Last week I attended an excellent talk on Taiwan and semiconductors given by my Hoover colleague Glenn Tiffert. It was at the Naval Postgraduate School.
It was quite informative and the questions from the students in the audience reminded me of what I miss most now that I’m retired from NPS.
At a couple of points in his talk. Glenn matter of factly referred to the large decline in manufacturing in the United States in the last few decades. I didn’t correct him for two reasons: (1) I wanted to ask other questions and not monopolize and (2) I think the point he was making might have followed simply from the decline in manufacturing employment, which certainly has happened.
I wrote him and attached two graphs. I haven’t heard back and so I’m writing this up and giving links to the graphs.
First, on manufacturing output, see this graph from FRED. As you can see, manufacturing output in the second quarter of 2023 was only 6 percent below its all-time peak, which it reached in the fourth quarter of 2007.
Second, manufacturing employment has plummeted. See this graph from FRED. Employment peaked at 19.553 million in July 1979 and was only 12.985 million in July 2023. That’s a drop of 33.6 percent. Necessarily implication: labor productivity has increased.
READER COMMENTS
Vivian Darkbloom
Aug 23 2023 at 9:09am
This article from the St Louis Fed (2017) reinforces the point: As a share of real GDP, manufacturing output has remained fairly steady over the past 70 years or so (roughly, 12-13 percent). However, as a share of nominal GDP, it has declined precipitously. This is yet another indication of the relatively larger increase in productivity in the manufacturing sector than other sectors of the US economy.
https://www.stlouisfed.org/on-the-economy/2017/april/us-manufacturing-really-declining
Jose Pablo
Aug 23 2023 at 2:13pm
It seems to be a contradiction between the St Louis FED graph on the evolution of “manufacturing share of real GDP” between 2005 and 2015 and FRED’s graph on “real sectoral output for all workers” for the same period.
In the St Louis FED graph the manufacturing share of real GDP seems to decline slightly for this period (I read, with some creativity, from 12.5% to 11.8%). Taking into account that real GDP in the us increased by around 17% from 2005 to 2015, that should mean that the quantity of manufacturing production grew more than 10% between these two years.
https://fred.stlouisfed.org/series/GDPC1
On the other hand, the “real sectoral output” on the FRED’s graph was basically flat for the same period 82005-15).
What am I doing wrong?
Vivian Darkbloom
Aug 23 2023 at 3:31pm
“Sectoral output” and “gdp” are not precisely the same measures, per the BEA:
https://www.bea.gov/help/faq/1197#:~:text=Value%20added%20is%20the%20difference,domestic%20product%20for%20the%20economy.
So, this may be comparing mandarines and tangerines….
Jose Pablo
Aug 23 2023 at 5:50pm
So, did the manufacturing sector’s output grow in real terms (in terms of quantities produced) between 2005 and 2015 or it didn’t?
Jon Murphy
Aug 23 2023 at 9:18am
In anticipation of some “but without trade, it would be even higher!” objections, I think it is also important to note that this increase in productivity did not happen in a vacuum but because of trade. Look at how quickly manufacturing grows after NAFTA and even after China joins the WTO.
Andrew_FL
Aug 23 2023 at 9:33am
Labor productivity in manufacturing is indeed much higher than in 1979. But since 2010-12 it has stagnated or declined. Even if we allow that production may have increased faster than measured because of quality improvements, unless that has only become a huge factor in the last ten years there’s still a huge change in the rate of manufacturing productivity growth. What happened? Did the Obama-Trump-Biden attempts to revive manufacturing employment result in a lot of relatively unproductive manufacturing jobs?
nobody.really
Aug 23 2023 at 10:40am
Since I read this blog, I already knew the punchline that manufacturing employment diverged from manufacturing output.
But I didn’t know that output had declined since 2017. What’s up (down) with that?
Richard W Fulmer
Aug 23 2023 at 11:32am
The way I read the FRED’s chart (OUTMS), the all-time peak was in Q4 2007, just before the Great Recession, and the next peak was in Q3 2018, not in 2017. That quibble aside, your point is still well taken. Q3 2018 was over a year before the pandemic hit. Why was there a significant decline (~5%) in manufacturing output in 2019?
Jon Murphy
Aug 23 2023 at 11:39am
The drop from 2018 (the peak) to 2020 is a combination of Trump’s trade war and the pandemic.
Richard W Fulmer
Aug 23 2023 at 12:12pm
I found several articles that chalked up the 2019 decline in manufacturing output to Trump’s trade war. It’s ironic, though not surprising, that his solution for increasing domestic manufacturing backfired. Industrial policy doesn’t seem to work any better when it comes from the right than when it comes from the left.
Richard W Fulmer
Aug 23 2023 at 8:04pm
Trump first announced the first of his tariffs in January 2018 and others followed that same year. Some of our trading partners began retaliating in 3Q 2018 – the same quarter in which U.S. manufacturing output began falling.
Trump has just promised to continue his attack on, er…, his support for American manufacturing with his announcement that he will impose a 10% tariff on all foreign goods.
Jon Murphy
Aug 24 2023 at 9:29am
If only there were some vast empirical literature that could have predicted this outcome.
Jose Pablo
Aug 24 2023 at 11:31am
Without Trump’s enthusiastic help, international trade (“globalization”) had been already flat or declining since 2009. Around this period the productivity of the manufacturing sector in the US also plateaued.
Since one way of increasing labor productivity is “composition” (offshoring low productivity manufacturing activities and developing at home high productivity ones), could this be the, at least, a contributing factor to the manufacturing productivity evolution?
Certainly, Trump policies didn’t help, but the trend was there before him. You can even say that he capitalized on this trend/mindset to win the elections (as any potential candidate will have to do if he/she wants to win the elections in 2024. Voters shape politician’s thinking, not the other way around)
Jon Murphy
Aug 24 2023 at 12:09pm
That’s not quite true. Obviously global trade fell in 2008-2010 because of the global recession. But it was rising again after that, surpassing the pre-recession peak in 2011. We see a substantial drop in 2015, but that’s because of the collapse in oil prices at the time, I think.
Jon Murphy
Aug 24 2023 at 12:15pm
(My point is there is a lot of variation in the trend. It’s misleading to say it has been declining since 2009).
Jose Pablo
Aug 24 2023 at 12:22pm
https://www.imf.org/en/Blogs/Articles/2023/02/08/charting-globalizations-turn-to-slowbalization-after-global-financial-crisis
The IMF seems to think otherwise
Jose Pablo
Aug 24 2023 at 1:32pm
In any case, it seems to be an inflection point in 2009 (it seems that from that point on, international trade plateau, grow at a slower pace or wave around).
My question was more about the “transmission” mechanisms of this fact into manufacturing labor productivity growth. Labor productivity is (mainly) driven by changes in TFP, capital intensity or the composition effect I was mentioning.
It could be interesting to understand how the different factors behind manufacturing labor productivity change over time (which should help to clarify what is causing what). Any interesting references on this topic?
Jon Murphy
Aug 24 2023 at 4:17pm
No, their graph agrees with me. The title of the post, though, is making the same linguistic mistake.
Jose Pablo
Aug 23 2023 at 1:52pm
Any data on the evolution of capital stock / capital services in manufacturing in the US for the same period?
Is this historical increase in manufacturing productivity a labor productivity increase or an increase in TFP?
[A quick internet search didn’t yield any interesting data on the TFP for “manufacturing” AND “the US”]
Mark Z
Aug 25 2023 at 12:10am
I think Susan Houseman argued on an econtalk episode that growth in productivity in manufacturing has been largely confined to high tech computer related industries, which don’t necessarily meet the conventional definition of manufacturing, e.g. software production I believe is classed as manufacturing. If one looks at traditional manufacturing, e.g. of physical goods like cars, she claimed, if I recall, that productivity has been pretty stagnant for decades and % of GDP has declined significantly.
Mike Burnson
Aug 27 2023 at 9:25am
I apologize for coming so late to this party.
As always, the devil is in the details. Note the significant increase in employment under Carter, who then was crushed by Reagan in 1980. What happened was that inflation had people withdrawing from savings then limited to 5.25% interest at S&Ls, and spending while the IRS subsidized personal debt. The result was an “increase” in GDP and employment that voters saw through easily.
Manufacturing underwent an enormous overhaul as EPA dramatically restricted production by increasing overhead. Enormous volumes of production did shift overseas. Reagan imposed quotas on imported autos, so most foreign motor companies built facilities in the USA. Automotive part production shifted overseas to avoid EPA cost problems, while car ASSEMBLY in the USA increased. Similarly, computer components were primarily manufactured domestically and shifted overseas as foreign capacity was increased. In these and other cases, what is defined as American manufacturing?
Another question is the unique distinction between “manufacturing” and “service”. I have worked 45 in metal finishing (plating, anodizing, painting). Auto companies stopped most of their in-house metal finishing (manufacturing) and jobbed it out to metal finishers (service). The exact same part was assembled into the auto, but its production shifted between manufacturing and service.
Comments are closed.