The paradox
Anthony Downs's An Economic Theory of Democracy set out the problem: the benefit of voting is the value of your preferred outcome multiplied by the probability your vote is decisive, minus the cost of voting. In any large election that probability is around one in tens of millions.
Riker and Ordeshook added a term for the satisfaction of performing a civic duty, which makes the equation fit the data by construction. That is the standard move, and it is also the standard criticism: a free parameter that absorbs whatever the model gets wrong.
The empirical regularity is that turnout responds to cost in exactly the way the model predicts — rain, distance to the polling station and registration deadlines all move it — while the level of turnout is far above what pure instrumentality implies.
The four families of explanation
Expressive voting (Brennan and Lomasky) treats the ballot as a way to express identity or values, at near-zero cost, because it will not change anything. On this account voters are not badly informed by accident; they are uninformed because the act is not instrumental.
Altruistic or pivotal voting (Edlin, Gelman and Kaplan) argues that a vote has social value even when it is not decisive, because it marginally raises the chance of the better outcome for everyone — so a small probability of a large benefit can justify the cost.
Group-based models treat turnout as a cooperative act within a group, where social rewards and sanctions operate and where the individual calculus is the wrong unit of analysis.
Behavioural accounts emphasise habit, norms and the doorstep conversation — the fact that turnout is sticky across a life and contagious within a household.
Cost matters: the strongest causal evidence comes from natural experiments — rain on election day reduces turnout, especially among marginal voters, and same-day registration raises it.
Habit matters: voting in one election strongly predicts voting in the next, which no single-shot cost-benefit model explains.
Duty is real but measured badly: self-reported civic duty correlates with turnout, but it is also the kind of variable that can explain any result after the fact.
Policy stakes: because marginal voters differ systematically from habitual ones, who turns out changes who wins — which is why the paradox is not merely academic.
The archive labels this live debate because the phenomenon is well measured and the explanation is genuinely unresolved.
Why it is in the archive
Because it is the cleanest case of a rational-choice model failing at its own game, and because the failure has real consequences for how democracies are designed.
Related: the median voter theorem, superrationality.
Sources for this page
- Wikipedia, “Paradox of voting” — the calculus and its criticsen.wikipedia.org
- Downs (1957), An Economic Theory of Democracyen.wikipedia.org
- Edlin, Gelman & Kaplan (2007), “Voting as a Rational Choice”doi.org
- Gomez, Hansford & Krause (2007), Journal of Politics — the effect of rain on turnoutdoi.org