The multiple-volunteers principle

Abstract

We present a class of simple transfer-free rules for assigning an unpleasant task among a group of agents: agents decide simultaneously whether or not to volunteer; if the number of volunteers exceeds a threshold number, the task is assigned to a volunteer; otherwise, the task is assigned to a non-volunteer. In particular, the rule may ask for multiple volunteers although one agent is sufficient to perform the task. In a setting in which agents care about who performs the task, any multiple-volunteers rule yields a strict interim Pareto improvement over random task assignment. Some volunteers rule is utilitarian optimal across all transfer-free binary mechanisms, and a rule with a large threshold reaches the first-best approximately if the group is large. Similar results hold for the problem of assigning a pleasant task. In that case, the task is assigned to a volunteer if and only if there are sufficiently few volunteers.