Scarce radio spectrum is assigned to mobile network operators (MNOs) by national regulatory authorities (NRAs). Spectrum is usually assigned by beauty contest or an auction. The process requires that winners make a payment to the government. MNOs seek scarce spectrum to enable the provision of wireless services for profit. While MNOs are imperfectly aware of their costs, NRAs rely solely on MNOs for this information. As such, NRAs set spectrum assignment conditions (including minimum bid price) largely ignorant of MNO operating conditions. This study examines the performance of 3G auction outcomes in terms of the prices paid by winners via an econometric analysis of a unique sample of national 3G spectrum auctions. These winning bids depend on national and mobile market conditions, spectrum package attributes, license process, and post-award operator requirements. Finally, model estimation accounts for the censored nature of these data.