What is your view of rational expectations models? Is this a useful abstraction, or is it too unrealistic?

It is irrelevant. Almost all the shocks that are experienced by the macroeconomy are idiosyncratic. It’s not like we live in an agrarian economy where you have regular weather shocks, where you have a well-defined probability distribution. You can’t analyse the world from a perspective such as ‘now we have a little bit less rain, now we have a little bit more rain, and what does that mean