Kenneth Arrow

Kenneth Joseph Arrow was born in New York on 23 August 1921 and died in Palo Alto, California, on 21 February 2017. He studied at City College and Columbia, served as a weather officer in the Second World War, and from 1946 combined graduate work at Columbia with research at the Cowles Commission in Chicago and at the RAND Corporation. A question that arose at RAND — how to define a coherent collective preference for a group of actors — led to his doctoral thesis, published in 1951 as Social Choice and Individual Values.

He taught mainly at Stanford and at Harvard. In 1972 he shared the Sveriges Riksbank Prize in Economic Sciences with John Hicks 'for their pioneering contributions to general economic equilibrium theory and welfare theory'. Besides social choice he made foundational contributions to general equilibrium (the Arrow–Debreu model), the economics of uncertainty and information, health economics and the theory of organisations.

• The impossibility theorem. Arrow asked whether there is a rule that turns any combination of individual rankings of alternatives into a consistent social ranking. He showed that with three or more alternatives no rule satisfies all of the following: it works for every logically possible profile of preferences (unrestricted domain); it always yields a transitive social ordering; if everyone prefers x to y, so does society (weak Pareto); the social ranking of x and y depends only on individual rankings of x and y (independence of irrelevant alternatives); and no individual is a dictator whose strict preferences always prevail.
• Generalising Condorcet. The theorem generalises Condorcet's paradox, in which majority voting over three options can produce a cycle. Arrow showed that the problem is not peculiar to majority rule: every ranking-based procedure must give up one of the conditions.
• Escape routes. The impossibility can be avoided by restricting preferences (Duncan Black showed that single-peaked preferences guarantee consistent majority outcomes), by using richer information than rankings — interpersonally comparable utilities or graded evaluations — or by relaxing transitivity or independence.
• Limits of organisation. In The Limits of Organization (1974) Arrow analysed how organisations economise on information, why authority and codes are needed, and how trust functions as an economic good.

Arrow's theorem founded modern social choice theory and changed how voting systems are studied. Every comparison of plurality, ranked-choice (instant-runoff), Condorcet, Borda, approval and score voting now proceeds by asking which of Arrow-type conditions each method satisfies and which it violates. Later results — the Gibbard–Satterthwaite theorem on strategic voting and Amartya Sen's 'liberal paradox' — extend the analysis.

The theorem also feeds a long debate about the meaning of democracy. William Riker's Liberalism Against Populism (1982) used it to argue that there is no coherent 'will of the people' and that democracy can at most allow voters to remove officials. Deliberative democrats and social-choice theorists such as Christian List and John Dryzek reply that deliberation can create structured, near single-peaked preferences, and others argue that evaluative ballots (approval, score or majority judgement) escape the theorem's framework. For designers of voting rules, DAOs, quadratic voting or AI-mediated preference aggregation, the theorem is a reminder that every aggregation rule embodies trade-offs that should be chosen openly.

• Practical relevance. Some scholars argue that cycles and paradoxes are rare in real electorates, so the theorem's practical significance is overstated; others counter that agenda control and strategic manipulation exploit them.
• Strength of conditions. The independence condition is often judged too strong, because it forbids using information about intensity or about other alternatives that seems relevant to fair collective choice.
• Ordinal framework. Critics of the ordinal, preference-based framework, including Sen, argue that welfare judgements need richer information than individual rankings.
• Over-reading. Philosophers caution against inferring from the theorem that democracy is meaningless; it shows only that no rule meets all of a specific set of conditions.

• 'A Difficulty in the Concept of Social Welfare', Journal of Political Economy 58(4) (1950) — DOI
• Social Choice and Individual Values (1951; 2nd ed. 1963)
• 'Uncertainty and the Welfare Economics of Medical Care', American Economic Review (1963)
• The Limits of Organization (1974)
• Overview: Michael Morreau, 'Arrow's Theorem', Stanford Encyclopedia of Philosophy; Nobel Prize facts