'experience — the kind of experience business people gather over years in the market –ceases to be a helpful guide, whereas models and deep historical knowledge have at least a hope of getting at what’s really going on'
I think he's right, up to a point. The problem is though, that what he's trying to model is a world in which most people have no access to the kind of abstract reasoning and deep historical knowledge he can draw on. And that can end up vitiating the models.
I can see a kind of circularity to all this. The models tell us what we should be doing, but they only work if we do what we should be doing. If we introduce the diversity of agents, then we can get somewhere: so assume very bounded rationality for consumers, investors and voters, and more 'rational rationality' on the part of policy makers, and you may get a guide to policy. But policy itself is just as subject to the boundedness of rationality.
A further irony occurs to me. The boundedness of rationality comes across in the ideologically driven hyper-rational models of 'fresh water' economists, who desperately try to cling on their rational expectations models, for (often) moralistic reasons.
I've got a headache now.