This paper develops a model that formalizes several connections between mergers, collusion, and competition policy. In equilibrium, firms may merge to make collusion sustainable when it cannot be sustained with the original set of firms. A rise in the probability of detecting and prosecuting collusion could induce a wave of mergers, so firms can sustain collusion again. Indeed, mergers could fully neutralize the procompetitive effect of an improvement in collusion detection and prosecution. From a normative perspective, we show that merger policy is crucial when cost synergies are small (or nonexistent) and the competition authority can only deter collusion by restricting mergers. Finally, we highlight that mergers could be more harmful (less beneficial) than expected if the impact that mergers have on the competition regime is properly considered, which suggests a decomposition of the welfare impact of mergers into unilateral and coordinated effects.