An options-based solution to the sequential auction problem

  • Authors:

  • Adam I. Juda;David C. Parkes

  • Affiliations:

  • School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA;School of Engineering and Applied Sciences, Harvard University, Cambridge, MA 02138, USA

  • Venue:
  • Artificial Intelligence
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

The sequential auction problem is commonplace in open, electronic marketplaces such as eBay. This is the problem where a buyer has no dominant strategy in bidding across multiple auctions when the buyer would have a simple, truth-revealing strategy if there was but a single auction event. Our model allows for multiple, distinct goods and market dynamics with buyers and sellers that arrive over time. Sellers each bring a single unit of a good to the market while buyers can have values on bundles of goods. We model each individual auction as a second-price (Vickrey) auction and propose an options-based, proxied solution to provide price and winner-determination coordination across auctions. While still allowing for temporally uncoordinated market participation, this options-based approach solves the sequential auction problem and provides truthful bidding as a weakly dominant strategy for buyers. An empirical study suggests that this coordination can enable a significant efficiency and revenue improvement over the current eBay market design, and highlights the effect on performance of complex buyer valuations (buyers with substitutes and complements valuations) and varying the market liquidity.