Do Academies work?

There is a good piece on Vox by Andrew Eyles and Steve Machin about the performance of the early tranche of academy schools. Unlike a lot of the late switchers these were mainly poorly performing schools. Their research - which has a very nice design  - finds that the KS4 performance of schools which switched to academy status improved  by roughly 0.2 of a SD (it would be nice to know what that means in terms of GCSE points) compared to relevant comparators. What is  intriguing is that this effect was almost entirely driven by improvement in so called community schools - in the UK context essentially non church schools. Also notable is that switching to academy status was quite highly correlated with a change of leadership in the school, which of course raises the question of what exactly the "treatment" was. I particularly liked the fact that Eyles and Machin are at pains to stress that early adopters of academy status are very different from late adopters.