(1) Measuring manufacturing employment is tricky. Manufacturers now use temp workers to man assembly lines. Even if they do the exact samething as fulltime workers they're still counted as 'service' workers.
I wonder if it was always hard to classify workers. Or if this is somehow an example of once you start using measurement tools its no longer valid
(2) Manafacturing output is measured as the value of outputs less the cost of inputs. So if US factories become better at sourcing cheap inputs from Mexico/China it looks the US has become more productive (this is similar to mismeasurement
(3) People talk about how US output has grown even as we outsource more to China so we must become more efficient. But almost all the growth in US mfg was from computer chips. And the value-add from computer chips goes up because the R&D/design for chips improved. Not the mfg process
Overall two interesting things coming out of the discussion. The first is maybe the idea that we lost jobs from automation more than outsourcing is exaggerated.
The other is in general there are three ways to improve your mfg. Become more efficient at sourcing inputs. Becoming better at designing outputs. And improving the process. People are biased towards thinking only the third thing is important. And I'm inclined to agree. Cheap sourced inputs made in China today could go away tomorrow.
But on the other hand wasn't an obsession with raw quantities and lack of attention on effeciency and desirability part of where the 5-year plans went wrong. And where the genius of capitalism is apparent?