Information sharing and collusion in Government debt auctions
Governments all over the world sell around 4 trillion USD worth of securities every year. For example, in 2014 Denmark sold government bonds and T-bills for approximately DKK 170.000 million.1 Given its size, the effciency of treasury bond auctions has been studied for decades. Nevertheless, to date there is no consensus on what is the optimal auction mechanism. Research on the effciency of auctions has focused on the underpricing relative to secondary markets. Both theoretical and empirical studies have arrived at conicting results as to which of the two main securities sales mechanisms, discriminatory or uni- form auctions, is best. In both auction mechanisms, bidders face quantity uncertainty. This encourages aggressive bidding, and more so for uniform auctions. But, as long as bidders' valuation of auctioned securities reects common values (e.g. their resale price in secondary markets), a winner's curse leads to less aggressive bidding, particular in discriminatory auctions. We believe that a first-order issue regarding the effciency of treasury auctions is whether bidders collude or not. There is no theoretical or empirical work that addresses this question, and we intend to fill this gap.