Here’s a great interview with inequality scholar Branko Milanovic wherein he brings a much-needed historical and international perspective to the debate (h/t: C. Marr). Many of Branko’s points are familiar to my readers: yes, increased trade has upsides, for both advanced and emerging economies. But it’s not hard to find significant swaths hurt by globalization, particularly workers in rich economies who’ve been placed into competition with those in poorer countries. The fact that little has been done to help them is one reason for president-elect Trump.

As Milanovic puts it:

The problems with globalization arise from the fact that gains from it are not (and can never be) evenly distributed. There would be always those who gain less than some others, or those who lose even in absolute terms. But to whom can they “appeal” for redress? Only to their national governments because this is how the world is politically organized. Thus national governments have to engage in “mop up” operations to fix the negative effects of globalization. And this they have not done well, led as they were by the belief that the trickle-down economics will take care of it. We know it did not.

But I’d like to focus on a related point from Branko’s interview, one that gets less attention: the question of whether it was really exposure to global trade or to labor-saving technology that is most responsible for displacing workers. What’s the real problem here: is it the trade deficit or the robots?

Branko cogently argues that “both technological change and economic polices responded to globalization. The nature of recent technological progress would have been different if you could not employ labor 10,000 miles away from your home base.” Their interaction makes their relative contributions hard to pull apart.

I’d argue that the rise of trade with China, from the 1990s to the 2007 crash, played a significant role in moving US manufacturing employment from its steady average of around 17 million factory jobs from around 1970 to 2000, to an average today that’s about 5 million less (see figure below; of course, manufacturing employment was falling as a share of total jobs over this entire period).

Source: BLS

The relative trade balances, shown in the next figure, underscore this point, as China’s surplus grew sharply in the 2000s while the deterioration of our trade deficit accelerated (source: Autor et al; I’ll discuss the reversal of these trends below).

Source: Autor et al

Finally, the “robots did it!” story requires an acceleration in productivity growth. The next figure shows manufacturing productivity growth (yr/yr) since 1987 (also, read Sue Houseman’s very important work for the full story here). It’s awfully jumpy, so I’ve plotted a 5-yr rolling average. There’s a bit of acceleration in the 1990s or 2000s, and I’m not denying automation has played a role in the pattern in the employment figure above. But it’s unlikely to be the whole story and that story is particularly strained today, as manufacturing productivity growth has crashed in recent years.

But my main point here is that it’s a mistake to either believe that trade and technology are clearly separable forces, or to think that it matters to workers which force is displacing them. Too often, policy makers seem to assert that, because “it’s not trade, it’s technology!”—typically offered without much evidence—displaced workers should somehow be assuaged. Hey, all they need to do is go from running a drill press to designing, building and programming drill-press-running robots!

True, trade and China are much more tangible targets, and I get that they sometimes carry more symbolic weight than they should. That’s particularly true regarding China right now, which is trade enemy #1 in the Trump playbook, despite the fact that their trade surplus has been falling and, if anything, they appear to be trying to prop up the value of their currency (the trade play is to depreciate).

But remember, automation and labor-saving technology are ongoing forces that have been with us forever. Historically, increased demand for the extra output created by faster productivity growth absorbed displaced workers, sometimes in new sectors.

We must understand why that hasn’t happened in recent years. The productivity evidence over the past decade contradicts an accelerating automation explanation. I think the evidence instead supports an explanation that exposes false assumptions:

–the winners from expanded trade would compensate the losers
–regressive tax cuts would trickle down
–trade deals centered on corporate interests would somehow help laborers
–full employment would automatically persist (even in the face of large, growth-draining trade deficits)
–austere fiscal policies amidst weak private demand would magically have anything other than their predicted, negative effect on growth
–the safety net, minimum wages, and other labor standards must be diminished to create the right micro-incentives

All of the above distribute income upward, and not just income, but just as importantly, power and the political influence that makes reversing these false assumptions an extremely heavy lift. The system effectively insulates itself from progressive change.

I don’t have the answer to the question of what breaks this extremely damaging chain, but I’m sure it involves some serious organizing of the many hurt by this dominant agenda against the minority helped by it. At any rate, a good place to start is a clear-eyed identification of the problem.