Duke University economist Ed Tower is on a private list of people to whom I send by e-mail my letters-to-the-editor. And so Ed received my letter today to the New York Times that I mention here  (but that I cannot yet post publicly).
Here’s the bulk of an e-mail that Ed sent to me after he received my letter; I share it with Ed’s kind permission:
Don, I think a casual reader of today’s letter could reasonably conclude that you would favor US protection if Mexican workers were paid one tenth of what American workers are paid if the Mexicans were equally productive. You seem to defend free trade on the grounds that Mexican labor is less productive. But Paul Krugman in his “What do Undergraduates Need to Know About Trade” argues that higher productivity abroad does not affect the US standard of living, aside from terms of trade effects. Suppose Mexican workers were equally productive as US workers and were paid one tenth as much. In a one commodity world, protection doesn’t buy or cost anything. In a multi-commodity world where Mexican productivity/US productivity differs between sectors, higher Mexican productivity buys a higher standard of living for Americans for it will improve the American terms of trade as Mexican exports flood the American market and Mexican imports drain the American market.
For example, suppose Mexican regulations mean that an industry we will label “dirty” is relatively more productive in Mexico than in the US as compared with another industry named “clean.” Then trade with Mexico leaves America with a higher material standard of living and a cleaner environment.
I agree with Ed. I emphatically do not believe that free trade is justified only if foreign workers are less productive than are American workers, or only if the low pay of foreign workers reflects those workers’ low productivity. That is, I would not support U.S. protectionism even if the productivity of foreign workers were as high as that of American workers while the wages of foreign workers were much lower.
For both economic and ethical reasons I strongly and unconditionally support a policy of unilateral free trade.
Just for the record (because I’m here writing about the matter): all talk of worker productivity (or, alternatively, of total factor productivity) at the level of countries is fraught with the possibility of great confusion. Talking about “the” productivity of (say) American workers can make sense, but it must be done and heard with care.
The workers of a country do not collectively produce a glob of a homogenous good, some portion of which is then exported. Instead, each worker assists in the production of particular and different outputs. The economic viability of Jones working in the shoe factory does not depend only on the value of the output that Jones helps to make possible by his working in that factory. Also relevant is the value of the highest-valued output that Jones does not produce by his working in the shoe factory.
If by working in the shoe factory Jones’s enables the market value of that factory’s output to rise by $50 per hour, we cannot, without more information, determine if, economically, Jones ‘should’ be working in that factory. To make this determination we must know the values of what Jones would produce were he to work at other tasks.
All of the above is an elliptical way of saying that although American workers as a group are, on average and by conventional measures, more productive than are (say) Mexican workers, what matters for determining trade patterns between the U.S. and Mexico is the economic productivity of specific American workers at producing each of countless particular goods and services compared to the economic productivity of specific Mexican workers at producing each of these countless particular goods and services.
Suppose an American worker can produce an additional $50 of value per hour producing shoes while his Mexican counterpart can produce only an additional $40 of value per hour producing shoes. It’s tempting to say that the American worker is “more productive” than the Mexican worker at producing shoes and, therefore, should work at supplying shoes to Americans and Mexicans alike.
But suppose that this same American worker can produce an additional $100 per hour of value producing chairs while his Mexican counterpart, if set to producing chairs, can produce an additional $30 per hour of value.
Clearly, everyone – Mexicans and Americans alike – are made better off if the Mexican worker specializes in producing shoes while the American worker specializes in producing chairs.
Contrary to superficial appearances, the American is not economically more productive than is the Mexican at producing shoes. The Mexican is more productive (compared to the American) at producing shoes. The American is more productive (compared to the Mexican) only at producing chairs.
This hypothetical example – which depicts nothing but the operation of comparative advantage  – makes plain that the details of trade patterns are not caused by overall productivity but by comparative advantage. It makes plain also the more fundamental reality that even if American workers were absolutely more productive (whatever this phrase might really mean) than are foreign workers at the production of every conceivable good and service, American workers would nevertheless have a comparative advantage only at producing a subset of these goods and services. Foreigners would have comparative advantages at producing the rest.