Roberts is Half Right
In his articles ("The Trade Question," The Washington Times 8/8/03 and "Second Thoughts on Free Trade," nytimes.com 1/6/04) attacking the application of the Ricardian argument in favor of free trade to the modern world, Paul Craig Roberts is half right.
It seems to me his critics such as Bruce Bartlett et al. have misconstrued his argument at least on one score. The Ricardian theory of comparative advantage does only apply under the assumption of absolutely rigid factor immobility between countries.
However, Roberts is also half wrong. Even though an outflow of domestic capital, all other things equal, causes real wages in the capital exporting country, e.g. the U.S., to fall or not to reach levels they would have reached in the short run, there is an increase of labor productivity and therefore output in the capital importing countries, e.g. China. China therefore develops an increased real demand for U. S. exports which raises prices and real wages in these industries.
In addition, the allocation of capital to more value productive uses throughout the world intensifies the division of labor and allows the U. S. labor force to specialize in even more value productive industries raising real wages further in the U.S. and throughout the world. Thus, in the late 19th century the enormous flow of capital from Great Britain and Germany to the industrializing U.S. made products more abundant and real wages higher for British and German workers as well as for American workers.
The same result occurred when capital and "jobs" flowed from the East coast to the West coast of the U.S. in the late 19th and first half of the 20th century. Somehow Roberts misses this point.
Three other empirical points that tell against Roberts' argument: 1. since the mid-1980's the U.S. has been a net importer not exporter of capital, which is the mirror image of our huge current account deficits, so this argument is empirically inapplicable in any case ; 2 it appears that Roberts focuses on the loss of jobs in one (manufacturing) sector and ignores the addition of jobs in U.S. export industries—but this redistribution of jobs is precisely what is to be expected from a worldwide intensification of the division of labor; 3. it is the change in a nation's real wages and not the number of jobs gained or lost in selected industries that indicates the economic optimality of any change in the economy. It is curious that Roberts, a pretty good economist from my reading of him, does not address these obvious points in his articles.