I wrote a diary / blog / whatever a little over a week ago pointing out that after 35 years of stable employment, the US shed over 5 million jobs in manufacturing, or about 1/3 of its total. This was the main graph I referenced:
I pointed out that the job losses are real and devastating to the communities they happen in, and they occured after NAFTA was passed. The blog was met with a high level of dismissiveness, much more than I expected. The usual reply was that there was a ‘wave of efficiency’ that resulted in the job losses, and that the impact of outsourcing was otherwise minimal. Others pointed out that it wasn’t until China was admitted to the WTO that the job losses hit. I’m not sure how / why that changes anything — if people are pissed about trade agreements leading to job losses, pointing out that they went to China not Mexico doesn’t change much. Besides, it was Bill Clinton who signed NAFTA, and everyone remembers that. Screwing around with TPP should be poison. But, many people seemed to differ.
But, there was one chart that was referenced in the comments that I have to admit threw me for a loop. It is Real Manufacturing Output from the St. Louis Fed:
As you can see, manufacturing output bounded back! See, everything is great! All those people at Trump rallies complaining about closed mills and closed plants are simply inevitable victims of efficiency. Output at or near all-time highs, 1/3 less people, more efficient!
Except, not. The good news, or sad news, is that after a whopping ten minutes of poking around the fed data, I traced the data to the BLS, and the efficiency myth is just that, a myth. Here is the thing, if you read the notes its not too hard to figure out that Real Manufacturing Output is just that — the value of the output of manufacturing. To see what is really going on, you have to look at the inputs. So check this out:
This chart plots real manufacturing output and all of its components. It isn’t inflation adjusted, so it rises sharper than the prior version, but we can work with it. If you look at the chart, all of the components trended similarly to the total. Then, right around 2000 (i.e. the same time that employment took a dive), one of the components broke from the pack and was clearly driving the total. And that component is (drum-roll please): Materials!
It is pretty clear from the shape of the ‘Materials’ line that it is now driving the total. In fact, if you subtract materials from the total, you get this:
Take out the cost of materials, and you get a different picture. You see growth until 2000, then a decline in the Net (‘Real Manufacturing Output’ minus ‘Materials’). Real Manufacturing Output continues to rise because it only focuses on output value. If the price of steel skyrockets, real output increases. If a cabinet maker stops manufacturing panels and instead only finish-assembles pre-cut boards shipped in from overseas, real output still rises. In both cases, stagnation or decline in activity is masked by increase in price of the final product. Metrics for economic activity developed 30 years ago no longer work. So, sorry to bust any bubbles, but manufacturing hasn’t rebounded and that ‘wave of efficiency and automation’ was little more than a bakery firing all of its workers and leaving a skeletal staff to write “Happy Birthday” on cakes shipped in from overseas.