Recent research has established that teachers matter for student achievements, albeit because of dimensions of 'teacher quality' that are largely unexplained. A less closely investigated issue is whether teacher turnover directly harms student academic achievement. In this paper, we examine whether teacher turnover affects academic achievement of 16 year old state secondary school students using a unique data set of linked students and teachers in England. Identification comes from either: a school fixed effects design which exploits year-on-year variation in turnover in different subject groups, within schools; or student fixed effect design that where the variation comes from the cross sectional variation in turnover in different subjects, in the same school, experienced by a student. Both methods give similar results, suggesting that a higher teacher entry rate reduces students' test scores, albeit by small amounts.