Microeconomics › Section J · Monopolistic competition and oligopoly › Topic J.3

The prisoners' dilemma and Nash equilibrium

When everyone's best move depends on everyone else's, good individual decisions can add up to a bad collective result. The prisoners' dilemma - the most famous game in economics - shows exactly how, and the Nash equilibrium names the outcome no player can improve alone.

What a game is

In game theory, a game is any situation with three ingredients: players (the decision makers), strategies (the choices available to each), and payoffs (what each player gets for every combination of choices). The players are assumed to want the best payoff for themselves - self-interest, as always in this course, is a modelling assumption, not a moral one. Oligopoly pricing was the previous topic's unfinished business precisely because it is a game: each firm's best price depends on what rivals charge. Game theory, developed by mathematicians such as John von Neumann and John Nash in the mid-20th century, supplies the tools.

Reading a payoff matrix

Two-player games are drawn as a payoff matrix: one player's strategies are the rows, the other's are the columns, and each cell holds both players' payoffs for that combination. By convention, the row player's payoff comes first. The matrix turns strategic guessing into bookkeeping: for each column, find the row player's best row (their best response); for each row, find the column player's best column. Everything the next cards do - dominance, dilemmas, equilibria - is this lookup performed carefully.

Dominant strategies

A dominant strategy is one that is best for a player NO MATTER what the others do. Check the matrix column by column: if the same row wins against every column, that row is the row player's dominant strategy. When every player has one, prediction is trivial - each plays it - and no psychology is needed. But not every game has dominant strategies: in many, the best choice genuinely depends on what the other does, and the analysis has to work harder. The next card presents a game where dominance exists - and where it does the damage.

The prisoners' dilemma

Two suspects are questioned in separate rooms. If both stay silent, each gets 2 years. If one confesses while the other stays silent, the confessor goes free and the silent one takes 12 years. If both confess, each gets 6 years. Put yourself in one prisoner's head: if my partner stays silent, confessing gets me 0 years against 2; if my partner confesses, confessing gets me 6 against 12. Either way, confessing is better - it is each prisoner's dominant strategy. So both confess and each serves 6 years, even though both staying silent would have cost them only 2 each. Rational individual choice; worse joint outcome. That contradiction is the dilemma.

Worked example: verifying the dilemma

Verify the confession story with the matrix. 1. If the partner stays SILENT: confessing yields 0 years, silence yields 2 - confess. 2. If the partner CONFESSES: confessing yields 6 years, silence yields 12 - confess. 3. Since confessing wins in every column, it is each prisoner's dominant strategy - and by symmetry, both confess. 4. Result: 6 years each, a total of 12 person-years, versus the 4 they could have 'agreed' on by mutual silence. 5. Note the agreement's worthlessness: even if both promised silence beforehand, each would confess anyway - promises have no payoff matrix behind them. The structure, not the characters, produces the outcome: change the payoffs and the dilemma vanishes. Assumptions: no communication, no future retaliation (the repeated-game topic relaxes both).

The lesson: rational individually, wasteful together

The dilemma's lesson generalises far beyond crime. Cartels break down (the previous topic's cheating arithmetic IS a prisoners' dilemma in numbers). Countries overfish shared seas - each benefits from the extra catch, everyone bears the collapsed fishery. Arms races drain budgets - each side matches the other's weapons, and both end poorer but no safer. In each case, the individually rational move is the collectively damaging one, and cooperation fails unless something changes the payoffs: trust, enforcement, or repetition - the subject of the next topic.

Nash equilibrium

The dilemma's outcome has a formal name. A Nash equilibrium is a combination of strategies such that no player can improve their payoff by changing strategy ALONE, holding the others' choices fixed. In the confession game, (confess, confess) is a Nash equilibrium: each prisoner, given the other's confession, loses by switching to silence (6 years becomes 12). Mutual silence, by contrast, is NOT an equilibrium - each gains by confessing alone. John Nash proved that every game with finitely many players and strategies has at least one equilibrium if players may randomise - a result so central that he shared the 1994 Nobel memorial prize in economics for game-theory work. Equilibrium means stable, not good: both prisoners would prefer mutual silence to the dilemma's equilibrium.

Worked example: an advertising war

Two rival cafés choose to advertise or not. If neither advertises, each earns $8 million. If both advertise, each earns $4 million. If one advertises while the other abstains, the advertiser earns $11 million and the abstainer $2 million. 1. Check café A's best responses: if B advertises, advertising earns 4 against 2 - advertise; if B abstains, advertising earns 11 against 8 - advertise. 2. Advertising is each café's dominant strategy. 3. Equilibrium: both advertise, earning $4 million each. 4. Compare with cooperation: both abstaining would pay $8 million each - the dilemma again, $8 million per café left on the table. 5. The industry's fix: an agreement not to advertise - which is why trade associations want them and why antitrust law looks hard at them (a joint pact to restrict advertising can count as collusion). Assumptions: payoffs as stated; a one-shot choice, no reputation effects.

Games without dominant strategies

Not every game punishes cooperation. In a coordination game, players simply want to match: which side of the road to drive on, which software standard to adopt, whether to meet at the café or the library. Here there is more than one Nash equilibrium - both left or both right, for instance - and the problem is not selfishness but agreement. Societies solve coordination games with conventions and laws: drive on the left, or the right, but everyone the same. The contrast sharpens the dilemma's moral: in coordination games the equilibrium problem is choosing among good outcomes; in the prisoners' dilemma the only equilibrium is the bad one.

Common mistakes to avoid

First, calling every bad outcome a prisoners' dilemma - the term means a specific payoff structure where the individually rational choice harms both, not any coordination failure. Second, confusing a Nash equilibrium with the best outcome - it is the UNILATERALLY stable one; in the dilemma both players prefer mutual silence to the equilibrium. Third, forgetting the no-communication, one-shot assumptions - real prisoners' dilemmas soften when players interact repeatedly, which the next topic shows. Fourth, mixing up the two payoffs when reading a matrix - row player first, always. Fifth, expecting dominant strategies to exist everywhere: many games' best responses depend on the rival, and dominance is a special convenience.

Key takeaways

A game has players, strategies and payoffs - readable from a matrix, and the right language for oligopoly's interdependence. A dominant strategy wins against every rival choice: confessing is one (0 versus 2 years if silent, 6 versus 12 if not), so both confess - 6 years each though mutual silence cost 2 - and the cafés' advertising game repeats the shape ($4 million each in equilibrium, $8 million available by restraint). The stable outcome is the Nash equilibrium: no player gains by changing strategy alone - John Nash proved such equilibria always exist for finite games with randomisation, work recognised in the 1994 Nobel memorial prize. The dilemma explains cartel breakdown, overfishing and arms races: rational self-interest, collectively wasteful. Coordination games are the benign cousin - multiple equilibria, solved by convention - while the dilemma leaves only the bad one.

Flashcards

Flashcard

Game

Players, strategies and payoffs

The three ingredients.

Test your understanding

Question 1 of 5

In the confession game (silent/silent = 2 years each; one confesses = 0 and 12; both confess = 6 each), what is each prisoner's dominant strategy?

Question 2 of 5

Why is (confess, confess) a Nash equilibrium?

Question 3 of 5

Two cafés face the advertising game (both abstain = $8m each; both advertise = $4m each; one advertises = $11m and $2m). What happens?

Question 4 of 5

What distinguishes a coordination game from a prisoners' dilemma?

Question 5 of 5

Why do promises to cooperate fail in a one-shot prisoners' dilemma?

← Previous topicNext topic →