| Dan Levin, Ohio State University |
| Multi-Unit Demand Auctions with Synergies: Behavior in Sealed-Bid versus Ascending-Bid Uniform Price Auctions |
| Session: C-12-1 Wednesday 16 August 2000 by Levin, Dan |
| We study uniform-price auctions with multiple units supply and demand with synergies. We compare sealed bid to ascending price private value auctions. In our experimental design supply is two units and there are n bidders with one unit demand who are represented by a computer. This is known to the single human bidder (h) with a demand for two units who competes with those n rivals. Our simplification eliminates strategic uncertainty for h in an already complicated environment. There are two additional economic incentives in such auctions that h needs to balance: demand reduction (DR) results from monopsony power that a bidder with multi units demand has dictates lowering the bid on the second unit; synergy, the fact that winning two units is valued exceeds the sum of values of two single units, dictates additional aggressiveness in order to win both units. Bidding behavior of h depends on the relative value she draws. In the sealed bid auction the balancing act is resolved in the following way: with low values the DR dominates synergy consideration and h ought to bid zero on the second unit in our design, thus gives up winning two units. At intermediate values, the two incentives are balanced with two equal bids set above the value of a single unit which may result in winning one unit and negative gains. At higher values, synergy dominates the DR incentive resulting in bidding high enough to assure winning both units. In the ascending clock auction, h can condition her dropping-out rule for each unit on the observed drop out prices of her rivals. As a result, her strategy is more refined than in the sealed bid auction resulting in different behavior and consequences (allocation and prices) in particular for intermediate values. We find substantial deviations from equilibrium in both auctions for multi-unit demand bidders, with (a) too much DR in ascending-bid auctions at intermediate values (in contrast to theory prediction of winning both units), (b) overly aggressive bidding in the sealed bid auctions at low valuation but not aggressive enough high valuations. The net outcome is that over some ranges of the value space sealed-bid auctions raise more revenue and have higher efficiencies than the ascending-bid auctions. Further, sealed-bid auctions never achieve significantly lower efficiencies and quite often significantly higher revenues than the ascending-bid auctions |