Superstitious Learning with Rare Strategic Decisions: Theory and Evidence from Corporate Acquisitions

In this paper, the notion of superstitious learning is applied to the context of rare and complex strategic decisions. I argue that superstitious learning is a particularly relevant problem for these types of decisions not only because causal linkages between actions and outcomes might be poorly inferred, but also, more basically, because their performance outcomes are often very difficult to assess in objective ways. I test these arguments with a sample of U.S. bank mergers and find evidence that managers' perceptions of success in previous acquisitions is negatively related to the actual performance of the focal merger, and that this effect increases , rather than decreases, as managers accumulate experience. Consistent with the theoretical arguments developed, the effect is significantly attenuated as the knowledge is systematically articulated and codified and the stock of experience becomes more heterogeneous.