Optimal Auctions for Constrained Buyers

2026-06-29

AI Summary

This study addresses the breakdown of multi-unit auction mechanisms under constraints such as the inability to incur interim debt. The authors develop a unified theoretical framework to analyze strategic behavior by constrained bidders and distinguish between two objectives: revenue maximization and consumer surplus maximization. Employing tools from measure theory, they establish that Myerson-style auctions remain optimal for revenue maximization, whereas for consumer surplus maximization, deterministic mechanisms that restrict bidder strategies can strictly outperform classic incentive-compatible designs. The work not only uncovers a fundamental divergence between these two objectives in constrained environments but also provides concrete constructions of optimal mechanisms applicable across a broad range of settings.

Abstract

We study multi-unit multi-buyer auctions, where buyers are subject to constraints that affect their bidding strategy. These may take the form of \emph{bidding constraints} (e.g., no-overbidding, in cases where bids are partially verifiable), or of \emph{outcome constraints} (e.g., in the case of auctions that unfold over time, an inability to go into debt even temporarily). The constraints fundamentally redefine the design space: the revelation principle, the envelope theorem, and Myerson's lemma do not apply in this constrained setting. Consequently, the space of implementable mechanisms expands significantly, admitting auctions that are incentive compatible with respect to the constrained buyers' utility, but would not be incentive compatible in the classical sense. In this paper we focus on monotone constraints where it is not the buyers' total budget that is restricted, but rather the manner in which they can bid or spend their budget. Our results show a separation between \emph{revenue-aligned} objectives (e.g., revenue, welfare, or any linear combination of the two) and \emph{consumer-aligned} objectives (e.g., consumer surplus). For revenue-aligned objectives, we rely on measure-theoretic tools to establish a unified theory parallel to Myerson, showing that despite the expanded design space, Myerson-style auctions remain optimal. For consumer-aligned objectives, the picture is different: we show that the seller can leverage the buyers' strategic limitations to strictly outperform classically incentive compatible mechanisms. We design the optimal deterministic auction for a wide class of instances, focusing in particular on buyers who cannot tolerate temporary debt. Overall, our work provides theoretical underpinnings for this area and shows that a rigorous and systematic approach can reveal general insights regarding optimal auction design.

Problem

• constrained buyers
• optimal auctions
• incentive compatibility
• revenue maximization
• consumer surplus

Innovation

• constrained buyers
• optimal auction design
• incentive compatibility
• revenue-aligned objectives
• measure-theoretic analysis