Quote:Revenue non-equivalence between auctions with soft and hard closing mechanisms: New evidence from Yahoo!
Abstract
We use a unique dataset to examine the revenue differences between auctions with a hard-close ending rule versus those with a soft-close ending rule. We find that selling items using the soft-close rule increases the selling price by an amount between $25 and $44 (or 13–20 percent) over the hard-close format. One possible theoretical explanation for these results is that the hard-close ending rule accommodates the practice of sniping, which leads to a lower expected selling price. We find empirically that a lack of experience could help to explain why, in spite of the revenue differences, some sellers select the hard-close ending rule.
Highlights
#9658; We use a unique dataset to examine the revenue differences between auctions with a hard-close ending rule versus those with a soft-close ending rule. #9658; We found that selling items using the soft-close rule increases the selling price by an amount between $25 and $44 (or 13–20 percent) over the hard-close format. #9658; One possible theoretical explanation for these results is that the hard-close ending rule accommodates the practice of sniping, which leads to a lower expected selling price. #9658; We find empirically that a lack of experience could help to explain why, in spite of the revenue differences, some sellers select the hard-close ending rule.
Introduction
Empirical investigations of online auctions have shown a common pattern of very late bidding. Bidders placing bids just seconds before the auction's close, a practice known as "sniping", is not only a commonplace, but has been made easy through the use of artificial bidding agents.2 Due to network problems, however, bids submitted this late in an auction have a positive probability of not being received. Roth and Ockenfels (2002) conclude that such a scenario of sniping, in which some bids are probabilistically not received, reduces an auction's expected closing price. Other authors have agreed with the Roth and Ockenfels (2002) assertion that sniping can be seen as a form of "implicit collusion" among bidders to lower the seller's expected revenues.3
Online auctions generally have two types of ending rules: hard-close and soft-close. Under a hard-close rule, the auction ends at a fixed and posted time. In a soft-close auction, the auction has a hypothetical posted closing time. A bid placed shortly before this posted closing time extends the auction for an additional period of time. Once extended, the auction is further extended with each additional bid and closes only after a pre-specified amount of time has elapsed without a bid.
This paper fills a gap in the literature by systematically testing if the soft-close ending rule yields higher seller revenues than its hard-close counterpart. We verify below that this is, indeed, the case. While the hard-close ending rule is a potentially valuable feature for bidders, it generates lower expected revenues for the seller than a soft-close ending rule.
We use data collected from Yahoo! Auctions, a site that allowed sellers to choose between soft- and hard-close ending rules, to directly investigate the revenue differences between auction ending rules. Giving sellers the choice of the auction's closing rule is a feature not commonly found on online auction sites. Since Yahoo! closed its online auction site in 2007, our empirical investigation is likely to remain unique for some time to come.4 Specifically, we record all single-item auctions of Apple 60 GB Video iPods that closed in a seven-month period in 2006. We hypothesize that, in reinforcing the incentive to snipe, the hard-close ending rule facilitates implicit bidder collusion and should consequently result in lower expected revenues than the soft-close ending rule.
We find the difference in revenues between soft- and hard-close auctions to be quite large. On average, iPods sold using the soft-close mechanism have selling prices between $25 and $44 (or 13–20 percent) higher than those sold under the hard-close format. Though consistent with the initial hypothesis, the magnitude of this estimate is rather surprising. Because we do not directly test for sniping, this is only one possible explanation for the observed results.
Such large revenue differences between closing formats naturally raises the question of which sellers choose the soft-close ending rule. Given our results, one would expect all Yahoo! sellers to select the soft-close format. Nonetheless, several sellers in our dataset select a hard-close format. One possible explanation is that these sellers are relatively inexperienced and do not realize that the closing format has an effect on expected revenues. In fact, we find that the only variable appearing to affect a seller's probability of choosing the soft-close ending rule is the seller's experience. The hard-close format is the default option when setting up an auction and naïve, inexperienced sellers might not realize that they would benefit from changing it. Our results also raise the question of why a soft-close ending rule is not more common among online auction sites. For example,
ebay and many other online auction sites only offer a hard-close ending rule.5
The remainder of the paper is organized as follows. The next section reviews the literature, focusing in particular on auction closing formats and sniping in online auctions. In Section 3, we describe the data and present our empirical analysis and results. Section 4 offers some concluding remarks.
Section snippets
Literature review
The prevalence of late bidding, seen in a striking number of online auctions, has challenged the predictions of auction theory and been recently examined in the literature from a variety of perspectives. An excellent survey of online auctions, with particular attention paid to late and incremental bidding, can be found in Ockenfels et al. (2006). Bajari and Hortaçsu (2004) provide another very useful survey of the literature concerning online auctions.
Vickrey (1961) shows that it is a dominant
Data analysis
Our dataset consists of all single item auctions of new Apple 60 GB Video iPods that closed in the seven month period from December 29, 2005 to July 26, 2006, on the Yahoo! Auctions website. Yahoo! allows the seller to select the type of closing rule used in the auction. This is a unique feature that results in both hard- and soft-close auctions taking place on a single site, which allows for a direct examination of the impact of the closing format on the selling price. In reinforcing the
Conclusion
We have presented empirical evidence that online auctions of hard- and soft-close formats are not revenue equivalent. Our unique dataset allows us to isolate the effect of the closing rule on a seller's expected revenues by controlling for much of the variation in other auction and seller characteristics. Apparently, the soft-close format affords the seller a substantial premium in expected revenues.
Posted 11/17/2022 2:19 pm
It hasn't happened often though, but I'd expect that, given just how rare the postmarks that I am looking for are. A small needle in a vast field of haystacks might be a good analogy.
