In classical social choice theory, participation is usually treated as a "free" action. Voters are assumed to cast ballots regardless of their preference strength, and candidates are assumed to be power-hungry agents who will run for office the moment they have a chance to win.
However, a growing body of research suggests that this "cost-free" assumption leads to overly pessimistic or unrealistic models. By incorporating the cost of participation—the time, effort, or money required to vote, run for office, or become informed—researchers have found that social choice systems behave in fundamentally different, and often more stable, ways.
The Pivotality Condition
The core intuition behind costly participation is the Calculus of Voting. In this framework, an agent only participates if the expected benefit of their action outweighs the cost. Because the benefit of a single vote or candidacy is only realized if that action changes the outcome, agents must weigh the cost against their probability of being "pivotal."
Fabricio Vasselai bridges this concept with iterative voting in the Iterative Calculus of Voting (ICV) framework. In ICV, voters don't just vote for their favorite; they simultaneously update their choices based on the likelihood that their vote will actually matter. This mechanism naturally leads to stable equilibria, often converging to a two-candidate race as predicted by Duverger’s Law, because voters realize that "wasting" effort on a non-pivotal candidate is irrational.
Why Abstention Improves Social Outcomes
If voting is costly, voters with weak preferences or those who feel "repulsed" by all viable candidates (Sylvain Bouveret) will simply stay home. While this might seem like a failure of democracy, it can actually lead to better social outcomes.
In traditional models where everyone must vote, the "distortion"—the gap between the winner's social cost and the ideal candidate's cost—is at least 3. However, Mohammad Ghodsi, Masoud Seddighin, and Mohamad Latifian proved in On the Distortion Value of the Elections with Abstention that when unmotivated voters abstain, the distortion drops to roughly 1.5. Essentially, costly participation acts as a filter: only those who feel strongly about the outcome participate, which naturally pushes the system toward candidates who minimize overall social dissatisfaction.
Reluctant Candidates and Stability
The cost of participation applies to candidates as well. Traditional models assume candidates are always willing to run, but Svetlana Obraztsova, Maria Polukarov, and Zinovi Rabinovich introduce the concept of "reluctance" or "laziness" in Strategic candidacy with reluctance.
A reluctant candidate only enters the race if their presence is pivotal to the outcome. This creates a more complex strategic landscape:
- Filtering: Every stable equilibrium in a costly model is also stable in a cost-free model, but many cost-free equilibria disappear because candidates would rather "stay home" than run in a race they don't influence.
- Instability: Reluctance can destroy the stability of an election, leading to cycles where candidates enter and exit based on the current field. Finding a stable set of candidates in this environment is a computationally hard (NP-complete) problem.
Costs as a Defense and a Driver
Participation costs aren't always a hurdle to be overcome; they can be a design feature. In online environments, "Sybil attacks" (where one person casts many votes) are a major threat. Liad Wagman and Vincent Conitzer showed in Optimal False-Name-Proof Voting Rules with Costly Voting that even a tiny cost for casting extra votes makes it possible to design "false-name-proof" rules that are highly responsive to voter preferences.
Similarly, in Liquid Democracy, the cost of becoming an informed voter drives the system's core mechanic: delegation. Daan Bloembergen, Davide Grossi, and Martin Lackner model delegation as a rational trade-off in On Rational Delegations in Liquid Democracy. Agents delegate their votes not just to increase accuracy, but to avoid the personal "effort cost" of direct participation. This rational avoidance of cost shapes the entire network structure of the democracy, determining whether the collective outcome is more or less accurate than a standard vote.