This paper investigates whether education is used as a signaling device for worker productivity in developing countries. To do such we employ a simple test of employer learning on Ghana manufacturing data. We find no evidence of educational signaling for individuals who were hired through direct contacts in the firm, and thus for workers for which employers arguably have more information about their true abilities. In contrast, education acts as signal for workers who were hired through more formal channels, although only for those that do not receive on-the-job-training.