Optimal Auctions 695on a number of factors: intrinsic interest in the various events and athletes and whether
the telecast is live. Other sources of revenue, such as reselling the rights for cable televi-
sion transmission, are also uncertain. Similarly, the cost of broadcasting the Olympics
(depending on inflation, exchange-rate changes, or predictions of future wages) has a
significant (unknown) common-value component. At the same time, revenues and costs
also can differ substantially across networks. Previous Olympic broadcasting experience
may enhance revenues (ABC was once regarded as theOlympic network) and reduce
costs. Broadcasting the Olympics gives a boost to a network’s ratings, thereby enhancing
the profitability of other programs. Analysts believe that this boost is more important to
the third-place network in the ratings, spurring an aggressive bid. In short, Olympic bid-
ding displays a mix of common-value and private-value elements.
The history of the bidding suggests that the three networks have engaged in vigor-
ous competition for the games. The norm has been for Olympics organizers to insist on
multiple rounds of bidding in an attempt to extract maximum revenue from the winner.
To the extent that multiple bidding rounds approximate an English auction, each buyer
should be willing to bid up to its estimate of what the games are worth, conditional on its
being the winner. How successfully did the networks carry out this strategy? The evidence
suggests that the winning networks have incurred losses, sometimes large ones, as fre-
quently as they have turned profits. ABC suffered a $65 million loss on the 1988 Calgary
games and a smaller loss on the 1984 Los Angeles games. CBS broke even on the 1992
Albertville games, whereas NBC made a $25 to $30 million profit on the 1988 Seoul games
(a much smaller profit than it had expected). NBC incurred a $100 million loss on the
1992 Barcelona games but made a $70 million profit on the 1996 Atlanta games and a
modest profit ($40 million or less) on the 2000 Sydney Olympics. In the last decade, the
profit record has been equally mixed. The Salt Lake City winter games of 2002 marked
NBC’s finest Olympics hour. Major events shown live in prime time, the strong showings
by American athletes, the figure skating judging scandal, and post–9/11 patriotism all
contributed to unusually high revenues, generating a $75 million profit for the network.
The 2004 Athens games generated a similar profit in the $60 to $70 million range. But the
2006 Turin Olympics and the 2008 Beijing Olympics barely broke even, while the 2010
Vancouver Olympics—which took place in the depths of the global recession—generated
a $223 million loss for the network.
On average, it appears that the networks have paid full value or even more than full
value for the Olympics telecasts. The networks’ mixed records might owe to a run of bad
luck (realized revenues unpredictably falling below estimates), but there is no hard evi-
dence to support this view. Did the networks tend to overestimate revenues (and/or under-
estimate costs)? The record suggests a qualified yes. Winning bids have been consistently
higher than organizers’ pre-auction predictions of what the broadcasting rights are worth.
(The organizers’ lower predictions have proved more accurate than the networks’.)
Indeed, NBC’s enormous Barcelona loss stemmed in part from the pay-per-view fiasco
(based on inflated revenue predictions as to the price cable viewers would pay for simul-
cast coverage) and in part on cost overruns. In addition, the winner’s curse played a major
role in NBC’s Barcelona loss. In the first bidding round, all three networks placed com-
parable bids between $350 and $360 million. In the second and final round, NBC raisedc16AuctionsandCompetitiveBidding.qxd 9/26/11 1:09 PM Page 695