Calendar mechanisms

I study the dynamic mechanism design problem of a monopolist selling a fixed number of service slots to randomly arriving, short-lived buyers with heterogeneous values. The fully optimal mechanism is a non-standard auction in which bidders' payoffs are non-monotone in their opponents' bids. Because its complexity may make the fully optimal mechanism too costly to implement, I also study the optimal mechanisms in restricted classes. The most restrictive are pure calendar mechanisms, which allocate service dates instead of contingent contracts. The optimal pure calendar mechanism is characterized by the opportunity costs of service slots and is implementable with a simple mechanism.