Premature deindustrialization proceeds apace

The South China Morning Post reports that iPhone maker Foxconn has replaced more than half of its workforce with robots since the launch of the iPhone 6. The figures were provided by the local government in Kunshan, where the company is based.

“The Foxconn factory has reduced its employee strength from 110,000 to 50,000, thanks to the introduction of robots. It has tasted success in reduction of labour costs,” said the department’s head Xu Yulian … 

Since September 2014, more than 500 companies in the Chinese province of Dongguan have spent a total of $630M on robot and AI technology to replace human workers. It was reported back in January that Foxconn had received a $12M government subsidy to help minimize layoffs in response to reduced iPhone orders.

Here is more, via Kurt Busboom.  And, via “anon from cl”, there is this from Adidas:

Adidas, the German maker of sportswear and equipment, has announced it will start marketing its first series of shoes manufactured by robots in Germany from 2017.

More than 20 years after Adidas ceased production activities in Germany and moved them to Asia, chief executive Herbert Hainer unveiled to the press the group’s new prototype “Speedfactory” in Ansbach, southern Germany.

The 4,600-square-metre plant is still being built but Adidas opened it to the press, pledging to automate shoe production – which is currently done mostly by hand in Asia – and enable the shoes to be made more quickly and closer to its sales outlets.

Here is my recent paper on premature deindustrialization.


Is "premature" deindustrialization (in this case caused by automation) an automatic outgrowth of globalization?

I suspect so....

This is not deindustrialization.

It is industrialization.

I'm baffled by the stupidity of this title. Using less labor is DE-industrialization?

So the luddites were actually promoting industrialization when they were smashing power looms?

It all depends on the rule you use to measure. It seems here industrialization means "% of manufacturing jobs in respect to total employment".


Do robots make suggestions for improvements to products and processes? Do they make complaints about processes that don't work very well for either product quality or efficiency? Do robots care about product quality or efficiency? It's common for economists to wear blinders, so they don't see what they don't wish to see. Robots don't need blinders, for they see only what they are programmed to see. Just like, you know, some economists.

Do robots form unions? Do robots complain about overtime? Do robots show up for work drunk? Do robots need bathroom breaks?

Right. I'd better get my pee and poop bags right away and undercut my salary by half before the robots replace me.

I think I had a free thought the other day. But I realized how painful that can be. I'm not that stupid. I refuse to think a free thought ever again.

Bigger better faster smarter. The future is coming. Embrace your slavery.

There is another way.

The way of enslaving the factory owners and forcing them to employ people and not buy robots?

That probably wouldn't be good for business. Or the economy. But I highly value the notion of upholding the principles of treating people as fundamentally human, worthy of good treatment (exceptions for openly and clearly justified means of discipline excepted, however, upon broad social approval).

You left out the most important value of workers, the value of workers eliminated by Reaganomics that eliminated the "on the other hand". Why no demand side in perfect balance with supply side economics?

The most important value of workers is buying what is produced.

But if you devalue workers by paying them less, you devalue the supply because the price of the supply can't exceed the income of workers which is the labor costs.

(Labor costs can be shifted in time with debt and savings.)

Honestly question did Ronald Reagan bang your prom date or something?

I have long believed Mulp was a PATCO member, is still on strike.

...neither do Chinese workers...

Not even close to true. Just the other day my assistant took a bathroom break. In the middle of work! I thought about calling the company to get her fired, but then I realized ... that's just a pretty human thing to do.

Also, she doesn't comply 100% with orders all the time, and some of the ways she "misunderstands" instructions are fearfully indicative of a capacity to free thought. But that's part of being human too.

Actually, I much prefer it that way. Considering that the alternative is the assumption that I too should be ruled over like a robot.

We're closing in on the time when the only mammalian entities at a typical factory will be a man and a dog.

The man's job is, obviously, to feed the dog. I don't get what the dog's job is.

"The factory of the future will have only two employees, a man and a dog. The man will be there to feed the dog. The dog will be there to keep the man from touching the equipment.
Read more at:"

I think it is under-discussed that there were 110,000 employees at one factory.

Think about that. How could that be anything other then a system much in need of optimization?

I'm going to call this late optimization rather than early deindustrialization.

Absolutely. what's going on here. But that's exactly why it is done in China: they cost so little and the cost of waste is so low, that it becomes acceptable.

But it is still extremely inefficiency, and extremely unproductive.

Which parts of their production process would you identify as being inefficient or unproductive?

PS, "you're so stupid" isn't a very good answer to that question.

background reading

I cannot imagine that 110,000 workers do not create continuous disruptions on each other. I wonder how many thousand are just trolley pushers?

Where do the robots come from? Are they leased or purchased? Are the robots being manufactured in plants by humans or are they built by other robots? What happens to a robot when a change in product or process occurs? Is there a market for used robots? Being mechanical devices robots must incur wear and start to operate out of tolerance, manufacturing stuff that doesn't pass quality control. How many robots could one human monitor or is no supervision necessary?

Capital costs are just labor costs by anot her name. The intrinsic value of capital is at most the labor cost depreciated over its useful life. Innovation destroys the intrinsic value of capital.

Capital gains are impossible without monopoly creating the opportunity for rent seeking that at low interest rates inflates the price of capital way beyond its intrinsic value.

Monopoly control over oil built over decades of crony capitalism that concentrated all the capital in the hands of big oil, and made government the sole source of new natural capital to pillage made the price of ExxonMobil market cap exceed it's capital intrinsic value by tens of billions of dollars.

Obama did not preserve the crony capitalist control over new capital to pillage, so upstarts drill baby drilled in Texas and Bakken, with investors feeling secure that Obama was not going to screw them by giving ExxonMobil the threat of flooding the oil market with free government oil by subsidizing drilling in Federal waters.

Remember, it's certainty that promotes investment, and the Bush-Cheney policies created uncertainty over the ability of big oil to increase supply from free government oil, but everyone was certain that Obama was going to stop big oil from producing more oil.

Thus, ExxonMobil et al had monopolies over oil supply that enabled them to charge rents of $50 a barrel of oil from their capital, inflating the price of capital that was falling in intrinsic value - after all, the capital was being sold to be burned.

But now investment in labor cost has created so much capital that, as Keynes put it, led to "the euthanasia of the rentier, and, consequently, the euthanasia of the cumulative oppressive power of the capitalist to exploit the scarcity-value of capital."

Likewise, Apple's inflated asset prices have fallen "destroying wealth" thanks to its rents being euthanized in China.

"Capital gains are impossible without monopoly creating the opportunity for rent seeking that at low interest rates inflates the price of capital way beyond its intrinsic value."

That doesn't make the slightest bit of sense. Total BS, I would add.

Capital...gains...are always "monopoly rents". So are labor gains. That is, the gains only happen when you take advantage of a unique market position. If there was no such thing, then there'd be no gains. You'd be in a "perfectly competitive" market with zero "gains:, i.e. profit, for the firm.

So yes, but it's not "monopoly" in the way you're thinking it. It has even less to do with rent seeking or interest rates or whatever other nonsense you wrote about there. It's in essence, markets that are not perfectly competitive, which describes every market on earth. Monopoly is the absolute extreme case of it, but you're talking about monopolistic rents.

I.e. to put in simple words...all profits come from doing or having something "different" and "better" then someone else. That's all you're saying here.

You two are taking opposite extremes of the argument. All that is necessary to earn a profit is that revenues exceed costs and that competition doesn't disrupt this.

It's not robots. It's chips. Whereas in the past you had to link several chips on a circuit board to create a functionality, now it is all on a chip, and just a few chips make a system.

And, the functionality will soon move to the cloud, with ever faster communications making it irrelevant where the computation occurs.

A few months ago we were in Japan. I read an article about the decline, or change, in the Japanese electronics industry. Whereas in the past electronics manufacturers created customized circuits, had customized software, and customized chips, today everyone uses commoditized chips (NVIDIA) to do a single functionality. Assembly of multiple unique components, and savings from not having to use persons to assemble because the function is on a chip, takes away the value added of an assembler and customized software/hardware manufacturer.

it's not the robots.

The chips are eating the robots.

'The chips are eating the robots.'

Not really - unless you think chips put shoes together, and not robots.

There are currently 20,000 unemployed robots for sale on Ebay right now, should we be concerned?

"The Foxconn factory has reduced its employee strength from 110,000 to 50,000, thanks to the introduction of robots."

Thus Apple has a significant reduction in Chinese consumers able to afford the high rent iPhone.

To better match the price to labor income, Chinese consumers are buying the brands priced at 99-101% of total labor cost (labor costs in the robots included) while the iPhone price has increased from 200+% to 300% of labor costs.

I'm sure Apple is complaining about the totally inadequate jack boot Chinese big government preventing cell phones from being sold in competition with Apple products.

"Thus Apple has a significant reduction in Chinese consumers able to afford the high rent iPhone."

Yeah, because those $5/day Chinese workers were buying $600 iPhones :)

Well if you double that and add more again, you might be approaching something. As the empiricist you are, why not actually check numbers before posting them? It is the exception to the rule that on some rare occasions you're not wrong to the tune of 100% or more.

Quadruple it and they're still not spending 10% of their gross income on a phone

It is strange that the result of that Reagan revolution was that everyone became much richer. I am willing to bet that next year, China will have more people able to buy iPhones, not less. Next year China will be richer.

Replacing workers by machines is a good thing. It makes everyone richer.

"It is strange that the result of that Reagan revolution was that everyone became much richer."
Ha ha ha. Oh, wait, was that serious?

Who writes these articles? Are the journalists writing them so clueless, that they need to use words like "robots" and "AI" describe what is in essence...automation?

Automation is hardly a new thing. Pretty much the entire value chain of the production of the iPhone has been automated and has always been automated. Producing the actual components etc.

The only part was the actual done by the assembly of the parts into the finished product. That is the lowest value-added activity in the entire value chain of the iPhone. Which is why it was also left to be done by low-skilled low-payed Chinese workers. Now, it's catching up with them too.

Love the stupid use of the term "AI" in the article. There's nothing "AI" about it.

I sometimes think "AI" is promoted as a word choice to get people paranoid and force them to work harder harder harder. It may serve as a whip of some sorts.

"Can you stay until 8pm tonight? Oh, by the way, I've been looking at the productivity numbers and we might have to replace some people with robots. Shucks. I forgot! You're going to have to come in on Sunday too. Unpaid."

I agree with you on this one. It's "automation" not "robots", and in the case of "AI", it's "machine learning algorithms". imo

I don't have time right now to read the paper, but "deindustrialization" is a weirdly poorly chosen word to describe automation.

It seems that the value in offshoring really isn't all that high when the latter country is using robots at such a margin.

None of the articles I've found on this factory even attempt to answer the most obvious question:

What is the ratio of shoes to employees in the new factory vs the comparable ratio in an existing Addidas factory?

Quoting the number of employees is meaningless without comparable productivity numbers.

I don't think it's correct to call this stuff "AI". It's machine learning algorithms. "AI", I think gives it too much credit, and in a sense sounds scary to people.

Tyler, please think about the title of this post for a minute. It is moronic.

Automation is what defines industrialization. Are you so attached to your thesis that you're willing call a very clear example of industrialization (greater use of capital goods), deindustrialization?

Not to defend Prof. Cowen, but he may believe that de-industrialization, particularly as seen in the U.S. over the last generation, means the loss of employment in the industrial sector.

What de-industrialization means in real terms is another discussion, but it seems as if he is not understanding the causal relationship, and thus describes increasing industrialization (note that nobody is buying robots to make iPhones or Adidas shoes in the U.S.) in terms of employment reduction. Which at least in the case of Adidas likely means that some highly paid German manufacturing workers who create the robots, and those that maintain and operate them, are the ones benefiting at the cost of a large number of lesser skilled and significantly lower paid workers - which is definitely the opposite of 'de-industrilization.'

This whole post is actually bizarre, considering how this web site trumpets the advantages to all, even those losing their jobs, when capital displaces labor using machinery/automation/robots.

In his cited paper on "premature deindustrialization" TC alludes to the economic virtues wrought globally by cell phones.

What paradigmatic status will cell phones enjoy globally if and/or when they are unambiguously implicated in the formation of troublesome and/or deadly tumors of various kinds, as was broadly suggested in a fresh study appearing late last week? ("E. T., call your oncologist, but find a working landline first.")

Whatever economic virtues wireless telephony has unleashed over the past quarter century, if tumor formations in human crania commence in earnest over the next quarter century, exactly what kind of economic utility was derived?

Citation or it didn't happen.

I doubt they cause tumors, but if they do we might have to launch massive R&D efforts into new innovations like phones with speakers or headphones with microphones. Keep your fingers crossed.

This, you mean?

The story came out late last week and still makes the rounds. While the data do not seem conclusive at this date, the asserted linkage between cellphone radiation and tumor formation continues to recur with some regularity and I suspect will be established or discredited unambiguously in the next decade or so: if the link IS ever established, though, one, two, and three decades of exposure will already have taken their toll, at which point the field of oncology can be underwritten by our tech giants and telecomm behemoths.

There seems to be some conflict in the comments regarding what de/industrialization means, to level the knowledge field I'd suggest reading "Premature Deindustrialization" by Dani Rodrik (2015) published in the Journal for Economic Growth: DOI 10.1007/s10887-015-9122-3

Comments for this post are closed