Short- and Long-term Effects of Class Assignment : Evidence from a Flagship University in Brazil

Given a discontinuity in the class assignment of a flagship university in Brazil, we find that the relative disadvantage students face in the better class reduces their absolute performance. Although their future earnings are unaffected, these students are more willing to switch majors and less likely to have a prestigious occupation. By checking the heterogeneity of the discontinuities, we find these results are not explained by differences in absolute peer quality, but instead by students' relative rank. The rank effects on the willingness to switch majors and time to graduation are stronger for men than for women. While for women a small increase in peer quality offsets the effect of a lower rank; for men the rank effect dominates the effect of peer quality on average. Therefore, assigning disadvantaged students to a slightly better group may undermine their aspirations, but this effect disappears if the peer quality is high enough