Responding to a comment on performance related pay in teaching, I wasn't talking about pay in general but specifically teaching. The results of teachers depend on their pupils. A set of bright motivated students may be much easier to teach than a more demoralised class. It can quite easly be the case that a teacher does a better job (I.e. makes more of a difference) by getting a struggling child to learn to read than another teacher does in getting an already achieving child to get straight As. The dynamic is therefore quite different from (say) achieving a sales target, although there is nothing wrong with that.
I guess I am arguing that what Michael Gove calls performance related pay bares no relation to actual performance. If the pay and performance are not linked, you are effectively not incentivising the teacher to change their behaviour, because they know that their behaviour is not linked to the "performance".