This paper aims to explain the high scores of 15-year-old native pupils in The Netherlands and Flanders by comparing them with the scores of pupils in countries with the same highly stratified educational system: Wallonia, the German "Lander," the Swiss German cantons, and Austria. We use the data from the Programme for International Pupil Assessment (PISA) 2006 together with the specific PISA data of Germany and Switzerland. We apply a multilevel model that takes into account the individual-, curriculum-, and system-level features in these highly stratified educational systems. The high scores of the Dutch pupils can be explained by the size of The Netherlands' vocational sector. The high Flemish scores can be partly explained by the high curriculum mobility. Central exit exams are not a good explanation of the high Dutch scores. Despite being limited to highly stratified systems, we still find educational policies and arrangements to have significant effects on the educational performance of pupils. (Contains 22 notes and 5 tables.)