Research on the role of behavioral biases in contract theory implicitly assumes that these are fixed. We show on the example of overconfidence that such biases may be endogenous to the incentives provided. Using a novel laboratory experimental design that allows to disentangle selection from incentive effects, we find that having an insurance against losses in a real effort induces individuals to overstate their performance relative to others. At the same time, we find no evidence that overconfidence plays a role in insurance choice.
Joint work with Joachim Winter and Martin Kocher (LMU)
