In the Austrian (as well as the German) education system students have to choose between different school tracks at the age of 10. We argue that early tracking creates inefficiencies because the earlier the track choice has to be made, the more it is influenced by factors other than innate ability. Recent evidence suggests that the relative age of a student within a grade is related to his or her achievement, and that this effect is decreasing over grades. Thus, age-related achievement differences probably translate into age-related differences in track choice if track choice has to be made early. In this paper we estimate the effect of observed age on the track choice after grade 4 using register data for a major Austrian city for the period 1984-2006. Since observed age at track choice is endogenous, we exploit the exogenous variation in birth month to identify the causal effect of age. We find a strong and significant positive effect of age on track choice in grades 5-8. Since after grade 8, students again have to make a track choice, we use additional data from PISA 2003 and 2006 to show that the effect is long-lasting in urban areas. Therefore, the education system fails to provide a mechanism that leads to an efficient allocation of students to tracks.