Abstract
We analyze online auction prices for premium Bordeaux reds sold on the German eBay site. Our data set reports a total of 2133 transactions completed during November and December 2003. First, we estimate a model which includes only fundamental quality attributes such as vintage conditions, appellation, vineyard classification, and chateau reputation. In a second step, we add quality and reputation indicators due to wine critic Robert Parker. In the third model, we add online-auction specific variables to see whether they have any impact on the final auction price. We evaluate seller reputation scores, auction length, timing of the auction end (day of week, time of day), number of bids, unknown bidder identities (private auctions), initial price per bottle and whether information on Parker ratings is stated in the offer. Our application is unique because eBay auctions (especially for premium wine) have not been analyzed in this detail in the literature. The inclusion of online-auction specific variables adds virtually nothing the overall goodness of fit. However, we observe significant but small effects for auction length, the number of bids, and private auctions. In contrast to many other studies, we estimate an insignificant seller reputation effect on final auction prices.