An oligopoly is a market run by a few firms: a handful of airlines on a route, a few wireless carriers, the two stations at the intersection. What makes it its own structure is not the headcount but the consequence of the headcount — mutual interdependence. When there are only a few sellers, each one is big enough that its move visibly moves the others. Drop your price and rivals feel it; raise it and they notice. You cannot pretend you are alone.

That single feature sets oligopoly apart from every structure so far. A monopolist could ignore rivals because it had none. A perfectly competitive firm could ignore rivals because there were too many to matter — one farmer's choice is a drop in the ocean. The oligopolist sits in the uncomfortable middle: few enough that each rival matters, too few to disappear into a crowd. So before it can choose anything, it has to ask a question the other structures never faced: what will they do if I do this?

This book has run on one idea since the first chapter: people respond to incentives. Oligopoly is where that idea folds back on itself. Your incentive now depends on your rival's incentive, which depends on yours. You can't read your best move off a cost curve, because the payoff to cutting your price is enormous if the rival holds steady and ruinous if the rival cuts too. The decision is strategic: it has to account for a thinking opponent who is, at this very moment, accounting for you.

Think back to the two gas stations. Holding the high price together is comfortable and profitable. But each owner can see a private prize: undercut while the other holds, and capture the whole street. That temptation is not a flaw in their character; it is the structure of their incentives. And because both owners see the same prize, both reach for it, and the comfortable outcome collapses into the price war neither wanted. To reason about this clearly — to see why rational moves lead somewhere irrational — we need a tool built exactly for interdependent incentives.

A payoff matrix lays out every combination of choices and the reward each player gets. Two firms, two choices each — hold the high price (cooperate) or cut it (cheat) — give four outcomes. Read each cell as the profit to Firm A and the profit to Firm B.

Now find each firm's best response. Look at Firm A. If B holds the high price, A earns $10 by holding but $14 by cutting — cutting wins. If B cuts, A earns $2 by holding but $5 by cutting — cutting wins again. Cutting is better for A no matter what B does, so it is A's dominant strategy. The game is symmetric, so cutting is B's dominant strategy too.

Both cut. The outcome is the bottom-right cell, $5 each — a Nash equilibrium, because neither firm can improve by changing its own choice while the other keeps cutting. Yet both firms would earn $10 in the top-left cell if they could just both hold. That gap — rational individual choices producing a worse shared outcome — is the prisoner's dilemma. It is the most important idea in this chapter, and it reaches far past gas stations: arms races, overfishing, doping in sports, and climate negotiations all share its shape.

If competition drives firms to the bad $5 cell, the obvious fix — for the firms, not for consumers — is to agree to sit in the good $10 cell. That agreement is collusion, and a formal version is a cartel: rivals coordinating to behave like a single monopolist, splitting the monopoly profit. In most places this is illegal precisely because it recreates monopoly harm without monopoly's occasional justifications.

But notice what the payoff matrix already told us: the cooperative cell is not a Nash equilibrium. The instant both firms agree to hold the high price, each one faces the same private temptation — cut a little, capture the street, jump from $10 to $14 while the partner is still playing nice. Every member of a cartel has, built into the deal, a personal incentive to break it. That is why cartels are chronically unstable. The same incentive that makes cooperation attractive is the incentive that detonates it.

So why do some cartels last? Because the gas-station story repeats. When the same firms meet day after day, cheating today invites retaliation tomorrow. A simple strategy — tit-for-tat: cooperate, then do whatever your rival just did — turns the long-run incentive around. Cheat once and you trade a single big day for a lasting price war. When the future matters enough, the threat of retaliation can hold the high price together. This is why the answer to “does oligopoly end in cooperation or in a price war?” is, honestly, it depends on the game — on how many firms, how visible their moves, how patient they are, how often they meet. Oligopoly has no single predicted outcome, and that indeterminacy is itself the finding.