Introduction
In modern representative democracies, citizens do not vote on every individual policy or law. Instead, they elect a committee—such as a parliament or a city council—to act as their proxy in future decision-making. A fundamental challenge in social choice theory is ensuring that this elected committee makes decisions that align with the preferences of the electorate, even when those future issues are unknown at the time of the election.
This paper investigates the "distortion" problem: the potential loss of utility that occurs because election rules typically only have access to voter rankings of candidates, rather than the underlying issue-by-issue preferences that drive those rankings. The authors ask: which multiwinner election rules result in a committee whose majority-rule decisions best align with the collective will of the voters?
The Voting Committee Model
To analyze this, the authors propose a formal "Voting Committee Model" with the following components:
- Binary Issue Space: Voters and candidates are represented as binary vectors (0s and 1s) across $p$ issues.
- Preference Derivation: A voter prefers candidate A over B if A agrees with the voter on a greater number of issues.
- Committee Decision-Making: Once elected, the committee resolves each issue using a simple majority rule.
- Utility Metrics: Performance is measured using two primary lenses:
- Utilitarian: Maximizing the total number of "correct" decisions across all voters.
- Proportional (Nash Welfare): A metric that favors committees providing balanced satisfaction across different segments of society, thereby reducing inequality.
Theoretical Insights: The "Infinity" Barrier and Polarization
The authors first establish a sobering theoretical baseline: in the most general case, the distortion of any ordinal voting rule is infinite. This means that without assumptions about societal structure, a rule relying solely on rankings can theoretically select a committee that makes the worst possible decisions for the electorate.
However, the problem simplifies significantly in polarized societies. When a society is strictly divided into two ideological camps, the $k$-Copeland rule—which selects candidates who win the most head-to-head matchups—is mathematically optimal, achieving a distortion of 1.
Simulation Analysis and the Robustness of STV
Beyond theoretical worst-case bounds, the authors conducted extensive computer simulations using various probabilistic models, such as the $t$-Poles model, where society is divided into $t$ different ideological groups.
A key finding is the trade-off between majoritarian and proportional goals. While $k$-Copeland excels in simple polarized settings, the Single Transferable Vote (STV) demonstrates remarkable robustness. Although designed for proportionality, STV consistently outperforms other rules in more complex societies with multiple ideological poles ($t > 2$), maintaining high utility across both utilitarian and proportional metrics.
Key Results Summary
- General Distortion: For any committee size $k \geq 3$, the distortion of any ordinal rule is theoretically $\infty$.
- Polarized Societies: $k$-Copeland is optimal (distortion = 1), while STV's distortion approaches 1 as committee size grows. In contrast, rules like SNTV perform poorly as the population size increases.
- STV’s Versatility: Simulations show that STV is a superior choice when the goal is to ensure diverse group representation (Nash Welfare) without sacrificing significant overall utility.
Conclusion and Future Directions
The research highlights that while no voting rule is perfect in every scenario, $k$-Copeland and STV are powerful tools for representative democracy. $k$-Copeland is ideal for maximizing total satisfaction in polarized environments, while STV provides a robust balance of proportionality and utility in complex, multi-faceted societies.
Future work may explore more nuanced policy spaces beyond binary issues and investigate how internal committee dynamics—such as bargaining and logrolling—affect the representativeness of the final decisions.
Summary updated