Summary: |
In this clinical study, we examine the winner’s curse hypothesis that rational agents, in an auction setting, will impute an inverse relation between bids and price uncertainty and between bids and the level of competition. To accomplish same, we avail of a unique data set of 548 successful real estate common value auctions with incomplete information during the period October 2004 to November 2005. A failure to incorporate conditional information price uncertainty and competition into bidding may invite a winner’s curse. This follows as the range of estimated prices (and the winning bid) tends to grow, relative to the mean estimate of value, with price uncertainty and competition. Winner’s curse arises when the winning bid exceeds the value of the auctioned asset or if the value of the auctioned asset is less than that which is anticipated on the part of the winning bidder. As a result, rational bidders will compensate for prices accordingly by adjusting bids downwards below their ex-ante estimate of the value of the asset so-called bid shading. If we infer a positive association between bids and price uncertainty and bids and the level of competition, these findings are indicative of the presence of de facto winner’s curse in our auction data. We adopt a novel methodology to quantify price uncertainty per unit area of residential real estate by constructing portfolios of properties with similar characterising traits, as indicated by hedonic asset pricing models. A property’s price uncertainty is estimated as the coefficient of variation of the replicating portfolio of which it is a constituent. Our calibration of the level of competition involves count data regarding the number of bids and the number of bidders. Initial findings do indicate a positive relation between bids and the level of competition, in line with the presence of winner’s curse in the observations. Our findings are robust to an extensive set of control variables and model specifications which include a variety of measurements of winning bid price relatives: to the opening bid; the reserve price and the expected auction revenues. Our set of control variables relates to the property’s location, size, condition and type, whether there was an editorial announcement as well as the gender and experience of the winning bidder with respect to the auctioned property. |