Much attention has been paid to the question of whether there is a modesty bias in East Asian employees' self-ratings of job performance (i.e., a tendency to self-rate their performance lower than supervisors rate it). However, empirical results are conflicting, with some studies supporting the modesty bias and others not supporting it. We suggest that moderators representing boundary conditions for the modesty bias effect may shed light on these conflicting results. In essence, the question should not be whether there is a modesty bias, but rather when is there a modesty bias? We propose three moderators: purpose of the ratings (administrative, developmental, or research), job performance dimension (task performance, organizational citizenship behavior, or leadership), and country-level in-group collectivism. Based on 40 studies (63 independent samples) with samples from East Asia (mainland China, Japan, South Korea, and Taiwan), we found no evidence of a modesty bias. That is, East Asian employees' self-ratings were, on average, higher than supervisor-ratings of job performance (i.e., a leniency bias). The one exception was when ratings were collected for research purposes; in this case, there was, on average, no mean difference between self- and supervisor-ratings. Thus, East Asian employees' research-purpose self-ratings are more modest, but this does not cross into a modesty bias. In all, our results do not support a modesty bias as a widespread cultural norm among East Asian employees.