Skip to content
Tutorials Center
My learning

Topics

Lesson 12 of 37 ยท Nash Equilibrium

Pareto Efficiency and Social Dilemmas

Learn Pareto efficiency and social welfare, measure the price of anarchy, and simulate the public goods game and tragedy of the commons to see why equilibria can be inefficient.

A Nash equilibrium tells you what is stable. It does not tell you what is good. To judge whether an outcome is good for the group, we need a second idea: efficiency. Economists use a standard of efficiency named after Vilfredo Pareto. This lesson defines it, measures how much an equilibrium loses compared with the best possible outcome (the "price of anarchy"), and applies the ideas to social dilemmas such as the public goods game, the tragedy of the commons and the prisoner's dilemma.

You should know how to find equilibria from Pure Nash Equilibrium. Here we put those equilibria next to the other outcomes of the game and ask whether everyone could have done better.

Pareto dominance and efficiency

Outcome x Pareto dominates outcome y if everyone is at least as well off in x as in y, and at least one person is strictly better off. An outcome is Pareto efficient (or Pareto optimal) if no other outcome Pareto dominates it. In a Pareto efficient outcome you cannot make someone better off without hurting somebody else.

Efficiency is a weak but useful notion. It says nothing about fairness: giving one player everything can be Pareto efficient. But the opposite, inefficiency, is a real defect: if an outcome is not Pareto efficient, then there exists a change that makes someone better off and nobody worse off, and the players leave value on the table.

Python Find Pareto-efficient outcomes

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Three of the four cells are efficient: (C, C), (C, D) and (D, C). The only inefficient cell is (D, D), which is dominated by (C, C) because both players prefer the latter. And (D, D) is the only Nash equilibrium. That is the definition of a social dilemma in one line: the equilibrium is not efficient.

Social welfare and the price of anarchy

To compare outcomes with a single number, we often use social welfare, the sum of the players' payoffs. (The sum assumes the payoffs are comparable across players; use it with care.) The optimum is the largest welfare over all outcomes. The price of anarchy is the ratio of the optimal welfare to the welfare of the worst Nash equilibrium (for payoffs that are positive costs the ratio is flipped). A price of anarchy of 1 means selfishness costs nothing. A large value means selfish behavior wastes a lot.

For games with positive welfare, price of anarchy = optimal welfare / worst equilibrium welfare.

Python Price of anarchy in the classic games

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

In the prisoner's dilemma the optimum is 6 (both cooperate) and the only equilibrium gives 4, so the price of anarchy is 1.5. In the stag hunt the best equilibrium reaches the optimum, but the worst one, (Hare, Hare), gives 6 instead of 8: the loss comes from a coordination failure, not from the structure of incentives alone. Chicken's equilibria give 6 against an optimum of 6 (the cooperative (Swerve, Swerve) gives 6 and so do the two asymmetric equilibria), so the ratio is 1.

The public goods game

Many real dilemmas involve more than two people. In a public goods game, n players each have an endowment of 10 coins. Each chooses how much to contribute to a common pot. The pot is multiplied by a factor m (with 1 < m < n) and shared equally. Your payoff is what you keep plus your share of the pot:

payoff_i = (10 - c_i) + (m / n) * sum of all contributions.

Each coin you contribute costs you 1 and returns only m/n < 1 to you, so withholding is the dominant strategy: contribute nothing. But if everybody contributes everything, each gets 10 m, which is more than 10. The group is better off with full contribution; yet the only Nash equilibrium is zero contribution.

flowchart LR
    C["Each player contributes c_i"] --> P["Pot x multiplier m"]
    P --> S["Share m/n of the pot to each"]
    S --> K["Keep what you did not contribute"]
    K --> R["Payoff = kept + share"]

Each coin you give returns less than a coin to you, so selfishly you give nothing.

Python A public goods game

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The best reply is 0 in both cases: even if everyone else contributes fully, you gain by keeping your coins (you earn 25 as a free rider instead of 20). So the unique equilibrium is zero contribution, with payoff 10 each, while full contribution would give 20 each.

Tragedy of the commons

A commons is a shared resource such as a fishery or a pasture. Each user gains the full benefit of taking more but shares the cost of depletion with everyone. The structure is the same as the public goods game in reverse: using the commons is the dominant action and everyone loses when everyone does it. The word "tragedy" refers to the fact that nobody is evil; each acts rationally.

Python A fishing commons

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The numbers show the usual pattern: at equilibrium each fisher catches more and earns less than at the group optimum. Solutions in the real world change the game: private ownership, quotas, taxes, or reputation that turns a one-shot dilemma into a repeated one. Repeated games show how long-run relationships can sustain cooperation without outside rules.

Note: Pareto efficiency and a Nash equilibrium can coincide. In many games, such as the harmony game or the market in a well-designed auction, the stable outcome is also efficient. The point of the dilemma games is that this match is not automatic.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Outcome x Pareto dominates outcome y when...

    Pareto dominance needs no loser and at least one strict winner.
  2. 2. In the prisoner's dilemma, which statement is correct?

    (Defect, Defect) is the unique equilibrium, and both players prefer (Cooperate, Cooperate).
  3. 3. In a public goods game with multiplier m < n, why is contributing nothing a dominant strategy?

    Contributing costs you one coin and returns you m/n of a coin, which is less than one, no matter what the others do.
  4. 4. If a game has a price of anarchy of 1, what does that mean?

    The ratio of the optimum to the worst equilibrium is 1 only when selfish play loses nothing.

Summary

  • An outcome is Pareto efficient if nobody can be made better off without making someone worse off.
  • A social dilemma is a game whose Nash equilibrium is Pareto dominated by another outcome.
  • The price of anarchy measures how much welfare selfish equilibrium play loses against the optimum.
  • Public goods games and commons problems are n-player prisoner's dilemmas with dominant free riding.
  • Changing the rules or the repetition of the game can restore efficiency.

Next: Multiple Equilibria and Selection