A costly Bayesian implementable social choice function may not be truthfully implementable

Wu, Haoyang (2016): A costly Bayesian implementable social choice function may not be truthfully implementable.

Abstract

The revelation principle is a fundamental theorem in many economics fields. In this paper, we construct a simple labor model to show that a social choice function which can be implemented costly in Bayesian Nash equilibrium may not be truthfully implementable. The key point is the strategy cost property given in Section 4: In the direct mechanism, each agent only reports a type and will not pay the strategy cost which would be paid by himself when playing strategies in the original indirect mechanism. As a result, the revelation principle may not hold when agents' strategies are costly in the indirect mechanism.

• Item Type: MPRA Paper
• Language: English
• Keywords: Revelation principle; Game theory; Mechanism design; Auction theory
• Subjects: D - Microeconomics > D7 - Analysis of Collective Decision-Making > D71 - Social Choice ; Clubs ; Committees ; Associations; D - Microeconomics > D8 - Information, Knowledge, and Uncertainty > D82 - Asymmetric and Private Information ; Mechanism Design

References

A. Mas-Colell, M.D. Whinston and J.R. Green, Microeconomic Theory, Oxford University Press, 1995.

A. Kephart and V. Conitzer, The revelation principle for mechanism design with reporting costs, In Proceedings of the ACM Conference on Electronic Commerce (EC), Maastricht, The Netherlands, 2016.

J. Green and J.J. Laffont, Partially verifiable information and mechanism design. Review of Economic Studies, vol.53, 447-456, 1986.

L. Yu, Mechanism design with partial verification and revelation principle. Autonomous Agents and Multi-Agent Systems, vol.22, 217-223, 2011.

V. Auletta, \emph{et al}, Alternatives to truthfulness are hard to recognize. Autonomous Agents and Multi-Agent Systems, vol.22, 200-216, 2011.