The core tenet of performance feedback theory is that the evaluation of past performance shapes future behaviours and choices: We repeat the actions that have seemingly led us to success and avoid those that we believe have led us to failure. However, this process is never straightforward and is subject to a myriad of cognitive biases, which significantly complicates learning from past performance.
In our study, published in the Academy of Management Journal, my co-authors* and I argue that self-serving attribution plays a critical role in how we learn from past performance. For example, if a firm attributes a failure to external factors over which it has little control (such as black swan events), it may conclude that its strategy is not responsible for the failure and does not need to be replaced. It will therefore persist with its current approach – even if it may be the wrong one.
Our research shows how different patterns of attributing successes and failures affect firms’ responses to performance outcomes. We focus on how firms’ prior alliance performance influences how they choose between an alliance and an acquisition as their next strategic move.
The effect of attribution
We investigated alliances and acquisitions, the two most popular corporate strategic options for firms to source external assets, knowledge and capabilities that cannot be easily acquired or purchased. While these are not direct or perfect substitutes for each other, firms often consider both options before choosing one.
Our research examines how alliance performance influences future decisions on making alliances vs. acquisitions. We obtained data on alliances and acquisitions carried out by Fortune 500 companies between 1994 and 2017.
Consistent with the standard prediction of performance feedback theory, our analysis showed that the lower the alliance performance, the more likely firms will subsequently pivot to making acquisitions. We then examined how this relationship is influenced by self-serving attribution – managers’ tendency to shift the blame for a failure by attributing it to external factors but take credit for a success by attributing it to their own actions and capabilities.
For example, if self-serving attribution makes managers attribute a failed alliance to their alliance partners, the failure will not discourage them from making more alliances because they believe it was not their fault, and that the next alliance will be successful as long as their partners do their part. Self-serving attribution may also make managers take more credit for a successful alliance than they deserve, which falsely boosts their confidence and further encourages them to make alliances.
Our analyses yielded three interesting findings. First, we found that firms are less likely to pivot to an acquisition after a failed alliance when the number of alliance partners is large. This is because it allows managers to easily diffuse the responsibility for failure across several partners, just as we feel less responsible for the failure of a group to which we belong when the group is large. Conversely, we found that firms are less likely to continue relying on alliances after a successful alliance when the number of partners is large. This is because the large number of partners makes it difficult for managers to attribute the success of an alliance entirely to their own capabilities.
Second, firms are more likely to turn to an acquisition after a failed alliance when the alliance partners are experienced and perceived to be highly capable. On the contrary, we found that a successful alliance reinforces the tendency that firms continue relying on the successful strategy (i.e. alliances) when the partners are inexperienced and perceived to be less capable. This is because managers are less likely to make self-serving attributions when their partners are perceived to be capable, as it is difficult for them to blame capable partners for failure or to claim all the credit for themselves.
Third, highly variable past alliance performance introduces ambiguity and noise and complicates managers’ ability to establish accurate causal relationships between their actions and the outcomes of past alliances. This can increase their tendency to make self-serving attributions. Indeed, in our study, highly variable past alliance performance made firms less likely to switch to an acquisition after both failed and successful alliances.
The importance of accounting for attribution
Our research not only provides evidence that attribution plays a significant role in organisations’ responses to past performance, but also demonstrates theoretical mechanisms that strengthen or mitigate the effect of attribution in learning from performance feedback.
These findings underscore the need to account for the attribution process in performance feedback research. In our study, failure did not lead to strategic change when the conditions promoting self-serving attribution (e.g. a large number of alliance partners) were present as much as when they were absent.
What’s more, strategic inertia – managers’ tendency to repeat successful strategies – was reduced when the conditions discouraging self-serving attribution (e.g. strong perceived capability of alliance partners) were present. Our research therefore helps explain enduring puzzles in performance feedback theory, which may require a critical revision if the role of performance attribution is properly considered.
Our study also contributes to the research on corporate strategy by being the first to provide a behavioural account of the choice between alliances and acquisitions. Previous studies on this topic have primarily relied on economic and agency perspectives. Departing from these approaches by portraying the decision between the two governance modes as a rational or opportunistic choice process, we show how performance feedback of a governance mode (alliances) shaped the choice of an alternative mode (acquisitions). We add that key governance choices (alliances vs. acquisitions) may not always be rational but may involve judgmental heuristics and biases, such as self-serving attribution.
Edited by:Rachel Eva Lim
About the research
Leave a Comment