This weekend's project: theories of trade
Following the president's protectionist State of the Union speech, and a lengthy discussion this blogger had with a friend who is an Andy Grove- or MIT Tech. Review-species of trade skeptic, I decided to do some reading into the foundational economic background on trade theory.
At some point a while ago, this sort of thing became my idea of fun. Well, my priors had always led me to make full-throated defenses of free trade, and my introductory micro textbook gave little time to "protectionism," except to dismiss the infant industry argument as a "more sophisticated argument" for it, implying obviously that it was specious and ultimately just as wrong.
After doing a lot of reading and mulling over, and I only link to here the best of my research finds, I remain, fundamentally a Ricardian--but not a fundamentalist Ricardian. Yes, I do think that, in general, the principle of comparative advantage and mutual gains from free trade do hold. And while I think that protectionism makes little sense--I'm thinking of Romney on currency manipulation, or Obama on trade dumping--I do see a place, albeit limited, for an American industrial policy in high-tech manufacturing.
It turns out, from my research, that there are two broad, and largely compatible, "schools of thought" on trade: classical (i.e. Ricardian) and New trade theories. The former emphasizes comparative advantage, the production possibilities frontier, and the plain-as-day conclusions of microeconomic social welfare analysis on tariffs and gains from trade. (If any of these terms are unfamiliar, as I imagine they are, they are only a Wikipedia search away.) Ricardian trade theory remains dominant to this day, but the New trade theory has added some very interesting insights, which have assimilated rather well.
What is "new" in New trade theory? First, it scraps a number of assumptions (perfect competition, zero transport costs, constant human capital). In their place, it assumes some nastier conditions -- and by "nastier," I mean less ideal in terms of mathematical theory, and more realistic: imperfect competition, non-negligible transport costs, "industrial commons" externalities on human capital and competitiveness. The conclusions are pretty interesting, as they imply that to the extent that the world behaves more closely to New trade assumptions rather than Ricardian ones, that industrial policy makes sense.
The way these theories are combined is by the fact that the rivaling assumptions are both true in certain senses, in certain industries, and at certain times. In other words, if high-tech electronics production fits more in with New trade economic models -- and it does so appear -- then an industrial policy may gain traction there, even as that same nation pursues a Ricardian strategy in services industries, where Ricardian assumptions seem more sturdy today than ever before. The world, by the way, is becoming increasingly Ricardian in reality, but there are important exceptions in industrial areas like Germany's Mittelstand, which is an economic ecosystem of small high-tech manufacturers which serve niche export markets.
Update (2/18/12): VoxEU has a new article which describes empirical success for an "industrial policy lite" in Britain, targeted at small firms which served national/international markets. Hmm. This seems to fit in very nicely with what I've discussed here.