> I think one of the epistemological problem with economics is that it's almost impossible to run experiments[...]
It's also true of cosmology, though cosmological errors won't ruin people lives :) If you haven't read "the shock doctrine", it states that actually there was many economic experiments, in Chile and other countries, usually for the worse... Well USSR made some large scale experiments too, from 1917 to 1921, then 1921 to 1927.
(Coming in a day late.) I'm not going to trust the Shock Doctrine after reading about how its "facts" were collected.
I think one of the epistemological problem with economics is that it's almost impossible to run experiments(except maybe college kids experiments), and it would probably be irresponsible to run experiments on whole nations. That why some economists approach it by using entirely deductive logic as a mean of study.