We model merger control procedures as a process of sequential acquisition of information in which mergers can be cleared after a first phase of investigation. We find that the enforceability of clearance decisions at the end of the first phase is unattractive to the extent that it prevents the authorities to use their expectations as to whether evidence gathered in the first phase will be confirmed in the second phase. This deprives the first phase of its potential as an effective screening mechanism. We also find that when clearance decisions in the first phase are enforceable, a different (higher) standard in the first phase is only desirable when Phase I decisions are captured by merging parties (as opposed to complainants).