Get More for Less? Experimental Evidence on Repeated Decision Making Under Non - Monotone Incentives

In this paper we test a dynamic agency model. The incentive compatible contract is non - monotone and does not award the highest pay to the most successful agent. Based on the assumption that agents are rational we expect them to choose high effort despite the non - monotonicity of the contract.Four treatments with differing information structures are analyzed; a neutral treamtent, a framing treatment, and two framing treatments with two players, agent and principal. Overall, the experiments confirm the model and show that the non-monotone contract has the desired incentive effects. Furthermore, principals strictly prefer the non - monotone contract over a monotone contract that would ex ante lead to a lower expected surplus for them. Observed learning behavior of agents accords with learning direction theory. Possible extensions of the experiments to analyze the impact of fairness and reciprocity in detail are discussed. In that respect, the study serves as a benchmark study to derive answers about the practical applicability of non - monotone pay schemes which are often viewed as non-plausible labor contracts