When auctions don't work

Last Friday, I wrote about investment in telecom infrastructure and I included a reference to a chart that I prepared a number of years ago that looked at the amount of capital sucked out of global wireless carriers in the year 2000 by (mainly European) 3G spectrum auctions.

Former CRTC vice-chair Richard French sent me a copy of a paper called "Governance and game theory: When do franchise auctions induce firms to overbid?" that has been accepted for publication later this year in Telecommunications Policy.

Rick's paper argues that overbidding, especially in the German and British auctions in 2000 aggravated the tech collapse which led to a loss of value by the successful bidders.

Canonical auction theory states that bidders make valuations of objects to be auctioned and bid rationally in the light of their valuations. In the case of spectrum auctions, this means that bidders make business plans and bid up to their minimum return on investment. In the ideal auction, they will bid away supranormal profits (economic rents) to the benefit of the public treasury.

While auction theory accounts for the possibility of a winner’s curse, in which the winning bidder is the one which overestimates the value of the auctioned object the most, it fails to account for a case in which not only the winning bidder(s), but also the last losing bidder, so grossly overvalue the object.

Rick suggests that in a small minority of cases, this kind of overbidding can be anticipated and should be planned for.

By the way, Rick tells me that he does not believe that Canada's AWS auction was a case of overbidding.

Write Rick for a copy of the paper.

Technorati Tags:
Rick French, spectrum auction