Resource Allocation with Ambiguous Needs
Speaker: Jingyi Xue (Singapore Management University)
Date: Sept 16, Wednesday, 2020
Time: 12:00 pm - 1:00 pm
Venue: Zoom (ID: 92908197751, you may join the zoom meeting via https://nyu.zoom.us/j/92908197751)
Abstract:
A group of agents have ambiguous needs on a one-dimensional and perfectly divisible resource, and the resource has to be allocated before ambiguity resolves. For example, an international emergency management organization distributes rescue forces/medical supplies among assistance centers across the world in preparation for unknown emergency strikes. A government allocates budgets to local authorities to finance local public facilities (public hospitals or roads) with a rough knowledge of local public demands. An academic institute divides grants among various departments to support research activities (seminars or conferences) based on an estimate of the number of participants. In this paper, the ambiguous need of an agent is modeled as a finite set of probability distributions over the non-negative real numbers. Indeed, by using different theories to estimate an agent’s need, one may end up with multiple probability distributions. I axiomatize a class of rules that select the allocations that maximize a particular Maxmin-expected-utility (MEU) utilitarian social welfare function which sums up the minimal expected utilities of agents. In the individual decision context, Gilboa and Schmeidler (1989) axiomatize the MEU decision rule with an ambiguity aversion axiom ---a convexity property. Surprisingly, in the social choice context, while none of my axioms resembles their convexity property, the interaction of the axioms implies a similar choice rule.
Speaker's Website