The Mathematics of Voting: Arrow's Theorem and Electoral Systems
An encyclopedic guide to the mathematics of voting — Arrow's Impossibility Theorem, electoral system design, Condorcet paradoxes, and the trade-offs embedded in democratic choice.
Mathematics and Democratic Choice
The mathematics of voting — formally called social choice theory — is the mathematical study of how individual preferences can be aggregated into collective decisions. It addresses questions that appear simple but reveal deep complexity: Given that voters have different preferences, what is the fairest way to determine an outcome? Can any voting system simultaneously satisfy all intuitive criteria of fairness? These questions sit at the intersection of mathematics, economics, and political philosophy. The field's most celebrated result — Arrow's Impossibility Theorem — demonstrates that no voting system can satisfy a set of seemingly reasonable criteria simultaneously, a finding with profound implications for the design of democratic institutions.
Basic Concepts
Social choice theory begins with a set of alternatives (candidates, options, policies), a set of voters, and each voter's preference ordering — a ranking of all alternatives from most to least preferred. A social welfare function takes all individual rankings as input and produces a collective ranking; a social choice function simply selects a winner.
Key desiderata for any voting system include:
- Pareto efficiency: If every voter prefers A over B, then A should collectively rank above B
- Independence of irrelevant alternatives (IIA): The collective ranking of A vs. B should depend only on how voters rank A relative to B — not on where a third candidate C stands
- Non-dictatorship: No single voter's preferences should always determine the collective outcome regardless of others' preferences
- Transitivity: If collectively A is preferred to B and B to C, then A should be preferred to C
- Universality (unrestricted domain): The system should work for any possible set of individual preference orderings
The Condorcet Paradox
Before Arrow, the Marquis de Condorcet identified a fundamental problem in the 18th century. A Condorcet winner is a candidate who beats every other candidate in pairwise majority comparisons. Condorcet argued such a candidate should win any election. However, majority voting over pairs of candidates can produce cyclical preferences even when all individual voters are rational (have transitive preferences).
Consider three voters and three candidates A, B, C:
- Voter 1 prefers A > B > C
- Voter 2 prefers B > C > A
- Voter 3 prefers C > A > B
Pairwise majority results: A beats B (Voters 1 and 3), B beats C (Voters 1 and 2), C beats A (Voters 2 and 3). The collective preference cycle A → B → C → A has no Condorcet winner. This Condorcet paradox (voting paradox) shows that majority rule over pairs can produce intransitive collective preferences from perfectly rational individual preferences — making the concept of a "will of the majority" ill-defined when more than two options exist.
Arrow's Impossibility Theorem
In 1951, economist Kenneth Arrow proved a landmark result: no social welfare function satisfying unrestricted domain, Pareto efficiency, independence of irrelevant alternatives, and transitivity can avoid being a dictatorship. In other words, any voting system (with three or more candidates) that meets four seemingly minimal fairness criteria must have a dictator — a voter whose preferences always become the collective outcome. Arrow received the Nobel Memorial Prize in Economic Sciences in 1972 partly for this work.
The theorem is not about any specific flaw in current systems; it is a mathematical proof that a perfect voting system — satisfying all natural axioms simultaneously — cannot exist with three or more candidates. Every real-world voting system violates at least one of Arrow's criteria. This forces system designers to choose which properties to sacrifice.
Major Electoral Systems Compared
| System | Method | Used By | Key Criterion Violated |
|---|---|---|---|
| Plurality (First Past the Post) | Candidate with most first-choice votes wins | US, UK, Canada, India | Fails Condorcet criterion; spoiler effect |
| Two-Round Runoff | Top two candidates advance to a runoff if no majority | France, Brazil | Still vulnerable to spoiler in first round |
| Ranked Choice Voting (IRV) | Eliminate lowest-ranked; redistribute votes iteratively | Australia, Alaska (US), Ireland | Fails monotonicity; can violate Condorcet criterion |
| Borda Count | Points assigned by rank position; highest total wins | Slovenia, Nauru; academic elections | Fails IIA; heavily influenced by irrelevant candidates |
| Condorcet Methods (e.g., Schulze) | Candidate who beats all others pairwise wins (if exists) | Some online elections, Wikimedia | Can produce ties when no Condorcet winner exists |
| Approval Voting | Voters approve any number of candidates; most approvals wins | Some US primaries, UN Secretary-General selection | Result depends on approval threshold strategy |
| Proportional Representation | Seats allocated proportional to vote share | Germany, Netherlands, Sweden | May require coalition formation; weaker majority government |
The Gibbard–Satterthwaite Theorem
A related result — the Gibbard–Satterthwaite theorem (independently proven by Allan Gibbard in 1973 and Mark Satterthwaite in 1975) — demonstrates that for any deterministic voting system with three or more candidates, either: (a) there is a dictator, (b) some candidate can never win regardless of votes, or (c) the system is manipulable — i.e., there exist situations where a voter can achieve a better outcome by voting insincerely (strategic voting) rather than according to true preferences.
This means that for any non-trivial voting rule, strategic voting cannot be entirely eliminated — voters will always sometimes have an incentive to misrepresent their preferences. The prevalence of tactical voting in plurality elections (voting for the "lesser of two evils" rather than a preferred third candidate) is the most visible practical expression of this theorem.
Proportionality vs. Majority Rule
A central design trade-off in electoral systems is between majoritarianism — where the majority's preference dominates — and proportionality — where the composition of a legislature reflects the full distribution of voter preferences.
| Property | Majoritarian Systems (FPTP, IRV) | Proportional Systems (PR, MMP) |
|---|---|---|
| Government formation | Typically single-party majority governments | Typically coalition governments |
| Minority representation | Often underrepresents minorities | Better reflects diverse preferences |
| Geographic representation | Strong local constituency links | Weaker in pure list PR systems |
| Policy stability | Higher (single-party control) | Variable (depends on coalition stability) |
| Voter wasted votes | High (votes for losing candidates count for nothing) | Low (most votes contribute to outcomes) |
Modern Developments
Contemporary social choice research has expanded Arrow's framework in several directions. Probabilistic voting rules can satisfy Arrow's criteria if the requirement of a deterministic outcome is relaxed. Judgment aggregation extends the framework beyond preferences to logical propositions. Computational social choice — combining social choice theory with computer science — addresses questions of algorithmic winner determination, preference elicitation, and the complexity of manipulation. Experiments with novel systems such as quadratic voting (where voters purchase votes at quadratically increasing cost, allowing intensity of preference to be expressed) and liquid democracy (where voters may delegate their vote to a trusted proxy) represent ongoing efforts to find better practical solutions within the boundaries Arrow's theorem defines.
Related Articles
applied mathematics
What Is Statistics? Descriptive, Inferential, Probability, and the Science of Data
A comprehensive introduction to statistics — descriptive vs. inferential statistics, probability and distributions, hypothesis testing, p-values, confidence intervals, correlation vs. causation, common statistical errors, and why statistical literacy is essential for understanding research and data.
8 min read
applied mathematics
Probability Theory Explained: Fundamentals, Rules, and Real-World Applications
A clear introduction to probability theory — from basic definitions and rules to conditional probability, Bayes' theorem, and how probability underpins everything from medicine to machine learning.
8 min read
applied mathematics
How Algorithms Work: Logic, Efficiency, and Applications
Understand what algorithms are, how they are designed and analyzed, key algorithm types including sorting and searching, and their role in modern computing.
8 min read
applied mathematics
How Fractals Work: Self-Similarity in Mathematics and Nature
Explore the mathematics of fractals — self-similar geometric patterns with fractional dimensions, from the Mandelbrot set to coastlines and biological systems.
8 min read