We develop a dynamic structural model of competitive bidding in multiunit sequential business-to-business auctions. Our model accounts for two notable characteristics of these auctions: (i) bidders have multiple purchase opportunities for the same product, and (ii) winning bidders in each round can acquire multiple units of the same product. We apply the model to bidding data from the world's largest flower wholesale market at which trades are facilitated through fast-paced, sequential, Dutch auctions. Using a two-step estimation approach, we are able to recover the structural parameters effectively and efficiently. We then conduct policy counterfactuals to evaluate the performance of alternative design choices. The results suggest that the current auction practice still has ample room for improvement. In light of this, we propose an optimization framework that can facilitate auctioneers' decisions in making the trade-off between revenue maximization and operational efficiency.
Bibliographical noteFunding Information:
The authors thank the department editor, associate editor, and reviewers for insightful comments and suggestions. This research would not have been possible without the collaboration of Royal FloraHolland via the Veilen met Advies project. The first author also acknowledges the support of Erasmus Research Institute of Management (ERIM) via the Talent Placement Program.
© 2019 INFORMS.
- Auction design
- Dynamic discrete games
- Sequential auction
- Structural modeling