I’ve noticed that many people reflexively blame trade for the decline of the Rustbelt. Here’s one example:
Apologists for the outgoing trade regime often ignore that its impact was felt most acutely in particular regions, like the American Midwest. Researchers John Russo and Sherry Linkon describe how the closure of a steel mill in Youngstown, Ohio – the first of a wave of closures in the region – undermined the sense of worth and optimism among residents. Many can still recall better days, when employment was high, jobs paid well, workers were protected by strong unions and industrial labour provided a source of pride – not only because it produced tangible goods, but also because it was recognised as challenging, dangerous and important.
But is that true? Does trade explain the decline of steel employment from roughly 190,000 to 84,000?
If trade explained the loss of employment in steel mills, then you would expect to have seen a precipitous decline in domestic steel production. In fact, there’s been very little change in steel output during a period where employment has plunged sharply:
This is not to deny that imports have had some impact on employment in manufacturing. But the primary cause of job loss has been automation. And with AI set to revolutionize manufacturing, employment in manufacturing will continue to decline sharply over the next few decades even if we were to entirely eliminate all imports.
PS. This post does a nice job of explaining why bringing back manufacturing jobs is harder than it looks.
READER COMMENTS
Jon Murphy
Apr 13 2025 at 9:26pm
There’s a paper in JPE (I think) from about 2 years ago that shows the rust belt was due more to unionization forcing jobs to other parts of the country rather than trade. I’ll post a link when I can.
Jon Murphy
Apr 13 2025 at 10:20pm
This is the paper.
Scott Sumner
Apr 14 2025 at 2:56pm
Thanks. Even within the Rustbelt, public policy differences explain why places like Indiana have done better than Michigan.
Arqiduka
Apr 14 2025 at 6:33am
Fair.
Now plot steel consumption please.
Mike Hammock
Apr 14 2025 at 8:05am
https://www.ceicdata.com/en/united-states/apparent-consumption/apparent-consumption-steel
Set it to “max”.
Travis Allison
Apr 14 2025 at 4:03pm
Interesting link. So in 1990, steel consumption was about 80M (tons?). In 2017, consumption was 100M. So if all of that production difference were done in the US, you might imagine 25% more steel jobs today than actually exist. Interestingly, that is the number that Krugman comes up with via a different way. Perhaps, just a funny coincidence.
https://paulkrugman.substack.com/p/a-note-on-trade-deficits-and-manufacturing
Scott Sumner
Apr 14 2025 at 2:58pm
Steel tariffs and quotas also hurt industries that use steel, such as autos, construction and energy.
Mark Barbieri
Apr 14 2025 at 8:56am
Here is a link to the article Scott referenced that doesn’t require you to go through X. Instead, it goes directly to the author’s website:
https://www.molsonhart.com/blog/america-underestimates-the-difficulty-of-bringing-manufacturing-back
Alan Goldhammer
Apr 14 2025 at 9:35am
Can’t the same be said for textile and furniture manufacturing? Both have largely disappeared. There are niche furniture manufacturers, but they are regional and produce limited stock. The textile mills of “Norma Rae” are no more and even finished garments are largely made ex-US. Auto parts used to all be made in the US but that is also now an international business. Specialty metal products such as the tubing used in the petroleum exploration industry now comes from Canada and is subject to tariffs.
I could go on with lots of other examples. It is not just trade competition, but also US CEOs looking to offshore production because of lower costs. As Walt Kelley in the cartoon strip ‘Pogo’ once said, “we have met the enemy and the enemy is us” So true today as it was in 1971 when the strip was published.
Mark Brophy
Apr 16 2025 at 6:37pm
Furniture exports haven’t disappeared. The United States buys $2.26b and sells $294m of furniture to Mexico.
David S
Apr 14 2025 at 9:39am
Nostalgia can be one of the most dangerous psychological conditions in the modern world. The phrase: “Well, back in my day….” can indicate deep misunderstandings of the changes that have taken place since “your day” and those memories are often suspect. When nostalgists are directing policy things can stay worse and get worse for generations.
Edward
Apr 22 2025 at 4:33pm
Scott, how come in some instances, technology and automation destroys jobs, while in others, it creates them? What’s the process at work?
Kevin Corcoran
Apr 25 2025 at 1:14pm
There’s many processes by which this can work, but I’ll just throw in a quick, big picture pass at this. Around the time of the American Revolution, something like 90% of people worked in agriculture. Right now, that number is just over 1%. (This, too, is relevant to people who claim that American manufacturing has been destroyed and “we don’t make things anymore” and try to justify this claim by looking at manufacturing employment rather than manufacturing output. As a percentage of population, agricultural employment has collapsed, but agricultural output has skyrocketed. The story with manufacturing is similar.)
These job losses in agriculture were driven by technological improvements, including automation. You can replace a lot of workers with a Combine harvester. But the vast majority of jobs that exist today couldn’t exist if we still needed 90% of the population to do agricultural work. By reducing the amount of labor needed to maintain (or indeed, massively increase) agricultural output, the door is opened for the division of labor to get more specialized, increasing the productivity of each worker, and for this increased level of specialization to also create new jobs and entirely new job fields. When keeping the population fed requires 9 out of 10 workers to work in agriculture, there’s not much room for new kinds of jobs to be created. When you only need 1% of the population to feed the remaining 99%, vastly more kinds of jobs and job specialties can be created, because so much labor has been freed up to focus on new things.
This is, of course, a particularly striking example, but the fundamental mechanism is basically the same even at a smaller scale.
Thomas L Hutcheson
Apr 25 2025 at 11:21pm
And as the effect of trade the most important trade policy decision was not tariff reductions in NAFTA or the WTO but the decision of GWB to create deficits by cutting taxes, drawing in capital, encouraging imports and taxing exports.
Warren Platts
Apr 26 2025 at 11:52am
The problem with that chart above is that the series only goes back to 1987 when the main damage occurred during the late 70’s and early 80’s. That’s when Youngstown collapsed. Steel production fell by 50% over that period. 1979 was also the peak of overall manufacturing employment. The fact that the steel industry is getting more automated (i.e., it takes fewer workers to produce the same amount of steel) can only be applauded. But that’s not what happened at Youngstown. Youngstown collapsed because the steel industry as a whole collapsed when production fell by half.
Comments are closed.