Introduction and Overview.

This paper reports on a large scale field experiment testing strategies available to a seller participating in simultaneous competitive sequential, ascending price automobile auctions. Every other week, the seller offered approximately 120 vehicles for sale in an auction environment in which several competitive sellers offered on the order of 3,000 vehicles. The experiment tested various sequences in which the seller could offer the vehicles, such as high values first or low values first. Surprisingly, and contrary to intuition drawn from the theory of single item and single seller auctions, the worst performing sequence from those tested is for the seller to order vehicles from highest to lowest values. The best sequence is to group the vehicles by type and offer the low valued vehicles first and then move to offer the higher valued vehicles. Our conjecture is that this sequence reduces the competition with other sellers for the attention of specialized buyers. JEL classification numbers: D440, L620, L810, C930, C900