Bounded rationality in principal-agent relationships

We conducted six treatments of a standard moral hazard experiment with hidden action. All treatments had identical Nash equilibria. However, the behavior in all treatments and periods was inconsistent with established agency theory (Nash equilibrium). In the early periods of the experiment, behavior differed significantly between treatments. This difference largely vanished in the final periods. We used logit equilibrium (LE) as a device to grasp boundedly rational behavior and found the following: (1) LE predictions are much closer to subjects' behavior in the laboratory; (2) LE probabilities of choosing between strategies and experimental behavior show remarkably similar patterns; and (3) profit-maximizing contract offers according to the LE are close to those derived from regressions.