Can Buyers Reveal for a Better Deal?

Daniel Halpern, Gregory Kehne, Jamie Tucker-Foltz

Abstract

While sharing private data with a seller can help a single buyer get a better price, this research reveals that the benefit often vanishes in more complex markets. In settings with multiple buyers or multiple goods, strategic transparency forces a choice between social efficiency and buyer savings. This paper demonstrates that finding the optimal way to reveal information is computationally NP-hard and that social efficiency is fundamentally incompatible with maximizing buyer welfare in multi-agent or multi-item settings.

1. Introduction

In a standard market, a seller trying to maximize revenue might set a price so high that some buyers are priced out, even if they value the item more than it cost the seller to produce. This results in a "deadweight loss"—an inefficient outcome. Recent economic theory (notably Ali et al., 2020) suggested a potential win-win solution: voluntary disclosure. If a buyer can credibly prove they have a low valuation (e.g., "I only have $10 in my budget"), a rational seller might lower the price to make the sale. In the simplest case of one buyer and one good, this can lead to "perfect" outcomes where the item is always sold and the buyer retains more surplus.

Trade-off between seller and buyer surplus and the Pareto efficiency frontier.

This paper investigates whether this optimism holds in more complex market dynamics, specifically when multiple buyers compete for an item or a single buyer evaluates multiple goods.

2. The Verifiable Disclosure Game

The authors model market interactions as a two-stage verifiable disclosure game:

  1. Disclosure Stage: Buyers observe their true private valuations and send a verifiable "message" to the seller. This message is a set of possible values that must include their true value, allowing for truthfulness with varying degrees of vagueness.
  2. Pricing/Auction Stage: The seller, utilizing this partial information, implements a revenue-maximizing mechanism (such as a Myerson auction) to determine prices and winners.

The analysis focuses on the Bayes-Nash equilibria of this game, identifying strategies that maximize buyer surplus—the difference between a buyer's valuation and the price paid. The study employs mathematical proofs for impossibility results and uses dynamic programming to evaluate the computational complexity of calculating optimal disclosure strategies.

3. The Conflict Between Efficiency and Welfare

A central finding is the "Asymmetry Surprise": in a two-buyer scenario, buyer welfare is maximized when only one buyer discloses information while the other remains silent. If both disclose, they often compete against each other's "virtual values," ultimately benefiting the seller more than themselves.

Comparison of disclosure strategies: (a) No disclosure, (b) Symmetric high/low disclosure, (c) Zeno's partition, and (d) Coarse asymmetric disclosure.

Furthermore, the paper proves that in multi-buyer or multi-good settings, social efficiency and buyer welfare are often mutually exclusive. To achieve a better deal, buyers may prefer an equilibrium where the seller occasionally fails to sell the item, rather than a socially efficient one where the seller extracts all surplus.

4. Theoretical and Computational Results

The paper establishes several key results regarding the limits of disclosure:

  • Theorem 3.3: For two buyers with uniform valuations, no "connected" disclosure strategy exists that is both socially efficient and provides savings to the buyers.
  • Theorem 4.1 (General Impossibility): In any setting with multiple goods or buyers, social efficiency is fundamentally incompatible with maximizing buyer welfare. For multiple goods, the buyer surplus in an efficient equilibrium is exactly zero.
  • Computational Complexity: Finding the optimal disclosure strategy is achievable in polynomial time if messages must be "connected" (intervals). However, if buyers can send arbitrary sets of values, the problem becomes weakly NP-hard.

5. Conclusion

While "data sovereignty" and transparency are often promoted as pro-consumer, this work serves as a cautionary tale. In competitive or complex markets, the seller is often the primary beneficiary of increased information. Future work may explore mixed-strategy equilibria and the impact of private disclosure models on these competitive dynamics.

Summary updated

Abstract

We study market interactions in which buyers are allowed to credibly reveal partial information about their types to the seller. Previous recent work has studied the special case of one buyer and one good, showing that such communication can simultaneously improve social welfare and ex ante buyer utility. However, with multiple buyers, we find that the buyer-optimal signalling schemes from the one-buyer case are actually harmful to buyer welfare. Moreover, we prove several impossibility results showing that, with either multiple i.i.d. buyers or multiple i.i.d. goods, maximizing buyer utility can be at odds with social efficiency, which is surprising in contrast with the one-buyer, one-good case. Finally, we investigate the computational tractability of implementing desirable equilibrium outcomes. We find that, even with one buyer and one good, optimizing buyer utility is generally NP-hard but tractable in a practical restricted setting.
BibTeX
@inproceedings{halpern2022can,
  author = {Daniel Halpern and Gregory Kehne and Jamie Tucker-Foltz},
  title = {{Can Buyers Reveal for a Better Deal?}},
  booktitle = {International Joint Conference on Artificial Intelligence},
  year = {2022},
  doi = {10.24963/ijcai.2022/45},
}
Figure