Superstitious learning in strategic events: theory and evidence from corporate acquisitions

In this paper, the notion of superstitious learning is extended to encompass both causal ambiguity (Levit and March, 1988) and outcome ambiguity phenomena. I argue that superstitious learning is a particularly relevant problem in strategic events not only because causal linkages between actions and outcomes might be poorly inferred, but for the more basic reason that their performance outcomes are often very difficult to assess in objective ways. Experience accumulation, in these conditions, not only might not be helpful but could even hurt the learning process, with negative performance consequences. The conditions defining the boundaries of the phenomenon are then considered: two mitigating factors are identified in the heterogeneity of the stock of accumulated experience and in deliberate learning processes. I test these arguments with a sample of US bank mergers and find evidence that managers' self-attributions of success in previous acquisitions is negatively related to the actual performance of the focal merger, and that this effect increases, rather than reducing, as they accumulate experience. Consistent with the theoretical arguments developed, the effect is significantly attenuated as the stock of experience becomes more heterogeneous and knowledge is systematically articulated and codified.