Thurow’s job-competition model implies that overeducation is contingent upon the differing skill endowments of employees. As yet, only rudimentary evidence has been furnished to confirm this relationship. In the present paper, we test the theory in a more sophisticated manner, by means of a more differentiated survey of the skill endowments of workers. Our analyses are based on the German Life History Study (GLHS), in which longitudinal biographical data was gathered for the West German cohorts born in 1964 and 1971 up until the year of the survey (1998). These data are analyzed using a trivariate probit model which takes into account the selective acquisition of school qualifications, and the selective choice of vocational training with varying levels of quality. Our findings confirm that type and grade of school leaving certificate both have a strong effect on the later risk of overeducation. The quality of the vocational training chosen only impacts on the overeducation risk when the strong selectivity effects in certain types of schools and types of vocational training are not taken into account. In line with existing literature, we find that the risk of overeducation decreases with increasing traditional skill measures such as experience, tenure, on-the-job-training, and further education. In sum, our results clearly confirm the capacity of the job-competition model to explain the persistence of overeducation in the labor market.