The summary of what has happened to manufacturing jobs in the United States over the past decade is that between 2008 and 2010 two million manufacturing jobs just vanished. Since 2010 we’ve recovered one million of those jobs. When you factor in the increase in population over the last six years, that isn’t much.
The explanations for that provided at Cerasis, from which the graph above was taken, are: technology, globalization, environmental concerns, and reclassification, which I find interesting:
With the rise of the digital age, more companies have entered an era of product-less or factory-less manufacturing. Ultimately, this means companies produce items that do not require physical space. This includes nearly every digital application available on your smart phone, tablet, or laptop. To further this point, some manufacturers may not have a factory-space, which removes them from being counted within the manufacturing industry. Apple has experienced this phenomenon first-hand following large sales of new technologies in the US, which were manufactured in China. Ultimately, the majority of the work on Apple technologies has been completed here in the US. These technical jobs, such as engineers, designers, computer-code analysts, and robotics experts, may not fall into the category of manufacturing, but their efforts are required for outside manufacturing to take place. This results in a confounding arena where US manufacturing jobs are hidden under other categories of the US Jobs Report.
Take the May 8th Report at slightly less than actual value. Consider how the overall job growth continued in a positive slope over the course of the past three years, speaking aside from slight growth and drop cycles experienced between 2009 and 2012—immediately following the housing market crash. The report shows that the manufacturing industry, while not making huge improvements from last month, continues to be adding more jobs than losing jobs over the course of the past month. Additionally, some manufacturing industries, such as oil and gas production, appear to have added significantly more jobs, but these jobs are the result of worker strikes coming to an end. With these factors playing into the overall analysis of the manufacturing industry, US manufacturing jobs appear obsolescent. However, you need to know where to look before making a decision of whether manufacturing is growing or slowing in the modern world. And, as we said in this post about why it is OK that US manufacturing jobs won’t ever get back to peak levels of past decades, in the end, it’s about output and the job multiplier effect manufacturing tends to have.
I think the conclusion is probably correct but I think there’s more to it. At the time I thought that the companies’ behavior couldn’t be explained by lost sales, technology, or globalization and I still believe that. I wrote incessantly about it at the time.
It might be that the companies just panicked but that doesn’t explain what happened in the aftermath of the Great Recession. I think that some companies exploited the panic during the recession to accomplish other goals. I’m open to explanations of what those goals were.
He has four sections. One is on robots. One is on how environmental concerns tend to favor robots. One is on reclassification of jobs that involve ‘manufacturing’ by using robots to make things (apps) for robots. The fourth point is that raw materials are cheaper overseas than here.
I don’t want to be too anthropist or robophobic, but it looks to me as if automation is a rather big part of what’s happening. So, after many years of telling me that we were not losing jobs to robots, are you now accepting the idea that automation is a major factor, quite probably the major factor, in job loss going forward?
I don’t want to mischaracterize your stand which tends toward more nuance than I’m capable of, but blaming low business investment when that investment would likely have gone to more robots rather than more humans, seems to miss the mark. In fact wouldn’t more investment likely result in low, no, or even negative growth in human employment? Fast food is getting by on low-wage immigrants now – if McD’s invested a few billion they could get rid of most of those humans. And clearly autonomous vehicles are seeing heavy investments and those are all job-killing robots.
I think it is time to make the leap and recognize that we are now in direct competition not just with sweatshop workers in China, but with robots, in a race to the bottom in terms of wages. The robots are coming not just for our jobs but for China’s as well, and they will win. We need to adapt to that reality.
For investment in robots to have reduced the need for human workers there would need to have been more investment in automation than there has been. Besides, what technological advance took place between 2009 and 2010 that would explain your preferred explanation? There wasn’t one. Neither technology nor globalization nor both together are sufficient to explain what’s actually happened.
I think you’re getting ahead of yourself. You’re looking at press releases about robots and phlegmatic growth in the number of jobs and leaping to the conclusion that the jobs are being taken by robots. That might happen some time in the future but it hasn’t happened yet.
In addition to technology, globalization, environmental concerns, and reclassification I’ll offer some other explanations:
1. American kids with college educations don’t want to work in factories.
2. The U. S. secondary educational system is so bad that kids without college educations don’t have the necessary knowledge, skills, and habits to work in today’s factories.
3. Too high a percentage of today’s kids don’t even have high school educations, particularly in big cities. Not to mention criminal records, nonexistent work habits, etc. And then there’s bigotry.
4. Managers are transitioning their companies away from unprotected sectors of the economy into protected and/or otherwise subsidized sectors.
That’s just a few. Multi-factorial.
BTW, the U. S. didn’t get out of the rare earths production business (a recent development) because we’d run out or because demand was too low. It was regulated out of existence. We need to get away from neoliberal models for solutions. Rather than punishing companies (like rare earths producers) for polluting, why not subsidize them to enable them to clean up? The outcome is better and the jobs stay here. I could repeat that same story for a dozen different industries.
That’s not quite it. I’m looking at capabilities. On the one hand an employee (human) that is expensive to raise, hard to educate, hard to discipline or control, vs. one which is hard to create but once that threshold is crossed, is under complete control and never whines or weakens or takes a day off. Robots are clearly superior at most assembly line jobs. They will soon be superior at most driving jobs, fast food jobs, many assessment gigs like reading X-rays and MRI’s. They’re even better at killing our enemies.
It is simply not possible, short of a Luddite uprising, that robots will not continue to surpass human capabilities. It’s what they are designed to do, they’ve been doing it, and there’s nothing to stop them doing it more.
As for effects already in place, I point again to your own assessment of infrastructure jobs. In the 30’s building a road meant thousands of jobs. Now it means three guys, one driving a grader, and two directing traffic. Robots (big yellow ones) have foreclosed the possibility that we can make millions of jobs with a big government infrastructure stimulus.
We have the capability of destroying the world. IMO acting as though it were about to happen would be premature to say the least.
Robots do not presently have the ability to replace 90% or 50% or even 25% of jobs. I doubt they will for the foreseeable future. More than 20 years ago Japan stepped back from automating its factories as much as it theoretically might have at the time because it was determined that it didn’t make any sense.
There’s a lot of chat these days about robots displacing designers. Maybe it will happen, maybe it won’t. It hasn’t happened yet. The history of such things has been that automation hasn’t replaced designers, it’s just made them more productive.
Retail salespersons, 4.48 million workers earning $25,370
(Already being replaced by Amazon which itself is hard at work on replacing its own workers with robots. Malls are dying. I can’t recall the last time I was in a bricks and mortar store aside from groceries, and I’m using Instacart for that more and more frequently.)
Cashiers 3.34 million workers earning $20,420
(Amazon’s new bricks and mortar stores will have automated check-out. Pick up your stuff, walk out the door, no cashiers. Apple stores have played with it as well. Self-check is already common in grocery stores.)
Food prep and serving staff, 3.02 million workers earning $18,880
(The only thing slowing automation here is the low pay of their human workers. The tech to create human-free fast food restaurants is already well-advanced.)
Customer service representatives, 2.39 million workers earning $33,370
(Hello, Siri? As much as I despise the automated systems they are getting better. I send a fair number of FedEx packages and I do it without human interaction in ordering.)
Laborers, and freight and material movers, 2.28 million workers earning $26,690
(Driverless trucks are coming. Billions are being invested in the project of replacing humans. Robots drive 24 hours a day, they don’t steal, and they won’t speed or get sick or pretend to have a mechanical problem so they can get a quickie with a truck stop hooker.)
These are not people who are going to be retrained as app developers or brain surgeons.
More business investment will drive more development of robots which will reduce still further the territory reserved to homo sapiens. Trade policies that cut us off from cheap foreign labor will also drive automation. Snobbery and safety fears will at first retard, then rapidly accelerate the development of robot vehicles.
It is in the nature of tools to do a thing better than we do. A claw hammer is better than a human fist at driving nails. Now we have a claw hammer that pounds its own nails. It defies logic to imagine that we would create more hammer-designing jobs than will be displaced by the robot hammer. And it is nothing but a statement of faith to assume we will discover new jobs only open to humans, especially manufacturing jobs. As many have pointed out, Apple earns billions every five minutes and somehow does it with about two dozen nerds in Cupertino and a few thousand Chinese.
We are in competition with our machines for jobs. They are the new illegals. You know the signs human workers sometimes hang, the one that says, “Better, Faster, Cheaper – pick two.” The robot doesn’t do that because the robot will do it better, faster and cheaper. And they don’t feel the need to be insolent to customers.
Here are some metrics that could persuade me to your view:
1. Unemployment goes up.
2. The labor force participation rate continues to decline.
3. Real wages go down.
All at the same time. Other factors that would be indicative:
– businesses stop asking for more H1-Bs
– Full or part-time college enrollments decline
– Flow of illegal immigrants across the southern border declines (increased federal government efforts could obscure this)
The problem with Malthusian or inverse-Malthusian arguments is that they never include a feedback loop(s).
People without jobs cannot buy products. Unpurchased products are not profitable no matter how cheaply they are made. Companies that cannot profit from producing products manufactured by robots will go out of business.
The US has been able to ship jobs to China and elsewhere for decades because there was a large surplus of wealth that was able to be leveraged into credit, and part of that credit was extended to those former job holders as debt. That debt allowed them to make up some of the difference between their old wages and their new lower ones.
It is not sustainable. When there are few jobs, people will breed less, and there will be fewer people out-of-work.
When Gutenberg began printing Bibles, a miniscule percentage of the population had the intellectual capability to write even a bad book. Today, the percentage of the population that is able to write even bad books, novels, newspaper/magazine articles, blog posts, comments, etc. is huge. The instruments may have changed, but there has not been any real great leaps in writing itself.
The Malthusian argument is not credible because the situation is not sustainable.
According to the International Federation of Robotics, a total of about 135,000 industrial robots were installed in the US between 2010 and 2015, most of those in the auto industry.
More data at the link. Usual caveats apply as the IFR is a trade group that promotes robotics.
Michael might be right about the future, but it appears that robotics is still relatively small and likely cannot explain a downturn in manufacturing employment.
Good pickup, Andy. I was looking at business investment. Robots cost money.
Here’s another sign: productivity. If automation were a significant factor, the productivity of human workers would be increasing. As it is we’re in a labor productivity dry patch. It isn’t increasing as would be expected.