This paper argues that grades cannot be interpreted only as a reward for a given academic achievement, but they also reflect teachers’ ratingsof pupils. Relative within-classroom differences in grades therefore contain valuable information about pupils’ own – usually unknown – ability, and could have an effect on subsequent academic achievement.Previous studies have found that school grading standards have a small but statistically significant positive effect on subsequent academic achievement.However, those studies have been less than forthcoming about why this is so.Here we analyse the impact of grades at the individual level and seek to answer the question of what happens if pupils receive better grades for the same academic achievement. Do better gradessubsequently motivate students to achieve more,or is the effect actually demotivating?Moreover, is the impact of grades a compositional effect of the classroom (who are the peers, how good are the teachers, how good is the school) or is it universal?In addition, this article seeks to cast light on why school grades are important in later academic achievement. By applying first-difference and fixedeffect estimators to various types of Hungarian educational panel datasets,we show that grades do have a positive effect on subsequent academic achievement.This holds true for those who have not switched classes. The estimated impact is independent of classroom composition, showing that the benefit arising from within-classroom differences in grades occurs in every classroom.The growthof self-confidence is offered as a possible underlying mechanismto explain this impact.