We exploit exogenous variation from a school desegregation policy to investigate the determinants of academic achievement and wellbeing of dual language learners. The policy buses some school starters with low host-country language proficiency from schools with many dual language learners, higher total budget for dual language learning and higher per-pupil spending to schools with fewer dual language learners but lower total budget for dual language learning and per-pupil spending. Assignment to busing is exogenous conditional on three observed individual characteristics, hence we compare bused and non-bused pupils conditional on these characteristics. We find that assignment to forced busing has a negative effect on the academic performance and wellbeing of dual language learners, which is at odds with findings for school desegregation in the US.