Schools have been named and shamed, targets have been set, weak teachers have been weeded out, hit squads have gone in and done whatever hit squads do (smiting, presumably), super heads have come and gone. There have been Office for Standards in Education reports, special measures, a strenuous amount of levering up and stretching, and strident demands (for exactly what I am never sure) from Ms Melanie Phillips. Now we have schools open at weekends, holidays and evenings, with both teachers and pupils bribed to attend. Yet nothing changes. Weak schools are weaker than ever. As last week’s TES reported, schools at the bottom of the league tables have indeed improved their performance in GCSE exams since 1997. Alas, schools at the top have improved even more.

You may say that this does not matter; that the point of reforms is to raise achievement across the board. But new Labour’s claim was not just that we could raise standards generally, but that we could stop schools under-performing in relation to others. Poor children, it argued, should not be abandoned to schools that were grossly inferior to others. If schools in middle-class areas did well, there was no reason why, with sufficient sticks and carrots, those in deprived areas should not do equally or nearly as well.

If this policy were sound, there should, at the very least, have been no change in the gap between the best and the worst schools. Instead, the gap has widened.

The miracle is that the weakest schools have improved at all. The publication of league tables and the growing differentiation of schools - beacon schools, city academies and specialist schools, for example - have made it easier for middle-class and aspirant working-class parents to pick out what they believe to be the best. “The best” nearly always means the schools that are already getting high exam scores. Those at the top therefore attract more and more pupils from favoured backgrounds.

The divide between the best and the worst in the state sector mirrors what happened years ago at the divide between state and private. As we moved from a traditional society, in which jobs went to the well-bred and well-connected, to a meritocratic one, in which jobs went to the well-qualified, the fee-charging schools dropped their emphasis on producing cultivated and sporting gentlemen, and became exam factories.

Since they were better-equipped than state schools, they inevitably outpaced them in the race for improvement.

Now, state schools know also that they stand or fall by the results, and will go to any lengths to get the best intake; and they will concentrate on exam success. All state schools will therefore improve. But those already at the top will improve more than those at the bottom.

This is an example of the classic problem of reformism: it makes things worse rather than better. The simplest way of getting rid of illiterate school-leavers, poor health and high crime levels is to get rid of poverty.

That would be cheaper than anything Charles Clarke is likely to invent. But though we are promised the abolition of child poverty, it lies 20 years in the future. In the meantime, politicians must act; wringing hands and saying nothing can change just now is not part of an acceptable political script.

So we have legislation against bad schools, bad hospitals and bad people, so that politicians can reassure themselves and everybody else that they are in control. Teachers, like doctors, will have to take their punishment.

And the public will probably have to continue moaning about “delivery”. But the only delivery that will really matter is that promise on child poverty.