[Paper Review] Simple and Near-Optimal Mechanisms For Market Intermediation

Rad Niazadeh, Yuan Yang+1 more|arXiv (Cornell University)|Sep 9, 2014

Auction Theory and Applications4 references4 citations

TL;DR

This paper proposes simple affine fee schedules for market intermediaries, showing they achieve constant-factor approximations to optimal revenue under mild conditions on buyer distributions—specifically, when the buyer's virtual valuation is linear (e.g., generalized Pareto or MHR distributions). It proves that a prior-independent affine fee schedule $ w(P) = P - \phi_{\mathcal{B}}(P) $ guarantees a constant approximation to both surplus and optimal revenue, with a worst-case approximation ratio of 4 for monotone PDFs, and establishes that proper affine mechanisms cannot achieve constant approximation in the worst-case seller distribution.

ABSTRACT

A prevalent market structure in the Internet economy consists of buyers and sellers connected by a platform (such as Amazon or eBay) that acts as an intermediary and keeps a share of the revenue of each transaction. While the optimal mechanism that maximizes the intermediary's profit in such a setting may be quite complicated, the mechanisms observed in reality are generally much simpler, e.g., applying an affine function to the price of the transaction as the intermediary's fee. Loertscher and Niedermayer [2007] initiated the study of such fee-setting mechanisms in two-sided markets, and we continue this investigation by addressing the question of when an affine fee schedule is approximately optimal for worst-case seller distribution. On one hand our work supplies non-trivial sufficient conditions on the buyer side (i.e. linearity of marginal revenue function, or MHR property of value and value minus cost distributions) under which an affine fee schedule can obtain a constant fraction of the intermediary's optimal profit for all seller distributions. On the other hand we complement our result by showing that proper affine fee-setting mechanisms (e.g. those used in eBay and Amazon selling plans) are unable to extract a constant fraction of optimal profit in the worst-case seller distribution. As subsidiary results we also show there exists a constant gap between maximum surplus and maximum revenue under the aforementioned conditions. Most of the mechanisms that we propose are also prior-independent with respect to the seller, which signifies the practical implications of our result.

Motivation & Objective

• To understand when simple, real-world fee-setting mechanisms—like those used by Amazon and eBay—can approximate the optimal revenue of a sophisticated intermediary mechanism.
• To identify sufficient conditions on buyer-side distributions (e.g., linearity of marginal revenue, MHR property) under which affine fee schedules achieve constant-factor approximation to optimal intermediary profit.
• To show that prior-independent affine mechanisms (dependent only on buyer distribution) can achieve constant approximation to both surplus and revenue, enhancing practical applicability.
• To demonstrate the limitations of proper affine mechanisms by proving they cannot achieve a constant approximation to optimal revenue in the worst-case seller distribution.

Proposed method

• Proposes a prior-independent affine fee schedule $ w(P) = P - \phi_{\mathcal{B}}(P) $, where $ \phi_{\mathcal{B}}(P) $ is derived from the buyer’s virtual valuation function, ensuring ex-post individual rationality for any seller distribution.
• Uses Bayesian mechanism design with worst-case approximation guarantees to evaluate performance relative to optimal revenue and maximum surplus.
• Applies extreme value theory and properties of generalized Pareto and MHR distributions to characterize when affine mechanisms yield constant approximation ratios.
• Employs a reduction from multi-buyer settings to a single effective buyer by taking the maximum value among i.i.d. buyers, preserving revenue and surplus expectations.
• Proves that for uniform $[0,1]$ buyer distributions, no proper affine fee schedule can achieve a constant approximation to optimal revenue in the worst-case seller distribution.
• Establishes a necessary and sufficient condition for prior-independent constant approximation: $ \alpha - \beta = 1 $, with $ \alpha = 2, \beta = 1 $ yielding the best approximation ratio of 8.

Experimental results

Research questions

• RQ1 Under what conditions on the buyer’s value distribution can affine fee schedules achieve a constant-factor approximation to the intermediary’s optimal revenue?
• RQ2 Can prior-independent affine fee schedules—dependent only on buyer distribution—achieve constant approximation to both surplus and revenue across all seller cost distributions?
• RQ3 Is it possible for proper affine fee schedules (e.g., $ w(P) = (1-\alpha)P + \beta $) to achieve a constant approximation to optimal revenue in the worst-case seller distribution?
• RQ4 How do the relationships between surplus and revenue behave under MHR or generalized Pareto assumptions on the buyer side?
• RQ5 What is the tightest possible approximation ratio achievable by affine mechanisms in the worst-case setting, and when is it attained?

Key findings

• A prior-independent affine fee schedule $ w(P) = P - \phi_{\mathcal{B}}(P) $ achieves a constant approximation to both maximum surplus and optimal revenue, with an approximation ratio of at most 4 when the buyer’s PDF is monotone.
• For the uniform $[0,1]$ buyer distribution, the optimal revenue is at least $ \frac{1}{8} $-approximated by the best proper affine fee schedule, and the surplus is at least $ \frac{1}{8} $-approximated in expectation.
• There exist regular seller distributions for which the ratio of optimal revenue to the best proper affine fee schedule’s revenue is arbitrarily large, proving that proper affine mechanisms cannot achieve constant approximation in the worst case.
• The only prior-independent affine fee schedules achieving constant approximation to surplus are those satisfying $ \alpha - \beta = 1 $, with $ \alpha = 2, \beta = 1 $ yielding the best approximation ratio of 8.
• In multi-buyer settings with i.i.d. MHR-distributed buyers, the maximum value among buyers is also MHR, allowing extension of the main results: a constant fee schedule achieves an $ e^2 $-approximation to optimal surplus and revenue in expectation.
• The paper establishes a constant gap between maximum surplus and optimal revenue under MHR or generalized Pareto assumptions, implying that surplus and revenue are within a constant factor of each other.