Via Brad DeLong and Mark Thoma, here is Mark quoting Brad:
But as Rodrik points out, "…saying that the impact of globalization
on advanced-country labor markets is quantitatively rather small in the
real world and is overshadowed by other phenomena (such as
technological change) is no different [in the Heckscher-Ohlin-Vanek
framework] from saying that the gains from trade have in practice been
small." There is a problem of cognitive dissonance here.
I worry about this, but I am not so sure that trade advocates have painted themselves into a corner. Trade can improve the global economy in at least three ways: a) factor price equalization and the resulting higher output, b) spreading innovations, new technologies, and new products, and c) by improving domestic politics. The existence of b) and c) means that the gains from trade can in principle be large while the factor price equalization effect is relatively small. Factor b) points us toward a very favorable opinion of trade. Rodrik of course also worries about c):
How does Rodrik believe that globalization undermines social democracy?
First, because globalization has undermined governments’ ability to
carry out social insurance programs.
I’m not sure that Rodrik’s view is so uniform on this question; for instance he has a JPE piece suggesting that more open economies are more likely to be interventionist. In any case my view is that the wonders of Sweden, Denmark and Norway rely very heavily on external trade. Note that "openness" and "smallness" are distinct but correlated variables here; most likely both qualities are necessary for welfare states of the Nordic kind. Or look at it from the other side. In recent times (though it has been changing) Brazil and India were relatively closed to trade, and I don’t see that it led them to take better care of their poor. International trade also makes countries more tolerant and it makes people more interesting. That’s the old classical liberal case for trade from Wilhelm von Humboldt and Richard Cobden; let’s not toss it out just because Heckscher-Ohlin came along.