Abstract:
In multi-object auction models with unitary demand agents, if agents’ utility functions satisfy quasi-linearity, three auction formats, sealed-bid auction, exact ascending auction, and approximate ascending auction, are known to identify the minimum price equilibrium (MPE), and exhibit elegant efficiency and incentive-compatibility. These auctions are conjured to preserve their properties beyond quasi-linearity. Nevertheless, we exemplify that with general utility functions, these auctions fail to identify the MPEs and are substantially inefficient and manipulatable. The implications of our negative results for multi-object auction models with agents with multi-unit demand, and matching with contracts models are also discussed.
