Gains-from-Trade in Bilateral Trade with a Broker

Ilya Hajiaghayi, MohammadTaghi Hajiaghayi, Gary Peng, Suho Shin

Abstract

In the classic bilateral trade problem, achieving perfect efficiency is impossible without external subsidies. This paper investigates the impact of a strategic, profit-maximizing broker on market efficiency. By analyzing the ratio between the Gains-from-Trade (GFT) achieved by a broker and the First-Best GFT (FB-GFT), the authors establish theoretical bounds across various valuation distributions. While a broker can guarantee constant-factor approximations under Monotone Hazard Rate (MHR) and Uniform distributions, the study reveals that in general settings, a broker's pursuit of profit can lead to an arbitrary loss of market efficiency.

Introduction

Bilateral trade involves a single buyer and a single seller, each holding a private valuation for an item. The fundamental goal is to facilitate a trade whenever the buyer's value exceeds the seller's, creating "Gains-from-Trade." However, the landmark Myerson-Satterthwaite theorem (1983) proved that no mechanism can be simultaneously efficient, budget-balanced, and incentive-compatible.

Recent literature has shifted toward "approximation," seeking mechanisms that guarantee a fraction of the maximum possible GFT. This paper introduces a strategic intermediary—a broker—into the equation. Unlike a neutral market designer, the broker aims to maximize their own profit by setting a spread between the buying and selling prices. This study explores whether the broker's inherent greed significantly degrades the social welfare of the transaction.

Theoretical Framework and Methodology

The authors employ theoretical mechanism design and probabilistic analysis to compare the GFT induced by a broker against the First-Best GFT (FB-GFT)—the surplus generated by an omniscient social planner. The analysis focuses on two primary mechanism types:

  1. Posted-Pricing: The broker sets a fixed buy price ($q$) and a fixed sell price ($p$).
  2. Optimal Mechanisms: The broker utilizes Bayesian Incentive Compatible and Individually Rational rules to maximize expected profit.

A central contribution of the paper is the Decomposition Theorem. This structural inequality partitions the GFT into the broker's profit and a remainder term, providing a mathematical pathway to prove that a profitable broker often—though not always—facilitates a meaningful degree of trade efficiency.

Key Findings

The "Good News": Constant-Factor Approximations

The research identifies several scenarios where the broker's presence does not destroy market utility:

  • Uniform Distributions: When valuations are drawn from uniform distributions, the broker induces at least a 1/2-approximation of the FB-GFT. This bound is proven to be tight.
  • MHR Distributions: For symmetric agents with Monotone Hazard Rate (MHR) valuations (e.g., Normal or Exponential distributions), an optimal posted-pricing mechanism achieves at least a 1/36-approximation.
  • The Single-Sample Mechanism: In a surprising result, the authors show that a broker does not require full knowledge of valuation distributions. By simply taking one random sample from the buyer’s distribution and one from the seller’s to set prices, the broker can guarantee a 1/12-approximation of the FB-GFT.
  • Quantile-Based Pricing: If a broker sets prices based on the $\alpha$-quantile for the buyer and $\beta$-quantile for the seller, they achieve a $\beta(1-\alpha)$ approximation of the first-best GFT.

The "Bad News": Inapproximability in General Settings

Despite the positive results for specific distributions, the paper highlights a significant "Greed Gap." In general settings without restricted distribution shapes, a profit-maximizing broker can be excessively strategic. The authors prove that the approximation factor can become arbitrarily close to zero. This holds true even in simplified scenarios, such as when one player's value is publicly known or when the agents are symmetric. In these cases, the broker may set a price spread so wide that the probability of trade vanishes, even when a trade would have been highly socially beneficial.

Discussion and Limitations

The study underscores that symmetry acts as a "safety net," making it less likely for a broker to break market efficiency. However, the gap between the constant-factor guarantees for MHR/Uniform distributions and the zero-efficiency results for general distributions remains a critical area of concern.

A primary limitation of this work is its focus on a single buyer and seller. While this provides a foundational understanding of strategic intermediation, real-world "Two-sided Markets" often involve multiple participants and complex financial exchanges (e.g., dark pools or real estate platforms). Furthermore, it remains an open question whether a constant-factor approximation is possible for the broader class of regular distributions, which are less restrictive than MHR but more common in economic modeling than general distributions.

Conclusion

This research provides a nuanced view of intermediaries in bilateral trade. While a strategic broker can facilitate efficient trade under favorable distributional conditions, their presence introduces the risk of significant market failure in more volatile or general environments. The findings suggest that the "cost of greed" is highly dependent on the predictability of the participants' valuations.

Summary updated

BibTeX
@inproceedings{hajiaghayi2025gains,
  author = {Ilya Hajiaghayi and MohammadTaghi Hajiaghayi and Gary Peng and Suho Shin},
  title = {{Gains-from-Trade in Bilateral Trade with a Broker}},
  booktitle = {SODA},
  year = {2025},
  doi = {10.1137/1.9781611978322.164},
}
Figure