others, too little for any one of them to react to. In each of those cases the firm can solve its own profit-maximization problem in isolation.

An oligopolist cannot. If one of three national mobile carriers cuts its monthly price by a fifth, the other two see their subscriber numbers fall immediately, and they will respond. So the profit from the price cut depends on their response, which depends on their reading of the first firm's likely counter-response. This mutual awareness is called strategic interdependence, and it is the reason oligopoly has no single agreed diagram. Instead, economists model it with several partial models, each capturing one behavior: the kinked demand curve for price rigidity, cartel models for cooperation, and game theory for strategic choice under uncertainty.

Barriers to entry: why the few stay few

If entry were easy, high profits would attract new firms and the industry would drift toward monopolistic competition. Oligopoly persists because entry is hard, usually for several reasons at once.

• Economies of scale. In industries where average cost keeps falling until a firm serves a large share of the market, a small entrant is stuck with high unit costs and cannot match incumbent prices. This is the most common source of oligopoly.
• Large sunk costs. Building a semiconductor fabrication plant, a national fiber network or an aircraft assembly line requires spending that cannot be recovered if the venture fails. That risk deters entry on its own.
• Patents and proprietary technology. Legal protection over a design, a molecule or a process locks rivals out for a fixed period.
• Brand loyalty and advertising. Decades of advertising build habits that an entrant must outspend to break, and the required spend is itself a barrier.
• Network effects and switching costs. A platform is more useful when more people use it, so a newcomer starts with a worse product for reasons that have nothing to do with its quality.
• Control of a key input or distribution channel. Owning the ore body, the landing slots or the shelf space limits what a rival can do.
• Strategic deterrence. Incumbents can hold prices below the level that would attract entry, or keep spare capacity as a visible threat to flood the market if anyone enters.

Measuring concentration: CR4 and the HHI

Because "a few large firms" is vague, economists make it operational with concentration measures.

The n-firm concentration ratio adds the market shares of the n largest firms. CR4, the four-firm ratio, is the version reported most often. As a rough guide, a CR4 above about 60 percent points to a tight oligopoly, and a CR4 below about 40 percent points to a more competitive market.

The Herfindahl-Hirschman Index (HHI) squares each firm's percentage market share and adds the results. Squaring gives extra weight to the largest firms, which is exactly what you want when the question is whether a few firms dominate. The index runs from close to zero in a fragmented market to 10,000 for a single firm with 100 percent of the market, since 100 squared is 10,000. The bands used in most textbooks treat an HHI below 1,000 as unconcentrated, 1,000 to 1,800 as moderately concentrated, and above 1,800 as highly concentrated.

Worked example

Market A has six firms with these shares: 35, 25, 20, 10, 6 and 4 percent.

Concentration ratio. CR4 = 35 + 25 + 20 + 10 = 90 percent. The four largest firms hold nine tenths of the market.

HHI. Square each share and add:

• 35 squared = 1,225
• 25 squared = 625
• 20 squared = 400
• 10 squared = 100
• 6 squared = 36
• 4 squared = 16

HHI = 1,225 + 625 + 400 + 100 + 36 + 16 = 2,402. That sits well above 1,800, so the market is highly concentrated, which matches the CR4 reading.

Now compare Market B, also with six firms, with shares of 25, 25, 20, 20, 5 and 5 percent. Its CR4 is 25 + 25 + 20 + 20 = 90 percent, identical to Market A. But its HHI is 625 + 625 + 400 + 400 + 25 + 25 = 2,100, noticeably lower. The two markets look the same through the concentration ratio and different through the HHI, because Market A has one firm with 35 percent while Market B's leaders are evenly matched. That is the standard exam point: the concentration ratio tells you how much the top firms hold in total, while the HHI also tells you how unevenly that total is split. You can run both measures on your own figures with the concentration ratio calculator and the HHI calculator.

Both measures share the same blind spots. They depend entirely on how the market is defined, they usually ignore imports and foreign competitors, they say nothing about the threat of entry, and they measure current shares rather than actual rivalry. A market with two firms that fight each other hard may deliver better outcomes than a market with eight firms that quietly avoid competing.

The kinked demand curve and why prices stick

The kinked demand curve model was built to explain a pattern economists noticed in oligopoly: list prices often stay unchanged for long stretches even when costs move.

The model rests on one behavioral assumption about how rivals react, and the assumption is deliberately asymmetric. If our firm raises its price, rivals hold theirs steady and happily take the customers who defect, so our firm loses a lot of sales. If our firm cuts its price, rivals match the cut to protect their share, so our firm gains very few extra sales.

That asymmetry produces a demand curve with a bend at the current price:

• Above the current price, demand is relatively elastic, because a price rise is not matched and drives customers away quickly.
• Below the current price, demand is relatively inelastic, because a price cut is matched and wins little new business.

Because the demand curve changes slope at the kink, the marginal revenue curve breaks. It has a vertical gap directly below the kink quantity. The firm still maximizes profit where marginal cost meets marginal revenue, and here is the payoff of the model: as long as the marginal cost curve shifts up or down within that vertical gap, the profit-maximizing quantity and price do not change at all. Costs can move and the price stays put. That is price rigidity, produced by the model rather than assumed. It also explains why oligopolists lean on non-price competition instead of price cuts.

The criticisms are as standard as the model itself. First, the model explains why the price at the kink is stable but never explains how that price got there in the first place, so it is descriptive rather than predictive. Second, the asymmetric reaction is assumed, not derived from any firm's optimization. Third, the empirical record is unfriendly: studies of actual oligopoly pricing have found that oligopolists change prices often, and that price changes are not obviously rarer in concentrated industries than elsewhere, especially during inflation, when everyone raises prices together. Fourth, if firms collude or follow a price leader, prices can be stable for reasons that have nothing to do with a kink. Learn it as a partial explanation of one observed behavior, present it with its criticisms, and do not treat it as the model of oligopoly. The kinked demand curve entry has the compact version.

Collusion and cartels: overt, tacit, and unstable

Since competing hard hurts every firm's profit, oligopolists have an obvious shared interest in not competing. When they act on it, they are engaging in collusion.

Overt collusion is an explicit agreement to fix prices, restrict output, divide territories or rig bids. A group of firms doing this formally is a cartel. A successful cartel behaves like a single monopoly: it sets a joint output where industry marginal revenue equals marginal cost, charges the corresponding high price, and splits the resulting profit through quotas. Overt price fixing between competitors is illegal in most jurisdictions, which pushes it underground and makes enforcement of the agreement harder, since members cannot go to court over a broken contract.

Tacit collusion achieves a similar result without any agreement. Firms watch each other's published prices and settle into a pattern where nobody undercuts, because everybody understands what undercutting would trigger. There is no meeting and no contract, only parallel behavior, which is why tacit collusion sits in a legal gray area in most competition regimes.

Why cartels fall apart. The same arithmetic that makes collusion attractive makes it fragile:

• Every member gains from cheating. At the cartel price, an individual member's marginal revenue on extra units exceeds its marginal cost, so quietly selling above quota, or shading the price a little, raises that member's own profit. The gain is private and the cost is spread across everyone.
• Cheating is hard to detect. Secret discounts, better credit terms, free delivery and quality upgrades are all price cuts that never show up on a list price.
• Firms are not identical. A low-cost member wants a higher quota and a lower price than a high-cost member. Bargaining over the split is where many agreements die.
• Profits attract entry. Unless barriers are very high, the cartel price invites new suppliers who are bound by no quota.
• More members means more instability. Coordination costs and the odds that at least one member defects both rise with the number of firms.
• Legal risk. Fines and leniency programs that reward the first firm to confess turn every member into a potential informant.

The oil exporters' cartel is the standard classroom example precisely because it shows both sides: it can raise prices when members hold to quotas, and it repeatedly struggles with members producing above them.

Game theory: the prisoner's dilemma with numbers

Because outcomes depend on rivals' choices, oligopoly is analyzed with game theory. The standard teaching case is the prisoner's dilemma, which shows why two firms that would both do better by cooperating end up competing anyway.

Two airlines, A and B, compete on the same route. Each chooses to hold its fare high or cut it. The cells show annual profit in millions for firm A first and firm B second.

Work through firm A's reasoning. If B holds its price high, A earns 80 by holding high and 100 by cutting, so A cuts. If B cuts, A earns 30 by holding high and 50 by cutting, so A cuts again. Cutting is better for A no matter what B does, which makes it a dominant strategy. The payoffs are symmetric, so cutting is dominant for B too.

Both firms cut, and the outcome is the bottom-right cell with 50 each. That cell is a Nash equilibrium: given what the other firm is doing, neither firm can improve its own payoff by changing course alone. The dilemma is visible in the numbers. Both would earn 80 in the top-left cell if they cooperated, but 80 is not stable, because from there either firm can jump to 100 by defecting. Individually rational choices produce a jointly worse result, which is exactly why cartels are unstable and why collusive prices need constant policing.

Two extensions matter for exams. First, repetition changes the game. When firms meet in the same market month after month with no known final round, a defector can be punished in later rounds, so cooperation becomes sustainable through strategies such as matching whatever the rival did last period, provided firms value future profit enough that the stream of collusive earnings outweighs the one-off gain from defecting. If instead the game has a known last period, backward induction unravels it and both firms defect from the start. That punishment threat is the formal reason tacit collusion can survive without a contract. Second, not every game has a dominant strategy for both players, so check each player separately and find the Nash equilibrium by testing each cell for a profitable one-sided deviation. Full worked matrices are in the game theory guide and the Nash equilibrium explainer.

Price leadership and other coordination without contracts

Price leadership is a common way oligopolists coordinate without breaking the law. One firm announces a price change and the rest follow. In dominant firm price leadership, the largest firm sets the price and smaller firms take it as given, supplying what they wish at that price. In barometric price leadership, the leader is not necessarily the biggest firm but the one with the best read on market conditions, so rivals treat its move as a signal about costs or demand. No agreement is made, no evidence of one exists, and the outcome can still look much like a cartel.

Firms coordinate through other soft signals too: public statements about future pricing on earnings calls, published list prices that rivals can observe daily, and most-favored-customer clauses that make discounting expensive for the firm that offers it. Competition authorities watch these practices closely because they can produce collusive outcomes without a smoking gun.

Non-price competition is the flip side. Where price cuts are quickly matched and only shrink everyone's margins, firms compete on things that are slower to copy: advertising and branding, loyalty programs, product features, packaging, warranties, store locations and service quality. That is why concentrated consumer markets tend to carry heavy advertising while the shelf prices of rival products move together.

Oligopoly in the real world

Airlines. Most individual routes are served by a small number of carriers, so fares on a route respond to what the other carriers do rather than to a textbook market price. Capacity decisions, frequent flyer schemes and slot ownership are all recognizable oligopoly behavior.

Mobile networks. After decades of consolidation, most countries are served by a handful of national networks. The capital cost of spectrum and infrastructure is a large sunk barrier, plans and prices track each other closely, and competition runs through handset subsidies, bundled streaming and data allowances rather than headline price wars.

Large commercial aircraft. The market for big passenger jets is close to a textbook duopoly. Development costs run into the billions, the learning curve is steep, certification is slow, and buyers are few and sophisticated. That combination has kept the number of viable manufacturers at roughly two for years.

Others worth naming in an essay. Supermarkets, fuel retailing, banking, soft drinks, breakfast cereals, credit card networks, cloud computing platforms and pharmacy chains are all commonly cited. The test in each case is not the number of brands you can name but whether a few firms are large enough that each one watches the others before deciding.

How oligopoly compares with the other three structures

Read the table by row and the boundary cases become clear. Oligopoly and monopolistic competition both feature differentiated products and heavy advertising, but only oligopoly has few enough firms for interdependence and high enough barriers for profit to survive in the long run. Oligopoly and monopoly both hold market power and can earn long-run profit, but the oligopolist's power is checked by rivals who can retaliate. The four-structure overview is set out in the complete market structures guide.

Is oligopoly efficient?

The verdict has a clear negative side and a genuine counterargument, and a good answer gives both.

The case against. Oligopolists face downward-sloping demand curves and set price above marginal cost, so the market is not allocatively efficient. Units that buyers value above the cost of producing them never get made, which is deadweight loss. Firms also do not generally produce at the minimum of average total cost, so they are not productively efficient either. Where entry barriers are high and rivalry is weak, comfortable incumbents can also let costs drift above the technical minimum, which is X-inefficiency. Add successful collusion and the outcome approaches monopoly pricing, with the extra harm that it is achieved by firms that officially compete.

The case for. Large scale can be genuinely cheaper. In industries with steep economies of scale, a fragmented market of small firms would produce at higher average cost, so the concentrated outcome can deliver lower prices than a competitive structure that the technology cannot support. Long-run profit also funds research and development, which is the dynamic efficiency argument: the profit that looks like a static loss may buy the next generation of products. Aircraft, semiconductors and pharmaceuticals are the usual illustrations. Where oligopolists compete hard rather than collude, buyers can get much of the benefit through better products and faster innovation.

The verdict. Oligopoly is statically inefficient on the standard criteria, since price exceeds marginal cost and output is below the competitive level, but its overall performance is contingent in a way monopoly's is not. It depends on whether the few firms actually compete. That is why competition policy focuses less on how many firms exist and more on their conduct: blocking mergers that would tip a market into easy coordination, prosecuting cartels, and watching the practices that make tacit collusion simple to sustain.

Common exam mistakes

Treating oligopoly as a graph question. There is no single oligopoly diagram in the way there is a monopoly diagram. If a question gives a payoff matrix, answer it with dominant strategies and the Nash equilibrium, not with cost curves.

Saying prices in an oligopoly are always rigid. Rigidity is the prediction of one model under one assumption about rivals' reactions. Real oligopoly prices move, sometimes in fast price wars.

Confusing a cartel with an oligopoly. Oligopoly is a market structure. A cartel is one behavior that firms in that structure sometimes adopt. Cartels almost always form in oligopoly, because coordinating a few firms is far easier than coordinating many, but most oligopolies are not cartels.

Assuming collusion is stable because it is profitable. The joint profit is why firms want the agreement, and the private gain from defecting is why it breaks. Say both things.

Forgetting the fringe. A market can hold dozens of small firms and still be an oligopoly if a few firms dominate output. Count the share, not the head count.

Where to go next

Work through payoff matrices until finding a dominant strategy is automatic in the oligopoly module, then compare the structures directly with oligopoly against monopoly, oligopoly against perfect competition, and cartels against oligopoly. If the pricing side interests you more than the strategy side, the way firms with market power split customers into groups is covered in price discrimination. The one-line definition to keep is short: a few large interdependent firms, high barriers to entry, and no way to choose a price without predicting what the others will do.