Do Uniform Price Auctions Trade‐off Higher Risk for Higher Return?*

Abstract

AbstractResults of the recent US Treasury experiment comparing uniform and discriminatory price auctions suggest that uniform price auctions achieve lower award concentration and higher average revenues but experience higher variance of revenues from auction to auction relative to discriminatory price auctions. In other words, uniform pricing offers higher risk for higher return. This apparent trade‐off may deter some sellers considering the switch from discriminatory to uniform pricing; thus, it is important to investigate whether these results can be generalized. The Treasury's results are driven by the tendency for bidders in the uniform auctions to submit multiple bids at a wide range of prices, essentially steep bid curves, instead of the flat bids submitted in the discriminatory auctions. However, theory and data from discriminatory auctions in Sweden and Portugal suggest that bidders in discriminatory auctions will submit steeper bids when there is more potential for winner's curse. It is possible that bidders in the US auctions, which are preceded by a when‐issued market, are relatively unconcerned with the winner's curse. We examine the auction formats in an environment much more likely to generate the winner's curse: the initial public offering market. We compare discriminatory price auctions in Japan and uniform price auctions in Israel on the basis of revenue and slope of the demand curves and test a number of divisible good auction theory implications. In contrast with the Treasury's results, we find that revenues are likely to be insignificantly different in the discriminatory and uniform price auctions and that bid curves may be equally steep under both formats in markets with information asymmetry. The winner's curse increases underpricing and steepens demand curves in Japan's discriminatory price auctions but not in Israel's uniform price auctions.