Conflict and the Design of Optimal Conflict

Abstract: One prevalent theme that runs through the management literature on compensation is the problem of reducing organizational conflict. However, theoretical research in economics emphasizes that while it is desirable to reduce destructive conflicts, there are situations in which functional conflicts are needed to ensure efficiency. MacLeod (2003) proposes a version of the principal-agent model in which output is not perfectly observable and the principal and agent receive noisy and private performance signals which are imperfectly correlated and can lead to disagreements. In this model, high effort equilibria require conflicts on the equilibrium path. Bonus payments can only be based on the principal’s private signal. Since the signals are only imperfectly correlated, there will be situations in which the agent does not get a bonus although he observes a good signal. In those cases, it is crucial that the agent initiates conflict which is costly for the principal to ensure that the principal is motivated to pay the bonus when she observes a good signal. To study the role of functional conflicts we implement MacLeod’s environment in an experiment. In one treatment we allow for conflicts while in another treatment we rule out conflicts. We observe that subjects fail to coordinate on the predicted high-effort equilibrium under conflict. Agents initiate too many conflicts which are not functional. These conflict patterns distort the incentives of the principal to pay the bonus as a function of his private signal. Thus, agents are not incentivized to choose high effort. Our results indicate that efficiency enhancing equilibria based on strategically initiated conflicts are hard to obtain. We plan to design instruments to institutionalise conflict. In particular, we intend to study a formal reporting system in which agents can file official complaints. We hope that behavioural forces such as lying aversion will help to better align behaviour by making conflict formal.