We analyze a setting common in privatizations, public tenders, and takeovers in which the ex post efficient allocation, i.e., the first best, is not implementable. Our first main result is that the open ascending auction is not second best because it is prone to rushes, i.e., all active bidders...