Thursday, February 21, 2008

Rethinking Comparative Advantage

I was watching the curling last night when a thought struck me. Our trade theory is generally premised on comparative advantage. This has certain merits. It has certain flaws.

Take two goods, A and B. Korea, China, Egypt, etc, etc, etc, may be able to produce both of them more cheaply than your favourite OECD country. Ricardo would have it that pareto-improving trade occurs as each country produces more of the good in which they have comparative advantage. Blah blah blah blah blah. We've heard this before.

Consider a world economy with a practically infinite number of industries, and firms restrict themselves to a single industry (or a small enough number that the fraction is negligible). Firms are non-cooperative. None of these firms care about comparative advantage - only absolute advantage.

Further suppose that the world is operating at less than full capacity (inside its PPF) - lots of unemployed, poor people with minimal capital. Realistic. Globalization suddenly makes capital less sticky. Each of our firms looks only to its own bottom line and relocates to the poorer, less costly, countries, ignoring the potential comparative advantage relationships. Firms don't care about the global planner perspective, only the absolute view.

I'm trying to reconcile this with other constraints to see if an 'overshooting' position is possible - whether we can observe in the short run an exodus of production to less developed countries above and beyond what we would in the very long run equilibrium in which comparative advantage holds strongly. This en-masse move of capital will realign factor prices and whatnot, and I think it relies upon very sticky prices/wages in rich countries. Unfortunately, I have no rigorous line of thought, it's all graphs and no equations. Interesting, though.

POSTSCRIPT: I forgot to mention the drawback of the BC carbon tax. If leaving the climate-change file to the provinces leads to a mishmash of different taxes, regulations, etc, that is a problem.

No comments: