This article examines to what extent the specific institutional arrangement of an education system moderates the influence of social background on students' track allocation, and whether this happens via primary or secondary effects of social origin. I tackle the problem of omitted-variable bias by analyzing subnational education systems in Switzerland, a country with a variety of cantonal school systems but otherwise homogeneous institutions. The results show a complex picture. First, even

more »

... gh the absolute transition probability to the highest track is higher in education systems with low stratification for students of higher social background, this does not translate into a relative advantage as in most cantons the odds of transitioning do not differ between high and low social background students. Second, in line with previous research, I observe that the secondary effect of social origin prevails in more stratified education systems. Third, it is not possible to conclude with certainty that specific features of the education system enable high social background parents to disproportionately influence their children's transition probabilities because the results are not robust.