Personally, I don't think something as old and as widespread as DSGE can be called "cutting-edge". Introductory macro, like any other introductory course, is a

Personally, I don't think something as old and as widespread as DSGE can be called "cutting-edge". Introductory macro, like any other introductory course, is always at least twenty years lagging behind standard theory, as used by large economic organizations (e.g. central banks, international agencies), and at least some twenty five to thirty behind the cutting-edge stuff. At least, that's the rule of thumb that I was taught. It's unfortunate for undergraduate students, who really don't get a taste of the hardcore econ. But I guess that's just how it is. But about Ricardian comparative advantage: it's a bad model. It's a gem of economic thinking, just like Ricardo's differential theory of rent. But as model, it just ignores so many things (pretty much all that you've pointed out) that it really isn't useful beyond a "first-order approximation" (much like IS/LM models). Ironically, it's a fine short-run model. You've asked about what the long-run is. Well, the long-run is when all inputs can be changed by the firms, whereas the short-run is when they can only change some of the inputs (e.g. labor and raw materials, as opposed to machinery). It doesn't really have a direct relationship with time, but the idea behind is that the varible inputs can be changed in a shorter time scale than the fixed inputs. But when everything but labor is fixed, that's just when comparative advantages kick in. On the other hand, comparative advantage simply can't handle stuff like technological advance, resource exhaustion, or pretty much anything which might change the productivity in any of the countries we're looking at. It wasn't conceived to deal with any of that, either. It was pretty much Ricardo's counter-example to the issue of whether trade would be profitable in the presence of absolute advantage, and one of his pet arguments against the Corn Laws.