I agree with these generally, especially the last part about manufacturing jobs not inherently being good jobs. But, on the point about manufacturing employment declining in the 2000s, I’m not sure the decline wasn’t specious. Even though the decline in manufacturing jobs coincides with an explosion in the trade deficit… so what? Perhaps there was also a one time spike in manufacturing productivity, or there was an accelerated shift to employment in IT as the Solow Paradox ceased to be a paradox.

I think the last sentence of that section reinforces the point— erasing trade deficits really wouldn’t do much to bring back manufacturing employment (Krugman’s calculation was that you might be able to get from 10% of GDP to 13% by eliminating trade deficits). And the last point further reinforces that we probably shouldn’t care about it at all, given that, as you correctly point out, unions made manufacturing jobs good, not some magic property of manufacturing.

So perhaps the messaging should continue to be that we should unionize Starbuckses and Amazon warehouses, not grasp at manufacturing as some kind of magic salve.