Via Greg Mankiw, we find Alan Blinder qualifying his support for free trade. Greg seems to be taking it personally. Rather, we should just take his suggestion (my emphasis below) to its natural conclusion:
Mr. Blinder's answer is not protectionism, a word he utters with the contempt that Cold Warriors reserved for communism. Rather, Mr. Blinder still believes the principle British economist David Ricardo introduced 200 years ago: Nations prosper by focusing on things they do best -- their "comparative advantage" -- and trading with other nations with different strengths. He accepts the economic logic that U.S. trade with large low-wage countries like India and China will make all of them richer -- eventually. He acknowledges that trade can create jobs in the U.S. and bolster productivity growth.
But he says the harm done when some lose jobs and others get them will be far more painful and disruptive than trade advocates acknowledge. He wants government to do far more for displaced workers than the few months of retraining it offers today. He thinks the U.S. education system must be revamped so it prepares workers for jobs that can't easily go overseas, and is contemplating changes to the tax code that would reward companies that produce jobs that stay in the U.S.
Fantastic. We can be a nation of barbers, gardeners, and custodians, and we can enforce this nirvana by favoring it in the tax code.
Jobs have many characteristics. It is true that for the purposes of talking about jobs in trade policy, economists often collapse these characteristics into a single characteristic, the wage at a point in time. Blinder is correctly pointing out that another characteristic is the risk associated with that wage in the future. That risk could come from a number of sources, of which foreign competition is just one.
When people choose jobs, they have the freedom to trade off among the characteristics embodied in each job. Standard economic analysis would suggest that, for jobs that require the same degree of skills, those that offered more risk would also have to offer higher average wages to compensate. Starting from an equilibrium in which workers have information that is no worse than the government about the terms of this tradeoff, a policy that favored less risky wages would necessarily generate lower expected wages. What's the compelling interest by the government to justify this shift in outcomes?
I can think of two. First, one could assert that the government has better information than the public about relative risks and rewards. Second, one could assert that the social cost of risky jobs is higher than the private cost, i.e. that the government bears a fiscal cost of people being employed in jobs with higher compensation risk. I am skeptical in both cases, but I would be interested in hearing other perspectives.