We document experimentally how biased self-assessments affect the outcome of labor markets. In the experiments, we exogenously manipulate the self-confidence of participants in the role of workers regarding their relative performance by employing hard and easy real-effort tasks. Participants in the role of firms can make offers before information about the workers’ performance has been revealed. Such early offers by firms are more often accepted by workers when the real-effort task is hard than when it is easy. We show that the treatment effect works through a shift in beliefs; that is, under-confident agents are more likely to accept early offers than overconfident agents. The experiment identifies a behavioral determinant of unraveling, namely biased self-assessments. The treatment with the hard task entails more unraveling and thereby leads to lower efficiency and less stability, and it shifts payoffs from high- to low-quality firms. This paper was accepted by Uri Gneezy, behavioral economics.