Maximal domain of preferences in the division problem (Q700081): Difference between revisions

In 1991 [Econometrica 59, 509-519 (1991; Zbl 0721.90012)] \textit{M. Sprumont} showed that, given a fixed amount \(M\) of a perfectly divisible good, if agents have single-peaked preferences over their shares, then the `uniform' allocation rule is the unique strategy-proof, efficient and anonymous rule. In this paper the authors show that there is a unique maximal domain of preferences, that includes the set of single-peaked preferences, for which there is at least one rule satisfying strategy-proofness, efficiency and strong symmetry. This maximal domain is characterized as the set of -- what the authors call -- feebly single-plateau preferences. This maximal domain depends crucially on both \(M\) and the number \(n\) of agents and includes only preferences whose set of best shares is an interval and which are weakly monotonic on an interval \(\Theta\), defined by the relative position of \(M\)/\(n\) and the set of best shares. This domain departs from the single-peaked domain in two respects: shares outside \(\Theta\) can be ordered freely and special intervals of indifference are allowed on \(\Theta\). This set of feebly single-plateaued preferences, given \(M\) and \(n\), is (much) larger than the set of single-plateau preferences studied by H. Moulin (1984) and D. Berga (1998), since the latter are strictly monotonic on both sides of the plateau. Furthermore, the intersection of all the maximal domains, when \(M\) varies from 0 to infinity, coincides with the single-plateaued domain.