I have often read or heard the assertion that there is no respectable work by economists that attributes an important part of rising inequality in the U.S. to international trade, with the implication being that it's all (or mostly) due to skill-biased technological change. Greg Mankiw has made this argument in the past, and Alan Blinder implies as much in his recent NYT article.
I have never understood why the work of Rob Feenstra and Gordon Hanson is overlooked in this context. These two are among the very best empirical trade economists today (and Feenstra is the author of the most widely used graduate-level textbook in trade). In a series of papers, they have argued that outsourcing and global production sharing act just like skill-biased technological change, and they have played an important role in shaping wage inequality. Their empirical work is careful and driven by a compelling theoretical model of within-industry specialization.
To get a flavor of their results, here is how they summarize their empirical work:
Thus, whether outsourcing is more or less important than computers depends of whether the latter are measured as a share of the capital stock, or as a share of investment. It is fair to conclude that both these variables are important explanations of the shift towards nonproduction labor, with their exact magnitudes depending on how they are measured.
Importantly, their framework helps explain how globalization contributes to inequality in both developed and developing countries.