From an interview of Ed Prescott: “People can quantify what gains there are from it [trade]. If you calibrate the models… most people want to get a big number, but a small number comes out.”
Ed explained that the importance of the difference between openness and free trade lies in explaining the big gains that “trade” generates. Empirically we know that periods of openness coincide with periods of strong economic growth and periods of protectionism coincide with recession. Yet the traditional models of trade don’t bear those big-gains results.
There are three theories traditionally used to explain trade, Ed explained:
The first is the Heckscher-Ohlin factor endowments model. China has a lot of low-skill workers so they produce goods that are labor-intensive, and since the U.S. has a lot of skilled workers, we produce goods that are skill-intensive. But the gains according to that model turn out to be small.
The second model is David Ricardo’s comparative advantage. It’s the textbook example: England had a comparative advantage in wool and Portugal had a comparative advantage producing wine, so England produces wool and trades for wine and Portugal produces win and trades for wool. But that model also yields small gains from trade.
Then there is the increasing returns to scale model from Paul Krugman, who use the Dixit-Stiglitz monopolistic competition model to explore the potential gains from increasing returns. Yet that, too, turned out small gains.
So clearly there’s got to be some other reason that trade yields big gains for the economies that engage in it. The answer is that “trade” is about much more than the exchange of goods. With openness, there is diffusion of knowledge.
[This isn't a transcript, but it's an accurate paraphrasing of the original audio.]
For two examples of such calculations, I’d look at Bernhofen & Brown (AER, 2005) and Broda & Weinstein (QJE, 2006). The former uses the minimal framework of putting an upper bound on the equivalent variation by looking at autarky prices and a counterfactual import vector. The latter impose more structure by using a CES demand system and look at the gains from new imported varieties.
Now, I won’t dispute that technological spillovers and knowledge diffusion are additional channels offering more gains from economic exchange. But how does Prescott know that the gains from trade are bigger than those estimated using the theories above? What is his benchmark? How could one quantify the gains from trade without using some theory?