This paper studies whether pupil performance gains in autonomous schools in England can be attributed to the strategic exclusion of poorly performing pupils. In England there were two phases of academy school introduction, the first in the 2000s being a school improvement programme for poorly per-forming schools, the second a mass academisation programme from 2010 for better-performing schools. Overall, exclusion rates are higher in academies, with the earlier programme featuring a much higher increase in the exclusion rates. However, rather than a means of test score manipulation, the higher exclusion rate reflects the rigorous discipline enforced by the pre-2010 academies.