Remorse, risk aversion and efficiency in eBay coin auctions.(ANTHOLOGY)

Internet auctions are popular. Their ease of entry, low transaction costs, and the pervasiveness of computers make these markets accessible to almost anyone and, ultimately, very competitive. Here is an intriguing laboratory for the observation of economic behavior. Collectible coins, such as the U.S. Morgan silver dollars in our sample, may be sold to a dealer, bought from a dealer, or may be auctioned to the public, avoiding the middleman. To voluntarily participate, auction participants either expect to realize a selling price above the dealer's bid or expect to pay a price below a dealer's ask price. We investigate these auctions for buyer's remorse (BR), when the winning bidder pays more than the dealer's ask price; seller's remorse (SR), when either the coin does not sell (the seller still incurs eBay fees), or when the selling price is below the dealer's bid price; and the incidence of efficient auction outcomes, defined as when the auction is advantageous to both buyer and seller because the winning bid is between the dealer's ask and bid prices.

We examined 184 randomly selected auctions of Morgan dollars that were active on eBay on July 7, 2007. For each coin, date and professionally-graded condition were recorded and coin dealers' bid and ask prices were found using The Coin Dealer Newsletter contemporary to the auction. Of the 184 auctions, 108 resulted in the sale of the coins while the remaining 76 sales never closed.

When categorized into auctions that closed (108) and auctions that did not (76), a significant difference was observed in the value of coins in the two groups: the mean dealer ask price of coins in the closed auctions was $215 and the average professional grade for these coins was mint state 59. These values were $954 and mint state (ms) 63.5 in non-closing auctions. Among all auctions, the incidence of SR is slightly higher than BR, 0.522 compared to 0.478. In completed auctions, BR is much more likely than is relative underpayment for a coin, BR occurring about 81% of the time in closed auctions versus only 19% resulting in SR. In conformance with the evidence that better and more valuable coins are included in the group of auctions that did not close, error coins with minting irregularities also appear more frequently in non-closed auctions. These results suggest that risk aversion is influencing both the buyers and sellers. Buyers may be reluctant to purchase high-valued coins sight-unseen and sellers may use high starting prices and reserves to ensure these coins will not sell at a low price.

It is surprising to note that the probability of observing either SR or BR equals one for the overall sample of auctions; therefore no selling prices fell within the dealer's bid and ask spread and in no auction did both buyer and seller simultaneously benefit. Thus, we observe no incidence of an efficient auction as defined here. We tested the reasonableness of the expectation that auctions will produce a mutually-beneficial efficient outcome by examining whether the actual price paid was statistically different from the bid-ask midrange. That hypothesis is rejected with 99% confidence, further calling into question the efficacy of participating in the auction compared to transacting with a dealer.

Using probit regression, a number of variables were tested as contributors to the likelihood of seller and buyer remorse. Included among these was the effect of winning bidders' reputational scores as reported by eBay. The score is a function of the number of transactions that a winner has successfully concluded without problems. Winner reputation is highly significant and negatively related to BR, indicating that experienced bidders are less likely to overbid. Similar to the winner's reputational score, being rated as a power seller by eBay may proxy for the sophistication of the party offering the coin. Probit results indicate greater likelihood of BR when the auction involves a "power seller". Here the use of bold typed descriptions and other marketing subtleties may be effectively employed by sophisticated sellers to elicit higher bids, increasing the chance of overpayment by the buyer.

In conclusion, buyers are affected by the uncertainty inherent in on-line coin auctions. It appears that buyers fear the Winner's Curse and attempt to minimize costly mistakes by avoiding auctions of high valued coins and coins with high grades for quality. Sellers suffer from buyers' aversion to risk as many auctions are unsuccessful or generate selling prices below dealers' bids. Consequently, it would be a prudent strategy for sellers to either not offer top quality or expensive coins on eBay or adopt a tactic for minimizing the perceived risk to buyers, such as offering a money-back guarantee. Alternatively, transacting with a dealer may prove to be the dominant strategy, especially considering the lack of efficient auctions in our sample.

JEL C90. G20. L10. L60

Published online: 28 August 2008

D. Friesner

College of Pharmacy, North Dakota State University, Fargo, ND 58105, USA

e-mail: Daniel.Friesner@NDSU.edu

K. Hickman ([mail]) * M. McPherson

School of Business, Gonzaga University, Spokane, WA 99258, USA

e-mail: hickman@jepson.gonzaga.edu

M. McPherson

e-mail: mcpherson@jepson.gonzaga.edu

Print Get the Mag Weekly Updates