I'm at the Brookings Papers on Economic Activity conference and listening to a presentation from Michael Elsby, Bart Hobijn, and Aysegul Sahin that in my opinion is a bit of a blockbuster. They are looking at the reduction in the "labor share" of national income over the past 20-30 years and boring down into detail.
One thing that they find is that the headline decline in this indicator is actually a bit overstated due to technical issues with the treatment of self-employment income. About a third of the total decline, they think, can be attributed to miscalculation. The blockbuster finding, however, is that the remainder is very heavily concentrated in industries that are newly composed to import competition. In other words, the labor share of national income has fallen because many more industries are exposed to foreign competition in a way that's systematically advantaged the owners of capital.
Now go to a rust belt town with this finding, and people are going to say: "That's news?! What the heck is wrong with you economists?!?!"
But the fact is that it is news. Over time, the labor share of national income tended to be constant. And most economists took that stylized fact and put it into their models. That led them to assure people that whatever happened when the country became more open to foreign trade, a fall in the labor share of income couldn't possibly be the result. Yes, some specific workers would lose out but workers on the whole wouldn't. You would see concentrated losses and diffuse benefits. Elsby, Hobijn, and Sahin are saying that's not what happened. The constant labor share was an empirical regularity, but not a law of nature. Joint policy changes in the United States and Asia were enough to substantially shift it and to meaningfully tilt the balance of power inside the American economy to owners of capital.