U.S. manufacturing employment does fine with more robots. It doesn’t with a skewed trade balance.
A common refrain in economic commentary on the nature of work, and the changes in manufacturing employment, is basically:
“Trade didn’t take those jobs. It was robots. The robots did it!”
Okay, maybe it’s not the terminators who are doing it. Rather, it’s industrial automation: We make more stuff with less people, because manufacturing is now performed by automated processes. This idea is breezily inserted into all kinds of articles, including ones with other interesting things to say. (And, in defense of those who spread this idea, there are studies that back this up.)
But this isn’t the entirety of thinking on the subject. David Autor – one of the academics behind the “China shock” hypothesis – has pumped the brakes on this idea. And others – from those at the Democratic-aligned Brookings Institute to President Trump’s hawkish chief of the National Trade Council, Peter Navarro – have pointed to the real-world example of Germany, which has put a lot of robots on the assembly line in recent years but hasn’t seen its substantial manufacturing workforce shrink.
So what gives? How does this refrain persist? Well, a new report by the Information Technology & Innovation Foundation (ITIF) takes a look at the “shaky foundations” it stands on:
First, while productivity growth in manufacturing compared to the rest of the economy was fairly consistent from 1990 to 2010 (25.8 percent in the 1990s, and 22.7 percent in the 2000s), job losses in the latter decade were 10 times greater than in the former.
That’s a very good point that has been made by others, albeit not pointed out enough. What else?
Second, government statistics significantly overstate manufacturing productivity growth. Government data shows an astonishing 179 percent increase in computer manufacturing output from 2000 to 2010. But companies actually produced fewer computers during this time, not more. The discrepancy is explained by the fact that the massive growth in productivity represents increasing computer processing speeds. The resulting mismeasurement of this subsector significantly skews manufacturing statistics overall. Indeed, after removing computer and electronic products, it turns out that real value added from U.S. manufacturing grew just 6.4 percent from 2000 to 2015, not the reported 19.3 percent.
The report’s author, Adams Nager, suggests that an explosion of trade with China – China, not Mexico – and its suite of mercantilist trade policies was a much more likely culprit for the sudden drop-off in American manufacturing employment.
Nager points to a number of other scholars who have drawn the same conclusion.
In sum: Yeah, the robots are coming, man. But are they actively stealing our jobs? It’s just as likely (if not more) that it’s a foreign government that’s trying to do that.