Lookers-on See Most of the Game: Decision-making for Others versus for Oneself in a Labor Market Experiment

Do lookers-on make better economic decisions and form more objective beliefs than individuals who are in the situation? We design a laboratory experiment studying the dynamic pattern of job applications and individuals' belief updating process in a labor market problem. Our experimental treatments feature decision-making for oneself versus for others. In both treatments, the probability of being accepted for a job position depends on external shocks as well as the individuals' ability ranking, which is unknown to the subjects. We elicit subjects' beliefs about their own and about others' ability ranking after the realization of each application outcome. We find that, when making decisions for themselves, subjects are more inclined to attribute failure to external shocks, while attributing success to their own ability being higher than others'. Estimation results from a reinforcement learning model show that, compared to decision-making for others, subjects tend to have a higher tolerance for failure and remain in applying for unsuitable jobs for longer periods when making decisions for themselves. Our findings provide supportive evidence for the presence of self-serving attributional bias when individuals' self-image affects their economic decisions. We show that decision-making for others could be used as a good benchmark to measure the degree of individuals' attributional bias.