Nationally, 18% of principals turn over each year, yet research has not yet credibly established the effects of this turnover on student and teacher outcomes. Using statewide data from Missouri and Tennessee, we employ a difference-in-differences model with a matched comparison group to estimate arguably causal effects. We find that principal turnover lowers school achievement by 0.03 SD in the next year, on average. Effects vary by transition type, with larger negative effects for transfers to other schools but no or even positive later effects of demotions of (presumably lower-performing) principals. Principal turnover also increases teacher turnover, but this does not explain the drop in student achievement. Replacement with an experienced successor can largely offset negative principal turnover effects.