Justifying Voting Rules via Distance to Consensus

20 concepts7 papers1 proposal

In an ideal election, the winner is indisputable. If every voter ranks Candidate A first, Candidate A wins. If a candidate beats every other competitor in head-to-head matchups, that candidate is the clear Condorcet winner.

Real-world elections, however, are rarely so neat. Votes are messy, fragmented, and conflicting. When no clear consensus exists, how do we decide which voting rule—whether it is Plurality, Borda Count, or something more complex—is the "right" one to aggregate these preferences?

Distance rationalizability provides a powerful geometric intuition to solve this problem. Instead of viewing voting rules as arbitrary procedural algorithms, this framework treats them as optimization problems. It imagines every possible election as a point in a vast mathematical space. Within this space lie pockets of "ideal" elections where the winner is obvious. To find the winner of a messy, real-world election, we simply measure the distance from our current position to the nearest ideal election zone. The winner of that nearby ideal election is declared the winner of the real one.

[ Messy Real Election ]  --- (Minimize Distance) --->  [ Ideal Consensus Election ]
       (No clear winner)                                      (Obvious Winner: Candidate A)
                                                                           |
                                                                           v
                                                               Candidate A Wins Real Election

The Two Knobs: Consensus and Distance

To construct or understand a voting rule through the lens of distance rationalizability, you only need to configure two conceptual components:

  1. The Consensus Class: This defines what an "ideal" election looks like. As explored by Elkind et al. in Good Rationalizations of Voting Rules, consensus class constraints establish formal targets for total agreement, such as strong unanimity or the existence of a Condorcet winner.
  2. The Distance Metric: This defines how we measure the "messiness" of an election by calculating how many modifications are required to transform the actual election profile into a consensus profile.

By pairing different consensus classes with different distance metrics, entirely different voting rules emerge. For example, in Rationalizations of Condorcet-Consistent Rules via Distances of Hamming Type, Elkind et al. demonstrate that classic mechanisms like the Maximin rule and Young's rule can be rationalized using specific geometric distances.

To see this intuition in action, consider the voter replacement rule analyzed by the authors. This rule uses a Hamming-type distance metric, meaning it measures distance by the number of individual voter preference profiles that must be completely swapped out. A candidate wins under this rule if they become the Condorcet winner in an altered election profile that requires the absolute minimum number of voter replacements from the original state. By framing rules this way, the authors were able to prove that the voter replacement rule is fundamentally distinct from Young's rule, correcting prior misconceptions in social choice theory.

Why Rationalization Matters: The High Stakes of Rule Selection

Is distance rationalizability just an elegant mathematical exercise, or does the choice of metric actually matter?

It matters immensely because subtle changes to how we aggregate votes can completely flip the outcome of an election. In How Likely Are Two Voting Rules Different?, Yu et al. proved that even closely related voting rules have a surprisingly high, constant probability ($\Theta(1)$) of producing drastically different winners. The assumption that common voting rules are largely interchangeable is a myth; minor design choices in scoring vectors or distance metrics fundamentally alter how voter preferences are processed.

This sensitivity is not confined to political elections. It manifests across diverse domains of collective decision-making:

  • Human Teams: Even when individuals share similar underlying values, the choice of aggregation rule (such as unanimity versus majority voting) shifts which preferences dominate the final group outcome, as shown by Villeval.
  • Goal-Based Multi-Agent Systems: When agents vote on complex, interconnected propositional goals rather than simple candidates, different variants of majority rules yield conflicting collective outcomes (Novaro et al.).
  • AI Alignment: In systems designed to align artificial intelligence with human values using "artificial democracy," the composition of AI juror panels and their voting rules heavily dictate the system's ethical boundaries (Freedman).

Distance rationalizability provides a rigorous framework to justify why we should trust one rule over another amidst this volatility. For instance, Elkind et al. used it to bridge social choice with Maximum Likelihood Estimation (MLE)—a framework that treats votes as noisy signals of an objective truth. They discovered that rules like the Kemeny rule sit at the gold standard of rationalizability because they perfectly align with both geometric distance minimization and statistical truth-tracking, whereas other popular rules like the Single Transferable Vote (STV) are remarkably difficult to justify under either framework.

Generalizing Beyond Standard Ballots

The ultimate strength of the distance-based perspective is its adaptability. Because it relies on the abstract concept of "distance to a consensus," it can be extended to unconventional voting structures where traditional rules break down.

Consider liquid democracy, where voters can dynamically delegate their votes to proxies. In Pairwise Liquid Democracy, Brill et al. introduce a setting where voters delegate specific pairwise comparisons of alternatives rather than their entire ballot. This flexibility often creates messy, incomplete delegation graphs and cyclical, intransitive preferences.

Standard voting rules cannot process such structured, incomplete data. However, by applying a distance rationalization approach, the authors generalized rules like Kemeny's to operate directly on these delegation graphs. The mechanism works by calculating the minimal "cost" (in terms of altering internal pairwise preferences) needed to resolve the graph into a clear consensus profile.

By defining election rules as the shortest path to agreement, distance rationalizability transforms the messy reality of collective choice into a clean, justifiable geometric map.

Key Foundational Work

Rationalizations of Condorcet-consistent rules via distances of hamming type (2010) — contributed 7 concept elements to this direction.
Good Rationalizations of Voting Rules (2010) — contributed 2 concept elements to this direction.
Are Teams Less Inequality Averse than Individuals? (2014) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_distance_rationalizability_65e8e3c5,
  title = {distance rationalizability},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/distance-rationalizability/rdr_65e8e3c5354643488b2e55482ea8e423},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-21},
}

This exact version (version 1, frozen text):

@misc{lacuna_distance_rationalizability_65e8e3c5_v1,
  title = {distance rationalizability},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/distance-rationalizability/rdr_65e8e3c5354643488b2e55482ea8e423/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-21},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake