Year of Graduation
Optimism bias in the trust game
Previous experiments with one-shot games have revealed systematic deviations from equilibrium predictions, and there is a long running interest in uncovering players’ motives. We have found a striking feature: senders tend to have much more optimistic beliefs about the receiver with whom they will be paired than about an average receiver. That phenomenon is called optimism bias and in our paper we show that it significantly affects sender’s decision. Optimism bias causes people to overestimate the probability of being paired with a trustworthy partner and as a result trust more. Senders behave irrationally – while they can forecast the number of receivers that return money quite well, they systematically overestimate their partner’s level of trustworthiness. In other words, overconfidence leads players to neglect the level of trustworthiness among receivers. We found that receivers also have optimism bias about senders’ level of trust. Moreover, children tend to have higher optimism bias than adolescents. We showed that if game is played sequentially, rather than simultaneously, levels of trustworthiness are smaller. Finally, we have found that senders’ decisions do not depend on type of payoff – players who played with money shows similar levels of trust as those who played with worthless carton cards. Behavior of receivers, on the contrary, crucially depends on type of payoff – under monetary payoffs levels of trustworthiness are much lower.