Online auctions differ from traditional auctions in several ways, but perhaps one of the most significant is the amount of time that bidders interact with the auction mechanism. This paper examines past findings in online auctions and discusses three psychological mechanisms through which bidders may increase their valuations of an item through interaction with the auction mechanism over time. In doing so, this paper provides a theoretical lens through which previous empirical studies of online auctions can be interpreted and serves as a guide for future research.
Keywords: Behavioral economics, prospect theory, online auctions, sniping, auction equivalence
With the rapid growth of the Internet, online auctions have become a popular and efficient way for both businesses and consumers to exchange goods while allowing businesses to reinvent the way that they manage their relationships with trading partners [Smith et al. 2000]. Consumers have been quick to adopt online auctions for the purchase and sale of everything from collectibles to automobiles. While these online auctions are typically structured around well-studied auction types, past research comparing online and offline auctions has yielded mixed findings [Lee 1998; Overby & Jap 2009]. As a result, numerous researchers have called for an examination of the theoretical and empirical impacts of moving auctions to the online environment [Klein & O'Keefe 1999; Pinker et al. 2003; Van Heck & Vervest 1998].
The increased use of information technology for both online and offline auctions has enabled empirical researchers new opportunities to analyze data on auction outcomes [Bajari & Hortacsu 2003; Katkar & Reiley 2006; Kuruzovich et al. 2010; Lee 1998; Lucking-Reiley et al. 2007] and conduct field experiments [Lucking-Reiley 1999]. These efforts have yielded a greater understanding of mechanism design and the behavior of bidders and sellers-providing empirical tests of auction theory and examining the impact of features unique to online auctions. These studies have presented three particularly intriguing findings not predicted by traditional auction theory. First, Lee  found that an electronic implementation of a wholesale auction resulted in higher overall prices than in the earlier offline market. Second, Lucking-Reiley  found a lack of equivalence between Dutch and first-price sealed bid auctions in an online field experiment-findings which differed both from theoretical predictions and prior offline experiments. Finally, several studies have observed the prevalence of bid "sniping," in which the majority of the bids arrive in the final minutes or seconds of the auction [e.g., Bapna 2003; Ely & Hossain 2009; Glasner 2002; Ockenfels & Roth 2002].
A key difference between the online auctions and the traditional "live" auctions is the length of time over which the auction occurs, and this paper suggests that this is important to understanding differences between online and offline auctions. Empirical studies of auctions have reported the significance of the amount of time in the determination of final price for items being auctioned [Lucking-Reiley et al. 2007]. The mechanism proposed to justify the higher revenues of longer auctions has been the arrival of bidders with higher valuation. This paper discusses alternative causal mechanisms through which the extended time of Internet auctions can influence bidder valuation and explain prior empirical studies. In order to help understand how consumers interact with auction mechanisms over time, this paper will draw upon the extensive research on individual decision-making originating in prospect theory [Kahneman & Tversky 1979] and the theory of mental accounting [Thaler 1980], in which individuals are found to make decisions in ways which differ from traditional utility theory.
Over the past 25 years, prospect theory has been recognized as a modification of traditional economic theory that in some cases more accurately explains how individuals actually make decisions. …