Experimental Evaluation of Mechanisms Designed to Achieve the Uniform Outcome in Rationing Problems

Concepts de base

The paper experimentally evaluates four mechanisms intended to achieve the Uniform outcome in rationing problems, finding that a mechanism providing non-binding feedback during the reporting period greatly improves performance compared to the dominant-strategy, direct-revelation mechanism.

Résumé

The paper experimentally evaluates four mechanisms designed to achieve the Uniform outcome in rationing problems:

1. Direct Revelation Uniform Rationing (DRU): Agents simultaneously report their types, and the Uniform allocation is assigned based on the reports.

2. Sequential Revelation Uniform Rationing (SRU): Agents report their types sequentially, with the second agent observing the first agent's report.

3. Obviously Strategy-proof Uniform Rationing (OSPU): A clock-auction type procedure where agents make a series of choices between their current temporary assignment and slightly modified assignments.

4. Pre-play Feedback Uniform Rationing (PFU): Identical to DRU, but agents can adjust their reports over a reporting period and are informed about their tentative assignments under the currently selected reports.

The key findings are:

• PFU greatly improves performance compared to the other mechanisms, achieving the highest rates of Uniform outcomes, efficient allocations, and share of efficient outcome earnings.

• SRU modestly improves performance compared to DRU, especially in cases where the Uniform allocation differs from equal division.

• Contrary to expectations, the additional simplicity of OSPU does not improve performance compared to DRU.

The authors characterize the behavior in the different mechanisms to understand the reasons behind this ranking. Subjects in DRU often achieve the Uniform outcome through non-dominant strategies, resembling weakly-dominated Nash equilibria. The informational advantages of SRU2 and the feedback in PFU lead to higher rates of dominant strategy play and Uniform outcomes. The potential shortcomings of the OSPU mechanism, such as the existence of multiple nodes per player and the prevention of weakly-dominated Nash equilibria that achieve the Uniform outcome, lower its performance relative to DRU.

Stats

"Subjects in DRU achieve the Uniform outcome in about half of all instances, though this outcome is often achieved via non-dominant strategies."
"Profiles of observed play in DRU often resemble weakly-dominated Nash equilibria, so subjects seem to be fairly successful in making correct conjectures about the play of others and responding accordingly."
"The tentative reports in PFU are highly correlated with final reports, giving both subjects a very similar informational advantage to the second mover in SRU."
"Most agents end up maximizing their payoffs in PFU, and dominant strategies seem to have a greater appeal because tentative reports vary and remain uncertain during the reporting period."

Citations

"The feedback mechanism, PFU, greatly improves performance. The sequentiality of SRU modestly improves performance. The additional simplicity of OSPU does not improve performance."
"Contrary to the expectations of Li and Pycia and Troyan (2023)'s theory, OSPU does not lead to better outcomes than DRU."

Idées clés tirées de

Testing the simplicity of strategy-proof mechanisms

by Alexander L.... à arxiv.org 04-19-2024

https://arxiv.org/pdf/2404.11883.pdf