Why Arrow's Impossibility Theorem Structurally Mandates That No Ranked-Preference Voting System Can Satisfy All Fairness Criteria

Every time you cast a ranked ballot in a local or national election, you are participating in a mathematical compromise that has already decided which democratic ideal to sacrifice. The assumption that a society can perfectly aggregate individual preferences into a single, fair collective decision is mathematically false. This is not a political failing or a consequence of partisan polarization, but a structural boundary proven by economists.[7]

The boundary was defined in 1951 by American economist Kenneth J. Arrow in his book Social Choice and Individual Values. Arrow demonstrated that when voters face three or more options, no ranked voting system can convert their individual preferences into a collective ranking while satisfying a specific set of basic fairness conditions.[1][4]

Arrow, who received the Nobel Memorial Prize in Economics in 1972 for this work, established five criteria that any rational and fair democratic system should theoretically meet. The first is Unrestricted Domain, meaning the system must be able to process any combination of voter preferences without breaking down. The second is Social Ordering, which requires the system to produce a clear ranking from best to worst, avoiding endless ties or cyclical loops.[1][3]

The third criterion is the Weak Pareto principle. This simply dictates that if every single voter prefers candidate A over candidate B, the final social ranking must also place A above B. The fourth is Non-Dictatorship, ensuring that no single voter's preferences dictate the outcome regardless of what anyone else wants.[1][4]

The final and most contentious criterion is the Independence of Irrelevant Alternatives (IIA). This principle states that the collective preference between two candidates should depend only on how voters rank those two specific candidates, not on how they rank a third, unrelated option. If society prefers A to B, introducing a new candidate C should not suddenly cause society to prefer B to A.[1][6]

Arrow's Impossibility Theorem proves that these five conditions are mutually exclusive. If a voting system satisfies the first four criteria, it will inevitably violate the Independence of Irrelevant Alternatives. If it satisfies IIA and the others, it must be a dictatorship.[1][5]

The root of this impossibility traces back to 1785, when the Marquis de Condorcet discovered the paradox of voting. Condorcet realized that majority preferences can be irrational even when individual voters are perfectly rational. If one third of an electorate prefers candidate X to Y to Z, another third prefers Y to Z to X, and the final third prefers Z to X to Y, a structural nightmare emerges.[3]

In that exact scenario, a two-thirds majority prefers X to Y, a two-thirds majority prefers Y to Z, and a two-thirds majority prefers Z to X. The result is a cycle where no candidate can definitively win a head-to-head matchup against all others. Condorcet's paradox showed that pairwise majority voting fails the Social Ordering criterion because it cannot produce a transitive result.[3][5]

Modern electoral systems attempt to bypass this paradox, but Arrow's theorem guarantees they only trade one failure for another. Plurality voting, where citizens simply choose one candidate, frequently violates the Independence of Irrelevant Alternatives by allowing spoiler candidates to split the vote and change the outcome between the frontrunners.[4][5]

Ranked-choice voting, specifically the instant-runoff method adopted by several municipalities, solves the spoiler effect in many scenarios but still mathematically violates IIA. Under instant-runoff rules, the elimination of a minor candidate can alter the sequence of vote transfers in a way that changes the relative standing of the remaining candidates, meaning the presence of an irrelevant alternative still dictates the winner.[5]

The Borda count, another proposed alternative where voters assign points to candidates based on their rank—for example, 3 points for first place, 2 for second, 1 for third—also fails IIA. Because the point values are relative to the total number of candidates, adding or removing a candidate shifts the mathematical weight of every other ranking on the ballot.[3][5]

Arrow's framework relies heavily on the assumption that voters can only provide ordinal information—ranking candidates first, second, and third—without measuring the intensity of those preferences. As Arrow himself wrote, "interpersonal comparison of utilities has no meaning and … there is no meaning relevant to welfare comparisons in the measurability of individual utility."[3]

This strict ordinalist approach is what makes the theorem so absolute. By refusing to weigh how much a voter loves their first choice compared to their second, the system is blind to the intensity of preference. Some economists argue that this restriction is the true flaw, suggesting that cardinal voting systems—where voters rate candidates on a scale of 1 to 10—can bypass Arrow's constraints entirely because they do not rely on relative rankings.[1][6]

However, within the realm of ranked preferences, the impossibility remains absolute. As the theorem dictates, any method that uses voters' ranked choices to decide an election with a single winner will violate at least one fairness criterion. Consequently, in any election, a losing candidate can point to a specific mathematical property of the voting system and correctly claim that the outcome is structurally unfair.[5]

The enduring legacy of Arrow's theorem is not that democracy is futile, but that electoral design is an exercise in choosing which flaws a society is willing to tolerate. Policymakers must decide whether it is more important to prevent spoiler candidates, ensure majority consensus, or maintain strict transitivity, knowing that achieving all of them simultaneously is a mathematical fiction.[1][7]