7 ms·

Is Automation Really to Blame for Lost Manufacturing Jobs?

41 points by jeffreyrogers 8y ago

pixl97 8y ago
Paywalled partial article.
madengr 8y ago
On NPR this morning, they said that USA manufacturing jobs have decreased this 1/4, despite total number of jobs increasing.
rossdavidh 8y ago
Well obviously automation is to blame, that's why the economy shifted to buying real goods made in China and Mexico, two nations that at the dawn of the 21st century were at the forefront of robotics. Also places like Bangladesh and Vietnam. It also explains why imports from places like Japan and South Korea, clearly laggards in robotics, declined over the last couple decades. Oh, wait...
- LarryDarrell 8y ago
  But we have cheap VCRs now.
- draw_down 8y ago
  Come on
- quxbar 8y ago
  The country where a technology was invented matters very little to the people making money from it. However, I think 'automation' is really just a bad proxy, technology is -augmenting- workers and re-architecting supply chains. Consider how transformative the standardized shipping container is for an import/export business - regardless of what countries they operate in.
- s_kilk 8y ago
  Turns out the Labour Theory of Value was correct.
- John_KZ 8y ago
  Although China is on the other side of the equation, it provides plenty of clear examples showing automation not eliminating jobs. It changes the available jobs, and generally improves them (ie digger operator instead of hand-shovel "operator") but jobs are still available.
  - Brockenstein 8y ago
    Well I wouldn't call an excavator automation exactly. But at the same time, you think ten guys digging a ditch at least some of them aren't made redundant when you get an excavator? You don't need ten excavators, and maybe you don't need 1 excavator driver and 9 ditch diggers... I mean I think you're engaging in selective reasoning.
    - sigstoat 8y ago
      > Well I wouldn't call an excavator automation exactly. go dig a trench by hand and get back to us.
- Brockenstein 8y ago
  Well paying someone a $1 a day is a lot cheaper than expensive robot and the infrastructure to maintain that robot. If cheap labor is available you use that. If you can't expert the work to a developing nation automation might be the next best option. There's no single boogey man to blame.
geggam 8y ago
Go tour a factory. Where there used to be 300+ people there are now 20. Guess what automated robotic things you now see that weren't there before ?
- madengr 8y ago
  I'm in low volume electronics manufacturing. Sometimes it's not even cost effective to pick-n-place circuit boards, rather have someone stuff a PCB by hand. The USA has probably moved to a lot of low-volume manufacturing. Could be that people are not buying as much stuff. I have read that credit card debt is back to 2008 levels. Maybe everyone is maxed out; not good.
  - geggam 8y ago
    Where I live people are on facebook selling home cooked food and offering to do handy man work. Jobs I grew up with in fast food that could be relied on to provide 38 or so hrs a week now provide my teenagers 16-20. The teenagers are competing with adults for the jobs. I am of the opinion outside metropolis areas the entire nation is running on Empty as far as finance goes. The people I know who live here and work normal blue collar jobs run just under the max on their credit cards and keep alive by paying minimum. Opening new lines of credit to pay for emergencies Granted this is only my personal experiences but its common enough its disturbing. When people run out of money to buy things what happens ?
    - benploni 8y ago
      Something about your post made be literally anxious, like to the point of affecting my breathing.
      - Brockenstein 8y ago
        Worried about the next big recession and whether or not you'll end up in a shanty town.
    - erikpukinskis 8y ago
salawat 8y ago
No. Not as the root cause. I mean it's definitely a contributing factor, but only because of the axioms that drive our definition of a good company. The ideal company is seen as an investment vehicle that creates shareholder value. Full stop. We don't measure well performance of a company in terms of hiring people. Nor do we even measure it in how well they do what they are there to do. It always getsmeasured in value returned to investors. Investors are generally not laborers in this day and age. And any company that can make more long term by automating and dumping the need for laborers will. You can blame Automation all you want, but it's something far more basic that makes automation a worthy goal in the first place. And no, neither automation nor the incentives that make it desirable are "easy" problems to fix.
- Finnucane 8y ago
  >The ideal company is seen as an investment vehicle that creates shareholder value. Full stop. Even if that were true, that's not a factor that just suddenly started to have force in the last couple of decades.
  - moosey 8y ago
    Massive tax cuts for the wealthy, particularly for capital gains and dividends, have reduced the resistance for money escaping corporations into the hands of investors, rather than the value add of paying employees more or improving working conditions and benefits. These changes started to occur in the 1980's, rather than the Dodge vs. Ford decision from 1919. While it is certainly true that nothing legally changed (edit: this isn't true: there is a full on attack on collective bargaining ongoing) in the more recent 30-40 years to allow for the massive accumulation of wealth and corporate practices that make the 'FULL STOP' in the original comment more accurate than before, the modification of tax law certainly has had a major impact.
  - Brockenstein 8y ago
    It doesn't have to be new, but we have new and more ways to increase shareholder value. Like vastly improved methods of automation that have appeared in the past few decades.
0xdeadbeefbabe 8y ago
Nope, speculation caused it.
phkahler 8y ago
>> If, for example, this year’s model sells for the same dollar price as last year’s but has more features that consumers want, its true price has, in effect, dropped. So government agencies produce statistics showing that output, adjusted for price, has risen. This is a terrible way to view it. If we pack twice as much on a chip vs last year, did output go up? But price went down by about the same amount. Should we measure output in transistors? Printed die area? Dollar price of chips? Number of man-hours required to produce things? Labor cost? If Intel spent a bunch of cash on a new fab, made circuits with twice as many transistors, reduced prices on the old model and sold the new ones at the old price, then sold the same number as last year, did "output" really change? If yo u think it did, then I ask if it changed in any way that is meaningful to economic measures - it didn't change GDP, it didn't change cost or amount of labor, it didn't even change the companies profits.
- pixl97 8y ago
  Maybe take a larger view of it. Instead of just looking at Intel, look at the companies buying intel chips, are they getting more output with the same dollar input?
  - phkahler 8y ago
    >> look at the companies buying intel chips, are they getting more output with the same dollar input? Probably not. The ones who were constrained by computing power are now doing higher quality work - FEA comes to mind as something that seems to produce more accurate results with increased computing capability. Games get "better" but their market doesn't get bigger - just more demanding. Hollywood does more CGI, but that doesn't actually increase their market. These are all things that don't affect GDP.
bsder 8y ago
Automation certainly isn't solely to blame. Our estimate of productivity is wrong. https://qz.com/1269172/the-epic-mistake-about-manufacturing-thats-cost-americans-millions-of-jobs/ https://qz.com/1269172/the-epic-mistake-about-manufacturing-...
ThomPete 8y ago
It's not just automation it's also digitalization which removes a lot of middle layers which used to be part of the cost. A much better to think about this is from the perspective of technology. The music industry is the best example of what happens. Basically, things that used to require all sorts of manufacturing and distribution layers are now more or less compressed into a single digital device.
milesvp 8y ago
It occurs to me as I think about this topic more and more (sadly article is paywalled so I can't speak to it directly), manufacturing is to blane for lost manufacturing jobs. Economies are self balancing, and since manufacturing makes goods cheaper, eventually people will do other things that are still expensive. This is prediminantly why people talk about transformation to service economies. Manufacturing is inherently about mass producing, which means it's always been about scale, which means doing more without adding proportionately to labor. Automation is just one of the many things that help scale, it's just one of the many things that requires capital investment. The bigger problem for manufacturing jobs has been globalization. When you can take your capital to a market with cheaper workers, then labor becomes a scaling tool (ironically?) in that you can throw labor at a problem to get your labor costs down. What's interesting about this, though, is that at some point you can't do this anymore. Textiles, which is considered level 0 for manufacturing, has nowhere left to go for cheaper labor than Bangladesh (without significant political instability issues).
clarkmoody 8y ago
Is automation to blame for lost agricultural jobs? Lost textile mill jobs? Yes, it is. Do we miss those jobs? No, we don't, because we're so much wealthier when we can replace human labor with machine labor. That frees up humans to do higher-order tasks, improving our standard of living. Yes, there are winners and losers with any change, but long-term we are far better off than we could have imagined at the dawn of the Industrial Revolution.
- sp527 8y ago
  > Yes, there are winners and losers with any change, but long-term we are far better off than we could have imagined at the dawn of the Industrial Revolution. This is a deeply ignorant statement that renders invisible over 2 billion completely disenfranchised people in the third world. Just because you don't experience the consequences - while you gorge on the bounty of human progress somewhere amongst the top decile of our species - doesn't mean they aren't there.
  - clarkmoody 8y ago
    I recommend that you go check out some of the data here: https://humanprogress.org/dws https://humanprogress.org/dws In the past 150 years, the human race has made extraordinary progress, lifting billions out of poverty. There are still many globally who are left out, and I am aware of that. But you can't argue against the widespread gains to humanity through improved technology.
- Aloha 8y ago
  I'd argue quite strongly that the communities who lost those jobs, do miss them dearly. They have not been replaced, and while you're correct we've lifted billions out of poverty, we've thrown hundreds of millions back into it, with no easy recourse out. Western nations, in general do a piss poor job with people in poverty, as we think ourselves beyond the problems of the third world, at least in the united states, its clearly not true - and dont have the right systems in place to allow folks to either fully participate in the low end of the market, or give them meaningful government assistance. Business right now is frightfully capital intensive - and has a high barrier to entry (paperwork mostly), and in many places there are no jobs available - what is someone to do?
bongo662 8y ago
It'd be interesting to see a breakdown of what 'forms' of manufacturing are leaving the nation. Its pretty obvious that as trade agreements rise up and its more cost effective to move manufacturing to other countries with lower payed workers that the company would do that but my guess would be that higher(?) quality manufacturing still goes on stateside rather than in China or Mexico for fear of IP theft. So for things like high grade medical devices - MRI, CT scanners would still be manufactured here because of the knowledge and technology necessary to construct it. I also think theirs a good Stephen Hawking quote I like to think of when people discuss how in a post scarcity world with full automation life will be a utopia and its something along the lines of "If machines produce everything we need, the outcome will depend on how things are distributed,"
Animats 8y ago
Actual paper from that author.[1] From 2016. Her real point is that the electronics industry is doing great, while some other parts of of US manufacturing are not. Fig. 2 is useful. Textiles, apparel, furniture, and paper products are the big losers. The apparel industry was totally crushed by imports from low-wage areas. That took out the US textile industry, since exporting textiles to be made into apparel and re-imported was pointless. US motor vehicle output is up, not down. That hasn't been crushed by imports. Primary metals (which is mostly steel and aluminum) are up. More breakdown of the computers and electronics sector would help. We'd probably see semiconductors through the roof and consumer electronics down. That data is available. This is useful. Trade is an industry-specific problem. The American apparel industry is mostly gone, but we have cheap clothing. The metals industry complains but isn't in real trouble. Motor vehicles are doing OK. A breakdown of consumer goods by category would help. That would show where to apply tariffs. Paper products are down because newspapers are obsolete, and they used vast amounts of paper, and because offices use far less paper. Not import-related. [1] https://upjohn.org/sites/default/files/pdf/state-of-american-mfg.pdf https://upjohn.org/sites/default/files/pdf/state-of-american...
- philipkglass 8y ago
  Another thing to consider, when pondering counterfactuals to free trade: offshoring destinations are also automating. China is now the largest market for industrial robots: https://www.cnbc.com/2017/10/26/china-is-the-largest-market-for-robots-in-the-world-right-now-abb-ceo.html https://www.cnbc.com/2017/10/26/china-is-the-largest-market-... Maybe when the local factory shut down in years past it transferred 500 jobs straight to China. That doesn't necessarily mean that it kept employing 500 people for the same output in the years since. Re-shoring the same volume of production may mean significantly fewer jobs today, in e.g. textiles: "U.S. Textile Plants Return, With Floors Largely Empty of People", 2013 https://www.nytimes.com/2013/09/20/business/us-textile-factories-return.html https://www.nytimes.com/2013/09/20/business/us-textile-facto... ... Take Parkdale: The mill here produces 2.5 million pounds of yarn a week with about 140 workers. In 1980, that production level would have required more than 2,000 people. ... Or in the case of fields that still are very little automated, like garment sewing, it would require pretty much the same number of workers now as 20 years ago. As you say, it's very product-specific. If we broaden the category a bit from "manufacturing jobs" to "industrial jobs" (including things like coal miners and workers at coal fired power plants), you also see strong effects from process-and-input changes. It takes far fewer people to produce a terawatt hour of electricity from a modern combined-cycle natural gas generating plant than from a modern (worse, built-in-1970) coal plant. It also takes fewer workers for a TWh from solar PV or wind power. The rapid growth of jobs in renewable energy is a temporary excursion, mostly an artifact of current rapid expansion. At steady-state it will take fewer people to supply electricity from a renewable-heavy electricity system than a coal-heavy system. That's one reason that I am bullish on the long term price-competitiveness of renewable electricity but skeptical that "green jobs" are a panacea for declining industrial employment in other sectors.

41 points by jeffreyrogers 8y ago

Paywalled partial article.

On NPR this morning, they said that USA manufacturing jobs have decreased this 1/4, despite total number of jobs increasing.

rossdavidh 8y ago

Well obviously automation is to blame, that's why the economy shifted to buying real goods made in China and Mexico, two nations that at the dawn of the 21st century were at the forefront of robotics. Also places like Bangladesh and Vietnam. It also explains why imports from places like Japan and South Korea, clearly laggards in robotics, declined over the last couple decades. Oh, wait...

LarryDarrell 8y ago
But we have cheap VCRs now.
draw_down 8y ago
Come on
quxbar 8y ago
The country where a technology was invented matters very little to the people making money from it. However, I think 'automation' is really just a bad proxy, technology is -augmenting- workers and re-architecting supply chains. Consider how transformative the standardized shipping container is for an import/export business - regardless of what countries they operate in.
s_kilk 8y ago
Turns out the Labour Theory of Value was correct.
John_KZ 8y ago
Although China is on the other side of the equation, it provides plenty of clear examples showing automation not eliminating jobs. It changes the available jobs, and generally improves them (ie digger operator instead of hand-shovel "operator") but jobs are still available.
- Brockenstein 8y ago
  Well I wouldn't call an excavator automation exactly. But at the same time, you think ten guys digging a ditch at least some of them aren't made redundant when you get an excavator? You don't need ten excavators, and maybe you don't need 1 excavator driver and 9 ditch diggers... I mean I think you're engaging in selective reasoning.
  - sigstoat 8y ago
    > Well I wouldn't call an excavator automation exactly. go dig a trench by hand and get back to us.
Brockenstein 8y ago
Well paying someone a $1 a day is a lot cheaper than expensive robot and the infrastructure to maintain that robot. If cheap labor is available you use that. If you can't expert the work to a developing nation automation might be the next best option. There's no single boogey man to blame.

geggam 8y ago

Go tour a factory. Where there used to be 300+ people there are now 20. Guess what automated robotic things you now see that weren't there before ?

madengr 8y ago
I'm in low volume electronics manufacturing. Sometimes it's not even cost effective to pick-n-place circuit boards, rather have someone stuff a PCB by hand. The USA has probably moved to a lot of low-volume manufacturing. Could be that people are not buying as much stuff. I have read that credit card debt is back to 2008 levels. Maybe everyone is maxed out; not good.
- geggam 8y ago
  Where I live people are on facebook selling home cooked food and offering to do handy man work. Jobs I grew up with in fast food that could be relied on to provide 38 or so hrs a week now provide my teenagers 16-20. The teenagers are competing with adults for the jobs. I am of the opinion outside metropolis areas the entire nation is running on Empty as far as finance goes. The people I know who live here and work normal blue collar jobs run just under the max on their credit cards and keep alive by paying minimum. Opening new lines of credit to pay for emergencies Granted this is only my personal experiences but its common enough its disturbing. When people run out of money to buy things what happens ?
  - benploni 8y ago
    Something about your post made be literally anxious, like to the point of affecting my breathing.
    - Brockenstein 8y ago
      Worried about the next big recession and whether or not you'll end up in a shanty town.
  - erikpukinskis 8y ago

salawat 8y ago

No. Not as the root cause. I mean it's definitely a contributing factor, but only because of the axioms that drive our definition of a good company. The ideal company is seen as an investment vehicle that creates shareholder value. Full stop. We don't measure well performance of a company in terms of hiring people. Nor do we even measure it in how well they do what they are there to do. It always getsmeasured in value returned to investors. Investors are generally not laborers in this day and age. And any company that can make more long term by automating and dumping the need for laborers will. You can blame Automation all you want, but it's something far more basic that makes automation a worthy goal in the first place. And no, neither automation nor the incentives that make it desirable are "easy" problems to fix.

Finnucane 8y ago
>The ideal company is seen as an investment vehicle that creates shareholder value. Full stop. Even if that were true, that's not a factor that just suddenly started to have force in the last couple of decades.
- moosey 8y ago
  Massive tax cuts for the wealthy, particularly for capital gains and dividends, have reduced the resistance for money escaping corporations into the hands of investors, rather than the value add of paying employees more or improving working conditions and benefits. These changes started to occur in the 1980's, rather than the Dodge vs. Ford decision from 1919. While it is certainly true that nothing legally changed (edit: this isn't true: there is a full on attack on collective bargaining ongoing) in the more recent 30-40 years to allow for the massive accumulation of wealth and corporate practices that make the 'FULL STOP' in the original comment more accurate than before, the modification of tax law certainly has had a major impact.
- Brockenstein 8y ago
  It doesn't have to be new, but we have new and more ways to increase shareholder value. Like vastly improved methods of automation that have appeared in the past few decades.

0xdeadbeefbabe 8y ago

Nope, speculation caused it.

phkahler 8y ago

>> If, for example, this year’s model sells for the same dollar price as last year’s but has more features that consumers want, its true price has, in effect, dropped. So government agencies produce statistics showing that output, adjusted for price, has risen. This is a terrible way to view it. If we pack twice as much on a chip vs last year, did output go up? But price went down by about the same amount. Should we measure output in transistors? Printed die area? Dollar price of chips? Number of man-hours required to produce things? Labor cost? If Intel spent a bunch of cash on a new fab, made circuits with twice as many transistors, reduced prices on the old model and sold the new ones at the old price, then sold the same number as last year, did "output" really change? If yo u think it did, then I ask if it changed in any way that is meaningful to economic measures - it didn't change GDP, it didn't change cost or amount of labor, it didn't even change the companies profits.

pixl97 8y ago
Maybe take a larger view of it. Instead of just looking at Intel, look at the companies buying intel chips, are they getting more output with the same dollar input?
- phkahler 8y ago
  >> look at the companies buying intel chips, are they getting more output with the same dollar input? Probably not. The ones who were constrained by computing power are now doing higher quality work - FEA comes to mind as something that seems to produce more accurate results with increased computing capability. Games get "better" but their market doesn't get bigger - just more demanding. Hollywood does more CGI, but that doesn't actually increase their market. These are all things that don't affect GDP.

bsder 8y ago

Automation certainly isn't solely to blame. Our estimate of productivity is wrong. https://qz.com/1269172/the-epic-mistake-about-manufacturing-thats-cost-americans-millions-of-jobs/ https://qz.com/1269172/the-epic-mistake-about-manufacturing-...

ThomPete 8y ago

It's not just automation it's also digitalization which removes a lot of middle layers which used to be part of the cost. A much better to think about this is from the perspective of technology. The music industry is the best example of what happens. Basically, things that used to require all sorts of manufacturing and distribution layers are now more or less compressed into a single digital device.

milesvp 8y ago

It occurs to me as I think about this topic more and more (sadly article is paywalled so I can't speak to it directly), manufacturing is to blane for lost manufacturing jobs. Economies are self balancing, and since manufacturing makes goods cheaper, eventually people will do other things that are still expensive. This is prediminantly why people talk about transformation to service economies. Manufacturing is inherently about mass producing, which means it's always been about scale, which means doing more without adding proportionately to labor. Automation is just one of the many things that help scale, it's just one of the many things that requires capital investment. The bigger problem for manufacturing jobs has been globalization. When you can take your capital to a market with cheaper workers, then labor becomes a scaling tool (ironically?) in that you can throw labor at a problem to get your labor costs down. What's interesting about this, though, is that at some point you can't do this anymore. Textiles, which is considered level 0 for manufacturing, has nowhere left to go for cheaper labor than Bangladesh (without significant political instability issues).

clarkmoody 8y ago

Is automation to blame for lost agricultural jobs? Lost textile mill jobs? Yes, it is. Do we miss those jobs? No, we don't, because we're so much wealthier when we can replace human labor with machine labor. That frees up humans to do higher-order tasks, improving our standard of living. Yes, there are winners and losers with any change, but long-term we are far better off than we could have imagined at the dawn of the Industrial Revolution.

sp527 8y ago
> Yes, there are winners and losers with any change, but long-term we are far better off than we could have imagined at the dawn of the Industrial Revolution. This is a deeply ignorant statement that renders invisible over 2 billion completely disenfranchised people in the third world. Just because you don't experience the consequences - while you gorge on the bounty of human progress somewhere amongst the top decile of our species - doesn't mean they aren't there.
- clarkmoody 8y ago
  I recommend that you go check out some of the data here: https://humanprogress.org/dws https://humanprogress.org/dws In the past 150 years, the human race has made extraordinary progress, lifting billions out of poverty. There are still many globally who are left out, and I am aware of that. But you can't argue against the widespread gains to humanity through improved technology.
Aloha 8y ago
I'd argue quite strongly that the communities who lost those jobs, do miss them dearly. They have not been replaced, and while you're correct we've lifted billions out of poverty, we've thrown hundreds of millions back into it, with no easy recourse out. Western nations, in general do a piss poor job with people in poverty, as we think ourselves beyond the problems of the third world, at least in the united states, its clearly not true - and dont have the right systems in place to allow folks to either fully participate in the low end of the market, or give them meaningful government assistance. Business right now is frightfully capital intensive - and has a high barrier to entry (paperwork mostly), and in many places there are no jobs available - what is someone to do?

bongo662 8y ago

It'd be interesting to see a breakdown of what 'forms' of manufacturing are leaving the nation. Its pretty obvious that as trade agreements rise up and its more cost effective to move manufacturing to other countries with lower payed workers that the company would do that but my guess would be that higher(?) quality manufacturing still goes on stateside rather than in China or Mexico for fear of IP theft. So for things like high grade medical devices - MRI, CT scanners would still be manufactured here because of the knowledge and technology necessary to construct it. I also think theirs a good Stephen Hawking quote I like to think of when people discuss how in a post scarcity world with full automation life will be a utopia and its something along the lines of "If machines produce everything we need, the outcome will depend on how things are distributed,"

Animats 8y ago

Actual paper from that author.[1] From 2016. Her real point is that the electronics industry is doing great, while some other parts of of US manufacturing are not. Fig. 2 is useful. Textiles, apparel, furniture, and paper products are the big losers. The apparel industry was totally crushed by imports from low-wage areas. That took out the US textile industry, since exporting textiles to be made into apparel and re-imported was pointless. US motor vehicle output is up, not down. That hasn't been crushed by imports. Primary metals (which is mostly steel and aluminum) are up. More breakdown of the computers and electronics sector would help. We'd probably see semiconductors through the roof and consumer electronics down. That data is available. This is useful. Trade is an industry-specific problem. The American apparel industry is mostly gone, but we have cheap clothing. The metals industry complains but isn't in real trouble. Motor vehicles are doing OK. A breakdown of consumer goods by category would help. That would show where to apply tariffs. Paper products are down because newspapers are obsolete, and they used vast amounts of paper, and because offices use far less paper. Not import-related. [1] https://upjohn.org/sites/default/files/pdf/state-of-american-mfg.pdf https://upjohn.org/sites/default/files/pdf/state-of-american...

philipkglass 8y ago
Another thing to consider, when pondering counterfactuals to free trade: offshoring destinations are also automating. China is now the largest market for industrial robots: https://www.cnbc.com/2017/10/26/china-is-the-largest-market-for-robots-in-the-world-right-now-abb-ceo.html https://www.cnbc.com/2017/10/26/china-is-the-largest-market-... Maybe when the local factory shut down in years past it transferred 500 jobs straight to China. That doesn't necessarily mean that it kept employing 500 people for the same output in the years since. Re-shoring the same volume of production may mean significantly fewer jobs today, in e.g. textiles: "U.S. Textile Plants Return, With Floors Largely Empty of People", 2013 https://www.nytimes.com/2013/09/20/business/us-textile-factories-return.html https://www.nytimes.com/2013/09/20/business/us-textile-facto... ... Take Parkdale: The mill here produces 2.5 million pounds of yarn a week with about 140 workers. In 1980, that production level would have required more than 2,000 people. ... Or in the case of fields that still are very little automated, like garment sewing, it would require pretty much the same number of workers now as 20 years ago. As you say, it's very product-specific. If we broaden the category a bit from "manufacturing jobs" to "industrial jobs" (including things like coal miners and workers at coal fired power plants), you also see strong effects from process-and-input changes. It takes far fewer people to produce a terawatt hour of electricity from a modern combined-cycle natural gas generating plant than from a modern (worse, built-in-1970) coal plant. It also takes fewer workers for a TWh from solar PV or wind power. The rapid growth of jobs in renewable energy is a temporary excursion, mostly an artifact of current rapid expansion. At steady-state it will take fewer people to supply electricity from a renewable-heavy electricity system than a coal-heavy system. That's one reason that I am bullish on the long term price-competitiveness of renewable electricity but skeptical that "green jobs" are a panacea for declining industrial employment in other sectors.