An Experimental Comparison of the FCC ’ s Combinatorial and Non-Combinatorial Simultaneous Multiple Round Auctions

Young for helping run the experiments, and to Dash Optimization for the free academic use of their Xpress-MP software. The authors wish to acknowledge FCC staff Mark Bykowsky, William Sharkey, and Martha Stancill for input and helpful comments. The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Federal Communications Commission, the Wireless Telecommunications Bureau, or other members of the Commission's staff. Furthermore, any views expressed in this report are not purported to reflect those of Caltech and the University of Virginia. Executive Summary Laboratory experiments were run to compare simultaneous, multi-round auctions (SMR) and simultaneous multi-round auctions with package bidding (SMRPB), using auction rules provided by FCC staff. The auction procedures were implemented with the jAuctions program developed by Jacob Goeree at Caltech. Each laboratory session consisted of a series of auctions in which the participants were assigned " national " or " regional " bidder roles. Bidders were provided with randomly generated license values and information about the extent to which combinations of licenses are worth more than the sum of the components (complementarities). High complementarities create an " exposure problem " in the sense that bidders may be hesitant to bid high on efficient networks of licenses if there is a risk of obtaining only part of the desired network. Package bidding is intended to minimize this exposure problem, although it may create other problems if efficient combinations of small bidders are unable to coordinate a response to an aggressive package bid by a large bidder, which is known as the " threshold problem. " For this reason, the revenue and efficiency effects of package bidding are an open question that is addressed with a series of laboratory experiments. The three treatment variations spelled out in the Design Report implement changes in (1) the degree of valuation complementarities, (2) the degree of overlapping strength of national bidders, and (3) the relative strengths of national and regional bidders. There are two levels for each treatment, so this 2x2x2 design yields 8 variations for each of the two auction types for a total of 16 treatments. This report is based on results from 64 laboratory sessions, each consisting of 6 to 10 auctions. Data are posted online at the jAuctions website described in the report that follows. The multi-dimensional experiment design permits an evaluation of the effects of …