Arrow's Impossibility Theorem illustration
Mathematical theorem / social choice theory
Mathematical theorem / social choice theory

Arrow's Impossibility Theorem

When three or more options are on the table, every voting system has trade-offs. You can improve fairness on one dimension, but not satisfy every reasonable fairness rule at once.

Popularity
Usefulness
Aliases
Arrow's Theorem / Arrow's Paradox / General Possibility Theorem / Impossibility Theorem
Domains
Economics, political science, voting theory, welfare economics, decision theory

Definition

  • Arrow's Impossibility Theorem says that with at least three options, no ranked voting rule can always turn individual preferences into a group ranking while meeting a full set of appealing fairness conditions at the same time.

Core Idea

  • A perfectly fair voting system for ranked preferences is impossible under Arrow's conditions.
  • The theorem does not say democracy is useless.
  • It says every collective decision rule must make trade-offs: for example, it may allow cycles, ignore some information, violate independence, restrict possible preferences, or behave like a dictatorship in the technical sense.

How It Works

  • Start with voters who rank the available options.

  • Then choose a rule that converts those personal rankings into one social ordering.

  • Arrow's result shows that no such rule can always keep all of these goals together:

  • Open preference domain: the method must handle any logically possible ranking pattern.

  • Unanimity / Pareto agreement: if everyone ranks A above B, the group should also rank A above B.

  • Independence of irrelevant alternatives: the social choice between A and B should depend on views about A and B, not on a third option entering or leaving the field.

  • No dictator: one voter cannot automatically determine the outcome every time.

  • Consistent group ordering: the final ranking should not contradict itself.

  • Once there are three or more options, you cannot guarantee all of these conditions simultaneously.

Usage Example

  • Suppose a team must choose between three project plans: A, B, and C.
  • Some members rank A > B > C.
  • Others rank B > C > A.
  • Others rank C > A > B.
  • Pairwise majority voting can produce a cycle: A beats B, B beats C, and C beats A.
  • This shows why a group can appear to have inconsistent preferences even when each individual voter is internally consistent.

Famous Example

  • Example: The Condorcet voting cycle with three alternatives.
  • Why it fits this rule: It illustrates how majority rule can generate circular collective preferences, which Arrow's theorem generalizes to a broader impossibility result for social choice rules.

Use Cases / Situations Where It Applies

  • Designing voting systems with ranked ballots.
  • Comparing electoral systems.
  • Understanding why no voting method is perfectly fair in every situation.
  • Analyzing committee decisions, public policy choices, constitutional design, and welfare economics.
  • Explaining why changing voting rules can change outcomes.

When Not to Use or Common Misuse

  • Do not use it to claim that all voting is meaningless.
  • Do not use it for simple two-option decisions; the theorem requires at least three alternatives.
  • Do not apply it directly to voting systems that do not use full ranked preferences unless the assumptions are carefully checked.
  • Do not treat it as an empirical psychological effect; it is a mathematical theorem.
  • Do not confuse it with the Condorcet paradox. The Condorcet paradox is an example of cyclic majority preferences; Arrow's theorem is a broader formal impossibility result.

Rule Invention / Origin

  • Invented by: Kenneth J. Arrow
  • Year of invention: 1950 for the paper "A Difficulty in the Concept of Social Welfare"; 1951 for the book Social Choice and Individual Values
  • Country / context of origin: United States; welfare economics and social choice theory

Short Practical Takeaway

  • No ranked voting system is perfect. If there are three or more options, any voting rule must sacrifice at least one reasonable fairness condition.