Asymmetric Optimal Auction Design with Loss-Averse Bidders *

52 Pages Posted: 23 Sep 2022

See all articles by Akitoshi Muramoto

Akitoshi Muramoto

Osaka University of Economics

Takeharu Sogo

SKEMA Business School; Université Côte d'Azur

Date Written: June 26, 2024

Abstract

We study optimal auctions with expectation-based loss-averse bidders. First, we consider the case of ex-ante identical bidders and establish sufficient conditions under which optimal mechanisms must be asymmetric. While symmetric designs are optimal for bidders with expected-utility preferences, expected revenues are higher in a onebidder optimal design (an extreme form of asymmetric design) than in any symmetric design with multiple bidders if the degree of loss aversion is sufficiently large relative to valuation variations. Second, we explicitly derive the optimal auction in a two-bidder setting. For ex-ante identical bidders, the optimal auction excludes one bidder if the degree of loss aversion is sufficiently large relative to valuation variations; otherwise, Myerson's symmetric two-bidder design is optimal. For ex-ante heterogeneous bidders, the optimal auction is modified from Myerson (1981) to reduce outcome uncertainty.

Keywords: JEL Classification: D03, D44, D81, D82, D84 Asymmetric Auctions, Optimal Auctions, Reference Dependence, Expectationbased Loss Aversion

JEL Classification: D03, D44, D81, D82, D84

Suggested Citation

Muramoto, Akitoshi and Sogo, Takeharu, Asymmetric Optimal Auction Design with Loss-Averse Bidders * (June 26, 2024). Available at SSRN: https://ssrn.com/abstract=4207422 or http://dx.doi.org/10.2139/ssrn.4207422

Akitoshi Muramoto

Osaka University of Economics ( email )

2-2-8 Osumi
Higashiyokogawa-ku
Osaka
Japan

Takeharu Sogo (Contact Author)

SKEMA Business School ( email )

Espl. Mona Lisa, Courbevoie, Office B208
Lille, 92400
France

Université Côte d'Azur ( email )

France

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
188
Abstract Views
1,065
Rank
348,042
PlumX Metrics