Allocation Scarce Resources: What Money Can’t Buy.” http://ssrn.com/abstract=1420899

In this paper I study the allocation of scarce and indivisible goods (or burdens) in cases where monetary transactions are unfeasible or prohibited. More precisely, I investigate the design of ex-ante Pareto optimal mechanisms for the allocation of m identical indivisible objects to n > m heterogenous risk-neutral agents with a private value and unitary demand for the goods, in environments where transfer payments are not allowed but agents are able to signal their value through some other costly activity (queuing, advertising, exerting effort, etc.). If money was available, the optimal mechanism would achieve the first best. It would allocate the goods efficiently, trough a second price auction and then redistribute the payments back to the agents (see d'Aspremont and Gerard-Varet (1979)). However, in a setting without transferable utility, the designer of the optimal mechanism faces a trade-off between ex-post efficiency and cost minimization. Increasing the efficiency of the allocation requires that agents are screened according to their private valuation for the goods, but eliciting private information comes at a cost. For example, suppose that the designer knows that two agents have values v 1 = x and v 2 = y, with x > y > 0, but does not know who is who. If he (directly or indirectly) impose a cost who gets the good (at least) equal to y, then only 1 will have an interest to claim it, and the surplus will be x − y