It is clear why there is grade inflation this year (12%?).
It is because when teachers gave their CAGs they could only say what the pupils would get if the exam went well. They could not assume the pupil would mess up, that would not be fair.
But each year a percentage of pupils in the real exam do mess up: maybe they are a bit ill, their pet died, they overslept, they didn't revise the topic in the high-mark essay question, etc etc.
So each year the final exam results will be a percentage lower than teacher predictions.
But the crucial issue that no-one knows in advance of the exams who those pupils will be.
It is just not possible to predict - it could be the top ranked pupil or the borderline pupil or a pupil in the middle of a grade range.
The grade inflation that people are talking about this year is largely not due to a teacher knowing a pupil is only capable of a B but giving them a CAG of A. It is because they gave a CAG of B.
If all teachers give the grade the pupil is actually capable of, overall the grades are higher because account is not taken of the percentage who underperform in the actual exam each year.
It is further crucial that in an exam year, there is a paper to mark so the result reflects the real actual performance. If anyone questions that mark, the paper can be dug out, reviewed, remarked and the pupil, parents and teacher will know the final mark is correct and fair.
As I said, this year there is no paper. No-one knows who the individual pupils are that would underperform from the predicted grade.
The only fair thing to do is the use the predicted grade - or CAG in this case.
Grade inflation is actually not really an issue for the following years ofl pupils who take the exam. Their exam moderation will be based on 2020 and then it is all fair.
12% grade inflation from 2020 onwards has been given undue significance in a year of mass tragedy and a change to normality in every single area of life.
Imagine of Ofqual had approached this from the opposite direction. Image they had said: this year there were no exams so we have no idea which pupils would have underperformed and which would have overperformed. There is no evidence of what exam performance would be for an individual pupil. So the fairest thing to do is use the teachers' predicted grades for these pupils.
...everyone would have been totally OK with that.