We investigate the predictive validity of teacher licensure tests using data from the Massachusetts Tests for Educator Licensure (MTEL). MTEL scores predict teachers' in-service performance ratings and contributions to student test scores (i.e., value added). We then explore whether these relationships vary for teacher candidates of color. Although candidates of color have lower first-time pass rates and are less likely to retake licensure tests, we do not find consistent evidence that MTEL scores are less predictive of student achievement gains. Finally, we find that some evidence that MTEL scores are more predictive of teacher performance ratings for teachers of color than for White teachers.