Import substitution, good or bad?

Free exchange and Dani Rodrik have both made intelligent posts on the issue of import substitution. Free exchange sticks to the common line that import substitution is bad, Dani says that there is evidence that it is good.

I know very little about any of this, but I’m going to say something anyway. As far as I can tell, trade policy should work off the idea of comparative advantage, implying that each country should make what the good they are ‘relatively’ better at making (specializing in goods with the lowest opportunity cost). As a result, government policy should react in ways that take advantage of this concept.

This might imply to some people that government should not intervene in trade, and just let the free market choose the most efficient industries, which will in turn trade with the rest of the world. However, I’m not sure I fully agree.

It is possible that an industry that would have a comparative advantage in trade terms may not have been founded given high fixed costs and the requirement of skilled and experienced labour which will only be created when the industry exists (infant industry type argument). If the government can recognise these industries, it can subsidise their creation until they become fully efficient, by which times they will be net exporters.

Now this form of intervention isn’t the same as import substitution (although it is often placed as a subset of it). Import substitution involves creating the goods you import at home, now if this is a good where another country has a comparative advantage then all you are doing is hurting yourself and the other country. Import substitution is a bad idea (unless there are security of supply or political issues), but government policy to develop domestic industries does have some potential.