Contributor Post Created with Sketch. Is Automation Really the Worst Enemy of the Middle Class?

 

This Axios headline is problematic: “Summers: Automation is the middle class’ worst enemy.”

The accompanying piece doesn’t actually quote economist Larry Summers making that declaration. Rather it summarizes an interview in which Summers indeed points out the challenge automation poses for workers. He’s right.​ Of course that’s been the case for the past 200 years and will likely be the case for the next 200. But in exchange for a degree of instability and disruption, technological progress has dramatically raised living standards for workers.

Automation is kind of​ like alcohol, which, ​as Homer Simpson puts it, is “t​he cause of, and solution to, all of life’s problems.” It’s the job of policymakers to make sure workers are read​y​ to climb to the next foothold or ledge as the waters of automation continue to rise. It’s also their job to make sure policy is as supportive as possible of innovation. Indeed, we need more tech progress, not less. “The U.S. economy currently suffers not from too much automation, but rather from too little investment in the sort of technology that would raise the country’s lackluster productivity,” writes​ Derek Thompson ​in an excellent new piece ​at ​t​he Atlantic.

Technology will erase jobs but also create them. Unfortunately, as Kevin Kelly writes, we can’t see those jobs from here “because we can’t yet see the machines and technologies that will make them.”

The piece also includes this chart, which shows lower US labor force participation than other advanced economies:

But I doubt whether Summers blames automation vs. the lack of US policies that center-left economists see as supporting workers, such as paid leave, and high US incarceration rates. And here is economist David Autor on the net impact of automation on jobs this century:

A final observation is that while much contemporary economic pessimism attributes the labor market woes of the past decade to the adverse impacts of computerization, I remain skeptical of this inference. Clearly, computerization has shaped the structure of occupational change and the evolution of skill demands. But it is harder to see the channel through which computerization could have dramatically reduced labor demand after 1999. … My suspicion is that the deceleration of the U.S. labor market after 2000, and further after 2007, is more closely associated with two other macroeconomic events. A first is the bursting of the “dot-com” bubble, followed by the collapse of the housing market and the ensuing financial crisis, both of which curtailed investment and innovative activity. A second is the employment dislocations in the U.S. labor market brought about by rapid globalization, particularly the sharp rise of import penetration from China following its accession to the World Trade Organization in 2001.

There are 14 comments.

Become a member to join the conversation. Or sign in if you're already a member.
  1. Doug Kimball Thatcher

    Ex-Harvard President should probably let everyone know that power tools are also the enemy of the middle class. Imagine construction without the power saw or drill? Imagine trenching without a backhoe, or even a shovel? In fact, all labor saving inventions back as far as the wheel or the lever, are the enemies of labor. In fact, if we eliminated fire, wouldn’t that encourage veganism? We would all be healthier and happier in our teepees and caves, if perhaps a little cold in the winter.

    These people are daft. Asking people to be less efficient, less inventive and less clever is like trying to tell teenagers to act responsibly. As soon as you turn around, they’re doing something crazy.

    You’d think Summers would have picked up a little common sense at some point, but no…

    And even if we are colder in our teepees and caves, there is Climate Change; so we have that going for us; which is nice…

    BTW Autors has it exactly right, except, I would point out, that both the Dot Com bubble and the Housing/Financial meltdowns have their genesis in Federal policy; the relaxation of the option exercise as a “neutral” transaction with respect to insider trading rules, the failure to recognize the risks inherent in credit default swaps and in the mortgage backed public bonds, the use of the Fed to create cheap money, hence leverage, the semi-privatization of Fannie and Freddy (stuffed with political hacks) and the egregious and arrogant, politically motivated dictatorial influence over mortgage underwriting.

    • #1
    • June 7, 2017, at 12:57 PM PDT
    • 7 likes
  2. Steve C. Member

    No, the Federal Reserve.

    • #2
    • June 7, 2017, at 1:33 PM PDT
    • 1 like
  3. DocJay Inactive

    Steve C. (View Comment):
    No, the Federal Reserve.

    Bingo!

    • #3
    • June 7, 2017, at 1:54 PM PDT
    • 1 like
  4. David Foster Member
    David Foster Joined in the first year of Ricochet Ricochet Charter Member

    The argument that “offshoring of jobs doesn’t matter anyhow because robots” is often being made these days, especially by Progs wanting to bash Trump, but also by tech-industry people who are dazzled by the potential of their own techchildren.

    In actuality, of course, *both* offshoring and automation have an impact on jobs. The impact of these factors is *less than additive*, because a more automated US factory (or other facility) will be more cost-competitive than a less-automated one.

    • #4
    • June 7, 2017, at 2:04 PM PDT
    • 2 likes
  5. David Foster Member
    David Foster Joined in the first year of Ricochet Ricochet Charter Member

    The David Autor piece looks very interesting, much better than most of the stuff currently being written on this topic.

    He cites statistics that “The share of information processing equipment and software in private, nonresidential investment rose from approximately 8 percent to more than 30 percent between 1950 and 2012, with the largest leap occurring between 1990 and 2000″….seems to me that this almost certainly understates the case. If a factory buys a batch of CNC machine tools, I doubt that the value of the information processing components embedded in this systems is being pulled out and counted in the info processing investment statistics…indeed, this would be impossible to do without detailed analysis of the CNC equipment manufacturer’s books. Similarly for elevators, air conditioning systems, aircraft autopilot and navigation systems, etc.

    • #5
    • June 7, 2017, at 2:13 PM PDT
    • 1 like
  6. John Walker Contributor

    James Pethokoukis:It’s the job of policymakers to make sure workers are read​y​ to climb to the next foothold or ledge as the waters of automation continue to rise.

    Technology will erase jobs but also create them.

    This is self evident, and it has always been the case, but the question is “Jobs for whom?”

    From the dawn of modern humans until the present day, human intelligence has been reasonably closely approximated by a normal distribution where IQ 100 is defined as the mean and 15 is the standard deviation. Most of the population falls within four standard deviations of the mean: IQ 70–130—less than 5% of the population is outside these extrema.

    Now, prior to the mechanisation of agriculture and, later, the first wave of automation in manufacturing, there were plenty of jobs for anybody on this cognitive spectrum; many people on the right side of the curve might not have had much chance to employ their intellect unless they were lucky enough to be born into circumstances which allowed them to be educated, but almost everybody was productively employed and able to raise and support a family.

    What we have seen since the early 20th century is a ratchet which has steadily been supplanting jobs previously performed by those on the low end of the intelligence spectrum with machines which are built, maintained, and operated by those on the higher end. Government social programmes have often accelerated this process by making it more expensive to employ humans and consequently making them less competitive against automated replacements.

    My estimate is that the employment IQ threshold in the developed world is presently crossing the IQ 100 level which means that it is difficult for around half the population to find productive work they’re capable of performing. This is something which has never before happened in human history, so we’re currently in uncharted terrain. This process will accelerate inexorably during the Roaring Twenties, where many jobs now considered occupied by the cognitive élite, such as routine legal and medical practice, will be automated.

    When trucks drive themselves, mass market restaurants are automated, retailing moves online with automated home delivery, and construction is increasingly based upon prefabricated modules made in automated factories, what will those on the left of the bell curve do for a living?

    I honestly don’t know. But if I had the ears of policy makers, while urging them to ponder the consequences of this, I’d also counsel spending a small amount of money on a wild card: how might we shift the mean of the bell curve one standard deviation to the right: to a mean IQ of 115? This would completely change the game and redefine the future, and almost nobody is working on it or even thinking about it.

    • #6
    • June 7, 2017, at 3:47 PM PDT
    • 6 likes
  7. Valiuth Member
    Valiuth Joined in the first year of Ricochet Ricochet Charter Member

    John Walker (View Comment):

    I honestly don’t know. But if I had the ears of policy makers, while urging them to ponder the consequences of this, I’d also counsel spending a small amount of money on a wild card: how might we shift the mean of the bell curve one standard deviation to the right: to a mean IQ of 115? This would completely change the game and redefine the future, and almost nobody is working on it or even thinking about it.

    So if I understand the IQ measure it always places the mean at 100. So how can you possibly shift it if you have no set standard of measure. Also what if your IQ test doesn’t actually measure the cognitive abilities that are most profitable in a new economy? But, putting aside issues of measurement.

    Do you have any guess how you might increase general IQ? Education? That seems like what government policy makers would gravitate to, but we have had public education in the West for over a century now. The only thing left is mandatory collage education. Which is where I think we will head to for better or worse.

    • #7
    • June 7, 2017, at 4:26 PM PDT
    • Like
  8. John Walker Contributor

    Valiuth (View Comment):
    So if I understand the IQ measure it always places the mean at 100. So how can you possibly shift it if you have no set standard of measure. Also what if your IQ test doesn’t actually measure the cognitive abilities that are most profitable in a new economy? But, putting aside issues of measurement.

    This is easily resolved. You simply test using original tests which have not been re-normed. Raven’s Progressive Matrices, for example, is a culturally-neutral test which measures the kind of basic cognition which humans would have evolved in their early history.

    Do you have any guess how you might increase general IQ? Education? That seems like what government policy makers would gravitate to, but we have had public education in the West for over a century now. The only thing left is mandatory collage education. Which is where I think we will head to for better or worse.

    I’m thinking nootropic drugs, infant nutrition, or genetic interventions. It seems like between 60–80% of IQ is baked in before you first test it, and subsequent interventions have little effect. Epochal change requires thinking outside the box. However, we’re facing epochal change whether we like it or not: the question is whether it will be for the better or worse.

    • #8
    • June 7, 2017, at 4:35 PM PDT
    • 1 like
  9. Locke On Member

    John Walker (View Comment):
    My estimate is that the employment IQ threshold in the developed world is presently crossing the IQ 100 level which means that it is difficult for around half the population to find productive work they’re capable of performing. This is something which has never before happened in human history, so we’re currently in uncharted terrain. This process will accelerate inexorably during the Roaring Twenties, where many jobs now considered occupied by the cognitive élite, such as routine legal and medical practice, will be automated.

    When trucks drive themselves, mass market restaurants are automated, retailing moves online with automated home delivery, and construction is increasingly based upon prefabricated modules made in automated factories, what will those on the left of the bell curve do for a living?

    This. 100 times this. It’s been staring me in the face for a lifetime in high tech, and there is still no answer.

    It’s not that there are overblown claims made for AI advances, or that leading edge automation is flaky, thus has it always been. It’s that the trailing edge – what’s easily accessible to any buyer – advances steadily, and has the effect which John highlights of successively making the less intelligent non-competitive for all but the meanest grunt labor – which is then denied them when regulation pushes the fixed costs of employment above their value.

    • #9
    • June 7, 2017, at 4:49 PM PDT
    • 4 likes
  10. Doug Kimball Thatcher

    John Walker (View Comment):

    James Pethokoukis:It’s the job of policymakers to make sure workers are read​y​ to climb to the next foothold or ledge as the waters of automation continue to rise.

    Technology will erase jobs but also create them.

    This is self evident, and it has always been the case, but the question is “Jobs for whom?”

    From the dawn of modern humans until the present day, human intelligence has been reasonably closely approximated by a normal distribution where IQ 100 is defined as the mean and 15 is the standard deviation. Most of the population falls within four standard deviations of the mean: IQ 70–130—less than 5% of the population is outside these extrema.

    Now, prior to the mechanisation of agriculture and, later, the first wave of automation in manufacturing, there were plenty of jobs for anybody on this cognitive spectrum; many people on the right side of the curve might not have had much chance to employ their intellect unless they were lucky enough to be born into circumstances which allowed them to be educated, but almost everybody was productively employed and able to raise and support a family.

    What we have seen since the early 20th century is a ratchet which has steadily been supplanting jobs previously performed by those on the low end of the intelligence spectrum with machines which are built, maintained, and operated by those on the higher end. Government social programmes have often accelerated this process by making it more expensive to employ humans and consequently making them less competitive against automated replacements.

    My estimate is that the employment IQ threshold in the developed world is presently crossing the IQ 100 level which means that it is difficult for around half the population to find productive work they’re capable of performing. This is something which has never before happened in human history, so we’re currently in uncharted terrain. This process will accelerate inexorably during the Roaring Twenties, where many jobs now considered occupied by the cognitive élite, such as routine legal and medical practice, will be automated.

    When trucks drive themselves, mass market restaurants are automated, retailing moves online with automated home delivery, and construction is increasingly based upon prefabricated modules made in automated factories, what will those on the left of the bell curve do for a living?

    I honestly don’t know. But if I had the ears of policy makers, while urging them to ponder the consequences of this, I’d also counsel spending a small amount of money on a wild card: how might we shift the mean of the bell curve one standard deviation to the right: to a mean IQ of 115? This would completely change the game and redefine the future, and almost nobody is working on it or even thinking about it.

    Is this the world of Vonnegut’s “Player Piano” and Orwell’s “1984”, arriving a bit late, where the world is segregated into the elite, the warriors and the essentially useless workers? Is Socialism merely the worker’s recognition that they have power in numbers, either exercised in Democracy, or where that fails, in revolution?

    There is still much work to be done in the trades, where skilled workers are much in demand across all types.

    No, what we need to do is to stop pretending that people have to continually reinvent themselves. A welder, a mechanic, a skilled carpenter, electrician, plumber, HVAC technician and any number of tradesmen have skills that are always in demand, that are never obsolete. Aside from technical training where they learn the basics of their trade, these people need small business training so that they can leverage their knowledge and train others to work when they inevitably wear down and are unable to perform eight hours of daily physical work.

    The problem is this. We’ve depreciated these skills and failed to encourage folks to pursue trade skills and small business opportunities in favor of government student debt. The search for promising work in the turbulent white collar job market is capricious, where skills, like technology, shift, move and become obsolete at an alarming rate, where everyone becomes a dinosaur in just a couple of years. Today’s coder is tomorrow’s barista. Bottom line, life would be far better for that guy slinging coffee had he become a diesel mechanic instead of getting a six year BA in comparative literature.

    • #10
    • June 7, 2017, at 4:49 PM PDT
    • 3 likes
  11. Weeping Member

    John Walker (View Comment):
    Now, prior to the mechanisation of agriculture and, later, the first wave of automation in manufacturing, there were plenty of jobs for anybody on this cognitive spectrum; many people on the right side of the curve might not have had much chance to employ their intellect unless they were lucky enough to be born into circumstances which allowed them to be educated, but almost everybody was productively employed and able to raise and support a family.

    What we have seen since the early 20th century is a ratchet which has steadily been supplanting jobs previously performed by those on the low end of the intelligence spectrum with machines which are built, maintained, and operated by those on the higher end. Government social programmes have often accelerated this process by making it more expensive to employ humans and consequently making them less competitive against automated replacements.

    My estimate is that the employment IQ threshold in the developed world is presently crossing the IQ 100 level which means that it is difficult for around half the population to find productive work they’re capable of performing. This is something which has never before happened in human history, so we’re currently in uncharted terrain. This process will accelerate inexorably during the Roaring Twenties, where many jobs now considered occupied by the cognitive élite, such as routine legal and medical practice, will be automated.

    When trucks drive themselves, mass market restaurants are automated, retailing moves online with automated home delivery, and construction is increasingly based upon prefabricated modules made in automated factories, what will those on the left of the bell curve do for a living?

    This is my major concern with the adoption of automation in every nook and cranny of the economy. Sure there will still be jobs, but will those who have been displaced have the intellect and/or temperament to perform them? Some will, sure. But all? No. So that leaves us with the question: How many will be able to make the change? And what happens to those who can’t?

    • #11
    • June 7, 2017, at 9:32 PM PDT
    • 2 likes
  12. RktSci Member

    There is a long piece on this topic at Reason by Ronald Bailey.

    In looking at some of the data he presented on how jobs are created amidst all the automation that we have seen, he cites the growth in paralegals from 85,000 in 1990 to 280,000. However, if you read Instapundit, you are quite aware that the job market for attorneys is shrinking. Have better paperwork systems that PCs allow pushed jobs down the job ladder? We have seen this before in medicine. In the 80s, in response to pressure from insurance companies to control costs, hospitals started pushing jobs down the hierarchy. Some physician tasks went to RNs, RN tasks went to LVNs, LVN tasks to techs. (This did flatline cost increases for a while, btw.) Now, many tasks that took a college degree to do are given to high school graduates with a few weeks training.

    The Atlantic piece cited above seems skeptical that we are on the edge of a new industrial revolution. I say we are. The reason is the difference in the automation that we are seeing. Past automation has mostly been mechanical – from tractors to mechanical looms to car assembly. We are now seeing a potential tidal wave of cognitive automation. The beginnings were in the 50s and 60s, when, as in the movie Hidden Figures, “computer” was a job title, not a device. A friend working at a medical informatics startup that has a system to monitor electrocardiograms in real time and predict upcoming heart attacks for certain ICU patients. Alexa and Siri can understand speech. Soon we will have real-time translation systems. (The three-letter agencies probably do already!)

    In the 1980s, I worked on AI research and we thought rule-based systems were the future – break a job down to the small cognitive tasks and have them programmed in as rules. Now, you throw data at a neural network and it learns the task. We’re seeing this in the automated driving systems. This alone would put most truck drivers and cab drivers out of work in the few years of the transition, leaving millions without jobs. I’m don’t think that it will be as fast as proponents say – I’ve seen 5 years from now predicted – as the legal issues dwarf technical ones, and getting the laws in place to allow driverless cars and settle where legal liability for accidents falls is going to take a while.

    So, are we headed to a world like Vonnegut’s Player Piano, of Managers, Engineers, and the unemployed proles? There might be new jobs spring up. Or the people on the left side of the cognitive bell curve may be stuck on the dole or frantically scrambling for one of the few jobs left they can do.

    Oh, and all this concentration on automation is ignoring the biotech revolution going on now. The 2oth century was the physics century, the 21st will be the biology century.

    • #12
    • June 8, 2017, at 8:35 AM PDT
    • 3 likes
  13. MJBubba Inactive

    John Walker has it right with his comment # 6.

    I live near Memphis, which is a transportation hub. As recently as the 1950s an illiterate man could feed a family, not well, but good enough by the standards of the “working poor,” just by working as a dockhand totin’ crates and barrels.

    All those jobs became jobs for forklift drivers by the end of the 1970s, and required an ability to read, at least enough to read labels and find numbered bins in a big warehouse array of shelving.

    By the end of the 1990s the forklift driver was carrying a blackberry with the manifest that listed the loads by row and shelf numbers. All the loads were palletized.

    Now the computer is built in to the forklift. The palletized loads are now shrink-wrapped and bar coded. The forklift operator has a wireless barcode reader wand and must enter every load movement into the computer. In some warehouses, the RFI reader mounted on the forklift does that for him automatically.

    By the end of the 2020s, the forklift will be a robot.

    There will be two million square foot warehouses with only four employees per shift. One tech guy to keep the computer system going, one mechanic for the robot forklifts and Roombas and conveyer belt systems, and two security guards. The phones are already answered in Bangalore, because customers don’t like talking to robots.

    Memphis is already in a jobs crisis. This is going to get worse.

    In the meantime, the productivity of the system is incredible. Depending on what it is that you want, if you place your order with Williams Sonoma, for example, by 7:00 pm Central time, they can find your Fathers Day gift order and get it moved two miles to the FedEx Hub at the midnight drop, and it can be delivered tomorrow morning, wherever you want it to go in the world.

    And our jail is full of illiterate men.

    • #13
    • June 8, 2017, at 9:59 AM PDT
    • 2 likes
  14. JimGoneWild Coolidge

    James Pethokoukis: It’s the job of policymakers to make sure workers are read​y​ to climb to the next foothold or ledge as the waters of automation continue to rise.

    No it’s not! This type of thinking is part of the problem. We need policymakers to leave people alone, stop write rules and regulations and get off the backs of business.

    • #14
    • June 8, 2017, at 1:33 PM PDT
    • Like

Comments are closed because this post is more than six months old. Please write a new post if you would like to continue this conversation.

Join the Conversation Created with Sketch.
Have Something to Say?

Join the Conversation