Perhaps the biggest question in American political economy right now is why middle-class wages have been falling. There are three main hypotheses. Roughly, these are: robots, unions and China.

The robots theory gets by far the most play in the news media, since it's by far the scariest — if automation is replacing big chunks of the human workforce, things are only going to get worse as robots become more capable and efficient. This interpretation has tentatively been embraced by many on the political right, since it doesn't imply a need for substantial government intervention in the economy (though it might imply a need for redistribution). The unions theory is favored by the political left, since it implies that giving more institutional power to this traditional liberal power bloc would shift the distribution of national income toward workers.

Neither side really wants to blame China. The right generally represents business interests and capital owners who have made a lot of money off of China, and hope to make a lot more. The left is afraid to go against the free-trade orthodoxy that has dominated postwar American economic thinking, and also fears a potential cold war with China.

But there's just one problem: The evidence may point to least favored answer being the right one.

A new National Bureau of Economic Research paper by economists Avraham Ebenstein, Ann Harrison and Margaret McMillan examines the impact of offshoring to China. They compare industries and occupations based on their exposure to Chinese offshoring after China's accession to the World Trade Organization in 2001. They find that when exposure is greater, wage declines are much bigger. They also find that competition from Chinese imports affects wages, but to a much smaller degree.

Another paper, by economists Michael Elsby, Bart Hobijn and Aysegul Sahin looks at the China story from a different angle. They ask why the share of income going to labor has decreased in the U. S. They examine two variants of the robots story and also the unions story, and find that these explain only a small part of the decline in the labor share. But when they look at industries exposed to imports, they find that import competition is responsible for most of the variation in the payroll share of value added. Our biggest new source of imports, of course, has been China.

And then there is the famous 2013 paper by economists David Autor, David Dorn and Gordon Hanson, entitled "The China Syndrome: Local Labor Market Effects of Import Competition in the United States." They compare areas of the U.S. based on how exposed they were to Chinese import competition from 1990 to 2007. Their abstract states their conclusion in no uncertain terms:

"Rising imports cause higher unemployment, lower labor force participation, and reduced wages in local labor markets that house import competing manufacturing industries...1/8I3/8mport competition explains one-quarter of the contemporaneous aggregate decline in US manufacturing employment."

In other words, there is a growing body of research showing that globalization — and, in particular, the rise of China — has been the biggest factor hollowing out the American middle class. Naturally, supporters of the robots explanation have challenged some of this research, but the papers keep piling up.

Meanwhile, the robots hypothesis is also starting to get serious pushback on other fronts. Celebrity economist Larry Summers, who has expressed concern over the possibility of automation replacing human jobs, has hedged his bets. He points out that productivity hasn't surged as fast as one might expect from a robot revolution. He also notes that if robots were replacing humans, we'd expect to see a temporary boom in human labor, since people would be needed to build and install the robots. We haven't seen that. Although Summers still believes robots are a factor, he points out some reasons to be skeptical of the story.

So if the U.S. middle class has been gutted because of China instead of robots or de-unionization, what do we do? Reshoring initiatives are becoming popular, but so far they have had limited effect. Trade barriers against China are unlikely to do much, since offshoring investment will simply shift to other low-wage countries — as it is already shifting as Chinese wages rise. And the globalization cat is probably simply out of the bag — now that markets and supply chains are global, walling off American industry will probably just cut American companies out of fast-growing global markets, and lead to slower growth in the U.S.

The only solution to the problem of globalization may be to wait. Chinese wages have risen a lot, and only India is big enough to take China's place. As global economic convergence proceeds, the U.S. will look more attractive as an investment destination, and reshoring will increase. That isn't an answer that people want to hear, but it may simply be the right one.

— Noah Smith is an assistant professor of finance at Stony Brook University and a freelance writer for finance and business publications.