Two essays on bidder behavior in simultaneous online auctions
MetadataShow full item record
The growing popularity of online auctions and the availability of rich and detailed bidding data have spawned many new empirical researches in auctions that were not possible to explore earlier. Particularly, more information is now known about bidder behavior in auctions than ever before. Still, there are certain auction characteristics that desire for more research. One such area is bidder competition in auctions. Current view on this topic considers each bidder to compete against all the other bidders as a whole, without distinguishing one competitor from another. This consideration is inadequate in explaining the effects of different levels of rivalry between bidder pairs. To overcome this limitation, this dissertation presents an innovative approach of looking at bidder competition at a dyadic bidder level. The context of the study is simultaneous online auctions selling modern Indian art. Simultaneous auction, a popular online auction format for selling highly complementary products is ideal for such study as bidders in these auctions typically compete for more than one lot (item) simultaneously, thus engaging in both within-lot and between-lot interaction (competition). This dissertation consists of two essays. The first essay shows that after controlling for aggregate level of competition, dyadic bidder interactions have significant effect on the auction outcome. Extending on these findings, a new approach of representing bidding data is presented. Each bidder is connected based on the dyadic interactions between them and a “bidder network” is formed. Using the analytical tools from Social Network Analysis, the key or important bidders in the network are identified and their effect on the auction outcome is analyzed. In the second essay, dyadic bidder interactions are used to analyze situations where bidders reevaluate or update their reservation value for the item. Using choice models, it is established that high level of between-lot interaction has a significant effect on the propensity of this behavior. The notion of dyadic interaction is further extended to analyze characteristics of bidders (termed here as Reactors) who change their pre-set reservation value for the item.