Distributionally Robust Optimal Allocation with Costly Verification

Allocation with Costly Verification under Ambiguity In “Distributionally Robust Optimal Allocation with Costly Verification,” Bayrak, Koçyiğit, Kuhn, and Pınar examine the mechanism design problem faced by a principal allocating a single good to one of several agents with costly verification and without monetary transfers, adopting a distributionally robust optimization perspective. Each agent desires the good and generates value for the principal, who does not initially know these values but can verify them at a cost. This problem arises in various scenarios, such as a venture capitalist selecting a start-up for funding or a procurement manager choosing a supplier. Many such allocation problems occur infrequently, leaving the principal with limited knowledge about the distribution of the agents’ values. The authors model these allocation problems as distributionally robust mechanism design problems, explicitly addressing distributional ambiguity. They introduce the first robustly (and Pareto-robustly) optimal mechanisms in the literature. These mechanisms are not only robust but also simple and interpretable.