Valuing Nostalgia: A Multi-Regression Analysis of Bidding Behaviors within a Bilateral Agent Framework for Online Collectible Auctions

Abstract

Public online auctions for collectible goods provide an effective way for sellers to connect with a wide-range of potential buyers. However, the identities of the winners of these auctions - whether they are genuine collectors or driven by the prospect of future profits - remains largely unknown. Using past bidding behavior to categorize winners into these two groups, I examine Base Set Charizard auctions on eBay using both a standard OLS regression model as well as a logit regression model. I find that winners categorized as collectors pay over the median price a significant portion of the time, and paradoxically, that winners that are categorized as collectors are able to find success in winning auctions at a discount to what collectors are paying.