This seems to me almost what we have now. Yes, there is a problem about defining "treating well".

Yup.

I agree that we should experiment more.

I'm all in favor of experimentation.

And if we're already experimenting to the limits of our existing regional independence, such that increased independence will relax the rate-limiting constraint on experimentation (which I doubt we are, but is I suppose possible), then yes, increased regional independence would make sense as a next step. Though perhaps it's best to do so in a small region, so it's easier to stop things if it goes horribly wrong.

Of course, if we believe for_ other_ reasons that decentralizing state power is a good idea, then we should endorse doing so for other reasons, but that's something of a nonsequitor.