On the Choice of Level or Ratio Targets in a Bonus Plan: Theory and Empirical Evidence From a Field Experiment

We study the choice of performance measures in a target setting context. In particular, we consider the choice between a level measure and a ratio measure that scales the target by the volume of general activity. Using the principles and assumptions of the agency theoretical paradigm, we first develop a theory and derive new results on the optimal choice of measure for an incentive plan that awards a bonus for exceeding a target. We show that incentives are stronger with ratio targets than level targets because they control for shocks to the aggregate level of activity and reduce the compensation risk borne by the agent. We then test our predictions using a field experiment in partnership with a leading hotel brand company. Contrary to our prediction, we find that both a level target and a ratio target combined with information about the expected level of performance required to meet the ratio target outperform a ratio target without level information. Our results show that the optimal choice of performance measure does not respond only to risk considerations.