It doesn't matter if a theory is correct when "nobody knows what that sentence means!"
Blyth isn't dismissing the concept of time inconsistency - he's pointing out that models that rely on people making "rational" decisions are total junk if nobody actually understands the required concepts. Models that assume people would change their behavior because of time inconsistency should never have been considered in the first place because most people don't know what the term even means.
In a similar manner, I'm suggesting that while it would be nice if we could set up "markets" that self correct or allow problems to be handled locally, this might not be possible in practice.