Title: Hide, Not Seek: Perceived Fairness in Envy-Free Allocations of Indivisible Goods

Abstract: Fair division provides a rich computational and mathematical framework for the allocation of indivisible goods, which has given rise to numerous fairness concepts and their relaxations. In recent years, much attention has been given to theoretical and computational aspects of various fairness concepts. Nonetheless, the choice of which fairness concept is in practice perceived to be fairer by individuals is not well understood. We consider two conceptually different relaxations of envy-freeness and investigate how individuals perceive the induced allocations as fair. In particular, we examine a well-studied relaxation of envy-freeness up to one good (EF1) which is based on counterfactual thinking that any pairwise envy can be eliminated by the hypothetical removal of a single good from the envied agent's bundle. In contrast, a recently proposed epistemic notion, namely envy-freeness up to $k$ hidden goods (HEF-$k$), provides a relaxation by hiding information about a small subset of $k$ goods. Through various crowdsourcing experiments, we empirically demonstrate that allocations achieved by withholding information are perceived to be fairer compared to two variants of EF1.